Fossil SCM

Sync with trunk.

florian 2024-12-12 16:58 timeline-keyboard-navigation merge

Commit 611f20e81708ee77c0d6a7b6103cb0723aa51731e8a6ec154fd4e0d45e6f74cd

Parent 00f7466addc2521…

157 files changed +1 -1 +4 -1 +1 -1 +7 +5 -3 +95 -28 +722 -241 +3 -3 +1 -1 +3370 -1991 +2361 -1405 +80 -13 +22 -17 +4 -1 +7 -6 +1 -1 +23 -6 +8 -1 +12 -4 +49 -1 +4 -3 -17 +3 -1 +9 -3 +114 -14 +221 -71 +28 -9 +13 -14 +52 -10 +52 -10 +50 -48 +146 -7 +1 -1 +8 -8 +26 -1 +2 +1 -1 +99 -9 +13 -5 +63 -35 +25 -5 +3 -1 +1 -1 +15 -3 +54 -6 +2 -2 +6 +4 -6 +4 -1 +152 -17 +2 -1 +16 -4 +13 -10 +2 -1 +7 -3 +465 -6 +581 +804 -168 +6 -2 +1 +12 -2 +9 -3 +4 -5 +7 -7 +83 -64 +16 -4 +7 -5 +5 -40 +9 +2 -1 +2 -1 +1 -1 +55 -3 +4 -4 +229 +40 -1 +20 +20 -20 +6 +2 +1 +2 -1 +1 -1 +1 +2 +4 -4 +3 -3 +2 -2 +2 -2 +3 -3 +2 -2 +6 -6 +2 -2 +1 -1 +1 -1 +2 -2 +26 -4 +2 -2 +1 -1 +4 -4 +1 -1 +10 -8 +13 -5 +3 -3 +1 -1 +3 -3 +6 -6 +1 -1 +1 -1 +9 -10 +1 -1 +3 -3 +2 -2 +1 -1 +1 -1 +1 -1 +1 -1 +4 -4 +4 -4 +2 -2 +5 -5 +1 -1 +3 -3 +1 -1 +1 -1 +2 -2 +2 -2 +8 -8 +7 -7 +10 -9 +1 -1 +2 -1 +6 -2 +4 -4 +2 -2 +1 -1 +1 -1 +2 -2 +1 -1 +1 -1 +1 -1 +1 -1 +1 -1 +2 -3 +2 -2 +2 -2 +6 -5 +33 -13 +1 -1 +5 -5 +2 -2 +2 -2 +2 -2 +1 -1 +5 -5 +7 -4 +1 -1

~ BUILD.txt ~ Makefile.in ~ VERSION ~ auto.def ~ autosetup/autosetup ~ autosetup/autosetup-config.guess ~ autosetup/autosetup-config.sub ~ autosetup/autosetup-find-tclsh ~ autosetup/cc.tcl ~ extsrc/shell.c ~ extsrc/sqlite3.c ~ extsrc/sqlite3.h ~ fossil.1 ~ skins/darkmode/css.txt ~ skins/default/css.txt ~ src/add.c ~ src/allrepo.c ~ src/attach.c ~ src/bisect.c ~ src/blob.c ~ src/branch.c ~ src/browse.c ~ src/builtin.c ~ src/chat.c ~ src/checkin.c ~ src/comformat.c ~ src/configure.c ~ src/db.c ~ src/default.css ~ src/default.css ~ src/delta.c ~ src/diff.c ~ src/diff.tcl ~ src/diffcmd.c ~ src/dispatch.c ~ src/doc.c ~ src/event.c ~ src/file.c ~ src/finfo.c ~ src/forum.c ~ src/fossil.diff.js ~ src/fossil.dom.js ~ src/fossil.page.chat.js ~ src/fossil.page.pikchrshowasm.js ~ src/fossil.pikchr.js ~ src/glob.c ~ src/http.c ~ src/http_ssl.c ~ src/http_transport.c ~ src/info.c ~ src/interwiki.c ~ src/login.c ~ src/main.c ~ src/main.mk ~ src/manifest.c ~ src/merge.c ~ src/merge.tcl ~ src/merge3.c ~ src/patch.c ~ src/path.c ~ src/pikchrshow.c ~ src/rebuild.c ~ src/rss.c ~ src/schema.c ~ src/search.c ~ src/setup.c ~ src/setupuser.c ~ src/sitemap.c ~ src/th_tcl.c ~ src/timeline.c ~ src/timeline.c ~ src/tkt.c ~ src/update.c ~ src/wiki.c ~ src/winfile.c ~ src/xfer.c ~ test/comment.test ~ test/glob.test ~ test/settings.test ~ test/tester.tcl ~ test/utf8-comment.txt ~ tools/makemake.tcl ~ tools/man_page_command_list.tcl ~ win/Makefile.mingw ~ win/Makefile.msc ~ www/aboutcgi.wiki ~ www/adding_code.wiki ~ www/alerts.md ~ www/antibot.wiki ~ www/backoffice.md ~ www/blame.wiki ~ www/bugtheory.wiki ~ www/build.wiki ~ www/caps/index.md ~ www/caps/ref.html ~ www/cgi.wiki ~ www/changes.wiki ~ www/chat.md ~ www/checkin_names.wiki ~ www/concepts.wiki ~ www/contribute.wiki ~ www/custom_ticket.wiki ~ www/embeddeddoc.wiki ~ www/env-opts.md ~ www/event.wiki ~ www/fileedit-page.md ~ www/fileformat.wiki ~ www/forum.wiki ~ www/fossil-is-not-relational.md ~ www/fossil-v-git.wiki ~ www/fossil_prompt.wiki ~ www/gitusers.md ~ www/glossary.md ~ www/grep.md ~ www/hashpolicy.wiki ~ www/hints.wiki ~ www/history.md ~ www/hooks.md ~ www/index.wiki ~ www/inout.wiki ~ www/interwiki.md ~ www/json-api/api-artifact.md ~ www/json-api/api-auth.md ~ www/json-api/api-checkout.md ~ www/json-api/api-finfo.md ~ www/json-api/api-query.md ~ www/json-api/api-tag.md ~ www/json-api/conventions.md ~ www/json-api/hacking.md ~ www/makefile.wiki ~ www/mdtest/test1.md ~ www/newrepo.wiki ~ www/password.wiki ~ www/patchcmd.md ~ www/pikchr.md ~ www/qandc.wiki ~ www/quickstart.wiki ~ www/rebaseharm.md ~ www/selfcheck.wiki ~ www/selfhost.wiki ~ www/server/debian/service.md ~ www/server/openbsd/fastcgi.md ~ www/server/windows/service.md ~ www/settings.wiki ~ www/shunning.wiki ~ www/stats.wiki ~ www/style.wiki ~ www/sync.wiki ~ www/tech_overview.wiki ~ www/th1.md ~ www/theory1.wiki ~ www/tickets.wiki ~ www/unvers.wiki ~ www/userlinks.wiki ~ www/webui.wiki ~ www/whyallinone.md ~ www/wikitheory.wiki

M BUILD.txt

+1 -1

		--- BUILD.txt
		+++ BUILD.txt
		@@ -42,11 +42,11 @@
42	42	mkdir build
43	43	cd build
44	44	../configure
45	45	make
46	46
47		-This will now keep all generates files separate from the maintained
	47	+This will now keep all generated files separate from the maintained
48	48	source code.
49	49
50	50	--------------------------------------------------------------------------
51	51
52	52	Here are some notes on what is happening behind the scenes:
53	53

	--- BUILD.txt
	+++ BUILD.txt
	@@ -42,11 +42,11 @@
42	mkdir build
43	cd build
44	../configure
45	make
46
47	This will now keep all generates files separate from the maintained
48	source code.
49
50	--------------------------------------------------------------------------
51
52	Here are some notes on what is happening behind the scenes:
53

	--- BUILD.txt
	+++ BUILD.txt
	@@ -42,11 +42,11 @@
42	mkdir build
43	cd build
44	../configure
45	make
46
47	This will now keep all generated files separate from the maintained
48	source code.
49
50	--------------------------------------------------------------------------
51
52	Here are some notes on what is happening behind the scenes:
53

M Makefile.in

+4 -1

		--- Makefile.in
		+++ Makefile.in
		@@ -51,11 +51,11 @@
51	51	BCCFLAGS = @CPPFLAGS@ $(CFLAGS)
52	52	TCCFLAGS = @EXTRA_CFLAGS@ @CPPFLAGS@ $(CFLAGS) -DHAVE_AUTOCONFIG_H
53	53	#
54	54	# Fuzzing may be enable by appending -fsanitize=fuzzer -DFOSSIL_FUZZ
55	55	# to the TCCFLAGS variable.
56		-# For more thorouth (but also slower) investigation
	56	+# For more thorough (but also slower) investigation
57	57	# -fsanitize=fuzzer,undefined,address
58	58	# might be more useful.
59	59
60	60	INSTALLDIR = $(DESTDIR)@prefix@/bin
61	61	USE_SYSTEM_SQLITE = @USE_SYSTEM_SQLITE@
		@@ -92,10 +92,13 @@
92	92	MAKE_COMPILATION_DB = @MAKE_COMPILATION_DB@
93	93
94	94	.PHONY: all tags
95	95
96	96	include $(SRCDIR)/main.mk
	97	+SQLITE_OPTIONS += @SQLITE_OPTIONS_EXT@
	98	+SHELL_OPTIONS += @SQLITE_OPTIONS_EXT@
	99	+# ^^^ must come after main.mk is included
97	100
98	101	distclean: clean
99	102	-rm -f autoconfig.h config.log Makefile
100	103	-rm -f cscope.out tags
101	104
102	105

	--- Makefile.in
	+++ Makefile.in
	@@ -51,11 +51,11 @@
51	BCCFLAGS = @CPPFLAGS@ $(CFLAGS)
52	TCCFLAGS = @EXTRA_CFLAGS@ @CPPFLAGS@ $(CFLAGS) -DHAVE_AUTOCONFIG_H
53	#
54	# Fuzzing may be enable by appending -fsanitize=fuzzer -DFOSSIL_FUZZ
55	# to the TCCFLAGS variable.
56	# For more thorouth (but also slower) investigation
57	# -fsanitize=fuzzer,undefined,address
58	# might be more useful.
59
60	INSTALLDIR = $(DESTDIR)@prefix@/bin
61	USE_SYSTEM_SQLITE = @USE_SYSTEM_SQLITE@
	@@ -92,10 +92,13 @@
92	MAKE_COMPILATION_DB = @MAKE_COMPILATION_DB@
93
94	.PHONY: all tags
95
96	include $(SRCDIR)/main.mk



97
98	distclean: clean
99	-rm -f autoconfig.h config.log Makefile
100	-rm -f cscope.out tags
101
102

	--- Makefile.in
	+++ Makefile.in
	@@ -51,11 +51,11 @@
51	BCCFLAGS = @CPPFLAGS@ $(CFLAGS)
52	TCCFLAGS = @EXTRA_CFLAGS@ @CPPFLAGS@ $(CFLAGS) -DHAVE_AUTOCONFIG_H
53	#
54	# Fuzzing may be enable by appending -fsanitize=fuzzer -DFOSSIL_FUZZ
55	# to the TCCFLAGS variable.
56	# For more thorough (but also slower) investigation
57	# -fsanitize=fuzzer,undefined,address
58	# might be more useful.
59
60	INSTALLDIR = $(DESTDIR)@prefix@/bin
61	USE_SYSTEM_SQLITE = @USE_SYSTEM_SQLITE@
	@@ -92,10 +92,13 @@
92	MAKE_COMPILATION_DB = @MAKE_COMPILATION_DB@
93
94	.PHONY: all tags
95
96	include $(SRCDIR)/main.mk
97	SQLITE_OPTIONS += @SQLITE_OPTIONS_EXT@
98	SHELL_OPTIONS += @SQLITE_OPTIONS_EXT@
99	# ^^^ must come after main.mk is included
100
101	distclean: clean
102	-rm -f autoconfig.h config.log Makefile
103	-rm -f cscope.out tags
104
105

M VERSION

+1 -1

		--- VERSION
		+++ VERSION
		@@ -1,1 +1,1 @@
1		-2.25
	1	+2.26
2	2

	--- VERSION
	+++ VERSION
	@@ -1,1 +1,1 @@
1	2.25
2

	--- VERSION
	+++ VERSION
	@@ -1,1 +1,1 @@
1	2.26
2

M auto.def

		--- auto.def
		+++ auto.def
		@@ -104,18 +104,21 @@
104	104	msg-result "WARNING: tclsh $v found; need >= 8.6 for 'make test'."
105	105	define TCLSH false ;# force "make test" failure via /usr/bin/false
106	106	}
107	107	}
108	108
	109	+define CFLAGS [get-env CFLAGS "-g -O2"]
109	110	define EXTRA_CFLAGS "-Wall"
110	111	define EXTRA_LDFLAGS ""
111	112	define USE_SYSTEM_SQLITE 0
112	113	define USE_LINENOISE 0
113	114	define USE_MMAN_H 0
114	115	define USE_SEE 0
115	116	define SQLITE3_ORIGIN 0
116	117	# SQLITE3_ORIGIN 0 = src/sqlite3, 1=src/sqlite3-see.c, 2=client-provided
	118	+define SQLITE_OPTIONS_EXT ""
	119	+# SQLITE_OPTIONS_EXT => build-dependent CFLAGS for sqlite3.c and shell.c
117	120
118	121	# Maintain the C89/C90-style order of variable declarations before statements.
119	122	# Check if the compiler supports the respective warning flag.
120	123	if {[cctest -cflags -Wdeclaration-after-statement]} {
121	124	define-append EXTRA_CFLAGS -Wdeclaration-after-statement
		@@ -401,10 +404,14 @@
401	404	msg-result "Linked to zlib via default library path"
402	405	} else {
403	406	define-append EXTRA_LDFLAGS "$lcheck"
404	407	msg-result "Linked to zlib via $lcheck"
405	408	}
	409	+ if {![check-function-in-lib compressBound z]} {
	410	+ puts "Notice: disabling zlib compression in the SQL shell"
	411	+ define-append SQLITE_OPTIONS_EXT {-USQLITE_HAVE_ZLIB}
	412	+ }
406	413	break
407	414	}
408	415	}
409	416	set ::zlib_lib -lz
410	417	}
411	418

	--- auto.def
	+++ auto.def
	@@ -104,18 +104,21 @@
104	msg-result "WARNING: tclsh $v found; need >= 8.6 for 'make test'."
105	define TCLSH false ;# force "make test" failure via /usr/bin/false
106	}
107	}
108

109	define EXTRA_CFLAGS "-Wall"
110	define EXTRA_LDFLAGS ""
111	define USE_SYSTEM_SQLITE 0
112	define USE_LINENOISE 0
113	define USE_MMAN_H 0
114	define USE_SEE 0
115	define SQLITE3_ORIGIN 0
116	# SQLITE3_ORIGIN 0 = src/sqlite3, 1=src/sqlite3-see.c, 2=client-provided


117
118	# Maintain the C89/C90-style order of variable declarations before statements.
119	# Check if the compiler supports the respective warning flag.
120	if {[cctest -cflags -Wdeclaration-after-statement]} {
121	define-append EXTRA_CFLAGS -Wdeclaration-after-statement
	@@ -401,10 +404,14 @@
401	msg-result "Linked to zlib via default library path"
402	} else {
403	define-append EXTRA_LDFLAGS "$lcheck"
404	msg-result "Linked to zlib via $lcheck"
405	}




406	break
407	}
408	}
409	set ::zlib_lib -lz
410	}
411

	--- auto.def
	+++ auto.def
	@@ -104,18 +104,21 @@
104	msg-result "WARNING: tclsh $v found; need >= 8.6 for 'make test'."
105	define TCLSH false ;# force "make test" failure via /usr/bin/false
106	}
107	}
108
109	define CFLAGS [get-env CFLAGS "-g -O2"]
110	define EXTRA_CFLAGS "-Wall"
111	define EXTRA_LDFLAGS ""
112	define USE_SYSTEM_SQLITE 0
113	define USE_LINENOISE 0
114	define USE_MMAN_H 0
115	define USE_SEE 0
116	define SQLITE3_ORIGIN 0
117	# SQLITE3_ORIGIN 0 = src/sqlite3, 1=src/sqlite3-see.c, 2=client-provided
118	define SQLITE_OPTIONS_EXT ""
119	# SQLITE_OPTIONS_EXT => build-dependent CFLAGS for sqlite3.c and shell.c
120
121	# Maintain the C89/C90-style order of variable declarations before statements.
122	# Check if the compiler supports the respective warning flag.
123	if {[cctest -cflags -Wdeclaration-after-statement]} {
124	define-append EXTRA_CFLAGS -Wdeclaration-after-statement
	@@ -401,10 +404,14 @@
404	msg-result "Linked to zlib via default library path"
405	} else {
406	define-append EXTRA_LDFLAGS "$lcheck"
407	msg-result "Linked to zlib via $lcheck"
408	}
409	if {![check-function-in-lib compressBound z]} {
410	puts "Notice: disabling zlib compression in the SQL shell"
411	define-append SQLITE_OPTIONS_EXT {-USQLITE_HAVE_ZLIB}
412	}
413	break
414	}
415	}
416	set ::zlib_lib -lz
417	}
418

M autosetup/autosetup

+5 -3

		--- autosetup/autosetup
		+++ autosetup/autosetup
		@@ -564,11 +564,14 @@
564	564	set max [max $max [string length $opt]]
565	565	}
566	566	set indent [string repeat " " [expr {$max+4}]]
567	567	set cols [getenv COLUMNS 80]
568	568	catch {
569		- lassign [exec stty size] rows cols
	569	+ lassign [exec stty size] _ sttycols
	570	+ if {[string is integer -strict $sttycols]} {
	571	+ set cols $sttycols
	572	+ }
570	573	}
571	574	incr cols -1
572	575	# Now output
573	576	foreach help $::autosetup(optionhelp) {
574	577	lassign $help opt module desc
		@@ -908,12 +911,11 @@
908	911	# @find-executable-path name
909	912	#
910	913	# Searches the path for an executable with the given name.
911	914	# Note that the name may include some parameters, e.g. 'cc -mbig-endian',
912	915	# in which case the parameters are ignored.
913		-# The full path to the executable if found, or "" if not found.
914		-# Returns 1 if found, or 0 if not.
	916	+# Returns the full path to the executable if found, or "" if not found.
915	917	#
916	918	proc find-executable-path {name} {
917	919	# Ignore any parameters
918	920	set name [lindex $name 0]
919	921	# The empty string is never a valid executable
920	922

	--- autosetup/autosetup
	+++ autosetup/autosetup
	@@ -564,11 +564,14 @@
564	set max [max $max [string length $opt]]
565	}
566	set indent [string repeat " " [expr {$max+4}]]
567	set cols [getenv COLUMNS 80]
568	catch {
569	lassign [exec stty size] rows cols



570	}
571	incr cols -1
572	# Now output
573	foreach help $::autosetup(optionhelp) {
574	lassign $help opt module desc
	@@ -908,12 +911,11 @@
908	# @find-executable-path name
909	#
910	# Searches the path for an executable with the given name.
911	# Note that the name may include some parameters, e.g. 'cc -mbig-endian',
912	# in which case the parameters are ignored.
913	# The full path to the executable if found, or "" if not found.
914	# Returns 1 if found, or 0 if not.
915	#
916	proc find-executable-path {name} {
917	# Ignore any parameters
918	set name [lindex $name 0]
919	# The empty string is never a valid executable
920

	--- autosetup/autosetup
	+++ autosetup/autosetup
	@@ -564,11 +564,14 @@
564	set max [max $max [string length $opt]]
565	}
566	set indent [string repeat " " [expr {$max+4}]]
567	set cols [getenv COLUMNS 80]
568	catch {
569	lassign [exec stty size] _ sttycols
570	if {[string is integer -strict $sttycols]} {
571	set cols $sttycols
572	}
573	}
574	incr cols -1
575	# Now output
576	foreach help $::autosetup(optionhelp) {
577	lassign $help opt module desc
	@@ -908,12 +911,11 @@
911	# @find-executable-path name
912	#
913	# Searches the path for an executable with the given name.
914	# Note that the name may include some parameters, e.g. 'cc -mbig-endian',
915	# in which case the parameters are ignored.
916	# Returns the full path to the executable if found, or "" if not found.

917	#
918	proc find-executable-path {name} {
919	# Ignore any parameters
920	set name [lindex $name 0]
921	# The empty string is never a valid executable
922

M autosetup/autosetup-config.guess

+95 -28

		--- autosetup/autosetup-config.guess
		+++ autosetup/autosetup-config.guess
		@@ -1,16 +1,16 @@
1	1	#! /bin/sh
2	2	# Attempt to guess a canonical system name.
3		-# Copyright 1992-2021 Free Software Foundation, Inc.
	3	+# Copyright 1992-2024 Free Software Foundation, Inc.
4	4
5	5	# shellcheck disable=SC2006,SC2268 # see below for rationale
6	6
7		-timestamp='2021-06-03'
	7	+timestamp='2024-07-27'
8	8
9	9	# This file is free software; you can redistribute it and/or modify it
10	10	# under the terms of the GNU General Public License as published by
11		-# the Free Software Foundation; either version 3 of the License, or
	11	+# the Free Software Foundation, either version 3 of the License, or
12	12	# (at your option) any later version.
13	13	#
14	14	# This program is distributed in the hope that it will be useful, but
15	15	# WITHOUT ANY WARRANTY; without even the implied warranty of
16	16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
		@@ -45,11 +45,11 @@
45	45	me=`echo "$0" \| sed -e 's,.*/,,'`
46	46
47	47	usage="\
48	48	Usage: $0 [OPTION]
49	49
50		-Output the configuration name of the system \`$me' is run on.
	50	+Output the configuration name of the system '$me' is run on.
51	51
52	52	Options:
53	53	-h, --help print this help, then exit
54	54	-t, --time-stamp print date of last modification, then exit
55	55	-v, --version print version number, then exit
		@@ -58,17 +58,17 @@
58	58
59	59	version="\
60	60	GNU config.guess ($timestamp)
61	61
62	62	Originally written by Per Bothner.
63		-Copyright 1992-2021 Free Software Foundation, Inc.
	63	+Copyright 1992-2024 Free Software Foundation, Inc.
64	64
65	65	This is free software; see the source for copying conditions. There is NO
66	66	warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
67	67
68	68	help="
69		-Try \`$me --help' for more information."
	69	+Try '$me --help' for more information."
70	70
71	71	# Parse command line
72	72	while test $# -gt 0 ; do
73	73	case $1 in
74	74	--time-stamp \| --time* \| -t )
		@@ -100,12 +100,12 @@
100	100	# CC_FOR_BUILD -- compiler used by this script. Note that the use of a
101	101	# compiler to aid in system detection is discouraged as it requires
102	102	# temporary files to be created and, as you can see below, it is a
103	103	# headache to deal with in a portable fashion.
104	104
105		-# Historically, `CC_FOR_BUILD' used to be named `HOST_CC'. We still
106		-# use `HOST_CC' if defined, but it is deprecated.
	105	+# Historically, 'CC_FOR_BUILD' used to be named 'HOST_CC'. We still
	106	+# use 'HOST_CC' if defined, but it is deprecated.
107	107
108	108	# Portable tmp directory creation inspired by the Autoconf team.
109	109
110	110	tmp=
111	111	# shellcheck disable=SC2172
		@@ -121,11 +121,11 @@
121	121	{ tmp=$TMPDIR/cg-$$ && (umask 077 && mkdir "$tmp" 2>/dev/null) && echo "Warning: creating insecure temp directory" >&2 ; } \|\|
122	122	{ echo "$me: cannot create a temporary directory in $TMPDIR" >&2 ; exit 1 ; }
123	123	dummy=$tmp/dummy
124	124	case ${CC_FOR_BUILD-},${HOST_CC-},${CC-} in
125	125	,,) echo "int x;" > "$dummy.c"
126		- for driver in cc gcc c89 c99 ; do
	126	+ for driver in cc gcc c17 c99 c89 ; do
127	127	if ($driver -c -o "$dummy.o" "$dummy.c") >/dev/null 2>&1 ; then
128	128	CC_FOR_BUILD=$driver
129	129	break
130	130	fi
131	131	done
		@@ -153,22 +153,28 @@
153	153	Linux\|GNU\|GNU/*)
154	154	LIBC=unknown
155	155
156	156	set_cc_for_build
157	157	cat <<-EOF > "$dummy.c"
	158	+ #if defined(__ANDROID__)
	159	+ LIBC=android
	160	+ #else
158	161	#include <features.h>
159	162	#if defined(__UCLIBC__)
160	163	LIBC=uclibc
161	164	#elif defined(__dietlibc__)
162	165	LIBC=dietlibc
163	166	#elif defined(__GLIBC__)
164	167	LIBC=gnu
	168	+ #elif defined(__LLVM_LIBC__)
	169	+ LIBC=llvm
165	170	#else
166	171	#include <stdarg.h>
167	172	/* First heuristic to detect musl libc. */
168	173	#ifdef __DEFINED_va_list
169	174	LIBC=musl
	175	+ #endif
170	176	#endif
171	177	#endif
172	178	EOF
173	179	cc_set_libc=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^LIBC' \| sed 's, ,,g'`
174	180	eval "$cc_set_libc"
		@@ -435,11 +441,11 @@
435	441	# If there is a compiler, see if it is configured for 64-bit objects.
436	442	# Note that the Sun cc does not turn __LP64__ into 1 like gcc does.
437	443	# This test works for both compilers.
438	444	if test "$CC_FOR_BUILD" != no_compiler_found; then
439	445	if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') \| \
440		- (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) \| \
	446	+ (CCOPTS="" $CC_FOR_BUILD -m64 -E - 2>/dev/null) \| \
441	447	grep IS_64BIT_ARCH >/dev/null
442	448	then
443	449	SUN_ARCH=x86_64
444	450	fi
445	451	fi
		@@ -457,11 +463,11 @@
457	463	case `/usr/bin/arch -k` in
458	464	Series\|S4)
459	465	UNAME_RELEASE=`uname -v`
460	466	;;
461	467	esac
462		- # Japanese Language versions have a version number like `4.1.3-JL'.
	468	+ # Japanese Language versions have a version number like '4.1.3-JL'.
463	469	SUN_REL=`echo "$UNAME_RELEASE" \| sed -e 's/-/_/'`
464	470	GUESS=sparc-sun-sunos$SUN_REL
465	471	;;
466	472	sun3:SunOS::*)
467	473	GUESS=m68k-sun-sunos$UNAME_RELEASE
		@@ -626,11 +632,12 @@
626	632	if grep bos325 /usr/include/stdio.h >/dev/null 2>&1; then
627	633	set_cc_for_build
628	634	sed 's/^ //' << EOF > "$dummy.c"
629	635	#include <sys/systemcfg.h>
630	636
631		- main()
	637	+ int
	638	+ main ()
632	639	{
633	640	if (!__power_pc())
634	641	exit(1);
635	642	puts("powerpc-ibm-aix3.2.5");
636	643	exit(0);
		@@ -710,11 +717,12 @@
710	717
711	718	#define _HPUX_SOURCE
712	719	#include <stdlib.h>
713	720	#include <unistd.h>
714	721
715		- int main ()
	722	+ int
	723	+ main ()
716	724	{
717	725	#if defined(_SC_KERNEL_BITS)
718	726	long bits = sysconf(_SC_KERNEL_BITS);
719	727	#endif
720	728	long cpu = sysconf (_SC_CPU_VERSION);
		@@ -902,11 +910,11 @@
902	910	FREEBSD_REL=`echo "$UNAME_RELEASE" \| sed -e 's/[-(].*//'`
903	911	GUESS=$UNAME_PROCESSOR-unknown-freebsd$FREEBSD_REL-gnueabihf
904	912	fi
905	913	;;
906	914	:FreeBSD::*)
907		- UNAME_PROCESSOR=`/usr/bin/uname -p`
	915	+ UNAME_PROCESSOR=`uname -p`
908	916	case $UNAME_PROCESSOR in
909	917	amd64)
910	918	UNAME_PROCESSOR=x86_64 ;;
911	919	i386)
912	920	UNAME_PROCESSOR=i586 ;;
		@@ -927,10 +935,13 @@
927	935	GUESS=$UNAME_MACHINE-pc-msys
928	936	;;
929	937	i:PW:*)
930	938	GUESS=$UNAME_MACHINE-pc-pw32
931	939	;;
	940	+ :SerenityOS::*)
	941	+ GUESS=$UNAME_MACHINE-pc-serenity
	942	+ ;;
932	943	:Interix:*)
933	944	case $UNAME_MACHINE in
934	945	x86)
935	946	GUESS=i586-pc-interix$UNAME_RELEASE
936	947	;;
		@@ -961,15 +972,41 @@
961	972	# other systems with GNU libc and userland
962	973	GNU_SYS=`echo "$UNAME_SYSTEM" \| sed 's,^[^/]*/,,' \| tr "[:upper:]" "[:lower:]"`
963	974	GNU_REL=`echo "$UNAME_RELEASE" \| sed -e 's/[-(].*//'`
964	975	GUESS=$UNAME_MACHINE-unknown-$GNU_SYS$GNU_REL-$LIBC
965	976	;;
	977	+ x86_64:[Mm]anagarm::\|i?86:[Mm]anagarm::)
	978	+ GUESS="$UNAME_MACHINE-pc-managarm-mlibc"
	979	+ ;;
	980	+ :[Mm]anagarm::*)
	981	+ GUESS="$UNAME_MACHINE-unknown-managarm-mlibc"
	982	+ ;;
966	983	:Minix::*)
967	984	GUESS=$UNAME_MACHINE-unknown-minix
968	985	;;
969	986	aarch64:Linux::)
970		- GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
	987	+ set_cc_for_build
	988	+ CPU=$UNAME_MACHINE
	989	+ LIBCABI=$LIBC
	990	+ if test "$CC_FOR_BUILD" != no_compiler_found; then
	991	+ ABI=64
	992	+ sed 's/^ //' << EOF > "$dummy.c"
	993	+ #ifdef __ARM_EABI__
	994	+ #ifdef __ARM_PCS_VFP
	995	+ ABI=eabihf
	996	+ #else
	997	+ ABI=eabi
	998	+ #endif
	999	+ #endif
	1000	+EOF
	1001	+ cc_set_abi=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^ABI' \| sed 's, ,,g'`
	1002	+ eval "$cc_set_abi"
	1003	+ case $ABI in
	1004	+ eabi \| eabihf) CPU=armv8l; LIBCABI=$LIBC$ABI ;;
	1005	+ esac
	1006	+ fi
	1007	+ GUESS=$CPU-unknown-linux-$LIBCABI
971	1008	;;
972	1009	aarch64_be:Linux::)
973	1010	UNAME_MACHINE=aarch64_be
974	1011	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
975	1012	;;
		@@ -1031,11 +1068,20 @@
1031	1068	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1032	1069	;;
1033	1070	k1om:Linux::)
1034	1071	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1035	1072	;;
1036		- loongarch32:Linux:: \| loongarch64:Linux:: \| loongarchx32:Linux::)
	1073	+ kvx:Linux::)
	1074	+ GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
	1075	+ ;;
	1076	+ kvx:cos::)
	1077	+ GUESS=$UNAME_MACHINE-unknown-cos
	1078	+ ;;
	1079	+ kvx:mbr::)
	1080	+ GUESS=$UNAME_MACHINE-unknown-mbr
	1081	+ ;;
	1082	+ loongarch32:Linux:: \| loongarch64:Linux::)
1037	1083	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1038	1084	;;
1039	1085	m32r:Linux::*)
1040	1086	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1041	1087	;;
		@@ -1146,20 +1192,31 @@
1146	1192	vax:Linux::)
1147	1193	GUESS=$UNAME_MACHINE-dec-linux-$LIBC
1148	1194	;;
1149	1195	x86_64:Linux::)
1150	1196	set_cc_for_build
	1197	+ CPU=$UNAME_MACHINE
1151	1198	LIBCABI=$LIBC
1152	1199	if test "$CC_FOR_BUILD" != no_compiler_found; then
1153		- if (echo '#ifdef __ILP32__'; echo IS_X32; echo '#endif') \| \
1154		- (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) \| \
1155		- grep IS_X32 >/dev/null
1156		- then
1157		- LIBCABI=${LIBC}x32
1158		- fi
1159		- fi
1160		- GUESS=$UNAME_MACHINE-pc-linux-$LIBCABI
	1200	+ ABI=64
	1201	+ sed 's/^ //' << EOF > "$dummy.c"
	1202	+ #ifdef __i386__
	1203	+ ABI=x86
	1204	+ #else
	1205	+ #ifdef __ILP32__
	1206	+ ABI=x32
	1207	+ #endif
	1208	+ #endif
	1209	+EOF
	1210	+ cc_set_abi=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^ABI' \| sed 's, ,,g'`
	1211	+ eval "$cc_set_abi"
	1212	+ case $ABI in
	1213	+ x86) CPU=i686 ;;
	1214	+ x32) LIBCABI=${LIBC}x32 ;;
	1215	+ esac
	1216	+ fi
	1217	+ GUESS=$CPU-pc-linux-$LIBCABI
1161	1218	;;
1162	1219	xtensa:Linux::*)
1163	1220	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1164	1221	;;
1165	1222	i86:DYNIX/ptx:4:*)
		@@ -1175,11 +1232,11 @@
1175	1232	# I just have to hope. -- rms.
1176	1233	# Use sysv4.2uw... so that sysv4* matches it.
1177	1234	GUESS=$UNAME_MACHINE-pc-sysv4.2uw$UNAME_VERSION
1178	1235	;;
1179	1236	i86:OS/2::*)
1180		- # If we were able to find `uname', then EMX Unix compatibility
	1237	+ # If we were able to find 'uname', then EMX Unix compatibility
1181	1238	# is probably installed.
1182	1239	GUESS=$UNAME_MACHINE-pc-os2-emx
1183	1240	;;
1184	1241	i86:XTS-300::STOP)
1185	1242	GUESS=$UNAME_MACHINE-unknown-stop
		@@ -1316,11 +1373,11 @@
1316	1373	GUESS=$UNAME_MACHINE-sni-sysv4
1317	1374	else
1318	1375	GUESS=ns32k-sni-sysv
1319	1376	fi
1320	1377	;;
1321		- PENTIUM::4.0:*) # Unisys `ClearPath HMP IX 4000' SVR4/MP effort
	1378	+ PENTIUM::4.0:*) # Unisys 'ClearPath HMP IX 4000' SVR4/MP effort
1322	1379	# says <[email protected]>
1323	1380	GUESS=i586-unisys-sysv4
1324	1381	;;
1325	1382	:UNIX_System_V:4:FTX*)
1326	1383	# From Gerald Hewes <[email protected]>.
		@@ -1362,12 +1419,15 @@
1362	1419	GUESS=i586-pc-beos
1363	1420	;;
1364	1421	BePC:Haiku::) # Haiku running on Intel PC compatible.
1365	1422	GUESS=i586-pc-haiku
1366	1423	;;
1367		- x86_64:Haiku::)
1368		- GUESS=x86_64-unknown-haiku
	1424	+ ppc:Haiku::) # Haiku running on Apple PowerPC
	1425	+ GUESS=powerpc-apple-haiku
	1426	+ ;;
	1427	+ :Haiku::*) # Haiku modern gcc (not bound by BeOS compat)
	1428	+ GUESS=$UNAME_MACHINE-unknown-haiku
1369	1429	;;
1370	1430	SX-4:SUPER-UX::)
1371	1431	GUESS=sx4-nec-superux$UNAME_RELEASE
1372	1432	;;
1373	1433	SX-5:SUPER-UX::)
		@@ -1520,10 +1580,13 @@
1520	1580	GUESS=$UNAME_MACHINE-pc-skyos$SKYOS_REL
1521	1581	;;
1522	1582	i86:rdos::*)
1523	1583	GUESS=$UNAME_MACHINE-pc-rdos
1524	1584	;;
	1585	+ i86:Fiwix::*)
	1586	+ GUESS=$UNAME_MACHINE-pc-fiwix
	1587	+ ;;
1525	1588	:AROS::*)
1526	1589	GUESS=$UNAME_MACHINE-unknown-aros
1527	1590	;;
1528	1591	x86_64:VMkernel::)
1529	1592	GUESS=$UNAME_MACHINE-unknown-esx
		@@ -1532,10 +1595,13 @@
1532	1595	GUESS=x86_64-unknown-onefs
1533	1596	;;
1534	1597	:Unleashed::*)
1535	1598	GUESS=$UNAME_MACHINE-unknown-unleashed$UNAME_RELEASE
1536	1599	;;
	1600	+ :Ironclad::*)
	1601	+ GUESS=$UNAME_MACHINE-unknown-ironclad
	1602	+ ;;
1537	1603	esac
1538	1604
1539	1605	# Do we have a guess based on uname results?
1540	1606	if test "x$GUESS" != x; then
1541	1607	echo "$GUESS"
		@@ -1555,10 +1621,11 @@
1555	1621	#if defined(_SIZE_T_) \|\| defined(SIGLOST)
1556	1622	#include <sys/utsname.h>
1557	1623	#endif
1558	1624	#endif
1559	1625	#endif
	1626	+int
1560	1627	main ()
1561	1628	{
1562	1629	#if defined (sony)
1563	1630	#if defined (MIPSEB)
1564	1631	/* BFD wants "bsd" instead of "newsos". Perhaps BFD should be changed,
1565	1632

	--- autosetup/autosetup-config.guess
	+++ autosetup/autosetup-config.guess
	@@ -1,16 +1,16 @@
1	#! /bin/sh
2	# Attempt to guess a canonical system name.
3	# Copyright 1992-2021 Free Software Foundation, Inc.
4
5	# shellcheck disable=SC2006,SC2268 # see below for rationale
6
7	timestamp='2021-06-03'
8
9	# This file is free software; you can redistribute it and/or modify it
10	# under the terms of the GNU General Public License as published by
11	# the Free Software Foundation; either version 3 of the License, or
12	# (at your option) any later version.
13	#
14	# This program is distributed in the hope that it will be useful, but
15	# WITHOUT ANY WARRANTY; without even the implied warranty of
16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	@@ -45,11 +45,11 @@
45	me=`echo "$0" \| sed -e 's,.*/,,'`
46
47	usage="\
48	Usage: $0 [OPTION]
49
50	Output the configuration name of the system \`$me' is run on.
51
52	Options:
53	-h, --help print this help, then exit
54	-t, --time-stamp print date of last modification, then exit
55	-v, --version print version number, then exit
	@@ -58,17 +58,17 @@
58
59	version="\
60	GNU config.guess ($timestamp)
61
62	Originally written by Per Bothner.
63	Copyright 1992-2021 Free Software Foundation, Inc.
64
65	This is free software; see the source for copying conditions. There is NO
66	warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
67
68	help="
69	Try \`$me --help' for more information."
70
71	# Parse command line
72	while test $# -gt 0 ; do
73	case $1 in
74	--time-stamp \| --time* \| -t )
	@@ -100,12 +100,12 @@
100	# CC_FOR_BUILD -- compiler used by this script. Note that the use of a
101	# compiler to aid in system detection is discouraged as it requires
102	# temporary files to be created and, as you can see below, it is a
103	# headache to deal with in a portable fashion.
104
105	# Historically, `CC_FOR_BUILD' used to be named `HOST_CC'. We still
106	# use `HOST_CC' if defined, but it is deprecated.
107
108	# Portable tmp directory creation inspired by the Autoconf team.
109
110	tmp=
111	# shellcheck disable=SC2172
	@@ -121,11 +121,11 @@
121	{ tmp=$TMPDIR/cg-$$ && (umask 077 && mkdir "$tmp" 2>/dev/null) && echo "Warning: creating insecure temp directory" >&2 ; } \|\|
122	{ echo "$me: cannot create a temporary directory in $TMPDIR" >&2 ; exit 1 ; }
123	dummy=$tmp/dummy
124	case ${CC_FOR_BUILD-},${HOST_CC-},${CC-} in
125	,,) echo "int x;" > "$dummy.c"
126	for driver in cc gcc c89 c99 ; do
127	if ($driver -c -o "$dummy.o" "$dummy.c") >/dev/null 2>&1 ; then
128	CC_FOR_BUILD=$driver
129	break
130	fi
131	done
	@@ -153,22 +153,28 @@
153	Linux\|GNU\|GNU/*)
154	LIBC=unknown
155
156	set_cc_for_build
157	cat <<-EOF > "$dummy.c"



158	#include <features.h>
159	#if defined(__UCLIBC__)
160	LIBC=uclibc
161	#elif defined(__dietlibc__)
162	LIBC=dietlibc
163	#elif defined(__GLIBC__)
164	LIBC=gnu


165	#else
166	#include <stdarg.h>
167	/* First heuristic to detect musl libc. */
168	#ifdef __DEFINED_va_list
169	LIBC=musl

170	#endif
171	#endif
172	EOF
173	cc_set_libc=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^LIBC' \| sed 's, ,,g'`
174	eval "$cc_set_libc"
	@@ -435,11 +441,11 @@
435	# If there is a compiler, see if it is configured for 64-bit objects.
436	# Note that the Sun cc does not turn __LP64__ into 1 like gcc does.
437	# This test works for both compilers.
438	if test "$CC_FOR_BUILD" != no_compiler_found; then
439	if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') \| \
440	(CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) \| \
441	grep IS_64BIT_ARCH >/dev/null
442	then
443	SUN_ARCH=x86_64
444	fi
445	fi
	@@ -457,11 +463,11 @@
457	case `/usr/bin/arch -k` in
458	Series\|S4)
459	UNAME_RELEASE=`uname -v`
460	;;
461	esac
462	# Japanese Language versions have a version number like `4.1.3-JL'.
463	SUN_REL=`echo "$UNAME_RELEASE" \| sed -e 's/-/_/'`
464	GUESS=sparc-sun-sunos$SUN_REL
465	;;
466	sun3:SunOS::*)
467	GUESS=m68k-sun-sunos$UNAME_RELEASE
	@@ -626,11 +632,12 @@
626	if grep bos325 /usr/include/stdio.h >/dev/null 2>&1; then
627	set_cc_for_build
628	sed 's/^ //' << EOF > "$dummy.c"
629	#include <sys/systemcfg.h>
630
631	main()

632	{
633	if (!__power_pc())
634	exit(1);
635	puts("powerpc-ibm-aix3.2.5");
636	exit(0);
	@@ -710,11 +717,12 @@
710
711	#define _HPUX_SOURCE
712	#include <stdlib.h>
713	#include <unistd.h>
714
715	int main ()

716	{
717	#if defined(_SC_KERNEL_BITS)
718	long bits = sysconf(_SC_KERNEL_BITS);
719	#endif
720	long cpu = sysconf (_SC_CPU_VERSION);
	@@ -902,11 +910,11 @@
902	FREEBSD_REL=`echo "$UNAME_RELEASE" \| sed -e 's/[-(].*//'`
903	GUESS=$UNAME_PROCESSOR-unknown-freebsd$FREEBSD_REL-gnueabihf
904	fi
905	;;
906	:FreeBSD::*)
907	UNAME_PROCESSOR=`/usr/bin/uname -p`
908	case $UNAME_PROCESSOR in
909	amd64)
910	UNAME_PROCESSOR=x86_64 ;;
911	i386)
912	UNAME_PROCESSOR=i586 ;;
	@@ -927,10 +935,13 @@
927	GUESS=$UNAME_MACHINE-pc-msys
928	;;
929	i:PW:*)
930	GUESS=$UNAME_MACHINE-pc-pw32
931	;;



932	:Interix:*)
933	case $UNAME_MACHINE in
934	x86)
935	GUESS=i586-pc-interix$UNAME_RELEASE
936	;;
	@@ -961,15 +972,41 @@
961	# other systems with GNU libc and userland
962	GNU_SYS=`echo "$UNAME_SYSTEM" \| sed 's,^[^/]*/,,' \| tr "[:upper:]" "[:lower:]"`
963	GNU_REL=`echo "$UNAME_RELEASE" \| sed -e 's/[-(].*//'`
964	GUESS=$UNAME_MACHINE-unknown-$GNU_SYS$GNU_REL-$LIBC
965	;;






966	:Minix::*)
967	GUESS=$UNAME_MACHINE-unknown-minix
968	;;
969	aarch64:Linux::)
970	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC




















971	;;
972	aarch64_be:Linux::)
973	UNAME_MACHINE=aarch64_be
974	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
975	;;
	@@ -1031,11 +1068,20 @@
1031	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1032	;;
1033	k1om:Linux::)
1034	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1035	;;
1036	loongarch32:Linux:: \| loongarch64:Linux:: \| loongarchx32:Linux::)









1037	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1038	;;
1039	m32r:Linux::*)
1040	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1041	;;
	@@ -1146,20 +1192,31 @@
1146	vax:Linux::)
1147	GUESS=$UNAME_MACHINE-dec-linux-$LIBC
1148	;;
1149	x86_64:Linux::)
1150	set_cc_for_build

1151	LIBCABI=$LIBC
1152	if test "$CC_FOR_BUILD" != no_compiler_found; then
1153	if (echo '#ifdef __ILP32__'; echo IS_X32; echo '#endif') \| \
1154	(CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) \| \
1155	grep IS_X32 >/dev/null
1156	then
1157	LIBCABI=${LIBC}x32
1158	fi
1159	fi
1160	GUESS=$UNAME_MACHINE-pc-linux-$LIBCABI










1161	;;
1162	xtensa:Linux::*)
1163	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1164	;;
1165	i86:DYNIX/ptx:4:*)
	@@ -1175,11 +1232,11 @@
1175	# I just have to hope. -- rms.
1176	# Use sysv4.2uw... so that sysv4* matches it.
1177	GUESS=$UNAME_MACHINE-pc-sysv4.2uw$UNAME_VERSION
1178	;;
1179	i86:OS/2::*)
1180	# If we were able to find `uname', then EMX Unix compatibility
1181	# is probably installed.
1182	GUESS=$UNAME_MACHINE-pc-os2-emx
1183	;;
1184	i86:XTS-300::STOP)
1185	GUESS=$UNAME_MACHINE-unknown-stop
	@@ -1316,11 +1373,11 @@
1316	GUESS=$UNAME_MACHINE-sni-sysv4
1317	else
1318	GUESS=ns32k-sni-sysv
1319	fi
1320	;;
1321	PENTIUM::4.0:*) # Unisys `ClearPath HMP IX 4000' SVR4/MP effort
1322	# says <[email protected]>
1323	GUESS=i586-unisys-sysv4
1324	;;
1325	:UNIX_System_V:4:FTX*)
1326	# From Gerald Hewes <[email protected]>.
	@@ -1362,12 +1419,15 @@
1362	GUESS=i586-pc-beos
1363	;;
1364	BePC:Haiku::) # Haiku running on Intel PC compatible.
1365	GUESS=i586-pc-haiku
1366	;;
1367	x86_64:Haiku::)
1368	GUESS=x86_64-unknown-haiku



1369	;;
1370	SX-4:SUPER-UX::)
1371	GUESS=sx4-nec-superux$UNAME_RELEASE
1372	;;
1373	SX-5:SUPER-UX::)
	@@ -1520,10 +1580,13 @@
1520	GUESS=$UNAME_MACHINE-pc-skyos$SKYOS_REL
1521	;;
1522	i86:rdos::*)
1523	GUESS=$UNAME_MACHINE-pc-rdos
1524	;;



1525	:AROS::*)
1526	GUESS=$UNAME_MACHINE-unknown-aros
1527	;;
1528	x86_64:VMkernel::)
1529	GUESS=$UNAME_MACHINE-unknown-esx
	@@ -1532,10 +1595,13 @@
1532	GUESS=x86_64-unknown-onefs
1533	;;
1534	:Unleashed::*)
1535	GUESS=$UNAME_MACHINE-unknown-unleashed$UNAME_RELEASE
1536	;;



1537	esac
1538
1539	# Do we have a guess based on uname results?
1540	if test "x$GUESS" != x; then
1541	echo "$GUESS"
	@@ -1555,10 +1621,11 @@
1555	#if defined(_SIZE_T_) \|\| defined(SIGLOST)
1556	#include <sys/utsname.h>
1557	#endif
1558	#endif
1559	#endif

1560	main ()
1561	{
1562	#if defined (sony)
1563	#if defined (MIPSEB)
1564	/* BFD wants "bsd" instead of "newsos". Perhaps BFD should be changed,
1565

	--- autosetup/autosetup-config.guess
	+++ autosetup/autosetup-config.guess
	@@ -1,16 +1,16 @@
1	#! /bin/sh
2	# Attempt to guess a canonical system name.
3	# Copyright 1992-2024 Free Software Foundation, Inc.
4
5	# shellcheck disable=SC2006,SC2268 # see below for rationale
6
7	timestamp='2024-07-27'
8
9	# This file is free software; you can redistribute it and/or modify it
10	# under the terms of the GNU General Public License as published by
11	# the Free Software Foundation, either version 3 of the License, or
12	# (at your option) any later version.
13	#
14	# This program is distributed in the hope that it will be useful, but
15	# WITHOUT ANY WARRANTY; without even the implied warranty of
16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	@@ -45,11 +45,11 @@
45	me=`echo "$0" \| sed -e 's,.*/,,'`
46
47	usage="\
48	Usage: $0 [OPTION]
49
50	Output the configuration name of the system '$me' is run on.
51
52	Options:
53	-h, --help print this help, then exit
54	-t, --time-stamp print date of last modification, then exit
55	-v, --version print version number, then exit
	@@ -58,17 +58,17 @@
58
59	version="\
60	GNU config.guess ($timestamp)
61
62	Originally written by Per Bothner.
63	Copyright 1992-2024 Free Software Foundation, Inc.
64
65	This is free software; see the source for copying conditions. There is NO
66	warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
67
68	help="
69	Try '$me --help' for more information."
70
71	# Parse command line
72	while test $# -gt 0 ; do
73	case $1 in
74	--time-stamp \| --time* \| -t )
	@@ -100,12 +100,12 @@
100	# CC_FOR_BUILD -- compiler used by this script. Note that the use of a
101	# compiler to aid in system detection is discouraged as it requires
102	# temporary files to be created and, as you can see below, it is a
103	# headache to deal with in a portable fashion.
104
105	# Historically, 'CC_FOR_BUILD' used to be named 'HOST_CC'. We still
106	# use 'HOST_CC' if defined, but it is deprecated.
107
108	# Portable tmp directory creation inspired by the Autoconf team.
109
110	tmp=
111	# shellcheck disable=SC2172
	@@ -121,11 +121,11 @@
121	{ tmp=$TMPDIR/cg-$$ && (umask 077 && mkdir "$tmp" 2>/dev/null) && echo "Warning: creating insecure temp directory" >&2 ; } \|\|
122	{ echo "$me: cannot create a temporary directory in $TMPDIR" >&2 ; exit 1 ; }
123	dummy=$tmp/dummy
124	case ${CC_FOR_BUILD-},${HOST_CC-},${CC-} in
125	,,) echo "int x;" > "$dummy.c"
126	for driver in cc gcc c17 c99 c89 ; do
127	if ($driver -c -o "$dummy.o" "$dummy.c") >/dev/null 2>&1 ; then
128	CC_FOR_BUILD=$driver
129	break
130	fi
131	done
	@@ -153,22 +153,28 @@
153	Linux\|GNU\|GNU/*)
154	LIBC=unknown
155
156	set_cc_for_build
157	cat <<-EOF > "$dummy.c"
158	#if defined(__ANDROID__)
159	LIBC=android
160	#else
161	#include <features.h>
162	#if defined(__UCLIBC__)
163	LIBC=uclibc
164	#elif defined(__dietlibc__)
165	LIBC=dietlibc
166	#elif defined(__GLIBC__)
167	LIBC=gnu
168	#elif defined(__LLVM_LIBC__)
169	LIBC=llvm
170	#else
171	#include <stdarg.h>
172	/* First heuristic to detect musl libc. */
173	#ifdef __DEFINED_va_list
174	LIBC=musl
175	#endif
176	#endif
177	#endif
178	EOF
179	cc_set_libc=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^LIBC' \| sed 's, ,,g'`
180	eval "$cc_set_libc"
	@@ -435,11 +441,11 @@
441	# If there is a compiler, see if it is configured for 64-bit objects.
442	# Note that the Sun cc does not turn __LP64__ into 1 like gcc does.
443	# This test works for both compilers.
444	if test "$CC_FOR_BUILD" != no_compiler_found; then
445	if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') \| \
446	(CCOPTS="" $CC_FOR_BUILD -m64 -E - 2>/dev/null) \| \
447	grep IS_64BIT_ARCH >/dev/null
448	then
449	SUN_ARCH=x86_64
450	fi
451	fi
	@@ -457,11 +463,11 @@
463	case `/usr/bin/arch -k` in
464	Series\|S4)
465	UNAME_RELEASE=`uname -v`
466	;;
467	esac
468	# Japanese Language versions have a version number like '4.1.3-JL'.
469	SUN_REL=`echo "$UNAME_RELEASE" \| sed -e 's/-/_/'`
470	GUESS=sparc-sun-sunos$SUN_REL
471	;;
472	sun3:SunOS::*)
473	GUESS=m68k-sun-sunos$UNAME_RELEASE
	@@ -626,11 +632,12 @@
632	if grep bos325 /usr/include/stdio.h >/dev/null 2>&1; then
633	set_cc_for_build
634	sed 's/^ //' << EOF > "$dummy.c"
635	#include <sys/systemcfg.h>
636
637	int
638	main ()
639	{
640	if (!__power_pc())
641	exit(1);
642	puts("powerpc-ibm-aix3.2.5");
643	exit(0);
	@@ -710,11 +717,12 @@
717
718	#define _HPUX_SOURCE
719	#include <stdlib.h>
720	#include <unistd.h>
721
722	int
723	main ()
724	{
725	#if defined(_SC_KERNEL_BITS)
726	long bits = sysconf(_SC_KERNEL_BITS);
727	#endif
728	long cpu = sysconf (_SC_CPU_VERSION);
	@@ -902,11 +910,11 @@
910	FREEBSD_REL=`echo "$UNAME_RELEASE" \| sed -e 's/[-(].*//'`
911	GUESS=$UNAME_PROCESSOR-unknown-freebsd$FREEBSD_REL-gnueabihf
912	fi
913	;;
914	:FreeBSD::*)
915	UNAME_PROCESSOR=`uname -p`
916	case $UNAME_PROCESSOR in
917	amd64)
918	UNAME_PROCESSOR=x86_64 ;;
919	i386)
920	UNAME_PROCESSOR=i586 ;;
	@@ -927,10 +935,13 @@
935	GUESS=$UNAME_MACHINE-pc-msys
936	;;
937	i:PW:*)
938	GUESS=$UNAME_MACHINE-pc-pw32
939	;;
940	:SerenityOS::*)
941	GUESS=$UNAME_MACHINE-pc-serenity
942	;;
943	:Interix:*)
944	case $UNAME_MACHINE in
945	x86)
946	GUESS=i586-pc-interix$UNAME_RELEASE
947	;;
	@@ -961,15 +972,41 @@
972	# other systems with GNU libc and userland
973	GNU_SYS=`echo "$UNAME_SYSTEM" \| sed 's,^[^/]*/,,' \| tr "[:upper:]" "[:lower:]"`
974	GNU_REL=`echo "$UNAME_RELEASE" \| sed -e 's/[-(].*//'`
975	GUESS=$UNAME_MACHINE-unknown-$GNU_SYS$GNU_REL-$LIBC
976	;;
977	x86_64:[Mm]anagarm::\|i?86:[Mm]anagarm::)
978	GUESS="$UNAME_MACHINE-pc-managarm-mlibc"
979	;;
980	:[Mm]anagarm::*)
981	GUESS="$UNAME_MACHINE-unknown-managarm-mlibc"
982	;;
983	:Minix::*)
984	GUESS=$UNAME_MACHINE-unknown-minix
985	;;
986	aarch64:Linux::)
987	set_cc_for_build
988	CPU=$UNAME_MACHINE
989	LIBCABI=$LIBC
990	if test "$CC_FOR_BUILD" != no_compiler_found; then
991	ABI=64
992	sed 's/^ //' << EOF > "$dummy.c"
993	#ifdef __ARM_EABI__
994	#ifdef __ARM_PCS_VFP
995	ABI=eabihf
996	#else
997	ABI=eabi
998	#endif
999	#endif
1000	EOF
1001	cc_set_abi=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^ABI' \| sed 's, ,,g'`
1002	eval "$cc_set_abi"
1003	case $ABI in
1004	eabi \| eabihf) CPU=armv8l; LIBCABI=$LIBC$ABI ;;
1005	esac
1006	fi
1007	GUESS=$CPU-unknown-linux-$LIBCABI
1008	;;
1009	aarch64_be:Linux::)
1010	UNAME_MACHINE=aarch64_be
1011	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1012	;;
	@@ -1031,11 +1068,20 @@
1068	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1069	;;
1070	k1om:Linux::)
1071	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1072	;;
1073	kvx:Linux::)
1074	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1075	;;
1076	kvx:cos::)
1077	GUESS=$UNAME_MACHINE-unknown-cos
1078	;;
1079	kvx:mbr::)
1080	GUESS=$UNAME_MACHINE-unknown-mbr
1081	;;
1082	loongarch32:Linux:: \| loongarch64:Linux::)
1083	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1084	;;
1085	m32r:Linux::*)
1086	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1087	;;
	@@ -1146,20 +1192,31 @@
1192	vax:Linux::)
1193	GUESS=$UNAME_MACHINE-dec-linux-$LIBC
1194	;;
1195	x86_64:Linux::)
1196	set_cc_for_build
1197	CPU=$UNAME_MACHINE
1198	LIBCABI=$LIBC
1199	if test "$CC_FOR_BUILD" != no_compiler_found; then
1200	ABI=64
1201	sed 's/^ //' << EOF > "$dummy.c"
1202	#ifdef __i386__
1203	ABI=x86
1204	#else
1205	#ifdef __ILP32__
1206	ABI=x32
1207	#endif
1208	#endif
1209	EOF
1210	cc_set_abi=`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null \| grep '^ABI' \| sed 's, ,,g'`
1211	eval "$cc_set_abi"
1212	case $ABI in
1213	x86) CPU=i686 ;;
1214	x32) LIBCABI=${LIBC}x32 ;;
1215	esac
1216	fi
1217	GUESS=$CPU-pc-linux-$LIBCABI
1218	;;
1219	xtensa:Linux::*)
1220	GUESS=$UNAME_MACHINE-unknown-linux-$LIBC
1221	;;
1222	i86:DYNIX/ptx:4:*)
	@@ -1175,11 +1232,11 @@
1232	# I just have to hope. -- rms.
1233	# Use sysv4.2uw... so that sysv4* matches it.
1234	GUESS=$UNAME_MACHINE-pc-sysv4.2uw$UNAME_VERSION
1235	;;
1236	i86:OS/2::*)
1237	# If we were able to find 'uname', then EMX Unix compatibility
1238	# is probably installed.
1239	GUESS=$UNAME_MACHINE-pc-os2-emx
1240	;;
1241	i86:XTS-300::STOP)
1242	GUESS=$UNAME_MACHINE-unknown-stop
	@@ -1316,11 +1373,11 @@
1373	GUESS=$UNAME_MACHINE-sni-sysv4
1374	else
1375	GUESS=ns32k-sni-sysv
1376	fi
1377	;;
1378	PENTIUM::4.0:*) # Unisys 'ClearPath HMP IX 4000' SVR4/MP effort
1379	# says <[email protected]>
1380	GUESS=i586-unisys-sysv4
1381	;;
1382	:UNIX_System_V:4:FTX*)
1383	# From Gerald Hewes <[email protected]>.
	@@ -1362,12 +1419,15 @@
1419	GUESS=i586-pc-beos
1420	;;
1421	BePC:Haiku::) # Haiku running on Intel PC compatible.
1422	GUESS=i586-pc-haiku
1423	;;
1424	ppc:Haiku::) # Haiku running on Apple PowerPC
1425	GUESS=powerpc-apple-haiku
1426	;;
1427	:Haiku::*) # Haiku modern gcc (not bound by BeOS compat)
1428	GUESS=$UNAME_MACHINE-unknown-haiku
1429	;;
1430	SX-4:SUPER-UX::)
1431	GUESS=sx4-nec-superux$UNAME_RELEASE
1432	;;
1433	SX-5:SUPER-UX::)
	@@ -1520,10 +1580,13 @@
1580	GUESS=$UNAME_MACHINE-pc-skyos$SKYOS_REL
1581	;;
1582	i86:rdos::*)
1583	GUESS=$UNAME_MACHINE-pc-rdos
1584	;;
1585	i86:Fiwix::*)
1586	GUESS=$UNAME_MACHINE-pc-fiwix
1587	;;
1588	:AROS::*)
1589	GUESS=$UNAME_MACHINE-unknown-aros
1590	;;
1591	x86_64:VMkernel::)
1592	GUESS=$UNAME_MACHINE-unknown-esx
	@@ -1532,10 +1595,13 @@
1595	GUESS=x86_64-unknown-onefs
1596	;;
1597	:Unleashed::*)
1598	GUESS=$UNAME_MACHINE-unknown-unleashed$UNAME_RELEASE
1599	;;
1600	:Ironclad::*)
1601	GUESS=$UNAME_MACHINE-unknown-ironclad
1602	;;
1603	esac
1604
1605	# Do we have a guess based on uname results?
1606	if test "x$GUESS" != x; then
1607	echo "$GUESS"
	@@ -1555,10 +1621,11 @@
1621	#if defined(_SIZE_T_) \|\| defined(SIGLOST)
1622	#include <sys/utsname.h>
1623	#endif
1624	#endif
1625	#endif
1626	int
1627	main ()
1628	{
1629	#if defined (sony)
1630	#if defined (MIPSEB)
1631	/* BFD wants "bsd" instead of "newsos". Perhaps BFD should be changed,
1632

M autosetup/autosetup-config.sub

+722 -241

		--- autosetup/autosetup-config.sub
		+++ autosetup/autosetup-config.sub
		@@ -1,16 +1,16 @@
1	1	#! /bin/sh
2	2	# Configuration validation subroutine script.
3		-# Copyright 1992-2021 Free Software Foundation, Inc.
	3	+# Copyright 1992-2024 Free Software Foundation, Inc.
4	4
5		-# shellcheck disable=SC2006,SC2268 # see below for rationale
	5	+# shellcheck disable=SC2006,SC2268,SC2162 # see below for rationale
6	6
7		-timestamp='2021-07-03'
	7	+timestamp='2024-05-27'
8	8
9	9	# This file is free software; you can redistribute it and/or modify it
10	10	# under the terms of the GNU General Public License as published by
11		-# the Free Software Foundation; either version 3 of the License, or
	11	+# the Free Software Foundation, either version 3 of the License, or
12	12	# (at your option) any later version.
13	13	#
14	14	# This program is distributed in the hope that it will be useful, but
15	15	# WITHOUT ANY WARRANTY; without even the implied warranty of
16	16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
		@@ -74,17 +74,17 @@
74	74	Report bugs and patches to <[email protected]>."
75	75
76	76	version="\
77	77	GNU config.sub ($timestamp)
78	78
79		-Copyright 1992-2021 Free Software Foundation, Inc.
	79	+Copyright 1992-2024 Free Software Foundation, Inc.
80	80
81	81	This is free software; see the source for copying conditions. There is NO
82	82	warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
83	83
84	84	help="
85		-Try \`$me --help' for more information."
	85	+Try '$me --help' for more information."
86	86
87	87	# Parse command line
88	88	while test $# -gt 0 ; do
89	89	case $1 in
90	90	--time-stamp \| --time* \| -t )
		@@ -118,19 +118,20 @@
118	118	*) echo "$me: too many arguments$help" >&2
119	119	exit 1;;
120	120	esac
121	121
122	122	# Split fields of configuration type
123		-# shellcheck disable=SC2162
	123	+saved_IFS=$IFS
124	124	IFS="-" read field1 field2 field3 field4 <<EOF
125	125	$1
126	126	EOF
	127	+IFS=$saved_IFS
127	128
128	129	# Separate into logical components for further validation
129	130	case $1 in
130	131	----*)
131		- echo Invalid configuration \`"$1"\': more than four components >&2
	132	+ echo "Invalid configuration '$1': more than four components" >&2
132	133	exit 1
133	134	;;
134	135	---)
135	136	basic_machine=$field1-$field2
136	137	basic_os=$field3-$field4
		@@ -138,14 +139,25 @@
138	139	--*)
139	140	# Ambiguous whether COMPANY is present, or skipped and KERNEL-OS is two
140	141	# parts
141	142	maybe_os=$field2-$field3
142	143	case $maybe_os in
143		- nto-qnx* \| linux-* \| uclinux-uclibc* \
144		- \| uclinux-gnu* \| kfreebsd-gnu \| knetbsd-gnu \| netbsd-gnu \
145		- \| netbsd-eabi \| kopensolaris-gnu \| cloudabi-eabi \
146		- \| storm-chaos* \| os2-emx* \| rtmk-nova*)
	144	+ cloudabi-eabi \
	145	+ \| kfreebsd-gnu \
	146	+ \| knetbsd-gnu \
	147	+ \| kopensolaris-gnu \
	148	+ \| linux-* \
	149	+ \| managarm-* \
	150	+ \| netbsd-eabi \
	151	+ \| netbsd-gnu \
	152	+ \| nto-qnx* \
	153	+ \| os2-emx* \
	154	+ \| rtmk-nova* \
	155	+ \| storm-chaos* \
	156	+ \| uclinux-gnu* \
	157	+ \| uclinux-uclibc* \
	158	+ \| windows-* )
147	159	basic_machine=$field1
148	160	basic_os=$maybe_os
149	161	;;
150	162	android-linux)
151	163	basic_machine=$field1-unknown
		@@ -156,37 +168,105 @@
156	168	basic_os=$field3
157	169	;;
158	170	esac
159	171	;;
160	172	-)
161		- # A lone config we happen to match not fitting any pattern
162	173	case $field1-$field2 in
	174	+ # Shorthands that happen to contain a single dash
	175	+ convex-c[12] \| convex-c3[248])
	176	+ basic_machine=$field2-convex
	177	+ basic_os=
	178	+ ;;
163	179	decstation-3100)
164	180	basic_machine=mips-dec
165	181	basic_os=
166	182	;;
167	183	-)
168	184	# Second component is usually, but not always the OS
169	185	case $field2 in
170		- # Prevent following clause from handling this valid os
	186	+ # Do not treat sunos as a manufacturer
171	187	sunos)
172	188	basic_machine=$field1
173	189	basic_os=$field2
174	190	;;
175	191	# Manufacturers
176		- dec* \| mips* \| sequent* \| encore* \| pc533* \| sgi* \| sony* \
177		- \| att* \| 7300* \| 3300* \| delta* \| motorola* \| sun[234]* \
178		- \| unicom* \| ibm* \| next \| hp \| isi* \| apollo \| altos* \
179		- \| convergent* \| ncr* \| news \| 32* \| 3600* \| 3100* \
180		- \| hitachi* \| c[123]* \| convex* \| sun \| crds \| omron* \| dg \
181		- \| ultra \| tti* \| harris \| dolphin \| highlevel \| gould \
182		- \| cbm \| ns \| masscomp \| apple \| axis \| knuth \| cray \
183		- \| microblaze* \| sim \| cisco \
184		- \| oki \| wec \| wrs \| winbond)
	192	+ 3100* \
	193	+ \| 32* \
	194	+ \| 3300* \
	195	+ \| 3600* \
	196	+ \| 7300* \
	197	+ \| acorn \
	198	+ \| altos* \
	199	+ \| apollo \
	200	+ \| apple \
	201	+ \| atari \
	202	+ \| att* \
	203	+ \| axis \
	204	+ \| be \
	205	+ \| bull \
	206	+ \| cbm \
	207	+ \| ccur \
	208	+ \| cisco \
	209	+ \| commodore \
	210	+ \| convergent* \
	211	+ \| convex* \
	212	+ \| cray \
	213	+ \| crds \
	214	+ \| dec* \
	215	+ \| delta* \
	216	+ \| dg \
	217	+ \| digital \
	218	+ \| dolphin \
	219	+ \| encore* \
	220	+ \| gould \
	221	+ \| harris \
	222	+ \| highlevel \
	223	+ \| hitachi* \
	224	+ \| hp \
	225	+ \| ibm* \
	226	+ \| intergraph \
	227	+ \| isi* \
	228	+ \| knuth \
	229	+ \| masscomp \
	230	+ \| microblaze* \
	231	+ \| mips* \
	232	+ \| motorola* \
	233	+ \| ncr* \
	234	+ \| news \
	235	+ \| next \
	236	+ \| ns \
	237	+ \| oki \
	238	+ \| omron* \
	239	+ \| pc533* \
	240	+ \| rebel \
	241	+ \| rom68k \
	242	+ \| rombug \
	243	+ \| semi \
	244	+ \| sequent* \
	245	+ \| siemens \
	246	+ \| sgi* \
	247	+ \| siemens \
	248	+ \| sim \
	249	+ \| sni \
	250	+ \| sony* \
	251	+ \| stratus \
	252	+ \| sun \
	253	+ \| sun[234]* \
	254	+ \| tektronix \
	255	+ \| tti* \
	256	+ \| ultra \
	257	+ \| unicom* \
	258	+ \| wec \
	259	+ \| winbond \
	260	+ \| wrs)
185	261	basic_machine=$field1-$field2
186	262	basic_os=
187	263	;;
	264	+ zephyr*)
	265	+ basic_machine=$field1-unknown
	266	+ basic_os=$field2
	267	+ ;;
188	268	*)
189	269	basic_machine=$field1
190	270	basic_os=$field2
191	271	;;
192	272	esac
		@@ -263,30 +343,10 @@
263	343	;;
264	344	cegcc)
265	345	basic_machine=arm-unknown
266	346	basic_os=cegcc
267	347	;;
268		- convex-c1)
269		- basic_machine=c1-convex
270		- basic_os=bsd
271		- ;;
272		- convex-c2)
273		- basic_machine=c2-convex
274		- basic_os=bsd
275		- ;;
276		- convex-c32)
277		- basic_machine=c32-convex
278		- basic_os=bsd
279		- ;;
280		- convex-c34)
281		- basic_machine=c34-convex
282		- basic_os=bsd
283		- ;;
284		- convex-c38)
285		- basic_machine=c38-convex
286		- basic_os=bsd
287		- ;;
288	348	cray)
289	349	basic_machine=j90-cray
290	350	basic_os=unicos
291	351	;;
292	352	crds \| unos)
		@@ -705,13 +765,21 @@
705	765	decsystem20* \| dec20*)
706	766	cpu=pdp10
707	767	vendor=dec
708	768	basic_os=tops20
709	769	;;
710		- delta \| 3300 \| motorola-3300 \| motorola-delta \
711		- \| 3300-motorola \| delta-motorola)
	770	+ delta \| 3300 \| delta-motorola \| 3300-motorola \| motorola-delta \| motorola-3300)
712	771	cpu=m68k
713	772	vendor=motorola
714	773	;;
715		- dpx2*)
	774	+ # This used to be dpx2*, but that gets the RS6000-based
	775	+ # DPX/20 and the x86-based DPX/2-100 wrong. See
	776	+ # https://oldskool.silicium.org/stations/bull_dpx20.htm
	777	+ # https://www.feb-patrimoine.com/english/bull_dpx2.htm
	778	+ # https://www.feb-patrimoine.com/english/unix_and_bull.htm
	779	+ dpx2 \| dpx2[23]00 \| dpx2[23]xx)
716	780	cpu=m68k
717	781	vendor=bull
	782	+ ;;
	783	+ dpx2100 \| dpx21xx)
	784	+ cpu=i386
	785	+ vendor=bull
		@@ -718,6 +786,9 @@
718		- basic_os=sysv3
	786	+ ;;
	787	+ dpx20)
	788	+ cpu=rs6000
	789	+ vendor=bull
719	790	;;
720	791	encore \| umax \| mmax)
721	792	cpu=ns32k
722	793	vendor=encore
723	794	;;
		@@ -828,22 +899,10 @@
828	899	basic_os=newsos
829	900	;;
830	901	next \| m*-next)
831	902	cpu=m68k
832	903	vendor=next
833		- case $basic_os in
834		- openstep*)
835		- ;;
836		- nextstep*)
837		- ;;
838		- ns2*)
839		- basic_os=nextstep2
840		- ;;
841		- *)
842		- basic_os=nextstep3
843		- ;;
844		- esac
845	904	;;
846	905	np1)
847	906	cpu=np1
848	907	vendor=gould
849	908	;;
		@@ -928,16 +987,17 @@
928	987	cpu=sparc
929	988	vendor=`echo "$basic_machine" \| sed 's/-.*//'`
930	989	;;
931	990
932	991	-)
933		- # shellcheck disable=SC2162
	992	+ saved_IFS=$IFS
934	993	IFS="-" read cpu vendor <<EOF
935	994	$basic_machine
936	995	EOF
	996	+ IFS=$saved_IFS
937	997	;;
938		- # We use `pc' rather than `unknown'
	998	+ # We use 'pc' rather than 'unknown'
939	999	# because (1) that's what they normally are, and
940	1000	# (2) the word "unknown" tends to confuse beginning users.
941	1001	i*86 \| x86_64)
942	1002	cpu=$basic_machine
943	1003	vendor=pc
		@@ -961,19 +1021,23 @@
961	1021
962	1022	unset -v basic_machine
963	1023
964	1024	# Decode basic machines in the full and proper CPU-Company form.
965	1025	case $cpu-$vendor in
966		- # Here we handle the default manufacturer of certain CPU types in canonical form. It is in
967		- # some cases the only manufacturer, in others, it is the most popular.
	1026	+ # Here we handle the default manufacturer of certain CPU types in canonical form.
	1027	+ # It is in some cases the only manufacturer, in others, it is the most popular.
	1028	+ c[12]-convex \| c[12]-unknown \| c3[248]-convex \| c3[248]-unknown)
	1029	+ vendor=convex
	1030	+ basic_os=${basic_os:-bsd}
	1031	+ ;;
968	1032	craynv-unknown)
969	1033	vendor=cray
970	1034	basic_os=${basic_os:-unicosmp}
971	1035	;;
972	1036	c90-unknown \| c90-cray)
973	1037	vendor=cray
974		- basic_os=${Basic_os:-unicos}
	1038	+ basic_os=${basic_os:-unicos}
975	1039	;;
976	1040	fx80-unknown)
977	1041	vendor=alliant
978	1042	;;
979	1043	romp-unknown)
		@@ -1010,23 +1074,46 @@
1010	1074	cpu=xps100
1011	1075	vendor=honeywell
1012	1076	;;
1013	1077
1014	1078	# Here we normalize CPU types with a missing or matching vendor
1015		- dpx20-unknown \| dpx20-bull)
1016		- cpu=rs6000
1017		- vendor=bull
	1079	+ armh-unknown \| armh-alt)
	1080	+ cpu=armv7l
	1081	+ vendor=alt
	1082	+ basic_os=${basic_os:-linux-gnueabihf}
	1083	+ ;;
	1084	+
	1085	+ # Normalized CPU+vendor pairs that imply an OS, if not otherwise specified
	1086	+ m68k-isi)
	1087	+ basic_os=${basic_os:-sysv}
	1088	+ ;;
	1089	+ m68k-sony)
	1090	+ basic_os=${basic_os:-newsos}
	1091	+ ;;
	1092	+ m68k-tektronix)
	1093	+ basic_os=${basic_os:-bsd}
	1094	+ ;;
	1095	+ m88k-harris)
	1096	+ basic_os=${basic_os:-sysv3}
	1097	+ ;;
	1098	+ i386-bull \| m68k-bull)
	1099	+ basic_os=${basic_os:-sysv3}
	1100	+ ;;
	1101	+ rs6000-bull)
1018	1102	basic_os=${basic_os:-bosx}
1019	1103	;;
	1104	+ mips-sni)
	1105	+ basic_os=${basic_os:-sysv4}
	1106	+ ;;
1020	1107
1021	1108	# Here we normalize CPU types irrespective of the vendor
1022	1109	amd64-*)
1023	1110	cpu=x86_64
1024	1111	;;
1025	1112	blackfin-*)
1026	1113	cpu=bfin
1027		- basic_os=linux
	1114	+ basic_os=${basic_os:-linux}
1028	1115	;;
1029	1116	c54x-*)
1030	1117	cpu=tic54x
1031	1118	;;
1032	1119	c55x-*)
		@@ -1045,37 +1132,34 @@
1045	1132	ms1-*)
1046	1133	cpu=mt
1047	1134	;;
1048	1135	m68knommu-*)
1049	1136	cpu=m68k
1050		- basic_os=linux
	1137	+ basic_os=${basic_os:-linux}
1051	1138	;;
1052	1139	m9s12z-* \| m68hcs12z-* \| hcs12z-* \| s12z-*)
1053	1140	cpu=s12z
1054	1141	;;
1055	1142	openrisc-*)
1056	1143	cpu=or32
1057	1144	;;
1058	1145	parisc-*)
1059	1146	cpu=hppa
1060		- basic_os=linux
	1147	+ basic_os=${basic_os:-linux}
1061	1148	;;
1062	1149	pentium-* \| p5-* \| k5-* \| k6-* \| nexgen-* \| viac3-*)
1063	1150	cpu=i586
1064	1151	;;
1065		- pentiumpro-* \| p6-* \| 6x86-* \| athlon-* \| athalon_-)
	1152	+ pentiumpro-* \| p6-* \| 6x86-* \| athlon-* \| athlon_-)
1066	1153	cpu=i686
1067	1154	;;
1068	1155	pentiumii-* \| pentium2-* \| pentiumiii-* \| pentium3-*)
1069	1156	cpu=i686
1070	1157	;;
1071	1158	pentium4-*)
1072	1159	cpu=i786
1073	1160	;;
1074		- pc98-*)
1075		- cpu=i386
1076		- ;;
1077	1161	ppc-* \| ppcbe-*)
1078	1162	cpu=powerpc
1079	1163	;;
1080	1164	ppcle-* \| powerpclittle-*)
1081	1165	cpu=powerpcle
		@@ -1105,17 +1189,14 @@
1105	1189	cpu=mipstx39
1106	1190	;;
1107	1191	tx39el-*)
1108	1192	cpu=mipstx39el
1109	1193	;;
1110		- x64-*)
1111		- cpu=x86_64
1112		- ;;
1113	1194	xscale-* \| xscalee[bl]-*)
1114	1195	cpu=`echo "$cpu" \| sed 's/^xscale/arm/'`
1115	1196	;;
1116		- arm64-*)
	1197	+ arm64-* \| aarch64le-*)
1117	1198	cpu=aarch64
1118	1199	;;
1119	1200
1120	1201	# Recognize the canonical CPU Types that limit and/or modify the
1121	1202	# company names they are paired with.
		@@ -1163,118 +1244,235 @@
1163	1244
1164	1245	*)
1165	1246	# Recognize the canonical CPU types that are allowed with any
1166	1247	# company name.
1167	1248	case $cpu in
1168		- 1750a \| 580 \
	1249	+ 1750a \
	1250	+ \| 580 \
	1251	+ \| [cjt]90 \
1169	1252	\| a29k \
1170		- \| aarch64 \| aarch64_be \
	1253	+ \| aarch64 \
	1254	+ \| aarch64_be \
	1255	+ \| aarch64c \
1171	1256	\| abacus \
1172		- \| alpha \| alphaev[4-8] \| alphaev56 \| alphaev6[78] \
1173		- \| alpha64 \| alpha64ev[4-8] \| alpha64ev56 \| alpha64ev6[78] \
1174		- \| alphapca5[67] \| alpha64pca5[67] \
	1257	+ \| alpha \
	1258	+ \| alpha64 \
	1259	+ \| alpha64ev56 \
	1260	+ \| alpha64ev6[78] \
	1261	+ \| alpha64ev[4-8] \
	1262	+ \| alpha64pca5[67] \
	1263	+ \| alphaev56 \
	1264	+ \| alphaev6[78] \
	1265	+ \| alphaev[4-8] \
	1266	+ \| alphapca5[67] \
1175	1267	\| am33_2.0 \
1176	1268	\| amdgcn \
1177		- \| arc \| arceb \| arc32 \| arc64 \
1178		- \| arm \| arm[lb]e \| arme[lb] \| armv* \
1179		- \| avr \| avr32 \
	1269	+ \| arc \
	1270	+ \| arc32 \
	1271	+ \| arc64 \
	1272	+ \| arceb \
	1273	+ \| arm \
	1274	+ \| arm64e \
	1275	+ \| arm64ec \
	1276	+ \| arm[lb]e \
	1277	+ \| arme[lb] \
	1278	+ \| armv* \
1180	1279	\| asmjs \
	1280	+ \| avr \
	1281	+ \| avr32 \
1181	1282	\| ba \
1182		- \| be32 \| be64 \
1183		- \| bfin \| bpf \| bs2000 \
1184		- \| c[123]* \| c30 \| [cjt]90 \| c4x \
1185		- \| c8051 \| clipper \| craynv \| csky \| cydra \
1186		- \| d10v \| d30v \| dlx \| dsp16xx \
1187		- \| e2k \| elxsi \| epiphany \
1188		- \| f30[01] \| f700 \| fido \| fr30 \| frv \| ft32 \| fx80 \
1189		- \| h8300 \| h8500 \
1190		- \| hppa \| hppa1.[01] \| hppa2.0 \| hppa2.0[nw] \| hppa64 \
	1283	+ \| be32 \
	1284	+ \| be64 \
	1285	+ \| bfin \
	1286	+ \| bpf \
	1287	+ \| bs2000 \
	1288	+ \| c30 \
	1289	+ \| c4x \
	1290	+ \| c8051 \
	1291	+ \| c[123]* \
	1292	+ \| clipper \
	1293	+ \| craynv \
	1294	+ \| csky \
	1295	+ \| cydra \
	1296	+ \| d10v \
	1297	+ \| d30v \
	1298	+ \| dlx \
	1299	+ \| dsp16xx \
	1300	+ \| e2k \
	1301	+ \| elxsi \
	1302	+ \| epiphany \
	1303	+ \| f30[01] \
	1304	+ \| f700 \
	1305	+ \| fido \
	1306	+ \| fr30 \
	1307	+ \| frv \
	1308	+ \| ft32 \
	1309	+ \| fx80 \
	1310	+ \| h8300 \
	1311	+ \| h8500 \
1191	1312	\| hexagon \
1192		- \| i370 \| i*86 \| i860 \| i960 \| ia16 \| ia64 \
1193		- \| ip2k \| iq2000 \
	1313	+ \| hppa \
	1314	+ \| hppa1.[01] \
	1315	+ \| hppa2.0 \
	1316	+ \| hppa2.0[nw] \
	1317	+ \| hppa64 \
	1318	+ \| i*86 \
	1319	+ \| i370 \
	1320	+ \| i860 \
	1321	+ \| i960 \
	1322	+ \| ia16 \
	1323	+ \| ia64 \
	1324	+ \| ip2k \
	1325	+ \| iq2000 \
	1326	+ \| javascript \
1194	1327	\| k1om \
1195		- \| le32 \| le64 \
	1328	+ \| kvx \
	1329	+ \| le32 \
	1330	+ \| le64 \
1196	1331	\| lm32 \
1197		- \| loongarch32 \| loongarch64 \| loongarchx32 \
1198		- \| m32c \| m32r \| m32rle \
1199		- \| m5200 \| m68000 \| m680[012346]0 \| m68360 \| m683?2 \| m68k \
1200		- \| m6811 \| m68hc11 \| m6812 \| m68hc12 \| m68hcs12x \
1201		- \| m88110 \| m88k \| maxq \| mb \| mcore \| mep \| metag \
1202		- \| microblaze \| microblazeel \
1203		- \| mips \| mipsbe \| mipseb \| mipsel \| mipsle \
1204		- \| mips16 \
1205		- \| mips64 \| mips64eb \| mips64el \
1206		- \| mips64octeon \| mips64octeonel \
1207		- \| mips64orion \| mips64orionel \
1208		- \| mips64r5900 \| mips64r5900el \
1209		- \| mips64vr \| mips64vrel \
1210		- \| mips64vr4100 \| mips64vr4100el \
1211		- \| mips64vr4300 \| mips64vr4300el \
1212		- \| mips64vr5000 \| mips64vr5000el \
1213		- \| mips64vr5900 \| mips64vr5900el \
1214		- \| mipsisa32 \| mipsisa32el \
1215		- \| mipsisa32r2 \| mipsisa32r2el \
1216		- \| mipsisa32r3 \| mipsisa32r3el \
1217		- \| mipsisa32r5 \| mipsisa32r5el \
1218		- \| mipsisa32r6 \| mipsisa32r6el \
1219		- \| mipsisa64 \| mipsisa64el \
1220		- \| mipsisa64r2 \| mipsisa64r2el \
1221		- \| mipsisa64r3 \| mipsisa64r3el \
1222		- \| mipsisa64r5 \| mipsisa64r5el \
1223		- \| mipsisa64r6 \| mipsisa64r6el \
1224		- \| mipsisa64sb1 \| mipsisa64sb1el \
1225		- \| mipsisa64sr71k \| mipsisa64sr71kel \
1226		- \| mipsr5900 \| mipsr5900el \
1227		- \| mipstx39 \| mipstx39el \
	1332	+ \| loongarch32 \
	1333	+ \| loongarch64 \
	1334	+ \| m32c \
	1335	+ \| m32r \
	1336	+ \| m32rle \
	1337	+ \| m5200 \
	1338	+ \| m68000 \
	1339	+ \| m680[012346]0 \
	1340	+ \| m6811 \
	1341	+ \| m6812 \
	1342	+ \| m68360 \
	1343	+ \| m683?2 \
	1344	+ \| m68hc11 \
	1345	+ \| m68hc12 \
	1346	+ \| m68hcs12x \
	1347	+ \| m68k \
	1348	+ \| m88110 \
	1349	+ \| m88k \
	1350	+ \| maxq \
	1351	+ \| mb \
	1352	+ \| mcore \
	1353	+ \| mep \
	1354	+ \| metag \
	1355	+ \| microblaze \
	1356	+ \| microblazeel \
	1357	+ \| mips* \
1228	1358	\| mmix \
1229		- \| mn10200 \| mn10300 \
	1359	+ \| mn10200 \
	1360	+ \| mn10300 \
1230	1361	\| moxie \
1231		- \| mt \
1232	1362	\| msp430 \
1233		- \| nds32 \| nds32le \| nds32be \
	1363	+ \| mt \
	1364	+ \| nanomips* \
	1365	+ \| nds32 \
	1366	+ \| nds32be \
	1367	+ \| nds32le \
1234	1368	\| nfp \
1235		- \| nios \| nios2 \| nios2eb \| nios2el \
1236		- \| none \| np1 \| ns16k \| ns32k \| nvptx \
	1369	+ \| nios \
	1370	+ \| nios2 \
	1371	+ \| nios2eb \
	1372	+ \| nios2el \
	1373	+ \| none \
	1374	+ \| np1 \
	1375	+ \| ns16k \
	1376	+ \| ns32k \
	1377	+ \| nvptx \
1237	1378	\| open8 \
1238	1379	\| or1k* \
1239	1380	\| or32 \
1240	1381	\| orion \
	1382	+ \| pdp10 \
	1383	+ \| pdp11 \
1241	1384	\| picochip \
1242		- \| pdp10 \| pdp11 \| pj \| pjl \| pn \| power \
1243		- \| powerpc \| powerpc64 \| powerpc64le \| powerpcle \| powerpcspe \
	1385	+ \| pj \
	1386	+ \| pjl \
	1387	+ \| pn \
	1388	+ \| power \
	1389	+ \| powerpc \
	1390	+ \| powerpc64 \
	1391	+ \| powerpc64le \
	1392	+ \| powerpcle \
	1393	+ \| powerpcspe \
1244	1394	\| pru \
1245	1395	\| pyramid \
1246		- \| riscv \| riscv32 \| riscv32be \| riscv64 \| riscv64be \
1247		- \| rl78 \| romp \| rs6000 \| rx \
1248		- \| s390 \| s390x \
	1396	+ \| riscv \
	1397	+ \| riscv32 \
	1398	+ \| riscv32be \
	1399	+ \| riscv64 \
	1400	+ \| riscv64be \
	1401	+ \| rl78 \
	1402	+ \| romp \
	1403	+ \| rs6000 \
	1404	+ \| rx \
	1405	+ \| s390 \
	1406	+ \| s390x \
1249	1407	\| score \
1250		- \| sh \| shl \
1251		- \| sh[1234] \| sh[24]a \| sh[24]ae[lb] \| sh[23]e \| she[lb] \| sh[lb]e \
1252		- \| sh[1234]e[lb] \| sh[12345][lb]e \| sh[23]ele \| sh64 \| sh64le \
1253		- \| sparc \| sparc64 \| sparc64b \| sparc64v \| sparc86x \| sparclet \
	1408	+ \| sh \
	1409	+ \| sh64 \
	1410	+ \| sh64le \
	1411	+ \| sh[12345][lb]e \
	1412	+ \| sh[1234] \
	1413	+ \| sh[1234]e[lb] \
	1414	+ \| sh[23]e \
	1415	+ \| sh[23]ele \
	1416	+ \| sh[24]a \
	1417	+ \| sh[24]ae[lb] \
	1418	+ \| sh[lb]e \
	1419	+ \| she[lb] \
	1420	+ \| shl \
	1421	+ \| sparc \
	1422	+ \| sparc64 \
	1423	+ \| sparc64b \
	1424	+ \| sparc64v \
	1425	+ \| sparc86x \
	1426	+ \| sparclet \
1254	1427	\| sparclite \
1255		- \| sparcv8 \| sparcv9 \| sparcv9b \| sparcv9v \| sv1 \| sx* \
	1428	+ \| sparcv8 \
	1429	+ \| sparcv9 \
	1430	+ \| sparcv9b \
	1431	+ \| sparcv9v \
1256	1432	\| spu \
	1433	+ \| sv1 \
	1434	+ \| sx* \
1257	1435	\| tahoe \
1258	1436	\| thumbv7* \
1259		- \| tic30 \| tic4x \| tic54x \| tic55x \| tic6x \| tic80 \
	1437	+ \| tic30 \
	1438	+ \| tic4x \
	1439	+ \| tic54x \
	1440	+ \| tic55x \
	1441	+ \| tic6x \
	1442	+ \| tic80 \
1260	1443	\| tron \
1261	1444	\| ubicom32 \
1262		- \| v70 \| v850 \| v850e \| v850e1 \| v850es \| v850e2 \| v850e2v3 \
	1445	+ \| v70 \
	1446	+ \| v810 \
	1447	+ \| v850 \
	1448	+ \| v850e \
	1449	+ \| v850e1 \
	1450	+ \| v850e2 \
	1451	+ \| v850e2v3 \
	1452	+ \| v850es \
1263	1453	\| vax \
	1454	+ \| vc4 \
1264	1455	\| visium \
1265	1456	\| w65 \
1266		- \| wasm32 \| wasm64 \
	1457	+ \| wasm32 \
	1458	+ \| wasm64 \
1267	1459	\| we32k \
1268		- \| x86 \| x86_64 \| xc16x \| xgate \| xps100 \
1269		- \| xstormy16 \| xtensa* \
	1460	+ \| x86 \
	1461	+ \| x86_64 \
	1462	+ \| xc16x \
	1463	+ \| xgate \
	1464	+ \| xps100 \
	1465	+ \| xstormy16 \
	1466	+ \| xtensa* \
1270	1467	\| ymp \
1271		- \| z8k \| z80)
	1468	+ \| z80 \
	1469	+ \| z8k)
1272	1470	;;
1273	1471
1274	1472	*)
1275		- echo Invalid configuration \`"$1"\': machine \`"$cpu-$vendor"\' not recognized 1>&2
	1473	+ echo "Invalid configuration '$1': machine '$cpu-$vendor' not recognized" 1>&2
1276	1474	exit 1
1277	1475	;;
1278	1476	esac
1279	1477	;;
1280	1478	esac
		@@ -1291,15 +1489,16 @@
1291	1489	;;
1292	1490	esac
1293	1491
1294	1492	# Decode manufacturer-specific aliases for certain operating systems.
1295	1493
1296		-if test x$basic_os != x
	1494	+if test x"$basic_os" != x
1297	1495	then
1298	1496
1299		-# First recognize some ad-hoc caes, or perhaps split kernel-os, or else just
	1497	+# First recognize some ad-hoc cases, or perhaps split kernel-os, or else just
1300	1498	# set os.
	1499	+obj=
1301	1500	case $basic_os in
1302	1501	gnu/linux*)
1303	1502	kernel=linux
1304	1503	os=`echo "$basic_os" \| sed -e 's\|gnu/linux\|gnu\|'`
1305	1504	;;
		@@ -1310,14 +1509,15 @@
1310	1509	nto-qnx*)
1311	1510	kernel=nto
1312	1511	os=`echo "$basic_os" \| sed -e 's\|nto-qnx\|qnx\|'`
1313	1512	;;
1314	1513	-)
1315		- # shellcheck disable=SC2162
	1514	+ saved_IFS=$IFS
1316	1515	IFS="-" read kernel os <<EOF
1317	1516	$basic_os
1318	1517	EOF
	1518	+ IFS=$saved_IFS
1319	1519	;;
1320	1520	# Default OS when just kernel was specified
1321	1521	nto*)
1322	1522	kernel=nto
1323	1523	os=`echo "$basic_os" \| sed -e 's\|nto\|qnx\|'`
		@@ -1324,10 +1524,14 @@
1324	1524	;;
1325	1525	linux*)
1326	1526	kernel=linux
1327	1527	os=`echo "$basic_os" \| sed -e 's\|linux\|gnu\|'`
1328	1528	;;
	1529	+ managarm*)
	1530	+ kernel=managarm
	1531	+ os=`echo "$basic_os" \| sed -e 's\|managarm\|mlibc\|'`
	1532	+ ;;
1329	1533	*)
1330	1534	kernel=
1331	1535	os=$basic_os
1332	1536	;;
1333	1537	esac
		@@ -1350,10 +1554,27 @@
1350	1554	solaris)
1351	1555	os=solaris2
1352	1556	;;
1353	1557	unixware*)
1354	1558	os=sysv4.2uw
	1559	+ ;;
	1560	+ # The marketing names for NeXT's operating systems were
	1561	+ # NeXTSTEP, NeXTSTEP 2, OpenSTEP 3, OpenSTEP 4. 'openstep' is
	1562	+ # mapped to 'openstep3', but 'openstep1' and 'openstep2' are
	1563	+ # mapped to 'nextstep' and 'nextstep2', consistent with the
	1564	+ # treatment of SunOS/Solaris.
	1565	+ ns \| ns1 \| nextstep \| nextstep1 \| openstep1)
	1566	+ os=nextstep
	1567	+ ;;
	1568	+ ns2 \| nextstep2 \| openstep2)
	1569	+ os=nextstep2
	1570	+ ;;
	1571	+ ns3 \| nextstep3 \| openstep \| openstep3)
	1572	+ os=openstep3
	1573	+ ;;
	1574	+ ns4 \| nextstep4 \| openstep4)
	1575	+ os=openstep4
1355	1576	;;
1356	1577	# es1800 is here to avoid being matched by es* (a different OS)
1357	1578	es1800*)
1358	1579	os=ose
1359	1580	;;
		@@ -1421,10 +1642,11 @@
1421	1642	wince*)
1422	1643	os=wince
1423	1644	;;
1424	1645	utek*)
1425	1646	os=bsd
	1647	+ vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|tektronix\|'`
1426	1648	;;
1427	1649	dynix*)
1428	1650	os=bsd
1429	1651	;;
1430	1652	acis*)
		@@ -1437,25 +1659,29 @@
1437	1659	os=syllable
1438	1660	;;
1439	1661	386bsd)
1440	1662	os=bsd
1441	1663	;;
1442		- ctix* \| uts*)
	1664	+ ctix*)
	1665	+ os=sysv
	1666	+ vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|convergent\|'`
	1667	+ ;;
	1668	+ uts*)
1443	1669	os=sysv
1444	1670	;;
1445	1671	nova*)
1446		- os=rtmk-nova
1447		- ;;
1448		- ns2)
1449		- os=nextstep2
	1672	+ kernel=rtmk
	1673	+ os=nova
1450	1674	;;
1451	1675	# Preserve the version number of sinix5.
1452	1676	sinix5.*)
1453	1677	os=`echo "$os" \| sed -e 's\|sinix\|sysv\|'`
	1678	+ vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|sni\|'`
1454	1679	;;
1455	1680	sinix*)
1456	1681	os=sysv4
	1682	+ vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|sni\|'`
1457	1683	;;
1458	1684	tpf*)
1459	1685	os=tpf
1460	1686	;;
1461	1687	triton*)
		@@ -1489,13 +1715,19 @@
1489	1715	case $cpu in
1490	1716	arm*)
1491	1717	os=eabi
1492	1718	;;
1493	1719	*)
1494		- os=elf
	1720	+ os=
	1721	+ obj=elf
1495	1722	;;
1496	1723	esac
	1724	+ ;;
	1725	+ aout* \| coff* \| elf* \| pe*)
	1726	+ # These are machine code file formats, not OSes
	1727	+ obj=$os
	1728	+ os=
1497	1729	;;
1498	1730	*)
1499	1731	# No normalization, but not necessarily accepted, that comes below.
1500	1732	;;
1501	1733	esac
		@@ -1511,47 +1743,57 @@
1511	1743	# that MANUFACTURER isn't an operating system. Otherwise, code above
1512	1744	# will signal an error saying that MANUFACTURER isn't an operating
1513	1745	# system, and we'll never get to this point.
1514	1746
1515	1747	kernel=
	1748	+obj=
1516	1749	case $cpu-$vendor in
1517	1750	score-*)
1518		- os=elf
	1751	+ os=
	1752	+ obj=elf
1519	1753	;;
1520	1754	spu-*)
1521		- os=elf
	1755	+ os=
	1756	+ obj=elf
1522	1757	;;
1523	1758	*-acorn)
1524	1759	os=riscix1.2
1525	1760	;;
1526	1761	arm*-rebel)
1527	1762	kernel=linux
1528	1763	os=gnu
1529	1764	;;
1530	1765	arm*-semi)
1531		- os=aout
	1766	+ os=
	1767	+ obj=aout
1532	1768	;;
1533	1769	c4x-* \| tic4x-*)
1534		- os=coff
	1770	+ os=
	1771	+ obj=coff
1535	1772	;;
1536	1773	c8051-*)
1537		- os=elf
	1774	+ os=
	1775	+ obj=elf
1538	1776	;;
1539	1777	clipper-intergraph)
1540	1778	os=clix
1541	1779	;;
1542	1780	hexagon-*)
1543		- os=elf
	1781	+ os=
	1782	+ obj=elf
1544	1783	;;
1545	1784	tic54x-*)
1546		- os=coff
	1785	+ os=
	1786	+ obj=coff
1547	1787	;;
1548	1788	tic55x-*)
1549		- os=coff
	1789	+ os=
	1790	+ obj=coff
1550	1791	;;
1551	1792	tic6x-*)
1552		- os=coff
	1793	+ os=
	1794	+ obj=coff
1553	1795	;;
1554	1796	# This must come before the *-dec entry.
1555	1797	pdp10-*)
1556	1798	os=tops20
1557	1799	;;
		@@ -1569,32 +1811,47 @@
1569	1811	;;
1570	1812	m68000-sun)
1571	1813	os=sunos3
1572	1814	;;
1573	1815	m68*-cisco)
1574		- os=aout
	1816	+ os=
	1817	+ obj=aout
1575	1818	;;
1576	1819	mep-*)
1577		- os=elf
	1820	+ os=
	1821	+ obj=elf
	1822	+ ;;
	1823	+ # The -sgi and -siemens entries must be before the mips- entry
	1824	+ # or we get the wrong os.
	1825	+ *-sgi)
	1826	+ os=irix
	1827	+ ;;
	1828	+ *-siemens)
	1829	+ os=sysv4
1578	1830	;;
1579	1831	mips*-cisco)
1580		- os=elf
	1832	+ os=
	1833	+ obj=elf
1581	1834	;;
1582		- mips-)
1583		- os=elf
	1835	+ mips-\|nanomips-)
	1836	+ os=
	1837	+ obj=elf
1584	1838	;;
1585	1839	or32-*)
1586		- os=coff
	1840	+ os=
	1841	+ obj=coff
1587	1842	;;
1588		- *-tti) # must be before sparc entry or we get the wrong os.
	1843	+ # This must be before the sparc-* entry or we get the wrong os.
	1844	+ *-tti)
1589	1845	os=sysv3
1590	1846	;;
1591	1847	sparc-* \| *-sun)
1592	1848	os=sunos4.1.1
1593	1849	;;
1594	1850	pru-*)
1595		- os=elf
	1851	+ os=
	1852	+ obj=elf
1596	1853	;;
1597	1854	*-be)
1598	1855	os=beos
1599	1856	;;
1600	1857	*-ibm)
		@@ -1614,11 +1871,11 @@
1614	1871	;;
1615	1872	*-hp)
1616	1873	os=hpux
1617	1874	;;
1618	1875	*-hitachi)
1619		- os=hiux
	1876	+ os=hiuxwe2
1620	1877	;;
1621	1878	i860-* \| -att \| -ncr \| -altos \| -motorola \| *-convergent)
1622	1879	os=sysv
1623	1880	;;
1624	1881	*-cbm)
		@@ -1658,27 +1915,23 @@
1658	1915	os=bsd
1659	1916	;;
1660	1917	*-encore)
1661	1918	os=bsd
1662	1919	;;
1663		- *-sgi)
1664		- os=irix
1665		- ;;
1666		- *-siemens)
1667		- os=sysv4
1668		- ;;
1669	1920	*-masscomp)
1670	1921	os=rtu
1671	1922	;;
1672	1923	f30[01]-fujitsu \| f700-fujitsu)
1673	1924	os=uxpv
1674	1925	;;
1675	1926	*-rom68k)
1676		- os=coff
	1927	+ os=
	1928	+ obj=coff
1677	1929	;;
1678	1930	-bug)
1679		- os=coff
	1931	+ os=
	1932	+ obj=coff
1680	1933	;;
1681	1934	*-apple)
1682	1935	os=macos
1683	1936	;;
1684	1937	-atari)
		@@ -1692,96 +1945,324 @@
1692	1945	;;
1693	1946	esac
1694	1947
1695	1948	fi
1696	1949
1697		-# Now, validate our (potentially fixed-up) OS.
	1950	+# Now, validate our (potentially fixed-up) individual pieces (OS, OBJ).
	1951	+
1698	1952	case $os in
1699	1953	# Sometimes we do "kernel-libc", so those need to count as OSes.
1700		- musl* \| newlib* \| uclibc*)
	1954	+ llvm* \| musl* \| newlib* \| relibc* \| uclibc*)
1701	1955	;;
1702	1956	# Likewise for "kernel-abi"
1703	1957	eabi* \| gnueabi*)
1704	1958	;;
1705	1959	# VxWorks passes extra cpu info in the 4th filed.
1706	1960	simlinux \| simwindows \| spe)
	1961	+ ;;
	1962	+ # See `case $cpu-$os` validation below
	1963	+ ghcjs)
1707	1964	;;
1708	1965	# Now accept the basic system types.
1709		- # The portable systems comes first.
1710	1966	# Each alternative MUST end in a * to match a version number.
1711		- gnu* \| android* \| bsd* \| mach* \| minix* \| genix* \| ultrix* \| irix* \
1712		- \| vms \| esix* \| aix* \| cnk* \| sunos \| sunos[34]* \
1713		- \| hpux* \| unos* \| osf* \| luna* \| dgux* \| auroraux* \| solaris* \
1714		- \| sym* \| plan9* \| psp* \| sim* \| xray* \| os68k* \| v88r* \
1715		- \| hiux* \| abug \| nacl* \| netware* \| windows* \
1716		- \| os9* \| macos* \| osx* \| ios* \
1717		- \| mpw* \| magic* \| mmixware* \| mon960* \| lnews* \
1718		- \| amigaos* \| amigados* \| msdos* \| newsos* \| unicos* \| aof* \
1719		- \| aos* \| aros* \| cloudabi* \| sortix* \| twizzler* \
1720		- \| nindy* \| vxsim* \| vxworks* \| ebmon* \| hms* \| mvs* \
1721		- \| clix* \| riscos* \| uniplus* \| iris* \| isc* \| rtu* \| xenix* \
1722		- \| mirbsd* \| netbsd* \| dicos* \| openedition* \| ose* \
1723		- \| bitrig* \| openbsd* \| secbsd* \| solidbsd* \| libertybsd* \| os108* \
1724		- \| ekkobsd* \| freebsd* \| riscix* \| lynxos* \| os400* \
1725		- \| bosx* \| nextstep* \| cxux* \| aout* \| elf* \| oabi* \
1726		- \| ptx* \| coff* \| ecoff* \| winnt* \| domain* \| vsta* \
1727		- \| udi* \| lites* \| ieee* \| go32* \| aux* \| hcos* \
1728		- \| chorusrdb* \| cegcc* \| glidix* \| serenity* \
1729		- \| cygwin* \| msys* \| pe* \| moss* \| proelf* \| rtems* \
1730		- \| midipix* \| mingw32* \| mingw64* \| mint* \
1731		- \| uxpv* \| beos* \| mpeix* \| udk* \| moxiebox* \
1732		- \| interix* \| uwin* \| mks* \| rhapsody* \| darwin* \
1733		- \| openstep* \| oskit* \| conix* \| pw32* \| nonstopux* \
1734		- \| storm-chaos* \| tops10* \| tenex* \| tops20* \| its* \
1735		- \| os2* \| vos* \| palmos* \| uclinux* \| nucleus* \| morphos* \
1736		- \| scout* \| superux* \| sysv* \| rtmk* \| tpf* \| windiss* \
1737		- \| powermax* \| dnix* \| nx6 \| nx7 \| sei* \| dragonfly* \
1738		- \| skyos* \| haiku* \| rdos* \| toppers* \| drops* \| es* \
1739		- \| onefs* \| tirtos* \| phoenix* \| fuchsia* \| redox* \| bme* \
1740		- \| midnightbsd* \| amdhsa* \| unleashed* \| emscripten* \| wasi* \
1741		- \| nsk* \| powerunix* \| genode* \| zvmoe* \| qnx* \| emx*)
	1967	+ abug \
	1968	+ \| aix* \
	1969	+ \| amdhsa* \
	1970	+ \| amigados* \
	1971	+ \| amigaos* \
	1972	+ \| android* \
	1973	+ \| aof* \
	1974	+ \| aos* \
	1975	+ \| aros* \
	1976	+ \| atheos* \
	1977	+ \| auroraux* \
	1978	+ \| aux* \
	1979	+ \| beos* \
	1980	+ \| bitrig* \
	1981	+ \| bme* \
	1982	+ \| bosx* \
	1983	+ \| bsd* \
	1984	+ \| cegcc* \
	1985	+ \| chorusos* \
	1986	+ \| chorusrdb* \
	1987	+ \| clix* \
	1988	+ \| cloudabi* \
	1989	+ \| cnk* \
	1990	+ \| conix* \
	1991	+ \| cos* \
	1992	+ \| cxux* \
	1993	+ \| cygwin* \
	1994	+ \| darwin* \
	1995	+ \| dgux* \
	1996	+ \| dicos* \
	1997	+ \| dnix* \
	1998	+ \| domain* \
	1999	+ \| dragonfly* \
	2000	+ \| drops* \
	2001	+ \| ebmon* \
	2002	+ \| ecoff* \
	2003	+ \| ekkobsd* \
	2004	+ \| emscripten* \
	2005	+ \| emx* \
	2006	+ \| es* \
	2007	+ \| fiwix* \
	2008	+ \| freebsd* \
	2009	+ \| fuchsia* \
	2010	+ \| genix* \
	2011	+ \| genode* \
	2012	+ \| glidix* \
	2013	+ \| gnu* \
	2014	+ \| go32* \
	2015	+ \| haiku* \
	2016	+ \| hcos* \
	2017	+ \| hiux* \
	2018	+ \| hms* \
	2019	+ \| hpux* \
	2020	+ \| ieee* \
	2021	+ \| interix* \
	2022	+ \| ios* \
	2023	+ \| iris* \
	2024	+ \| irix* \
	2025	+ \| ironclad* \
	2026	+ \| isc* \
	2027	+ \| its* \
	2028	+ \| l4re* \
	2029	+ \| libertybsd* \
	2030	+ \| lites* \
	2031	+ \| lnews* \
	2032	+ \| luna* \
	2033	+ \| lynxos* \
	2034	+ \| mach* \
	2035	+ \| macos* \
	2036	+ \| magic* \
	2037	+ \| mbr* \
	2038	+ \| midipix* \
	2039	+ \| midnightbsd* \
	2040	+ \| mingw32* \
	2041	+ \| mingw64* \
	2042	+ \| minix* \
	2043	+ \| mint* \
	2044	+ \| mirbsd* \
	2045	+ \| mks* \
	2046	+ \| mlibc* \
	2047	+ \| mmixware* \
	2048	+ \| mon960* \
	2049	+ \| morphos* \
	2050	+ \| moss* \
	2051	+ \| moxiebox* \
	2052	+ \| mpeix* \
	2053	+ \| mpw* \
	2054	+ \| msdos* \
	2055	+ \| msys* \
	2056	+ \| mvs* \
	2057	+ \| nacl* \
	2058	+ \| netbsd* \
	2059	+ \| netware* \
	2060	+ \| newsos* \
	2061	+ \| nextstep* \
	2062	+ \| nindy* \
	2063	+ \| nonstopux* \
	2064	+ \| nova* \
	2065	+ \| nsk* \
	2066	+ \| nucleus* \
	2067	+ \| nx6 \
	2068	+ \| nx7 \
	2069	+ \| oabi* \
	2070	+ \| ohos* \
	2071	+ \| onefs* \
	2072	+ \| openbsd* \
	2073	+ \| openedition* \
	2074	+ \| openstep* \
	2075	+ \| os108* \
	2076	+ \| os2* \
	2077	+ \| os400* \
	2078	+ \| os68k* \
	2079	+ \| os9* \
	2080	+ \| ose* \
	2081	+ \| osf* \
	2082	+ \| oskit* \
	2083	+ \| osx* \
	2084	+ \| palmos* \
	2085	+ \| phoenix* \
	2086	+ \| plan9* \
	2087	+ \| powermax* \
	2088	+ \| powerunix* \
	2089	+ \| proelf* \
	2090	+ \| psos* \
	2091	+ \| psp* \
	2092	+ \| ptx* \
	2093	+ \| pw32* \
	2094	+ \| qnx* \
	2095	+ \| rdos* \
	2096	+ \| redox* \
	2097	+ \| rhapsody* \
	2098	+ \| riscix* \
	2099	+ \| riscos* \
	2100	+ \| rtems* \
	2101	+ \| rtmk* \
	2102	+ \| rtu* \
	2103	+ \| scout* \
	2104	+ \| secbsd* \
	2105	+ \| sei* \
	2106	+ \| serenity* \
	2107	+ \| sim* \
	2108	+ \| skyos* \
	2109	+ \| solaris* \
	2110	+ \| solidbsd* \
	2111	+ \| sortix* \
	2112	+ \| storm-chaos* \
	2113	+ \| sunos \
	2114	+ \| sunos[34]* \
	2115	+ \| superux* \
	2116	+ \| syllable* \
	2117	+ \| sym* \
	2118	+ \| sysv* \
	2119	+ \| tenex* \
	2120	+ \| tirtos* \
	2121	+ \| toppers* \
	2122	+ \| tops10* \
	2123	+ \| tops20* \
	2124	+ \| tpf* \
	2125	+ \| tvos* \
	2126	+ \| twizzler* \
	2127	+ \| uclinux* \
	2128	+ \| udi* \
	2129	+ \| udk* \
	2130	+ \| ultrix* \
	2131	+ \| unicos* \
	2132	+ \| uniplus* \
	2133	+ \| unleashed* \
	2134	+ \| unos* \
	2135	+ \| uwin* \
	2136	+ \| uxpv* \
	2137	+ \| v88r* \
	2138	+ \|vms \
	2139	+ \| vos* \
	2140	+ \| vsta* \
	2141	+ \| vxsim* \
	2142	+ \| vxworks* \
	2143	+ \| wasi* \
	2144	+ \| watchos* \
	2145	+ \| wince* \
	2146	+ \| windiss* \
	2147	+ \| windows* \
	2148	+ \| winnt* \
	2149	+ \| xenix* \
	2150	+ \| xray* \
	2151	+ \| zephyr* \
	2152	+ \| zvmoe* )
1742	2153	;;
1743	2154	# This one is extra strict with allowed versions
1744	2155	sco3.2v2 \| sco3.2v[4-9]* \| sco5v6*)
1745	2156	# Don't forget version if it is 3.2v4 or newer.
1746	2157	;;
	2158	+ # This refers to builds using the UEFI calling convention
	2159	+ # (which depends on the architecture) and PE file format.
	2160	+ # Note that this is both a different calling convention and
	2161	+ # different file format than that of GNU-EFI
	2162	+ # (x86_64-w64-mingw32).
	2163	+ uefi)
	2164	+ ;;
1747	2165	none)
1748	2166	;;
	2167	+ kernel* \| msvc* )
	2168	+ # Restricted further below
	2169	+ ;;
	2170	+ '')
	2171	+ if test x"$obj" = x
	2172	+ then
	2173	+ echo "Invalid configuration '$1': Blank OS only allowed with explicit machine code file format" 1>&2
	2174	+ fi
	2175	+ ;;
	2176	+ *)
	2177	+ echo "Invalid configuration '$1': OS '$os' not recognized" 1>&2
	2178	+ exit 1
	2179	+ ;;
	2180	+esac
	2181	+
	2182	+case $obj in
	2183	+ aout* \| coff* \| elf* \| pe*)
	2184	+ ;;
	2185	+ '')
	2186	+ # empty is fine
	2187	+ ;;
1749	2188	*)
1750		- echo Invalid configuration \`"$1"\': OS \`"$os"\' not recognized 1>&2
	2189	+ echo "Invalid configuration '$1': Machine code format '$obj' not recognized" 1>&2
	2190	+ exit 1
	2191	+ ;;
	2192	+esac
	2193	+
	2194	+# Here we handle the constraint that a (synthetic) cpu and os are
	2195	+# valid only in combination with each other and nowhere else.
	2196	+case $cpu-$os in
	2197	+ # The "javascript-unknown-ghcjs" triple is used by GHC; we
	2198	+ # accept it here in order to tolerate that, but reject any
	2199	+ # variations.
	2200	+ javascript-ghcjs)
	2201	+ ;;
	2202	+ javascript-* \| *-ghcjs)
	2203	+ echo "Invalid configuration '$1': cpu '$cpu' is not valid with os '$os$obj'" 1>&2
1751	2204	exit 1
1752	2205	;;
1753	2206	esac
1754	2207
1755	2208	# As a final step for OS-related things, validate the OS-kernel combination
1756	2209	# (given a valid OS), if there is a kernel.
1757		-case $kernel-$os in
1758		- linux-gnu* \| linux-dietlibc* \| linux-android* \| linux-newlib* \| linux-musl* \| linux-uclibc* )
	2210	+case $kernel-$os-$obj in
	2211	+ linux-gnu- \| linux-android- \| linux-dietlibc- \| linux-llvm- \
	2212	+ \| linux-mlibc- \| linux-musl- \| linux-newlib*- \
	2213	+ \| linux-relibc- \| linux-uclibc- \| linux-ohos*- )
	2214	+ ;;
	2215	+ uclinux-uclibc- \| uclinux-gnu- )
	2216	+ ;;
	2217	+ managarm-mlibc- \| managarm-kernel- )
1759	2218	;;
1760		- uclinux-uclibc* )
	2219	+ windows-msvc-)
1761	2220	;;
1762		- -dietlibc* \| -newlib* \| -musl* \| -uclibc* )
	2221	+ -dietlibc- \| -llvm- \| -mlibc- \| -musl- \| -newlib- \| -relibc- \
	2222	+ \| -uclibc*- )
1763	2223	# These are just libc implementations, not actual OSes, and thus
1764	2224	# require a kernel.
1765		- echo "Invalid configuration \`$1': libc \`$os' needs explicit kernel." 1>&2
	2225	+ echo "Invalid configuration '$1': libc '$os' needs explicit kernel." 1>&2
	2226	+ exit 1
	2227	+ ;;
	2228	+ -kernel*- )
	2229	+ echo "Invalid configuration '$1': '$os' needs explicit kernel." 1>&2
	2230	+ exit 1
	2231	+ ;;
	2232	+ -kernel- )
	2233	+ echo "Invalid configuration '$1': '$kernel' does not support '$os'." 1>&2
	2234	+ exit 1
	2235	+ ;;
	2236	+ -msvc- )
	2237	+ echo "Invalid configuration '$1': '$os' needs 'windows'." 1>&2
1766	2238	exit 1
1767	2239	;;
1768		- kfreebsd-gnu \| kopensolaris-gnu)
1769		- ;;
1770		- vxworks-simlinux \| vxworks-simwindows \| vxworks-spe)
1771		- ;;
1772		- nto-qnx*)
1773		- ;;
1774		- os2-emx)
1775		- ;;
1776		- -eabi \| -gnueabi)
1777		- ;;
1778		- -*)
	2240	+ kfreebsd-gnu- \| knetbsd-gnu- \| netbsd-gnu- \| kopensolaris-gnu-)
	2241	+ ;;
	2242	+ vxworks-simlinux- \| vxworks-simwindows- \| vxworks-spe-)
	2243	+ ;;
	2244	+ nto-qnx*-)
	2245	+ ;;
	2246	+ os2-emx-)
	2247	+ ;;
	2248	+ rtmk-nova-)
	2249	+ ;;
	2250	+ -eabi- \| -gnueabi-)
	2251	+ ;;
	2252	+ none--*)
	2253	+ # None (no kernel, i.e. freestanding / bare metal),
	2254	+ # can be paired with an machine code file format
	2255	+ ;;
	2256	+ -*-)
1779	2257	# Blank kernel with real OS is always fine.
1780	2258	;;
1781		- -)
1782		- echo "Invalid configuration \`$1': Kernel \`$kernel' not known to work with OS \`$os'." 1>&2
	2259	+ --*)
	2260	+ # Blank kernel and OS with real machine code file format is always fine.
	2261	+ ;;
	2262	+ --*)
	2263	+ echo "Invalid configuration '$1': Kernel '$kernel' not known to work with OS '$os'." 1>&2
1783	2264	exit 1
1784	2265	;;
1785	2266	esac
1786	2267
1787	2268	# Here we handle the case where we know the os, and the CPU type, but not the
		@@ -1790,11 +2271,11 @@
1790	2271	unknown)
1791	2272	case $cpu-$os in
1792	2273	-riscix)
1793	2274	vendor=acorn
1794	2275	;;
1795		- -sunos)
	2276	+ -sunos \| -solaris)
1796	2277	vendor=sun
1797	2278	;;
1798	2279	-cnk \| -aix)
1799	2280	vendor=ibm
1800	2281	;;
		@@ -1860,14 +2341,14 @@
1860	2341	;;
1861	2342	esac
1862	2343	;;
1863	2344	esac
1864	2345
1865		-echo "$cpu-$vendor-${kernel:+$kernel-}$os"
	2346	+echo "$cpu-$vendor${kernel:+-$kernel}${os:+-$os}${obj:+-$obj}"
1866	2347	exit
1867	2348
1868	2349	# Local variables:
1869	2350	# eval: (add-hook 'before-save-hook 'time-stamp)
1870	2351	# time-stamp-start: "timestamp='"
1871	2352	# time-stamp-format: "%:y-%02m-%02d"
1872	2353	# time-stamp-end: "'"
1873	2354	# End:
1874	2355

	--- autosetup/autosetup-config.sub
	+++ autosetup/autosetup-config.sub
	@@ -1,16 +1,16 @@
1	#! /bin/sh
2	# Configuration validation subroutine script.
3	# Copyright 1992-2021 Free Software Foundation, Inc.
4
5	# shellcheck disable=SC2006,SC2268 # see below for rationale
6
7	timestamp='2021-07-03'
8
9	# This file is free software; you can redistribute it and/or modify it
10	# under the terms of the GNU General Public License as published by
11	# the Free Software Foundation; either version 3 of the License, or
12	# (at your option) any later version.
13	#
14	# This program is distributed in the hope that it will be useful, but
15	# WITHOUT ANY WARRANTY; without even the implied warranty of
16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	@@ -74,17 +74,17 @@
74	Report bugs and patches to <[email protected]>."
75
76	version="\
77	GNU config.sub ($timestamp)
78
79	Copyright 1992-2021 Free Software Foundation, Inc.
80
81	This is free software; see the source for copying conditions. There is NO
82	warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
83
84	help="
85	Try \`$me --help' for more information."
86
87	# Parse command line
88	while test $# -gt 0 ; do
89	case $1 in
90	--time-stamp \| --time* \| -t )
	@@ -118,19 +118,20 @@
118	*) echo "$me: too many arguments$help" >&2
119	exit 1;;
120	esac
121
122	# Split fields of configuration type
123	# shellcheck disable=SC2162
124	IFS="-" read field1 field2 field3 field4 <<EOF
125	$1
126	EOF

127
128	# Separate into logical components for further validation
129	case $1 in
130	----*)
131	echo Invalid configuration \`"$1"\': more than four components >&2
132	exit 1
133	;;
134	---)
135	basic_machine=$field1-$field2
136	basic_os=$field3-$field4
	@@ -138,14 +139,25 @@
138	--*)
139	# Ambiguous whether COMPANY is present, or skipped and KERNEL-OS is two
140	# parts
141	maybe_os=$field2-$field3
142	case $maybe_os in
143	nto-qnx* \| linux-* \| uclinux-uclibc* \
144	\| uclinux-gnu* \| kfreebsd-gnu \| knetbsd-gnu \| netbsd-gnu \
145	\| netbsd-eabi \| kopensolaris-gnu \| cloudabi-eabi \
146	\| storm-chaos* \| os2-emx* \| rtmk-nova*)











147	basic_machine=$field1
148	basic_os=$maybe_os
149	;;
150	android-linux)
151	basic_machine=$field1-unknown
	@@ -156,37 +168,105 @@
156	basic_os=$field3
157	;;
158	esac
159	;;
160	-)
161	# A lone config we happen to match not fitting any pattern
162	case $field1-$field2 in





163	decstation-3100)
164	basic_machine=mips-dec
165	basic_os=
166	;;
167	-)
168	# Second component is usually, but not always the OS
169	case $field2 in
170	# Prevent following clause from handling this valid os
171	sunos)
172	basic_machine=$field1
173	basic_os=$field2
174	;;
175	# Manufacturers
176	dec* \| mips* \| sequent* \| encore* \| pc533* \| sgi* \| sony* \
177	\| att* \| 7300* \| 3300* \| delta* \| motorola* \| sun[234]* \
178	\| unicom* \| ibm* \| next \| hp \| isi* \| apollo \| altos* \
179	\| convergent* \| ncr* \| news \| 32* \| 3600* \| 3100* \
180	\| hitachi* \| c[123]* \| convex* \| sun \| crds \| omron* \| dg \
181	\| ultra \| tti* \| harris \| dolphin \| highlevel \| gould \
182	\| cbm \| ns \| masscomp \| apple \| axis \| knuth \| cray \
183	\| microblaze* \| sim \| cisco \
184	\| oki \| wec \| wrs \| winbond)




























































185	basic_machine=$field1-$field2
186	basic_os=
187	;;




188	*)
189	basic_machine=$field1
190	basic_os=$field2
191	;;
192	esac
	@@ -263,30 +343,10 @@
263	;;
264	cegcc)
265	basic_machine=arm-unknown
266	basic_os=cegcc
267	;;
268	convex-c1)
269	basic_machine=c1-convex
270	basic_os=bsd
271	;;
272	convex-c2)
273	basic_machine=c2-convex
274	basic_os=bsd
275	;;
276	convex-c32)
277	basic_machine=c32-convex
278	basic_os=bsd
279	;;
280	convex-c34)
281	basic_machine=c34-convex
282	basic_os=bsd
283	;;
284	convex-c38)
285	basic_machine=c38-convex
286	basic_os=bsd
287	;;
288	cray)
289	basic_machine=j90-cray
290	basic_os=unicos
291	;;
292	crds \| unos)
	@@ -705,13 +765,21 @@
705	decsystem20* \| dec20*)
706	cpu=pdp10
707	vendor=dec
708	basic_os=tops20
709	;;
710	delta \| 3300 \| motorola-3300 \| motorola-delta \
711	\| 3300-motorola \| delta-motorola)
712	cpu=m68k
713	vendor=motorola
714	;;
715	dpx2*)





716	cpu=m68k
717	vendor=bull




	@@ -718,6 +786,9 @@
718	basic_os=sysv3



719	;;
720	encore \| umax \| mmax)
721	cpu=ns32k
722	vendor=encore
723	;;
	@@ -828,22 +899,10 @@
828	basic_os=newsos
829	;;
830	next \| m*-next)
831	cpu=m68k
832	vendor=next
833	case $basic_os in
834	openstep*)
835	;;
836	nextstep*)
837	;;
838	ns2*)
839	basic_os=nextstep2
840	;;
841	*)
842	basic_os=nextstep3
843	;;
844	esac
845	;;
846	np1)
847	cpu=np1
848	vendor=gould
849	;;
	@@ -928,16 +987,17 @@
928	cpu=sparc
929	vendor=`echo "$basic_machine" \| sed 's/-.*//'`
930	;;
931
932	-)
933	# shellcheck disable=SC2162
934	IFS="-" read cpu vendor <<EOF
935	$basic_machine
936	EOF

937	;;
938	# We use `pc' rather than `unknown'
939	# because (1) that's what they normally are, and
940	# (2) the word "unknown" tends to confuse beginning users.
941	i*86 \| x86_64)
942	cpu=$basic_machine
943	vendor=pc
	@@ -961,19 +1021,23 @@
961
962	unset -v basic_machine
963
964	# Decode basic machines in the full and proper CPU-Company form.
965	case $cpu-$vendor in
966	# Here we handle the default manufacturer of certain CPU types in canonical form. It is in
967	# some cases the only manufacturer, in others, it is the most popular.




968	craynv-unknown)
969	vendor=cray
970	basic_os=${basic_os:-unicosmp}
971	;;
972	c90-unknown \| c90-cray)
973	vendor=cray
974	basic_os=${Basic_os:-unicos}
975	;;
976	fx80-unknown)
977	vendor=alliant
978	;;
979	romp-unknown)
	@@ -1010,23 +1074,46 @@
1010	cpu=xps100
1011	vendor=honeywell
1012	;;
1013
1014	# Here we normalize CPU types with a missing or matching vendor
1015	dpx20-unknown \| dpx20-bull)
1016	cpu=rs6000
1017	vendor=bull




















1018	basic_os=${basic_os:-bosx}
1019	;;



1020
1021	# Here we normalize CPU types irrespective of the vendor
1022	amd64-*)
1023	cpu=x86_64
1024	;;
1025	blackfin-*)
1026	cpu=bfin
1027	basic_os=linux
1028	;;
1029	c54x-*)
1030	cpu=tic54x
1031	;;
1032	c55x-*)
	@@ -1045,37 +1132,34 @@
1045	ms1-*)
1046	cpu=mt
1047	;;
1048	m68knommu-*)
1049	cpu=m68k
1050	basic_os=linux
1051	;;
1052	m9s12z-* \| m68hcs12z-* \| hcs12z-* \| s12z-*)
1053	cpu=s12z
1054	;;
1055	openrisc-*)
1056	cpu=or32
1057	;;
1058	parisc-*)
1059	cpu=hppa
1060	basic_os=linux
1061	;;
1062	pentium-* \| p5-* \| k5-* \| k6-* \| nexgen-* \| viac3-*)
1063	cpu=i586
1064	;;
1065	pentiumpro-* \| p6-* \| 6x86-* \| athlon-* \| athalon_-)
1066	cpu=i686
1067	;;
1068	pentiumii-* \| pentium2-* \| pentiumiii-* \| pentium3-*)
1069	cpu=i686
1070	;;
1071	pentium4-*)
1072	cpu=i786
1073	;;
1074	pc98-*)
1075	cpu=i386
1076	;;
1077	ppc-* \| ppcbe-*)
1078	cpu=powerpc
1079	;;
1080	ppcle-* \| powerpclittle-*)
1081	cpu=powerpcle
	@@ -1105,17 +1189,14 @@
1105	cpu=mipstx39
1106	;;
1107	tx39el-*)
1108	cpu=mipstx39el
1109	;;
1110	x64-*)
1111	cpu=x86_64
1112	;;
1113	xscale-* \| xscalee[bl]-*)
1114	cpu=`echo "$cpu" \| sed 's/^xscale/arm/'`
1115	;;
1116	arm64-*)
1117	cpu=aarch64
1118	;;
1119
1120	# Recognize the canonical CPU Types that limit and/or modify the
1121	# company names they are paired with.
	@@ -1163,118 +1244,235 @@
1163
1164	*)
1165	# Recognize the canonical CPU types that are allowed with any
1166	# company name.
1167	case $cpu in
1168	1750a \| 580 \


1169	\| a29k \
1170	\| aarch64 \| aarch64_be \


1171	\| abacus \
1172	\| alpha \| alphaev[4-8] \| alphaev56 \| alphaev6[78] \
1173	\| alpha64 \| alpha64ev[4-8] \| alpha64ev56 \| alpha64ev6[78] \
1174	\| alphapca5[67] \| alpha64pca5[67] \







1175	\| am33_2.0 \
1176	\| amdgcn \
1177	\| arc \| arceb \| arc32 \| arc64 \
1178	\| arm \| arm[lb]e \| arme[lb] \| armv* \
1179	\| avr \| avr32 \







1180	\| asmjs \


1181	\| ba \
1182	\| be32 \| be64 \
1183	\| bfin \| bpf \| bs2000 \
1184	\| c[123]* \| c30 \| [cjt]90 \| c4x \
1185	\| c8051 \| clipper \| craynv \| csky \| cydra \
1186	\| d10v \| d30v \| dlx \| dsp16xx \
1187	\| e2k \| elxsi \| epiphany \
1188	\| f30[01] \| f700 \| fido \| fr30 \| frv \| ft32 \| fx80 \
1189	\| h8300 \| h8500 \
1190	\| hppa \| hppa1.[01] \| hppa2.0 \| hppa2.0[nw] \| hppa64 \




















1191	\| hexagon \
1192	\| i370 \| i*86 \| i860 \| i960 \| ia16 \| ia64 \
1193	\| ip2k \| iq2000 \












1194	\| k1om \
1195	\| le32 \| le64 \


1196	\| lm32 \
1197	\| loongarch32 \| loongarch64 \| loongarchx32 \
1198	\| m32c \| m32r \| m32rle \
1199	\| m5200 \| m68000 \| m680[012346]0 \| m68360 \| m683?2 \| m68k \
1200	\| m6811 \| m68hc11 \| m6812 \| m68hc12 \| m68hcs12x \
1201	\| m88110 \| m88k \| maxq \| mb \| mcore \| mep \| metag \
1202	\| microblaze \| microblazeel \
1203	\| mips \| mipsbe \| mipseb \| mipsel \| mipsle \
1204	\| mips16 \
1205	\| mips64 \| mips64eb \| mips64el \
1206	\| mips64octeon \| mips64octeonel \
1207	\| mips64orion \| mips64orionel \
1208	\| mips64r5900 \| mips64r5900el \
1209	\| mips64vr \| mips64vrel \
1210	\| mips64vr4100 \| mips64vr4100el \
1211	\| mips64vr4300 \| mips64vr4300el \
1212	\| mips64vr5000 \| mips64vr5000el \
1213	\| mips64vr5900 \| mips64vr5900el \
1214	\| mipsisa32 \| mipsisa32el \
1215	\| mipsisa32r2 \| mipsisa32r2el \
1216	\| mipsisa32r3 \| mipsisa32r3el \
1217	\| mipsisa32r5 \| mipsisa32r5el \
1218	\| mipsisa32r6 \| mipsisa32r6el \
1219	\| mipsisa64 \| mipsisa64el \
1220	\| mipsisa64r2 \| mipsisa64r2el \
1221	\| mipsisa64r3 \| mipsisa64r3el \
1222	\| mipsisa64r5 \| mipsisa64r5el \
1223	\| mipsisa64r6 \| mipsisa64r6el \
1224	\| mipsisa64sb1 \| mipsisa64sb1el \
1225	\| mipsisa64sr71k \| mipsisa64sr71kel \
1226	\| mipsr5900 \| mipsr5900el \
1227	\| mipstx39 \| mipstx39el \
1228	\| mmix \
1229	\| mn10200 \| mn10300 \

1230	\| moxie \
1231	\| mt \
1232	\| msp430 \
1233	\| nds32 \| nds32le \| nds32be \




1234	\| nfp \
1235	\| nios \| nios2 \| nios2eb \| nios2el \
1236	\| none \| np1 \| ns16k \| ns32k \| nvptx \







1237	\| open8 \
1238	\| or1k* \
1239	\| or32 \
1240	\| orion \


1241	\| picochip \
1242	\| pdp10 \| pdp11 \| pj \| pjl \| pn \| power \
1243	\| powerpc \| powerpc64 \| powerpc64le \| powerpcle \| powerpcspe \







1244	\| pru \
1245	\| pyramid \
1246	\| riscv \| riscv32 \| riscv32be \| riscv64 \| riscv64be \
1247	\| rl78 \| romp \| rs6000 \| rx \
1248	\| s390 \| s390x \








1249	\| score \
1250	\| sh \| shl \
1251	\| sh[1234] \| sh[24]a \| sh[24]ae[lb] \| sh[23]e \| she[lb] \| sh[lb]e \
1252	\| sh[1234]e[lb] \| sh[12345][lb]e \| sh[23]ele \| sh64 \| sh64le \
1253	\| sparc \| sparc64 \| sparc64b \| sparc64v \| sparc86x \| sparclet \















1254	\| sparclite \
1255	\| sparcv8 \| sparcv9 \| sparcv9b \| sparcv9v \| sv1 \| sx* \



1256	\| spu \


1257	\| tahoe \
1258	\| thumbv7* \
1259	\| tic30 \| tic4x \| tic54x \| tic55x \| tic6x \| tic80 \





1260	\| tron \
1261	\| ubicom32 \
1262	\| v70 \| v850 \| v850e \| v850e1 \| v850es \| v850e2 \| v850e2v3 \







1263	\| vax \

1264	\| visium \
1265	\| w65 \
1266	\| wasm32 \| wasm64 \

1267	\| we32k \
1268	\| x86 \| x86_64 \| xc16x \| xgate \| xps100 \
1269	\| xstormy16 \| xtensa* \





1270	\| ymp \
1271	\| z8k \| z80)

1272	;;
1273
1274	*)
1275	echo Invalid configuration \`"$1"\': machine \`"$cpu-$vendor"\' not recognized 1>&2
1276	exit 1
1277	;;
1278	esac
1279	;;
1280	esac
	@@ -1291,15 +1489,16 @@
1291	;;
1292	esac
1293
1294	# Decode manufacturer-specific aliases for certain operating systems.
1295
1296	if test x$basic_os != x
1297	then
1298
1299	# First recognize some ad-hoc caes, or perhaps split kernel-os, or else just
1300	# set os.

1301	case $basic_os in
1302	gnu/linux*)
1303	kernel=linux
1304	os=`echo "$basic_os" \| sed -e 's\|gnu/linux\|gnu\|'`
1305	;;
	@@ -1310,14 +1509,15 @@
1310	nto-qnx*)
1311	kernel=nto
1312	os=`echo "$basic_os" \| sed -e 's\|nto-qnx\|qnx\|'`
1313	;;
1314	-)
1315	# shellcheck disable=SC2162
1316	IFS="-" read kernel os <<EOF
1317	$basic_os
1318	EOF

1319	;;
1320	# Default OS when just kernel was specified
1321	nto*)
1322	kernel=nto
1323	os=`echo "$basic_os" \| sed -e 's\|nto\|qnx\|'`
	@@ -1324,10 +1524,14 @@
1324	;;
1325	linux*)
1326	kernel=linux
1327	os=`echo "$basic_os" \| sed -e 's\|linux\|gnu\|'`
1328	;;




1329	*)
1330	kernel=
1331	os=$basic_os
1332	;;
1333	esac
	@@ -1350,10 +1554,27 @@
1350	solaris)
1351	os=solaris2
1352	;;
1353	unixware*)
1354	os=sysv4.2uw

















1355	;;
1356	# es1800 is here to avoid being matched by es* (a different OS)
1357	es1800*)
1358	os=ose
1359	;;
	@@ -1421,10 +1642,11 @@
1421	wince*)
1422	os=wince
1423	;;
1424	utek*)
1425	os=bsd

1426	;;
1427	dynix*)
1428	os=bsd
1429	;;
1430	acis*)
	@@ -1437,25 +1659,29 @@
1437	os=syllable
1438	;;
1439	386bsd)
1440	os=bsd
1441	;;
1442	ctix* \| uts*)




1443	os=sysv
1444	;;
1445	nova*)
1446	os=rtmk-nova
1447	;;
1448	ns2)
1449	os=nextstep2
1450	;;
1451	# Preserve the version number of sinix5.
1452	sinix5.*)
1453	os=`echo "$os" \| sed -e 's\|sinix\|sysv\|'`

1454	;;
1455	sinix*)
1456	os=sysv4

1457	;;
1458	tpf*)
1459	os=tpf
1460	;;
1461	triton*)
	@@ -1489,13 +1715,19 @@
1489	case $cpu in
1490	arm*)
1491	os=eabi
1492	;;
1493	*)
1494	os=elf

1495	;;
1496	esac





1497	;;
1498	*)
1499	# No normalization, but not necessarily accepted, that comes below.
1500	;;
1501	esac
	@@ -1511,47 +1743,57 @@
1511	# that MANUFACTURER isn't an operating system. Otherwise, code above
1512	# will signal an error saying that MANUFACTURER isn't an operating
1513	# system, and we'll never get to this point.
1514
1515	kernel=

1516	case $cpu-$vendor in
1517	score-*)
1518	os=elf

1519	;;
1520	spu-*)
1521	os=elf

1522	;;
1523	*-acorn)
1524	os=riscix1.2
1525	;;
1526	arm*-rebel)
1527	kernel=linux
1528	os=gnu
1529	;;
1530	arm*-semi)
1531	os=aout

1532	;;
1533	c4x-* \| tic4x-*)
1534	os=coff

1535	;;
1536	c8051-*)
1537	os=elf

1538	;;
1539	clipper-intergraph)
1540	os=clix
1541	;;
1542	hexagon-*)
1543	os=elf

1544	;;
1545	tic54x-*)
1546	os=coff

1547	;;
1548	tic55x-*)
1549	os=coff

1550	;;
1551	tic6x-*)
1552	os=coff

1553	;;
1554	# This must come before the *-dec entry.
1555	pdp10-*)
1556	os=tops20
1557	;;
	@@ -1569,32 +1811,47 @@
1569	;;
1570	m68000-sun)
1571	os=sunos3
1572	;;
1573	m68*-cisco)
1574	os=aout

1575	;;
1576	mep-*)
1577	os=elf









1578	;;
1579	mips*-cisco)
1580	os=elf

1581	;;
1582	mips-)
1583	os=elf

1584	;;
1585	or32-*)
1586	os=coff

1587	;;
1588	*-tti) # must be before sparc entry or we get the wrong os.

1589	os=sysv3
1590	;;
1591	sparc-* \| *-sun)
1592	os=sunos4.1.1
1593	;;
1594	pru-*)
1595	os=elf

1596	;;
1597	*-be)
1598	os=beos
1599	;;
1600	*-ibm)
	@@ -1614,11 +1871,11 @@
1614	;;
1615	*-hp)
1616	os=hpux
1617	;;
1618	*-hitachi)
1619	os=hiux
1620	;;
1621	i860-* \| -att \| -ncr \| -altos \| -motorola \| *-convergent)
1622	os=sysv
1623	;;
1624	*-cbm)
	@@ -1658,27 +1915,23 @@
1658	os=bsd
1659	;;
1660	*-encore)
1661	os=bsd
1662	;;
1663	*-sgi)
1664	os=irix
1665	;;
1666	*-siemens)
1667	os=sysv4
1668	;;
1669	*-masscomp)
1670	os=rtu
1671	;;
1672	f30[01]-fujitsu \| f700-fujitsu)
1673	os=uxpv
1674	;;
1675	*-rom68k)
1676	os=coff

1677	;;
1678	-bug)
1679	os=coff

1680	;;
1681	*-apple)
1682	os=macos
1683	;;
1684	-atari)
	@@ -1692,96 +1945,324 @@
1692	;;
1693	esac
1694
1695	fi
1696
1697	# Now, validate our (potentially fixed-up) OS.

1698	case $os in
1699	# Sometimes we do "kernel-libc", so those need to count as OSes.
1700	musl* \| newlib* \| uclibc*)
1701	;;
1702	# Likewise for "kernel-abi"
1703	eabi* \| gnueabi*)
1704	;;
1705	# VxWorks passes extra cpu info in the 4th filed.
1706	simlinux \| simwindows \| spe)



1707	;;
1708	# Now accept the basic system types.
1709	# The portable systems comes first.
1710	# Each alternative MUST end in a * to match a version number.
1711	gnu* \| android* \| bsd* \| mach* \| minix* \| genix* \| ultrix* \| irix* \
1712	\| vms \| esix* \| aix* \| cnk* \| sunos \| sunos[34]* \
1713	\| hpux* \| unos* \| osf* \| luna* \| dgux* \| auroraux* \| solaris* \
1714	\| sym* \| plan9* \| psp* \| sim* \| xray* \| os68k* \| v88r* \
1715	\| hiux* \| abug \| nacl* \| netware* \| windows* \
1716	\| os9* \| macos* \| osx* \| ios* \
1717	\| mpw* \| magic* \| mmixware* \| mon960* \| lnews* \
1718	\| amigaos* \| amigados* \| msdos* \| newsos* \| unicos* \| aof* \
1719	\| aos* \| aros* \| cloudabi* \| sortix* \| twizzler* \
1720	\| nindy* \| vxsim* \| vxworks* \| ebmon* \| hms* \| mvs* \
1721	\| clix* \| riscos* \| uniplus* \| iris* \| isc* \| rtu* \| xenix* \
1722	\| mirbsd* \| netbsd* \| dicos* \| openedition* \| ose* \
1723	\| bitrig* \| openbsd* \| secbsd* \| solidbsd* \| libertybsd* \| os108* \
1724	\| ekkobsd* \| freebsd* \| riscix* \| lynxos* \| os400* \
1725	\| bosx* \| nextstep* \| cxux* \| aout* \| elf* \| oabi* \
1726	\| ptx* \| coff* \| ecoff* \| winnt* \| domain* \| vsta* \
1727	\| udi* \| lites* \| ieee* \| go32* \| aux* \| hcos* \
1728	\| chorusrdb* \| cegcc* \| glidix* \| serenity* \
1729	\| cygwin* \| msys* \| pe* \| moss* \| proelf* \| rtems* \
1730	\| midipix* \| mingw32* \| mingw64* \| mint* \
1731	\| uxpv* \| beos* \| mpeix* \| udk* \| moxiebox* \
1732	\| interix* \| uwin* \| mks* \| rhapsody* \| darwin* \
1733	\| openstep* \| oskit* \| conix* \| pw32* \| nonstopux* \
1734	\| storm-chaos* \| tops10* \| tenex* \| tops20* \| its* \
1735	\| os2* \| vos* \| palmos* \| uclinux* \| nucleus* \| morphos* \
1736	\| scout* \| superux* \| sysv* \| rtmk* \| tpf* \| windiss* \
1737	\| powermax* \| dnix* \| nx6 \| nx7 \| sei* \| dragonfly* \
1738	\| skyos* \| haiku* \| rdos* \| toppers* \| drops* \| es* \
1739	\| onefs* \| tirtos* \| phoenix* \| fuchsia* \| redox* \| bme* \
1740	\| midnightbsd* \| amdhsa* \| unleashed* \| emscripten* \| wasi* \
1741	\| nsk* \| powerunix* \| genode* \| zvmoe* \| qnx* \| emx*)



























































































































































1742	;;
1743	# This one is extra strict with allowed versions
1744	sco3.2v2 \| sco3.2v[4-9]* \| sco5v6*)
1745	# Don't forget version if it is 3.2v4 or newer.
1746	;;







1747	none)
1748	;;





















1749	*)
1750	echo Invalid configuration \`"$1"\': OS \`"$os"\' not recognized 1>&2














1751	exit 1
1752	;;
1753	esac
1754
1755	# As a final step for OS-related things, validate the OS-kernel combination
1756	# (given a valid OS), if there is a kernel.
1757	case $kernel-$os in
1758	linux-gnu* \| linux-dietlibc* \| linux-android* \| linux-newlib* \| linux-musl* \| linux-uclibc* )






1759	;;
1760	uclinux-uclibc* )
1761	;;
1762	-dietlibc* \| -newlib* \| -musl* \| -uclibc* )

1763	# These are just libc implementations, not actual OSes, and thus
1764	# require a kernel.
1765	echo "Invalid configuration \`$1': libc \`$os' needs explicit kernel." 1>&2












1766	exit 1
1767	;;
1768	kfreebsd-gnu \| kopensolaris-gnu)
1769	;;
1770	vxworks-simlinux \| vxworks-simwindows \| vxworks-spe)
1771	;;
1772	nto-qnx*)
1773	;;
1774	os2-emx)
1775	;;
1776	-eabi \| -gnueabi)
1777	;;
1778	-*)






1779	# Blank kernel with real OS is always fine.
1780	;;
1781	-)
1782	echo "Invalid configuration \`$1': Kernel \`$kernel' not known to work with OS \`$os'." 1>&2



1783	exit 1
1784	;;
1785	esac
1786
1787	# Here we handle the case where we know the os, and the CPU type, but not the
	@@ -1790,11 +2271,11 @@
1790	unknown)
1791	case $cpu-$os in
1792	-riscix)
1793	vendor=acorn
1794	;;
1795	-sunos)
1796	vendor=sun
1797	;;
1798	-cnk \| -aix)
1799	vendor=ibm
1800	;;
	@@ -1860,14 +2341,14 @@
1860	;;
1861	esac
1862	;;
1863	esac
1864
1865	echo "$cpu-$vendor-${kernel:+$kernel-}$os"
1866	exit
1867
1868	# Local variables:
1869	# eval: (add-hook 'before-save-hook 'time-stamp)
1870	# time-stamp-start: "timestamp='"
1871	# time-stamp-format: "%:y-%02m-%02d"
1872	# time-stamp-end: "'"
1873	# End:
1874

	--- autosetup/autosetup-config.sub
	+++ autosetup/autosetup-config.sub
	@@ -1,16 +1,16 @@
1	#! /bin/sh
2	# Configuration validation subroutine script.
3	# Copyright 1992-2024 Free Software Foundation, Inc.
4
5	# shellcheck disable=SC2006,SC2268,SC2162 # see below for rationale
6
7	timestamp='2024-05-27'
8
9	# This file is free software; you can redistribute it and/or modify it
10	# under the terms of the GNU General Public License as published by
11	# the Free Software Foundation, either version 3 of the License, or
12	# (at your option) any later version.
13	#
14	# This program is distributed in the hope that it will be useful, but
15	# WITHOUT ANY WARRANTY; without even the implied warranty of
16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	@@ -74,17 +74,17 @@
74	Report bugs and patches to <[email protected]>."
75
76	version="\
77	GNU config.sub ($timestamp)
78
79	Copyright 1992-2024 Free Software Foundation, Inc.
80
81	This is free software; see the source for copying conditions. There is NO
82	warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
83
84	help="
85	Try '$me --help' for more information."
86
87	# Parse command line
88	while test $# -gt 0 ; do
89	case $1 in
90	--time-stamp \| --time* \| -t )
	@@ -118,19 +118,20 @@
118	*) echo "$me: too many arguments$help" >&2
119	exit 1;;
120	esac
121
122	# Split fields of configuration type
123	saved_IFS=$IFS
124	IFS="-" read field1 field2 field3 field4 <<EOF
125	$1
126	EOF
127	IFS=$saved_IFS
128
129	# Separate into logical components for further validation
130	case $1 in
131	----*)
132	echo "Invalid configuration '$1': more than four components" >&2
133	exit 1
134	;;
135	---)
136	basic_machine=$field1-$field2
137	basic_os=$field3-$field4
	@@ -138,14 +139,25 @@
139	--*)
140	# Ambiguous whether COMPANY is present, or skipped and KERNEL-OS is two
141	# parts
142	maybe_os=$field2-$field3
143	case $maybe_os in
144	cloudabi-eabi \
145	\| kfreebsd-gnu \
146	\| knetbsd-gnu \
147	\| kopensolaris-gnu \
148	\| linux-* \
149	\| managarm-* \
150	\| netbsd-eabi \
151	\| netbsd-gnu \
152	\| nto-qnx* \
153	\| os2-emx* \
154	\| rtmk-nova* \
155	\| storm-chaos* \
156	\| uclinux-gnu* \
157	\| uclinux-uclibc* \
158	\| windows-* )
159	basic_machine=$field1
160	basic_os=$maybe_os
161	;;
162	android-linux)
163	basic_machine=$field1-unknown
	@@ -156,37 +168,105 @@
168	basic_os=$field3
169	;;
170	esac
171	;;
172	-)

173	case $field1-$field2 in
174	# Shorthands that happen to contain a single dash
175	convex-c[12] \| convex-c3[248])
176	basic_machine=$field2-convex
177	basic_os=
178	;;
179	decstation-3100)
180	basic_machine=mips-dec
181	basic_os=
182	;;
183	-)
184	# Second component is usually, but not always the OS
185	case $field2 in
186	# Do not treat sunos as a manufacturer
187	sunos)
188	basic_machine=$field1
189	basic_os=$field2
190	;;
191	# Manufacturers
192	3100* \
193	\| 32* \
194	\| 3300* \
195	\| 3600* \
196	\| 7300* \
197	\| acorn \
198	\| altos* \
199	\| apollo \
200	\| apple \
201	\| atari \
202	\| att* \
203	\| axis \
204	\| be \
205	\| bull \
206	\| cbm \
207	\| ccur \
208	\| cisco \
209	\| commodore \
210	\| convergent* \
211	\| convex* \
212	\| cray \
213	\| crds \
214	\| dec* \
215	\| delta* \
216	\| dg \
217	\| digital \
218	\| dolphin \
219	\| encore* \
220	\| gould \
221	\| harris \
222	\| highlevel \
223	\| hitachi* \
224	\| hp \
225	\| ibm* \
226	\| intergraph \
227	\| isi* \
228	\| knuth \
229	\| masscomp \
230	\| microblaze* \
231	\| mips* \
232	\| motorola* \
233	\| ncr* \
234	\| news \
235	\| next \
236	\| ns \
237	\| oki \
238	\| omron* \
239	\| pc533* \
240	\| rebel \
241	\| rom68k \
242	\| rombug \
243	\| semi \
244	\| sequent* \
245	\| siemens \
246	\| sgi* \
247	\| siemens \
248	\| sim \
249	\| sni \
250	\| sony* \
251	\| stratus \
252	\| sun \
253	\| sun[234]* \
254	\| tektronix \
255	\| tti* \
256	\| ultra \
257	\| unicom* \
258	\| wec \
259	\| winbond \
260	\| wrs)
261	basic_machine=$field1-$field2
262	basic_os=
263	;;
264	zephyr*)
265	basic_machine=$field1-unknown
266	basic_os=$field2
267	;;
268	*)
269	basic_machine=$field1
270	basic_os=$field2
271	;;
272	esac
	@@ -263,30 +343,10 @@
343	;;
344	cegcc)
345	basic_machine=arm-unknown
346	basic_os=cegcc
347	;;




















348	cray)
349	basic_machine=j90-cray
350	basic_os=unicos
351	;;
352	crds \| unos)
	@@ -705,13 +765,21 @@
765	decsystem20* \| dec20*)
766	cpu=pdp10
767	vendor=dec
768	basic_os=tops20
769	;;
770	delta \| 3300 \| delta-motorola \| 3300-motorola \| motorola-delta \| motorola-3300)

771	cpu=m68k
772	vendor=motorola
773	;;
774	# This used to be dpx2*, but that gets the RS6000-based
775	# DPX/20 and the x86-based DPX/2-100 wrong. See
776	# https://oldskool.silicium.org/stations/bull_dpx20.htm
777	# https://www.feb-patrimoine.com/english/bull_dpx2.htm
778	# https://www.feb-patrimoine.com/english/unix_and_bull.htm
779	dpx2 \| dpx2[23]00 \| dpx2[23]xx)
780	cpu=m68k
781	vendor=bull
782	;;
783	dpx2100 \| dpx21xx)
784	cpu=i386
785	vendor=bull
	@@ -718,6 +786,9 @@
786	;;
787	dpx20)
788	cpu=rs6000
789	vendor=bull
790	;;
791	encore \| umax \| mmax)
792	cpu=ns32k
793	vendor=encore
794	;;
	@@ -828,22 +899,10 @@
899	basic_os=newsos
900	;;
901	next \| m*-next)
902	cpu=m68k
903	vendor=next












904	;;
905	np1)
906	cpu=np1
907	vendor=gould
908	;;
	@@ -928,16 +987,17 @@
987	cpu=sparc
988	vendor=`echo "$basic_machine" \| sed 's/-.*//'`
989	;;
990
991	-)
992	saved_IFS=$IFS
993	IFS="-" read cpu vendor <<EOF
994	$basic_machine
995	EOF
996	IFS=$saved_IFS
997	;;
998	# We use 'pc' rather than 'unknown'
999	# because (1) that's what they normally are, and
1000	# (2) the word "unknown" tends to confuse beginning users.
1001	i*86 \| x86_64)
1002	cpu=$basic_machine
1003	vendor=pc
	@@ -961,19 +1021,23 @@
1021
1022	unset -v basic_machine
1023
1024	# Decode basic machines in the full and proper CPU-Company form.
1025	case $cpu-$vendor in
1026	# Here we handle the default manufacturer of certain CPU types in canonical form.
1027	# It is in some cases the only manufacturer, in others, it is the most popular.
1028	c[12]-convex \| c[12]-unknown \| c3[248]-convex \| c3[248]-unknown)
1029	vendor=convex
1030	basic_os=${basic_os:-bsd}
1031	;;
1032	craynv-unknown)
1033	vendor=cray
1034	basic_os=${basic_os:-unicosmp}
1035	;;
1036	c90-unknown \| c90-cray)
1037	vendor=cray
1038	basic_os=${basic_os:-unicos}
1039	;;
1040	fx80-unknown)
1041	vendor=alliant
1042	;;
1043	romp-unknown)
	@@ -1010,23 +1074,46 @@
1074	cpu=xps100
1075	vendor=honeywell
1076	;;
1077
1078	# Here we normalize CPU types with a missing or matching vendor
1079	armh-unknown \| armh-alt)
1080	cpu=armv7l
1081	vendor=alt
1082	basic_os=${basic_os:-linux-gnueabihf}
1083	;;
1084
1085	# Normalized CPU+vendor pairs that imply an OS, if not otherwise specified
1086	m68k-isi)
1087	basic_os=${basic_os:-sysv}
1088	;;
1089	m68k-sony)
1090	basic_os=${basic_os:-newsos}
1091	;;
1092	m68k-tektronix)
1093	basic_os=${basic_os:-bsd}
1094	;;
1095	m88k-harris)
1096	basic_os=${basic_os:-sysv3}
1097	;;
1098	i386-bull \| m68k-bull)
1099	basic_os=${basic_os:-sysv3}
1100	;;
1101	rs6000-bull)
1102	basic_os=${basic_os:-bosx}
1103	;;
1104	mips-sni)
1105	basic_os=${basic_os:-sysv4}
1106	;;
1107
1108	# Here we normalize CPU types irrespective of the vendor
1109	amd64-*)
1110	cpu=x86_64
1111	;;
1112	blackfin-*)
1113	cpu=bfin
1114	basic_os=${basic_os:-linux}
1115	;;
1116	c54x-*)
1117	cpu=tic54x
1118	;;
1119	c55x-*)
	@@ -1045,37 +1132,34 @@
1132	ms1-*)
1133	cpu=mt
1134	;;
1135	m68knommu-*)
1136	cpu=m68k
1137	basic_os=${basic_os:-linux}
1138	;;
1139	m9s12z-* \| m68hcs12z-* \| hcs12z-* \| s12z-*)
1140	cpu=s12z
1141	;;
1142	openrisc-*)
1143	cpu=or32
1144	;;
1145	parisc-*)
1146	cpu=hppa
1147	basic_os=${basic_os:-linux}
1148	;;
1149	pentium-* \| p5-* \| k5-* \| k6-* \| nexgen-* \| viac3-*)
1150	cpu=i586
1151	;;
1152	pentiumpro-* \| p6-* \| 6x86-* \| athlon-* \| athlon_-)
1153	cpu=i686
1154	;;
1155	pentiumii-* \| pentium2-* \| pentiumiii-* \| pentium3-*)
1156	cpu=i686
1157	;;
1158	pentium4-*)
1159	cpu=i786
1160	;;



1161	ppc-* \| ppcbe-*)
1162	cpu=powerpc
1163	;;
1164	ppcle-* \| powerpclittle-*)
1165	cpu=powerpcle
	@@ -1105,17 +1189,14 @@
1189	cpu=mipstx39
1190	;;
1191	tx39el-*)
1192	cpu=mipstx39el
1193	;;



1194	xscale-* \| xscalee[bl]-*)
1195	cpu=`echo "$cpu" \| sed 's/^xscale/arm/'`
1196	;;
1197	arm64-* \| aarch64le-*)
1198	cpu=aarch64
1199	;;
1200
1201	# Recognize the canonical CPU Types that limit and/or modify the
1202	# company names they are paired with.
	@@ -1163,118 +1244,235 @@
1244
1245	*)
1246	# Recognize the canonical CPU types that are allowed with any
1247	# company name.
1248	case $cpu in
1249	1750a \
1250	\| 580 \
1251	\| [cjt]90 \
1252	\| a29k \
1253	\| aarch64 \
1254	\| aarch64_be \
1255	\| aarch64c \
1256	\| abacus \
1257	\| alpha \
1258	\| alpha64 \
1259	\| alpha64ev56 \
1260	\| alpha64ev6[78] \
1261	\| alpha64ev[4-8] \
1262	\| alpha64pca5[67] \
1263	\| alphaev56 \
1264	\| alphaev6[78] \
1265	\| alphaev[4-8] \
1266	\| alphapca5[67] \
1267	\| am33_2.0 \
1268	\| amdgcn \
1269	\| arc \
1270	\| arc32 \
1271	\| arc64 \
1272	\| arceb \
1273	\| arm \
1274	\| arm64e \
1275	\| arm64ec \
1276	\| arm[lb]e \
1277	\| arme[lb] \
1278	\| armv* \
1279	\| asmjs \
1280	\| avr \
1281	\| avr32 \
1282	\| ba \
1283	\| be32 \
1284	\| be64 \
1285	\| bfin \
1286	\| bpf \
1287	\| bs2000 \
1288	\| c30 \
1289	\| c4x \
1290	\| c8051 \
1291	\| c[123]* \
1292	\| clipper \
1293	\| craynv \
1294	\| csky \
1295	\| cydra \
1296	\| d10v \
1297	\| d30v \
1298	\| dlx \
1299	\| dsp16xx \
1300	\| e2k \
1301	\| elxsi \
1302	\| epiphany \
1303	\| f30[01] \
1304	\| f700 \
1305	\| fido \
1306	\| fr30 \
1307	\| frv \
1308	\| ft32 \
1309	\| fx80 \
1310	\| h8300 \
1311	\| h8500 \
1312	\| hexagon \
1313	\| hppa \
1314	\| hppa1.[01] \
1315	\| hppa2.0 \
1316	\| hppa2.0[nw] \
1317	\| hppa64 \
1318	\| i*86 \
1319	\| i370 \
1320	\| i860 \
1321	\| i960 \
1322	\| ia16 \
1323	\| ia64 \
1324	\| ip2k \
1325	\| iq2000 \
1326	\| javascript \
1327	\| k1om \
1328	\| kvx \
1329	\| le32 \
1330	\| le64 \
1331	\| lm32 \
1332	\| loongarch32 \
1333	\| loongarch64 \
1334	\| m32c \
1335	\| m32r \
1336	\| m32rle \
1337	\| m5200 \
1338	\| m68000 \
1339	\| m680[012346]0 \
1340	\| m6811 \
1341	\| m6812 \
1342	\| m68360 \
1343	\| m683?2 \
1344	\| m68hc11 \
1345	\| m68hc12 \
1346	\| m68hcs12x \
1347	\| m68k \
1348	\| m88110 \
1349	\| m88k \
1350	\| maxq \
1351	\| mb \
1352	\| mcore \
1353	\| mep \
1354	\| metag \
1355	\| microblaze \
1356	\| microblazeel \
1357	\| mips* \





1358	\| mmix \
1359	\| mn10200 \
1360	\| mn10300 \
1361	\| moxie \

1362	\| msp430 \
1363	\| mt \
1364	\| nanomips* \
1365	\| nds32 \
1366	\| nds32be \
1367	\| nds32le \
1368	\| nfp \
1369	\| nios \
1370	\| nios2 \
1371	\| nios2eb \
1372	\| nios2el \
1373	\| none \
1374	\| np1 \
1375	\| ns16k \
1376	\| ns32k \
1377	\| nvptx \
1378	\| open8 \
1379	\| or1k* \
1380	\| or32 \
1381	\| orion \
1382	\| pdp10 \
1383	\| pdp11 \
1384	\| picochip \
1385	\| pj \
1386	\| pjl \
1387	\| pn \
1388	\| power \
1389	\| powerpc \
1390	\| powerpc64 \
1391	\| powerpc64le \
1392	\| powerpcle \
1393	\| powerpcspe \
1394	\| pru \
1395	\| pyramid \
1396	\| riscv \
1397	\| riscv32 \
1398	\| riscv32be \
1399	\| riscv64 \
1400	\| riscv64be \
1401	\| rl78 \
1402	\| romp \
1403	\| rs6000 \
1404	\| rx \
1405	\| s390 \
1406	\| s390x \
1407	\| score \
1408	\| sh \
1409	\| sh64 \
1410	\| sh64le \
1411	\| sh[12345][lb]e \
1412	\| sh[1234] \
1413	\| sh[1234]e[lb] \
1414	\| sh[23]e \
1415	\| sh[23]ele \
1416	\| sh[24]a \
1417	\| sh[24]ae[lb] \
1418	\| sh[lb]e \
1419	\| she[lb] \
1420	\| shl \
1421	\| sparc \
1422	\| sparc64 \
1423	\| sparc64b \
1424	\| sparc64v \
1425	\| sparc86x \
1426	\| sparclet \
1427	\| sparclite \
1428	\| sparcv8 \
1429	\| sparcv9 \
1430	\| sparcv9b \
1431	\| sparcv9v \
1432	\| spu \
1433	\| sv1 \
1434	\| sx* \
1435	\| tahoe \
1436	\| thumbv7* \
1437	\| tic30 \
1438	\| tic4x \
1439	\| tic54x \
1440	\| tic55x \
1441	\| tic6x \
1442	\| tic80 \
1443	\| tron \
1444	\| ubicom32 \
1445	\| v70 \
1446	\| v810 \
1447	\| v850 \
1448	\| v850e \
1449	\| v850e1 \
1450	\| v850e2 \
1451	\| v850e2v3 \
1452	\| v850es \
1453	\| vax \
1454	\| vc4 \
1455	\| visium \
1456	\| w65 \
1457	\| wasm32 \
1458	\| wasm64 \
1459	\| we32k \
1460	\| x86 \
1461	\| x86_64 \
1462	\| xc16x \
1463	\| xgate \
1464	\| xps100 \
1465	\| xstormy16 \
1466	\| xtensa* \
1467	\| ymp \
1468	\| z80 \
1469	\| z8k)
1470	;;
1471
1472	*)
1473	echo "Invalid configuration '$1': machine '$cpu-$vendor' not recognized" 1>&2
1474	exit 1
1475	;;
1476	esac
1477	;;
1478	esac
	@@ -1291,15 +1489,16 @@
1489	;;
1490	esac
1491
1492	# Decode manufacturer-specific aliases for certain operating systems.
1493
1494	if test x"$basic_os" != x
1495	then
1496
1497	# First recognize some ad-hoc cases, or perhaps split kernel-os, or else just
1498	# set os.
1499	obj=
1500	case $basic_os in
1501	gnu/linux*)
1502	kernel=linux
1503	os=`echo "$basic_os" \| sed -e 's\|gnu/linux\|gnu\|'`
1504	;;
	@@ -1310,14 +1509,15 @@
1509	nto-qnx*)
1510	kernel=nto
1511	os=`echo "$basic_os" \| sed -e 's\|nto-qnx\|qnx\|'`
1512	;;
1513	-)
1514	saved_IFS=$IFS
1515	IFS="-" read kernel os <<EOF
1516	$basic_os
1517	EOF
1518	IFS=$saved_IFS
1519	;;
1520	# Default OS when just kernel was specified
1521	nto*)
1522	kernel=nto
1523	os=`echo "$basic_os" \| sed -e 's\|nto\|qnx\|'`
	@@ -1324,10 +1524,14 @@
1524	;;
1525	linux*)
1526	kernel=linux
1527	os=`echo "$basic_os" \| sed -e 's\|linux\|gnu\|'`
1528	;;
1529	managarm*)
1530	kernel=managarm
1531	os=`echo "$basic_os" \| sed -e 's\|managarm\|mlibc\|'`
1532	;;
1533	*)
1534	kernel=
1535	os=$basic_os
1536	;;
1537	esac
	@@ -1350,10 +1554,27 @@
1554	solaris)
1555	os=solaris2
1556	;;
1557	unixware*)
1558	os=sysv4.2uw
1559	;;
1560	# The marketing names for NeXT's operating systems were
1561	# NeXTSTEP, NeXTSTEP 2, OpenSTEP 3, OpenSTEP 4. 'openstep' is
1562	# mapped to 'openstep3', but 'openstep1' and 'openstep2' are
1563	# mapped to 'nextstep' and 'nextstep2', consistent with the
1564	# treatment of SunOS/Solaris.
1565	ns \| ns1 \| nextstep \| nextstep1 \| openstep1)
1566	os=nextstep
1567	;;
1568	ns2 \| nextstep2 \| openstep2)
1569	os=nextstep2
1570	;;
1571	ns3 \| nextstep3 \| openstep \| openstep3)
1572	os=openstep3
1573	;;
1574	ns4 \| nextstep4 \| openstep4)
1575	os=openstep4
1576	;;
1577	# es1800 is here to avoid being matched by es* (a different OS)
1578	es1800*)
1579	os=ose
1580	;;
	@@ -1421,10 +1642,11 @@
1642	wince*)
1643	os=wince
1644	;;
1645	utek*)
1646	os=bsd
1647	vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|tektronix\|'`
1648	;;
1649	dynix*)
1650	os=bsd
1651	;;
1652	acis*)
	@@ -1437,25 +1659,29 @@
1659	os=syllable
1660	;;
1661	386bsd)
1662	os=bsd
1663	;;
1664	ctix*)
1665	os=sysv
1666	vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|convergent\|'`
1667	;;
1668	uts*)
1669	os=sysv
1670	;;
1671	nova*)
1672	kernel=rtmk
1673	os=nova


1674	;;
1675	# Preserve the version number of sinix5.
1676	sinix5.*)
1677	os=`echo "$os" \| sed -e 's\|sinix\|sysv\|'`
1678	vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|sni\|'`
1679	;;
1680	sinix*)
1681	os=sysv4
1682	vendor=`echo "$vendor" \| sed -e 's\|^unknown$\|sni\|'`
1683	;;
1684	tpf*)
1685	os=tpf
1686	;;
1687	triton*)
	@@ -1489,13 +1715,19 @@
1715	case $cpu in
1716	arm*)
1717	os=eabi
1718	;;
1719	*)
1720	os=
1721	obj=elf
1722	;;
1723	esac
1724	;;
1725	aout* \| coff* \| elf* \| pe*)
1726	# These are machine code file formats, not OSes
1727	obj=$os
1728	os=
1729	;;
1730	*)
1731	# No normalization, but not necessarily accepted, that comes below.
1732	;;
1733	esac
	@@ -1511,47 +1743,57 @@
1743	# that MANUFACTURER isn't an operating system. Otherwise, code above
1744	# will signal an error saying that MANUFACTURER isn't an operating
1745	# system, and we'll never get to this point.
1746
1747	kernel=
1748	obj=
1749	case $cpu-$vendor in
1750	score-*)
1751	os=
1752	obj=elf
1753	;;
1754	spu-*)
1755	os=
1756	obj=elf
1757	;;
1758	*-acorn)
1759	os=riscix1.2
1760	;;
1761	arm*-rebel)
1762	kernel=linux
1763	os=gnu
1764	;;
1765	arm*-semi)
1766	os=
1767	obj=aout
1768	;;
1769	c4x-* \| tic4x-*)
1770	os=
1771	obj=coff
1772	;;
1773	c8051-*)
1774	os=
1775	obj=elf
1776	;;
1777	clipper-intergraph)
1778	os=clix
1779	;;
1780	hexagon-*)
1781	os=
1782	obj=elf
1783	;;
1784	tic54x-*)
1785	os=
1786	obj=coff
1787	;;
1788	tic55x-*)
1789	os=
1790	obj=coff
1791	;;
1792	tic6x-*)
1793	os=
1794	obj=coff
1795	;;
1796	# This must come before the *-dec entry.
1797	pdp10-*)
1798	os=tops20
1799	;;
	@@ -1569,32 +1811,47 @@
1811	;;
1812	m68000-sun)
1813	os=sunos3
1814	;;
1815	m68*-cisco)
1816	os=
1817	obj=aout
1818	;;
1819	mep-*)
1820	os=
1821	obj=elf
1822	;;
1823	# The -sgi and -siemens entries must be before the mips- entry
1824	# or we get the wrong os.
1825	*-sgi)
1826	os=irix
1827	;;
1828	*-siemens)
1829	os=sysv4
1830	;;
1831	mips*-cisco)
1832	os=
1833	obj=elf
1834	;;
1835	mips-\|nanomips-)
1836	os=
1837	obj=elf
1838	;;
1839	or32-*)
1840	os=
1841	obj=coff
1842	;;
1843	# This must be before the sparc-* entry or we get the wrong os.
1844	*-tti)
1845	os=sysv3
1846	;;
1847	sparc-* \| *-sun)
1848	os=sunos4.1.1
1849	;;
1850	pru-*)
1851	os=
1852	obj=elf
1853	;;
1854	*-be)
1855	os=beos
1856	;;
1857	*-ibm)
	@@ -1614,11 +1871,11 @@
1871	;;
1872	*-hp)
1873	os=hpux
1874	;;
1875	*-hitachi)
1876	os=hiuxwe2
1877	;;
1878	i860-* \| -att \| -ncr \| -altos \| -motorola \| *-convergent)
1879	os=sysv
1880	;;
1881	*-cbm)
	@@ -1658,27 +1915,23 @@
1915	os=bsd
1916	;;
1917	*-encore)
1918	os=bsd
1919	;;






1920	*-masscomp)
1921	os=rtu
1922	;;
1923	f30[01]-fujitsu \| f700-fujitsu)
1924	os=uxpv
1925	;;
1926	*-rom68k)
1927	os=
1928	obj=coff
1929	;;
1930	-bug)
1931	os=
1932	obj=coff
1933	;;
1934	*-apple)
1935	os=macos
1936	;;
1937	-atari)
	@@ -1692,96 +1945,324 @@
1945	;;
1946	esac
1947
1948	fi
1949
1950	# Now, validate our (potentially fixed-up) individual pieces (OS, OBJ).
1951
1952	case $os in
1953	# Sometimes we do "kernel-libc", so those need to count as OSes.
1954	llvm* \| musl* \| newlib* \| relibc* \| uclibc*)
1955	;;
1956	# Likewise for "kernel-abi"
1957	eabi* \| gnueabi*)
1958	;;
1959	# VxWorks passes extra cpu info in the 4th filed.
1960	simlinux \| simwindows \| spe)
1961	;;
1962	# See `case $cpu-$os` validation below
1963	ghcjs)
1964	;;
1965	# Now accept the basic system types.

1966	# Each alternative MUST end in a * to match a version number.
1967	abug \
1968	\| aix* \
1969	\| amdhsa* \
1970	\| amigados* \
1971	\| amigaos* \
1972	\| android* \
1973	\| aof* \
1974	\| aos* \
1975	\| aros* \
1976	\| atheos* \
1977	\| auroraux* \
1978	\| aux* \
1979	\| beos* \
1980	\| bitrig* \
1981	\| bme* \
1982	\| bosx* \
1983	\| bsd* \
1984	\| cegcc* \
1985	\| chorusos* \
1986	\| chorusrdb* \
1987	\| clix* \
1988	\| cloudabi* \
1989	\| cnk* \
1990	\| conix* \
1991	\| cos* \
1992	\| cxux* \
1993	\| cygwin* \
1994	\| darwin* \
1995	\| dgux* \
1996	\| dicos* \
1997	\| dnix* \
1998	\| domain* \
1999	\| dragonfly* \
2000	\| drops* \
2001	\| ebmon* \
2002	\| ecoff* \
2003	\| ekkobsd* \
2004	\| emscripten* \
2005	\| emx* \
2006	\| es* \
2007	\| fiwix* \
2008	\| freebsd* \
2009	\| fuchsia* \
2010	\| genix* \
2011	\| genode* \
2012	\| glidix* \
2013	\| gnu* \
2014	\| go32* \
2015	\| haiku* \
2016	\| hcos* \
2017	\| hiux* \
2018	\| hms* \
2019	\| hpux* \
2020	\| ieee* \
2021	\| interix* \
2022	\| ios* \
2023	\| iris* \
2024	\| irix* \
2025	\| ironclad* \
2026	\| isc* \
2027	\| its* \
2028	\| l4re* \
2029	\| libertybsd* \
2030	\| lites* \
2031	\| lnews* \
2032	\| luna* \
2033	\| lynxos* \
2034	\| mach* \
2035	\| macos* \
2036	\| magic* \
2037	\| mbr* \
2038	\| midipix* \
2039	\| midnightbsd* \
2040	\| mingw32* \
2041	\| mingw64* \
2042	\| minix* \
2043	\| mint* \
2044	\| mirbsd* \
2045	\| mks* \
2046	\| mlibc* \
2047	\| mmixware* \
2048	\| mon960* \
2049	\| morphos* \
2050	\| moss* \
2051	\| moxiebox* \
2052	\| mpeix* \
2053	\| mpw* \
2054	\| msdos* \
2055	\| msys* \
2056	\| mvs* \
2057	\| nacl* \
2058	\| netbsd* \
2059	\| netware* \
2060	\| newsos* \
2061	\| nextstep* \
2062	\| nindy* \
2063	\| nonstopux* \
2064	\| nova* \
2065	\| nsk* \
2066	\| nucleus* \
2067	\| nx6 \
2068	\| nx7 \
2069	\| oabi* \
2070	\| ohos* \
2071	\| onefs* \
2072	\| openbsd* \
2073	\| openedition* \
2074	\| openstep* \
2075	\| os108* \
2076	\| os2* \
2077	\| os400* \
2078	\| os68k* \
2079	\| os9* \
2080	\| ose* \
2081	\| osf* \
2082	\| oskit* \
2083	\| osx* \
2084	\| palmos* \
2085	\| phoenix* \
2086	\| plan9* \
2087	\| powermax* \
2088	\| powerunix* \
2089	\| proelf* \
2090	\| psos* \
2091	\| psp* \
2092	\| ptx* \
2093	\| pw32* \
2094	\| qnx* \
2095	\| rdos* \
2096	\| redox* \
2097	\| rhapsody* \
2098	\| riscix* \
2099	\| riscos* \
2100	\| rtems* \
2101	\| rtmk* \
2102	\| rtu* \
2103	\| scout* \
2104	\| secbsd* \
2105	\| sei* \
2106	\| serenity* \
2107	\| sim* \
2108	\| skyos* \
2109	\| solaris* \
2110	\| solidbsd* \
2111	\| sortix* \
2112	\| storm-chaos* \
2113	\| sunos \
2114	\| sunos[34]* \
2115	\| superux* \
2116	\| syllable* \
2117	\| sym* \
2118	\| sysv* \
2119	\| tenex* \
2120	\| tirtos* \
2121	\| toppers* \
2122	\| tops10* \
2123	\| tops20* \
2124	\| tpf* \
2125	\| tvos* \
2126	\| twizzler* \
2127	\| uclinux* \
2128	\| udi* \
2129	\| udk* \
2130	\| ultrix* \
2131	\| unicos* \
2132	\| uniplus* \
2133	\| unleashed* \
2134	\| unos* \
2135	\| uwin* \
2136	\| uxpv* \
2137	\| v88r* \
2138	\|vms \
2139	\| vos* \
2140	\| vsta* \
2141	\| vxsim* \
2142	\| vxworks* \
2143	\| wasi* \
2144	\| watchos* \
2145	\| wince* \
2146	\| windiss* \
2147	\| windows* \
2148	\| winnt* \
2149	\| xenix* \
2150	\| xray* \
2151	\| zephyr* \
2152	\| zvmoe* )
2153	;;
2154	# This one is extra strict with allowed versions
2155	sco3.2v2 \| sco3.2v[4-9]* \| sco5v6*)
2156	# Don't forget version if it is 3.2v4 or newer.
2157	;;
2158	# This refers to builds using the UEFI calling convention
2159	# (which depends on the architecture) and PE file format.
2160	# Note that this is both a different calling convention and
2161	# different file format than that of GNU-EFI
2162	# (x86_64-w64-mingw32).
2163	uefi)
2164	;;
2165	none)
2166	;;
2167	kernel* \| msvc* )
2168	# Restricted further below
2169	;;
2170	'')
2171	if test x"$obj" = x
2172	then
2173	echo "Invalid configuration '$1': Blank OS only allowed with explicit machine code file format" 1>&2
2174	fi
2175	;;
2176	*)
2177	echo "Invalid configuration '$1': OS '$os' not recognized" 1>&2
2178	exit 1
2179	;;
2180	esac
2181
2182	case $obj in
2183	aout* \| coff* \| elf* \| pe*)
2184	;;
2185	'')
2186	# empty is fine
2187	;;
2188	*)
2189	echo "Invalid configuration '$1': Machine code format '$obj' not recognized" 1>&2
2190	exit 1
2191	;;
2192	esac
2193
2194	# Here we handle the constraint that a (synthetic) cpu and os are
2195	# valid only in combination with each other and nowhere else.
2196	case $cpu-$os in
2197	# The "javascript-unknown-ghcjs" triple is used by GHC; we
2198	# accept it here in order to tolerate that, but reject any
2199	# variations.
2200	javascript-ghcjs)
2201	;;
2202	javascript-* \| *-ghcjs)
2203	echo "Invalid configuration '$1': cpu '$cpu' is not valid with os '$os$obj'" 1>&2
2204	exit 1
2205	;;
2206	esac
2207
2208	# As a final step for OS-related things, validate the OS-kernel combination
2209	# (given a valid OS), if there is a kernel.
2210	case $kernel-$os-$obj in
2211	linux-gnu- \| linux-android- \| linux-dietlibc- \| linux-llvm- \
2212	\| linux-mlibc- \| linux-musl- \| linux-newlib*- \
2213	\| linux-relibc- \| linux-uclibc- \| linux-ohos*- )
2214	;;
2215	uclinux-uclibc- \| uclinux-gnu- )
2216	;;
2217	managarm-mlibc- \| managarm-kernel- )
2218	;;
2219	windows-msvc-)
2220	;;
2221	-dietlibc- \| -llvm- \| -mlibc- \| -musl- \| -newlib- \| -relibc- \
2222	\| -uclibc*- )
2223	# These are just libc implementations, not actual OSes, and thus
2224	# require a kernel.
2225	echo "Invalid configuration '$1': libc '$os' needs explicit kernel." 1>&2
2226	exit 1
2227	;;
2228	-kernel*- )
2229	echo "Invalid configuration '$1': '$os' needs explicit kernel." 1>&2
2230	exit 1
2231	;;
2232	-kernel- )
2233	echo "Invalid configuration '$1': '$kernel' does not support '$os'." 1>&2
2234	exit 1
2235	;;
2236	-msvc- )
2237	echo "Invalid configuration '$1': '$os' needs 'windows'." 1>&2
2238	exit 1
2239	;;
2240	kfreebsd-gnu- \| knetbsd-gnu- \| netbsd-gnu- \| kopensolaris-gnu-)
2241	;;
2242	vxworks-simlinux- \| vxworks-simwindows- \| vxworks-spe-)
2243	;;
2244	nto-qnx*-)
2245	;;
2246	os2-emx-)
2247	;;
2248	rtmk-nova-)
2249	;;
2250	-eabi- \| -gnueabi-)
2251	;;
2252	none--*)
2253	# None (no kernel, i.e. freestanding / bare metal),
2254	# can be paired with an machine code file format
2255	;;
2256	-*-)
2257	# Blank kernel with real OS is always fine.
2258	;;
2259	--*)
2260	# Blank kernel and OS with real machine code file format is always fine.
2261	;;
2262	--*)
2263	echo "Invalid configuration '$1': Kernel '$kernel' not known to work with OS '$os'." 1>&2
2264	exit 1
2265	;;
2266	esac
2267
2268	# Here we handle the case where we know the os, and the CPU type, but not the
	@@ -1790,11 +2271,11 @@
2271	unknown)
2272	case $cpu-$os in
2273	-riscix)
2274	vendor=acorn
2275	;;
2276	-sunos \| -solaris)
2277	vendor=sun
2278	;;
2279	-cnk \| -aix)
2280	vendor=ibm
2281	;;
	@@ -1860,14 +2341,14 @@
2341	;;
2342	esac
2343	;;
2344	esac
2345
2346	echo "$cpu-$vendor${kernel:+-$kernel}${os:+-$os}${obj:+-$obj}"
2347	exit
2348
2349	# Local variables:
2350	# eval: (add-hook 'before-save-hook 'time-stamp)
2351	# time-stamp-start: "timestamp='"
2352	# time-stamp-format: "%:y-%02m-%02d"
2353	# time-stamp-end: "'"
2354	# End:
2355

M autosetup/autosetup-find-tclsh

+3 -3

		--- autosetup/autosetup-find-tclsh
		+++ autosetup/autosetup-find-tclsh
		@@ -3,14 +3,14 @@
3	3	# If not found, builds a bootstrap jimsh in current dir from source
4	4	# Prefer $autosetup_tclsh if is set in the environment (unless ./jimsh0 works)
5	5	# If an argument is given, use that as the test instead of autosetup-test-tclsh
6	6	d="`dirname "$0"`"
7	7	for tclsh in ./jimsh0 $autosetup_tclsh jimsh tclsh tclsh8.5 tclsh8.6 tclsh8.7; do
8		- { $tclsh "$d/${1-autosetup-test-tclsh}"; } 2>/dev/null && exit 0
	8	+ { $tclsh "$d/${1-autosetup-test-tclsh}"; } 2>/dev/null && exit 0
9	9	done
10	10	echo 1>&2 "No installed jimsh or tclsh, building local bootstrap jimsh0"
11	11	for cc in ${CC_FOR_BUILD:-cc} gcc; do
12		- { $cc -o jimsh0 "$d/jimsh0.c"; } 2>&1 >/dev/null \|\| continue
13		- ./jimsh0 "$d/${1-autosetup-test-tclsh}" && exit 0
	12	+ { $cc -o jimsh0 "$d/jimsh0.c"; } >/dev/null 2>&1 \|\| continue
	13	+ ./jimsh0 "$d/${1-autosetup-test-tclsh}" && exit 0
14	14	done
15	15	echo 1>&2 "No working C compiler found. Tried ${CC_FOR_BUILD:-cc} and gcc."
16	16	echo false
17	17

	--- autosetup/autosetup-find-tclsh
	+++ autosetup/autosetup-find-tclsh
	@@ -3,14 +3,14 @@
3	# If not found, builds a bootstrap jimsh in current dir from source
4	# Prefer $autosetup_tclsh if is set in the environment (unless ./jimsh0 works)
5	# If an argument is given, use that as the test instead of autosetup-test-tclsh
6	d="`dirname "$0"`"
7	for tclsh in ./jimsh0 $autosetup_tclsh jimsh tclsh tclsh8.5 tclsh8.6 tclsh8.7; do
8	{ $tclsh "$d/${1-autosetup-test-tclsh}"; } 2>/dev/null && exit 0
9	done
10	echo 1>&2 "No installed jimsh or tclsh, building local bootstrap jimsh0"
11	for cc in ${CC_FOR_BUILD:-cc} gcc; do
12	{ $cc -o jimsh0 "$d/jimsh0.c"; } 2>&1 >/dev/null \|\| continue
13	./jimsh0 "$d/${1-autosetup-test-tclsh}" && exit 0
14	done
15	echo 1>&2 "No working C compiler found. Tried ${CC_FOR_BUILD:-cc} and gcc."
16	echo false
17

	--- autosetup/autosetup-find-tclsh
	+++ autosetup/autosetup-find-tclsh
	@@ -3,14 +3,14 @@
3	# If not found, builds a bootstrap jimsh in current dir from source
4	# Prefer $autosetup_tclsh if is set in the environment (unless ./jimsh0 works)
5	# If an argument is given, use that as the test instead of autosetup-test-tclsh
6	d="`dirname "$0"`"
7	for tclsh in ./jimsh0 $autosetup_tclsh jimsh tclsh tclsh8.5 tclsh8.6 tclsh8.7; do
8	{ $tclsh "$d/${1-autosetup-test-tclsh}"; } 2>/dev/null && exit 0
9	done
10	echo 1>&2 "No installed jimsh or tclsh, building local bootstrap jimsh0"
11	for cc in ${CC_FOR_BUILD:-cc} gcc; do
12	{ $cc -o jimsh0 "$d/jimsh0.c"; } >/dev/null 2>&1 \|\| continue
13	./jimsh0 "$d/${1-autosetup-test-tclsh}" && exit 0
14	done
15	echo 1>&2 "No working C compiler found. Tried ${CC_FOR_BUILD:-cc} and gcc."
16	echo false
17

M autosetup/cc.tcl

+1 -1

		--- autosetup/cc.tcl
		+++ autosetup/cc.tcl
		@@ -3,11 +3,11 @@
3	3
4	4	# @synopsis:
5	5	#
6	6	# The 'cc' module supports checking various 'features' of the C or C++
7	7	# compiler/linker environment. Common commands are 'cc-check-includes',
8		-# 'cc-check-types', 'cc-check-functions', 'cc-with', 'make-config-header' and 'make-template'.
	8	+# 'cc-check-types', 'cc-check-functions', 'cc-with' and 'make-config-header'
9	9	#
10	10	# The following environment variables are used if set:
11	11	#
12	12	## CC - C compiler
13	13	## CXX - C++ compiler
14	14

	--- autosetup/cc.tcl
	+++ autosetup/cc.tcl
	@@ -3,11 +3,11 @@
3
4	# @synopsis:
5	#
6	# The 'cc' module supports checking various 'features' of the C or C++
7	# compiler/linker environment. Common commands are 'cc-check-includes',
8	# 'cc-check-types', 'cc-check-functions', 'cc-with', 'make-config-header' and 'make-template'.
9	#
10	# The following environment variables are used if set:
11	#
12	## CC - C compiler
13	## CXX - C++ compiler
14

	--- autosetup/cc.tcl
	+++ autosetup/cc.tcl
	@@ -3,11 +3,11 @@
3
4	# @synopsis:
5	#
6	# The 'cc' module supports checking various 'features' of the C or C++
7	# compiler/linker environment. Common commands are 'cc-check-includes',
8	# 'cc-check-types', 'cc-check-functions', 'cc-with' and 'make-config-header'
9	#
10	# The following environment variables are used if set:
11	#
12	## CC - C compiler
13	## CXX - C++ compiler
14

M extsrc/shell.c

+3370 -1991

		--- extsrc/shell.c
		+++ extsrc/shell.c
		@@ -120,13 +120,10 @@
120	120	#include <math.h>
121	121	#include "sqlite3.h"
122	122	typedef sqlite3_int64 i64;
123	123	typedef sqlite3_uint64 u64;
124	124	typedef unsigned char u8;
125		-#if SQLITE_USER_AUTHENTICATION
126		-# include "sqlite3userauth.h"
127		-#endif
128	125	#include <ctype.h>
129	126	#include <stdarg.h>
130	127
131	128	#if !defined(_WIN32) && !defined(WIN32)
132	129	# include <signal.h>
		@@ -208,12 +205,10 @@
208	205	# define unlink _unlink
209	206	# endif
210	207	# ifndef strdup
211	208	# define strdup _strdup
212	209	# endif
213		-# undef popen
214		-# define popen _popen
215	210	# undef pclose
216	211	# define pclose _pclose
217	212	# endif
218	213	#else
219	214	/* Make sure isatty() has a prototype. */
		@@ -253,10 +248,370 @@
253	248	/* string conversion routines only needed on Win32 */
254	249	extern char *sqlite3_win32_unicode_to_utf8(LPCWSTR);
255	250	extern LPWSTR sqlite3_win32_utf8_to_unicode(const char *zText);
256	251	#endif
257	252
	253	+/*********************** Begin ../ext/misc/sqlite3_stdio.h ****************/
	254	+/*
	255	+** 2024-09-24
	256	+**
	257	+** The author disclaims copyright to this source code. In place of
	258	+** a legal notice, here is a blessing:
	259	+**
	260	+** May you do good and not evil.
	261	+** May you find forgiveness for yourself and forgive others.
	262	+** May you share freely, never taking more than you give.
	263	+**
	264	+*************************************************************************
	265	+**
	266	+** This header file contains definitions of interfaces that provide
	267	+** cross-platform I/O for UTF-8 content.
	268	+**
	269	+** On most platforms, the interfaces definitions in this file are
	270	+** just #defines. For example sqlite3_fopen() is a macro that resolves
	271	+** to the standard fopen() in the C-library.
	272	+**
	273	+** But Windows does not have a standard C-library, at least not one that
	274	+** can handle UTF-8. So for windows build, the interfaces resolve to new
	275	+** C-language routines contained in the separate sqlite3_stdio.c source file.
	276	+**
	277	+** So on all non-Windows platforms, simply #include this header file and
	278	+** use the interfaces defined herein. Then to run your application on Windows,
	279	+** also link in the accompanying sqlite3_stdio.c source file when compiling
	280	+** to get compatible interfaces.
	281	+*/
	282	+#ifndef _SQLITE3_STDIO_H_
	283	+#define _SQLITE3_STDIO_H_ 1
	284	+#ifdef _WIN32
	285	+/** Definitions For Windows **/
	286	+#include <stdio.h>
	287	+#include <windows.h>
	288	+
	289	+FILE sqlite3_fopen(const char zFilename, const char *zMode);
	290	+FILE sqlite3_popen(const char zCommand, const char *type);
	291	+char sqlite3_fgets(char s, int size, FILE *stream);
	292	+int sqlite3_fputs(const char s, FILE stream);
	293	+int sqlite3_fprintf(FILE stream, const char format, ...);
	294	+void sqlite3_fsetmode(FILE *stream, int mode);
	295	+
	296	+
	297	+#else
	298	+/** Definitions For All Other Platforms **/
	299	+#include <stdio.h>
	300	+#define sqlite3_fopen fopen
	301	+#define sqlite3_popen popen
	302	+#define sqlite3_fgets fgets
	303	+#define sqlite3_fputs fputs
	304	+#define sqlite3_fprintf fprintf
	305	+#define sqlite3_fsetmode(F,X) /no-op/
	306	+
	307	+#endif
	308	+#endif /* _SQLITE3_STDIO_H_ */
	309	+
	310	+/*********************** End ../ext/misc/sqlite3_stdio.h ******************/
	311	+/*********************** Begin ../ext/misc/sqlite3_stdio.c ****************/
	312	+/*
	313	+** 2024-09-24
	314	+**
	315	+** The author disclaims copyright to this source code. In place of
	316	+** a legal notice, here is a blessing:
	317	+**
	318	+** May you do good and not evil.
	319	+** May you find forgiveness for yourself and forgive others.
	320	+** May you share freely, never taking more than you give.
	321	+**
	322	+*************************************************************************
	323	+**
	324	+** Implementation of standard I/O interfaces for UTF-8 that are missing
	325	+** on Windows.
	326	+*/
	327	+#ifdef _WIN32 /* This file is a no-op on all platforms except Windows */
	328	+#ifndef _SQLITE3_STDIO_H_
	329	+/* #include "sqlite3_stdio.h" */
	330	+#endif
	331	+#undef WIN32_LEAN_AND_MEAN
	332	+#define WIN32_LEAN_AND_MEAN
	333	+#include <windows.h>
	334	+#include <stdlib.h>
	335	+#include <string.h>
	336	+#include <stdio.h>
	337	+#include <assert.h>
	338	+/* #include "sqlite3.h" */
	339	+#include <ctype.h>
	340	+#include <stdarg.h>
	341	+#include <io.h>
	342	+#include <fcntl.h>
	343	+
	344	+/*
	345	+** If the SQLITE_U8TEXT_ONLY option is defined, then use O_U8TEXT
	346	+** when appropriate on all output. (Sometimes use O_BINARY when
	347	+** rendering ASCII text in cases where NL-to-CRLF expansion would
	348	+** not be correct.)
	349	+**
	350	+** If the SQLITE_U8TEXT_STDIO option is defined, then use O_U8TEXT
	351	+** when appropriate when writing to stdout or stderr. Use O_BINARY
	352	+** or O_TEXT (depending on things like the .mode and the .crlf setting
	353	+** in the CLI, or other context clues in other applications) for all
	354	+** other output channels.
	355	+**
	356	+** The default behavior, if neither of the above is defined is to
	357	+** use O_U8TEXT when writing to the Windows console (or anything
	358	+** else for which _isatty() returns true) and to use O_BINARY or O_TEXT
	359	+** for all other output channels.
	360	+*/
	361	+#if defined(SQLITE_U8TEXT_ONLY)
	362	+# define UseWtextForOutput(fd) 1
	363	+# define UseWtextForInput(fd) 1
	364	+# define IsConsole(fd) _isatty(_fileno(fd))
	365	+#elif defined(SQLITE_U8TEXT_STDIO)
	366	+# define UseWtextForOutput(fd) ((fd)==stdout \|\| (fd)==stderr)
	367	+# define UseWtextForInput(fd) ((fd)==stdin)
	368	+# define IsConsole(fd) _isatty(_fileno(fd))
	369	+#else
	370	+# define UseWtextForOutput(fd) _isatty(_fileno(fd))
	371	+# define UseWtextForInput(fd) _isatty(_fileno(fd))
	372	+# define IsConsole(fd) 1
	373	+#endif
	374	+
	375	+/*
	376	+** Global variables determine if simulated O_BINARY mode is to be
	377	+** used for stdout or other, respectively. Simulated O_BINARY mode
	378	+** means the mode is usually O_BINARY, but switches to O_U8TEXT for
	379	+** unicode characters U+0080 or greater (any character that has a
	380	+** multi-byte representation in UTF-8). This is the only way we
	381	+** have found to render Unicode characters on a Windows console while
	382	+** at the same time avoiding undesirable \n to \r\n translation.
	383	+*/
	384	+static int simBinaryStdout = 0;
	385	+static int simBinaryOther = 0;
	386	+
	387	+
	388	+/*
	389	+** Determine if simulated binary mode should be used for output to fd
	390	+*/
	391	+static int UseBinaryWText(FILE *fd){
	392	+ if( fd==stdout \|\| fd==stderr ){
	393	+ return simBinaryStdout;
	394	+ }else{
	395	+ return simBinaryOther;
	396	+ }
	397	+}
	398	+
	399	+
	400	+/*
	401	+** Work-alike for the fopen() routine from the standard C library.
	402	+*/
	403	+FILE sqlite3_fopen(const char zFilename, const char *zMode){
	404	+ FILE *fp = 0;
	405	+ wchar_t b1, b2;
	406	+ int sz1, sz2;
	407	+
	408	+ sz1 = (int)strlen(zFilename);
	409	+ sz2 = (int)strlen(zMode);
	410	+ b1 = sqlite3_malloc( (sz1+1)*sizeof(b1[0]) );
	411	+ b2 = sqlite3_malloc( (sz2+1)*sizeof(b1[0]) );
	412	+ if( b1 && b2 ){
	413	+ sz1 = MultiByteToWideChar(CP_UTF8, 0, zFilename, sz1, b1, sz1);
	414	+ b1[sz1] = 0;
	415	+ sz2 = MultiByteToWideChar(CP_UTF8, 0, zMode, sz2, b2, sz2);
	416	+ b2[sz2] = 0;
	417	+ fp = _wfopen(b1, b2);
	418	+ }
	419	+ sqlite3_free(b1);
	420	+ sqlite3_free(b2);
	421	+ simBinaryOther = 0;
	422	+ return fp;
	423	+}
	424	+
	425	+
	426	+/*
	427	+** Work-alike for the popen() routine from the standard C library.
	428	+*/
	429	+FILE sqlite3_popen(const char zCommand, const char *zMode){
	430	+ FILE *fp = 0;
	431	+ wchar_t b1, b2;
	432	+ int sz1, sz2;
	433	+
	434	+ sz1 = (int)strlen(zCommand);
	435	+ sz2 = (int)strlen(zMode);
	436	+ b1 = sqlite3_malloc( (sz1+1)*sizeof(b1[0]) );
	437	+ b2 = sqlite3_malloc( (sz2+1)*sizeof(b1[0]) );
	438	+ if( b1 && b2 ){
	439	+ sz1 = MultiByteToWideChar(CP_UTF8, 0, zCommand, sz1, b1, sz1);
	440	+ b1[sz1] = 0;
	441	+ sz2 = MultiByteToWideChar(CP_UTF8, 0, zMode, sz2, b2, sz2);
	442	+ b2[sz2] = 0;
	443	+ fp = _wpopen(b1, b2);
	444	+ }
	445	+ sqlite3_free(b1);
	446	+ sqlite3_free(b2);
	447	+ return fp;
	448	+}
	449	+
	450	+/*
	451	+** Work-alike for fgets() from the standard C library.
	452	+*/
	453	+char sqlite3_fgets(char buf, int sz, FILE *in){
	454	+ if( UseWtextForInput(in) ){
	455	+ /* When reading from the command-prompt in Windows, it is necessary
	456	+ ** to use _O_WTEXT input mode to read UTF-16 characters, then translate
	457	+ ** that into UTF-8. Otherwise, non-ASCII characters all get translated
	458	+ ** into '?'.
	459	+ */
	460	+ wchar_t b1 = sqlite3_malloc( szsizeof(wchar_t) );
	461	+ if( b1==0 ) return 0;
	462	+#ifndef SQLITE_USE_STDIO_FOR_CONSOLE
	463	+ DWORD nRead = 0;
	464	+ if( IsConsole(in)
	465	+ && ReadConsoleW(GetStdHandle(STD_INPUT_HANDLE), b1, sz, &nRead, 0)
	466	+ ){
	467	+ b1[nRead] = 0;
	468	+ }else
	469	+#endif
	470	+ {
	471	+ _setmode(_fileno(in), IsConsole(in) ? _O_WTEXT : _O_U8TEXT);
	472	+ if( fgetws(b1, sz/4, in)==0 ){
	473	+ sqlite3_free(b1);
	474	+ return 0;
	475	+ }
	476	+ }
	477	+ WideCharToMultiByte(CP_UTF8, 0, b1, -1, buf, sz, 0, 0);
	478	+ sqlite3_free(b1);
	479	+ return buf;
	480	+ }else{
	481	+ /* Reading from a file or other input source, just read bytes without
	482	+ ** any translation. */
	483	+ return fgets(buf, sz, in);
	484	+ }
	485	+}
	486	+
	487	+/*
	488	+** Send ASCII text as O_BINARY. But for Unicode characters U+0080 and
	489	+** greater, switch to O_U8TEXT.
	490	+*/
	491	+static void piecemealOutput(wchar_t b1, int sz, FILE out){
	492	+ int i;
	493	+ wchar_t c;
	494	+ while( sz>0 ){
	495	+ for(i=0; i<sz && b1[i]>=0x80; i++){}
	496	+ if( i>0 ){
	497	+ c = b1[i];
	498	+ b1[i] = 0;
	499	+ fflush(out);
	500	+ _setmode(_fileno(out), _O_U8TEXT);
	501	+ fputws(b1, out);
	502	+ fflush(out);
	503	+ b1 += i;
	504	+ b1[0] = c;
	505	+ sz -= i;
	506	+ }else{
	507	+ fflush(out);
	508	+ _setmode(_fileno(out), _O_TEXT);
	509	+ _setmode(_fileno(out), _O_BINARY);
	510	+ fwrite(&b1[0], 1, 1, out);
	511	+ for(i=1; i<sz && b1[i]<0x80; i++){
	512	+ fwrite(&b1[i], 1, 1, out);
	513	+ }
	514	+ fflush(out);
	515	+ _setmode(_fileno(out), _O_U8TEXT);
	516	+ b1 += i;
	517	+ sz -= i;
	518	+ }
	519	+ }
	520	+}
	521	+
	522	+/*
	523	+** Work-alike for fputs() from the standard C library.
	524	+*/
	525	+int sqlite3_fputs(const char z, FILE out){
	526	+ if( !UseWtextForOutput(out) ){
	527	+ /* Writing to a file or other destination, just write bytes without
	528	+ ** any translation. */
	529	+ return fputs(z, out);
	530	+ }else{
	531	+ /* One must use UTF16 in order to get unicode support when writing
	532	+ ** to the console on Windows.
	533	+ */
	534	+ int sz = (int)strlen(z);
	535	+ wchar_t b1 = sqlite3_malloc( (sz+1)sizeof(wchar_t) );
	536	+ if( b1==0 ) return 0;
	537	+ sz = MultiByteToWideChar(CP_UTF8, 0, z, sz, b1, sz);
	538	+ b1[sz] = 0;
	539	+
	540	+#ifndef SQLITE_STDIO_FOR_CONSOLE
	541	+ DWORD nWr = 0;
	542	+ if( IsConsole(out)
	543	+ && WriteConsoleW(GetStdHandle(STD_OUTPUT_HANDLE),b1,sz,&nWr,0)
	544	+ ){
	545	+ /* If writing to the console, then the WriteConsoleW() is all we
	546	+ ** need to do. */
	547	+ }else
	548	+#endif
	549	+ {
	550	+ /* For non-console I/O, or if SQLITE_USE_STDIO_FOR_CONSOLE is defined
	551	+ ** then write using the standard library. */
	552	+ _setmode(_fileno(out), _O_U8TEXT);
	553	+ if( UseBinaryWText(out) ){
	554	+ piecemealOutput(b1, sz, out);
	555	+ }else{
	556	+ fputws(b1, out);
	557	+ }
	558	+ }
	559	+ sqlite3_free(b1);
	560	+ return 0;
	561	+ }
	562	+}
	563	+
	564	+
	565	+/*
	566	+** Work-alike for fprintf() from the standard C library.
	567	+*/
	568	+int sqlite3_fprintf(FILE out, const char zFormat, ...){
	569	+ int rc;
	570	+ if( UseWtextForOutput(out) ){
	571	+ /* When writing to the command-prompt in Windows, it is necessary
	572	+ ** to use _O_WTEXT input mode and write UTF-16 characters.
	573	+ */
	574	+ char *z;
	575	+ va_list ap;
	576	+
	577	+ va_start(ap, zFormat);
	578	+ z = sqlite3_vmprintf(zFormat, ap);
	579	+ va_end(ap);
	580	+ sqlite3_fputs(z, out);
	581	+ rc = (int)strlen(z);
	582	+ sqlite3_free(z);
	583	+ }else{
	584	+ /* Writing to a file or other destination, just write bytes without
	585	+ ** any translation. */
	586	+ va_list ap;
	587	+ va_start(ap, zFormat);
	588	+ rc = vfprintf(out, zFormat, ap);
	589	+ va_end(ap);
	590	+ }
	591	+ return rc;
	592	+}
	593	+
	594	+/*
	595	+** Set the mode for an output stream. mode argument is typically _O_BINARY or
	596	+** _O_TEXT.
	597	+*/
	598	+void sqlite3_fsetmode(FILE *fp, int mode){
	599	+ if( !UseWtextForOutput(fp) ){
	600	+ fflush(fp);
	601	+ _setmode(_fileno(fp), mode);
	602	+ }else if( fp==stdout \|\| fp==stderr ){
	603	+ simBinaryStdout = (mode==_O_BINARY);
	604	+ }else{
	605	+ simBinaryOther = (mode==_O_BINARY);
	606	+ }
	607	+}
	608	+
	609	+#endif /* defined(_WIN32) */
	610	+
	611	+/*********************** End ../ext/misc/sqlite3_stdio.c ******************/
	612	+
258	613	/* Use console I/O package as a direct INCLUDE. */
259	614	#define SQLITE_INTERNAL_LINKAGE static
260	615
261	616	#ifdef SQLITE_SHELL_FIDDLE
262	617	/* Deselect most features from the console I/O package for Fiddle. */
		@@ -264,1118 +619,13 @@
264	619	# define SQLITE_CIO_NO_CLASSIFY
265	620	# define SQLITE_CIO_NO_TRANSLATE
266	621	# define SQLITE_CIO_NO_SETMODE
267	622	# define SQLITE_CIO_NO_FLUSH
268	623	#endif
269		-/*********************** Begin ../ext/consio/console_io.h ****************/
270		-/*
271		-** 2023 November 1
272		-**
273		-** The author disclaims copyright to this source code. In place of
274		-** a legal notice, here is a blessing:
275		-**
276		-** May you do good and not evil.
277		-** May you find forgiveness for yourself and forgive others.
278		-** May you share freely, never taking more than you give.
279		-**
280		-********************************************************************************
281		-** This file exposes various interfaces used for console and other I/O
282		-** by the SQLite project command-line tools. These interfaces are used
283		-** at either source conglomeration time, compilation time, or run time.
284		-** This source provides for either inclusion into conglomerated,
285		-** "single-source" forms or separate compilation then linking.
286		-**
287		-** Platform dependencies are "hidden" here by various stratagems so
288		-** that, provided certain conditions are met, the programs using this
289		-** source or object code compiled from it need no explicit conditional
290		-** compilation in their source for their console and stream I/O.
291		-**
292		-** The symbols and functionality exposed here are not a public API.
293		-** This code may change in tandem with other project code as needed.
294		-**
295		-** When this .h file and its companion .c are directly incorporated into
296		-** a source conglomeration (such as shell.c), the preprocessor symbol
297		-** CIO_WIN_WC_XLATE is defined as 0 or 1, reflecting whether console I/O
298		-** translation for Windows is effected for the build.
299		-*/
300		-#define HAVE_CONSOLE_IO_H 1
301		-#ifndef SQLITE_INTERNAL_LINKAGE
302		-# define SQLITE_INTERNAL_LINKAGE extern /* external to translation unit */
303		-# include <stdio.h>
304		-#else
305		-# define SHELL_NO_SYSINC /* Better yet, modify mkshellc.tcl for this. */
306		-#endif
307		-
308		-#ifndef SQLITE3_H
309		-/* # include "sqlite3.h" */
310		-#endif
311		-
312		-#ifndef SQLITE_CIO_NO_CLASSIFY
313		-
314		-/* Define enum for use with following function. */
315		-typedef enum StreamsAreConsole {
316		- SAC_NoConsole = 0,
317		- SAC_InConsole = 1, SAC_OutConsole = 2, SAC_ErrConsole = 4,
318		- SAC_AnyConsole = 0x7
319		-} StreamsAreConsole;
320		-
321		-/*
322		-** Classify the three standard I/O streams according to whether
323		-** they are connected to a console attached to the process.
324		-**
325		-** Returns the bit-wise OR of SAC_{In,Out,Err}Console values,
326		-** or SAC_NoConsole if none of the streams reaches a console.
327		-**
328		-** This function should be called before any I/O is done with
329		-** the given streams. As a side-effect, the given inputs are
330		-** recorded so that later I/O operations on them may be done
331		-** differently than the C library FILE* I/O would be done,
332		-** iff the stream is used for the I/O functions that follow,
333		-** and to support the ones that use an implicit stream.
334		-**
335		-** On some platforms, stream or console mode alteration (aka
336		-** "Setup") may be made which is undone by consoleRestore().
337		-*/
338		-SQLITE_INTERNAL_LINKAGE StreamsAreConsole
339		-consoleClassifySetup( FILE pfIn, FILE pfOut, FILE *pfErr );
340		-/* A usual call for convenience: */
341		-#define SQLITE_STD_CONSOLE_INIT() consoleClassifySetup(stdin,stdout,stderr)
342		-
343		-/*
344		-** After an initial call to consoleClassifySetup(...), renew
345		-** the same setup it effected. (A call not after is an error.)
346		-** This will restore state altered by consoleRestore();
347		-**
348		-** Applications which run an inferior (child) process which
349		-** inherits the same I/O streams may call this function after
350		-** such a process exits to guard against console mode changes.
351		-*/
352		-SQLITE_INTERNAL_LINKAGE void consoleRenewSetup(void);
353		-
354		-/*
355		-** Undo any side-effects left by consoleClassifySetup(...).
356		-**
357		-** This should be called after consoleClassifySetup() and
358		-** before the process terminates normally. It is suitable
359		-** for use with the atexit() C library procedure. After
360		-** this call, no console I/O should be done until one of
361		-** console{Classify or Renew}Setup(...) is called again.
362		-**
363		-** Applications which run an inferior (child) process that
364		-** inherits the same I/O streams might call this procedure
365		-** before so that said process will have a console setup
366		-** however users have configured it or come to expect.
367		-*/
368		-SQLITE_INTERNAL_LINKAGE void SQLITE_CDECL consoleRestore( void );
369		-
370		-#else /* defined(SQLITE_CIO_NO_CLASSIFY) */
371		-# define consoleClassifySetup(i,o,e)
372		-# define consoleRenewSetup()
373		-# define consoleRestore()
374		-#endif /* defined(SQLITE_CIO_NO_CLASSIFY) */
375		-
376		-#ifndef SQLITE_CIO_NO_REDIRECT
377		-/*
378		-** Set stream to be used for the functions below which write
379		-** to "the designated X stream", where X is Output or Error.
380		-** Returns the previous value.
381		-**
382		-** Alternatively, pass the special value, invalidFileStream,
383		-** to get the designated stream value without setting it.
384		-**
385		-** Before the designated streams are set, they default to
386		-** those passed to consoleClassifySetup(...), and before
387		-** that is called they default to stdout and stderr.
388		-**
389		-** It is error to close a stream so designated, then, without
390		-** designating another, use the corresponding {o,e}Emit(...).
391		-*/
392		-SQLITE_INTERNAL_LINKAGE FILE *invalidFileStream;
393		-SQLITE_INTERNAL_LINKAGE FILE setOutputStream(FILE pf);
394		-# ifdef CONSIO_SET_ERROR_STREAM
395		-SQLITE_INTERNAL_LINKAGE FILE setErrorStream(FILE pf);
396		-# endif
397		-#else
398		-# define setOutputStream(pf)
399		-# define setErrorStream(pf)
400		-#endif /* !defined(SQLITE_CIO_NO_REDIRECT) */
401		-
402		-#ifndef SQLITE_CIO_NO_TRANSLATE
403		-/*
404		-** Emit output like fprintf(). If the output is going to the
405		-** console and translation from UTF-8 is necessary, perform
406		-** the needed translation. Otherwise, write formatted output
407		-** to the provided stream almost as-is, possibly with newline
408		-** translation as specified by set{Binary,Text}Mode().
409		-*/
410		-SQLITE_INTERNAL_LINKAGE int fPrintfUtf8(FILE pfO, const char zFormat, ...);
411		-/* Like fPrintfUtf8 except stream is always the designated output. */
412		-SQLITE_INTERNAL_LINKAGE int oPrintfUtf8(const char *zFormat, ...);
413		-/* Like fPrintfUtf8 except stream is always the designated error. */
414		-SQLITE_INTERNAL_LINKAGE int ePrintfUtf8(const char *zFormat, ...);
415		-
416		-/*
417		-** Emit output like fputs(). If the output is going to the
418		-** console and translation from UTF-8 is necessary, perform
419		-** the needed translation. Otherwise, write given text to the
420		-** provided stream almost as-is, possibly with newline
421		-** translation as specified by set{Binary,Text}Mode().
422		-*/
423		-SQLITE_INTERNAL_LINKAGE int fPutsUtf8(const char z, FILE pfO);
424		-/* Like fPutsUtf8 except stream is always the designated output. */
425		-SQLITE_INTERNAL_LINKAGE int oPutsUtf8(const char *z);
426		-/* Like fPutsUtf8 except stream is always the designated error. */
427		-SQLITE_INTERNAL_LINKAGE int ePutsUtf8(const char *z);
428		-
429		-/*
430		-** Emit output like fPutsUtf8(), except that the length of the
431		-** accepted char or character sequence is limited by nAccept.
432		-**
433		-** Returns the number of accepted char values.
434		-*/
435		-#ifdef CONSIO_SPUTB
436		-SQLITE_INTERNAL_LINKAGE int
437		-fPutbUtf8(FILE pfOut, const char cBuf, int nAccept);
438		-/* Like fPutbUtf8 except stream is always the designated output. */
439		-#endif
440		-SQLITE_INTERNAL_LINKAGE int
441		-oPutbUtf8(const char *cBuf, int nAccept);
442		-/* Like fPutbUtf8 except stream is always the designated error. */
443		-#ifdef CONSIO_EPUTB
444		-SQLITE_INTERNAL_LINKAGE int
445		-ePutbUtf8(const char *cBuf, int nAccept);
446		-#endif
447		-
448		-/*
449		-** Flush the given output stream. Return non-zero for success, else 0.
450		-*/
451		-#if !defined(SQLITE_CIO_NO_FLUSH) && !defined(SQLITE_CIO_NO_SETMODE)
452		-SQLITE_INTERNAL_LINKAGE int
453		-fFlushBuffer(FILE *pfOut);
454		-#endif
455		-
456		-/*
457		-** Collect input like fgets(...) with special provisions for input
458		-** from the console on such platforms as require same. Newline
459		-** translation may be done as set by set{Binary,Text}Mode().
460		-** As a convenience, pfIn==NULL is treated as stdin.
461		-*/
462		-SQLITE_INTERNAL_LINKAGE char* fGetsUtf8(char cBuf, int ncMax, FILE pfIn);
463		-/* Like fGetsUtf8 except stream is always the designated input. */
464		-/* SQLITE_INTERNAL_LINKAGE char* iGetsUtf8(char cBuf, int ncMax); /
465		-
466		-#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
467		-
468		-#ifndef SQLITE_CIO_NO_SETMODE
469		-/*
470		-** Set given stream for binary mode, where newline translation is
471		-** not done, or for text mode where, for some platforms, newlines
472		-** are translated to the platform's conventional char sequence.
473		-** If bFlush true, flush the stream.
474		-**
475		-** An additional side-effect is that if the stream is one passed
476		-** to consoleClassifySetup() as an output, it is flushed first.
477		-**
478		-** Note that binary/text mode has no effect on console I/O
479		-** translation. On all platforms, newline to the console starts
480		-** a new line and CR,LF chars from the console become a newline.
481		-*/
482		-SQLITE_INTERNAL_LINKAGE void setBinaryMode(FILE *, short bFlush);
483		-SQLITE_INTERNAL_LINKAGE void setTextMode(FILE *, short bFlush);
484		-#endif
485		-
486		-#ifdef SQLITE_CIO_PROMPTED_IN
487		-typedef struct Prompts {
488		- int numPrompts;
489		- const char **azPrompts;
490		-} Prompts;
491		-
492		-/*
493		-** Macros for use of a line editor.
494		-**
495		-** The following macros define operations involving use of a
496		-** line-editing library or simple console interaction.
497		-** A "T" argument is a text (char *) buffer or filename.
498		-** A "N" argument is an integer.
499		-**
500		-** SHELL_ADD_HISTORY(T) // Record text as line(s) of history.
501		-** SHELL_READ_HISTORY(T) // Read history from file named by T.
502		-** SHELL_WRITE_HISTORY(T) // Write history to file named by T.
503		-** SHELL_STIFLE_HISTORY(N) // Limit history to N entries.
504		-**
505		-** A console program which does interactive console input is
506		-** expected to call:
507		-** SHELL_READ_HISTORY(T) before collecting such input;
508		-** SHELL_ADD_HISTORY(T) as record-worthy input is taken;
509		-** SHELL_STIFLE_HISTORY(N) after console input ceases; then
510		-** SHELL_WRITE_HISTORY(T) before the program exits.
511		-*/
512		-
513		-/*
514		-** Retrieve a single line of input text from an input stream.
515		-**
516		-** If pfIn is the input stream passed to consoleClassifySetup(),
517		-** and azPrompt is not NULL, then a prompt is issued before the
518		-** line is collected, as selected by the isContinuation flag.
519		-** Array azPrompt[{0,1}] holds the {main,continuation} prompt.
520		-**
521		-** If zBufPrior is not NULL then it is a buffer from a prior
522		-** call to this routine that can be reused, or will be freed.
523		-**
524		-** The result is stored in space obtained from malloc() and
525		-** must either be freed by the caller or else passed back to
526		-** this function as zBufPrior for reuse.
527		-**
528		-** This function may call upon services of a line-editing
529		-** library to interactively collect line edited input.
530		-*/
531		-SQLITE_INTERNAL_LINKAGE char *
532		-shellGetLine(FILE pfIn, char zBufPrior, int nLen,
533		- short isContinuation, Prompts azPrompt);
534		-#endif /* defined(SQLITE_CIO_PROMPTED_IN) */
535		-/*
536		-** TBD: Define an interface for application(s) to generate
537		-** completion candidates for use by the line-editor.
538		-**
539		-** This may be premature; the CLI is the only application
540		-** that does this. Yet, getting line-editing melded into
541		-** console I/O is desirable because a line-editing library
542		-** may have to establish console operating mode, possibly
543		-** in a way that interferes with the above functionality.
544		-*/
545		-
546		-#if !(defined(SQLITE_CIO_NO_UTF8SCAN)&&defined(SQLITE_CIO_NO_TRANSLATE))
547		-/* Skip over as much z[] input char sequence as is valid UTF-8,
548		-** limited per nAccept char's or whole characters and containing
549		-** no char cn such that ((1<<cn) & ccm)!=0. On return, the
550		-** sequence z:return (inclusive:exclusive) is validated UTF-8.
551		-** Limit: nAccept>=0 => char count, nAccept<0 => character
552		- */
553		-SQLITE_INTERNAL_LINKAGE const char*
554		-zSkipValidUtf8(const char *z, int nAccept, long ccm);
555		-
556		-#endif
557		-
558		-/*********************** End ../ext/consio/console_io.h ******************/
559		-/*********************** Begin ../ext/consio/console_io.c ****************/
560		-/*
561		-** 2023 November 4
562		-**
563		-** The author disclaims copyright to this source code. In place of
564		-** a legal notice, here is a blessing:
565		-**
566		-** May you do good and not evil.
567		-** May you find forgiveness for yourself and forgive others.
568		-** May you share freely, never taking more than you give.
569		-**
570		-********************************************************************************
571		-** This file implements various interfaces used for console and stream I/O
572		-** by the SQLite project command-line tools, as explained in console_io.h .
573		-** Functions prefixed by "SQLITE_INTERNAL_LINKAGE" behave as described there.
574		-*/
575		-
576		-#ifndef SQLITE_CDECL
577		-# define SQLITE_CDECL
578		-#endif
579		-
580		-#ifndef SHELL_NO_SYSINC
581		-# include <stdarg.h>
582		-# include <string.h>
583		-# include <stdlib.h>
584		-# include <limits.h>
585		-# include <assert.h>
586		-/* # include "sqlite3.h" */
587		-#endif
588		-#ifndef HAVE_CONSOLE_IO_H
589		-# include "console_io.h"
590		-#endif
591		-#if defined(_MSC_VER)
592		-# pragma warning(disable : 4204)
593		-#endif
594		-
595		-#ifndef SQLITE_CIO_NO_TRANSLATE
596		-# if (defined(_WIN32) \|\| defined(WIN32)) && !SQLITE_OS_WINRT
597		-# ifndef SHELL_NO_SYSINC
598		-# include <io.h>
599		-# include <fcntl.h>
600		-# undef WIN32_LEAN_AND_MEAN
601		-# define WIN32_LEAN_AND_MEAN
602		-# include <windows.h>
603		-# endif
604		-# define CIO_WIN_WC_XLATE 1 /* Use WCHAR Windows APIs for console I/O */
605		-# else
606		-# ifndef SHELL_NO_SYSINC
607		-# include <unistd.h>
608		-# endif
609		-# define CIO_WIN_WC_XLATE 0 /* Use plain C library stream I/O at console */
610		-# endif
611		-#else
612		-# define CIO_WIN_WC_XLATE 0 /* Not exposing translation routines at all */
613		-#endif
614		-
615		-#if CIO_WIN_WC_XLATE
616		-static HANDLE handleOfFile(FILE *pf){
617		- int fileDesc = _fileno(pf);
618		- union { intptr_t osfh; HANDLE fh; } fid = {
619		- (fileDesc>=0)? _get_osfhandle(fileDesc) : (intptr_t)INVALID_HANDLE_VALUE
620		- };
621		- return fid.fh;
622		-}
623		-#endif
624		-
625		-#ifndef SQLITE_CIO_NO_TRANSLATE
626		-typedef struct PerStreamTags {
627		-# if CIO_WIN_WC_XLATE
628		- HANDLE hx;
629		- DWORD consMode;
630		- char acIncomplete[4];
631		-# else
632		- short reachesConsole;
633		-# endif
634		- FILE *pf;
635		-} PerStreamTags;
636		-
637		-/* Define NULL-like value for things which can validly be 0. */
638		-# define SHELL_INVALID_FILE_PTR ((FILE *)~0)
639		-# if CIO_WIN_WC_XLATE
640		-# define SHELL_INVALID_CONS_MODE 0xFFFF0000
641		-# endif
642		-
643		-# if CIO_WIN_WC_XLATE
644		-# define PST_INITIALIZER { INVALID_HANDLE_VALUE, SHELL_INVALID_CONS_MODE, \
645		- {0,0,0,0}, SHELL_INVALID_FILE_PTR }
646		-# else
647		-# define PST_INITIALIZER { 0, SHELL_INVALID_FILE_PTR }
648		-# endif
649		-
650		-/* Quickly say whether a known output is going to the console. */
651		-# if CIO_WIN_WC_XLATE
652		-static short pstReachesConsole(PerStreamTags *ppst){
653		- return (ppst->hx != INVALID_HANDLE_VALUE);
654		-}
655		-# else
656		-# define pstReachesConsole(ppst) 0
657		-# endif
658		-
659		-# if CIO_WIN_WC_XLATE
660		-static void restoreConsoleArb(PerStreamTags *ppst){
661		- if( pstReachesConsole(ppst) ) SetConsoleMode(ppst->hx, ppst->consMode);
662		-}
663		-# else
664		-# define restoreConsoleArb(ppst)
665		-# endif
666		-
667		-/* Say whether FILE* appears to be a console, collect associated info. */
668		-static short streamOfConsole(FILE pf, / out / PerStreamTags ppst){
669		-# if CIO_WIN_WC_XLATE
670		- short rv = 0;
671		- DWORD dwCM = SHELL_INVALID_CONS_MODE;
672		- HANDLE fh = handleOfFile(pf);
673		- ppst->pf = pf;
674		- if( INVALID_HANDLE_VALUE != fh ){
675		- rv = (GetFileType(fh) == FILE_TYPE_CHAR && GetConsoleMode(fh,&dwCM));
676		- }
677		- ppst->hx = (rv)? fh : INVALID_HANDLE_VALUE;
678		- ppst->consMode = dwCM;
679		- return rv;
680		-# else
681		- ppst->pf = pf;
682		- ppst->reachesConsole = ( (short)isatty(fileno(pf)) );
683		- return ppst->reachesConsole;
684		-# endif
685		-}
686		-
687		-# ifndef ENABLE_VIRTUAL_TERMINAL_PROCESSING
688		-# define ENABLE_VIRTUAL_TERMINAL_PROCESSING (0x4)
689		-# endif
690		-
691		-# if CIO_WIN_WC_XLATE
692		-/* Define console modes for use with the Windows Console API. */
693		-# define SHELL_CONI_MODE \
694		- (ENABLE_ECHO_INPUT \| ENABLE_INSERT_MODE \| ENABLE_LINE_INPUT \| 0x80 \
695		- \| ENABLE_QUICK_EDIT_MODE \| ENABLE_EXTENDED_FLAGS \| ENABLE_PROCESSED_INPUT)
696		-# define SHELL_CONO_MODE (ENABLE_PROCESSED_OUTPUT \| ENABLE_WRAP_AT_EOL_OUTPUT \
697		- \| ENABLE_VIRTUAL_TERMINAL_PROCESSING)
698		-# endif
699		-
700		-typedef struct ConsoleInfo {
701		- PerStreamTags pstSetup[3];
702		- PerStreamTags pstDesignated[3];
703		- StreamsAreConsole sacSetup;
704		-} ConsoleInfo;
705		-
706		-static short isValidStreamInfo(PerStreamTags *ppst){
707		- return (ppst->pf != SHELL_INVALID_FILE_PTR);
708		-}
709		-
710		-static ConsoleInfo consoleInfo = {
711		- { /* pstSetup */ PST_INITIALIZER, PST_INITIALIZER, PST_INITIALIZER },
712		- { /* pstDesignated[] */ PST_INITIALIZER, PST_INITIALIZER, PST_INITIALIZER },
713		- SAC_NoConsole /* sacSetup */
714		-};
715		-
716		-SQLITE_INTERNAL_LINKAGE FILE* invalidFileStream = (FILE *)~0;
717		-
718		-# if CIO_WIN_WC_XLATE
719		-static void maybeSetupAsConsole(PerStreamTags *ppst, short odir){
720		- if( pstReachesConsole(ppst) ){
721		- DWORD cm = odir? SHELL_CONO_MODE : SHELL_CONI_MODE;
722		- SetConsoleMode(ppst->hx, cm);
723		- }
724		-}
725		-# else
726		-# define maybeSetupAsConsole(ppst,odir)
727		-# endif
728		-
729		-SQLITE_INTERNAL_LINKAGE void consoleRenewSetup(void){
730		-# if CIO_WIN_WC_XLATE
731		- int ix = 0;
732		- while( ix < 6 ){
733		- PerStreamTags *ppst = (ix<3)?
734		- &consoleInfo.pstSetup[ix] : &consoleInfo.pstDesignated[ix-3];
735		- maybeSetupAsConsole(ppst, (ix % 3)>0);
736		- ++ix;
737		- }
738		-# endif
739		-}
740		-
741		-SQLITE_INTERNAL_LINKAGE StreamsAreConsole
742		-consoleClassifySetup( FILE pfIn, FILE pfOut, FILE *pfErr ){
743		- StreamsAreConsole rv = SAC_NoConsole;
744		- FILE* apf[3] = { pfIn, pfOut, pfErr };
745		- int ix;
746		- for( ix = 2; ix >= 0; --ix ){
747		- PerStreamTags *ppst = &consoleInfo.pstSetup[ix];
748		- if( streamOfConsole(apf[ix], ppst) ){
749		- rv \|= (SAC_InConsole<<ix);
750		- }
751		- consoleInfo.pstDesignated[ix] = *ppst;
752		- if( ix > 0 ) fflush(apf[ix]);
753		- }
754		- consoleInfo.sacSetup = rv;
755		- consoleRenewSetup();
756		- return rv;
757		-}
758		-
759		-SQLITE_INTERNAL_LINKAGE void SQLITE_CDECL consoleRestore( void ){
760		-# if CIO_WIN_WC_XLATE
761		- static ConsoleInfo *pci = &consoleInfo;
762		- if( pci->sacSetup ){
763		- int ix;
764		- for( ix=0; ix<3; ++ix ){
765		- if( pci->sacSetup & (SAC_InConsole<<ix) ){
766		- PerStreamTags *ppst = &pci->pstSetup[ix];
767		- SetConsoleMode(ppst->hx, ppst->consMode);
768		- }
769		- }
770		- }
771		-# endif
772		-}
773		-#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
774		-
775		-#ifdef SQLITE_CIO_INPUT_REDIR
776		-/* Say whether given FILE* is among those known, via either
777		-** consoleClassifySetup() or set{Output,Error}Stream, as
778		-** readable, and return an associated PerStreamTags pointer
779		-** if so. Otherwise, return 0.
780		-*/
781		-static PerStreamTags * isKnownReadable(FILE *pf){
782		- static PerStreamTags *apst[] = {
783		- &consoleInfo.pstDesignated[0], &consoleInfo.pstSetup[0], 0
784		- };
785		- int ix = 0;
786		- do {
787		- if( apst[ix]->pf == pf ) break;
788		- } while( apst[++ix] != 0 );
789		- return apst[ix];
790		-}
791		-#endif
792		-
793		-#ifndef SQLITE_CIO_NO_TRANSLATE
794		-/* Say whether given FILE* is among those known, via either
795		-** consoleClassifySetup() or set{Output,Error}Stream, as
796		-** writable, and return an associated PerStreamTags pointer
797		-** if so. Otherwise, return 0.
798		-*/
799		-static PerStreamTags * isKnownWritable(FILE *pf){
800		- static PerStreamTags *apst[] = {
801		- &consoleInfo.pstDesignated[1], &consoleInfo.pstDesignated[2],
802		- &consoleInfo.pstSetup[1], &consoleInfo.pstSetup[2], 0
803		- };
804		- int ix = 0;
805		- do {
806		- if( apst[ix]->pf == pf ) break;
807		- } while( apst[++ix] != 0 );
808		- return apst[ix];
809		-}
810		-
811		-static FILE designateEmitStream(FILE pf, unsigned chix){
812		- FILE *rv = consoleInfo.pstDesignated[chix].pf;
813		- if( pf == invalidFileStream ) return rv;
814		- else{
815		- /* Setting a possibly new output stream. */
816		- PerStreamTags *ppst = isKnownWritable(pf);
817		- if( ppst != 0 ){
818		- PerStreamTags pst = *ppst;
819		- consoleInfo.pstDesignated[chix] = pst;
820		- }else streamOfConsole(pf, &consoleInfo.pstDesignated[chix]);
821		- }
822		- return rv;
823		-}
824		-
825		-SQLITE_INTERNAL_LINKAGE FILE setOutputStream(FILE pf){
826		- return designateEmitStream(pf, 1);
827		-}
828		-# ifdef CONSIO_SET_ERROR_STREAM
829		-SQLITE_INTERNAL_LINKAGE FILE setErrorStream(FILE pf){
830		- return designateEmitStream(pf, 2);
831		-}
832		-# endif
833		-#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
834		-
835		-#ifndef SQLITE_CIO_NO_SETMODE
836		-# if CIO_WIN_WC_XLATE
837		-static void setModeFlushQ(FILE *pf, short bFlush, int mode){
838		- if( bFlush ) fflush(pf);
839		- _setmode(_fileno(pf), mode);
840		-}
841		-# else
842		-# define setModeFlushQ(f, b, m) if(b) fflush(f)
843		-# endif
844		-
845		-SQLITE_INTERNAL_LINKAGE void setBinaryMode(FILE *pf, short bFlush){
846		- setModeFlushQ(pf, bFlush, _O_BINARY);
847		-}
848		-SQLITE_INTERNAL_LINKAGE void setTextMode(FILE *pf, short bFlush){
849		- setModeFlushQ(pf, bFlush, _O_TEXT);
850		-}
851		-# undef setModeFlushQ
852		-
853		-#else /* defined(SQLITE_CIO_NO_SETMODE) */
854		-# define setBinaryMode(f, bFlush) do{ if((bFlush)) fflush(f); }while(0)
855		-# define setTextMode(f, bFlush) do{ if((bFlush)) fflush(f); }while(0)
856		-#endif /* defined(SQLITE_CIO_NO_SETMODE) */
857		-
858		-#ifndef SQLITE_CIO_NO_TRANSLATE
859		-# if CIO_WIN_WC_XLATE
860		-/* Write buffer cBuf as output to stream known to reach console,
861		-** limited to ncTake char's. Return ncTake on success, else 0. */
862		-static int conZstrEmit(PerStreamTags ppst, const char z, int ncTake){
863		- int rv = 0;
864		- if( z!=NULL ){
865		- int nwc = MultiByteToWideChar(CP_UTF8,0, z,ncTake, 0,0);
866		- if( nwc > 0 ){
867		- WCHAR zw = sqlite3_malloc64(nwcsizeof(WCHAR));
868		- if( zw!=NULL ){
869		- nwc = MultiByteToWideChar(CP_UTF8,0, z,ncTake, zw,nwc);
870		- if( nwc > 0 ){
871		- /* Translation from UTF-8 to UTF-16, then WCHARs out. */
872		- if( WriteConsoleW(ppst->hx, zw,nwc, 0, NULL) ){
873		- rv = ncTake;
874		- }
875		- }
876		- sqlite3_free(zw);
877		- }
878		- }
879		- }
880		- return rv;
881		-}
882		-
883		-/* For {f,o,e}PrintfUtf8() when stream is known to reach console. */
884		-static int conioVmPrintf(PerStreamTags ppst, const char zFormat, va_list ap){
885		- char *z = sqlite3_vmprintf(zFormat, ap);
886		- if( z ){
887		- int rv = conZstrEmit(ppst, z, (int)strlen(z));
888		- sqlite3_free(z);
889		- return rv;
890		- }else return 0;
891		-}
892		-# endif /* CIO_WIN_WC_XLATE */
893		-
894		-# ifdef CONSIO_GET_EMIT_STREAM
895		-static PerStreamTags * getDesignatedEmitStream(FILE *pf, unsigned chix,
896		- PerStreamTags *ppst){
897		- PerStreamTags *rv = isKnownWritable(pf);
898		- short isValid = (rv!=0)? isValidStreamInfo(rv) : 0;
899		- if( rv != 0 && isValid ) return rv;
900		- streamOfConsole(pf, ppst);
901		- return ppst;
902		-}
903		-# endif
904		-
905		-/* Get stream info, either for designated output or error stream when
906		-** chix equals 1 or 2, or for an arbitrary stream when chix == 0.
907		-** In either case, ppst references a caller-owned PerStreamTags
908		-** struct which may be filled in if none of the known writable
909		-** streams is being held by consoleInfo. The ppf parameter is a
910		-** byref output when chix!=0 and a byref input when chix==0.
911		- */
912		-static PerStreamTags *
913		-getEmitStreamInfo(unsigned chix, PerStreamTags *ppst,
914		- /* in/out / FILE *ppf){
915		- PerStreamTags *ppstTry;
916		- FILE *pfEmit;
917		- if( chix > 0 ){
918		- ppstTry = &consoleInfo.pstDesignated[chix];
919		- if( !isValidStreamInfo(ppstTry) ){
920		- ppstTry = &consoleInfo.pstSetup[chix];
921		- pfEmit = ppst->pf;
922		- }else pfEmit = ppstTry->pf;
923		- if( !isValidStreamInfo(ppstTry) ){
924		- pfEmit = (chix > 1)? stderr : stdout;
925		- ppstTry = ppst;
926		- streamOfConsole(pfEmit, ppstTry);
927		- }
928		- *ppf = pfEmit;
929		- }else{
930		- ppstTry = isKnownWritable(*ppf);
931		- if( ppstTry != 0 ) return ppstTry;
932		- streamOfConsole(*ppf, ppst);
933		- return ppst;
934		- }
935		- return ppstTry;
936		-}
937		-
938		-SQLITE_INTERNAL_LINKAGE int oPrintfUtf8(const char *zFormat, ...){
939		- va_list ap;
940		- int rv;
941		- FILE *pfOut;
942		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
943		-# if CIO_WIN_WC_XLATE
944		- PerStreamTags *ppst = getEmitStreamInfo(1, &pst, &pfOut);
945		-# else
946		- getEmitStreamInfo(1, &pst, &pfOut);
947		-# endif
948		- assert(zFormat!=0);
949		- va_start(ap, zFormat);
950		-# if CIO_WIN_WC_XLATE
951		- if( pstReachesConsole(ppst) ){
952		- rv = conioVmPrintf(ppst, zFormat, ap);
953		- }else{
954		-# endif
955		- rv = vfprintf(pfOut, zFormat, ap);
956		-# if CIO_WIN_WC_XLATE
957		- }
958		-# endif
959		- va_end(ap);
960		- return rv;
961		-}
962		-
963		-SQLITE_INTERNAL_LINKAGE int ePrintfUtf8(const char *zFormat, ...){
964		- va_list ap;
965		- int rv;
966		- FILE *pfErr;
967		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
968		-# if CIO_WIN_WC_XLATE
969		- PerStreamTags *ppst = getEmitStreamInfo(2, &pst, &pfErr);
970		-# else
971		- getEmitStreamInfo(2, &pst, &pfErr);
972		-# endif
973		- assert(zFormat!=0);
974		- va_start(ap, zFormat);
975		-# if CIO_WIN_WC_XLATE
976		- if( pstReachesConsole(ppst) ){
977		- rv = conioVmPrintf(ppst, zFormat, ap);
978		- }else{
979		-# endif
980		- rv = vfprintf(pfErr, zFormat, ap);
981		-# if CIO_WIN_WC_XLATE
982		- }
983		-# endif
984		- va_end(ap);
985		- return rv;
986		-}
987		-
988		-SQLITE_INTERNAL_LINKAGE int fPrintfUtf8(FILE pfO, const char zFormat, ...){
989		- va_list ap;
990		- int rv;
991		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
992		-# if CIO_WIN_WC_XLATE
993		- PerStreamTags *ppst = getEmitStreamInfo(0, &pst, &pfO);
994		-# else
995		- getEmitStreamInfo(0, &pst, &pfO);
996		-# endif
997		- assert(zFormat!=0);
998		- va_start(ap, zFormat);
999		-# if CIO_WIN_WC_XLATE
1000		- if( pstReachesConsole(ppst) ){
1001		- maybeSetupAsConsole(ppst, 1);
1002		- rv = conioVmPrintf(ppst, zFormat, ap);
1003		- if( 0 == isKnownWritable(ppst->pf) ) restoreConsoleArb(ppst);
1004		- }else{
1005		-# endif
1006		- rv = vfprintf(pfO, zFormat, ap);
1007		-# if CIO_WIN_WC_XLATE
1008		- }
1009		-# endif
1010		- va_end(ap);
1011		- return rv;
1012		-}
1013		-
1014		-SQLITE_INTERNAL_LINKAGE int fPutsUtf8(const char z, FILE pfO){
1015		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1016		-# if CIO_WIN_WC_XLATE
1017		- PerStreamTags *ppst = getEmitStreamInfo(0, &pst, &pfO);
1018		-# else
1019		- getEmitStreamInfo(0, &pst, &pfO);
1020		-# endif
1021		- assert(z!=0);
1022		-# if CIO_WIN_WC_XLATE
1023		- if( pstReachesConsole(ppst) ){
1024		- int rv;
1025		- maybeSetupAsConsole(ppst, 1);
1026		- rv = conZstrEmit(ppst, z, (int)strlen(z));
1027		- if( 0 == isKnownWritable(ppst->pf) ) restoreConsoleArb(ppst);
1028		- return rv;
1029		- }else {
1030		-# endif
1031		- return (fputs(z, pfO)<0)? 0 : (int)strlen(z);
1032		-# if CIO_WIN_WC_XLATE
1033		- }
1034		-# endif
1035		-}
1036		-
1037		-SQLITE_INTERNAL_LINKAGE int ePutsUtf8(const char *z){
1038		- FILE *pfErr;
1039		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1040		-# if CIO_WIN_WC_XLATE
1041		- PerStreamTags *ppst = getEmitStreamInfo(2, &pst, &pfErr);
1042		-# else
1043		- getEmitStreamInfo(2, &pst, &pfErr);
1044		-# endif
1045		- assert(z!=0);
1046		-# if CIO_WIN_WC_XLATE
1047		- if( pstReachesConsole(ppst) ) return conZstrEmit(ppst, z, (int)strlen(z));
1048		- else {
1049		-# endif
1050		- return (fputs(z, pfErr)<0)? 0 : (int)strlen(z);
1051		-# if CIO_WIN_WC_XLATE
1052		- }
1053		-# endif
1054		-}
1055		-
1056		-SQLITE_INTERNAL_LINKAGE int oPutsUtf8(const char *z){
1057		- FILE *pfOut;
1058		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1059		-# if CIO_WIN_WC_XLATE
1060		- PerStreamTags *ppst = getEmitStreamInfo(1, &pst, &pfOut);
1061		-# else
1062		- getEmitStreamInfo(1, &pst, &pfOut);
1063		-# endif
1064		- assert(z!=0);
1065		-# if CIO_WIN_WC_XLATE
1066		- if( pstReachesConsole(ppst) ) return conZstrEmit(ppst, z, (int)strlen(z));
1067		- else {
1068		-# endif
1069		- return (fputs(z, pfOut)<0)? 0 : (int)strlen(z);
1070		-# if CIO_WIN_WC_XLATE
1071		- }
1072		-# endif
1073		-}
1074		-
1075		-#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
1076		-
1077		-#if !(defined(SQLITE_CIO_NO_UTF8SCAN) && defined(SQLITE_CIO_NO_TRANSLATE))
1078		-/* Skip over as much z[] input char sequence as is valid UTF-8,
1079		-** limited per nAccept char's or whole characters and containing
1080		-** no char cn such that ((1<<cn) & ccm)!=0. On return, the
1081		-** sequence z:return (inclusive:exclusive) is validated UTF-8.
1082		-** Limit: nAccept>=0 => char count, nAccept<0 => character
1083		- */
1084		-SQLITE_INTERNAL_LINKAGE const char*
1085		-zSkipValidUtf8(const char *z, int nAccept, long ccm){
1086		- int ng = (nAccept<0)? -nAccept : 0;
1087		- const char *pcLimit = (nAccept>=0)? z+nAccept : 0;
1088		- assert(z!=0);
1089		- while( (pcLimit)? (z<pcLimit) : (ng-- != 0) ){
1090		- char c = *z;
1091		- if( (c & 0x80) == 0 ){
1092		- if( ccm != 0L && c < 0x20 && ((1L<<c) & ccm) != 0 ) return z;
1093		- ++z; /* ASCII */
1094		- }else if( (c & 0xC0) != 0xC0 ) return z; /* not a lead byte */
1095		- else{
1096		- const char zt = z+1; / Got lead byte, look at trail bytes.*/
1097		- do{
1098		- if( pcLimit && zt >= pcLimit ) return z;
1099		- else{
1100		- char ct = *zt++;
1101		- if( ct==0 \|\| (zt-z)>4 \|\| (ct & 0xC0)!=0x80 ){
1102		- /* Trailing bytes are too few, too many, or invalid. */
1103		- return z;
1104		- }
1105		- }
1106		- } while( ((c <<= 1) & 0x40) == 0x40 ); /* Eat lead byte's count. */
1107		- z = zt;
1108		- }
1109		- }
1110		- return z;
1111		-}
1112		-#endif /!(defined(SQLITE_CIO_NO_UTF8SCAN)&&defined(SQLITE_CIO_NO_TRANSLATE))/
1113		-
1114		-#ifndef SQLITE_CIO_NO_TRANSLATE
1115		-# ifdef CONSIO_SPUTB
1116		-SQLITE_INTERNAL_LINKAGE int
1117		-fPutbUtf8(FILE pfO, const char cBuf, int nAccept){
1118		- assert(pfO!=0);
1119		-# if CIO_WIN_WC_XLATE
1120		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1121		- PerStreamTags *ppst = getEmitStreamInfo(0, &pst, &pfO);
1122		- if( pstReachesConsole(ppst) ){
1123		- int rv;
1124		- maybeSetupAsConsole(ppst, 1);
1125		- rv = conZstrEmit(ppst, cBuf, nAccept);
1126		- if( 0 == isKnownWritable(ppst->pf) ) restoreConsoleArb(ppst);
1127		- return rv;
1128		- }else {
1129		-# endif
1130		- return (int)fwrite(cBuf, 1, nAccept, pfO);
1131		-# if CIO_WIN_WC_XLATE
1132		- }
1133		-# endif
1134		-}
1135		-# endif
1136		-
1137		-SQLITE_INTERNAL_LINKAGE int
1138		-oPutbUtf8(const char *cBuf, int nAccept){
1139		- FILE *pfOut;
1140		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1141		-# if CIO_WIN_WC_XLATE
1142		- PerStreamTags *ppst = getEmitStreamInfo(1, &pst, &pfOut);
1143		-# else
1144		- getEmitStreamInfo(1, &pst, &pfOut);
1145		-# endif
1146		-# if CIO_WIN_WC_XLATE
1147		- if( pstReachesConsole(ppst) ){
1148		- return conZstrEmit(ppst, cBuf, nAccept);
1149		- }else {
1150		-# endif
1151		- return (int)fwrite(cBuf, 1, nAccept, pfOut);
1152		-# if CIO_WIN_WC_XLATE
1153		- }
1154		-# endif
1155		-}
1156		-
1157		-/*
1158		-** Flush the given output stream. Return non-zero for success, else 0.
1159		-*/
1160		-#if !defined(SQLITE_CIO_NO_FLUSH) && !defined(SQLITE_CIO_NO_SETMODE)
1161		-SQLITE_INTERNAL_LINKAGE int
1162		-fFlushBuffer(FILE *pfOut){
1163		-# if CIO_WIN_WC_XLATE && !defined(SHELL_OMIT_FIO_DUPE)
1164		- return FlushFileBuffers(handleOfFile(pfOut))? 1 : 0;
1165		-# else
1166		- return fflush(pfOut);
1167		-# endif
1168		-}
1169		-#endif
1170		-
1171		-#if CIO_WIN_WC_XLATE \
1172		- && !defined(SHELL_OMIT_FIO_DUPE) \
1173		- && defined(SQLITE_USE_ONLY_WIN32)
1174		-static struct FileAltIds {
1175		- int fd;
1176		- HANDLE fh;
1177		-} altIdsOfFile(FILE *pf){
1178		- struct FileAltIds rv = { _fileno(pf) };
1179		- union { intptr_t osfh; HANDLE fh; } fid = {
1180		- (rv.fd>=0)? _get_osfhandle(rv.fd) : (intptr_t)INVALID_HANDLE_VALUE
1181		- };
1182		- rv.fh = fid.fh;
1183		- return rv;
1184		-}
1185		-
1186		-SQLITE_INTERNAL_LINKAGE size_t
1187		-cfWrite(const void buf, size_t osz, size_t ocnt, FILE pf){
1188		- size_t rv = 0;
1189		- struct FileAltIds fai = altIdsOfFile(pf);
1190		- int fmode = _setmode(fai.fd, _O_BINARY);
1191		- _setmode(fai.fd, fmode);
1192		- while( rv < ocnt ){
1193		- size_t nbo = osz;
1194		- while( nbo > 0 ){
1195		- DWORD dwno = (nbo>(1L<<24))? 1L<<24 : (DWORD)nbo;
1196		- BOOL wrc = TRUE;
1197		- BOOL genCR = (fmode & _O_TEXT)!=0;
1198		- if( genCR ){
1199		- const char pnl = (const char)memchr(buf, '\n', nbo);
1200		- if( pnl ) nbo = pnl - (const char*)buf;
1201		- else genCR = 0;
1202		- }
1203		- if( dwno>0 ) wrc = WriteFile(fai.fh, buf, dwno, 0,0);
1204		- if( genCR && wrc ){
1205		- wrc = WriteFile(fai.fh, "\r\n", 2, 0,0);
1206		- ++dwno; /* Skip over the LF */
1207		- }
1208		- if( !wrc ) return rv;
1209		- buf = (const char*)buf + dwno;
1210		- nbo += dwno;
1211		- }
1212		- ++rv;
1213		- }
1214		- return rv;
1215		-}
1216		-
1217		-SQLITE_INTERNAL_LINKAGE char *
1218		-cfGets(char cBuf, int n, FILE pf){
1219		- int nci = 0;
1220		- struct FileAltIds fai = altIdsOfFile(pf);
1221		- int fmode = _setmode(fai.fd, _O_BINARY);
1222		- BOOL eatCR = (fmode & _O_TEXT)!=0;
1223		- _setmode(fai.fd, fmode);
1224		- while( nci < n-1 ){
1225		- DWORD nr;
1226		- if( !ReadFile(fai.fh, cBuf+nci, 1, &nr, 0) \|\| nr==0 ) break;
1227		- if( nr>0 && (!eatCR \|\| cBuf[nci]!='\r') ) nci += nr;
1228		- }
1229		- if( nci < n ) cBuf[nci] = 0;
1230		- return (nci>0)? cBuf : 0;
1231		-}
1232		-# else
1233		-# define cfWrite(b,os,no,f) fwrite(b,os,no,f)
1234		-# define cfGets(b,n,f) fgets(b,n,f)
1235		-# endif
1236		-
1237		-# ifdef CONSIO_EPUTB
1238		-SQLITE_INTERNAL_LINKAGE int
1239		-ePutbUtf8(const char *cBuf, int nAccept){
1240		- FILE *pfErr;
1241		- PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1242		- PerStreamTags *ppst = getEmitStreamInfo(2, &pst, &pfErr);
1243		-# if CIO_WIN_WC_XLATE
1244		- if( pstReachesConsole(ppst) ){
1245		- return conZstrEmit(ppst, cBuf, nAccept);
1246		- }else {
1247		-# endif
1248		- return (int)cfWrite(cBuf, 1, nAccept, pfErr);
1249		-# if CIO_WIN_WC_XLATE
1250		- }
1251		-# endif
1252		-}
1253		-# endif /* defined(CONSIO_EPUTB) */
1254		-
1255		-SQLITE_INTERNAL_LINKAGE char* fGetsUtf8(char cBuf, int ncMax, FILE pfIn){
1256		- if( pfIn==0 ) pfIn = stdin;
1257		-# if CIO_WIN_WC_XLATE
1258		- if( pfIn == consoleInfo.pstSetup[0].pf
1259		- && (consoleInfo.sacSetup & SAC_InConsole)!=0 ){
1260		-# if CIO_WIN_WC_XLATE==1
1261		-# define SHELL_GULP 150 /* Count of WCHARS to be gulped at a time */
1262		- WCHAR wcBuf[SHELL_GULP+1];
1263		- int lend = 0, noc = 0;
1264		- if( ncMax > 0 ) cBuf[0] = 0;
1265		- while( noc < ncMax-8-1 && !lend ){
1266		- /* There is room for at least 2 more characters and a 0-terminator. */
1267		- int na = (ncMax > SHELL_GULP*4+1 + noc)? SHELL_GULP : (ncMax-1 - noc)/4;
1268		-# undef SHELL_GULP
1269		- DWORD nbr = 0;
1270		- BOOL bRC = ReadConsoleW(consoleInfo.pstSetup[0].hx, wcBuf, na, &nbr, 0);
1271		- if( bRC && nbr>0 && (wcBuf[nbr-1]&0xF800)==0xD800 ){
1272		- /* Last WHAR read is first of a UTF-16 surrogate pair. Grab its mate. */
1273		- DWORD nbrx;
1274		- bRC &= ReadConsoleW(consoleInfo.pstSetup[0].hx, wcBuf+nbr, 1, &nbrx, 0);
1275		- if( bRC ) nbr += nbrx;
1276		- }
1277		- if( !bRC \|\| (noc==0 && nbr==0) ) return 0;
1278		- if( nbr > 0 ){
1279		- int nmb = WideCharToMultiByte(CP_UTF8, 0, wcBuf,nbr,0,0,0,0);
1280		- if( nmb != 0 && noc+nmb <= ncMax ){
1281		- int iseg = noc;
1282		- nmb = WideCharToMultiByte(CP_UTF8, 0, wcBuf,nbr,cBuf+noc,nmb,0,0);
1283		- noc += nmb;
1284		- /* Fixup line-ends as coded by Windows for CR (or "Enter".)
1285		- ** This is done without regard for any setMode{Text,Binary}()
1286		- ** call that might have been done on the interactive input.
1287		- */
1288		- if( noc > 0 ){
1289		- if( cBuf[noc-1]=='\n' ){
1290		- lend = 1;
1291		- if( noc > 1 && cBuf[noc-2]=='\r' ) cBuf[--noc-1] = '\n';
1292		- }
1293		- }
1294		- /* Check for ^Z (anywhere in line) too, to act as EOF. */
1295		- while( iseg < noc ){
1296		- if( cBuf[iseg]=='\x1a' ){
1297		- noc = iseg; /* Chop ^Z and anything following. */
1298		- lend = 1; /* Counts as end of line too. */
1299		- break;
1300		- }
1301		- ++iseg;
1302		- }
1303		- }else break; /* Drop apparent garbage in. (Could assert.) */
1304		- }else break;
1305		- }
1306		- /* If got nothing, (after ^Z chop), must be at end-of-file. */
1307		- if( noc > 0 ){
1308		- cBuf[noc] = 0;
1309		- return cBuf;
1310		- }else return 0;
1311		-# endif
1312		- }else{
1313		-# endif
1314		- return cfGets(cBuf, ncMax, pfIn);
1315		-# if CIO_WIN_WC_XLATE
1316		- }
1317		-# endif
1318		-}
1319		-#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
1320		-
1321		-#if defined(_MSC_VER)
1322		-# pragma warning(default : 4204)
1323		-#endif
1324		-
1325		-#undef SHELL_INVALID_FILE_PTR
1326		-
1327		-/*********************** End ../ext/consio/console_io.c ******************/
1328		-
1329		-#ifndef SQLITE_SHELL_FIDDLE
1330		-
1331		-/* From here onward, fgets() is redirected to the console_io library. */
1332		-# define fgets(b,n,f) fGetsUtf8(b,n,f)
1333		-/*
1334		- * Define macros for emitting output text in various ways:
1335		- * sputz(s, z) => emit 0-terminated string z to given stream s
1336		- * sputf(s, f, ...) => emit varargs per format f to given stream s
1337		- * oputz(z) => emit 0-terminated string z to default stream
1338		- * oputf(f, ...) => emit varargs per format f to default stream
1339		- * eputz(z) => emit 0-terminated string z to error stream
1340		- * eputf(f, ...) => emit varargs per format f to error stream
1341		- * oputb(b, n) => emit char buffer b[0..n-1] to default stream
1342		- *
1343		- * Note that the default stream is whatever has been last set via:
1344		- * setOutputStream(FILE *pf)
1345		- * This is normally the stream that CLI normal output goes to.
1346		- * For the stand-alone CLI, it is stdout with no .output redirect.
1347		- *
1348		- * The ?putz(z) forms are required for the Fiddle builds for string literal
1349		- * output, in aid of enforcing format string to argument correspondence.
1350		- */
1351		-# define sputz(s,z) fPutsUtf8(z,s)
1352		-# define sputf fPrintfUtf8
1353		-# define oputz(z) oPutsUtf8(z)
1354		-# define oputf oPrintfUtf8
1355		-# define eputz(z) ePutsUtf8(z)
1356		-# define eputf ePrintfUtf8
1357		-# define oputb(buf,na) oPutbUtf8(buf,na)
1358		-# define fflush(s) fFlushBuffer(s);
1359		-
1360		-#else
1361		-/* For Fiddle, all console handling and emit redirection is omitted. */
1362		-/* These next 3 macros are for emitting formatted output. When complaints
1363		- * from the WASM build are issued for non-formatted output, when a mere
1364		- * string literal is to be emitted, the ?putz(z) forms should be used.
1365		- * (This permits compile-time checking of format string / argument mismatch.)
1366		- */
1367		-# define oputf(fmt, ...) printf(fmt,__VA_ARGS__)
1368		-# define eputf(fmt, ...) fprintf(stderr,fmt,__VA_ARGS__)
1369		-# define sputf(fp,fmt, ...) fprintf(fp,fmt,__VA_ARGS__)
1370		-/* These next 3 macros are for emitting simple string literals. */
1371		-# define oputz(z) fputs(z,stdout)
1372		-# define eputz(z) fputs(z,stderr)
1373		-# define sputz(fp,z) fputs(z,fp)
1374		-# define oputb(buf,na) fwrite(buf,1,na,stdout)
1375		-# undef fflush
1376		-#endif
	624	+
	625	+#define eputz(z) sqlite3_fputs(z,stderr)
	626	+#define sputz(fp,z) sqlite3_fputs(z,fp)
1377	627
1378	628	/* True if the timer is enabled */
1379	629	static int enableTimer = 0;
1380	630
1381	631	/* A version of strcmp() that works with NULL values */
		@@ -1416,10 +666,11 @@
1416	666	struct timeval ru_utime; /* user CPU time used */
1417	667	struct timeval ru_stime; /* system CPU time used */
1418	668	};
1419	669	#define getrusage(A,B) memset(B,0,sizeof(*B))
1420	670	#endif
	671	+
1421	672
1422	673	/* Saved resource information for the beginning of an operation */
1423	674	static struct rusage sBegin; /* CPU time at start */
1424	675	static sqlite3_int64 iBegin; /* Wall-clock time at start */
1425	676
		@@ -1440,24 +691,24 @@
1440	691	}
1441	692
1442	693	/*
1443	694	** Print the timing results.
1444	695	*/
1445		-static void endTimer(void){
	696	+static void endTimer(FILE *out){
1446	697	if( enableTimer ){
1447	698	sqlite3_int64 iEnd = timeOfDay();
1448	699	struct rusage sEnd;
1449	700	getrusage(RUSAGE_SELF, &sEnd);
1450		- sputf(stdout, "Run Time: real %.3f user %f sys %f\n",
	701	+ sqlite3_fprintf(out, "Run Time: real %.3f user %f sys %f\n",
1451	702	(iEnd - iBegin)*0.001,
1452	703	timeDiff(&sBegin.ru_utime, &sEnd.ru_utime),
1453	704	timeDiff(&sBegin.ru_stime, &sEnd.ru_stime));
1454	705	}
1455	706	}
1456	707
1457	708	#define BEGIN_TIMER beginTimer()
1458		-#define END_TIMER endTimer()
	709	+#define END_TIMER(X) endTimer(X)
1459	710	#define HAS_TIMER 1
1460	711
1461	712	#elif (defined(_WIN32) \|\| defined(WIN32))
1462	713
1463	714	/* Saved resource information for the beginning of an operation */
		@@ -1519,29 +770,29 @@
1519	770	}
1520	771
1521	772	/*
1522	773	** Print the timing results.
1523	774	*/
1524		-static void endTimer(void){
	775	+static void endTimer(FILE *out){
1525	776	if( enableTimer && getProcessTimesAddr){
1526	777	FILETIME ftCreation, ftExit, ftKernelEnd, ftUserEnd;
1527	778	sqlite3_int64 ftWallEnd = timeOfDay();
1528	779	getProcessTimesAddr(hProcess,&ftCreation,&ftExit,&ftKernelEnd,&ftUserEnd);
1529		- sputf(stdout, "Run Time: real %.3f user %f sys %f\n",
	780	+ sqlite3_fprintf(out, "Run Time: real %.3f user %f sys %f\n",
1530	781	(ftWallEnd - ftWallBegin)*0.001,
1531	782	timeDiff(&ftUserBegin, &ftUserEnd),
1532	783	timeDiff(&ftKernelBegin, &ftKernelEnd));
1533	784	}
1534	785	}
1535	786
1536	787	#define BEGIN_TIMER beginTimer()
1537		-#define END_TIMER endTimer()
	788	+#define END_TIMER(X) endTimer(X)
1538	789	#define HAS_TIMER hasTimer()
1539	790
1540	791	#else
1541	792	#define BEGIN_TIMER
1542		-#define END_TIMER
	793	+#define END_TIMER(X) /no-op/
1543	794	#define HAS_TIMER 0
1544	795	#endif
1545	796
1546	797	/*
1547	798	** Used to prevent warnings about unused parameters
		@@ -1738,41 +989,216 @@
1738	989	char *z;
1739	990	if( iotrace==0 ) return;
1740	991	va_start(ap, zFormat);
1741	992	z = sqlite3_vmprintf(zFormat, ap);
1742	993	va_end(ap);
1743		- sputf(iotrace, "%s", z);
	994	+ sqlite3_fprintf(iotrace, "%s", z);
1744	995	sqlite3_free(z);
1745	996	}
1746	997	#endif
1747	998
	999	+/* Lookup table to estimate the number of columns consumed by a Unicode
	1000	+** character.
	1001	+*/
	1002	+static const struct {
	1003	+ unsigned char w; /* Width of the character in columns */
	1004	+ int iFirst; /* First character in a span having this width */
	1005	+} aUWidth[] = {
	1006	+ /* {1, 0x00000}, */
	1007	+ {0, 0x00300}, {1, 0x00370}, {0, 0x00483}, {1, 0x00487}, {0, 0x00488},
	1008	+ {1, 0x0048a}, {0, 0x00591}, {1, 0x005be}, {0, 0x005bf}, {1, 0x005c0},
	1009	+ {0, 0x005c1}, {1, 0x005c3}, {0, 0x005c4}, {1, 0x005c6}, {0, 0x005c7},
	1010	+ {1, 0x005c8}, {0, 0x00600}, {1, 0x00604}, {0, 0x00610}, {1, 0x00616},
	1011	+ {0, 0x0064b}, {1, 0x0065f}, {0, 0x00670}, {1, 0x00671}, {0, 0x006d6},
	1012	+ {1, 0x006e5}, {0, 0x006e7}, {1, 0x006e9}, {0, 0x006ea}, {1, 0x006ee},
	1013	+ {0, 0x0070f}, {1, 0x00710}, {0, 0x00711}, {1, 0x00712}, {0, 0x00730},
	1014	+ {1, 0x0074b}, {0, 0x007a6}, {1, 0x007b1}, {0, 0x007eb}, {1, 0x007f4},
	1015	+ {0, 0x00901}, {1, 0x00903}, {0, 0x0093c}, {1, 0x0093d}, {0, 0x00941},
	1016	+ {1, 0x00949}, {0, 0x0094d}, {1, 0x0094e}, {0, 0x00951}, {1, 0x00955},
	1017	+ {0, 0x00962}, {1, 0x00964}, {0, 0x00981}, {1, 0x00982}, {0, 0x009bc},
	1018	+ {1, 0x009bd}, {0, 0x009c1}, {1, 0x009c5}, {0, 0x009cd}, {1, 0x009ce},
	1019	+ {0, 0x009e2}, {1, 0x009e4}, {0, 0x00a01}, {1, 0x00a03}, {0, 0x00a3c},
	1020	+ {1, 0x00a3d}, {0, 0x00a41}, {1, 0x00a43}, {0, 0x00a47}, {1, 0x00a49},
	1021	+ {0, 0x00a4b}, {1, 0x00a4e}, {0, 0x00a70}, {1, 0x00a72}, {0, 0x00a81},
	1022	+ {1, 0x00a83}, {0, 0x00abc}, {1, 0x00abd}, {0, 0x00ac1}, {1, 0x00ac6},
	1023	+ {0, 0x00ac7}, {1, 0x00ac9}, {0, 0x00acd}, {1, 0x00ace}, {0, 0x00ae2},
	1024	+ {1, 0x00ae4}, {0, 0x00b01}, {1, 0x00b02}, {0, 0x00b3c}, {1, 0x00b3d},
	1025	+ {0, 0x00b3f}, {1, 0x00b40}, {0, 0x00b41}, {1, 0x00b44}, {0, 0x00b4d},
	1026	+ {1, 0x00b4e}, {0, 0x00b56}, {1, 0x00b57}, {0, 0x00b82}, {1, 0x00b83},
	1027	+ {0, 0x00bc0}, {1, 0x00bc1}, {0, 0x00bcd}, {1, 0x00bce}, {0, 0x00c3e},
	1028	+ {1, 0x00c41}, {0, 0x00c46}, {1, 0x00c49}, {0, 0x00c4a}, {1, 0x00c4e},
	1029	+ {0, 0x00c55}, {1, 0x00c57}, {0, 0x00cbc}, {1, 0x00cbd}, {0, 0x00cbf},
	1030	+ {1, 0x00cc0}, {0, 0x00cc6}, {1, 0x00cc7}, {0, 0x00ccc}, {1, 0x00cce},
	1031	+ {0, 0x00ce2}, {1, 0x00ce4}, {0, 0x00d41}, {1, 0x00d44}, {0, 0x00d4d},
	1032	+ {1, 0x00d4e}, {0, 0x00dca}, {1, 0x00dcb}, {0, 0x00dd2}, {1, 0x00dd5},
	1033	+ {0, 0x00dd6}, {1, 0x00dd7}, {0, 0x00e31}, {1, 0x00e32}, {0, 0x00e34},
	1034	+ {1, 0x00e3b}, {0, 0x00e47}, {1, 0x00e4f}, {0, 0x00eb1}, {1, 0x00eb2},
	1035	+ {0, 0x00eb4}, {1, 0x00eba}, {0, 0x00ebb}, {1, 0x00ebd}, {0, 0x00ec8},
	1036	+ {1, 0x00ece}, {0, 0x00f18}, {1, 0x00f1a}, {0, 0x00f35}, {1, 0x00f36},
	1037	+ {0, 0x00f37}, {1, 0x00f38}, {0, 0x00f39}, {1, 0x00f3a}, {0, 0x00f71},
	1038	+ {1, 0x00f7f}, {0, 0x00f80}, {1, 0x00f85}, {0, 0x00f86}, {1, 0x00f88},
	1039	+ {0, 0x00f90}, {1, 0x00f98}, {0, 0x00f99}, {1, 0x00fbd}, {0, 0x00fc6},
	1040	+ {1, 0x00fc7}, {0, 0x0102d}, {1, 0x01031}, {0, 0x01032}, {1, 0x01033},
	1041	+ {0, 0x01036}, {1, 0x01038}, {0, 0x01039}, {1, 0x0103a}, {0, 0x01058},
	1042	+ {1, 0x0105a}, {2, 0x01100}, {0, 0x01160}, {1, 0x01200}, {0, 0x0135f},
	1043	+ {1, 0x01360}, {0, 0x01712}, {1, 0x01715}, {0, 0x01732}, {1, 0x01735},
	1044	+ {0, 0x01752}, {1, 0x01754}, {0, 0x01772}, {1, 0x01774}, {0, 0x017b4},
	1045	+ {1, 0x017b6}, {0, 0x017b7}, {1, 0x017be}, {0, 0x017c6}, {1, 0x017c7},
	1046	+ {0, 0x017c9}, {1, 0x017d4}, {0, 0x017dd}, {1, 0x017de}, {0, 0x0180b},
	1047	+ {1, 0x0180e}, {0, 0x018a9}, {1, 0x018aa}, {0, 0x01920}, {1, 0x01923},
	1048	+ {0, 0x01927}, {1, 0x01929}, {0, 0x01932}, {1, 0x01933}, {0, 0x01939},
	1049	+ {1, 0x0193c}, {0, 0x01a17}, {1, 0x01a19}, {0, 0x01b00}, {1, 0x01b04},
	1050	+ {0, 0x01b34}, {1, 0x01b35}, {0, 0x01b36}, {1, 0x01b3b}, {0, 0x01b3c},
	1051	+ {1, 0x01b3d}, {0, 0x01b42}, {1, 0x01b43}, {0, 0x01b6b}, {1, 0x01b74},
	1052	+ {0, 0x01dc0}, {1, 0x01dcb}, {0, 0x01dfe}, {1, 0x01e00}, {0, 0x0200b},
	1053	+ {1, 0x02010}, {0, 0x0202a}, {1, 0x0202f}, {0, 0x02060}, {1, 0x02064},
	1054	+ {0, 0x0206a}, {1, 0x02070}, {0, 0x020d0}, {1, 0x020f0}, {2, 0x02329},
	1055	+ {1, 0x0232b}, {2, 0x02e80}, {0, 0x0302a}, {2, 0x03030}, {1, 0x0303f},
	1056	+ {2, 0x03040}, {0, 0x03099}, {2, 0x0309b}, {1, 0x0a4d0}, {0, 0x0a806},
	1057	+ {1, 0x0a807}, {0, 0x0a80b}, {1, 0x0a80c}, {0, 0x0a825}, {1, 0x0a827},
	1058	+ {2, 0x0ac00}, {1, 0x0d7a4}, {2, 0x0f900}, {1, 0x0fb00}, {0, 0x0fb1e},
	1059	+ {1, 0x0fb1f}, {0, 0x0fe00}, {2, 0x0fe10}, {1, 0x0fe1a}, {0, 0x0fe20},
	1060	+ {1, 0x0fe24}, {2, 0x0fe30}, {1, 0x0fe70}, {0, 0x0feff}, {2, 0x0ff00},
	1061	+ {1, 0x0ff61}, {2, 0x0ffe0}, {1, 0x0ffe7}, {0, 0x0fff9}, {1, 0x0fffc},
	1062	+ {0, 0x10a01}, {1, 0x10a04}, {0, 0x10a05}, {1, 0x10a07}, {0, 0x10a0c},
	1063	+ {1, 0x10a10}, {0, 0x10a38}, {1, 0x10a3b}, {0, 0x10a3f}, {1, 0x10a40},
	1064	+ {0, 0x1d167}, {1, 0x1d16a}, {0, 0x1d173}, {1, 0x1d183}, {0, 0x1d185},
	1065	+ {1, 0x1d18c}, {0, 0x1d1aa}, {1, 0x1d1ae}, {0, 0x1d242}, {1, 0x1d245},
	1066	+ {2, 0x20000}, {1, 0x2fffe}, {2, 0x30000}, {1, 0x3fffe}, {0, 0xe0001},
	1067	+ {1, 0xe0002}, {0, 0xe0020}, {1, 0xe0080}, {0, 0xe0100}, {1, 0xe01f0}
	1068	+};
	1069	+
1748	1070	/*
1749		-** Output string zUtf to Out stream as w characters. If w is negative,
	1071	+** Return an estimate of the width, in columns, for the single Unicode
	1072	+** character c. For normal characters, the answer is always 1. But the
	1073	+** estimate might be 0 or 2 for zero-width and double-width characters.
	1074	+**
	1075	+** Different display devices display unicode using different widths. So
	1076	+** it is impossible to know that true display width with 100% accuracy.
	1077	+** Inaccuracies in the width estimates might cause columns to be misaligned.
	1078	+** Unfortunately, there is nothing we can do about that.
	1079	+*/
	1080	+int cli_wcwidth(int c){
	1081	+ int iFirst, iLast;
	1082	+
	1083	+ /* Fast path for common characters */
	1084	+ if( c<=0x300 ) return 1;
	1085	+
	1086	+ /* The general case */
	1087	+ iFirst = 0;
	1088	+ iLast = sizeof(aUWidth)/sizeof(aUWidth[0]) - 1;
	1089	+ while( iFirst<iLast-1 ){
	1090	+ int iMid = (iFirst+iLast)/2;
	1091	+ int cMid = aUWidth[iMid].iFirst;
	1092	+ if( cMid < c ){
	1093	+ iFirst = iMid;
	1094	+ }else if( cMid > c ){
	1095	+ iLast = iMid - 1;
	1096	+ }else{
	1097	+ return aUWidth[iMid].w;
	1098	+ }
	1099	+ }
	1100	+ if( aUWidth[iLast].iFirst > c ) return aUWidth[iFirst].w;
	1101	+ return aUWidth[iLast].w;
	1102	+}
	1103	+
	1104	+/*
	1105	+** Compute the value and length of a multi-byte UTF-8 character that
	1106	+** begins at z[0]. Return the length. Write the Unicode value into *pU.
	1107	+**
	1108	+** This routine only works for multi-byte UTF-8 characters.
	1109	+*/
	1110	+static int decodeUtf8(const unsigned char z, int pU){
	1111	+ if( (z[0] & 0xe0)==0xc0 && (z[1] & 0xc0)==0x80 ){
	1112	+ *pU = ((z[0] & 0x1f)<<6) \| (z[1] & 0x3f);
	1113	+ return 2;
	1114	+ }
	1115	+ if( (z[0] & 0xf0)==0xe0 && (z[1] & 0xc0)==0x80 && (z[2] & 0xc0)==0x80 ){
	1116	+ *pU = ((z[0] & 0x0f)<<12) \| ((z[1] & 0x3f)<<6) \| (z[2] & 0x3f);
	1117	+ return 3;
	1118	+ }
	1119	+ if( (z[0] & 0xf8)==0xf0 && (z[1] & 0xc0)==0x80 && (z[2] & 0xc0)==0x80
	1120	+ && (z[3] & 0xc0)==0x80
	1121	+ ){
	1122	+ *pU = ((z[0] & 0x0f)<<18) \| ((z[1] & 0x3f)<<12) \| ((z[2] & 0x3f))<<6
	1123	+ \| (z[4] & 0x3f);
	1124	+ return 4;
	1125	+ }
	1126	+ *pU = 0;
	1127	+ return 1;
	1128	+}
	1129	+
	1130	+
	1131	+#if 0 /* NOT USED */
	1132	+/*
	1133	+** Return the width, in display columns, of a UTF-8 string.
	1134	+**
	1135	+** Each normal character counts as 1. Zero-width characters count
	1136	+** as zero, and double-width characters count as 2.
	1137	+*/
	1138	+int cli_wcswidth(const char *z){
	1139	+ const unsigned char a = (const unsigned char)z;
	1140	+ int n = 0;
	1141	+ int i = 0;
	1142	+ unsigned char c;
	1143	+ while( (c = a[i])!=0 ){
	1144	+ if( c>=0xc0 ){
	1145	+ int u;
	1146	+ int len = decodeUtf8(&a[i], &u);
	1147	+ i += len;
	1148	+ n += cli_wcwidth(u);
	1149	+ }else if( c>=' ' ){
	1150	+ n++;
	1151	+ i++;
	1152	+ }else{
	1153	+ i++;
	1154	+ }
	1155	+ }
	1156	+ return n;
	1157	+}
	1158	+#endif
	1159	+
	1160	+/*
	1161	+** Output string zUtf to stdout as w characters. If w is negative,
1750	1162	** then right-justify the text. W is the width in UTF-8 characters, not
1751	1163	** in bytes. This is different from the %.s specification in printf
1752	1164	** since with %.s the width is measured in bytes, not characters.
	1165	+**
	1166	+** Take into account zero-width and double-width Unicode characters.
	1167	+** In other words, a zero-width character does not count toward the
	1168	+** the w limit. A double-width character counts as two.
1753	1169	*/
1754		-static void utf8_width_print(int w, const char *zUtf){
1755		- int i;
1756		- int n;
	1170	+static void utf8_width_print(FILE out, int w, const char zUtf){
	1171	+ const unsigned char a = (const unsigned char)zUtf;
	1172	+ unsigned char c;
	1173	+ int i = 0;
	1174	+ int n = 0;
1757	1175	int aw = w<0 ? -w : w;
1758	1176	if( zUtf==0 ) zUtf = "";
1759		- for(i=n=0; zUtf[i]; i++){
1760		- if( (zUtf[i]&0xc0)!=0x80 ){
	1177	+ while( (c = a[i])!=0 ){
	1178	+ if( (c&0xc0)==0xc0 ){
	1179	+ int u;
	1180	+ int len = decodeUtf8(a+i, &u);
	1181	+ int x = cli_wcwidth(u);
	1182	+ if( x+n>aw ){
	1183	+ break;
	1184	+ }
	1185	+ i += len;
	1186	+ n += x;
	1187	+ }else if( n>=aw ){
	1188	+ break;
	1189	+ }else{
1761	1190	n++;
1762		- if( n==aw ){
1763		- do{ i++; }while( (zUtf[i]&0xc0)==0x80 );
1764		- break;
1765		- }
	1191	+ i++;
1766	1192	}
1767	1193	}
1768	1194	if( n>=aw ){
1769		- oputf("%.*s", i, zUtf);
	1195	+ sqlite3_fprintf(out, "%.*s", i, zUtf);
1770	1196	}else if( w<0 ){
1771		- oputf("%*s%s", aw-n, "", zUtf);
	1197	+ sqlite3_fprintf(out, "%*s%s", aw-n, "", zUtf);
1772	1198	}else{
1773		- oputf("%s%*s", zUtf, aw-n, "");
	1199	+ sqlite3_fprintf(out, "%s%*s", zUtf, aw-n, "");
1774	1200	}
1775	1201	}
1776	1202
1777	1203
1778	1204	/*
		@@ -1834,11 +1260,11 @@
1834	1260	struct __stat64 x = {0};
1835	1261	# define STAT_CHR_SRC(mode) ((mode & (_S_IFCHR\|_S_IFIFO\|_S_IFREG))!=0)
1836	1262	/* On Windows, open first, then check the stream nature. This order
1837	1263	** is necessary because _stat() and sibs, when checking a named pipe,
1838	1264	** effectively break the pipe as its supplier sees it. */
1839		- FILE *rv = fopen(zFile, "rb");
	1265	+ FILE *rv = sqlite3_fopen(zFile, "rb");
1840	1266	if( rv==0 ) return 0;
1841	1267	if( _fstat64(_fileno(rv), &x) != 0
1842	1268	\|\| !STAT_CHR_SRC(x.st_mode)){
1843	1269	fclose(rv);
1844	1270	rv = 0;
		@@ -1848,11 +1274,11 @@
1848	1274	struct stat x = {0};
1849	1275	int rc = stat(zFile, &x);
1850	1276	# define STAT_CHR_SRC(mode) (S_ISREG(mode)\|\|S_ISFIFO(mode)\|\|S_ISCHR(mode))
1851	1277	if( rc!=0 ) return 0;
1852	1278	if( STAT_CHR_SRC(x.st_mode) ){
1853		- return fopen(zFile, "rb");
	1279	+ return sqlite3_fopen(zFile, "rb");
1854	1280	}else{
1855	1281	return 0;
1856	1282	}
1857	1283	#endif
1858	1284	#undef STAT_CHR_SRC
		@@ -1875,11 +1301,11 @@
1875	1301	if( n+100>nLine ){
1876	1302	nLine = nLine*2 + 100;
1877	1303	zLine = realloc(zLine, nLine);
1878	1304	shell_check_oom(zLine);
1879	1305	}
1880		- if( fgets(&zLine[n], nLine - n, in)==0 ){
	1306	+ if( sqlite3_fgets(&zLine[n], nLine - n, in)==0 ){
1881	1307	if( n==0 ){
1882	1308	free(zLine);
1883	1309	return 0;
1884	1310	}
1885	1311	zLine[n] = 0;
		@@ -2941,11 +2367,11 @@
2941	2367	**
2942	2368	******************************************************************************
2943	2369	**
2944	2370	** This SQLite extension implements functions that compute SHA3 hashes
2945	2371	** in the way described by the (U.S.) NIST FIPS 202 SHA-3 Standard.
2946		-** Two SQL functions are implemented:
	2372	+** Three SQL functions are implemented:
2947	2373	**
2948	2374	** sha3(X,SIZE)
2949	2375	** sha3_agg(Y,SIZE)
2950	2376	** sha3_query(Z,SIZE)
2951	2377	**
		@@ -3783,10 +3209,422 @@
3783	3209	}
3784	3210	return rc;
3785	3211	}
3786	3212
3787	3213	/*********************** End ../ext/misc/shathree.c ******************/
	3214	+/*********************** Begin ../ext/misc/sha1.c ****************/
	3215	+/*
	3216	+** 2017-01-27
	3217	+**
	3218	+** The author disclaims copyright to this source code. In place of
	3219	+** a legal notice, here is a blessing:
	3220	+**
	3221	+** May you do good and not evil.
	3222	+** May you find forgiveness for yourself and forgive others.
	3223	+** May you share freely, never taking more than you give.
	3224	+**
	3225	+******************************************************************************
	3226	+**
	3227	+** This SQLite extension implements functions that compute SHA1 hashes.
	3228	+** Two SQL functions are implemented:
	3229	+**
	3230	+** sha1(X)
	3231	+** sha1_query(Y)
	3232	+**
	3233	+** The sha1(X) function computes the SHA1 hash of the input X, or NULL if
	3234	+** X is NULL.
	3235	+**
	3236	+** The sha1_query(Y) function evalutes all queries in the SQL statements of Y
	3237	+** and returns a hash of their results.
	3238	+*/
	3239	+/* #include "sqlite3ext.h" */
	3240	+SQLITE_EXTENSION_INIT1
	3241	+#include <assert.h>
	3242	+#include <string.h>
	3243	+#include <stdarg.h>
	3244	+
	3245	+/******************************************************************************
	3246	+** The Hash Engine
	3247	+*/
	3248	+/* Context for the SHA1 hash */
	3249	+typedef struct SHA1Context SHA1Context;
	3250	+struct SHA1Context {
	3251	+ unsigned int state[5];
	3252	+ unsigned int count[2];
	3253	+ unsigned char buffer[64];
	3254	+};
	3255	+
	3256	+#define SHA_ROT(x,l,r) ((x) << (l) \| (x) >> (r))
	3257	+#define rol(x,k) SHA_ROT(x,k,32-(k))
	3258	+#define ror(x,k) SHA_ROT(x,32-(k),k)
	3259	+
	3260	+#define blk0le(i) (block[i] = (ror(block[i],8)&0xFF00FF00) \
	3261	+ \|(rol(block[i],8)&0x00FF00FF))
	3262	+#define blk0be(i) block[i]
	3263	+#define blk(i) (block[i&15] = rol(block[(i+13)&15]^block[(i+8)&15] \
	3264	+ ^block[(i+2)&15]^block[i&15],1))
	3265	+
	3266	+/*
	3267	+ * (R0+R1), R2, R3, R4 are the different operations (rounds) used in SHA1
	3268	+ *
	3269	+ * Rl0() for little-endian and Rb0() for big-endian. Endianness is
	3270	+ * determined at run-time.
	3271	+ */
	3272	+#define Rl0(v,w,x,y,z,i) \
	3273	+ z+=((w&(x^y))^y)+blk0le(i)+0x5A827999+rol(v,5);w=ror(w,2);
	3274	+#define Rb0(v,w,x,y,z,i) \
	3275	+ z+=((w&(x^y))^y)+blk0be(i)+0x5A827999+rol(v,5);w=ror(w,2);
	3276	+#define R1(v,w,x,y,z,i) \
	3277	+ z+=((w&(x^y))^y)+blk(i)+0x5A827999+rol(v,5);w=ror(w,2);
	3278	+#define R2(v,w,x,y,z,i) \
	3279	+ z+=(w^x^y)+blk(i)+0x6ED9EBA1+rol(v,5);w=ror(w,2);
	3280	+#define R3(v,w,x,y,z,i) \
	3281	+ z+=(((w\|x)&y)\|(w&x))+blk(i)+0x8F1BBCDC+rol(v,5);w=ror(w,2);
	3282	+#define R4(v,w,x,y,z,i) \
	3283	+ z+=(w^x^y)+blk(i)+0xCA62C1D6+rol(v,5);w=ror(w,2);
	3284	+
	3285	+/*
	3286	+ * Hash a single 512-bit block. This is the core of the algorithm.
	3287	+ */
	3288	+static void SHA1Transform(unsigned int state[5], const unsigned char buffer[64]){
	3289	+ unsigned int qq[5]; /* a, b, c, d, e; */
	3290	+ static int one = 1;
	3291	+ unsigned int block[16];
	3292	+ memcpy(block, buffer, 64);
	3293	+ memcpy(qq,state,5*sizeof(unsigned int));
	3294	+
	3295	+#define a qq[0]
	3296	+#define b qq[1]
	3297	+#define c qq[2]
	3298	+#define d qq[3]
	3299	+#define e qq[4]
	3300	+
	3301	+ /* Copy p->state[] to working vars */
	3302	+ /*
	3303	+ a = state[0];
	3304	+ b = state[1];
	3305	+ c = state[2];
	3306	+ d = state[3];
	3307	+ e = state[4];
	3308	+ */
	3309	+
	3310	+ /* 4 rounds of 20 operations each. Loop unrolled. */
	3311	+ if( 1 == (unsigned char)&one ){
	3312	+ Rl0(a,b,c,d,e, 0); Rl0(e,a,b,c,d, 1); Rl0(d,e,a,b,c, 2); Rl0(c,d,e,a,b, 3);
	3313	+ Rl0(b,c,d,e,a, 4); Rl0(a,b,c,d,e, 5); Rl0(e,a,b,c,d, 6); Rl0(d,e,a,b,c, 7);
	3314	+ Rl0(c,d,e,a,b, 8); Rl0(b,c,d,e,a, 9); Rl0(a,b,c,d,e,10); Rl0(e,a,b,c,d,11);
	3315	+ Rl0(d,e,a,b,c,12); Rl0(c,d,e,a,b,13); Rl0(b,c,d,e,a,14); Rl0(a,b,c,d,e,15);
	3316	+ }else{
	3317	+ Rb0(a,b,c,d,e, 0); Rb0(e,a,b,c,d, 1); Rb0(d,e,a,b,c, 2); Rb0(c,d,e,a,b, 3);
	3318	+ Rb0(b,c,d,e,a, 4); Rb0(a,b,c,d,e, 5); Rb0(e,a,b,c,d, 6); Rb0(d,e,a,b,c, 7);
	3319	+ Rb0(c,d,e,a,b, 8); Rb0(b,c,d,e,a, 9); Rb0(a,b,c,d,e,10); Rb0(e,a,b,c,d,11);
	3320	+ Rb0(d,e,a,b,c,12); Rb0(c,d,e,a,b,13); Rb0(b,c,d,e,a,14); Rb0(a,b,c,d,e,15);
	3321	+ }
	3322	+ R1(e,a,b,c,d,16); R1(d,e,a,b,c,17); R1(c,d,e,a,b,18); R1(b,c,d,e,a,19);
	3323	+ R2(a,b,c,d,e,20); R2(e,a,b,c,d,21); R2(d,e,a,b,c,22); R2(c,d,e,a,b,23);
	3324	+ R2(b,c,d,e,a,24); R2(a,b,c,d,e,25); R2(e,a,b,c,d,26); R2(d,e,a,b,c,27);
	3325	+ R2(c,d,e,a,b,28); R2(b,c,d,e,a,29); R2(a,b,c,d,e,30); R2(e,a,b,c,d,31);
	3326	+ R2(d,e,a,b,c,32); R2(c,d,e,a,b,33); R2(b,c,d,e,a,34); R2(a,b,c,d,e,35);
	3327	+ R2(e,a,b,c,d,36); R2(d,e,a,b,c,37); R2(c,d,e,a,b,38); R2(b,c,d,e,a,39);
	3328	+ R3(a,b,c,d,e,40); R3(e,a,b,c,d,41); R3(d,e,a,b,c,42); R3(c,d,e,a,b,43);
	3329	+ R3(b,c,d,e,a,44); R3(a,b,c,d,e,45); R3(e,a,b,c,d,46); R3(d,e,a,b,c,47);
	3330	+ R3(c,d,e,a,b,48); R3(b,c,d,e,a,49); R3(a,b,c,d,e,50); R3(e,a,b,c,d,51);
	3331	+ R3(d,e,a,b,c,52); R3(c,d,e,a,b,53); R3(b,c,d,e,a,54); R3(a,b,c,d,e,55);
	3332	+ R3(e,a,b,c,d,56); R3(d,e,a,b,c,57); R3(c,d,e,a,b,58); R3(b,c,d,e,a,59);
	3333	+ R4(a,b,c,d,e,60); R4(e,a,b,c,d,61); R4(d,e,a,b,c,62); R4(c,d,e,a,b,63);
	3334	+ R4(b,c,d,e,a,64); R4(a,b,c,d,e,65); R4(e,a,b,c,d,66); R4(d,e,a,b,c,67);
	3335	+ R4(c,d,e,a,b,68); R4(b,c,d,e,a,69); R4(a,b,c,d,e,70); R4(e,a,b,c,d,71);
	3336	+ R4(d,e,a,b,c,72); R4(c,d,e,a,b,73); R4(b,c,d,e,a,74); R4(a,b,c,d,e,75);
	3337	+ R4(e,a,b,c,d,76); R4(d,e,a,b,c,77); R4(c,d,e,a,b,78); R4(b,c,d,e,a,79);
	3338	+
	3339	+ /* Add the working vars back into context.state[] */
	3340	+ state[0] += a;
	3341	+ state[1] += b;
	3342	+ state[2] += c;
	3343	+ state[3] += d;
	3344	+ state[4] += e;
	3345	+
	3346	+#undef a
	3347	+#undef b
	3348	+#undef c
	3349	+#undef d
	3350	+#undef e
	3351	+}
	3352	+
	3353	+
	3354	+/* Initialize a SHA1 context */
	3355	+static void hash_init(SHA1Context *p){
	3356	+ /* SHA1 initialization constants */
	3357	+ p->state[0] = 0x67452301;
	3358	+ p->state[1] = 0xEFCDAB89;
	3359	+ p->state[2] = 0x98BADCFE;
	3360	+ p->state[3] = 0x10325476;
	3361	+ p->state[4] = 0xC3D2E1F0;
	3362	+ p->count[0] = p->count[1] = 0;
	3363	+}
	3364	+
	3365	+/* Add new content to the SHA1 hash */
	3366	+static void hash_step(
	3367	+ SHA1Context p, / Add content to this context */
	3368	+ const unsigned char data, / Data to be added */
	3369	+ unsigned int len /* Number of bytes in data */
	3370	+){
	3371	+ unsigned int i, j;
	3372	+
	3373	+ j = p->count[0];
	3374	+ if( (p->count[0] += len << 3) < j ){
	3375	+ p->count[1] += (len>>29)+1;
	3376	+ }
	3377	+ j = (j >> 3) & 63;
	3378	+ if( (j + len) > 63 ){
	3379	+ (void)memcpy(&p->buffer[j], data, (i = 64-j));
	3380	+ SHA1Transform(p->state, p->buffer);
	3381	+ for(; i + 63 < len; i += 64){
	3382	+ SHA1Transform(p->state, &data[i]);
	3383	+ }
	3384	+ j = 0;
	3385	+ }else{
	3386	+ i = 0;
	3387	+ }
	3388	+ (void)memcpy(&p->buffer[j], &data[i], len - i);
	3389	+}
	3390	+
	3391	+/* Compute a string using sqlite3_vsnprintf() and hash it */
	3392	+static void hash_step_vformat(
	3393	+ SHA1Context p, / Add content to this context */
	3394	+ const char *zFormat,
	3395	+ ...
	3396	+){
	3397	+ va_list ap;
	3398	+ int n;
	3399	+ char zBuf[50];
	3400	+ va_start(ap, zFormat);
	3401	+ sqlite3_vsnprintf(sizeof(zBuf),zBuf,zFormat,ap);
	3402	+ va_end(ap);
	3403	+ n = (int)strlen(zBuf);
	3404	+ hash_step(p, (unsigned char*)zBuf, n);
	3405	+}
	3406	+
	3407	+
	3408	+/* Add padding and compute the message digest. Render the
	3409	+** message digest as lower-case hexadecimal and put it into
	3410	+** zOut[]. zOut[] must be at least 41 bytes long. */
	3411	+static void hash_finish(
	3412	+ SHA1Context p, / The SHA1 context to finish and render */
	3413	+ char zOut, / Store hex or binary hash here */
	3414	+ int bAsBinary /* 1 for binary hash, 0 for hex hash */
	3415	+){
	3416	+ unsigned int i;
	3417	+ unsigned char finalcount[8];
	3418	+ unsigned char digest[20];
	3419	+ static const char zEncode[] = "0123456789abcdef";
	3420	+
	3421	+ for (i = 0; i < 8; i++){
	3422	+ finalcount[i] = (unsigned char)((p->count[(i >= 4 ? 0 : 1)]
	3423	+ >> ((3-(i & 3)) * 8) ) & 255); /* Endian independent */
	3424	+ }
	3425	+ hash_step(p, (const unsigned char *)"\200", 1);
	3426	+ while ((p->count[0] & 504) != 448){
	3427	+ hash_step(p, (const unsigned char *)"\0", 1);
	3428	+ }
	3429	+ hash_step(p, finalcount, 8); /* Should cause a SHA1Transform() */
	3430	+ for (i = 0; i < 20; i++){
	3431	+ digest[i] = (unsigned char)((p->state[i>>2] >> ((3-(i & 3)) * 8) ) & 255);
	3432	+ }
	3433	+ if( bAsBinary ){
	3434	+ memcpy(zOut, digest, 20);
	3435	+ }else{
	3436	+ for(i=0; i<20; i++){
	3437	+ zOut[i*2] = zEncode[(digest[i]>>4)&0xf];
	3438	+ zOut[i*2+1] = zEncode[digest[i] & 0xf];
	3439	+ }
	3440	+ zOut[i*2]= 0;
	3441	+ }
	3442	+}
	3443	+/* End of the hashing logic
	3444	+*****************************************************************************/
	3445	+
	3446	+/*
	3447	+** Implementation of the sha1(X) function.
	3448	+**
	3449	+** Return a lower-case hexadecimal rendering of the SHA1 hash of the
	3450	+** argument X. If X is a BLOB, it is hashed as is. For all other
	3451	+** types of input, X is converted into a UTF-8 string and the string
	3452	+** is hash without the trailing 0x00 terminator. The hash of a NULL
	3453	+** value is NULL.
	3454	+*/
	3455	+static void sha1Func(
	3456	+ sqlite3_context *context,
	3457	+ int argc,
	3458	+ sqlite3_value **argv
	3459	+){
	3460	+ SHA1Context cx;
	3461	+ int eType = sqlite3_value_type(argv[0]);
	3462	+ int nByte = sqlite3_value_bytes(argv[0]);
	3463	+ char zOut[44];
	3464	+
	3465	+ assert( argc==1 );
	3466	+ if( eType==SQLITE_NULL ) return;
	3467	+ hash_init(&cx);
	3468	+ if( eType==SQLITE_BLOB ){
	3469	+ hash_step(&cx, sqlite3_value_blob(argv[0]), nByte);
	3470	+ }else{
	3471	+ hash_step(&cx, sqlite3_value_text(argv[0]), nByte);
	3472	+ }
	3473	+ if( sqlite3_user_data(context)!=0 ){
	3474	+ hash_finish(&cx, zOut, 1);
	3475	+ sqlite3_result_blob(context, zOut, 20, SQLITE_TRANSIENT);
	3476	+ }else{
	3477	+ hash_finish(&cx, zOut, 0);
	3478	+ sqlite3_result_blob(context, zOut, 40, SQLITE_TRANSIENT);
	3479	+ }
	3480	+}
	3481	+
	3482	+/*
	3483	+** Implementation of the sha1_query(SQL) function.
	3484	+**
	3485	+** This function compiles and runs the SQL statement(s) given in the
	3486	+** argument. The results are hashed using SHA1 and that hash is returned.
	3487	+**
	3488	+** The original SQL text is included as part of the hash.
	3489	+**
	3490	+** The hash is not just a concatenation of the outputs. Each query
	3491	+** is delimited and each row and value within the query is delimited,
	3492	+** with all values being marked with their datatypes.
	3493	+*/
	3494	+static void sha1QueryFunc(
	3495	+ sqlite3_context *context,
	3496	+ int argc,
	3497	+ sqlite3_value **argv
	3498	+){
	3499	+ sqlite3 *db = sqlite3_context_db_handle(context);
	3500	+ const char zSql = (const char)sqlite3_value_text(argv[0]);
	3501	+ sqlite3_stmt *pStmt = 0;
	3502	+ int nCol; /* Number of columns in the result set */
	3503	+ int i; /* Loop counter */
	3504	+ int rc;
	3505	+ int n;
	3506	+ const char *z;
	3507	+ SHA1Context cx;
	3508	+ char zOut[44];
	3509	+
	3510	+ assert( argc==1 );
	3511	+ if( zSql==0 ) return;
	3512	+ hash_init(&cx);
	3513	+ while( zSql[0] ){
	3514	+ rc = sqlite3_prepare_v2(db, zSql, -1, &pStmt, &zSql);
	3515	+ if( rc ){
	3516	+ char *zMsg = sqlite3_mprintf("error SQL statement [%s]: %s",
	3517	+ zSql, sqlite3_errmsg(db));
	3518	+ sqlite3_finalize(pStmt);
	3519	+ sqlite3_result_error(context, zMsg, -1);
	3520	+ sqlite3_free(zMsg);
	3521	+ return;
	3522	+ }
	3523	+ if( !sqlite3_stmt_readonly(pStmt) ){
	3524	+ char *zMsg = sqlite3_mprintf("non-query: [%s]", sqlite3_sql(pStmt));
	3525	+ sqlite3_finalize(pStmt);
	3526	+ sqlite3_result_error(context, zMsg, -1);
	3527	+ sqlite3_free(zMsg);
	3528	+ return;
	3529	+ }
	3530	+ nCol = sqlite3_column_count(pStmt);
	3531	+ z = sqlite3_sql(pStmt);
	3532	+ n = (int)strlen(z);
	3533	+ hash_step_vformat(&cx,"S%d:",n);
	3534	+ hash_step(&cx,(unsigned char*)z,n);
	3535	+
	3536	+ /* Compute a hash over the result of the query */
	3537	+ while( SQLITE_ROW==sqlite3_step(pStmt) ){
	3538	+ hash_step(&cx,(const unsigned char*)"R",1);
	3539	+ for(i=0; i<nCol; i++){
	3540	+ switch( sqlite3_column_type(pStmt,i) ){
	3541	+ case SQLITE_NULL: {
	3542	+ hash_step(&cx, (const unsigned char*)"N",1);
	3543	+ break;
	3544	+ }
	3545	+ case SQLITE_INTEGER: {
	3546	+ sqlite3_uint64 u;
	3547	+ int j;
	3548	+ unsigned char x[9];
	3549	+ sqlite3_int64 v = sqlite3_column_int64(pStmt,i);
	3550	+ memcpy(&u, &v, 8);
	3551	+ for(j=8; j>=1; j--){
	3552	+ x[j] = u & 0xff;
	3553	+ u >>= 8;
	3554	+ }
	3555	+ x[0] = 'I';
	3556	+ hash_step(&cx, x, 9);
	3557	+ break;
	3558	+ }
	3559	+ case SQLITE_FLOAT: {
	3560	+ sqlite3_uint64 u;
	3561	+ int j;
	3562	+ unsigned char x[9];
	3563	+ double r = sqlite3_column_double(pStmt,i);
	3564	+ memcpy(&u, &r, 8);
	3565	+ for(j=8; j>=1; j--){
	3566	+ x[j] = u & 0xff;
	3567	+ u >>= 8;
	3568	+ }
	3569	+ x[0] = 'F';
	3570	+ hash_step(&cx,x,9);
	3571	+ break;
	3572	+ }
	3573	+ case SQLITE_TEXT: {
	3574	+ int n2 = sqlite3_column_bytes(pStmt, i);
	3575	+ const unsigned char *z2 = sqlite3_column_text(pStmt, i);
	3576	+ hash_step_vformat(&cx,"T%d:",n2);
	3577	+ hash_step(&cx, z2, n2);
	3578	+ break;
	3579	+ }
	3580	+ case SQLITE_BLOB: {
	3581	+ int n2 = sqlite3_column_bytes(pStmt, i);
	3582	+ const unsigned char *z2 = sqlite3_column_blob(pStmt, i);
	3583	+ hash_step_vformat(&cx,"B%d:",n2);
	3584	+ hash_step(&cx, z2, n2);
	3585	+ break;
	3586	+ }
	3587	+ }
	3588	+ }
	3589	+ }
	3590	+ sqlite3_finalize(pStmt);
	3591	+ }
	3592	+ hash_finish(&cx, zOut, 0);
	3593	+ sqlite3_result_text(context, zOut, 40, SQLITE_TRANSIENT);
	3594	+}
	3595	+
	3596	+
	3597	+#ifdef _WIN32
	3598	+
	3599	+#endif
	3600	+int sqlite3_sha_init(
	3601	+ sqlite3 *db,
	3602	+ char **pzErrMsg,
	3603	+ const sqlite3_api_routines *pApi
	3604	+){
	3605	+ int rc = SQLITE_OK;
	3606	+ static int one = 1;
	3607	+ SQLITE_EXTENSION_INIT2(pApi);
	3608	+ (void)pzErrMsg; /* Unused parameter */
	3609	+ rc = sqlite3_create_function(db, "sha1", 1,
	3610	+ SQLITE_UTF8 \| SQLITE_INNOCUOUS \| SQLITE_DETERMINISTIC,
	3611	+ 0, sha1Func, 0, 0);
	3612	+ if( rc==SQLITE_OK ){
	3613	+ rc = sqlite3_create_function(db, "sha1b", 1,
	3614	+ SQLITE_UTF8 \| SQLITE_INNOCUOUS \| SQLITE_DETERMINISTIC,
	3615	+ (void*)&one, sha1Func, 0, 0);
	3616	+ }
	3617	+ if( rc==SQLITE_OK ){
	3618	+ rc = sqlite3_create_function(db, "sha1_query", 1,
	3619	+ SQLITE_UTF8\|SQLITE_DIRECTONLY, 0,
	3620	+ sha1QueryFunc, 0, 0);
	3621	+ }
	3622	+ return rc;
	3623	+}
	3624	+
	3625	+/*********************** End ../ext/misc/sha1.c ******************/
3788	3626	/*********************** Begin ../ext/misc/uint.c ****************/
3789	3627	/*
3790	3628	** 2020-04-14
3791	3629	**
3792	3630	** The author disclaims copyright to this source code. In place of
		@@ -5075,11 +4913,11 @@
5075	4913	}else if( !p->bSorted \|\| y>=p->a[p->nUsed-1] ){
5076	4914	p->a[p->nUsed++] = y;
5077	4915	}else if( p->bKeepSorted ){
5078	4916	int i;
5079	4917	i = percentBinarySearch(p, y, 0);
5080		- if( i<p->nUsed ){
	4918	+ if( i<(int)p->nUsed ){
5081	4919	memmove(&p->a[i+1], &p->a[i], (p->nUsed-i)*sizeof(p->a[0]));
5082	4920	}
5083	4921	p->a[i] = y;
5084	4922	p->nUsed++;
5085	4923	}else{
		@@ -5191,11 +5029,11 @@
5191	5029
5192	5030	/* Find and remove the row */
5193	5031	i = percentBinarySearch(p, y, 1);
5194	5032	if( i>=0 ){
5195	5033	p->nUsed--;
5196		- if( i<p->nUsed ){
	5034	+ if( i<(int)p->nUsed ){
5197	5035	memmove(&p->a[i], &p->a[i+1], (p->nUsed - i)*sizeof(p->a[0]));
5198	5036	}
5199	5037	}
5200	5038	}
5201	5039
		@@ -5251,15 +5089,15 @@
5251	5089	sqlite3 *db,
5252	5090	char **pzErrMsg,
5253	5091	const sqlite3_api_routines *pApi
5254	5092	){
5255	5093	int rc = SQLITE_OK;
5256		- int i;
5257		-#if defined(SQLITE3_H) \|\| defined(SQLITE_STATIC_PERCENTILE)
5258		- (void)pApi; /* Unused parameter */
5259		-#else
	5094	+ unsigned int i;
	5095	+#ifdef SQLITE3EXT_H
5260	5096	SQLITE_EXTENSION_INIT2(pApi);
	5097	+#else
	5098	+ (void)pApi; /* Unused parameter */
5261	5099	#endif
5262	5100	(void)pzErrMsg; /* Unused parameter */
5263	5101	for(i=0; i<sizeof(aPercentFunc)/sizeof(aPercentFunc[0]); i++){
5264	5102	rc = sqlite3_create_window_function(db,
5265	5103	aPercentFunc[i].zName,
		@@ -7012,50 +6850,58 @@
7012	6850	aIdx[4] = i;
7013	6851	idxNum \|= 0x40;
7014	6852	}
7015	6853	continue;
7016	6854	}
7017		- if( pConstraint->iColumn==SERIES_COLUMN_VALUE ){
7018		- switch( op ){
7019		- case SQLITE_INDEX_CONSTRAINT_EQ:
7020		- case SQLITE_INDEX_CONSTRAINT_IS: {
7021		- idxNum \|= 0x0080;
7022		- idxNum &= ~0x3300;
7023		- aIdx[5] = i;
7024		- aIdx[6] = -1;
7025		- bStartSeen = 1;
7026		- break;
7027		- }
7028		- case SQLITE_INDEX_CONSTRAINT_GE: {
7029		- if( idxNum & 0x0080 ) break;
7030		- idxNum \|= 0x0100;
7031		- idxNum &= ~0x0200;
7032		- aIdx[5] = i;
7033		- bStartSeen = 1;
7034		- break;
7035		- }
7036		- case SQLITE_INDEX_CONSTRAINT_GT: {
7037		- if( idxNum & 0x0080 ) break;
7038		- idxNum \|= 0x0200;
7039		- idxNum &= ~0x0100;
7040		- aIdx[5] = i;
7041		- bStartSeen = 1;
7042		- break;
7043		- }
7044		- case SQLITE_INDEX_CONSTRAINT_LE: {
7045		- if( idxNum & 0x0080 ) break;
7046		- idxNum \|= 0x1000;
7047		- idxNum &= ~0x2000;
7048		- aIdx[6] = i;
7049		- break;
7050		- }
7051		- case SQLITE_INDEX_CONSTRAINT_LT: {
7052		- if( idxNum & 0x0080 ) break;
7053		- idxNum \|= 0x2000;
7054		- idxNum &= ~0x1000;
7055		- aIdx[6] = i;
7056		- break;
	6855	+ if( pConstraint->iColumn<SERIES_COLUMN_START ){
	6856	+ if( pConstraint->iColumn==SERIES_COLUMN_VALUE && pConstraint->usable ){
	6857	+ switch( op ){
	6858	+ case SQLITE_INDEX_CONSTRAINT_EQ:
	6859	+ case SQLITE_INDEX_CONSTRAINT_IS: {
	6860	+ idxNum \|= 0x0080;
	6861	+ idxNum &= ~0x3300;
	6862	+ aIdx[5] = i;
	6863	+ aIdx[6] = -1;
	6864	+#ifndef ZERO_ARGUMENT_GENERATE_SERIES
	6865	+ bStartSeen = 1;
	6866	+#endif
	6867	+ break;
	6868	+ }
	6869	+ case SQLITE_INDEX_CONSTRAINT_GE: {
	6870	+ if( idxNum & 0x0080 ) break;
	6871	+ idxNum \|= 0x0100;
	6872	+ idxNum &= ~0x0200;
	6873	+ aIdx[5] = i;
	6874	+#ifndef ZERO_ARGUMENT_GENERATE_SERIES
	6875	+ bStartSeen = 1;
	6876	+#endif
	6877	+ break;
	6878	+ }
	6879	+ case SQLITE_INDEX_CONSTRAINT_GT: {
	6880	+ if( idxNum & 0x0080 ) break;
	6881	+ idxNum \|= 0x0200;
	6882	+ idxNum &= ~0x0100;
	6883	+ aIdx[5] = i;
	6884	+#ifndef ZERO_ARGUMENT_GENERATE_SERIES
	6885	+ bStartSeen = 1;
	6886	+#endif
	6887	+ break;
	6888	+ }
	6889	+ case SQLITE_INDEX_CONSTRAINT_LE: {
	6890	+ if( idxNum & 0x0080 ) break;
	6891	+ idxNum \|= 0x1000;
	6892	+ idxNum &= ~0x2000;
	6893	+ aIdx[6] = i;
	6894	+ break;
	6895	+ }
	6896	+ case SQLITE_INDEX_CONSTRAINT_LT: {
	6897	+ if( idxNum & 0x0080 ) break;
	6898	+ idxNum \|= 0x2000;
	6899	+ idxNum &= ~0x1000;
	6900	+ aIdx[6] = i;
	6901	+ break;
	6902	+ }
7057	6903	}
7058	6904	}
7059	6905	continue;
7060	6906	}
7061	6907	iCol = pConstraint->iColumn - SERIES_COLUMN_START;
		@@ -8115,11 +7961,11 @@
8115	7961	** If the optional MTIME argument is present, then it is interpreted
8116	7962	** as an integer - the number of seconds since the unix epoch. The
8117	7963	** modification-time of the target file is set to this value before
8118	7964	** returning.
8119	7965	**
8120		-** If three or more arguments are passed to this function and an
	7966	+** If five or more arguments are passed to this function and an
8121	7967	** error is encountered, an exception is raised.
8122	7968	**
8123	7969	** READFILE(FILE):
8124	7970	**
8125	7971	** Read and return the contents of file FILE (type blob) from disk.
		@@ -8185,10 +8031,17 @@
8185	8031	# define lstat(path,buf) stat(path,buf)
8186	8032	#endif
8187	8033	#include <time.h>
8188	8034	#include <errno.h>
8189	8035
	8036	+/* When used as part of the CLI, the sqlite3_stdio.h module will have
	8037	+** been included before this one. In that case use the sqlite3_stdio.h
	8038	+** #defines. If not, create our own for fopen().
	8039	+*/
	8040	+#ifndef _SQLITE3_STDIO_H_
	8041	+# define sqlite3_fopen fopen
	8042	+#endif
8190	8043
8191	8044	/*
8192	8045	** Structure of the fsdir() table-valued function
8193	8046	*/
8194	8047	/* 0 1 2 3 4 5 */
		@@ -8217,11 +8070,11 @@
8217	8070	sqlite3_int64 nIn;
8218	8071	void *pBuf;
8219	8072	sqlite3 *db;
8220	8073	int mxBlob;
8221	8074
8222		- in = fopen(zName, "rb");
	8075	+ in = sqlite3_fopen(zName, "rb");
8223	8076	if( in==0 ){
8224	8077	/* File does not exist or is unreadable. Leave the result set to NULL. */
8225	8078	return;
8226	8079	}
8227	8080	fseek(in, 0, SEEK_END);
		@@ -8472,11 +8325,11 @@
8472	8325	}
8473	8326	}else{
8474	8327	sqlite3_int64 nWrite = 0;
8475	8328	const char *z;
8476	8329	int rc = 0;
8477		- FILE *out = fopen(zFile, "wb");
	8330	+ FILE *out = sqlite3_fopen(zFile, "wb");
8478	8331	if( out==0 ) return 1;
8479	8332	z = (const char*)sqlite3_value_blob(pData);
8480	8333	if( z ){
8481	8334	sqlite3_int64 n = fwrite(z, 1, sqlite3_value_bytes(pData), out);
8482	8335	nWrite = sqlite3_value_bytes(pData);
		@@ -10333,10 +10186,18 @@
10333	10186	#ifndef SQLITE_NO_STDINT
10334	10187	# include <stdint.h>
10335	10188	#endif
10336	10189
10337	10190	#include <zlib.h>
	10191	+
	10192	+/* When used as part of the CLI, the sqlite3_stdio.h module will have
	10193	+** been included before this one. In that case use the sqlite3_stdio.h
	10194	+** #defines. If not, create our own for fopen().
	10195	+*/
	10196	+#ifndef _SQLITE3_STDIO_H_
	10197	+# define sqlite3_fopen fopen
	10198	+#endif
10338	10199
10339	10200	#ifndef SQLITE_OMIT_VIRTUALTABLE
10340	10201
10341	10202	#ifndef SQLITE_AMALGAMATION
10342	10203
		@@ -11590,11 +11451,11 @@
11590	11451	}else{
11591	11452	zFile = (const char*)sqlite3_value_text(argv[0]);
11592	11453	}
11593	11454
11594	11455	if( 0==pTab->pWriteFd && 0==bInMemory ){
11595		- pCsr->pFile = zFile ? fopen(zFile, "rb") : 0;
	11456	+ pCsr->pFile = zFile ? sqlite3_fopen(zFile, "rb") : 0;
11596	11457	if( pCsr->pFile==0 ){
11597	11458	zipfileCursorErr(pCsr, "cannot open file: %s", zFile);
11598	11459	rc = SQLITE_ERROR;
11599	11460	}else{
11600	11461	rc = zipfileReadEOCD(pTab, 0, 0, pCsr->pFile, &pCsr->eocd);
		@@ -11780,11 +11641,11 @@
11780	11641
11781	11642	/* Open a write fd on the file. Also load the entire central directory
11782	11643	** structure into memory. During the transaction any new file data is
11783	11644	** appended to the archive file, but the central directory is accumulated
11784	11645	** in main-memory until the transaction is committed. */
11785		- pTab->pWriteFd = fopen(pTab->zFile, "ab+");
	11646	+ pTab->pWriteFd = sqlite3_fopen(pTab->zFile, "ab+");
11786	11647	if( pTab->pWriteFd==0 ){
11787	11648	pTab->base.zErrMsg = sqlite3_mprintf(
11788	11649	"zipfile: failed to open file %s for writing", pTab->zFile
11789	11650	);
11790	11651	rc = SQLITE_ERROR;
		@@ -14233,10 +14094,70 @@
14233	14094	}
14234	14095
14235	14096	return rc;
14236	14097	}
14237	14098
	14099	+/*
	14100	+** This function tests if the schema of the main database of database handle
	14101	+** db contains an object named zTab. Assuming no error occurs, output parameter
	14102	+** (*pbContains) is set to true if zTab exists, or false if it does not.
	14103	+**
	14104	+** Or, if an error occurs, an SQLite error code is returned. The final value
	14105	+** of (*pbContains) is undefined in this case.
	14106	+*/
	14107	+static int expertDbContainsObject(
	14108	+ sqlite3 *db,
	14109	+ const char *zTab,
	14110	+ int pbContains / OUT: True if object exists */
	14111	+){
	14112	+ const char *zSql = "SELECT 1 FROM sqlite_schema WHERE name = ?";
	14113	+ sqlite3_stmt *pSql = 0;
	14114	+ int rc = SQLITE_OK;
	14115	+ int ret = 0;
	14116	+
	14117	+ rc = sqlite3_prepare_v2(db, zSql, -1, &pSql, 0);
	14118	+ if( rc==SQLITE_OK ){
	14119	+ sqlite3_bind_text(pSql, 1, zTab, -1, SQLITE_STATIC);
	14120	+ if( SQLITE_ROW==sqlite3_step(pSql) ){
	14121	+ ret = 1;
	14122	+ }
	14123	+ rc = sqlite3_finalize(pSql);
	14124	+ }
	14125	+
	14126	+ *pbContains = ret;
	14127	+ return rc;
	14128	+}
	14129	+
	14130	+/*
	14131	+** Execute SQL command zSql using database handle db. If no error occurs,
	14132	+** set (*pzErr) to NULL and return SQLITE_OK.
	14133	+**
	14134	+** If an error does occur, return an SQLite error code and set (*pzErr) to
	14135	+** point to a buffer containing an English language error message. Except,
	14136	+** if the error message begins with "no such module:", then ignore the
	14137	+** error and return as if the SQL statement had succeeded.
	14138	+**
	14139	+** This is used to copy as much of the database schema as possible while
	14140	+** ignoring any errors related to missing virtual table modules.
	14141	+*/
	14142	+static int expertSchemaSql(sqlite3 db, const char zSql, char **pzErr){
	14143	+ int rc = SQLITE_OK;
	14144	+ char *zErr = 0;
	14145	+
	14146	+ rc = sqlite3_exec(db, zSql, 0, 0, &zErr);
	14147	+ if( rc!=SQLITE_OK && zErr ){
	14148	+ int nErr = STRLEN(zErr);
	14149	+ if( nErr>=15 && memcmp(zErr, "no such module:", 15)==0 ){
	14150	+ sqlite3_free(zErr);
	14151	+ rc = SQLITE_OK;
	14152	+ zErr = 0;
	14153	+ }
	14154	+ }
	14155	+
	14156	+ *pzErr = zErr;
	14157	+ return rc;
	14158	+}
14238	14159
14239	14160	static int idxCreateVtabSchema(sqlite3expert p, char *pzErrmsg){
14240	14161	int rc = idxRegisterVtab(p);
14241	14162	sqlite3_stmt *pSchema = 0;
14242	14163
		@@ -14244,30 +14165,39 @@
14244	14165	**
14245	14166	** 1) Add an entry to the p->pTable list, and
14246	14167	** 2) Create the equivalent virtual table in dbv.
14247	14168	*/
14248	14169	rc = idxPrepareStmt(p->db, &pSchema, pzErrmsg,
14249		- "SELECT type, name, sql, 1 FROM sqlite_schema "
14250		- "WHERE type IN ('table','view') AND name NOT LIKE 'sqlite_%%' "
	14170	+ "SELECT type, name, sql, 1, "
	14171	+ " substr(sql,1,14)=='create virtual' COLLATE nocase "
	14172	+ "FROM sqlite_schema "
	14173	+ "WHERE type IN ('table','view') AND "
	14174	+ " substr(name,1,7)!='sqlite_' COLLATE nocase "
14251	14175	" UNION ALL "
14252		- "SELECT type, name, sql, 2 FROM sqlite_schema "
	14176	+ "SELECT type, name, sql, 2, 0 FROM sqlite_schema "
14253	14177	"WHERE type = 'trigger'"
14254	14178	" AND tbl_name IN(SELECT name FROM sqlite_schema WHERE type = 'view') "
14255		- "ORDER BY 4, 1"
	14179	+ "ORDER BY 4, 5 DESC, 1"
14256	14180	);
14257	14181	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSchema) ){
14258	14182	const char zType = (const char)sqlite3_column_text(pSchema, 0);
14259	14183	const char zName = (const char)sqlite3_column_text(pSchema, 1);
14260	14184	const char zSql = (const char)sqlite3_column_text(pSchema, 2);
	14185	+ int bVirtual = sqlite3_column_int(pSchema, 4);
	14186	+ int bExists = 0;
14261	14187
14262	14188	if( zType==0 \|\| zName==0 ) continue;
14263		- if( zType[0]=='v' \|\| zType[1]=='r' ){
14264		- if( zSql ) rc = sqlite3_exec(p->dbv, zSql, 0, 0, pzErrmsg);
	14189	+ rc = expertDbContainsObject(p->dbv, zName, &bExists);
	14190	+ if( rc \|\| bExists ) continue;
	14191	+
	14192	+ if( zType[0]=='v' \|\| zType[1]=='r' \|\| bVirtual ){
	14193	+ /* A view. Or a trigger on a view. */
	14194	+ if( zSql ) rc = expertSchemaSql(p->dbv, zSql, pzErrmsg);
14265	14195	}else{
14266	14196	IdxTable *pTab;
14267	14197	rc = idxGetTableInfo(p->db, zName, &pTab, pzErrmsg);
14268		- if( rc==SQLITE_OK ){
	14198	+ if( rc==SQLITE_OK && ALWAYS(pTab!=0) ){
14269	14199	int i;
14270	14200	char *zInner = 0;
14271	14201	char *zOuter = 0;
14272	14202	pTab->pNext = p->pTable;
14273	14203	p->pTable = pTab;
		@@ -14464,10 +14394,16 @@
14464	14394	sqlite3_bind_text(pIndexXInfo, 1, zIdx, -1, SQLITE_STATIC);
14465	14395	while( SQLITE_OK==rc && SQLITE_ROW==sqlite3_step(pIndexXInfo) ){
14466	14396	const char *zComma = zCols==0 ? "" : ", ";
14467	14397	const char zName = (const char)sqlite3_column_text(pIndexXInfo, 0);
14468	14398	const char zColl = (const char)sqlite3_column_text(pIndexXInfo, 1);
	14399	+ if( zName==0 ){
	14400	+ /* This index contains an expression. Ignore it. */
	14401	+ sqlite3_free(zCols);
	14402	+ sqlite3_free(zOrder);
	14403	+ return sqlite3_reset(pIndexXInfo);
	14404	+ }
14469	14405	zCols = idxAppendText(&rc, zCols,
14470	14406	"%sx.%Q IS sqlite_expert_rem(%d, x.%Q) COLLATE %s",
14471	14407	zComma, zName, nCol, zName, zColl
14472	14408	);
14473	14409	zOrder = idxAppendText(&rc, zOrder, "%s%d", zComma, ++nCol);
		@@ -14792,16 +14728,22 @@
14792	14728
14793	14729	/* Copy the entire schema of database [db] into [dbm]. */
14794	14730	if( rc==SQLITE_OK ){
14795	14731	sqlite3_stmt *pSql = 0;
14796	14732	rc = idxPrintfPrepareStmt(pNew->db, &pSql, pzErrmsg,
14797		- "SELECT sql FROM sqlite_schema WHERE name NOT LIKE 'sqlite_%%'"
14798		- " AND sql NOT LIKE 'CREATE VIRTUAL %%' ORDER BY rowid"
	14733	+ "SELECT sql, name, substr(sql,1,14)=='create virtual' COLLATE nocase"
	14734	+ " FROM sqlite_schema WHERE substr(name,1,7)!='sqlite_' COLLATE nocase"
	14735	+ " ORDER BY 3 DESC, rowid"
14799	14736	);
14800	14737	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
14801	14738	const char zSql = (const char)sqlite3_column_text(pSql, 0);
14802		- if( zSql ) rc = sqlite3_exec(pNew->dbm, zSql, 0, 0, pzErrmsg);
	14739	+ const char zName = (const char)sqlite3_column_text(pSql, 1);
	14740	+ int bExists = 0;
	14741	+ rc = expertDbContainsObject(pNew->dbm, zName, &bExists);
	14742	+ if( rc==SQLITE_OK && zSql && bExists==0 ){
	14743	+ rc = expertSchemaSql(pNew->dbm, zSql, pzErrmsg);
	14744	+ }
14803	14745	}
14804	14746	idxFinalize(&rc, pSql);
14805	14747	}
14806	14748
14807	14749	/* Create the vtab schema */
		@@ -16211,10 +16153,1197 @@
16211	16153	}
16212	16154	return rc;
16213	16155	}
16214	16156
16215	16157	/*********************** End ../ext/misc/stmtrand.c ******************/
	16158	+/*********************** Begin ../ext/misc/vfstrace.c ****************/
	16159	+/*
	16160	+** 2011 March 16
	16161	+**
	16162	+** The author disclaims copyright to this source code. In place of
	16163	+** a legal notice, here is a blessing:
	16164	+**
	16165	+** May you do good and not evil.
	16166	+** May you find forgiveness for yourself and forgive others.
	16167	+** May you share freely, never taking more than you give.
	16168	+**
	16169	+******************************************************************************
	16170	+**
	16171	+** This file contains code implements a VFS shim that writes diagnostic
	16172	+** output for each VFS call, similar to "strace".
	16173	+**
	16174	+** USAGE:
	16175	+**
	16176	+** This source file exports a single symbol which is the name of a
	16177	+** function:
	16178	+**
	16179	+** int vfstrace_register(
	16180	+** const char *zTraceName, // Name of the newly constructed VFS
	16181	+** const char *zOldVfsName, // Name of the underlying VFS
	16182	+** int (xOut)(const char,void*), // Output routine. ex: fputs
	16183	+** void *pOutArg, // 2nd argument to xOut. ex: stderr
	16184	+** int makeDefault // Make the new VFS the default
	16185	+** );
	16186	+**
	16187	+** Applications that want to trace their VFS usage must provide a callback
	16188	+** function with this prototype:
	16189	+**
	16190	+** int traceOutput(const char zMessage, void pAppData);
	16191	+**
	16192	+** This function will "output" the trace messages, where "output" can
	16193	+** mean different things to different applications. The traceOutput function
	16194	+** for the command-line shell (see shell.c) is "fputs" from the standard
	16195	+** library, which means that all trace output is written on the stream
	16196	+** specified by the second argument. In the case of the command-line shell
	16197	+** the second argument is stderr. Other applications might choose to output
	16198	+** trace information to a file, over a socket, or write it into a buffer.
	16199	+**
	16200	+** The vfstrace_register() function creates a new "shim" VFS named by
	16201	+** the zTraceName parameter. A "shim" VFS is an SQLite backend that does
	16202	+** not really perform the duties of a true backend, but simply filters or
	16203	+** interprets VFS calls before passing them off to another VFS which does
	16204	+** the actual work. In this case the other VFS - the one that does the
	16205	+** real work - is identified by the second parameter, zOldVfsName. If
	16206	+** the 2nd parameter is NULL then the default VFS is used. The common
	16207	+** case is for the 2nd parameter to be NULL.
	16208	+**
	16209	+** The third and fourth parameters are the pointer to the output function
	16210	+** and the second argument to the output function. For the SQLite
	16211	+** command-line shell, when the -vfstrace option is used, these parameters
	16212	+** are fputs and stderr, respectively.
	16213	+**
	16214	+** The fifth argument is true (non-zero) to cause the newly created VFS
	16215	+** to become the default VFS. The common case is for the fifth parameter
	16216	+** to be true.
	16217	+**
	16218	+** The call to vfstrace_register() simply creates the shim VFS that does
	16219	+** tracing. The application must also arrange to use the new VFS for
	16220	+** all database connections that are created and for which tracing is
	16221	+** desired. This can be done by specifying the trace VFS using URI filename
	16222	+** notation, or by specifying the trace VFS as the 4th parameter to
	16223	+** sqlite3_open_v2() or by making the trace VFS be the default (by setting
	16224	+** the 5th parameter of vfstrace_register() to 1).
	16225	+**
	16226	+**
	16227	+** ENABLING VFSTRACE IN A COMMAND-LINE SHELL
	16228	+**
	16229	+** The SQLite command line shell implemented by the shell.c source file
	16230	+** can be used with this module. To compile in -vfstrace support, first
	16231	+** gather this file (test_vfstrace.c), the shell source file (shell.c),
	16232	+** and the SQLite amalgamation source files (sqlite3.c, sqlite3.h) into
	16233	+** the working directory. Then compile using a command like the following:
	16234	+**
	16235	+** gcc -o sqlite3 -Os -I. -DSQLITE_ENABLE_VFSTRACE \
	16236	+** -DSQLITE_THREADSAFE=0 -DSQLITE_ENABLE_FTS3 -DSQLITE_ENABLE_RTREE \
	16237	+** -DHAVE_READLINE -DHAVE_USLEEP=1 \
	16238	+** shell.c test_vfstrace.c sqlite3.c -ldl -lreadline -lncurses
	16239	+**
	16240	+** The gcc command above works on Linux and provides (in addition to the
	16241	+** -vfstrace option) support for FTS3 and FTS4, RTREE, and command-line
	16242	+** editing using the readline library. The command-line shell does not
	16243	+** use threads so we added -DSQLITE_THREADSAFE=0 just to make the code
	16244	+** run a little faster. For compiling on a Mac, you'll probably need
	16245	+** to omit the -DHAVE_READLINE, the -lreadline, and the -lncurses options.
	16246	+** The compilation could be simplified to just this:
	16247	+**
	16248	+** gcc -DSQLITE_ENABLE_VFSTRACE \
	16249	+** shell.c test_vfstrace.c sqlite3.c -ldl -lpthread
	16250	+**
	16251	+** In this second example, all unnecessary options have been removed
	16252	+** Note that since the code is now threadsafe, we had to add the -lpthread
	16253	+** option to pull in the pthreads library.
	16254	+**
	16255	+** To cross-compile for windows using MinGW, a command like this might
	16256	+** work:
	16257	+**
	16258	+** /opt/mingw/bin/i386-mingw32msvc-gcc -o sqlite3.exe -Os -I \
	16259	+** -DSQLITE_THREADSAFE=0 -DSQLITE_ENABLE_VFSTRACE \
	16260	+** shell.c test_vfstrace.c sqlite3.c
	16261	+**
	16262	+** Similar compiler commands will work on different systems. The key
	16263	+** invariants are (1) you must have -DSQLITE_ENABLE_VFSTRACE so that
	16264	+** the shell.c source file will know to include the -vfstrace command-line
	16265	+** option and (2) you must compile and link the three source files
	16266	+** shell,c, test_vfstrace.c, and sqlite3.c.
	16267	+**
	16268	+** RUNTIME CONTROL OF VFSTRACE OUTPUT
	16269	+**
	16270	+** The application can use the "vfstrace" pragma to control which VFS
	16271	+** APIs are traced. To disable all output:
	16272	+**
	16273	+** PRAGMA vfstrace('-all');
	16274	+**
	16275	+** To enable all output (which is the default setting):
	16276	+**
	16277	+** PRAGMA vfstrace('+all');
	16278	+**
	16279	+** Individual APIs can be enabled or disabled by name, with or without
	16280	+** the initial "x" character. For example, to set up for tracing lock
	16281	+** primatives only:
	16282	+**
	16283	+** PRAGMA vfstrace('-all, +Lock,Unlock,ShmLock');
	16284	+**
	16285	+** The argument to the vfstrace pragma ignores capitalization and any
	16286	+** characters other than alphabetics, '+', and '-'.
	16287	+*/
	16288	+#include <stdlib.h>
	16289	+#include <string.h>
	16290	+/* #include "sqlite3.h" */
	16291	+
	16292	+/*
	16293	+** An instance of this structure is attached to the each trace VFS to
	16294	+** provide auxiliary information.
	16295	+*/
	16296	+typedef struct vfstrace_info vfstrace_info;
	16297	+struct vfstrace_info {
	16298	+ sqlite3_vfs pRootVfs; / The underlying real VFS */
	16299	+ int (xOut)(const char, void); / Send output here */
	16300	+ unsigned int mTrace; /* Mask of interfaces to trace */
	16301	+ u8 bOn; /* Tracing on/off */
	16302	+ void pOutArg; / First argument to xOut */
	16303	+ const char zVfsName; / Name of this trace-VFS */
	16304	+ sqlite3_vfs pTraceVfs; / Pointer back to the trace VFS */
	16305	+};
	16306	+
	16307	+/*
	16308	+** The sqlite3_file object for the trace VFS
	16309	+*/
	16310	+typedef struct vfstrace_file vfstrace_file;
	16311	+struct vfstrace_file {
	16312	+ sqlite3_file base; /* Base class. Must be first */
	16313	+ vfstrace_info pInfo; / The trace-VFS to which this file belongs */
	16314	+ const char zFName; / Base name of the file */
	16315	+ sqlite3_file pReal; / The real underlying file */
	16316	+};
	16317	+
	16318	+/*
	16319	+** Bit values for vfstrace_info.mTrace.
	16320	+*/
	16321	+#define VTR_CLOSE 0x00000001
	16322	+#define VTR_READ 0x00000002
	16323	+#define VTR_WRITE 0x00000004
	16324	+#define VTR_TRUNC 0x00000008
	16325	+#define VTR_SYNC 0x00000010
	16326	+#define VTR_FSIZE 0x00000020
	16327	+#define VTR_LOCK 0x00000040
	16328	+#define VTR_UNLOCK 0x00000080
	16329	+#define VTR_CRL 0x00000100
	16330	+#define VTR_FCTRL 0x00000200
	16331	+#define VTR_SECSZ 0x00000400
	16332	+#define VTR_DEVCHAR 0x00000800
	16333	+#define VTR_SHMLOCK 0x00001000
	16334	+#define VTR_SHMMAP 0x00002000
	16335	+#define VTR_SHMBAR 0x00004000
	16336	+#define VTR_SHMUNMAP 0x00008000
	16337	+#define VTR_OPEN 0x00010000
	16338	+#define VTR_DELETE 0x00020000
	16339	+#define VTR_ACCESS 0x00040000
	16340	+#define VTR_FULLPATH 0x00080000
	16341	+#define VTR_DLOPEN 0x00100000
	16342	+#define VTR_DLERR 0x00200000
	16343	+#define VTR_DLSYM 0x00400000
	16344	+#define VTR_DLCLOSE 0x00800000
	16345	+#define VTR_RAND 0x01000000
	16346	+#define VTR_SLEEP 0x02000000
	16347	+#define VTR_CURTIME 0x04000000
	16348	+#define VTR_LASTERR 0x08000000
	16349	+
	16350	+/*
	16351	+** Method declarations for vfstrace_file.
	16352	+*/
	16353	+static int vfstraceClose(sqlite3_file*);
	16354	+static int vfstraceRead(sqlite3_file, void, int iAmt, sqlite3_int64 iOfst);
	16355	+static int vfstraceWrite(sqlite3_file,const void,int iAmt, sqlite3_int64);
	16356	+static int vfstraceTruncate(sqlite3_file*, sqlite3_int64 size);
	16357	+static int vfstraceSync(sqlite3_file*, int flags);
	16358	+static int vfstraceFileSize(sqlite3_file, sqlite3_int64 pSize);
	16359	+static int vfstraceLock(sqlite3_file*, int);
	16360	+static int vfstraceUnlock(sqlite3_file*, int);
	16361	+static int vfstraceCheckReservedLock(sqlite3_file, int );
	16362	+static int vfstraceFileControl(sqlite3_file, int op, void pArg);
	16363	+static int vfstraceSectorSize(sqlite3_file*);
	16364	+static int vfstraceDeviceCharacteristics(sqlite3_file*);
	16365	+static int vfstraceShmLock(sqlite3_file*,int,int,int);
	16366	+static int vfstraceShmMap(sqlite3_file,int,int,int, void volatile *);
	16367	+static void vfstraceShmBarrier(sqlite3_file*);
	16368	+static int vfstraceShmUnmap(sqlite3_file*,int);
	16369	+
	16370	+/*
	16371	+** Method declarations for vfstrace_vfs.
	16372	+*/
	16373	+static int vfstraceOpen(sqlite3_vfs, const char , sqlite3_file, int , int );
	16374	+static int vfstraceDelete(sqlite3_vfs, const char zName, int syncDir);
	16375	+static int vfstraceAccess(sqlite3_vfs, const char zName, int flags, int *);
	16376	+static int vfstraceFullPathname(sqlite3_vfs, const char zName, int, char *);
	16377	+static void vfstraceDlOpen(sqlite3_vfs, const char *zFilename);
	16378	+static void vfstraceDlError(sqlite3_vfs, int nByte, char zErrMsg);
	16379	+static void (vfstraceDlSym(sqlite3_vfs,void, const char zSymbol))(void);
	16380	+static void vfstraceDlClose(sqlite3_vfs, void);
	16381	+static int vfstraceRandomness(sqlite3_vfs, int nByte, char zOut);
	16382	+static int vfstraceSleep(sqlite3_vfs*, int microseconds);
	16383	+static int vfstraceCurrentTime(sqlite3_vfs, double);
	16384	+static int vfstraceGetLastError(sqlite3_vfs, int, char);
	16385	+static int vfstraceCurrentTimeInt64(sqlite3_vfs, sqlite3_int64);
	16386	+static int vfstraceSetSystemCall(sqlite3_vfs,const char, sqlite3_syscall_ptr);
	16387	+static sqlite3_syscall_ptr vfstraceGetSystemCall(sqlite3_vfs, const char );
	16388	+static const char vfstraceNextSystemCall(sqlite3_vfs, const char *zName);
	16389	+
	16390	+/*
	16391	+** Return a pointer to the tail of the pathname. Examples:
	16392	+**
	16393	+** /home/drh/xyzzy.txt -> xyzzy.txt
	16394	+** xyzzy.txt -> xyzzy.txt
	16395	+*/
	16396	+static const char fileTail(const char z){
	16397	+ size_t i;
	16398	+ if( z==0 ) return 0;
	16399	+ i = strlen(z)-1;
	16400	+ while( i>0 && z[i-1]!='/' ){ i--; }
	16401	+ return &z[i];
	16402	+}
	16403	+
	16404	+/*
	16405	+** Send trace output defined by zFormat and subsequent arguments.
	16406	+*/
	16407	+static void vfstrace_printf(
	16408	+ vfstrace_info *pInfo,
	16409	+ const char *zFormat,
	16410	+ ...
	16411	+){
	16412	+ va_list ap;
	16413	+ char *zMsg;
	16414	+ if( pInfo->bOn ){
	16415	+ va_start(ap, zFormat);
	16416	+ zMsg = sqlite3_vmprintf(zFormat, ap);
	16417	+ va_end(ap);
	16418	+ pInfo->xOut(zMsg, pInfo->pOutArg);
	16419	+ sqlite3_free(zMsg);
	16420	+ }
	16421	+}
	16422	+
	16423	+/*
	16424	+** Try to convert an error code into a symbolic name for that error code.
	16425	+*/
	16426	+static const char *vfstrace_errcode_name(int rc ){
	16427	+ const char *zVal = 0;
	16428	+ switch( rc ){
	16429	+ case SQLITE_OK: zVal = "SQLITE_OK"; break;
	16430	+ case SQLITE_INTERNAL: zVal = "SQLITE_INTERNAL"; break;
	16431	+ case SQLITE_ERROR: zVal = "SQLITE_ERROR"; break;
	16432	+ case SQLITE_PERM: zVal = "SQLITE_PERM"; break;
	16433	+ case SQLITE_ABORT: zVal = "SQLITE_ABORT"; break;
	16434	+ case SQLITE_BUSY: zVal = "SQLITE_BUSY"; break;
	16435	+ case SQLITE_LOCKED: zVal = "SQLITE_LOCKED"; break;
	16436	+ case SQLITE_NOMEM: zVal = "SQLITE_NOMEM"; break;
	16437	+ case SQLITE_READONLY: zVal = "SQLITE_READONLY"; break;
	16438	+ case SQLITE_INTERRUPT: zVal = "SQLITE_INTERRUPT"; break;
	16439	+ case SQLITE_IOERR: zVal = "SQLITE_IOERR"; break;
	16440	+ case SQLITE_CORRUPT: zVal = "SQLITE_CORRUPT"; break;
	16441	+ case SQLITE_NOTFOUND: zVal = "SQLITE_NOTFOUND"; break;
	16442	+ case SQLITE_FULL: zVal = "SQLITE_FULL"; break;
	16443	+ case SQLITE_CANTOPEN: zVal = "SQLITE_CANTOPEN"; break;
	16444	+ case SQLITE_PROTOCOL: zVal = "SQLITE_PROTOCOL"; break;
	16445	+ case SQLITE_EMPTY: zVal = "SQLITE_EMPTY"; break;
	16446	+ case SQLITE_SCHEMA: zVal = "SQLITE_SCHEMA"; break;
	16447	+ case SQLITE_TOOBIG: zVal = "SQLITE_TOOBIG"; break;
	16448	+ case SQLITE_CONSTRAINT: zVal = "SQLITE_CONSTRAINT"; break;
	16449	+ case SQLITE_MISMATCH: zVal = "SQLITE_MISMATCH"; break;
	16450	+ case SQLITE_MISUSE: zVal = "SQLITE_MISUSE"; break;
	16451	+ case SQLITE_NOLFS: zVal = "SQLITE_NOLFS"; break;
	16452	+ case SQLITE_IOERR_READ: zVal = "SQLITE_IOERR_READ"; break;
	16453	+ case SQLITE_IOERR_SHORT_READ: zVal = "SQLITE_IOERR_SHORT_READ"; break;
	16454	+ case SQLITE_IOERR_WRITE: zVal = "SQLITE_IOERR_WRITE"; break;
	16455	+ case SQLITE_IOERR_FSYNC: zVal = "SQLITE_IOERR_FSYNC"; break;
	16456	+ case SQLITE_IOERR_DIR_FSYNC: zVal = "SQLITE_IOERR_DIR_FSYNC"; break;
	16457	+ case SQLITE_IOERR_TRUNCATE: zVal = "SQLITE_IOERR_TRUNCATE"; break;
	16458	+ case SQLITE_IOERR_FSTAT: zVal = "SQLITE_IOERR_FSTAT"; break;
	16459	+ case SQLITE_IOERR_UNLOCK: zVal = "SQLITE_IOERR_UNLOCK"; break;
	16460	+ case SQLITE_IOERR_RDLOCK: zVal = "SQLITE_IOERR_RDLOCK"; break;
	16461	+ case SQLITE_IOERR_DELETE: zVal = "SQLITE_IOERR_DELETE"; break;
	16462	+ case SQLITE_IOERR_BLOCKED: zVal = "SQLITE_IOERR_BLOCKED"; break;
	16463	+ case SQLITE_IOERR_NOMEM: zVal = "SQLITE_IOERR_NOMEM"; break;
	16464	+ case SQLITE_IOERR_ACCESS: zVal = "SQLITE_IOERR_ACCESS"; break;
	16465	+ case SQLITE_IOERR_CHECKRESERVEDLOCK:
	16466	+ zVal = "SQLITE_IOERR_CHECKRESERVEDLOCK"; break;
	16467	+ case SQLITE_IOERR_LOCK: zVal = "SQLITE_IOERR_LOCK"; break;
	16468	+ case SQLITE_IOERR_CLOSE: zVal = "SQLITE_IOERR_CLOSE"; break;
	16469	+ case SQLITE_IOERR_DIR_CLOSE: zVal = "SQLITE_IOERR_DIR_CLOSE"; break;
	16470	+ case SQLITE_IOERR_SHMOPEN: zVal = "SQLITE_IOERR_SHMOPEN"; break;
	16471	+ case SQLITE_IOERR_SHMSIZE: zVal = "SQLITE_IOERR_SHMSIZE"; break;
	16472	+ case SQLITE_IOERR_SHMLOCK: zVal = "SQLITE_IOERR_SHMLOCK"; break;
	16473	+ case SQLITE_IOERR_SHMMAP: zVal = "SQLITE_IOERR_SHMMAP"; break;
	16474	+ case SQLITE_IOERR_SEEK: zVal = "SQLITE_IOERR_SEEK"; break;
	16475	+ case SQLITE_IOERR_GETTEMPPATH: zVal = "SQLITE_IOERR_GETTEMPPATH"; break;
	16476	+ case SQLITE_IOERR_CONVPATH: zVal = "SQLITE_IOERR_CONVPATH"; break;
	16477	+ case SQLITE_READONLY_DBMOVED: zVal = "SQLITE_READONLY_DBMOVED"; break;
	16478	+ case SQLITE_LOCKED_SHAREDCACHE: zVal = "SQLITE_LOCKED_SHAREDCACHE"; break;
	16479	+ case SQLITE_BUSY_RECOVERY: zVal = "SQLITE_BUSY_RECOVERY"; break;
	16480	+ case SQLITE_CANTOPEN_NOTEMPDIR: zVal = "SQLITE_CANTOPEN_NOTEMPDIR"; break;
	16481	+ }
	16482	+ return zVal;
	16483	+}
	16484	+
	16485	+/*
	16486	+** Convert value rc into a string and print it using zFormat. zFormat
	16487	+** should have exactly one %s
	16488	+*/
	16489	+static void vfstrace_print_errcode(
	16490	+ vfstrace_info *pInfo,
	16491	+ const char *zFormat,
	16492	+ int rc
	16493	+){
	16494	+ const char *zVal;
	16495	+ char zBuf[50];
	16496	+ zVal = vfstrace_errcode_name(rc);
	16497	+ if( zVal==0 ){
	16498	+ zVal = vfstrace_errcode_name(rc&0xff);
	16499	+ if( zVal ){
	16500	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "%s \| 0x%x", zVal, rc&0xffff00);
	16501	+ }else{
	16502	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "%d (0x%x)", rc, rc);
	16503	+ }
	16504	+ zVal = zBuf;
	16505	+ }
	16506	+ vfstrace_printf(pInfo, zFormat, zVal);
	16507	+}
	16508	+
	16509	+/*
	16510	+** Append to a buffer.
	16511	+*/
	16512	+static void strappend(char z, int pI, const char *zAppend){
	16513	+ int i = *pI;
	16514	+ while( zAppend[0] ){ z[i++] = *(zAppend++); }
	16515	+ z[i] = 0;
	16516	+ *pI = i;
	16517	+}
	16518	+
	16519	+/*
	16520	+** Turn tracing output on or off according to mMask.
	16521	+*/
	16522	+static void vfstraceOnOff(vfstrace_info *pInfo, unsigned int mMask){
	16523	+ pInfo->bOn = (pInfo->mTrace & mMask)!=0;
	16524	+}
	16525	+
	16526	+/*
	16527	+** Close an vfstrace-file.
	16528	+*/
	16529	+static int vfstraceClose(sqlite3_file *pFile){
	16530	+ vfstrace_file p = (vfstrace_file )pFile;
	16531	+ vfstrace_info *pInfo = p->pInfo;
	16532	+ int rc;
	16533	+ vfstraceOnOff(pInfo, VTR_CLOSE);
	16534	+ vfstrace_printf(pInfo, "%s.xClose(%s)", pInfo->zVfsName, p->zFName);
	16535	+ rc = p->pReal->pMethods->xClose(p->pReal);
	16536	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16537	+ if( rc==SQLITE_OK ){
	16538	+ sqlite3_free((void*)p->base.pMethods);
	16539	+ p->base.pMethods = 0;
	16540	+ }
	16541	+ return rc;
	16542	+}
	16543	+
	16544	+/*
	16545	+** Read data from an vfstrace-file.
	16546	+*/
	16547	+static int vfstraceRead(
	16548	+ sqlite3_file *pFile,
	16549	+ void *zBuf,
	16550	+ int iAmt,
	16551	+ sqlite_int64 iOfst
	16552	+){
	16553	+ vfstrace_file p = (vfstrace_file )pFile;
	16554	+ vfstrace_info *pInfo = p->pInfo;
	16555	+ int rc;
	16556	+ vfstraceOnOff(pInfo, VTR_READ);
	16557	+ vfstrace_printf(pInfo, "%s.xRead(%s,n=%d,ofst=%lld)",
	16558	+ pInfo->zVfsName, p->zFName, iAmt, iOfst);
	16559	+ rc = p->pReal->pMethods->xRead(p->pReal, zBuf, iAmt, iOfst);
	16560	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16561	+ return rc;
	16562	+}
	16563	+
	16564	+/*
	16565	+** Write data to an vfstrace-file.
	16566	+*/
	16567	+static int vfstraceWrite(
	16568	+ sqlite3_file *pFile,
	16569	+ const void *zBuf,
	16570	+ int iAmt,
	16571	+ sqlite_int64 iOfst
	16572	+){
	16573	+ vfstrace_file p = (vfstrace_file )pFile;
	16574	+ vfstrace_info *pInfo = p->pInfo;
	16575	+ int rc;
	16576	+ vfstraceOnOff(pInfo, VTR_WRITE);
	16577	+ vfstrace_printf(pInfo, "%s.xWrite(%s,n=%d,ofst=%lld)",
	16578	+ pInfo->zVfsName, p->zFName, iAmt, iOfst);
	16579	+ rc = p->pReal->pMethods->xWrite(p->pReal, zBuf, iAmt, iOfst);
	16580	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16581	+ return rc;
	16582	+}
	16583	+
	16584	+/*
	16585	+** Truncate an vfstrace-file.
	16586	+*/
	16587	+static int vfstraceTruncate(sqlite3_file *pFile, sqlite_int64 size){
	16588	+ vfstrace_file p = (vfstrace_file )pFile;
	16589	+ vfstrace_info *pInfo = p->pInfo;
	16590	+ int rc;
	16591	+ vfstraceOnOff(pInfo, VTR_TRUNC);
	16592	+ vfstrace_printf(pInfo, "%s.xTruncate(%s,%lld)", pInfo->zVfsName, p->zFName,
	16593	+ size);
	16594	+ rc = p->pReal->pMethods->xTruncate(p->pReal, size);
	16595	+ vfstrace_printf(pInfo, " -> %d\n", rc);
	16596	+ return rc;
	16597	+}
	16598	+
	16599	+/*
	16600	+** Sync an vfstrace-file.
	16601	+*/
	16602	+static int vfstraceSync(sqlite3_file *pFile, int flags){
	16603	+ vfstrace_file p = (vfstrace_file )pFile;
	16604	+ vfstrace_info *pInfo = p->pInfo;
	16605	+ int rc;
	16606	+ int i;
	16607	+ char zBuf[100];
	16608	+ memcpy(zBuf, "\|0", 3);
	16609	+ i = 0;
	16610	+ if( flags & SQLITE_SYNC_FULL ) strappend(zBuf, &i, "\|FULL");
	16611	+ else if( flags & SQLITE_SYNC_NORMAL ) strappend(zBuf, &i, "\|NORMAL");
	16612	+ if( flags & SQLITE_SYNC_DATAONLY ) strappend(zBuf, &i, "\|DATAONLY");
	16613	+ if( flags & ~(SQLITE_SYNC_FULL\|SQLITE_SYNC_DATAONLY) ){
	16614	+ sqlite3_snprintf(sizeof(zBuf)-i, &zBuf[i], "\|0x%x", flags);
	16615	+ }
	16616	+ vfstraceOnOff(pInfo, VTR_SYNC);
	16617	+ vfstrace_printf(pInfo, "%s.xSync(%s,%s)", pInfo->zVfsName, p->zFName,
	16618	+ &zBuf[1]);
	16619	+ rc = p->pReal->pMethods->xSync(p->pReal, flags);
	16620	+ vfstrace_printf(pInfo, " -> %d\n", rc);
	16621	+ return rc;
	16622	+}
	16623	+
	16624	+/*
	16625	+** Return the current file-size of an vfstrace-file.
	16626	+*/
	16627	+static int vfstraceFileSize(sqlite3_file pFile, sqlite_int64 pSize){
	16628	+ vfstrace_file p = (vfstrace_file )pFile;
	16629	+ vfstrace_info *pInfo = p->pInfo;
	16630	+ int rc;
	16631	+ vfstraceOnOff(pInfo, VTR_FSIZE);
	16632	+ vfstrace_printf(pInfo, "%s.xFileSize(%s)", pInfo->zVfsName, p->zFName);
	16633	+ rc = p->pReal->pMethods->xFileSize(p->pReal, pSize);
	16634	+ vfstrace_print_errcode(pInfo, " -> %s,", rc);
	16635	+ vfstrace_printf(pInfo, " size=%lld\n", *pSize);
	16636	+ return rc;
	16637	+}
	16638	+
	16639	+/*
	16640	+** Return the name of a lock.
	16641	+*/
	16642	+static const char *lockName(int eLock){
	16643	+ const char *azLockNames[] = {
	16644	+ "NONE", "SHARED", "RESERVED", "PENDING", "EXCLUSIVE"
	16645	+ };
	16646	+ if( eLock<0 \|\| eLock>=(int)(sizeof(azLockNames)/sizeof(azLockNames[0])) ){
	16647	+ return "???";
	16648	+ }else{
	16649	+ return azLockNames[eLock];
	16650	+ }
	16651	+}
	16652	+
	16653	+/*
	16654	+** Lock an vfstrace-file.
	16655	+*/
	16656	+static int vfstraceLock(sqlite3_file *pFile, int eLock){
	16657	+ vfstrace_file p = (vfstrace_file )pFile;
	16658	+ vfstrace_info *pInfo = p->pInfo;
	16659	+ int rc;
	16660	+ vfstraceOnOff(pInfo, VTR_LOCK);
	16661	+ vfstrace_printf(pInfo, "%s.xLock(%s,%s)", pInfo->zVfsName, p->zFName,
	16662	+ lockName(eLock));
	16663	+ rc = p->pReal->pMethods->xLock(p->pReal, eLock);
	16664	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16665	+ return rc;
	16666	+}
	16667	+
	16668	+/*
	16669	+** Unlock an vfstrace-file.
	16670	+*/
	16671	+static int vfstraceUnlock(sqlite3_file *pFile, int eLock){
	16672	+ vfstrace_file p = (vfstrace_file )pFile;
	16673	+ vfstrace_info *pInfo = p->pInfo;
	16674	+ int rc;
	16675	+ vfstraceOnOff(pInfo, VTR_UNLOCK);
	16676	+ vfstrace_printf(pInfo, "%s.xUnlock(%s,%s)", pInfo->zVfsName, p->zFName,
	16677	+ lockName(eLock));
	16678	+ rc = p->pReal->pMethods->xUnlock(p->pReal, eLock);
	16679	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16680	+ return rc;
	16681	+}
	16682	+
	16683	+/*
	16684	+** Check if another file-handle holds a RESERVED lock on an vfstrace-file.
	16685	+*/
	16686	+static int vfstraceCheckReservedLock(sqlite3_file pFile, int pResOut){
	16687	+ vfstrace_file p = (vfstrace_file )pFile;
	16688	+ vfstrace_info *pInfo = p->pInfo;
	16689	+ int rc;
	16690	+ vfstraceOnOff(pInfo, VTR_CRL);
	16691	+ vfstrace_printf(pInfo, "%s.xCheckReservedLock(%s,%d)",
	16692	+ pInfo->zVfsName, p->zFName);
	16693	+ rc = p->pReal->pMethods->xCheckReservedLock(p->pReal, pResOut);
	16694	+ vfstrace_print_errcode(pInfo, " -> %s", rc);
	16695	+ vfstrace_printf(pInfo, ", out=%d\n", *pResOut);
	16696	+ return rc;
	16697	+}
	16698	+
	16699	+/*
	16700	+** File control method. For custom operations on an vfstrace-file.
	16701	+*/
	16702	+static int vfstraceFileControl(sqlite3_file pFile, int op, void pArg){
	16703	+ vfstrace_file p = (vfstrace_file )pFile;
	16704	+ vfstrace_info *pInfo = p->pInfo;
	16705	+ int rc;
	16706	+ char zBuf[100];
	16707	+ char zBuf2[100];
	16708	+ char *zOp;
	16709	+ char *zRVal = 0;
	16710	+ vfstraceOnOff(pInfo, VTR_FCTRL);
	16711	+ switch( op ){
	16712	+ case SQLITE_FCNTL_LOCKSTATE: zOp = "LOCKSTATE"; break;
	16713	+ case SQLITE_GET_LOCKPROXYFILE: zOp = "GET_LOCKPROXYFILE"; break;
	16714	+ case SQLITE_SET_LOCKPROXYFILE: zOp = "SET_LOCKPROXYFILE"; break;
	16715	+ case SQLITE_LAST_ERRNO: zOp = "LAST_ERRNO"; break;
	16716	+ case SQLITE_FCNTL_SIZE_HINT: {
	16717	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "SIZE_HINT,%lld",
	16718	+ (sqlite3_int64)pArg);
	16719	+ zOp = zBuf;
	16720	+ break;
	16721	+ }
	16722	+ case SQLITE_FCNTL_CHUNK_SIZE: {
	16723	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "CHUNK_SIZE,%d", (int)pArg);
	16724	+ zOp = zBuf;
	16725	+ break;
	16726	+ }
	16727	+ case SQLITE_FCNTL_FILE_POINTER: zOp = "FILE_POINTER"; break;
	16728	+ case SQLITE_FCNTL_WIN32_AV_RETRY: zOp = "WIN32_AV_RETRY"; break;
	16729	+ case SQLITE_FCNTL_PERSIST_WAL: {
	16730	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "PERSIST_WAL,%d", (int)pArg);
	16731	+ zOp = zBuf;
	16732	+ break;
	16733	+ }
	16734	+ case SQLITE_FCNTL_OVERWRITE: zOp = "OVERWRITE"; break;
	16735	+ case SQLITE_FCNTL_VFSNAME: zOp = "VFSNAME"; break;
	16736	+ case SQLITE_FCNTL_POWERSAFE_OVERWRITE: zOp = "POWERSAFE_OVERWRITE"; break;
	16737	+ case SQLITE_FCNTL_PRAGMA: {
	16738	+ const char const a = (const charconst)pArg;
	16739	+ if( a[1] && strcmp(a[1],"vfstrace")==0 && a[2] ){
	16740	+ const u8 zArg = (const u8)a[2];
	16741	+ if( zArg[0]>='0' && zArg[0]<=9 ){
	16742	+ pInfo->mTrace = (sqlite3_uint64)strtoll(a[2], 0, 0);
	16743	+ }else{
	16744	+ static const struct {
	16745	+ const char *z;
	16746	+ unsigned int m;
	16747	+ } aKw[] = {
	16748	+ { "all", 0xffffffff },
	16749	+ { "close", VTR_CLOSE },
	16750	+ { "read", VTR_READ },
	16751	+ { "write", VTR_WRITE },
	16752	+ { "truncate", VTR_TRUNC },
	16753	+ { "sync", VTR_SYNC },
	16754	+ { "filesize", VTR_FSIZE },
	16755	+ { "lock", VTR_LOCK },
	16756	+ { "unlock", VTR_UNLOCK },
	16757	+ { "checkreservedlock", VTR_CRL },
	16758	+ { "filecontrol", VTR_FCTRL },
	16759	+ { "sectorsize", VTR_SECSZ },
	16760	+ { "devicecharacteristics", VTR_DEVCHAR },
	16761	+ { "shmlock", VTR_SHMLOCK },
	16762	+ { "shmmap", VTR_SHMMAP },
	16763	+ { "shmummap", VTR_SHMUNMAP },
	16764	+ { "shmbarrier", VTR_SHMBAR },
	16765	+ { "open", VTR_OPEN },
	16766	+ { "delete", VTR_DELETE },
	16767	+ { "access", VTR_ACCESS },
	16768	+ { "fullpathname", VTR_FULLPATH },
	16769	+ { "dlopen", VTR_DLOPEN },
	16770	+ { "dlerror", VTR_DLERR },
	16771	+ { "dlsym", VTR_DLSYM },
	16772	+ { "dlclose", VTR_DLCLOSE },
	16773	+ { "randomness", VTR_RAND },
	16774	+ { "sleep", VTR_SLEEP },
	16775	+ { "currenttime", VTR_CURTIME },
	16776	+ { "currenttimeint64", VTR_CURTIME },
	16777	+ { "getlasterror", VTR_LASTERR },
	16778	+ };
	16779	+ int onOff = 1;
	16780	+ while( zArg[0] ){
	16781	+ int jj, n;
	16782	+ while( zArg[0]!=0 && zArg[0]!='-' && zArg[0]!='+'
	16783	+ && !isalpha(zArg[0]) ) zArg++;
	16784	+ if( zArg[0]==0 ) break;
	16785	+ if( zArg[0]=='-' ){
	16786	+ onOff = 0;
	16787	+ zArg++;
	16788	+ }else if( zArg[0]=='+' ){
	16789	+ onOff = 1;
	16790	+ zArg++;
	16791	+ }
	16792	+ while( !isalpha(zArg[0]) ){
	16793	+ if( zArg[0]==0 ) break;
	16794	+ zArg++;
	16795	+ }
	16796	+ if( zArg[0]=='x' && isalpha(zArg[1]) ) zArg++;
	16797	+ for(n=0; isalpha(zArg[n]); n++){}
	16798	+ for(jj=0; jj<(int)(sizeof(aKw)/sizeof(aKw[0])); jj++){
	16799	+ if( sqlite3_strnicmp(aKw[jj].z,(const char*)zArg,n)==0 ){
	16800	+ if( onOff ){
	16801	+ pInfo->mTrace \|= aKw[jj].m;
	16802	+ }else{
	16803	+ pInfo->mTrace &= ~aKw[jj].m;
	16804	+ }
	16805	+ break;
	16806	+ }
	16807	+ }
	16808	+ zArg += n;
	16809	+ }
	16810	+ }
	16811	+ }
	16812	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "PRAGMA,[%s,%s]",a[1],a[2]);
	16813	+ zOp = zBuf;
	16814	+ break;
	16815	+ }
	16816	+ case SQLITE_FCNTL_BUSYHANDLER: zOp = "BUSYHANDLER"; break;
	16817	+ case SQLITE_FCNTL_TEMPFILENAME: zOp = "TEMPFILENAME"; break;
	16818	+ case SQLITE_FCNTL_MMAP_SIZE: {
	16819	+ sqlite3_int64 iMMap = (sqlite3_int64)pArg;
	16820	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "MMAP_SIZE,%lld",iMMap);
	16821	+ zOp = zBuf;
	16822	+ break;
	16823	+ }
	16824	+ case SQLITE_FCNTL_TRACE: zOp = "TRACE"; break;
	16825	+ case SQLITE_FCNTL_HAS_MOVED: zOp = "HAS_MOVED"; break;
	16826	+ case SQLITE_FCNTL_SYNC: zOp = "SYNC"; break;
	16827	+ case SQLITE_FCNTL_COMMIT_PHASETWO: zOp = "COMMIT_PHASETWO"; break;
	16828	+ case SQLITE_FCNTL_WIN32_SET_HANDLE: zOp = "WIN32_SET_HANDLE"; break;
	16829	+ case SQLITE_FCNTL_WAL_BLOCK: zOp = "WAL_BLOCK"; break;
	16830	+ case SQLITE_FCNTL_ZIPVFS: zOp = "ZIPVFS"; break;
	16831	+ case SQLITE_FCNTL_RBU: zOp = "RBU"; break;
	16832	+ case SQLITE_FCNTL_VFS_POINTER: zOp = "VFS_POINTER"; break;
	16833	+ case SQLITE_FCNTL_JOURNAL_POINTER: zOp = "JOURNAL_POINTER"; break;
	16834	+ case SQLITE_FCNTL_WIN32_GET_HANDLE: zOp = "WIN32_GET_HANDLE"; break;
	16835	+ case SQLITE_FCNTL_PDB: zOp = "PDB"; break;
	16836	+ case SQLITE_FCNTL_BEGIN_ATOMIC_WRITE: zOp = "BEGIN_ATOMIC_WRITE"; break;
	16837	+ case SQLITE_FCNTL_COMMIT_ATOMIC_WRITE: zOp = "COMMIT_ATOMIC_WRITE"; break;
	16838	+ case SQLITE_FCNTL_ROLLBACK_ATOMIC_WRITE: {
	16839	+ zOp = "ROLLBACK_ATOMIC_WRITE";
	16840	+ break;
	16841	+ }
	16842	+ case SQLITE_FCNTL_LOCK_TIMEOUT: {
	16843	+ sqlite3_snprintf(sizeof(zBuf), zBuf, "LOCK_TIMEOUT,%d", (int)pArg);
	16844	+ zOp = zBuf;
	16845	+ break;
	16846	+ }
	16847	+ case SQLITE_FCNTL_DATA_VERSION: zOp = "DATA_VERSION"; break;
	16848	+ case SQLITE_FCNTL_SIZE_LIMIT: zOp = "SIZE_LIMIT"; break;
	16849	+ case SQLITE_FCNTL_CKPT_DONE: zOp = "CKPT_DONE"; break;
	16850	+ case SQLITE_FCNTL_RESERVE_BYTES: zOp = "RESERVED_BYTES"; break;
	16851	+ case SQLITE_FCNTL_CKPT_START: zOp = "CKPT_START"; break;
	16852	+ case SQLITE_FCNTL_EXTERNAL_READER: zOp = "EXTERNAL_READER"; break;
	16853	+ case SQLITE_FCNTL_CKSM_FILE: zOp = "CKSM_FILE"; break;
	16854	+ case SQLITE_FCNTL_RESET_CACHE: zOp = "RESET_CACHE"; break;
	16855	+ case 0xca093fa0: zOp = "DB_UNCHANGED"; break;
	16856	+ default: {
	16857	+ sqlite3_snprintf(sizeof zBuf, zBuf, "%d", op);
	16858	+ zOp = zBuf;
	16859	+ break;
	16860	+ }
	16861	+ }
	16862	+ vfstrace_printf(pInfo, "%s.xFileControl(%s,%s)",
	16863	+ pInfo->zVfsName, p->zFName, zOp);
	16864	+ rc = p->pReal->pMethods->xFileControl(p->pReal, op, pArg);
	16865	+ if( rc==SQLITE_OK ){
	16866	+ switch( op ){
	16867	+ case SQLITE_FCNTL_VFSNAME: {
	16868	+ (char*)pArg = sqlite3_mprintf("vfstrace.%s/%z",
	16869	+ pInfo->zVfsName, (char*)pArg);
	16870	+ zRVal = (char*)pArg;
	16871	+ break;
	16872	+ }
	16873	+ case SQLITE_FCNTL_MMAP_SIZE: {
	16874	+ sqlite3_snprintf(sizeof(zBuf2), zBuf2, "%lld", (sqlite3_int64)pArg);
	16875	+ zRVal = zBuf2;
	16876	+ break;
	16877	+ }
	16878	+ case SQLITE_FCNTL_HAS_MOVED:
	16879	+ case SQLITE_FCNTL_PERSIST_WAL: {
	16880	+ sqlite3_snprintf(sizeof(zBuf2), zBuf2, "%d", (int)pArg);
	16881	+ zRVal = zBuf2;
	16882	+ break;
	16883	+ }
	16884	+ case SQLITE_FCNTL_PRAGMA:
	16885	+ case SQLITE_FCNTL_TEMPFILENAME: {
	16886	+ zRVal = (char*)pArg;
	16887	+ break;
	16888	+ }
	16889	+ }
	16890	+ }
	16891	+ if( zRVal ){
	16892	+ vfstrace_print_errcode(pInfo, " -> %s", rc);
	16893	+ vfstrace_printf(pInfo, ", %s\n", zRVal);
	16894	+ }else{
	16895	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16896	+ }
	16897	+ return rc;
	16898	+}
	16899	+
	16900	+/*
	16901	+** Return the sector-size in bytes for an vfstrace-file.
	16902	+*/
	16903	+static int vfstraceSectorSize(sqlite3_file *pFile){
	16904	+ vfstrace_file p = (vfstrace_file )pFile;
	16905	+ vfstrace_info *pInfo = p->pInfo;
	16906	+ int rc;
	16907	+ vfstraceOnOff(pInfo, VTR_SECSZ);
	16908	+ vfstrace_printf(pInfo, "%s.xSectorSize(%s)", pInfo->zVfsName, p->zFName);
	16909	+ rc = p->pReal->pMethods->xSectorSize(p->pReal);
	16910	+ vfstrace_printf(pInfo, " -> %d\n", rc);
	16911	+ return rc;
	16912	+}
	16913	+
	16914	+/*
	16915	+** Return the device characteristic flags supported by an vfstrace-file.
	16916	+*/
	16917	+static int vfstraceDeviceCharacteristics(sqlite3_file *pFile){
	16918	+ vfstrace_file p = (vfstrace_file )pFile;
	16919	+ vfstrace_info *pInfo = p->pInfo;
	16920	+ int rc;
	16921	+ vfstraceOnOff(pInfo, VTR_DEVCHAR);
	16922	+ vfstrace_printf(pInfo, "%s.xDeviceCharacteristics(%s)",
	16923	+ pInfo->zVfsName, p->zFName);
	16924	+ rc = p->pReal->pMethods->xDeviceCharacteristics(p->pReal);
	16925	+ vfstrace_printf(pInfo, " -> 0x%08x\n", rc);
	16926	+ return rc;
	16927	+}
	16928	+
	16929	+/*
	16930	+** Shared-memory operations.
	16931	+*/
	16932	+static int vfstraceShmLock(sqlite3_file *pFile, int ofst, int n, int flags){
	16933	+ static const char *azLockName[] = {
	16934	+ "WRITE",
	16935	+ "CKPT",
	16936	+ "RECOVER",
	16937	+ "READ0",
	16938	+ "READ1",
	16939	+ "READ2",
	16940	+ "READ3",
	16941	+ "READ4",
	16942	+ };
	16943	+ vfstrace_file p = (vfstrace_file )pFile;
	16944	+ vfstrace_info *pInfo = p->pInfo;
	16945	+ int rc;
	16946	+ char zLck[100];
	16947	+ int i = 0;
	16948	+ vfstraceOnOff(pInfo, VTR_SHMLOCK);
	16949	+ memcpy(zLck, "\|0", 3);
	16950	+ if( flags & SQLITE_SHM_UNLOCK ) strappend(zLck, &i, "\|UNLOCK");
	16951	+ if( flags & SQLITE_SHM_LOCK ) strappend(zLck, &i, "\|LOCK");
	16952	+ if( flags & SQLITE_SHM_SHARED ) strappend(zLck, &i, "\|SHARED");
	16953	+ if( flags & SQLITE_SHM_EXCLUSIVE ) strappend(zLck, &i, "\|EXCLUSIVE");
	16954	+ if( flags & ~(0xf) ){
	16955	+ sqlite3_snprintf(sizeof(zLck)-i, &zLck[i], "\|0x%x", flags);
	16956	+ }
	16957	+ if( ofst>=0 && ofst<(int)(sizeof(azLockName)/sizeof(azLockName[0])) ){
	16958	+ vfstrace_printf(pInfo, "%s.xShmLock(%s,ofst=%d(%s),n=%d,%s)",
	16959	+ pInfo->zVfsName, p->zFName, ofst, azLockName[ofst],
	16960	+ n, &zLck[1]);
	16961	+ }else{
	16962	+ vfstrace_printf(pInfo, "%s.xShmLock(%s,ofst=5d,n=%d,%s)",
	16963	+ pInfo->zVfsName, p->zFName, ofst,
	16964	+ n, &zLck[1]);
	16965	+ }
	16966	+ rc = p->pReal->pMethods->xShmLock(p->pReal, ofst, n, flags);
	16967	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16968	+ return rc;
	16969	+}
	16970	+static int vfstraceShmMap(
	16971	+ sqlite3_file *pFile,
	16972	+ int iRegion,
	16973	+ int szRegion,
	16974	+ int isWrite,
	16975	+ void volatile **pp
	16976	+){
	16977	+ vfstrace_file p = (vfstrace_file )pFile;
	16978	+ vfstrace_info *pInfo = p->pInfo;
	16979	+ int rc;
	16980	+ vfstraceOnOff(pInfo, VTR_SHMMAP);
	16981	+ vfstrace_printf(pInfo, "%s.xShmMap(%s,iRegion=%d,szRegion=%d,isWrite=%d,*)",
	16982	+ pInfo->zVfsName, p->zFName, iRegion, szRegion, isWrite);
	16983	+ rc = p->pReal->pMethods->xShmMap(p->pReal, iRegion, szRegion, isWrite, pp);
	16984	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	16985	+ return rc;
	16986	+}
	16987	+static void vfstraceShmBarrier(sqlite3_file *pFile){
	16988	+ vfstrace_file p = (vfstrace_file )pFile;
	16989	+ vfstrace_info *pInfo = p->pInfo;
	16990	+ vfstraceOnOff(pInfo, VTR_SHMBAR);
	16991	+ vfstrace_printf(pInfo, "%s.xShmBarrier(%s)\n", pInfo->zVfsName, p->zFName);
	16992	+ p->pReal->pMethods->xShmBarrier(p->pReal);
	16993	+}
	16994	+static int vfstraceShmUnmap(sqlite3_file *pFile, int delFlag){
	16995	+ vfstrace_file p = (vfstrace_file )pFile;
	16996	+ vfstrace_info *pInfo = p->pInfo;
	16997	+ int rc;
	16998	+ vfstraceOnOff(pInfo, VTR_SHMUNMAP);
	16999	+ vfstrace_printf(pInfo, "%s.xShmUnmap(%s,delFlag=%d)",
	17000	+ pInfo->zVfsName, p->zFName, delFlag);
	17001	+ rc = p->pReal->pMethods->xShmUnmap(p->pReal, delFlag);
	17002	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	17003	+ return rc;
	17004	+}
	17005	+
	17006	+
	17007	+
	17008	+/*
	17009	+** Open an vfstrace file handle.
	17010	+*/
	17011	+static int vfstraceOpen(
	17012	+ sqlite3_vfs *pVfs,
	17013	+ const char *zName,
	17014	+ sqlite3_file *pFile,
	17015	+ int flags,
	17016	+ int *pOutFlags
	17017	+){
	17018	+ int rc;
	17019	+ vfstrace_file p = (vfstrace_file )pFile;
	17020	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17021	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17022	+ p->pInfo = pInfo;
	17023	+ p->zFName = zName ? fileTail(zName) : "<temp>";
	17024	+ p->pReal = (sqlite3_file *)&p[1];
	17025	+ rc = pRoot->xOpen(pRoot, zName, p->pReal, flags, pOutFlags);
	17026	+ vfstraceOnOff(pInfo, VTR_OPEN);
	17027	+ vfstrace_printf(pInfo, "%s.xOpen(%s,flags=0x%x)",
	17028	+ pInfo->zVfsName, p->zFName, flags);
	17029	+ if( p->pReal->pMethods ){
	17030	+ sqlite3_io_methods pNew = sqlite3_malloc( sizeof(pNew) );
	17031	+ const sqlite3_io_methods *pSub = p->pReal->pMethods;
	17032	+ memset(pNew, 0, sizeof(*pNew));
	17033	+ pNew->iVersion = pSub->iVersion;
	17034	+ pNew->xClose = vfstraceClose;
	17035	+ pNew->xRead = vfstraceRead;
	17036	+ pNew->xWrite = vfstraceWrite;
	17037	+ pNew->xTruncate = vfstraceTruncate;
	17038	+ pNew->xSync = vfstraceSync;
	17039	+ pNew->xFileSize = vfstraceFileSize;
	17040	+ pNew->xLock = vfstraceLock;
	17041	+ pNew->xUnlock = vfstraceUnlock;
	17042	+ pNew->xCheckReservedLock = vfstraceCheckReservedLock;
	17043	+ pNew->xFileControl = vfstraceFileControl;
	17044	+ pNew->xSectorSize = vfstraceSectorSize;
	17045	+ pNew->xDeviceCharacteristics = vfstraceDeviceCharacteristics;
	17046	+ if( pNew->iVersion>=2 ){
	17047	+ pNew->xShmMap = pSub->xShmMap ? vfstraceShmMap : 0;
	17048	+ pNew->xShmLock = pSub->xShmLock ? vfstraceShmLock : 0;
	17049	+ pNew->xShmBarrier = pSub->xShmBarrier ? vfstraceShmBarrier : 0;
	17050	+ pNew->xShmUnmap = pSub->xShmUnmap ? vfstraceShmUnmap : 0;
	17051	+ }
	17052	+ pFile->pMethods = pNew;
	17053	+ }
	17054	+ vfstrace_print_errcode(pInfo, " -> %s", rc);
	17055	+ if( pOutFlags ){
	17056	+ vfstrace_printf(pInfo, ", outFlags=0x%x\n", *pOutFlags);
	17057	+ }else{
	17058	+ vfstrace_printf(pInfo, "\n");
	17059	+ }
	17060	+ return rc;
	17061	+}
	17062	+
	17063	+/*
	17064	+** Delete the file located at zPath. If the dirSync argument is true,
	17065	+** ensure the file-system modifications are synced to disk before
	17066	+** returning.
	17067	+*/
	17068	+static int vfstraceDelete(sqlite3_vfs pVfs, const char zPath, int dirSync){
	17069	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17070	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17071	+ int rc;
	17072	+ vfstraceOnOff(pInfo, VTR_DELETE);
	17073	+ vfstrace_printf(pInfo, "%s.xDelete(\"%s\",%d)",
	17074	+ pInfo->zVfsName, zPath, dirSync);
	17075	+ rc = pRoot->xDelete(pRoot, zPath, dirSync);
	17076	+ vfstrace_print_errcode(pInfo, " -> %s\n", rc);
	17077	+ return rc;
	17078	+}
	17079	+
	17080	+/*
	17081	+** Test for access permissions. Return true if the requested permission
	17082	+** is available, or false otherwise.
	17083	+*/
	17084	+static int vfstraceAccess(
	17085	+ sqlite3_vfs *pVfs,
	17086	+ const char *zPath,
	17087	+ int flags,
	17088	+ int *pResOut
	17089	+){
	17090	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17091	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17092	+ int rc;
	17093	+ vfstraceOnOff(pInfo, VTR_ACCESS);
	17094	+ vfstrace_printf(pInfo, "%s.xAccess(\"%s\",%d)",
	17095	+ pInfo->zVfsName, zPath, flags);
	17096	+ rc = pRoot->xAccess(pRoot, zPath, flags, pResOut);
	17097	+ vfstrace_print_errcode(pInfo, " -> %s", rc);
	17098	+ vfstrace_printf(pInfo, ", out=%d\n", *pResOut);
	17099	+ return rc;
	17100	+}
	17101	+
	17102	+/*
	17103	+** Populate buffer zOut with the full canonical pathname corresponding
	17104	+** to the pathname in zPath. zOut is guaranteed to point to a buffer
	17105	+** of at least (DEVSYM_MAX_PATHNAME+1) bytes.
	17106	+*/
	17107	+static int vfstraceFullPathname(
	17108	+ sqlite3_vfs *pVfs,
	17109	+ const char *zPath,
	17110	+ int nOut,
	17111	+ char *zOut
	17112	+){
	17113	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17114	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17115	+ int rc;
	17116	+ vfstraceOnOff(pInfo, VTR_FULLPATH);
	17117	+ vfstrace_printf(pInfo, "%s.xFullPathname(\"%s\")",
	17118	+ pInfo->zVfsName, zPath);
	17119	+ rc = pRoot->xFullPathname(pRoot, zPath, nOut, zOut);
	17120	+ vfstrace_print_errcode(pInfo, " -> %s", rc);
	17121	+ vfstrace_printf(pInfo, ", out=\"%.*s\"\n", nOut, zOut);
	17122	+ return rc;
	17123	+}
	17124	+
	17125	+/*
	17126	+** Open the dynamic library located at zPath and return a handle.
	17127	+*/
	17128	+static void vfstraceDlOpen(sqlite3_vfs pVfs, const char *zPath){
	17129	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17130	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17131	+ vfstraceOnOff(pInfo, VTR_DLOPEN);
	17132	+ vfstrace_printf(pInfo, "%s.xDlOpen(\"%s\")\n", pInfo->zVfsName, zPath);
	17133	+ return pRoot->xDlOpen(pRoot, zPath);
	17134	+}
	17135	+
	17136	+/*
	17137	+** Populate the buffer zErrMsg (size nByte bytes) with a human readable
	17138	+** utf-8 string describing the most recent error encountered associated
	17139	+** with dynamic libraries.
	17140	+*/
	17141	+static void vfstraceDlError(sqlite3_vfs pVfs, int nByte, char zErrMsg){
	17142	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17143	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17144	+ vfstraceOnOff(pInfo, VTR_DLERR);
	17145	+ vfstrace_printf(pInfo, "%s.xDlError(%d)", pInfo->zVfsName, nByte);
	17146	+ pRoot->xDlError(pRoot, nByte, zErrMsg);
	17147	+ vfstrace_printf(pInfo, " -> \"%s\"", zErrMsg);
	17148	+}
	17149	+
	17150	+/*
	17151	+** Return a pointer to the symbol zSymbol in the dynamic library pHandle.
	17152	+*/
	17153	+static void (vfstraceDlSym(sqlite3_vfs pVfs,void p,const char zSym))(void){
	17154	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17155	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17156	+ vfstrace_printf(pInfo, "%s.xDlSym(\"%s\")\n", pInfo->zVfsName, zSym);
	17157	+ return pRoot->xDlSym(pRoot, p, zSym);
	17158	+}
	17159	+
	17160	+/*
	17161	+** Close the dynamic library handle pHandle.
	17162	+*/
	17163	+static void vfstraceDlClose(sqlite3_vfs pVfs, void pHandle){
	17164	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17165	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17166	+ vfstraceOnOff(pInfo, VTR_DLCLOSE);
	17167	+ vfstrace_printf(pInfo, "%s.xDlOpen()\n", pInfo->zVfsName);
	17168	+ pRoot->xDlClose(pRoot, pHandle);
	17169	+}
	17170	+
	17171	+/*
	17172	+** Populate the buffer pointed to by zBufOut with nByte bytes of
	17173	+** random data.
	17174	+*/
	17175	+static int vfstraceRandomness(sqlite3_vfs pVfs, int nByte, char zBufOut){
	17176	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17177	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17178	+ vfstraceOnOff(pInfo, VTR_RAND);
	17179	+ vfstrace_printf(pInfo, "%s.xRandomness(%d)\n", pInfo->zVfsName, nByte);
	17180	+ return pRoot->xRandomness(pRoot, nByte, zBufOut);
	17181	+}
	17182	+
	17183	+/*
	17184	+** Sleep for nMicro microseconds. Return the number of microseconds
	17185	+** actually slept.
	17186	+*/
	17187	+static int vfstraceSleep(sqlite3_vfs *pVfs, int nMicro){
	17188	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17189	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17190	+ vfstraceOnOff(pInfo, VTR_SLEEP);
	17191	+ vfstrace_printf(pInfo, "%s.xSleep(%d)\n", pInfo->zVfsName, nMicro);
	17192	+ return pRoot->xSleep(pRoot, nMicro);
	17193	+}
	17194	+
	17195	+/*
	17196	+** Return the current time as a Julian Day number in *pTimeOut.
	17197	+*/
	17198	+static int vfstraceCurrentTime(sqlite3_vfs pVfs, double pTimeOut){
	17199	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17200	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17201	+ int rc;
	17202	+ vfstraceOnOff(pInfo, VTR_CURTIME);
	17203	+ vfstrace_printf(pInfo, "%s.xCurrentTime()", pInfo->zVfsName);
	17204	+ rc = pRoot->xCurrentTime(pRoot, pTimeOut);
	17205	+ vfstrace_printf(pInfo, " -> %.17g\n", *pTimeOut);
	17206	+ return rc;
	17207	+}
	17208	+static int vfstraceCurrentTimeInt64(sqlite3_vfs pVfs, sqlite3_int64 pTimeOut){
	17209	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17210	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17211	+ int rc;
	17212	+ vfstraceOnOff(pInfo, VTR_CURTIME);
	17213	+ vfstrace_printf(pInfo, "%s.xCurrentTimeInt64()", pInfo->zVfsName);
	17214	+ rc = pRoot->xCurrentTimeInt64(pRoot, pTimeOut);
	17215	+ vfstrace_printf(pInfo, " -> %lld\n", *pTimeOut);
	17216	+ return rc;
	17217	+}
	17218	+
	17219	+/*
	17220	+** Return the most recent error code and message
	17221	+*/
	17222	+static int vfstraceGetLastError(sqlite3_vfs pVfs, int nErr, char zErr){
	17223	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17224	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17225	+ int rc;
	17226	+ vfstraceOnOff(pInfo, VTR_LASTERR);
	17227	+ vfstrace_printf(pInfo, "%s.xGetLastError(%d,zBuf)", pInfo->zVfsName, nErr);
	17228	+ if( nErr ) zErr[0] = 0;
	17229	+ rc = pRoot->xGetLastError(pRoot, nErr, zErr);
	17230	+ vfstrace_printf(pInfo, " -> zBuf[] = \"%s\", rc = %d\n", nErr?zErr:"", rc);
	17231	+ return rc;
	17232	+}
	17233	+
	17234	+/*
	17235	+** Override system calls.
	17236	+*/
	17237	+static int vfstraceSetSystemCall(
	17238	+ sqlite3_vfs *pVfs,
	17239	+ const char *zName,
	17240	+ sqlite3_syscall_ptr pFunc
	17241	+){
	17242	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17243	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17244	+ return pRoot->xSetSystemCall(pRoot, zName, pFunc);
	17245	+}
	17246	+static sqlite3_syscall_ptr vfstraceGetSystemCall(
	17247	+ sqlite3_vfs *pVfs,
	17248	+ const char *zName
	17249	+){
	17250	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17251	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17252	+ return pRoot->xGetSystemCall(pRoot, zName);
	17253	+}
	17254	+static const char vfstraceNextSystemCall(sqlite3_vfs pVfs, const char *zName){
	17255	+ vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
	17256	+ sqlite3_vfs *pRoot = pInfo->pRootVfs;
	17257	+ return pRoot->xNextSystemCall(pRoot, zName);
	17258	+}
	17259	+
	17260	+
	17261	+/*
	17262	+** Clients invoke this routine to construct a new trace-vfs shim.
	17263	+**
	17264	+** Return SQLITE_OK on success.
	17265	+**
	17266	+** SQLITE_NOMEM is returned in the case of a memory allocation error.
	17267	+** SQLITE_NOTFOUND is returned if zOldVfsName does not exist.
	17268	+*/
	17269	+int vfstrace_register(
	17270	+ const char zTraceName, / Name of the newly constructed VFS */
	17271	+ const char zOldVfsName, / Name of the underlying VFS */
	17272	+ int (xOut)(const char,void), / Output routine. ex: fputs */
	17273	+ void pOutArg, / 2nd argument to xOut. ex: stderr */
	17274	+ int makeDefault /* True to make the new VFS the default */
	17275	+){
	17276	+ sqlite3_vfs *pNew;
	17277	+ sqlite3_vfs *pRoot;
	17278	+ vfstrace_info *pInfo;
	17279	+ size_t nName;
	17280	+ size_t nByte;
	17281	+
	17282	+ pRoot = sqlite3_vfs_find(zOldVfsName);
	17283	+ if( pRoot==0 ) return SQLITE_NOTFOUND;
	17284	+ nName = strlen(zTraceName);
	17285	+ nByte = sizeof(pNew) + sizeof(pInfo) + nName + 1;
	17286	+ pNew = sqlite3_malloc64( nByte );
	17287	+ if( pNew==0 ) return SQLITE_NOMEM;
	17288	+ memset(pNew, 0, nByte);
	17289	+ pInfo = (vfstrace_info*)&pNew[1];
	17290	+ pNew->iVersion = pRoot->iVersion;
	17291	+ pNew->szOsFile = pRoot->szOsFile + sizeof(vfstrace_file);
	17292	+ pNew->mxPathname = pRoot->mxPathname;
	17293	+ pNew->zName = (char*)&pInfo[1];
	17294	+ memcpy((char*)&pInfo[1], zTraceName, nName+1);
	17295	+ pNew->pAppData = pInfo;
	17296	+ pNew->xOpen = vfstraceOpen;
	17297	+ pNew->xDelete = vfstraceDelete;
	17298	+ pNew->xAccess = vfstraceAccess;
	17299	+ pNew->xFullPathname = vfstraceFullPathname;
	17300	+ pNew->xDlOpen = pRoot->xDlOpen==0 ? 0 : vfstraceDlOpen;
	17301	+ pNew->xDlError = pRoot->xDlError==0 ? 0 : vfstraceDlError;
	17302	+ pNew->xDlSym = pRoot->xDlSym==0 ? 0 : vfstraceDlSym;
	17303	+ pNew->xDlClose = pRoot->xDlClose==0 ? 0 : vfstraceDlClose;
	17304	+ pNew->xRandomness = vfstraceRandomness;
	17305	+ pNew->xSleep = vfstraceSleep;
	17306	+ pNew->xCurrentTime = vfstraceCurrentTime;
	17307	+ pNew->xGetLastError = pRoot->xGetLastError==0 ? 0 : vfstraceGetLastError;
	17308	+ if( pNew->iVersion>=2 ){
	17309	+ pNew->xCurrentTimeInt64 = pRoot->xCurrentTimeInt64==0 ? 0 :
	17310	+ vfstraceCurrentTimeInt64;
	17311	+ if( pNew->iVersion>=3 ){
	17312	+ pNew->xSetSystemCall = pRoot->xSetSystemCall==0 ? 0 :
	17313	+ vfstraceSetSystemCall;
	17314	+ pNew->xGetSystemCall = pRoot->xGetSystemCall==0 ? 0 :
	17315	+ vfstraceGetSystemCall;
	17316	+ pNew->xNextSystemCall = pRoot->xNextSystemCall==0 ? 0 :
	17317	+ vfstraceNextSystemCall;
	17318	+ }
	17319	+ }
	17320	+ pInfo->pRootVfs = pRoot;
	17321	+ pInfo->xOut = xOut;
	17322	+ pInfo->pOutArg = pOutArg;
	17323	+ pInfo->zVfsName = pNew->zName;
	17324	+ pInfo->pTraceVfs = pNew;
	17325	+ pInfo->mTrace = 0xffffffff;
	17326	+ pInfo->bOn = 1;
	17327	+ vfstrace_printf(pInfo, "%s.enabled_for(\"%s\")\n",
	17328	+ pInfo->zVfsName, pRoot->zName);
	17329	+ return sqlite3_vfs_register(pNew, makeDefault);
	17330	+}
	17331	+
	17332	+/*
	17333	+** Look for the named VFS. If it is a TRACEVFS, then unregister it
	17334	+** and delete it.
	17335	+*/
	17336	+void vfstrace_unregister(const char *zTraceName){
	17337	+ sqlite3_vfs *pVfs = sqlite3_vfs_find(zTraceName);
	17338	+ if( pVfs==0 ) return;
	17339	+ if( pVfs->xOpen!=vfstraceOpen ) return;
	17340	+ sqlite3_vfs_unregister(pVfs);
	17341	+ sqlite3_free(pVfs);
	17342	+}
	17343	+
	17344	+/*********************** End ../ext/misc/vfstrace.c ******************/
16216	17345
16217	17346	#if !defined(SQLITE_OMIT_VIRTUALTABLE) && defined(SQLITE_ENABLE_DBPAGE_VTAB)
16218	17347	#define SQLITE_SHELL_HAVE_RECOVER 1
16219	17348	#else
16220	17349	#define SQLITE_SHELL_HAVE_RECOVER 0
		@@ -18241,11 +19370,11 @@
18241	19370	}while( strstr(z,zBuf)!=0 );
18242	19371	return zBuf;
18243	19372	}
18244	19373
18245	19374	/*
18246		-** Implementation of scalar SQL function "escape_crnl". The argument passed to
	19375	+** Implementation of scalar SQL function "escape_crlf". The argument passed to
18247	19376	** this function is the output of built-in function quote(). If the first
18248	19377	** character of the input is "'", indicating that the value passed to quote()
18249	19378	** was a text value, then this function searches the input for "\n" and "\r"
18250	19379	** characters and adds a wrapper similar to the following:
18251	19380	**
		@@ -18252,11 +19381,11 @@
18252	19381	** replace(replace(<input>, '\n', char(10), '\r', char(13));
18253	19382	**
18254	19383	** Or, if the first character of the input is not "'", then a copy of the input
18255	19384	** is returned.
18256	19385	*/
18257		-static void recoverEscapeCrnl(
	19386	+static void recoverEscapeCrlf(
18258	19387	sqlite3_context *context,
18259	19388	int argc,
18260	19389	sqlite3_value **argv
18261	19390	){
18262	19391	const char zText = (const char)sqlite3_value_text(argv[0]);
		@@ -18467,11 +19596,11 @@
18467	19596	void (xFunc)(sqlite3_context,int,sqlite3_value **);
18468	19597	} aFunc[] = {
18469	19598	{ "getpage", 1, recoverGetPage },
18470	19599	{ "page_is_used", 1, recoverPageIsUsed },
18471	19600	{ "read_i32", 2, recoverReadI32 },
18472		- { "escape_crnl", 1, recoverEscapeCrnl },
	19601	+ { "escape_crlf", 1, recoverEscapeCrlf },
18473	19602	};
18474	19603
18475	19604	const int flags = SQLITE_OPEN_URI\|SQLITE_OPEN_CREATE\|SQLITE_OPEN_READWRITE;
18476	19605	sqlite3 db = 0; / New database handle */
18477	19606	int ii; /* For iterating through aFunc[] */
		@@ -18820,11 +19949,11 @@
18820	19949	assert( pTab->aCol[ii].iField>=0 && pTab->aCol[ii].iBind>=1 );
18821	19950	zSql = recoverMPrintf(p, "%z%s%Q", zSql, zSep, pTab->aCol[ii].zCol);
18822	19951
18823	19952	if( bSql ){
18824	19953	zBind = recoverMPrintf(p,
18825		- "%z%sescape_crnl(quote(?%d))", zBind, zSqlSep, pTab->aCol[ii].iBind
	19954	+ "%z%sescape_crlf(quote(?%d))", zBind, zSqlSep, pTab->aCol[ii].iBind
18826	19955	);
18827	19956	zSqlSep = "\|\|', '\|\|";
18828	19957	}else{
18829	19958	zBind = recoverMPrintf(p, "%z%s?%d", zBind, zSep, pTab->aCol[ii].iBind);
18830	19959	}
		@@ -19322,10 +20451,12 @@
19322	20451	apVal[iField] = sqlite3_value_dup( pVal );
19323	20452	if( apVal[iField]==0 ){
19324	20453	recoverError(p, SQLITE_NOMEM, 0);
19325	20454	}
19326	20455	p1->nVal = iField+1;
	20456	+ }else if( pTab->nCol==0 ){
	20457	+ p1->nVal = pTab->nCol;
19327	20458	}
19328	20459	p1->iPrevCell = iCell;
19329	20460	p1->iPrevPage = iPage;
19330	20461	}
19331	20462	}else{
		@@ -20437,10 +21568,11 @@
20437	21568	u8 nEqpLevel; /* Depth of the EQP output graph */
20438	21569	u8 eTraceType; /* SHELL_TRACE_* value for type of trace */
20439	21570	u8 bSafeMode; /* True to prohibit unsafe operations */
20440	21571	u8 bSafeModePersist; /* The long-term value of bSafeMode */
20441	21572	u8 eRestoreState; /* See comments above doAutoDetectRestore() */
	21573	+ u8 crlfMode; /* Do NL-to-CRLF translations when enabled (maybe) */
20442	21574	ColModeOpts cmOpts; /* Option values affecting columnar mode output */
20443	21575	unsigned statsOn; /* True to display memory stats before each finalize */
20444	21576	unsigned mEqpLines; /* Mask of vertical lines in the EQP output graph */
20445	21577	int inputNesting; /* Track nesting level of .read and other redirects */
20446	21578	int outCount; /* Revert to stdout when reaching zero */
		@@ -20582,10 +21714,11 @@
20582	21714	#define MODE_Table 15 /* MySQL-style table formatting */
20583	21715	#define MODE_Box 16 /* Unicode box-drawing characters */
20584	21716	#define MODE_Count 17 /* Output only a count of the rows of output */
20585	21717	#define MODE_Off 18 /* No query output shown */
20586	21718	#define MODE_ScanExp 19 /* Like MODE_Explain, but for ".scanstats vm" */
	21719	+#define MODE_Www 20 /* Full web-page output */
20587	21720
20588	21721	static const char *modeDescr[] = {
20589	21722	"line",
20590	21723	"column",
20591	21724	"list",
		@@ -20602,11 +21735,13 @@
20602	21735	"json",
20603	21736	"markdown",
20604	21737	"table",
20605	21738	"box",
20606	21739	"count",
20607		- "off"
	21740	+ "off",
	21741	+ "scanexp",
	21742	+ "www",
20608	21743	};
20609	21744
20610	21745	/*
20611	21746	** These are the column/row/line separators used by the various
20612	21747	** import/export modes.
		@@ -20630,11 +21765,11 @@
20630	21765	** A callback for the sqlite3_log() interface.
20631	21766	*/
20632	21767	static void shellLog(void pArg, int iErrCode, const char zMsg){
20633	21768	ShellState p = (ShellState)pArg;
20634	21769	if( p->pLog==0 ) return;
20635		- sputf(p->pLog, "(%d) %s\n", iErrCode, zMsg);
	21770	+ sqlite3_fprintf(p->pLog, "(%d) %s\n", iErrCode, zMsg);
20636	21771	fflush(p->pLog);
20637	21772	}
20638	21773
20639	21774	/*
20640	21775	** SQL function: shell_putsnl(X)
		@@ -20645,13 +21780,13 @@
20645	21780	static void shellPutsFunc(
20646	21781	sqlite3_context *pCtx,
20647	21782	int nVal,
20648	21783	sqlite3_value **apVal
20649	21784	){
20650		- /* Unused: (ShellState)sqlite3_user_data(pCtx); /
	21785	+ ShellState p = (ShellState)sqlite3_user_data(pCtx);
20651	21786	(void)nVal;
20652		- oputf("%s\n", sqlite3_value_text(apVal[0]));
	21787	+ sqlite3_fprintf(p->out, "%s\n", sqlite3_value_text(apVal[0]));
20653	21788	sqlite3_result_value(pCtx, apVal[0]);
20654	21789	}
20655	21790
20656	21791	/*
20657	21792	** If in safe mode, print an error message described by the arguments
		@@ -20666,11 +21801,11 @@
20666	21801	va_list ap;
20667	21802	char *zMsg;
20668	21803	va_start(ap, zErrMsg);
20669	21804	zMsg = sqlite3_vmprintf(zErrMsg, ap);
20670	21805	va_end(ap);
20671		- eputf("line %d: %s\n", p->lineno, zMsg);
	21806	+ sqlite3_fprintf(stderr, "line %d: %s\n", p->lineno, zMsg);
20672	21807	exit(1);
20673	21808	}
20674	21809	}
20675	21810
20676	21811	/*
		@@ -20699,11 +21834,11 @@
20699	21834	char *zTempFile = 0;
20700	21835	sqlite3 *db;
20701	21836	char *zCmd = 0;
20702	21837	int bBin;
20703	21838	int rc;
20704		- int hasCRNL = 0;
	21839	+ int hasCRLF = 0;
20705	21840	FILE *f = 0;
20706	21841	sqlite3_int64 sz;
20707	21842	sqlite3_int64 x;
20708	21843	unsigned char *p = 0;
20709	21844
		@@ -20733,11 +21868,11 @@
20733	21868	}
20734	21869	}
20735	21870	bBin = sqlite3_value_type(argv[0])==SQLITE_BLOB;
20736	21871	/* When writing the file to be edited, do \n to \r\n conversions on systems
20737	21872	** that want \r\n line endings */
20738		- f = fopen(zTempFile, bBin ? "wb" : "w");
	21873	+ f = sqlite3_fopen(zTempFile, bBin ? "wb" : "w");
20739	21874	if( f==0 ){
20740	21875	sqlite3_result_error(context, "edit() cannot open temp file", -1);
20741	21876	goto edit_func_end;
20742	21877	}
20743	21878	sz = sqlite3_value_bytes(argv[0]);
		@@ -20744,11 +21879,11 @@
20744	21879	if( bBin ){
20745	21880	x = fwrite(sqlite3_value_blob(argv[0]), 1, (size_t)sz, f);
20746	21881	}else{
20747	21882	const char z = (const char)sqlite3_value_text(argv[0]);
20748	21883	/* Remember whether or not the value originally contained \r\n */
20749		- if( z && strstr(z,"\r\n")!=0 ) hasCRNL = 1;
	21884	+ if( z && strstr(z,"\r\n")!=0 ) hasCRLF = 1;
20750	21885	x = fwrite(sqlite3_value_text(argv[0]), 1, (size_t)sz, f);
20751	21886	}
20752	21887	fclose(f);
20753	21888	f = 0;
20754	21889	if( x!=sz ){
		@@ -20764,11 +21899,11 @@
20764	21899	sqlite3_free(zCmd);
20765	21900	if( rc ){
20766	21901	sqlite3_result_error(context, "EDITOR returned non-zero", -1);
20767	21902	goto edit_func_end;
20768	21903	}
20769		- f = fopen(zTempFile, "rb");
	21904	+ f = sqlite3_fopen(zTempFile, "rb");
20770	21905	if( f==0 ){
20771	21906	sqlite3_result_error(context,
20772	21907	"edit() cannot reopen temp file after edit", -1);
20773	21908	goto edit_func_end;
20774	21909	}
		@@ -20789,11 +21924,11 @@
20789	21924	}
20790	21925	if( bBin ){
20791	21926	sqlite3_result_blob64(context, p, sz, sqlite3_free);
20792	21927	}else{
20793	21928	sqlite3_int64 i, j;
20794		- if( hasCRNL ){
	21929	+ if( hasCRLF ){
20795	21930	/* If the original contains \r\n then do no conversions back to \n */
20796	21931	}else{
20797	21932	/* If the file did not originally contain \r\n then convert any new
20798	21933	** \r\n back into \n */
20799	21934	p[sz] = 0;
		@@ -20830,15 +21965,30 @@
20830	21965	p->mode = p->modePrior;
20831	21966	p->shellFlgs = p->priorShFlgs;
20832	21967	memcpy(p->colSeparator, p->colSepPrior, sizeof(p->colSeparator));
20833	21968	memcpy(p->rowSeparator, p->rowSepPrior, sizeof(p->rowSeparator));
20834	21969	}
	21970	+
	21971	+/*
	21972	+** Set output mode to text or binary for Windows.
	21973	+*/
	21974	+static void setCrlfMode(ShellState *p){
	21975	+#ifdef _WIN32
	21976	+ if( p->crlfMode ){
	21977	+ sqlite3_fsetmode(p->out, _O_TEXT);
	21978	+ }else{
	21979	+ sqlite3_fsetmode(p->out, _O_BINARY);
	21980	+ }
	21981	+#else
	21982	+ UNUSED_PARAMETER(p);
	21983	+#endif
	21984	+}
20835	21985
20836	21986	/*
20837	21987	** Output the given string as a hex-encoded blob (eg. X'1234' )
20838	21988	*/
20839		-static void output_hex_blob(const void *pBlob, int nBlob){
	21989	+static void output_hex_blob(FILE out, const void pBlob, int nBlob){
20840	21990	int i;
20841	21991	unsigned char aBlob = (unsigned char)pBlob;
20842	21992
20843	21993	char zStr = sqlite3_malloc(nBlob2 + 1);
20844	21994	shell_check_oom(zStr);
		@@ -20851,11 +22001,11 @@
20851	22001	zStr[i*2] = aHex[ (aBlob[i] >> 4) ];
20852	22002	zStr[i*2+1] = aHex[ (aBlob[i] & 0x0F) ];
20853	22003	}
20854	22004	zStr[i*2] = '\0';
20855	22005
20856		- oputf("X'%s'", zStr);
	22006	+ sqlite3_fprintf(out, "X'%s'", zStr);
20857	22007	sqlite3_free(zStr);
20858	22008	}
20859	22009
20860	22010	/*
20861	22011	** Find a string that is not found anywhere in z[]. Return a pointer
		@@ -20881,46 +22031,40 @@
20881	22031	/*
20882	22032	** Output the given string as a quoted string using SQL quoting conventions.
20883	22033	**
20884	22034	** See also: output_quoted_escaped_string()
20885	22035	*/
20886		-static void output_quoted_string(const char *z){
	22036	+static void output_quoted_string(ShellState p, const char z){
20887	22037	int i;
20888	22038	char c;
20889		-#ifndef SQLITE_SHELL_FIDDLE
20890		- FILE *pfO = setOutputStream(invalidFileStream);
20891		- setBinaryMode(pfO, 1);
20892		-#endif
	22039	+ FILE *out = p->out;
	22040	+ sqlite3_fsetmode(out, _O_BINARY);
20893	22041	if( z==0 ) return;
20894	22042	for(i=0; (c = z[i])!=0 && c!='\''; i++){}
20895	22043	if( c==0 ){
20896		- oputf("'%s'",z);
	22044	+ sqlite3_fprintf(out, "'%s'",z);
20897	22045	}else{
20898		- oputz("'");
	22046	+ sqlite3_fputs("'", out);
20899	22047	while( *z ){
20900	22048	for(i=0; (c = z[i])!=0 && c!='\''; i++){}
20901	22049	if( c=='\'' ) i++;
20902	22050	if( i ){
20903		- oputf("%.*s", i, z);
	22051	+ sqlite3_fprintf(out, "%.*s", i, z);
20904	22052	z += i;
20905	22053	}
20906	22054	if( c=='\'' ){
20907		- oputz("'");
	22055	+ sqlite3_fputs("'", out);
20908	22056	continue;
20909	22057	}
20910	22058	if( c==0 ){
20911	22059	break;
20912	22060	}
20913	22061	z++;
20914	22062	}
20915		- oputz("'");
	22063	+ sqlite3_fputs("'", out);
20916	22064	}
20917		-#ifndef SQLITE_SHELL_FIDDLE
20918		- setTextMode(pfO, 1);
20919		-#else
20920		- setTextMode(stdout, 1);
20921		-#endif
	22065	+ setCrlfMode(p);
20922	22066	}
20923	22067
20924	22068	/*
20925	22069	** Output the given string as a quoted string using SQL quoting conventions.
20926	22070	** Additionallly , escape the "\n" and "\r" characters so that they do not
		@@ -20928,20 +22072,18 @@
20928	22072	** systems.
20929	22073	**
20930	22074	** This is like output_quoted_string() but with the addition of the \r\n
20931	22075	** escape mechanism.
20932	22076	*/
20933		-static void output_quoted_escaped_string(const char *z){
	22077	+static void output_quoted_escaped_string(ShellState p, const char z){
20934	22078	int i;
20935	22079	char c;
20936		-#ifndef SQLITE_SHELL_FIDDLE
20937		- FILE *pfO = setOutputStream(invalidFileStream);
20938		- setBinaryMode(pfO, 1);
20939		-#endif
	22080	+ FILE *out = p->out;
	22081	+ sqlite3_fsetmode(out, _O_BINARY);
20940	22082	for(i=0; (c = z[i])!=0 && c!='\'' && c!='\n' && c!='\r'; i++){}
20941	22083	if( c==0 ){
20942		- oputf("'%s'",z);
	22084	+ sqlite3_fprintf(out, "'%s'",z);
20943	22085	}else{
20944	22086	const char *zNL = 0;
20945	22087	const char *zCR = 0;
20946	22088	int nNL = 0;
20947	22089	int nCR = 0;
		@@ -20949,52 +22091,48 @@
20949	22091	for(i=0; z[i]; i++){
20950	22092	if( z[i]=='\n' ) nNL++;
20951	22093	if( z[i]=='\r' ) nCR++;
20952	22094	}
20953	22095	if( nNL ){
20954		- oputz("replace(");
	22096	+ sqlite3_fputs("replace(", out);
20955	22097	zNL = unused_string(z, "\\n", "\\012", zBuf1);
20956	22098	}
20957	22099	if( nCR ){
20958		- oputz("replace(");
	22100	+ sqlite3_fputs("replace(", out);
20959	22101	zCR = unused_string(z, "\\r", "\\015", zBuf2);
20960	22102	}
20961		- oputz("'");
	22103	+ sqlite3_fputs("'", out);
20962	22104	while( *z ){
20963	22105	for(i=0; (c = z[i])!=0 && c!='\n' && c!='\r' && c!='\''; i++){}
20964	22106	if( c=='\'' ) i++;
20965	22107	if( i ){
20966		- oputf("%.*s", i, z);
	22108	+ sqlite3_fprintf(out, "%.*s", i, z);
20967	22109	z += i;
20968	22110	}
20969	22111	if( c=='\'' ){
20970		- oputz("'");
	22112	+ sqlite3_fputs("'", out);
20971	22113	continue;
20972	22114	}
20973	22115	if( c==0 ){
20974	22116	break;
20975	22117	}
20976	22118	z++;
20977	22119	if( c=='\n' ){
20978		- oputz(zNL);
	22120	+ sqlite3_fputs(zNL, out);
20979	22121	continue;
20980	22122	}
20981		- oputz(zCR);
	22123	+ sqlite3_fputs(zCR, out);
20982	22124	}
20983		- oputz("'");
	22125	+ sqlite3_fputs("'", out);
20984	22126	if( nCR ){
20985		- oputf(",'%s',char(13))", zCR);
	22127	+ sqlite3_fprintf(out, ",'%s',char(13))", zCR);
20986	22128	}
20987	22129	if( nNL ){
20988		- oputf(",'%s',char(10))", zNL);
	22130	+ sqlite3_fprintf(out, ",'%s',char(10))", zNL);
20989	22131	}
20990	22132	}
20991		-#ifndef SQLITE_SHELL_FIDDLE
20992		- setTextMode(pfO, 1);
20993		-#else
20994		- setTextMode(stdout, 1);
20995		-#endif
	22133	+ setCrlfMode(p);
20996	22134	}
20997	22135
20998	22136	/*
20999	22137	** Find earliest of chars within s specified in zAny.
21000	22138	** With ns == ~0, is like strpbrk(s,zAny) and s must be 0-terminated.
		@@ -21010,26 +22148,64 @@
21010	22148	}
21011	22149	++zAny;
21012	22150	}
21013	22151	return pcFirst;
21014	22152	}
	22153	+
	22154	+/* Skip over as much z[] input char sequence as is valid UTF-8,
	22155	+** limited per nAccept char's or whole characters and containing
	22156	+** no char cn such that ((1<<cn) & ccm)!=0. On return, the
	22157	+** sequence z:return (inclusive:exclusive) is validated UTF-8.
	22158	+** Limit: nAccept>=0 => char count, nAccept<0 => character
	22159	+ */
	22160	+const char zSkipValidUtf8(const char z, int nAccept, long ccm){
	22161	+ int ng = (nAccept<0)? -nAccept : 0;
	22162	+ const char *pcLimit = (nAccept>=0)? z+nAccept : 0;
	22163	+ assert(z!=0);
	22164	+ while( (pcLimit)? (z<pcLimit) : (ng-- != 0) ){
	22165	+ unsigned char c = (u8)z;
	22166	+ if( c<0x7f ){
	22167	+ if( ccm != 0L && c < 0x20 && ((1L<<c) & ccm) != 0 ) return z;
	22168	+ ++z; /* ASCII */
	22169	+ }else if( (c & 0xC0) != 0xC0 ) return z; /* not a lead byte */
	22170	+ else{
	22171	+ const char zt = z+1; / Got lead byte, look at trail bytes.*/
	22172	+ do{
	22173	+ if( pcLimit && zt >= pcLimit ) return z;
	22174	+ else{
	22175	+ char ct = *zt++;
	22176	+ if( ct==0 \|\| (zt-z)>4 \|\| (ct & 0xC0)!=0x80 ){
	22177	+ /* Trailing bytes are too few, too many, or invalid. */
	22178	+ return z;
	22179	+ }
	22180	+ }
	22181	+ } while( ((c <<= 1) & 0x40) == 0x40 ); /* Eat lead byte's count. */
	22182	+ z = zt;
	22183	+ }
	22184	+ }
	22185	+ return z;
	22186	+}
	22187	+
	22188	+
21015	22189	/*
21016	22190	** Output the given string as a quoted according to C or TCL quoting rules.
21017	22191	*/
21018		-static void output_c_string(const char *z){
	22192	+static void output_c_string(FILE out, const char z){
21019	22193	char c;
21020	22194	static const char *zq = "\"";
21021	22195	static long ctrlMask = ~0L;
21022	22196	static const char zDQBSRO = "\"\\\x7f"; / double-quote, backslash, rubout */
21023	22197	char ace[3] = "\\?";
21024	22198	char cbsSay;
21025		- oputz(zq);
	22199	+ sqlite3_fputs(zq, out);
21026	22200	while( *z!=0 ){
21027	22201	const char *pcDQBSRO = anyOfInStr(z, zDQBSRO, ~(size_t)0);
21028	22202	const char *pcPast = zSkipValidUtf8(z, INT_MAX, ctrlMask);
21029	22203	const char *pcEnd = (pcDQBSRO && pcDQBSRO < pcPast)? pcDQBSRO : pcPast;
21030		- if( pcEnd > z ) oputb(z, (int)(pcEnd-z));
	22204	+ if( pcEnd > z ){
	22205	+ sqlite3_fprintf(out, "%.*s", (int)(pcEnd-z), z);
	22206	+ }
21031	22207	if( (c = *pcEnd)==0 ) break;
21032	22208	++pcEnd;
21033	22209	switch( c ){
21034	22210	case '\\': case '"':
21035	22211	cbsSay = (char)c;
		@@ -21040,47 +22216,47 @@
21040	22216	case '\f': cbsSay = 'f'; break;
21041	22217	default: cbsSay = 0; break;
21042	22218	}
21043	22219	if( cbsSay ){
21044	22220	ace[1] = cbsSay;
21045		- oputz(ace);
	22221	+ sqlite3_fputs(ace, out);
21046	22222	}else if( !isprint(c&0xff) ){
21047		- oputf("\\%03o", c&0xff);
	22223	+ sqlite3_fprintf(out, "\\%03o", c&0xff);
21048	22224	}else{
21049	22225	ace[1] = (char)c;
21050		- oputz(ace+1);
	22226	+ sqlite3_fputs(ace+1, out);
21051	22227	}
21052	22228	z = pcEnd;
21053	22229	}
21054		- oputz(zq);
	22230	+ sqlite3_fputs(zq, out);
21055	22231	}
21056	22232
21057	22233	/*
21058		-** Output the given string as a quoted according to JSON quoting rules.
	22234	+** Output the given string as quoted according to JSON quoting rules.
21059	22235	*/
21060		-static void output_json_string(const char *z, i64 n){
21061		- char c;
	22236	+static void output_json_string(FILE out, const char z, i64 n){
	22237	+ unsigned char c;
21062	22238	static const char *zq = "\"";
21063	22239	static long ctrlMask = ~0L;
21064	22240	static const char *zDQBS = "\"\\";
21065	22241	const char *pcLimit;
21066	22242	char ace[3] = "\\?";
21067	22243	char cbsSay;
21068	22244
21069	22245	if( z==0 ) z = "";
21070	22246	pcLimit = z + ((n<0)? strlen(z) : (size_t)n);
21071		- oputz(zq);
	22247	+ sqlite3_fputs(zq, out);
21072	22248	while( z < pcLimit ){
21073	22249	const char *pcDQBS = anyOfInStr(z, zDQBS, pcLimit-z);
21074	22250	const char *pcPast = zSkipValidUtf8(z, (int)(pcLimit-z), ctrlMask);
21075	22251	const char *pcEnd = (pcDQBS && pcDQBS < pcPast)? pcDQBS : pcPast;
21076	22252	if( pcEnd > z ){
21077		- oputb(z, (int)(pcEnd-z));
	22253	+ sqlite3_fprintf(out, "%.*s", (int)(pcEnd-z), z);
21078	22254	z = pcEnd;
21079	22255	}
21080	22256	if( z >= pcLimit ) break;
21081		- c = *(z++);
	22257	+ c = (unsigned char)*(z++);
21082	22258	switch( c ){
21083	22259	case '"': case '\\':
21084	22260	cbsSay = (char)c;
21085	22261	break;
21086	22262	case '\b': cbsSay = 'b'; break;
		@@ -21090,26 +22266,26 @@
21090	22266	case '\t': cbsSay = 't'; break;
21091	22267	default: cbsSay = 0; break;
21092	22268	}
21093	22269	if( cbsSay ){
21094	22270	ace[1] = cbsSay;
21095		- oputz(ace);
21096		- }else if( c<=0x1f ){
21097		- oputf("u%04x", c);
	22271	+ sqlite3_fputs(ace, out);
	22272	+ }else if( c<=0x1f \|\| c>=0x7f ){
	22273	+ sqlite3_fprintf(out, "\\u%04x", c);
21098	22274	}else{
21099	22275	ace[1] = (char)c;
21100		- oputz(ace+1);
	22276	+ sqlite3_fputs(ace+1, out);
21101	22277	}
21102	22278	}
21103		- oputz(zq);
	22279	+ sqlite3_fputs(zq, out);
21104	22280	}
21105	22281
21106	22282	/*
21107	22283	** Output the given string with characters that are special to
21108	22284	** HTML escaped.
21109	22285	*/
21110		-static void output_html_string(const char *z){
	22286	+static void output_html_string(FILE out, const char z){
21111	22287	int i;
21112	22288	if( z==0 ) z = "";
21113	22289	while( *z ){
21114	22290	for(i=0; z[i]
21115	22291	&& z[i]!='<'
		@@ -21117,22 +22293,22 @@
21117	22293	&& z[i]!='>'
21118	22294	&& z[i]!='\"'
21119	22295	&& z[i]!='\'';
21120	22296	i++){}
21121	22297	if( i>0 ){
21122		- oputf("%.*s",i,z);
	22298	+ sqlite3_fprintf(out, "%.*s",i,z);
21123	22299	}
21124	22300	if( z[i]=='<' ){
21125		- oputz("<");
	22301	+ sqlite3_fputs("<", out);
21126	22302	}else if( z[i]=='&' ){
21127		- oputz("&");
	22303	+ sqlite3_fputs("&", out);
21128	22304	}else if( z[i]=='>' ){
21129		- oputz(">");
	22305	+ sqlite3_fputs(">", out);
21130	22306	}else if( z[i]=='\"' ){
21131		- oputz(""");
	22307	+ sqlite3_fputs(""", out);
21132	22308	}else if( z[i]=='\'' ){
21133		- oputz("'");
	22309	+ sqlite3_fputs("'", out);
21134	22310	}else{
21135	22311	break;
21136	22312	}
21137	22313	z += i + 1;
21138	22314	}
		@@ -21167,11 +22343,11 @@
21167	22343	** the null value. Strings are quoted if necessary. The separator
21168	22344	** is only issued if bSep is true.
21169	22345	*/
21170	22346	static void output_csv(ShellState p, const char z, int bSep){
21171	22347	if( z==0 ){
21172		- oputf("%s",p->nullValue);
	22348	+ sqlite3_fprintf(p->out, "%s",p->nullValue);
21173	22349	}else{
21174	22350	unsigned i;
21175	22351	for(i=0; z[i]; i++){
21176	22352	if( needCsvQuote[((unsigned char*)z)[i]] ){
21177	22353	i = 0;
		@@ -21179,18 +22355,18 @@
21179	22355	}
21180	22356	}
21181	22357	if( i==0 \|\| strstr(z, p->colSeparator)!=0 ){
21182	22358	char *zQuoted = sqlite3_mprintf("\"%w\"", z);
21183	22359	shell_check_oom(zQuoted);
21184		- oputz(zQuoted);
	22360	+ sqlite3_fputs(zQuoted, p->out);
21185	22361	sqlite3_free(zQuoted);
21186	22362	}else{
21187		- oputz(z);
	22363	+ sqlite3_fputs(z, p->out);
21188	22364	}
21189	22365	}
21190	22366	if( bSep ){
21191		- oputz(p->colSeparator);
	22367	+ sqlite3_fputs(p->colSeparator, p->out);
21192	22368	}
21193	22369	}
21194	22370
21195	22371	/*
21196	22372	** This routine runs when the user presses Ctrl-C
		@@ -21294,20 +22470,20 @@
21294	22470	const char *az[4];
21295	22471	az[0] = zA1;
21296	22472	az[1] = zA2;
21297	22473	az[2] = zA3;
21298	22474	az[3] = zA4;
21299		- oputf("authorizer: %s", azAction[op]);
	22475	+ sqlite3_fprintf(p->out, "authorizer: %s", azAction[op]);
21300	22476	for(i=0; i<4; i++){
21301		- oputz(" ");
	22477	+ sqlite3_fputs(" ", p->out);
21302	22478	if( az[i] ){
21303		- output_c_string(az[i]);
	22479	+ output_c_string(p->out, az[i]);
21304	22480	}else{
21305		- oputz("NULL");
	22481	+ sqlite3_fputs("NULL", p->out);
21306	22482	}
21307	22483	}
21308		- oputz("\n");
	22484	+ sqlite3_fputs("\n", p->out);
21309	22485	if( p->bSafeMode ) (void)safeModeAuth(pClientData, op, zA1, zA2, zA3, zA4);
21310	22486	return SQLITE_OK;
21311	22487	}
21312	22488	#endif
21313	22489
		@@ -21319,11 +22495,11 @@
21319	22495	**
21320	22496	** If the schema statement in z[] contains a start-of-comment and if
21321	22497	** sqlite3_complete() returns false, try to terminate the comment before
21322	22498	** printing the result. https://sqlite.org/forum/forumpost/d7be961c5c
21323	22499	*/
21324		-static void printSchemaLine(const char z, const char zTail){
	22500	+static void printSchemaLine(FILE out, const char z, const char *zTail){
21325	22501	char *zToFree = 0;
21326	22502	if( z==0 ) return;
21327	22503	if( zTail==0 ) return;
21328	22504	if( zTail[0]==';' && (strstr(z, "/*")!=0 \|\| strstr(z,"--")!=0) ){
21329	22505	const char *zOrig = z;
		@@ -21341,20 +22517,20 @@
21341	22517	}
21342	22518	sqlite3_free(zNew);
21343	22519	}
21344	22520	}
21345	22521	if( sqlite3_strglob("CREATE TABLE ['\"]*", z)==0 ){
21346		- oputf("CREATE TABLE IF NOT EXISTS %s%s", z+13, zTail);
	22522	+ sqlite3_fprintf(out, "CREATE TABLE IF NOT EXISTS %s%s", z+13, zTail);
21347	22523	}else{
21348		- oputf("%s%s", z, zTail);
	22524	+ sqlite3_fprintf(out, "%s%s", z, zTail);
21349	22525	}
21350	22526	sqlite3_free(zToFree);
21351	22527	}
21352		-static void printSchemaLineN(char z, int n, const char zTail){
	22528	+static void printSchemaLineN(FILE out, char z, int n, const char *zTail){
21353	22529	char c = z[n];
21354	22530	z[n] = 0;
21355		- printSchemaLine(z, zTail);
	22531	+ printSchemaLine(out, z, zTail);
21356	22532	z[n] = c;
21357	22533	}
21358	22534
21359	22535	/*
21360	22536	** Return true if string z[] has nothing but whitespace and comments to the
		@@ -21378,11 +22554,11 @@
21378	22554	EQPGraphRow *pNew;
21379	22555	i64 nText;
21380	22556	if( zText==0 ) return;
21381	22557	nText = strlen(zText);
21382	22558	if( p->autoEQPtest ){
21383		- oputf("%d,%d,%s\n", iEqpId, p2, zText);
	22559	+ sqlite3_fprintf(p->out, "%d,%d,%s\n", iEqpId, p2, zText);
21384	22560	}
21385	22561	pNew = sqlite3_malloc64( sizeof(*pNew) + nText );
21386	22562	shell_check_oom(pNew);
21387	22563	pNew->iEqpId = iEqpId;
21388	22564	pNew->iParentId = p2;
		@@ -21426,11 +22602,12 @@
21426	22602	i64 n = strlen(p->sGraph.zPrefix);
21427	22603	char *z;
21428	22604	for(pRow = eqp_next_row(p, iEqpId, 0); pRow; pRow = pNext){
21429	22605	pNext = eqp_next_row(p, iEqpId, pRow);
21430	22606	z = pRow->zText;
21431		- oputf("%s%s%s\n", p->sGraph.zPrefix, pNext ? "\|--" : "`--", z);
	22607	+ sqlite3_fprintf(p->out, "%s%s%s\n", p->sGraph.zPrefix,
	22608	+ pNext ? "\|--" : "`--", z);
21432	22609	if( n<(i64)sizeof(p->sGraph.zPrefix)-7 ){
21433	22610	memcpy(&p->sGraph.zPrefix[n], pNext ? "\| " : " ", 4);
21434	22611	eqp_render_level(p, pRow->iEqpId);
21435	22612	p->sGraph.zPrefix[n] = 0;
21436	22613	}
		@@ -21446,17 +22623,17 @@
21446	22623	if( pRow->zText[0]=='-' ){
21447	22624	if( pRow->pNext==0 ){
21448	22625	eqp_reset(p);
21449	22626	return;
21450	22627	}
21451		- oputf("%s\n", pRow->zText+3);
	22628	+ sqlite3_fprintf(p->out, "%s\n", pRow->zText+3);
21452	22629	p->sGraph.pRow = pRow->pNext;
21453	22630	sqlite3_free(pRow);
21454	22631	}else if( nCycle>0 ){
21455		- oputf("QUERY PLAN (cycles=%lld [100%%])\n", nCycle);
	22632	+ sqlite3_fprintf(p->out, "QUERY PLAN (cycles=%lld [100%%])\n", nCycle);
21456	22633	}else{
21457		- oputz("QUERY PLAN\n");
	22634	+ sqlite3_fputs("QUERY PLAN\n", p->out);
21458	22635	}
21459	22636	p->sGraph.zPrefix[0] = 0;
21460	22637	eqp_render_level(p, 0);
21461	22638	eqp_reset(p);
21462	22639	}
		@@ -21468,33 +22645,33 @@
21468	22645	*/
21469	22646	static int progress_handler(void *pClientData) {
21470	22647	ShellState p = (ShellState)pClientData;
21471	22648	p->nProgress++;
21472	22649	if( p->nProgress>=p->mxProgress && p->mxProgress>0 ){
21473		- oputf("Progress limit reached (%u)\n", p->nProgress);
	22650	+ sqlite3_fprintf(p->out, "Progress limit reached (%u)\n", p->nProgress);
21474	22651	if( p->flgProgress & SHELL_PROGRESS_RESET ) p->nProgress = 0;
21475	22652	if( p->flgProgress & SHELL_PROGRESS_ONCE ) p->mxProgress = 0;
21476	22653	return 1;
21477	22654	}
21478	22655	if( (p->flgProgress & SHELL_PROGRESS_QUIET)==0 ){
21479		- oputf("Progress %u\n", p->nProgress);
	22656	+ sqlite3_fprintf(p->out, "Progress %u\n", p->nProgress);
21480	22657	}
21481	22658	return 0;
21482	22659	}
21483	22660	#endif /* SQLITE_OMIT_PROGRESS_CALLBACK */
21484	22661
21485	22662	/*
21486	22663	** Print N dashes
21487	22664	*/
21488		-static void print_dashes(int N){
	22665	+static void print_dashes(FILE *out, int N){
21489	22666	const char zDash[] = "--------------------------------------------------";
21490	22667	const int nDash = sizeof(zDash) - 1;
21491	22668	while( N>nDash ){
21492		- oputz(zDash);
	22669	+ sqlite3_fputs(zDash, out);
21493	22670	N -= nDash;
21494	22671	}
21495		- oputf("%.*s", N, zDash);
	22672	+ sqlite3_fprintf(out, "%.*s", N, zDash);
21496	22673	}
21497	22674
21498	22675	/*
21499	22676	** Print a markdown or table-style row separator using ascii-art
21500	22677	*/
		@@ -21503,19 +22680,19 @@
21503	22680	int nArg,
21504	22681	const char *zSep
21505	22682	){
21506	22683	int i;
21507	22684	if( nArg>0 ){
21508		- oputz(zSep);
21509		- print_dashes(p->actualWidth[0]+2);
	22685	+ sqlite3_fputs(zSep, p->out);
	22686	+ print_dashes(p->out, p->actualWidth[0]+2);
21510	22687	for(i=1; i<nArg; i++){
21511		- oputz(zSep);
21512		- print_dashes(p->actualWidth[i]+2);
	22688	+ sqlite3_fputs(zSep, p->out);
	22689	+ print_dashes(p->out, p->actualWidth[i]+2);
21513	22690	}
21514		- oputz(zSep);
	22691	+ sqlite3_fputs(zSep, p->out);
21515	22692	}
21516		- oputz("\n");
	22693	+ sqlite3_fputs("\n", p->out);
21517	22694	}
21518	22695
21519	22696	/*
21520	22697	** This is the callback routine that the shell
21521	22698	** invokes for each row of a query result.
		@@ -21541,13 +22718,13 @@
21541	22718	if( azArg==0 ) break;
21542	22719	for(i=0; i<nArg; i++){
21543	22720	int len = strlen30(azCol[i] ? azCol[i] : "");
21544	22721	if( len>w ) w = len;
21545	22722	}
21546		- if( p->cnt++>0 ) oputz(p->rowSeparator);
	22723	+ if( p->cnt++>0 ) sqlite3_fputs(p->rowSeparator, p->out);
21547	22724	for(i=0; i<nArg; i++){
21548		- oputf("%*s = %s%s", w, azCol[i],
	22725	+ sqlite3_fprintf(p->out, "%*s = %s%s", w, azCol[i],
21549	22726	azArg[i] ? azArg[i] : p->nullValue, p->rowSeparator);
21550	22727	}
21551	22728	break;
21552	22729	}
21553	22730	case MODE_ScanExp:
		@@ -21571,16 +22748,16 @@
21571	22748	if( nArg>nWidth ) nArg = nWidth;
21572	22749
21573	22750	/* If this is the first row seen, print out the headers */
21574	22751	if( p->cnt++==0 ){
21575	22752	for(i=0; i<nArg; i++){
21576		- utf8_width_print(aWidth[i], azCol[ aMap[i] ]);
21577		- oputz(i==nArg-1 ? "\n" : " ");
	22753	+ utf8_width_print(p->out, aWidth[i], azCol[ aMap[i] ]);
	22754	+ sqlite3_fputs(i==nArg-1 ? "\n" : " ", p->out);
21578	22755	}
21579	22756	for(i=0; i<nArg; i++){
21580		- print_dashes(aWidth[i]);
21581		- oputz(i==nArg-1 ? "\n" : " ");
	22757	+ print_dashes(p->out, aWidth[i]);
	22758	+ sqlite3_fputs(i==nArg-1 ? "\n" : " ", p->out);
21582	22759	}
21583	22760	}
21584	22761
21585	22762	/* If there is no data, exit early. */
21586	22763	if( azArg==0 ) break;
		@@ -21594,21 +22771,21 @@
21594	22771	w = strlenChar(zVal);
21595	22772	zSep = " ";
21596	22773	}
21597	22774	if( i==iIndent && p->aiIndent && p->pStmt ){
21598	22775	if( p->iIndent<p->nIndent ){
21599		- oputf("%*.s", p->aiIndent[p->iIndent], "");
	22776	+ sqlite3_fprintf(p->out, "%*.s", p->aiIndent[p->iIndent], "");
21600	22777	}
21601	22778	p->iIndent++;
21602	22779	}
21603		- utf8_width_print(w, zVal ? zVal : p->nullValue);
21604		- oputz(i==nArg-1 ? "\n" : zSep);
	22780	+ utf8_width_print(p->out, w, zVal ? zVal : p->nullValue);
	22781	+ sqlite3_fputs(i==nArg-1 ? "\n" : zSep, p->out);
21605	22782	}
21606	22783	break;
21607	22784	}
21608	22785	case MODE_Semi: { /* .schema and .fullschema output */
21609		- printSchemaLine(azArg[0], ";\n");
	22786	+ printSchemaLine(p->out, azArg[0], ";\n");
21610	22787	break;
21611	22788	}
21612	22789	case MODE_Pretty: { /* .schema and .fullschema with --indent */
21613	22790	char *z;
21614	22791	int j;
		@@ -21619,11 +22796,11 @@
21619	22796	assert( nArg==1 );
21620	22797	if( azArg[0]==0 ) break;
21621	22798	if( sqlite3_strlike("CREATE VIEW%", azArg[0], 0)==0
21622	22799	\|\| sqlite3_strlike("CREATE TRIG%", azArg[0], 0)==0
21623	22800	){
21624		- oputf("%s;\n", azArg[0]);
	22801	+ sqlite3_fprintf(p->out, "%s;\n", azArg[0]);
21625	22802	break;
21626	22803	}
21627	22804	z = sqlite3_mprintf("%s", azArg[0]);
21628	22805	shell_check_oom(z);
21629	22806	j = 0;
		@@ -21652,255 +22829,265 @@
21652	22829	}else if( c=='(' ){
21653	22830	nParen++;
21654	22831	}else if( c==')' ){
21655	22832	nParen--;
21656	22833	if( nLine>0 && nParen==0 && j>0 ){
21657		- printSchemaLineN(z, j, "\n");
	22834	+ printSchemaLineN(p->out, z, j, "\n");
21658	22835	j = 0;
21659	22836	}
21660	22837	}
21661	22838	z[j++] = c;
21662	22839	if( nParen==1 && cEnd==0
21663	22840	&& (c=='(' \|\| c=='\n' \|\| (c==',' && !wsToEol(z+i+1)))
21664	22841	){
21665	22842	if( c=='\n' ) j--;
21666		- printSchemaLineN(z, j, "\n ");
	22843	+ printSchemaLineN(p->out, z, j, "\n ");
21667	22844	j = 0;
21668	22845	nLine++;
21669	22846	while( IsSpace(z[i+1]) ){ i++; }
21670	22847	}
21671	22848	}
21672	22849	z[j] = 0;
21673	22850	}
21674		- printSchemaLine(z, ";\n");
	22851	+ printSchemaLine(p->out, z, ";\n");
21675	22852	sqlite3_free(z);
21676	22853	break;
21677	22854	}
21678	22855	case MODE_List: {
21679	22856	if( p->cnt++==0 && p->showHeader ){
21680	22857	for(i=0; i<nArg; i++){
21681		- oputf("%s%s",azCol[i], i==nArg-1 ? p->rowSeparator : p->colSeparator);
	22858	+ sqlite3_fprintf(p->out, "%s%s", azCol[i],
	22859	+ i==nArg-1 ? p->rowSeparator : p->colSeparator);
21682	22860	}
21683	22861	}
21684	22862	if( azArg==0 ) break;
21685	22863	for(i=0; i<nArg; i++){
21686	22864	char *z = azArg[i];
21687	22865	if( z==0 ) z = p->nullValue;
21688		- oputz(z);
21689		- oputz((i<nArg-1)? p->colSeparator : p->rowSeparator);
	22866	+ sqlite3_fputs(z, p->out);
	22867	+ sqlite3_fputs((i<nArg-1)? p->colSeparator : p->rowSeparator, p->out);
21690	22868	}
21691	22869	break;
21692	22870	}
	22871	+ case MODE_Www:
21693	22872	case MODE_Html: {
21694		- if( p->cnt++==0 && p->showHeader ){
21695		- oputz("<TR>");
	22873	+ if( p->cnt==0 && p->cMode==MODE_Www ){
	22874	+ sqlite3_fputs(
	22875	+ "</PRE>\n"
	22876	+ "<TABLE border='1' cellspacing='0' cellpadding='2'>\n"
	22877	+ ,p->out
	22878	+ );
	22879	+ }
	22880	+ if( p->cnt==0 && (p->showHeader \|\| p->cMode==MODE_Www) ){
	22881	+ sqlite3_fputs("<TR>", p->out);
21696	22882	for(i=0; i<nArg; i++){
21697		- oputz("<TH>");
21698		- output_html_string(azCol[i]);
21699		- oputz("</TH>\n");
	22883	+ sqlite3_fputs("<TH>", p->out);
	22884	+ output_html_string(p->out, azCol[i]);
	22885	+ sqlite3_fputs("</TH>\n", p->out);
21700	22886	}
21701		- oputz("</TR>\n");
	22887	+ sqlite3_fputs("</TR>\n", p->out);
21702	22888	}
	22889	+ p->cnt++;
21703	22890	if( azArg==0 ) break;
21704		- oputz("<TR>");
	22891	+ sqlite3_fputs("<TR>", p->out);
21705	22892	for(i=0; i<nArg; i++){
21706		- oputz("<TD>");
21707		- output_html_string(azArg[i] ? azArg[i] : p->nullValue);
21708		- oputz("</TD>\n");
	22893	+ sqlite3_fputs("<TD>", p->out);
	22894	+ output_html_string(p->out, azArg[i] ? azArg[i] : p->nullValue);
	22895	+ sqlite3_fputs("</TD>\n", p->out);
21709	22896	}
21710		- oputz("</TR>\n");
	22897	+ sqlite3_fputs("</TR>\n", p->out);
21711	22898	break;
21712	22899	}
21713	22900	case MODE_Tcl: {
21714	22901	if( p->cnt++==0 && p->showHeader ){
21715	22902	for(i=0; i<nArg; i++){
21716		- output_c_string(azCol[i] ? azCol[i] : "");
21717		- if(i<nArg-1) oputz(p->colSeparator);
	22903	+ output_c_string(p->out, azCol[i] ? azCol[i] : "");
	22904	+ if(i<nArg-1) sqlite3_fputs(p->colSeparator, p->out);
21718	22905	}
21719		- oputz(p->rowSeparator);
	22906	+ sqlite3_fputs(p->rowSeparator, p->out);
21720	22907	}
21721	22908	if( azArg==0 ) break;
21722	22909	for(i=0; i<nArg; i++){
21723		- output_c_string(azArg[i] ? azArg[i] : p->nullValue);
21724		- if(i<nArg-1) oputz(p->colSeparator);
	22910	+ output_c_string(p->out, azArg[i] ? azArg[i] : p->nullValue);
	22911	+ if(i<nArg-1) sqlite3_fputs(p->colSeparator, p->out);
21725	22912	}
21726		- oputz(p->rowSeparator);
	22913	+ sqlite3_fputs(p->rowSeparator, p->out);
21727	22914	break;
21728	22915	}
21729	22916	case MODE_Csv: {
21730		- setBinaryMode(p->out, 1);
	22917	+ sqlite3_fsetmode(p->out, _O_BINARY);
21731	22918	if( p->cnt++==0 && p->showHeader ){
21732	22919	for(i=0; i<nArg; i++){
21733	22920	output_csv(p, azCol[i] ? azCol[i] : "", i<nArg-1);
21734	22921	}
21735		- oputz(p->rowSeparator);
	22922	+ sqlite3_fputs(p->rowSeparator, p->out);
21736	22923	}
21737	22924	if( nArg>0 ){
21738	22925	for(i=0; i<nArg; i++){
21739	22926	output_csv(p, azArg[i], i<nArg-1);
21740	22927	}
21741		- oputz(p->rowSeparator);
	22928	+ sqlite3_fputs(p->rowSeparator, p->out);
21742	22929	}
21743		- setTextMode(p->out, 1);
	22930	+ setCrlfMode(p);
21744	22931	break;
21745	22932	}
21746	22933	case MODE_Insert: {
21747	22934	if( azArg==0 ) break;
21748		- oputf("INSERT INTO %s",p->zDestTable);
	22935	+ sqlite3_fprintf(p->out, "INSERT INTO %s",p->zDestTable);
21749	22936	if( p->showHeader ){
21750		- oputz("(");
	22937	+ sqlite3_fputs("(", p->out);
21751	22938	for(i=0; i<nArg; i++){
21752		- if( i>0 ) oputz(",");
	22939	+ if( i>0 ) sqlite3_fputs(",", p->out);
21753	22940	if( quoteChar(azCol[i]) ){
21754	22941	char *z = sqlite3_mprintf("\"%w\"", azCol[i]);
21755	22942	shell_check_oom(z);
21756		- oputz(z);
	22943	+ sqlite3_fputs(z, p->out);
21757	22944	sqlite3_free(z);
21758	22945	}else{
21759		- oputf("%s", azCol[i]);
	22946	+ sqlite3_fprintf(p->out, "%s", azCol[i]);
21760	22947	}
21761	22948	}
21762		- oputz(")");
	22949	+ sqlite3_fputs(")", p->out);
21763	22950	}
21764	22951	p->cnt++;
21765	22952	for(i=0; i<nArg; i++){
21766		- oputz(i>0 ? "," : " VALUES(");
	22953	+ sqlite3_fputs(i>0 ? "," : " VALUES(", p->out);
21767	22954	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
21768		- oputz("NULL");
	22955	+ sqlite3_fputs("NULL", p->out);
21769	22956	}else if( aiType && aiType[i]==SQLITE_TEXT ){
21770	22957	if( ShellHasFlag(p, SHFLG_Newlines) ){
21771		- output_quoted_string(azArg[i]);
	22958	+ output_quoted_string(p, azArg[i]);
21772	22959	}else{
21773		- output_quoted_escaped_string(azArg[i]);
	22960	+ output_quoted_escaped_string(p, azArg[i]);
21774	22961	}
21775	22962	}else if( aiType && aiType[i]==SQLITE_INTEGER ){
21776		- oputz(azArg[i]);
	22963	+ sqlite3_fputs(azArg[i], p->out);
21777	22964	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
21778	22965	char z[50];
21779	22966	double r = sqlite3_column_double(p->pStmt, i);
21780	22967	sqlite3_uint64 ur;
21781	22968	memcpy(&ur,&r,sizeof(r));
21782	22969	if( ur==0x7ff0000000000000LL ){
21783		- oputz("9.0e+999");
	22970	+ sqlite3_fputs("9.0e+999", p->out);
21784	22971	}else if( ur==0xfff0000000000000LL ){
21785		- oputz("-9.0e+999");
	22972	+ sqlite3_fputs("-9.0e+999", p->out);
21786	22973	}else{
21787	22974	sqlite3_int64 ir = (sqlite3_int64)r;
21788	22975	if( r==(double)ir ){
21789	22976	sqlite3_snprintf(50,z,"%lld.0", ir);
21790	22977	}else{
21791	22978	sqlite3_snprintf(50,z,"%!.20g", r);
21792	22979	}
21793		- oputz(z);
	22980	+ sqlite3_fputs(z, p->out);
21794	22981	}
21795	22982	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
21796	22983	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
21797	22984	int nBlob = sqlite3_column_bytes(p->pStmt, i);
21798		- output_hex_blob(pBlob, nBlob);
	22985	+ output_hex_blob(p->out, pBlob, nBlob);
21799	22986	}else if( isNumber(azArg[i], 0) ){
21800		- oputz(azArg[i]);
	22987	+ sqlite3_fputs(azArg[i], p->out);
21801	22988	}else if( ShellHasFlag(p, SHFLG_Newlines) ){
21802		- output_quoted_string(azArg[i]);
	22989	+ output_quoted_string(p, azArg[i]);
21803	22990	}else{
21804		- output_quoted_escaped_string(azArg[i]);
	22991	+ output_quoted_escaped_string(p, azArg[i]);
21805	22992	}
21806	22993	}
21807		- oputz(");\n");
	22994	+ sqlite3_fputs(");\n", p->out);
21808	22995	break;
21809	22996	}
21810	22997	case MODE_Json: {
21811	22998	if( azArg==0 ) break;
21812	22999	if( p->cnt==0 ){
21813		- fputs("[{", p->out);
	23000	+ sqlite3_fputs("[{", p->out);
21814	23001	}else{
21815		- fputs(",\n{", p->out);
	23002	+ sqlite3_fputs(",\n{", p->out);
21816	23003	}
21817	23004	p->cnt++;
21818	23005	for(i=0; i<nArg; i++){
21819		- output_json_string(azCol[i], -1);
21820		- oputz(":");
	23006	+ output_json_string(p->out, azCol[i], -1);
	23007	+ sqlite3_fputs(":", p->out);
21821	23008	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
21822		- oputz("null");
	23009	+ sqlite3_fputs("null", p->out);
21823	23010	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
21824	23011	char z[50];
21825	23012	double r = sqlite3_column_double(p->pStmt, i);
21826	23013	sqlite3_uint64 ur;
21827	23014	memcpy(&ur,&r,sizeof(r));
21828	23015	if( ur==0x7ff0000000000000LL ){
21829		- oputz("9.0e+999");
	23016	+ sqlite3_fputs("9.0e+999", p->out);
21830	23017	}else if( ur==0xfff0000000000000LL ){
21831		- oputz("-9.0e+999");
	23018	+ sqlite3_fputs("-9.0e+999", p->out);
21832	23019	}else{
21833	23020	sqlite3_snprintf(50,z,"%!.20g", r);
21834		- oputz(z);
	23021	+ sqlite3_fputs(z, p->out);
21835	23022	}
21836	23023	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
21837	23024	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
21838	23025	int nBlob = sqlite3_column_bytes(p->pStmt, i);
21839		- output_json_string(pBlob, nBlob);
	23026	+ output_json_string(p->out, pBlob, nBlob);
21840	23027	}else if( aiType && aiType[i]==SQLITE_TEXT ){
21841		- output_json_string(azArg[i], -1);
	23028	+ output_json_string(p->out, azArg[i], -1);
21842	23029	}else{
21843		- oputz(azArg[i]);
	23030	+ sqlite3_fputs(azArg[i], p->out);
21844	23031	}
21845	23032	if( i<nArg-1 ){
21846		- oputz(",");
	23033	+ sqlite3_fputs(",", p->out);
21847	23034	}
21848	23035	}
21849		- oputz("}");
	23036	+ sqlite3_fputs("}", p->out);
21850	23037	break;
21851	23038	}
21852	23039	case MODE_Quote: {
21853	23040	if( azArg==0 ) break;
21854	23041	if( p->cnt==0 && p->showHeader ){
21855	23042	for(i=0; i<nArg; i++){
21856		- if( i>0 ) fputs(p->colSeparator, p->out);
21857		- output_quoted_string(azCol[i]);
	23043	+ if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
	23044	+ output_quoted_string(p, azCol[i]);
21858	23045	}
21859		- fputs(p->rowSeparator, p->out);
	23046	+ sqlite3_fputs(p->rowSeparator, p->out);
21860	23047	}
21861	23048	p->cnt++;
21862	23049	for(i=0; i<nArg; i++){
21863		- if( i>0 ) fputs(p->colSeparator, p->out);
	23050	+ if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
21864	23051	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
21865		- oputz("NULL");
	23052	+ sqlite3_fputs("NULL", p->out);
21866	23053	}else if( aiType && aiType[i]==SQLITE_TEXT ){
21867		- output_quoted_string(azArg[i]);
	23054	+ output_quoted_string(p, azArg[i]);
21868	23055	}else if( aiType && aiType[i]==SQLITE_INTEGER ){
21869		- oputz(azArg[i]);
	23056	+ sqlite3_fputs(azArg[i], p->out);
21870	23057	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
21871	23058	char z[50];
21872	23059	double r = sqlite3_column_double(p->pStmt, i);
21873	23060	sqlite3_snprintf(50,z,"%!.20g", r);
21874		- oputz(z);
	23061	+ sqlite3_fputs(z, p->out);
21875	23062	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
21876	23063	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
21877	23064	int nBlob = sqlite3_column_bytes(p->pStmt, i);
21878		- output_hex_blob(pBlob, nBlob);
	23065	+ output_hex_blob(p->out, pBlob, nBlob);
21879	23066	}else if( isNumber(azArg[i], 0) ){
21880		- oputz(azArg[i]);
	23067	+ sqlite3_fputs(azArg[i], p->out);
21881	23068	}else{
21882		- output_quoted_string(azArg[i]);
	23069	+ output_quoted_string(p, azArg[i]);
21883	23070	}
21884	23071	}
21885		- fputs(p->rowSeparator, p->out);
	23072	+ sqlite3_fputs(p->rowSeparator, p->out);
21886	23073	break;
21887	23074	}
21888	23075	case MODE_Ascii: {
21889	23076	if( p->cnt++==0 && p->showHeader ){
21890	23077	for(i=0; i<nArg; i++){
21891		- if( i>0 ) oputz(p->colSeparator);
21892		- oputz(azCol[i] ? azCol[i] : "");
	23078	+ if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
	23079	+ sqlite3_fputs(azCol[i] ? azCol[i] : "", p->out);
21893	23080	}
21894		- oputz(p->rowSeparator);
	23081	+ sqlite3_fputs(p->rowSeparator, p->out);
21895	23082	}
21896	23083	if( azArg==0 ) break;
21897	23084	for(i=0; i<nArg; i++){
21898		- if( i>0 ) oputz(p->colSeparator);
21899		- oputz(azArg[i] ? azArg[i] : p->nullValue);
	23085	+ if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
	23086	+ sqlite3_fputs(azArg[i] ? azArg[i] : p->nullValue, p->out);
21900	23087	}
21901		- oputz(p->rowSeparator);
	23088	+ sqlite3_fputs(p->rowSeparator, p->out);
21902	23089	break;
21903	23090	}
21904	23091	case MODE_EQP: {
21905	23092	eqp_append(p, atoi(azArg[0]), atoi(azArg[1]), azArg[3]);
21906	23093	break;
		@@ -21975,11 +23162,11 @@
21975	23162	"INSERT INTO selftest(tno,op,cmd,ans)"
21976	23163	" SELECT rowid*10,op,cmd,ans FROM [_shell$self];\n"
21977	23164	"DROP TABLE [_shell$self];"
21978	23165	,0,0,&zErrMsg);
21979	23166	if( zErrMsg ){
21980		- eputf("SELFTEST initialization failure: %s\n", zErrMsg);
	23167	+ sqlite3_fprintf(stderr, "SELFTEST initialization failure: %s\n", zErrMsg);
21981	23168	sqlite3_free(zErrMsg);
21982	23169	}
21983	23170	sqlite3_exec(p->db, "RELEASE selftest_init",0,0,0);
21984	23171	}
21985	23172
		@@ -22078,36 +23265,37 @@
22078	23265	int i;
22079	23266	const char *z;
22080	23267	rc = sqlite3_prepare_v2(p->db, zSelect, -1, &pSelect, 0);
22081	23268	if( rc!=SQLITE_OK \|\| !pSelect ){
22082	23269	char *zContext = shell_error_context(zSelect, p->db);
22083		- oputf("/** ERROR: (%d) %s ***/\n%s",
	23270	+ sqlite3_fprintf(p->out, "/** ERROR: (%d) %s ***/\n%s",
22084	23271	rc, sqlite3_errmsg(p->db), zContext);
22085	23272	sqlite3_free(zContext);
22086	23273	if( (rc&0xff)!=SQLITE_CORRUPT ) p->nErr++;
22087	23274	return rc;
22088	23275	}
22089	23276	rc = sqlite3_step(pSelect);
22090	23277	nResult = sqlite3_column_count(pSelect);
22091	23278	while( rc==SQLITE_ROW ){
22092	23279	z = (const char*)sqlite3_column_text(pSelect, 0);
22093		- oputf("%s", z);
	23280	+ sqlite3_fprintf(p->out, "%s", z);
22094	23281	for(i=1; i<nResult; i++){
22095		- oputf(",%s", sqlite3_column_text(pSelect, i));
	23282	+ sqlite3_fprintf(p->out, ",%s", sqlite3_column_text(pSelect, i));
22096	23283	}
22097	23284	if( z==0 ) z = "";
22098	23285	while( z[0] && (z[0]!='-' \|\| z[1]!='-') ) z++;
22099	23286	if( z[0] ){
22100		- oputz("\n;\n");
	23287	+ sqlite3_fputs("\n;\n", p->out);
22101	23288	}else{
22102		- oputz(";\n");
	23289	+ sqlite3_fputs(";\n", p->out);
22103	23290	}
22104	23291	rc = sqlite3_step(pSelect);
22105	23292	}
22106	23293	rc = sqlite3_finalize(pSelect);
22107	23294	if( rc!=SQLITE_OK ){
22108		- oputf("/** ERROR: (%d) %s ***/\n", rc, sqlite3_errmsg(p->db));
	23295	+ sqlite3_fprintf(p->out, "/** ERROR: (%d) %s ***/\n",
	23296	+ rc, sqlite3_errmsg(p->db));
22109	23297	if( (rc&0xff)!=SQLITE_CORRUPT ) p->nErr++;
22110	23298	}
22111	23299	return rc;
22112	23300	}
22113	23301
		@@ -22139,17 +23327,17 @@
22139	23327
22140	23328	#ifdef __linux__
22141	23329	/*
22142	23330	** Attempt to display I/O stats on Linux using /proc/PID/io
22143	23331	*/
22144		-static void displayLinuxIoStats(void){
	23332	+static void displayLinuxIoStats(FILE *out){
22145	23333	FILE *in;
22146	23334	char z[200];
22147	23335	sqlite3_snprintf(sizeof(z), z, "/proc/%d/io", getpid());
22148		- in = fopen(z, "rb");
	23336	+ in = sqlite3_fopen(z, "rb");
22149	23337	if( in==0 ) return;
22150		- while( fgets(z, sizeof(z), in)!=0 ){
	23338	+ while( sqlite3_fgets(z, sizeof(z), in)!=0 ){
22151	23339	static const struct {
22152	23340	const char *zPattern;
22153	23341	const char *zDesc;
22154	23342	} aTrans[] = {
22155	23343	{ "rchar: ", "Bytes received by read():" },
		@@ -22162,11 +23350,11 @@
22162	23350	};
22163	23351	int i;
22164	23352	for(i=0; i<ArraySize(aTrans); i++){
22165	23353	int n = strlen30(aTrans[i].zPattern);
22166	23354	if( cli_strncmp(aTrans[i].zPattern, z, n)==0 ){
22167		- oputf("%-36s %s", aTrans[i].zDesc, &z[n]);
	23355	+ sqlite3_fprintf(out, "%-36s %s", aTrans[i].zDesc, &z[n]);
22168	23356	break;
22169	23357	}
22170	23358	}
22171	23359	}
22172	23360	fclose(in);
		@@ -22175,10 +23363,11 @@
22175	23363
22176	23364	/*
22177	23365	** Display a single line of status using 64-bit values.
22178	23366	*/
22179	23367	static void displayStatLine(
	23368	+ FILE out, / Write to this channel */
22180	23369	char zLabel, / Label for this one line */
22181	23370	char zFormat, / Format for the result */
22182	23371	int iStatusCtrl, /* Which status to display */
22183	23372	int bReset /* True to reset the stats */
22184	23373	){
		@@ -22193,11 +23382,11 @@
22193	23382	if( nPercent>1 ){
22194	23383	sqlite3_snprintf(sizeof(zLine), zLine, zFormat, iCur, iHiwtr);
22195	23384	}else{
22196	23385	sqlite3_snprintf(sizeof(zLine), zLine, zFormat, iHiwtr);
22197	23386	}
22198		- oputf("%-36s %s\n", zLabel, zLine);
	23387	+ sqlite3_fprintf(out, "%-36s %s\n", zLabel, zLine);
22199	23388	}
22200	23389
22201	23390	/*
22202	23391	** Display memory stats.
22203	23392	*/
		@@ -22206,130 +23395,152 @@
22206	23395	ShellState pArg, / Pointer to ShellState */
22207	23396	int bReset /* True to reset the stats */
22208	23397	){
22209	23398	int iCur;
22210	23399	int iHiwtr;
	23400	+ FILE *out;
22211	23401	if( pArg==0 \|\| pArg->out==0 ) return 0;
	23402	+ out = pArg->out;
22212	23403
22213	23404	if( pArg->pStmt && pArg->statsOn==2 ){
22214	23405	int nCol, i, x;
22215	23406	sqlite3_stmt *pStmt = pArg->pStmt;
22216	23407	char z[100];
22217	23408	nCol = sqlite3_column_count(pStmt);
22218		- oputf("%-36s %d\n", "Number of output columns:", nCol);
	23409	+ sqlite3_fprintf(out, "%-36s %d\n", "Number of output columns:", nCol);
22219	23410	for(i=0; i<nCol; i++){
22220	23411	sqlite3_snprintf(sizeof(z),z,"Column %d %nname:", i, &x);
22221		- oputf("%-36s %s\n", z, sqlite3_column_name(pStmt,i));
	23412	+ sqlite3_fprintf(out, "%-36s %s\n", z, sqlite3_column_name(pStmt,i));
22222	23413	#ifndef SQLITE_OMIT_DECLTYPE
22223	23414	sqlite3_snprintf(30, z+x, "declared type:");
22224		- oputf("%-36s %s\n", z, sqlite3_column_decltype(pStmt, i));
	23415	+ sqlite3_fprintf(out, "%-36s %s\n", z, sqlite3_column_decltype(pStmt, i));
22225	23416	#endif
22226	23417	#ifdef SQLITE_ENABLE_COLUMN_METADATA
22227	23418	sqlite3_snprintf(30, z+x, "database name:");
22228		- oputf("%-36s %s\n", z, sqlite3_column_database_name(pStmt,i));
	23419	+ sqlite3_fprintf(out, "%-36s %s\n", z,
	23420	+ sqlite3_column_database_name(pStmt,i));
22229	23421	sqlite3_snprintf(30, z+x, "table name:");
22230		- oputf("%-36s %s\n", z, sqlite3_column_table_name(pStmt,i));
	23422	+ sqlite3_fprintf(out, "%-36s %s\n", z, sqlite3_column_table_name(pStmt,i));
22231	23423	sqlite3_snprintf(30, z+x, "origin name:");
22232		- oputf("%-36s %s\n", z, sqlite3_column_origin_name(pStmt,i));
	23424	+ sqlite3_fprintf(out, "%-36s %s\n", z,sqlite3_column_origin_name(pStmt,i));
22233	23425	#endif
22234	23426	}
22235	23427	}
22236	23428
22237	23429	if( pArg->statsOn==3 ){
22238	23430	if( pArg->pStmt ){
22239	23431	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_VM_STEP,bReset);
22240		- oputf("VM-steps: %d\n", iCur);
	23432	+ sqlite3_fprintf(out, "VM-steps: %d\n", iCur);
22241	23433	}
22242	23434	return 0;
22243	23435	}
22244	23436
22245		- displayStatLine("Memory Used:",
	23437	+ displayStatLine(out, "Memory Used:",
22246	23438	"%lld (max %lld) bytes", SQLITE_STATUS_MEMORY_USED, bReset);
22247		- displayStatLine("Number of Outstanding Allocations:",
	23439	+ displayStatLine(out, "Number of Outstanding Allocations:",
22248	23440	"%lld (max %lld)", SQLITE_STATUS_MALLOC_COUNT, bReset);
22249	23441	if( pArg->shellFlgs & SHFLG_Pagecache ){
22250		- displayStatLine("Number of Pcache Pages Used:",
	23442	+ displayStatLine(out, "Number of Pcache Pages Used:",
22251	23443	"%lld (max %lld) pages", SQLITE_STATUS_PAGECACHE_USED, bReset);
22252	23444	}
22253		- displayStatLine("Number of Pcache Overflow Bytes:",
	23445	+ displayStatLine(out, "Number of Pcache Overflow Bytes:",
22254	23446	"%lld (max %lld) bytes", SQLITE_STATUS_PAGECACHE_OVERFLOW, bReset);
22255		- displayStatLine("Largest Allocation:",
	23447	+ displayStatLine(out, "Largest Allocation:",
22256	23448	"%lld bytes", SQLITE_STATUS_MALLOC_SIZE, bReset);
22257		- displayStatLine("Largest Pcache Allocation:",
	23449	+ displayStatLine(out, "Largest Pcache Allocation:",
22258	23450	"%lld bytes", SQLITE_STATUS_PAGECACHE_SIZE, bReset);
22259	23451	#ifdef YYTRACKMAXSTACKDEPTH
22260		- displayStatLine("Deepest Parser Stack:",
	23452	+ displayStatLine(out, "Deepest Parser Stack:",
22261	23453	"%lld (max %lld)", SQLITE_STATUS_PARSER_STACK, bReset);
22262	23454	#endif
22263	23455
22264	23456	if( db ){
22265	23457	if( pArg->shellFlgs & SHFLG_Lookaside ){
22266	23458	iHiwtr = iCur = -1;
22267	23459	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_USED,
22268	23460	&iCur, &iHiwtr, bReset);
22269		- oputf("Lookaside Slots Used: %d (max %d)\n", iCur, iHiwtr);
	23461	+ sqlite3_fprintf(out,
	23462	+ "Lookaside Slots Used: %d (max %d)\n", iCur, iHiwtr);
22270	23463	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_HIT,
22271	23464	&iCur, &iHiwtr, bReset);
22272		- oputf("Successful lookaside attempts: %d\n", iHiwtr);
	23465	+ sqlite3_fprintf(out,
	23466	+ "Successful lookaside attempts: %d\n", iHiwtr);
22273	23467	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_MISS_SIZE,
22274	23468	&iCur, &iHiwtr, bReset);
22275		- oputf("Lookaside failures due to size: %d\n", iHiwtr);
	23469	+ sqlite3_fprintf(out,
	23470	+ "Lookaside failures due to size: %d\n", iHiwtr);
22276	23471	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_MISS_FULL,
22277	23472	&iCur, &iHiwtr, bReset);
22278		- oputf("Lookaside failures due to OOM: %d\n", iHiwtr);
	23473	+ sqlite3_fprintf(out,
	23474	+ "Lookaside failures due to OOM: %d\n", iHiwtr);
22279	23475	}
22280	23476	iHiwtr = iCur = -1;
22281	23477	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_USED, &iCur, &iHiwtr, bReset);
22282		- oputf("Pager Heap Usage: %d bytes\n", iCur);
	23478	+ sqlite3_fprintf(out,
	23479	+ "Pager Heap Usage: %d bytes\n", iCur);
22283	23480	iHiwtr = iCur = -1;
22284	23481	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_HIT, &iCur, &iHiwtr, 1);
22285		- oputf("Page cache hits: %d\n", iCur);
	23482	+ sqlite3_fprintf(out,
	23483	+ "Page cache hits: %d\n", iCur);
22286	23484	iHiwtr = iCur = -1;
22287	23485	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_MISS, &iCur, &iHiwtr, 1);
22288		- oputf("Page cache misses: %d\n", iCur);
	23486	+ sqlite3_fprintf(out,
	23487	+ "Page cache misses: %d\n", iCur);
22289	23488	iHiwtr = iCur = -1;
22290	23489	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_WRITE, &iCur, &iHiwtr, 1);
22291		- oputf("Page cache writes: %d\n", iCur);
	23490	+ sqlite3_fprintf(out,
	23491	+ "Page cache writes: %d\n", iCur);
22292	23492	iHiwtr = iCur = -1;
22293	23493	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_SPILL, &iCur, &iHiwtr, 1);
22294		- oputf("Page cache spills: %d\n", iCur);
	23494	+ sqlite3_fprintf(out,
	23495	+ "Page cache spills: %d\n", iCur);
22295	23496	iHiwtr = iCur = -1;
22296	23497	sqlite3_db_status(db, SQLITE_DBSTATUS_SCHEMA_USED, &iCur, &iHiwtr, bReset);
22297		- oputf("Schema Heap Usage: %d bytes\n", iCur);
	23498	+ sqlite3_fprintf(out,
	23499	+ "Schema Heap Usage: %d bytes\n", iCur);
22298	23500	iHiwtr = iCur = -1;
22299	23501	sqlite3_db_status(db, SQLITE_DBSTATUS_STMT_USED, &iCur, &iHiwtr, bReset);
22300		- oputf("Statement Heap/Lookaside Usage: %d bytes\n", iCur);
	23502	+ sqlite3_fprintf(out,
	23503	+ "Statement Heap/Lookaside Usage: %d bytes\n", iCur);
22301	23504	}
22302	23505
22303	23506	if( pArg->pStmt ){
22304	23507	int iHit, iMiss;
22305	23508	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FULLSCAN_STEP,
22306	23509	bReset);
22307		- oputf("Fullscan Steps: %d\n", iCur);
	23510	+ sqlite3_fprintf(out,
	23511	+ "Fullscan Steps: %d\n", iCur);
22308	23512	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_SORT, bReset);
22309		- oputf("Sort Operations: %d\n", iCur);
	23513	+ sqlite3_fprintf(out,
	23514	+ "Sort Operations: %d\n", iCur);
22310	23515	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_AUTOINDEX,bReset);
22311		- oputf("Autoindex Inserts: %d\n", iCur);
	23516	+ sqlite3_fprintf(out,
	23517	+ "Autoindex Inserts: %d\n", iCur);
22312	23518	iHit = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FILTER_HIT,
22313	23519	bReset);
22314	23520	iMiss = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FILTER_MISS,
22315	23521	bReset);
22316	23522	if( iHit \|\| iMiss ){
22317		- oputf("Bloom filter bypass taken: %d/%d\n", iHit, iHit+iMiss);
	23523	+ sqlite3_fprintf(out,
	23524	+ "Bloom filter bypass taken: %d/%d\n", iHit, iHit+iMiss);
22318	23525	}
22319	23526	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_VM_STEP, bReset);
22320		- oputf("Virtual Machine Steps: %d\n", iCur);
	23527	+ sqlite3_fprintf(out,
	23528	+ "Virtual Machine Steps: %d\n", iCur);
22321	23529	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_REPREPARE,bReset);
22322		- oputf("Reprepare operations: %d\n", iCur);
	23530	+ sqlite3_fprintf(out,
	23531	+ "Reprepare operations: %d\n", iCur);
22323	23532	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_RUN, bReset);
22324		- oputf("Number of times run: %d\n", iCur);
	23533	+ sqlite3_fprintf(out,
	23534	+ "Number of times run: %d\n", iCur);
22325	23535	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_MEMUSED, bReset);
22326		- oputf("Memory used by prepared stmt: %d\n", iCur);
	23536	+ sqlite3_fprintf(out,
	23537	+ "Memory used by prepared stmt: %d\n", iCur);
22327	23538	}
22328	23539
22329	23540	#ifdef __linux__
22330		- displayLinuxIoStats();
	23541	+ displayLinuxIoStats(pArg->out);
22331	23542	#endif
22332	23543
22333	23544	/* Do not remove this machine readable comment: extra-stats-output-here */
22334	23545
22335	23546	return 0;
		@@ -22721,21 +23932,21 @@
22721	23932	#define BOX_1234 "\342\224\274" /* U+253c -\|- */
22722	23933
22723	23934	/* Draw horizontal line N characters long using unicode box
22724	23935	** characters
22725	23936	*/
22726		-static void print_box_line(int N){
	23937	+static void print_box_line(FILE *out, int N){
22727	23938	const char zDash[] =
22728	23939	BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24
22729	23940	BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24;
22730	23941	const int nDash = sizeof(zDash) - 1;
22731	23942	N *= 3;
22732	23943	while( N>nDash ){
22733		- oputz(zDash);
	23944	+ sqlite3_fputs(zDash, out);
22734	23945	N -= nDash;
22735	23946	}
22736		- oputf("%.*s", N, zDash);
	23947	+ sqlite3_fprintf(out, "%.*s", N, zDash);
22737	23948	}
22738	23949
22739	23950	/*
22740	23951	** Draw a horizontal separator for a MODE_Box table.
22741	23952	*/
		@@ -22746,19 +23957,19 @@
22746	23957	const char *zSep2,
22747	23958	const char *zSep3
22748	23959	){
22749	23960	int i;
22750	23961	if( nArg>0 ){
22751		- oputz(zSep1);
22752		- print_box_line(p->actualWidth[0]+2);
	23962	+ sqlite3_fputs(zSep1, p->out);
	23963	+ print_box_line(p->out, p->actualWidth[0]+2);
22753	23964	for(i=1; i<nArg; i++){
22754		- oputz(zSep2);
22755		- print_box_line(p->actualWidth[i]+2);
	23965	+ sqlite3_fputs(zSep2, p->out);
	23966	+ print_box_line(p->out, p->actualWidth[i]+2);
22756	23967	}
22757		- oputz(zSep3);
	23968	+ sqlite3_fputs(zSep3, p->out);
22758	23969	}
22759		- oputz("\n");
	23970	+ sqlite3_fputs("\n", p->out);
22760	23971	}
22761	23972
22762	23973	/*
22763	23974	** z[] is a line of text that is to be displayed the .mode box or table or
22764	23975	** similar tabular formats. z[] might contain control characters such
		@@ -22788,16 +23999,26 @@
22788	23999	}
22789	24000	if( mxWidth<0 ) mxWidth = -mxWidth;
22790	24001	if( mxWidth==0 ) mxWidth = 1000000;
22791	24002	i = j = n = 0;
22792	24003	while( n<mxWidth ){
22793		- if( z[i]>=' ' ){
	24004	+ unsigned char c = z[i];
	24005	+ if( c>=0xc0 ){
	24006	+ int u;
	24007	+ int len = decodeUtf8(&z[i], &u);
	24008	+ i += len;
	24009	+ j += len;
	24010	+ n += cli_wcwidth(u);
	24011	+ continue;
	24012	+ }
	24013	+ if( c>=' ' ){
22794	24014	n++;
22795		- do{ i++; j++; }while( (z[i]&0xc0)==0x80 );
	24015	+ i++;
	24016	+ j++;
22796	24017	continue;
22797	24018	}
22798		- if( z[i]=='\t' ){
	24019	+ if( c=='\t' ){
22799	24020	do{
22800	24021	n++;
22801	24022	j++;
22802	24023	}while( (n&7)!=0 && n<mxWidth );
22803	24024	i++;
		@@ -22835,13 +24056,21 @@
22835	24056	}
22836	24057	zOut = malloc( j+1 );
22837	24058	shell_check_oom(zOut);
22838	24059	i = j = n = 0;
22839	24060	while( i<k ){
22840		- if( z[i]>=' ' ){
	24061	+ unsigned char c = z[i];
	24062	+ if( c>=0xc0 ){
	24063	+ int u;
	24064	+ int len = decodeUtf8(&z[i], &u);
	24065	+ do{ zOut[j++] = z[i++]; }while( (--len)>0 );
	24066	+ n += cli_wcwidth(u);
	24067	+ continue;
	24068	+ }
	24069	+ if( c>=' ' ){
22841	24070	n++;
22842		- do{ zOut[j++] = z[i++]; }while( (z[i]&0xc0)==0x80 );
	24071	+ zOut[j++] = z[i++];
22843	24072	continue;
22844	24073	}
22845	24074	if( z[i]=='\t' ){
22846	24075	do{
22847	24076	n++;
		@@ -23017,97 +24246,99 @@
23017	24246	rowSep = "\n";
23018	24247	if( p->showHeader ){
23019	24248	for(i=0; i<nColumn; i++){
23020	24249	w = p->actualWidth[i];
23021	24250	if( p->colWidth[i]<0 ) w = -w;
23022		- utf8_width_print(w, azData[i]);
23023		- fputs(i==nColumn-1?"\n":" ", p->out);
	24251	+ utf8_width_print(p->out, w, azData[i]);
	24252	+ sqlite3_fputs(i==nColumn-1?"\n":" ", p->out);
23024	24253	}
23025	24254	for(i=0; i<nColumn; i++){
23026		- print_dashes(p->actualWidth[i]);
23027		- fputs(i==nColumn-1?"\n":" ", p->out);
	24255	+ print_dashes(p->out, p->actualWidth[i]);
	24256	+ sqlite3_fputs(i==nColumn-1?"\n":" ", p->out);
23028	24257	}
23029	24258	}
23030	24259	break;
23031	24260	}
23032	24261	case MODE_Table: {
23033	24262	colSep = " \| ";
23034	24263	rowSep = " \|\n";
23035	24264	print_row_separator(p, nColumn, "+");
23036		- fputs("\| ", p->out);
	24265	+ sqlite3_fputs("\| ", p->out);
23037	24266	for(i=0; i<nColumn; i++){
23038	24267	w = p->actualWidth[i];
23039	24268	n = strlenChar(azData[i]);
23040		- oputf("%s%s%s", (w-n)/2, "", azData[i], (w-n+1)/2, "");
23041		- oputz(i==nColumn-1?" \|\n":" \| ");
	24269	+ sqlite3_fprintf(p->out, "%s%s%s", (w-n)/2, "",
	24270	+ azData[i], (w-n+1)/2, "");
	24271	+ sqlite3_fputs(i==nColumn-1?" \|\n":" \| ", p->out);
23042	24272	}
23043	24273	print_row_separator(p, nColumn, "+");
23044	24274	break;
23045	24275	}
23046	24276	case MODE_Markdown: {
23047	24277	colSep = " \| ";
23048	24278	rowSep = " \|\n";
23049		- fputs("\| ", p->out);
	24279	+ sqlite3_fputs("\| ", p->out);
23050	24280	for(i=0; i<nColumn; i++){
23051	24281	w = p->actualWidth[i];
23052	24282	n = strlenChar(azData[i]);
23053		- oputf("%s%s%s", (w-n)/2, "", azData[i], (w-n+1)/2, "");
23054		- oputz(i==nColumn-1?" \|\n":" \| ");
	24283	+ sqlite3_fprintf(p->out, "%s%s%s", (w-n)/2, "",
	24284	+ azData[i], (w-n+1)/2, "");
	24285	+ sqlite3_fputs(i==nColumn-1?" \|\n":" \| ", p->out);
23055	24286	}
23056	24287	print_row_separator(p, nColumn, "\|");
23057	24288	break;
23058	24289	}
23059	24290	case MODE_Box: {
23060	24291	colSep = " " BOX_13 " ";
23061	24292	rowSep = " " BOX_13 "\n";
23062	24293	print_box_row_separator(p, nColumn, BOX_23, BOX_234, BOX_34);
23063		- oputz(BOX_13 " ");
	24294	+ sqlite3_fputs(BOX_13 " ", p->out);
23064	24295	for(i=0; i<nColumn; i++){
23065	24296	w = p->actualWidth[i];
23066	24297	n = strlenChar(azData[i]);
23067		- oputf("%s%s%s%s",
	24298	+ sqlite3_fprintf(p->out, "%s%s%s%s",
23068	24299	(w-n)/2, "", azData[i], (w-n+1)/2, "",
23069	24300	i==nColumn-1?" "BOX_13"\n":" "BOX_13" ");
23070	24301	}
23071	24302	print_box_row_separator(p, nColumn, BOX_123, BOX_1234, BOX_134);
23072	24303	break;
23073	24304	}
23074	24305	}
23075	24306	for(i=nColumn, j=0; i<nTotal; i++, j++){
23076	24307	if( j==0 && p->cMode!=MODE_Column ){
23077		- oputz(p->cMode==MODE_Box?BOX_13" ":"\| ");
	24308	+ sqlite3_fputs(p->cMode==MODE_Box?BOX_13" ":"\| ", p->out);
23078	24309	}
23079	24310	z = azData[i];
23080	24311	if( z==0 ) z = p->nullValue;
23081	24312	w = p->actualWidth[j];
23082	24313	if( p->colWidth[j]<0 ) w = -w;
23083		- utf8_width_print(w, z);
	24314	+ utf8_width_print(p->out, w, z);
23084	24315	if( j==nColumn-1 ){
23085		- oputz(rowSep);
	24316	+ sqlite3_fputs(rowSep, p->out);
23086	24317	if( bMultiLineRowExists && abRowDiv[i/nColumn-1] && i+1<nTotal ){
23087	24318	if( p->cMode==MODE_Table ){
23088	24319	print_row_separator(p, nColumn, "+");
23089	24320	}else if( p->cMode==MODE_Box ){
23090	24321	print_box_row_separator(p, nColumn, BOX_123, BOX_1234, BOX_134);
23091	24322	}else if( p->cMode==MODE_Column ){
23092		- oputz("\n");
	24323	+ sqlite3_fputs("\n", p->out);
23093	24324	}
23094	24325	}
23095	24326	j = -1;
23096	24327	if( seenInterrupt ) goto columnar_end;
23097	24328	}else{
23098		- oputz(colSep);
	24329	+ sqlite3_fputs(colSep, p->out);
23099	24330	}
23100	24331	}
23101	24332	if( p->cMode==MODE_Table ){
23102	24333	print_row_separator(p, nColumn, "+");
23103	24334	}else if( p->cMode==MODE_Box ){
23104	24335	print_box_row_separator(p, nColumn, BOX_12, BOX_124, BOX_14);
23105	24336	}
23106	24337	columnar_end:
23107	24338	if( seenInterrupt ){
23108		- oputz("Interrupt\n");
	24339	+ sqlite3_fputs("Interrupt\n", p->out);
23109	24340	}
23110	24341	nData = (nRow+1)*nColumn;
23111	24342	for(i=0; i<nData; i++){
23112	24343	z = azData[i];
23113	24344	if( z!=zEmpty && z!=zShowNull ) free(azData[i]);
		@@ -23190,11 +24421,13 @@
23190	24421	}
23191	24422	}
23192	24423	} while( SQLITE_ROW == rc );
23193	24424	sqlite3_free(pData);
23194	24425	if( pArg->cMode==MODE_Json ){
23195		- fputs("]\n", pArg->out);
	24426	+ sqlite3_fputs("]\n", pArg->out);
	24427	+ }else if( pArg->cMode==MODE_Www ){
	24428	+ sqlite3_fputs("</TABLE>\n<PRE>\n", pArg->out);
23196	24429	}else if( pArg->cMode==MODE_Count ){
23197	24430	char zBuf[200];
23198	24431	sqlite3_snprintf(sizeof(zBuf), zBuf, "%llu row%s\n",
23199	24432	nRow, nRow!=1 ? "s" : "");
23200	24433	printf("%s", zBuf);
		@@ -23239,10 +24472,11 @@
23239	24472	int bCancel,
23240	24473	char **pzErr
23241	24474	){
23242	24475	int rc = SQLITE_OK;
23243	24476	sqlite3expert *p = pState->expert.pExpert;
	24477	+ FILE *out = pState->out;
23244	24478	assert( p );
23245	24479	assert( bCancel \|\| pzErr==0 \|\| *pzErr==0 );
23246	24480	if( bCancel==0 ){
23247	24481	int bVerbose = pState->expert.bVerbose;
23248	24482
		@@ -23251,24 +24485,25 @@
23251	24485	int nQuery = sqlite3_expert_count(p);
23252	24486	int i;
23253	24487
23254	24488	if( bVerbose ){
23255	24489	const char *zCand = sqlite3_expert_report(p,0,EXPERT_REPORT_CANDIDATES);
23256		- oputz("-- Candidates -----------------------------\n");
23257		- oputf("%s\n", zCand);
	24490	+ sqlite3_fputs("-- Candidates -----------------------------\n", out);
	24491	+ sqlite3_fprintf(out, "%s\n", zCand);
23258	24492	}
23259	24493	for(i=0; i<nQuery; i++){
23260	24494	const char *zSql = sqlite3_expert_report(p, i, EXPERT_REPORT_SQL);
23261	24495	const char *zIdx = sqlite3_expert_report(p, i, EXPERT_REPORT_INDEXES);
23262	24496	const char *zEQP = sqlite3_expert_report(p, i, EXPERT_REPORT_PLAN);
23263	24497	if( zIdx==0 ) zIdx = "(no new indexes)\n";
23264	24498	if( bVerbose ){
23265		- oputf("-- Query %d --------------------------------\n",i+1);
23266		- oputf("%s\n\n", zSql);
	24499	+ sqlite3_fprintf(out,
	24500	+ "-- Query %d --------------------------------\n"
	24501	+ "%s\n\n"
	24502	+ ,i+1, zSql);
23267	24503	}
23268		- oputf("%s\n", zIdx);
23269		- oputf("%s\n", zEQP);
	24504	+ sqlite3_fprintf(out, "%s\n%s\n", zIdx, zEQP);
23270	24505	}
23271	24506	}
23272	24507	}
23273	24508	sqlite3_expert_destroy(p);
23274	24509	pState->expert.pExpert = 0;
		@@ -23299,30 +24534,31 @@
23299	24534	if( n>=2 && 0==cli_strncmp(z, "-verbose", n) ){
23300	24535	pState->expert.bVerbose = 1;
23301	24536	}
23302	24537	else if( n>=2 && 0==cli_strncmp(z, "-sample", n) ){
23303	24538	if( i==(nArg-1) ){
23304		- eputf("option requires an argument: %s\n", z);
	24539	+ sqlite3_fprintf(stderr, "option requires an argument: %s\n", z);
23305	24540	rc = SQLITE_ERROR;
23306	24541	}else{
23307	24542	iSample = (int)integerValue(azArg[++i]);
23308	24543	if( iSample<0 \|\| iSample>100 ){
23309		- eputf("value out of range: %s\n", azArg[i]);
	24544	+ sqlite3_fprintf(stderr,"value out of range: %s\n", azArg[i]);
23310	24545	rc = SQLITE_ERROR;
23311	24546	}
23312	24547	}
23313	24548	}
23314	24549	else{
23315		- eputf("unknown option: %s\n", z);
	24550	+ sqlite3_fprintf(stderr,"unknown option: %s\n", z);
23316	24551	rc = SQLITE_ERROR;
23317	24552	}
23318	24553	}
23319	24554
23320	24555	if( rc==SQLITE_OK ){
23321	24556	pState->expert.pExpert = sqlite3_expert_new(pState->db, &zErr);
23322	24557	if( pState->expert.pExpert==0 ){
23323		- eputf("sqlite3_expert_new: %s\n", zErr ? zErr : "out of memory");
	24558	+ sqlite3_fprintf(stderr,
	24559	+ "sqlite3_expert_new: %s\n", zErr ? zErr : "out of memory");
23324	24560	rc = SQLITE_ERROR;
23325	24561	}else{
23326	24562	sqlite3_expert_config(
23327	24563	pState->expert.pExpert, EXPERT_CONFIG_SAMPLE, iSample
23328	24564	);
		@@ -23647,33 +24883,33 @@
23647	24883	if( zType==0 ) return 0;
23648	24884	dataOnly = (p->shellFlgs & SHFLG_DumpDataOnly)!=0;
23649	24885	noSys = (p->shellFlgs & SHFLG_DumpNoSys)!=0;
23650	24886
23651	24887	if( cli_strcmp(zTable, "sqlite_sequence")==0 && !noSys ){
23652		- if( !dataOnly ) oputz("DELETE FROM sqlite_sequence;\n");
	24888	+ /* no-op */
23653	24889	}else if( sqlite3_strglob("sqlite_stat?", zTable)==0 && !noSys ){
23654		- if( !dataOnly ) oputz("ANALYZE sqlite_schema;\n");
	24890	+ if( !dataOnly ) sqlite3_fputs("ANALYZE sqlite_schema;\n", p->out);
23655	24891	}else if( cli_strncmp(zTable, "sqlite_", 7)==0 ){
23656	24892	return 0;
23657	24893	}else if( dataOnly ){
23658	24894	/* no-op */
23659	24895	}else if( cli_strncmp(zSql, "CREATE VIRTUAL TABLE", 20)==0 ){
23660	24896	char *zIns;
23661	24897	if( !p->writableSchema ){
23662		- oputz("PRAGMA writable_schema=ON;\n");
	24898	+ sqlite3_fputs("PRAGMA writable_schema=ON;\n", p->out);
23663	24899	p->writableSchema = 1;
23664	24900	}
23665	24901	zIns = sqlite3_mprintf(
23666	24902	"INSERT INTO sqlite_schema(type,name,tbl_name,rootpage,sql)"
23667	24903	"VALUES('table','%q','%q',0,'%q');",
23668	24904	zTable, zTable, zSql);
23669	24905	shell_check_oom(zIns);
23670		- oputf("%s\n", zIns);
	24906	+ sqlite3_fprintf(p->out, "%s\n", zIns);
23671	24907	sqlite3_free(zIns);
23672	24908	return 0;
23673	24909	}else{
23674		- printSchemaLine(zSql, ";\n");
	24910	+ printSchemaLine(p->out, zSql, ";\n");
23675	24911	}
23676	24912
23677	24913	if( cli_strcmp(zType, "table")==0 ){
23678	24914	ShellText sSelect;
23679	24915	ShellText sTable;
		@@ -23727,11 +24963,11 @@
23727	24963	savedMode = p->mode;
23728	24964	p->zDestTable = sTable.z;
23729	24965	p->mode = p->cMode = MODE_Insert;
23730	24966	rc = shell_exec(p, sSelect.z, 0);
23731	24967	if( (rc&0xff)==SQLITE_CORRUPT ){
23732		- oputz("/**** CORRUPTION ERROR *****/\n");
	24968	+ sqlite3_fputs("/**** CORRUPTION ERROR *****/\n", p->out);
23733	24969	toggleSelectOrder(p->db);
23734	24970	shell_exec(p, sSelect.z, 0);
23735	24971	toggleSelectOrder(p->db);
23736	24972	}
23737	24973	p->zDestTable = savedDestTable;
		@@ -23758,28 +24994,28 @@
23758	24994	char *zErr = 0;
23759	24995	rc = sqlite3_exec(p->db, zQuery, dump_callback, p, &zErr);
23760	24996	if( rc==SQLITE_CORRUPT ){
23761	24997	char *zQ2;
23762	24998	int len = strlen30(zQuery);
23763		- oputz("/**** CORRUPTION ERROR *****/\n");
	24999	+ sqlite3_fputs("/**** CORRUPTION ERROR *****/\n", p->out);
23764	25000	if( zErr ){
23765		- oputf("/**** %s ****/\n", zErr);
	25001	+ sqlite3_fprintf(p->out, "/**** %s ****/\n", zErr);
23766	25002	sqlite3_free(zErr);
23767	25003	zErr = 0;
23768	25004	}
23769	25005	zQ2 = malloc( len+100 );
23770	25006	if( zQ2==0 ) return rc;
23771	25007	sqlite3_snprintf(len+100, zQ2, "%s ORDER BY rowid DESC", zQuery);
23772	25008	rc = sqlite3_exec(p->db, zQ2, dump_callback, p, &zErr);
23773	25009	if( rc ){
23774		- oputf("/**** ERROR: %s ****/\n", zErr);
	25010	+ sqlite3_fprintf(p->out, "/**** ERROR: %s ****/\n", zErr);
23775	25011	}else{
23776	25012	rc = SQLITE_CORRUPT;
23777	25013	}
23778		- sqlite3_free(zErr);
23779	25014	free(zQ2);
23780	25015	}
	25016	+ sqlite3_free(zErr);
23781	25017	return rc;
23782	25018	}
23783	25019
23784	25020	/*
23785	25021	** Text of help messages.
		@@ -23832,18 +25068,17 @@
23832	25068	#ifndef SQLITE_SHELL_FIDDLE
23833	25069	".check GLOB Fail if output since .testcase does not match",
23834	25070	".clone NEWDB Clone data into NEWDB from the existing database",
23835	25071	#endif
23836	25072	".connection [close] [#] Open or close an auxiliary database connection",
23837		-#if defined(_WIN32) \|\| defined(WIN32)
23838		- ".crnl on\|off Translate \\n to \\r\\n. Default ON",
23839		-#endif
	25073	+ ".crlf ?on\|off? Whether or not to use \\r\\n line endings",
23840	25074	".databases List names and files of attached databases",
23841	25075	".dbconfig ?op? ?val? List or change sqlite3_db_config() options",
23842	25076	#if SQLITE_SHELL_HAVE_RECOVER
23843	25077	".dbinfo ?DB? Show status information about the database",
23844	25078	#endif
	25079	+ ".dbtotxt Hex dump of the database file",
23845	25080	".dump ?OBJECTS? Render database content as SQL",
23846	25081	" Options:",
23847	25082	" --data-only Output only INSERT statements",
23848	25083	" --newlines Allow unescaped newline characters in output",
23849	25084	" --nosys Omit system tables (ex: \"sqlite_stat1\")",
		@@ -23940,13 +25175,15 @@
23940	25175	#endif
23941	25176	".nullvalue STRING Use STRING in place of NULL values",
23942	25177	#ifndef SQLITE_SHELL_FIDDLE
23943	25178	".once ?OPTIONS? ?FILE? Output for the next SQL command only to FILE",
23944	25179	" If FILE begins with '\|' then open as a pipe",
23945		- " --bom Put a UTF8 byte-order mark at the beginning",
23946		- " -e Send output to the system text editor",
23947		- " -x Send output as CSV to a spreadsheet (same as \".excel\")",
	25180	+ " --bom Put a UTF8 byte-order mark at the beginning",
	25181	+ " -e Send output to the system text editor",
	25182	+ " --plain Use text/plain output instead of HTML for -w option",
	25183	+ " -w Send output as HTML to a web browser (same as \".www\")",
	25184	+ " -x Send output as CSV to a spreadsheet (same as \".excel\")",
23948	25185	/* Note that .open is (partially) available in WASM builds but is
23949	25186	** currently only intended to be used by the fiddle tool, not
23950	25187	** end users, so is "undocumented." */
23951	25188	".open ?OPTIONS? ?FILE? Close existing database and reopen FILE",
23952	25189	" Options:",
		@@ -23965,10 +25202,12 @@
23965	25202	".output ?FILE? Send output to FILE or stdout if FILE is omitted",
23966	25203	" If FILE begins with '\|' then open it as a pipe.",
23967	25204	" Options:",
23968	25205	" --bom Prefix output with a UTF8 byte-order mark",
23969	25206	" -e Send output to the system text editor",
	25207	+ " --plain Use text/plain for -w option",
	25208	+ " -w Send output to a web browser",
23970	25209	" -x Send output as CSV to a spreadsheet",
23971	25210	#endif
23972	25211	".parameter CMD ... Manage SQL parameter bindings",
23973	25212	" clear Erase all bindings",
23974	25213	" init Initialize the TEMP table that holds bindings",
		@@ -24078,10 +25317,14 @@
24078	25317	".vfsinfo ?AUX? Information about the top-level VFS",
24079	25318	".vfslist List all available VFSes",
24080	25319	".vfsname ?AUX? Print the name of the VFS stack",
24081	25320	".width NUM1 NUM2 ... Set minimum column widths for columnar output",
24082	25321	" Negative values right-justify",
	25322	+#ifndef SQLITE_SHELL_FIDDLE
	25323	+ ".www Display output of the next command in web browser",
	25324	+ " --plain Show results as text/plain, not as HTML",
	25325	+#endif
24083	25326	};
24084	25327
24085	25328	/*
24086	25329	** Output help text.
24087	25330	**
		@@ -24126,24 +25369,24 @@
24126	25369	hh &= ~HH_Summary;
24127	25370	break;
24128	25371	}
24129	25372	if( ((hw^hh)&HH_Undoc)==0 ){
24130	25373	if( (hh&HH_Summary)!=0 ){
24131		- sputf(out, ".%s\n", azHelp[i]+1);
	25374	+ sqlite3_fprintf(out, ".%s\n", azHelp[i]+1);
24132	25375	++n;
24133	25376	}else if( (hw&HW_SummaryOnly)==0 ){
24134		- sputf(out, "%s\n", azHelp[i]);
	25377	+ sqlite3_fprintf(out, "%s\n", azHelp[i]);
24135	25378	}
24136	25379	}
24137	25380	}
24138	25381	}else{
24139	25382	/* Seek documented commands for which zPattern is an exact prefix */
24140	25383	zPat = sqlite3_mprintf(".%s*", zPattern);
24141	25384	shell_check_oom(zPat);
24142	25385	for(i=0; i<ArraySize(azHelp); i++){
24143	25386	if( sqlite3_strglob(zPat, azHelp[i])==0 ){
24144		- sputf(out, "%s\n", azHelp[i]);
	25387	+ sqlite3_fprintf(out, "%s\n", azHelp[i]);
24145	25388	j = i+1;
24146	25389	n++;
24147	25390	}
24148	25391	}
24149	25392	sqlite3_free(zPat);
		@@ -24150,11 +25393,11 @@
24150	25393	if( n ){
24151	25394	if( n==1 ){
24152	25395	/* when zPattern is a prefix of exactly one command, then include
24153	25396	** the details of that command, which should begin at offset j */
24154	25397	while( j<ArraySize(azHelp)-1 && azHelp[j][0]==' ' ){
24155		- sputf(out, "%s\n", azHelp[j]);
	25398	+ sqlite3_fprintf(out, "%s\n", azHelp[j]);
24156	25399	j++;
24157	25400	}
24158	25401	}
24159	25402	return n;
24160	25403	}
		@@ -24167,14 +25410,14 @@
24167	25410	while( i<ArraySize(azHelp)-1 && azHelp[i+1][0]==' ' ) ++i;
24168	25411	continue;
24169	25412	}
24170	25413	if( azHelp[i][0]=='.' ) j = i;
24171	25414	if( sqlite3_strlike(zPat, azHelp[i], 0)==0 ){
24172		- sputf(out, "%s\n", azHelp[j]);
	25415	+ sqlite3_fprintf(out, "%s\n", azHelp[j]);
24173	25416	while( j<ArraySize(azHelp)-1 && azHelp[j+1][0]==' ' ){
24174	25417	j++;
24175		- sputf(out, "%s\n", azHelp[j]);
	25418	+ sqlite3_fprintf(out, "%s\n", azHelp[j]);
24176	25419	}
24177	25420	i = j;
24178	25421	n++;
24179	25422	}
24180	25423	}
		@@ -24200,35 +25443,35 @@
24200	25443	**
24201	25444	** NULL is returned if any error is encountered. The final value of *pnByte
24202	25445	** is undefined in this case.
24203	25446	*/
24204	25447	static char readFile(const char zName, int *pnByte){
24205		- FILE *in = fopen(zName, "rb");
	25448	+ FILE *in = sqlite3_fopen(zName, "rb");
24206	25449	long nIn;
24207	25450	size_t nRead;
24208	25451	char *pBuf;
24209	25452	int rc;
24210	25453	if( in==0 ) return 0;
24211	25454	rc = fseek(in, 0, SEEK_END);
24212	25455	if( rc!=0 ){
24213		- eputf("Error: '%s' not seekable\n", zName);
	25456	+ sqlite3_fprintf(stderr,"Error: '%s' not seekable\n", zName);
24214	25457	fclose(in);
24215	25458	return 0;
24216	25459	}
24217	25460	nIn = ftell(in);
24218	25461	rewind(in);
24219	25462	pBuf = sqlite3_malloc64( nIn+1 );
24220	25463	if( pBuf==0 ){
24221		- eputz("Error: out of memory\n");
	25464	+ sqlite3_fputs("Error: out of memory\n", stderr);
24222	25465	fclose(in);
24223	25466	return 0;
24224	25467	}
24225	25468	nRead = fread(pBuf, nIn, 1, in);
24226	25469	fclose(in);
24227	25470	if( nRead!=1 ){
24228	25471	sqlite3_free(pBuf);
24229		- eputf("Error: cannot read '%s'\n", zName);
	25472	+ sqlite3_fprintf(stderr,"Error: cannot read '%s'\n", zName);
24230	25473	return 0;
24231	25474	}
24232	25475	pBuf[nIn] = 0;
24233	25476	if( pnByte ) *pnByte = nIn;
24234	25477	return pBuf;
		@@ -24290,11 +25533,11 @@
24290	25533	** archive and the dfltZip flag is true, then assume it is a ZIP archive.
24291	25534	** Otherwise, assume an ordinary database regardless of the filename if
24292	25535	** the type cannot be determined from content.
24293	25536	*/
24294	25537	int deduceDatabaseType(const char *zName, int dfltZip){
24295		- FILE *f = fopen(zName, "rb");
	25538	+ FILE *f = sqlite3_fopen(zName, "rb");
24296	25539	size_t n;
24297	25540	int rc = SHELL_OPEN_UNSPEC;
24298	25541	char zBuf[100];
24299	25542	if( f==0 ){
24300	25543	if( dfltZip && sqlite3_strlike("%.zip",zName,0)==0 ){
		@@ -24343,13 +25586,13 @@
24343	25586	FILE *in;
24344	25587	const char *zDbFilename = p->pAuxDb->zDbFilename;
24345	25588	unsigned int x[16];
24346	25589	char zLine[1000];
24347	25590	if( zDbFilename ){
24348		- in = fopen(zDbFilename, "r");
	25591	+ in = sqlite3_fopen(zDbFilename, "r");
24349	25592	if( in==0 ){
24350		- eputf("cannot open \"%s\" for reading\n", zDbFilename);
	25593	+ sqlite3_fprintf(stderr,"cannot open \"%s\" for reading\n", zDbFilename);
24351	25594	return 0;
24352	25595	}
24353	25596	nLine = 0;
24354	25597	}else{
24355	25598	in = p->in;
		@@ -24356,24 +25599,24 @@
24356	25599	nLine = p->lineno;
24357	25600	if( in==0 ) in = stdin;
24358	25601	}
24359	25602	*pnData = 0;
24360	25603	nLine++;
24361		- if( fgets(zLine, sizeof(zLine), in)==0 ) goto readHexDb_error;
	25604	+ if( sqlite3_fgets(zLine, sizeof(zLine), in)==0 ) goto readHexDb_error;
24362	25605	rc = sscanf(zLine, "\| size %d pagesize %d", &n, &pgsz);
24363	25606	if( rc!=2 ) goto readHexDb_error;
24364	25607	if( n<0 ) goto readHexDb_error;
24365	25608	if( pgsz<512 \|\| pgsz>65536 \|\| (pgsz&(pgsz-1))!=0 ) goto readHexDb_error;
24366	25609	n = (n+pgsz-1)&~(pgsz-1); /* Round n up to the next multiple of pgsz */
24367	25610	a = sqlite3_malloc( n ? n : 1 );
24368	25611	shell_check_oom(a);
24369	25612	memset(a, 0, n);
24370	25613	if( pgsz<512 \|\| pgsz>65536 \|\| (pgsz & (pgsz-1))!=0 ){
24371		- eputz("invalid pagesize\n");
	25614	+ sqlite3_fputs("invalid pagesize\n", stderr);
24372	25615	goto readHexDb_error;
24373	25616	}
24374		- for(nLine++; fgets(zLine, sizeof(zLine), in)!=0; nLine++){
	25617	+ for(nLine++; sqlite3_fgets(zLine, sizeof(zLine), in)!=0; nLine++){
24375	25618	rc = sscanf(zLine, "\| page %d offset %d", &j, &k);
24376	25619	if( rc==2 ){
24377	25620	iOffset = k;
24378	25621	continue;
24379	25622	}
		@@ -24401,18 +25644,18 @@
24401	25644
24402	25645	readHexDb_error:
24403	25646	if( in!=p->in ){
24404	25647	fclose(in);
24405	25648	}else{
24406		- while( fgets(zLine, sizeof(zLine), p->in)!=0 ){
	25649	+ while( sqlite3_fgets(zLine, sizeof(zLine), p->in)!=0 ){
24407	25650	nLine++;
24408	25651	if(cli_strncmp(zLine, "\| end ", 6)==0 ) break;
24409	25652	}
24410	25653	p->lineno = nLine;
24411	25654	}
24412	25655	sqlite3_free(a);
24413		- eputf("Error on line %d of --hexdb input\n", nLine);
	25656	+ sqlite3_fprintf(stderr,"Error on line %d of --hexdb input\n", nLine);
24414	25657	return 0;
24415	25658	}
24416	25659	#endif /* SQLITE_OMIT_DESERIALIZE */
24417	25660
24418	25661	/*
		@@ -24483,22 +25726,24 @@
24483	25726	SQLITE_OPEN_READWRITE\|SQLITE_OPEN_CREATE\|p->openFlags, 0);
24484	25727	break;
24485	25728	}
24486	25729	}
24487	25730	if( p->db==0 \|\| SQLITE_OK!=sqlite3_errcode(p->db) ){
24488		- eputf("Error: unable to open database \"%s\": %s\n",
	25731	+ sqlite3_fprintf(stderr,"Error: unable to open database \"%s\": %s\n",
24489	25732	zDbFilename, sqlite3_errmsg(p->db));
24490	25733	if( (openFlags & OPEN_DB_KEEPALIVE)==0 ){
24491	25734	exit(1);
24492	25735	}
24493	25736	sqlite3_close(p->db);
24494	25737	sqlite3_open(":memory:", &p->db);
24495	25738	if( p->db==0 \|\| SQLITE_OK!=sqlite3_errcode(p->db) ){
24496		- eputz("Also: unable to open substitute in-memory database.\n");
	25739	+ sqlite3_fputs("Also: unable to open substitute in-memory database.\n",
	25740	+ stderr);
24497	25741	exit(1);
24498	25742	}else{
24499		- eputf("Notice: using substitute in-memory database instead of \"%s\"\n",
	25743	+ sqlite3_fprintf(stderr,
	25744	+ "Notice: using substitute in-memory database instead of \"%s\"\n",
24500	25745	zDbFilename);
24501	25746	}
24502	25747	}
24503	25748	globalDb = p->db;
24504	25749	sqlite3_db_config(p->db, SQLITE_DBCONFIG_STMT_SCANSTATUS, (int)0, (int*)0);
		@@ -24511,10 +25756,11 @@
24511	25756	}
24512	25757
24513	25758	#ifndef SQLITE_OMIT_LOAD_EXTENSION
24514	25759	sqlite3_enable_load_extension(p->db, 1);
24515	25760	#endif
	25761	+ sqlite3_sha_init(p->db, 0, 0);
24516	25762	sqlite3_shathree_init(p->db, 0, 0);
24517	25763	sqlite3_uint_init(p->db, 0, 0);
24518	25764	sqlite3_stmtrand_init(p->db, 0, 0);
24519	25765	sqlite3_decimal_init(p->db, 0, 0);
24520	25766	sqlite3_percentile_init(p->db, 0, 0);
		@@ -24605,11 +25851,11 @@
24605	25851	}
24606	25852	rc = sqlite3_deserialize(p->db, "main", aData, nData, nData,
24607	25853	SQLITE_DESERIALIZE_RESIZEABLE \|
24608	25854	SQLITE_DESERIALIZE_FREEONCLOSE);
24609	25855	if( rc ){
24610		- eputf("Error: sqlite3_deserialize() returns %d\n", rc);
	25856	+ sqlite3_fprintf(stderr,"Error: sqlite3_deserialize() returns %d\n", rc);
24611	25857	}
24612	25858	if( p->szMax>0 ){
24613	25859	sqlite3_file_control(p->db, "main", SQLITE_FCNTL_SIZE_LIMIT, &p->szMax);
24614	25860	}
24615	25861	}
		@@ -24629,15 +25875,17 @@
24629	25875	** Attempt to close the database connection. Report errors.
24630	25876	*/
24631	25877	void close_db(sqlite3 *db){
24632	25878	int rc = sqlite3_close(db);
24633	25879	if( rc ){
24634		- eputf("Error: sqlite3_close() returns %d: %s\n", rc, sqlite3_errmsg(db));
	25880	+ sqlite3_fprintf(stderr,
	25881	+ "Error: sqlite3_close() returns %d: %s\n", rc, sqlite3_errmsg(db));
24635	25882	}
24636	25883	}
24637	25884
24638		-#if HAVE_READLINE \|\| HAVE_EDITLINE
	25885	+#if (HAVE_READLINE \|\| HAVE_EDITLINE) \
	25886	+ && !defined(SQLITE_OMIT_READLINE_COMPLETION)
24639	25887	/*
24640	25888	** Readline completion callbacks
24641	25889	*/
24642	25890	static char readline_completion_generator(const char text, int state){
24643	25891	static sqlite3_stmt *pStmt = 0;
		@@ -24671,19 +25919,26 @@
24671	25919	#elif HAVE_LINENOISE
24672	25920	/*
24673	25921	** Linenoise completion callback. Note that the 3rd argument is from
24674	25922	** the "msteveb" version of linenoise, not the "antirez" version.
24675	25923	*/
24676		-static void linenoise_completion(const char zLine, linenoiseCompletions lc,
24677		- void *pUserData){
	25924	+static void linenoise_completion(
	25925	+ const char *zLine,
	25926	+ linenoiseCompletions *lc
	25927	+#if HAVE_LINENOISE==2
	25928	+ ,void *pUserData
	25929	+#endif
	25930	+){
24678	25931	i64 nLine = strlen(zLine);
24679	25932	i64 i, iStart;
24680	25933	sqlite3_stmt *pStmt = 0;
24681	25934	char *zSql;
24682	25935	char zBuf[1000];
24683	25936
	25937	+#if HAVE_LINENOISE==2
24684	25938	UNUSED_PARAMETER(pUserData);
	25939	+#endif
24685	25940	if( nLine>(i64)sizeof(zBuf)-30 ) return;
24686	25941	if( zLine[0]=='.' \|\| zLine[0]=='#') return;
24687	25942	for(i=nLine-1; i>=0 && (isalnum(zLine[i]) \|\| zLine[i]=='_'); i--){}
24688	25943	if( i==nLine-1 ) return;
24689	25944	iStart = i+1;
		@@ -24793,11 +26048,12 @@
24793	26048	return 1;
24794	26049	}
24795	26050	if( sqlite3_stricmp(zArg, "off")==0 \|\| sqlite3_stricmp(zArg,"no")==0 ){
24796	26051	return 0;
24797	26052	}
24798		- eputf("ERROR: Not a boolean value: \"%s\". Assuming \"no\".\n", zArg);
	26053	+ sqlite3_fprintf(stderr,
	26054	+ "ERROR: Not a boolean value: \"%s\". Assuming \"no\".\n", zArg);
24799	26055	return 0;
24800	26056	}
24801	26057
24802	26058	/*
24803	26059	** Set or clear a shell flag according to a boolean value.
		@@ -24820,22 +26076,22 @@
24820	26076	/*
24821	26077	** Try to open an output file. The names "stdout" and "stderr" are
24822	26078	** recognized and do the right thing. NULL is returned if the output
24823	26079	** filename is "off".
24824	26080	*/
24825		-static FILE output_file_open(const char zFile, int bTextMode){
	26081	+static FILE output_file_open(const char zFile){
24826	26082	FILE *f;
24827	26083	if( cli_strcmp(zFile,"stdout")==0 ){
24828	26084	f = stdout;
24829	26085	}else if( cli_strcmp(zFile, "stderr")==0 ){
24830	26086	f = stderr;
24831	26087	}else if( cli_strcmp(zFile, "off")==0 ){
24832	26088	f = 0;
24833	26089	}else{
24834		- f = fopen(zFile, bTextMode ? "w" : "wb");
	26090	+ f = sqlite3_fopen(zFile, "w");
24835	26091	if( f==0 ){
24836		- eputf("Error: cannot open \"%s\"\n", zFile);
	26092	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zFile);
24837	26093	}
24838	26094	}
24839	26095	return f;
24840	26096	}
24841	26097
		@@ -24884,16 +26140,17 @@
24884	26140	if( nSql>1000000000 ) nSql = 1000000000;
24885	26141	while( nSql>0 && zSql[nSql-1]==';' ){ nSql--; }
24886	26142	switch( mType ){
24887	26143	case SQLITE_TRACE_ROW:
24888	26144	case SQLITE_TRACE_STMT: {
24889		- sputf(p->traceOut, "%.*s;\n", (int)nSql, zSql);
	26145	+ sqlite3_fprintf(p->traceOut, "%.*s;\n", (int)nSql, zSql);
24890	26146	break;
24891	26147	}
24892	26148	case SQLITE_TRACE_PROFILE: {
24893	26149	sqlite3_int64 nNanosec = pX ? (sqlite3_int64)pX : 0;
24894		- sputf(p->traceOut, "%.*s; -- %lld ns\n", (int)nSql, zSql, nNanosec);
	26150	+ sqlite3_fprintf(p->traceOut,
	26151	+ "%.*s; -- %lld ns\n", (int)nSql, zSql, nNanosec);
24895	26152	break;
24896	26153	}
24897	26154	}
24898	26155	return 0;
24899	26156	}
		@@ -24996,14 +26253,15 @@
24996	26253	do{ p->n--; }while( p->z[p->n]!=cQuote );
24997	26254	p->cTerm = c;
24998	26255	break;
24999	26256	}
25000	26257	if( pc==cQuote && c!='\r' ){
25001		- eputf("%s:%d: unescaped %c character\n", p->zFile, p->nLine, cQuote);
	26258	+ sqlite3_fprintf(stderr,"%s:%d: unescaped %c character\n",
	26259	+ p->zFile, p->nLine, cQuote);
25002	26260	}
25003	26261	if( c==EOF ){
25004		- eputf("%s:%d: unterminated %c-quoted field\n",
	26262	+ sqlite3_fprintf(stderr,"%s:%d: unterminated %c-quoted field\n",
25005	26263	p->zFile, startLine, cQuote);
25006	26264	p->cTerm = c;
25007	26265	break;
25008	26266	}
25009	26267	import_append_char(p, c);
		@@ -25098,11 +26356,11 @@
25098	26356
25099	26357	zQuery = sqlite3_mprintf("SELECT * FROM \"%w\"", zTable);
25100	26358	shell_check_oom(zQuery);
25101	26359	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25102	26360	if( rc ){
25103		- eputf("Error %d: %s on [%s]\n",
	26361	+ sqlite3_fprintf(stderr,"Error %d: %s on [%s]\n",
25104	26362	sqlite3_extended_errcode(p->db), sqlite3_errmsg(p->db), zQuery);
25105	26363	goto end_data_xfer;
25106	26364	}
25107	26365	n = sqlite3_column_count(pQuery);
25108	26366	zInsert = sqlite3_malloc64(200 + nTable + n*3);
		@@ -25115,11 +26373,11 @@
25115	26373	i += 2;
25116	26374	}
25117	26375	memcpy(zInsert+i, ");", 3);
25118	26376	rc = sqlite3_prepare_v2(newDb, zInsert, -1, &pInsert, 0);
25119	26377	if( rc ){
25120		- eputf("Error %d: %s on [%s]\n",
	26378	+ sqlite3_fprintf(stderr,"Error %d: %s on [%s]\n",
25121	26379	sqlite3_extended_errcode(newDb), sqlite3_errmsg(newDb), zInsert);
25122	26380	goto end_data_xfer;
25123	26381	}
25124	26382	for(k=0; k<2; k++){
25125	26383	while( (rc = sqlite3_step(pQuery))==SQLITE_ROW ){
		@@ -25151,11 +26409,11 @@
25151	26409	}
25152	26410	}
25153	26411	} /* End for */
25154	26412	rc = sqlite3_step(pInsert);
25155	26413	if( rc!=SQLITE_OK && rc!=SQLITE_ROW && rc!=SQLITE_DONE ){
25156		- eputf("Error %d: %s\n",
	26414	+ sqlite3_fprintf(stderr,"Error %d: %s\n",
25157	26415	sqlite3_extended_errcode(newDb), sqlite3_errmsg(newDb));
25158	26416	}
25159	26417	sqlite3_reset(pInsert);
25160	26418	cnt++;
25161	26419	if( (cnt%spinRate)==0 ){
		@@ -25169,11 +26427,11 @@
25169	26427	zQuery = sqlite3_mprintf("SELECT * FROM \"%w\" ORDER BY rowid DESC;",
25170	26428	zTable);
25171	26429	shell_check_oom(zQuery);
25172	26430	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25173	26431	if( rc ){
25174		- eputf("Warning: cannot step \"%s\" backwards", zTable);
	26432	+ sqlite3_fprintf(stderr,"Warning: cannot step \"%s\" backwards", zTable);
25175	26433	break;
25176	26434	}
25177	26435	} /* End for(k=0...) */
25178	26436
25179	26437	end_data_xfer:
		@@ -25206,23 +26464,24 @@
25206	26464	zQuery = sqlite3_mprintf("SELECT name, sql FROM sqlite_schema"
25207	26465	" WHERE %s ORDER BY rowid ASC", zWhere);
25208	26466	shell_check_oom(zQuery);
25209	26467	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25210	26468	if( rc ){
25211		- eputf("Error: (%d) %s on [%s]\n", sqlite3_extended_errcode(p->db),
	26469	+ sqlite3_fprintf(stderr,
	26470	+ "Error: (%d) %s on [%s]\n", sqlite3_extended_errcode(p->db),
25212	26471	sqlite3_errmsg(p->db), zQuery);
25213	26472	goto end_schema_xfer;
25214	26473	}
25215	26474	while( (rc = sqlite3_step(pQuery))==SQLITE_ROW ){
25216	26475	zName = sqlite3_column_text(pQuery, 0);
25217	26476	zSql = sqlite3_column_text(pQuery, 1);
25218	26477	if( zName==0 \|\| zSql==0 ) continue;
25219	26478	if( sqlite3_stricmp((char*)zName, "sqlite_sequence")!=0 ){
25220		- sputf(stdout, "%s... ", zName); fflush(stdout);
	26479	+ sqlite3_fprintf(stdout, "%s... ", zName); fflush(stdout);
25221	26480	sqlite3_exec(newDb, (const char*)zSql, 0, 0, &zErrMsg);
25222	26481	if( zErrMsg ){
25223		- eputf("Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
	26482	+ sqlite3_fprintf(stderr,"Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
25224	26483	sqlite3_free(zErrMsg);
25225	26484	zErrMsg = 0;
25226	26485	}
25227	26486	}
25228	26487	if( xForEach ){
		@@ -25236,23 +26495,23 @@
25236	26495	zQuery = sqlite3_mprintf("SELECT name, sql FROM sqlite_schema"
25237	26496	" WHERE %s ORDER BY rowid DESC", zWhere);
25238	26497	shell_check_oom(zQuery);
25239	26498	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25240	26499	if( rc ){
25241		- eputf("Error: (%d) %s on [%s]\n",
	26500	+ sqlite3_fprintf(stderr,"Error: (%d) %s on [%s]\n",
25242	26501	sqlite3_extended_errcode(p->db), sqlite3_errmsg(p->db), zQuery);
25243	26502	goto end_schema_xfer;
25244	26503	}
25245	26504	while( sqlite3_step(pQuery)==SQLITE_ROW ){
25246	26505	zName = sqlite3_column_text(pQuery, 0);
25247	26506	zSql = sqlite3_column_text(pQuery, 1);
25248	26507	if( zName==0 \|\| zSql==0 ) continue;
25249	26508	if( sqlite3_stricmp((char*)zName, "sqlite_sequence")==0 ) continue;
25250		- sputf(stdout, "%s... ", zName); fflush(stdout);
	26509	+ sqlite3_fprintf(stdout, "%s... ", zName); fflush(stdout);
25251	26510	sqlite3_exec(newDb, (const char*)zSql, 0, 0, &zErrMsg);
25252	26511	if( zErrMsg ){
25253		- eputf("Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
	26512	+ sqlite3_fprintf(stderr,"Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
25254	26513	sqlite3_free(zErrMsg);
25255	26514	zErrMsg = 0;
25256	26515	}
25257	26516	if( xForEach ){
25258	26517	xForEach(p, newDb, (const char*)zName);
		@@ -25272,16 +26531,17 @@
25272	26531	*/
25273	26532	static void tryToClone(ShellState p, const char zNewDb){
25274	26533	int rc;
25275	26534	sqlite3 *newDb = 0;
25276	26535	if( access(zNewDb,0)==0 ){
25277		- eputf("File \"%s\" already exists.\n", zNewDb);
	26536	+ sqlite3_fprintf(stderr,"File \"%s\" already exists.\n", zNewDb);
25278	26537	return;
25279	26538	}
25280	26539	rc = sqlite3_open(zNewDb, &newDb);
25281	26540	if( rc ){
25282		- eputf("Cannot create output database: %s\n", sqlite3_errmsg(newDb));
	26541	+ sqlite3_fprintf(stderr,
	26542	+ "Cannot create output database: %s\n", sqlite3_errmsg(newDb));
25283	26543	}else{
25284	26544	sqlite3_exec(p->db, "PRAGMA writable_schema=ON;", 0, 0, 0);
25285	26545	sqlite3_exec(newDb, "BEGIN EXCLUSIVE;", 0, 0, 0);
25286	26546	tryToCloneSchema(p, newDb, "type='table'", tryToCloneData);
25287	26547	tryToCloneSchema(p, newDb, "type!='table'", 0);
		@@ -25294,14 +26554,22 @@
25294	26554	#ifndef SQLITE_SHELL_FIDDLE
25295	26555	/*
25296	26556	** Change the output stream (file or pipe or console) to something else.
25297	26557	*/
25298	26558	static void output_redir(ShellState p, FILE pfNew){
25299		- if( p->out != stdout ) eputz("Output already redirected.\n");
25300		- else{
	26559	+ if( p->out != stdout ){
	26560	+ sqlite3_fputs("Output already redirected.\n", stderr);
	26561	+ }else{
25301	26562	p->out = pfNew;
25302		- setOutputStream(pfNew);
	26563	+ setCrlfMode(p);
	26564	+ if( p->mode==MODE_Www ){
	26565	+ sqlite3_fputs(
	26566	+ "<!DOCTYPE html>\n"
	26567	+ "<HTML><BODY><PRE>\n",
	26568	+ p->out
	26569	+ );
	26570	+ }
25303	26571	}
25304	26572	}
25305	26573
25306	26574	/*
25307	26575	** Change the output file back to stdout.
		@@ -25314,10 +26582,13 @@
25314	26582	if( p->outfile[0]=='\|' ){
25315	26583	#ifndef SQLITE_OMIT_POPEN
25316	26584	pclose(p->out);
25317	26585	#endif
25318	26586	}else{
	26587	+ if( p->mode==MODE_Www ){
	26588	+ sqlite3_fputs("</PRE></BODY></HTML>\n", p->out);
	26589	+ }
25319	26590	output_file_close(p->out);
25320	26591	#ifndef SQLITE_NOHAVE_SYSTEM
25321	26592	if( p->doXdgOpen ){
25322	26593	const char *zXdgOpenCmd =
25323	26594	#if defined(_WIN32)
		@@ -25328,11 +26599,11 @@
25328	26599	"xdg-open";
25329	26600	#endif
25330	26601	char *zCmd;
25331	26602	zCmd = sqlite3_mprintf("%s %s", zXdgOpenCmd, p->zTempFile);
25332	26603	if( system(zCmd) ){
25333		- eputf("Failed: [%s]\n", zCmd);
	26604	+ sqlite3_fprintf(stderr,"Failed: [%s]\n", zCmd);
25334	26605	}else{
25335	26606	/* Give the start/open/xdg-open command some time to get
25336	26607	** going before we continue, and potential delete the
25337	26608	** p->zTempFile data file out from under it */
25338	26609	sqlite3_sleep(2000);
		@@ -25343,28 +26614,34 @@
25343	26614	}
25344	26615	#endif /* !defined(SQLITE_NOHAVE_SYSTEM) */
25345	26616	}
25346	26617	p->outfile[0] = 0;
25347	26618	p->out = stdout;
25348		- setOutputStream(stdout);
	26619	+ setCrlfMode(p);
25349	26620	}
25350	26621	#else
25351	26622	# define output_redir(SS,pfO)
25352	26623	# define output_reset(SS)
25353	26624	#endif
25354	26625
25355	26626	/*
25356	26627	** Run an SQL command and return the single integer result.
25357	26628	*/
25358		-static int db_int(sqlite3 db, const char zSql){
	26629	+static int db_int(sqlite3 db, const char zSql, ...){
25359	26630	sqlite3_stmt *pStmt;
25360	26631	int res = 0;
25361		- sqlite3_prepare_v2(db, zSql, -1, &pStmt, 0);
	26632	+ char *z;
	26633	+ va_list ap;
	26634	+ va_start(ap, zSql);
	26635	+ z = sqlite3_vmprintf(zSql, ap);
	26636	+ va_end(ap);
	26637	+ sqlite3_prepare_v2(db, z, -1, &pStmt, 0);
25362	26638	if( pStmt && sqlite3_step(pStmt)==SQLITE_ROW ){
25363	26639	res = sqlite3_column_int(pStmt,0);
25364	26640	}
25365	26641	sqlite3_finalize(pStmt);
	26642	+ sqlite3_free(z);
25366	26643	return res;
25367	26644	}
25368	26645
25369	26646	#if SQLITE_SHELL_HAVE_RECOVER
25370	26647	/*
		@@ -25419,11 +26696,11 @@
25419	26696	if( p->db==0 ) return 1;
25420	26697	rc = sqlite3_prepare_v2(p->db,
25421	26698	"SELECT data FROM sqlite_dbpage(?1) WHERE pgno=1",
25422	26699	-1, &pStmt, 0);
25423	26700	if( rc ){
25424		- eputf("error: %s\n", sqlite3_errmsg(p->db));
	26701	+ sqlite3_fprintf(stderr,"error: %s\n", sqlite3_errmsg(p->db));
25425	26702	sqlite3_finalize(pStmt);
25426	26703	return 1;
25427	26704	}
25428	26705	sqlite3_bind_text(pStmt, 1, zDb, -1, SQLITE_STATIC);
25429	26706	if( sqlite3_step(pStmt)==SQLITE_ROW
		@@ -25432,58 +26709,149 @@
25432	26709	const u8 *pb = sqlite3_column_blob(pStmt,0);
25433	26710	shell_check_oom(pb);
25434	26711	memcpy(aHdr, pb, 100);
25435	26712	sqlite3_finalize(pStmt);
25436	26713	}else{
25437		- eputz("unable to read database header\n");
	26714	+ sqlite3_fputs("unable to read database header\n", stderr);
25438	26715	sqlite3_finalize(pStmt);
25439	26716	return 1;
25440	26717	}
25441	26718	i = get2byteInt(aHdr+16);
25442	26719	if( i==1 ) i = 65536;
25443		- oputf("%-20s %d\n", "database page size:", i);
25444		- oputf("%-20s %d\n", "write format:", aHdr[18]);
25445		- oputf("%-20s %d\n", "read format:", aHdr[19]);
25446		- oputf("%-20s %d\n", "reserved bytes:", aHdr[20]);
	26720	+ sqlite3_fprintf(p->out, "%-20s %d\n", "database page size:", i);
	26721	+ sqlite3_fprintf(p->out, "%-20s %d\n", "write format:", aHdr[18]);
	26722	+ sqlite3_fprintf(p->out, "%-20s %d\n", "read format:", aHdr[19]);
	26723	+ sqlite3_fprintf(p->out, "%-20s %d\n", "reserved bytes:", aHdr[20]);
25447	26724	for(i=0; i<ArraySize(aField); i++){
25448	26725	int ofst = aField[i].ofst;
25449	26726	unsigned int val = get4byteInt(aHdr + ofst);
25450		- oputf("%-20s %u", aField[i].zName, val);
	26727	+ sqlite3_fprintf(p->out, "%-20s %u", aField[i].zName, val);
25451	26728	switch( ofst ){
25452	26729	case 56: {
25453		- if( val==1 ) oputz(" (utf8)");
25454		- if( val==2 ) oputz(" (utf16le)");
25455		- if( val==3 ) oputz(" (utf16be)");
	26730	+ if( val==1 ) sqlite3_fputs(" (utf8)", p->out);
	26731	+ if( val==2 ) sqlite3_fputs(" (utf16le)", p->out);
	26732	+ if( val==3 ) sqlite3_fputs(" (utf16be)", p->out);
25456	26733	}
25457	26734	}
25458		- oputz("\n");
	26735	+ sqlite3_fputs("\n", p->out);
25459	26736	}
25460	26737	if( zDb==0 ){
25461	26738	zSchemaTab = sqlite3_mprintf("main.sqlite_schema");
25462	26739	}else if( cli_strcmp(zDb,"temp")==0 ){
25463	26740	zSchemaTab = sqlite3_mprintf("%s", "sqlite_temp_schema");
25464	26741	}else{
25465	26742	zSchemaTab = sqlite3_mprintf("\"%w\".sqlite_schema", zDb);
25466	26743	}
25467	26744	for(i=0; i<ArraySize(aQuery); i++){
25468		- char *zSql = sqlite3_mprintf(aQuery[i].zSql, zSchemaTab);
25469		- int val = db_int(p->db, zSql);
25470		- sqlite3_free(zSql);
25471		- oputf("%-20s %d\n", aQuery[i].zName, val);
	26745	+ int val = db_int(p->db, aQuery[i].zSql, zSchemaTab);
	26746	+ sqlite3_fprintf(p->out, "%-20s %d\n", aQuery[i].zName, val);
25472	26747	}
25473	26748	sqlite3_free(zSchemaTab);
25474	26749	sqlite3_file_control(p->db, zDb, SQLITE_FCNTL_DATA_VERSION, &iDataVersion);
25475		- oputf("%-20s %u\n", "data version", iDataVersion);
	26750	+ sqlite3_fprintf(p->out, "%-20s %u\n", "data version", iDataVersion);
25476	26751	return 0;
25477	26752	}
25478	26753	#endif /* SQLITE_SHELL_HAVE_RECOVER */
	26754	+
	26755	+/*
	26756	+** Implementation of the ".dbtotxt" command.
	26757	+**
	26758	+** Return 1 on error, 2 to exit, and 0 otherwise.
	26759	+*/
	26760	+static int shell_dbtotxt_command(ShellState p, int nArg, char *azArg){
	26761	+ sqlite3_stmt *pStmt = 0;
	26762	+ sqlite3_int64 nPage = 0;
	26763	+ int pgSz = 0;
	26764	+ const char *zTail;
	26765	+ char *zName = 0;
	26766	+ int rc, i, j;
	26767	+ unsigned char bShow[256]; /* Characters ok to display */
	26768	+
	26769	+ UNUSED_PARAMETER(nArg);
	26770	+ UNUSED_PARAMETER(azArg);
	26771	+ memset(bShow, '.', sizeof(bShow));
	26772	+ for(i=' '; i<='~'; i++){
	26773	+ if( i!='{' && i!='}' && i!='"' && i!='\\' ) bShow[i] = (unsigned char)i;
	26774	+ }
	26775	+ rc = sqlite3_prepare_v2(p->db, "PRAGMA page_size", -1, &pStmt, 0);
	26776	+ if( rc ) goto dbtotxt_error;
	26777	+ rc = 0;
	26778	+ if( sqlite3_step(pStmt)!=SQLITE_ROW ) goto dbtotxt_error;
	26779	+ pgSz = sqlite3_column_int(pStmt, 0);
	26780	+ sqlite3_finalize(pStmt);
	26781	+ pStmt = 0;
	26782	+ if( pgSz<512 \|\| pgSz>65536 \|\| (pgSz&(pgSz-1))!=0 ) goto dbtotxt_error;
	26783	+ rc = sqlite3_prepare_v2(p->db, "PRAGMA page_count", -1, &pStmt, 0);
	26784	+ if( rc ) goto dbtotxt_error;
	26785	+ rc = 0;
	26786	+ if( sqlite3_step(pStmt)!=SQLITE_ROW ) goto dbtotxt_error;
	26787	+ nPage = sqlite3_column_int64(pStmt, 0);
	26788	+ sqlite3_finalize(pStmt);
	26789	+ pStmt = 0;
	26790	+ if( nPage<1 ) goto dbtotxt_error;
	26791	+ rc = sqlite3_prepare_v2(p->db, "PRAGMA databases", -1, &pStmt, 0);
	26792	+ if( rc ) goto dbtotxt_error;
	26793	+ if( sqlite3_step(pStmt)!=SQLITE_ROW ){
	26794	+ zTail = "unk.db";
	26795	+ }else{
	26796	+ const char zFilename = (const char)sqlite3_column_text(pStmt, 2);
	26797	+ if( zFilename==0 \|\| zFilename[0]==0 ) zFilename = "unk.db";
	26798	+ zTail = strrchr(zFilename, '/');
	26799	+#if defined(_WIN32)
	26800	+ if( zTail==0 ) zTail = strrchr(zFilename, '\\');
	26801	+#endif
	26802	+ }
	26803	+ zName = strdup(zTail);
	26804	+ shell_check_oom(zName);
	26805	+ sqlite3_fprintf(p->out, "\| size %lld pagesize %d filename %s\n",
	26806	+ nPage*pgSz, pgSz, zName);
	26807	+ sqlite3_finalize(pStmt);
	26808	+ pStmt = 0;
	26809	+ rc = sqlite3_prepare_v2(p->db,
	26810	+ "SELECT pgno, data FROM sqlite_dbpage ORDER BY pgno", -1, &pStmt, 0);
	26811	+ if( rc ) goto dbtotxt_error;
	26812	+ while( sqlite3_step(pStmt)==SQLITE_ROW ){
	26813	+ sqlite3_int64 pgno = sqlite3_column_int64(pStmt, 0);
	26814	+ const u8 *aData = sqlite3_column_blob(pStmt, 1);
	26815	+ int seenPageLabel = 0;
	26816	+ for(i=0; i<pgSz; i+=16){
	26817	+ const u8 *aLine = aData+i;
	26818	+ for(j=0; j<16 && aLine[j]==0; j++){}
	26819	+ if( j==16 ) continue;
	26820	+ if( !seenPageLabel ){
	26821	+ sqlite3_fprintf(p->out, "\| page %lld offset %lld\n", pgno, pgno*pgSz);
	26822	+ seenPageLabel = 1;
	26823	+ }
	26824	+ sqlite3_fprintf(p->out, "\| %5d:", i);
	26825	+ for(j=0; j<16; j++) sqlite3_fprintf(p->out, " %02x", aLine[j]);
	26826	+ sqlite3_fprintf(p->out, " ");
	26827	+ for(j=0; j<16; j++){
	26828	+ unsigned char c = (unsigned char)aLine[j];
	26829	+ sqlite3_fprintf(p->out, "%c", bShow[c]);
	26830	+ }
	26831	+ sqlite3_fprintf(p->out, "\n");
	26832	+ }
	26833	+ }
	26834	+ sqlite3_finalize(pStmt);
	26835	+ sqlite3_fprintf(p->out, "\| end %s\n", zName);
	26836	+ free(zName);
	26837	+ return 0;
	26838	+
	26839	+dbtotxt_error:
	26840	+ if( rc ){
	26841	+ sqlite3_fprintf(stderr, "ERROR: %s\n", sqlite3_errmsg(p->db));
	26842	+ }
	26843	+ sqlite3_finalize(pStmt);
	26844	+ free(zName);
	26845	+ return 1;
	26846	+}
25479	26847
25480	26848	/*
25481	26849	** Print the given string as an error message.
25482	26850	*/
25483	26851	static void shellEmitError(const char *zErr){
25484		- eputf("Error: %s\n", zErr);
	26852	+ sqlite3_fprintf(stderr,"Error: %s\n", zErr);
25485	26853	}
25486	26854	/*
25487	26855	** Print the current sqlite3_errmsg() value to stderr and return 1.
25488	26856	*/
25489	26857	static int shellDatabaseError(sqlite3 *db){
		@@ -25726,10 +27094,11 @@
25726	27094	int bGroupByParent = 0; /* If -groupbyparent is present */
25727	27095	int i; /* To iterate through azArg[] */
25728	27096	const char zIndent = ""; / How much to indent CREATE INDEX by */
25729	27097	int rc; /* Return code */
25730	27098	sqlite3_stmt pSql = 0; / Compiled version of SQL statement below */
	27099	+ FILE out = pState->out; / Send output here */
25731	27100
25732	27101	/*
25733	27102	** This SELECT statement returns one row for each foreign key constraint
25734	27103	** in the schema of the main database. The column values are:
25735	27104	**
		@@ -25801,11 +27170,12 @@
25801	27170	else if( n>1 && sqlite3_strnicmp("-groupbyparent", azArg[i], n)==0 ){
25802	27171	bGroupByParent = 1;
25803	27172	zIndent = " ";
25804	27173	}
25805	27174	else{
25806		- eputf("Usage: %s %s ?-verbose? ?-groupbyparent?\n", azArg[0], azArg[1]);
	27175	+ sqlite3_fprintf(stderr,
	27176	+ "Usage: %s %s ?-verbose? ?-groupbyparent?\n", azArg[0], azArg[1]);
25807	27177	return SQLITE_ERROR;
25808	27178	}
25809	27179	}
25810	27180
25811	27181	/* Register the fkey_collate_clause() SQL function */
		@@ -25845,44 +27215,45 @@
25845	27215	}
25846	27216	rc = sqlite3_finalize(pExplain);
25847	27217	if( rc!=SQLITE_OK ) break;
25848	27218
25849	27219	if( res<0 ){
25850		- eputz("Error: internal error");
	27220	+ sqlite3_fputs("Error: internal error", stderr);
25851	27221	break;
25852	27222	}else{
25853	27223	if( bGroupByParent
25854	27224	&& (bVerbose \|\| res==0)
25855	27225	&& (zPrev==0 \|\| sqlite3_stricmp(zParent, zPrev))
25856	27226	){
25857		- oputf("-- Parent table %s\n", zParent);
	27227	+ sqlite3_fprintf(out, "-- Parent table %s\n", zParent);
25858	27228	sqlite3_free(zPrev);
25859	27229	zPrev = sqlite3_mprintf("%s", zParent);
25860	27230	}
25861	27231
25862	27232	if( res==0 ){
25863		- oputf("%s%s --> %s\n", zIndent, zCI, zTarget);
	27233	+ sqlite3_fprintf(out, "%s%s --> %s\n", zIndent, zCI, zTarget);
25864	27234	}else if( bVerbose ){
25865		- oputf("%s/* no extra indexes required for %s -> %s */\n",
	27235	+ sqlite3_fprintf(out,
	27236	+ "%s/* no extra indexes required for %s -> %s */\n",
25866	27237	zIndent, zFrom, zTarget
25867	27238	);
25868	27239	}
25869	27240	}
25870	27241	}
25871	27242	sqlite3_free(zPrev);
25872	27243
25873	27244	if( rc!=SQLITE_OK ){
25874		- eputf("%s\n", sqlite3_errmsg(db));
	27245	+ sqlite3_fprintf(stderr,"%s\n", sqlite3_errmsg(db));
25875	27246	}
25876	27247
25877	27248	rc2 = sqlite3_finalize(pSql);
25878	27249	if( rc==SQLITE_OK && rc2!=SQLITE_OK ){
25879	27250	rc = rc2;
25880		- eputf("%s\n", sqlite3_errmsg(db));
	27251	+ sqlite3_fprintf(stderr,"%s\n", sqlite3_errmsg(db));
25881	27252	}
25882	27253	}else{
25883		- eputf("%s\n", sqlite3_errmsg(db));
	27254	+ sqlite3_fprintf(stderr,"%s\n", sqlite3_errmsg(db));
25884	27255	}
25885	27256
25886	27257	return rc;
25887	27258	}
25888	27259
		@@ -25898,13 +27269,13 @@
25898	27269	n = (nArg>=2 ? strlen30(azArg[1]) : 0);
25899	27270	if( n<1 \|\| sqlite3_strnicmp(azArg[1], "fkey-indexes", n) ) goto usage;
25900	27271	return lintFkeyIndexes(pState, azArg, nArg);
25901	27272
25902	27273	usage:
25903		- eputf("Usage %s sub-command ?switches...?\n", azArg[0]);
25904		- eputz("Where sub-commands are:\n");
25905		- eputz(" fkey-indexes\n");
	27274	+ sqlite3_fprintf(stderr,"Usage %s sub-command ?switches...?\n", azArg[0]);
	27275	+ sqlite3_fprintf(stderr, "Where sub-commands are:\n");
	27276	+ sqlite3_fprintf(stderr, " fkey-indexes\n");
25906	27277	return SQLITE_ERROR;
25907	27278	}
25908	27279
25909	27280	static void shellPrepare(
25910	27281	sqlite3 *db,
		@@ -25914,11 +27285,12 @@
25914	27285	){
25915	27286	*ppStmt = 0;
25916	27287	if( *pRc==SQLITE_OK ){
25917	27288	int rc = sqlite3_prepare_v2(db, zSql, -1, ppStmt, 0);
25918	27289	if( rc!=SQLITE_OK ){
25919		- eputf("sql error: %s (%d)\n", sqlite3_errmsg(db), sqlite3_errcode(db));
	27290	+ sqlite3_fprintf(stderr,
	27291	+ "sql error: %s (%d)\n", sqlite3_errmsg(db), sqlite3_errcode(db));
25920	27292	*pRc = rc;
25921	27293	}
25922	27294	}
25923	27295	}
25924	27296
		@@ -25958,11 +27330,11 @@
25958	27330	if( pStmt ){
25959	27331	sqlite3 *db = sqlite3_db_handle(pStmt);
25960	27332	int rc = sqlite3_finalize(pStmt);
25961	27333	if( *pRc==SQLITE_OK ){
25962	27334	if( rc!=SQLITE_OK ){
25963		- eputf("SQL error: %s\n", sqlite3_errmsg(db));
	27335	+ sqlite3_fprintf(stderr,"SQL error: %s\n", sqlite3_errmsg(db));
25964	27336	}
25965	27337	*pRc = rc;
25966	27338	}
25967	27339	}
25968	27340	}
		@@ -25980,11 +27352,11 @@
25980	27352	){
25981	27353	int rc = sqlite3_reset(pStmt);
25982	27354	if( *pRc==SQLITE_OK ){
25983	27355	if( rc!=SQLITE_OK ){
25984	27356	sqlite3 *db = sqlite3_db_handle(pStmt);
25985		- eputf("SQL error: %s\n", sqlite3_errmsg(db));
	27357	+ sqlite3_fprintf(stderr,"SQL error: %s\n", sqlite3_errmsg(db));
25986	27358	}
25987	27359	*pRc = rc;
25988	27360	}
25989	27361	}
25990	27362	#endif /* !defined SQLITE_OMIT_VIRTUALTABLE */
		@@ -26009,10 +27381,11 @@
26009	27381	char zSrcTable; / "sqlar", "zipfile($file)" or "zip" */
26010	27382	const char zFile; / --file argument, or NULL */
26011	27383	const char zDir; / --directory argument, or NULL */
26012	27384	char *azArg; / Array of command arguments */
26013	27385	ShellState p; / Shell state */
	27386	+ FILE out; / Output to this stream */
26014	27387	sqlite3 db; / Database containing the archive */
26015	27388	};
26016	27389
26017	27390	/*
26018	27391	** Print a usage message for the .ar command to stderr and return SQLITE_ERROR.
		@@ -26032,13 +27405,13 @@
26032	27405	va_start(ap, zFmt);
26033	27406	z = sqlite3_vmprintf(zFmt, ap);
26034	27407	va_end(ap);
26035	27408	shellEmitError(z);
26036	27409	if( pAr->fromCmdLine ){
26037		- eputz("Use \"-A\" for more help\n");
	27410	+ sqlite3_fputs("Use \"-A\" for more help\n", stderr);
26038	27411	}else{
26039		- eputz("Use \".archive --help\" for more help\n");
	27412	+ sqlite3_fputs("Use \".archive --help\" for more help\n", stderr);
26040	27413	}
26041	27414	sqlite3_free(z);
26042	27415	return SQLITE_ERROR;
26043	27416	}
26044	27417
		@@ -26134,11 +27507,11 @@
26134	27507	};
26135	27508	int nSwitch = sizeof(aSwitch) / sizeof(struct ArSwitch);
26136	27509	struct ArSwitch *pEnd = &aSwitch[nSwitch];
26137	27510
26138	27511	if( nArg<=1 ){
26139		- eputz("Wrong number of arguments. Usage:\n");
	27512	+ sqlite3_fprintf(stderr, "Wrong number of arguments. Usage:\n");
26140	27513	return arUsage(stderr);
26141	27514	}else{
26142	27515	char *z = azArg[1];
26143	27516	if( z[0]!='-' ){
26144	27517	/* Traditional style [tar] invocation */
		@@ -26240,11 +27613,11 @@
26240	27613	}
26241	27614	}
26242	27615	}
26243	27616	}
26244	27617	if( pAr->eCmd==0 ){
26245		- eputz("Required argument missing. Usage:\n");
	27618	+ sqlite3_fprintf(stderr, "Required argument missing. Usage:\n");
26246	27619	return arUsage(stderr);
26247	27620	}
26248	27621	return SQLITE_OK;
26249	27622	}
26250	27623
		@@ -26283,11 +27656,11 @@
26283	27656	if( SQLITE_ROW==sqlite3_step(pTest) ){
26284	27657	bOk = 1;
26285	27658	}
26286	27659	shellReset(&rc, pTest);
26287	27660	if( rc==SQLITE_OK && bOk==0 ){
26288		- eputf("not found in archive: %s\n", z);
	27661	+ sqlite3_fprintf(stderr,"not found in archive: %s\n", z);
26289	27662	rc = SQLITE_ERROR;
26290	27663	}
26291	27664	}
26292	27665	shellFinalize(&rc, pTest);
26293	27666	}
		@@ -26350,19 +27723,19 @@
26350	27723	arWhereClause(&rc, pAr, &zWhere);
26351	27724
26352	27725	shellPreparePrintf(pAr->db, &rc, &pSql, zSql, azCols[pAr->bVerbose],
26353	27726	pAr->zSrcTable, zWhere);
26354	27727	if( pAr->bDryRun ){
26355		- oputf("%s\n", sqlite3_sql(pSql));
	27728	+ sqlite3_fprintf(pAr->out, "%s\n", sqlite3_sql(pSql));
26356	27729	}else{
26357	27730	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
26358	27731	if( pAr->bVerbose ){
26359		- oputf("%s % 10d %s %s\n",
	27732	+ sqlite3_fprintf(pAr->out, "%s % 10d %s %s\n",
26360	27733	sqlite3_column_text(pSql, 0), sqlite3_column_int(pSql, 1),
26361	27734	sqlite3_column_text(pSql, 2),sqlite3_column_text(pSql, 3));
26362	27735	}else{
26363		- oputf("%s\n", sqlite3_column_text(pSql, 0));
	27736	+ sqlite3_fprintf(pAr->out, "%s\n", sqlite3_column_text(pSql, 0));
26364	27737	}
26365	27738	}
26366	27739	}
26367	27740	shellFinalize(&rc, pSql);
26368	27741	sqlite3_free(zWhere);
		@@ -26385,11 +27758,11 @@
26385	27758	}
26386	27759	if( rc==SQLITE_OK ){
26387	27760	zSql = sqlite3_mprintf("DELETE FROM %s WHERE %s;",
26388	27761	pAr->zSrcTable, zWhere);
26389	27762	if( pAr->bDryRun ){
26390		- oputf("%s\n", zSql);
	27763	+ sqlite3_fprintf(pAr->out, "%s\n", zSql);
26391	27764	}else{
26392	27765	char *zErr = 0;
26393	27766	rc = sqlite3_exec(pAr->db, "SAVEPOINT ar;", 0, 0, 0);
26394	27767	if( rc==SQLITE_OK ){
26395	27768	rc = sqlite3_exec(pAr->db, zSql, 0, 0, &zErr);
		@@ -26398,11 +27771,11 @@
26398	27771	}else{
26399	27772	rc = sqlite3_exec(pAr->db, "RELEASE ar;", 0, 0, 0);
26400	27773	}
26401	27774	}
26402	27775	if( zErr ){
26403		- sputf(stdout, "ERROR: %s\n", zErr); /* stdout? */
	27776	+ sqlite3_fprintf(stdout, "ERROR: %s\n", zErr); /* stdout? */
26404	27777	sqlite3_free(zErr);
26405	27778	}
26406	27779	}
26407	27780	}
26408	27781	sqlite3_free(zWhere);
		@@ -26462,15 +27835,15 @@
26462	27835	** populating them changes the timestamp). */
26463	27836	for(i=0; i<2; i++){
26464	27837	j = sqlite3_bind_parameter_index(pSql, "$dirOnly");
26465	27838	sqlite3_bind_int(pSql, j, i);
26466	27839	if( pAr->bDryRun ){
26467		- oputf("%s\n", sqlite3_sql(pSql));
	27840	+ sqlite3_fprintf(pAr->out, "%s\n", sqlite3_sql(pSql));
26468	27841	}else{
26469	27842	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
26470	27843	if( i==0 && pAr->bVerbose ){
26471		- oputf("%s\n", sqlite3_column_text(pSql, 0));
	27844	+ sqlite3_fprintf(pAr->out, "%s\n", sqlite3_column_text(pSql, 0));
26472	27845	}
26473	27846	}
26474	27847	}
26475	27848	shellReset(&rc, pSql);
26476	27849	}
		@@ -26486,17 +27859,17 @@
26486	27859	** Run the SQL statement in zSql. Or if doing a --dryrun, merely print it out.
26487	27860	*/
26488	27861	static int arExecSql(ArCommand pAr, const char zSql){
26489	27862	int rc;
26490	27863	if( pAr->bDryRun ){
26491		- oputf("%s\n", zSql);
	27864	+ sqlite3_fprintf(pAr->out, "%s\n", zSql);
26492	27865	rc = SQLITE_OK;
26493	27866	}else{
26494	27867	char *zErr = 0;
26495	27868	rc = sqlite3_exec(pAr->db, zSql, 0, 0, &zErr);
26496	27869	if( zErr ){
26497		- sputf(stdout, "ERROR: %s\n", zErr);
	27870	+ sqlite3_fprintf(stdout, "ERROR: %s\n", zErr);
26498	27871	sqlite3_free(zErr);
26499	27872	}
26500	27873	}
26501	27874	return rc;
26502	27875	}
		@@ -26641,10 +28014,11 @@
26641	28014	cmd.fromCmdLine = fromCmdLine;
26642	28015	rc = arParseCommand(azArg, nArg, &cmd);
26643	28016	if( rc==SQLITE_OK ){
26644	28017	int eDbType = SHELL_OPEN_UNSPEC;
26645	28018	cmd.p = pState;
	28019	+ cmd.out = pState->out;
26646	28020	cmd.db = pState->db;
26647	28021	if( cmd.zFile ){
26648	28022	eDbType = deduceDatabaseType(cmd.zFile, 1);
26649	28023	}else{
26650	28024	eDbType = pState->openMode;
		@@ -26667,17 +28041,18 @@
26667	28041	}else{
26668	28042	flags = SQLITE_OPEN_READONLY;
26669	28043	}
26670	28044	cmd.db = 0;
26671	28045	if( cmd.bDryRun ){
26672		- oputf("-- open database '%s'%s\n", cmd.zFile,
	28046	+ sqlite3_fprintf(cmd.out, "-- open database '%s'%s\n", cmd.zFile,
26673	28047	eDbType==SHELL_OPEN_APPENDVFS ? " using 'apndvfs'" : "");
26674	28048	}
26675	28049	rc = sqlite3_open_v2(cmd.zFile, &cmd.db, flags,
26676	28050	eDbType==SHELL_OPEN_APPENDVFS ? "apndvfs" : 0);
26677	28051	if( rc!=SQLITE_OK ){
26678		- eputf("cannot open file: %s (%s)\n", cmd.zFile, sqlite3_errmsg(cmd.db));
	28052	+ sqlite3_fprintf(stderr, "cannot open file: %s (%s)\n",
	28053	+ cmd.zFile, sqlite3_errmsg(cmd.db));
26679	28054	goto end_ar_command;
26680	28055	}
26681	28056	sqlite3_fileio_init(cmd.db, 0, 0);
26682	28057	sqlite3_sqlar_init(cmd.db, 0, 0);
26683	28058	sqlite3_create_function(cmd.db, "shell_putsnl", 1, SQLITE_UTF8, cmd.p,
		@@ -26686,11 +28061,11 @@
26686	28061	}
26687	28062	if( cmd.zSrcTable==0 && cmd.bZip==0 && cmd.eCmd!=AR_CMD_HELP ){
26688	28063	if( cmd.eCmd!=AR_CMD_CREATE
26689	28064	&& sqlite3_table_column_metadata(cmd.db,0,"sqlar","name",0,0,0,0,0)
26690	28065	){
26691		- eputz("database does not contain an 'sqlar' table\n");
	28066	+ sqlite3_fprintf(stderr, "database does not contain an 'sqlar' table\n");
26692	28067	rc = SQLITE_ERROR;
26693	28068	goto end_ar_command;
26694	28069	}
26695	28070	cmd.zSrcTable = sqlite3_mprintf("sqlar");
26696	28071	}
		@@ -26744,11 +28119,11 @@
26744	28119	** This function is used as a callback by the recover extension. Simply
26745	28120	** print the supplied SQL statement to stdout.
26746	28121	*/
26747	28122	static int recoverSqlCb(void pCtx, const char zSql){
26748	28123	ShellState pState = (ShellState)pCtx;
26749		- sputf(pState->out, "%s;\n", zSql);
	28124	+ sqlite3_fprintf(pState->out, "%s;\n", zSql);
26750	28125	return SQLITE_OK;
26751	28126	}
26752	28127
26753	28128	/*
26754	28129	** This function is called to recover data from the database. A script
		@@ -26787,11 +28162,11 @@
26787	28162	}else
26788	28163	if( n<=10 && memcmp("-no-rowids", z, n)==0 ){
26789	28164	bRowids = 0;
26790	28165	}
26791	28166	else{
26792		- eputf("unexpected option: %s\n", azArg[i]);
	28167	+ sqlite3_fprintf(stderr,"unexpected option: %s\n", azArg[i]);
26793	28168	showHelp(pState->out, azArg[0]);
26794	28169	return 1;
26795	28170	}
26796	28171	}
26797	28172
		@@ -26806,11 +28181,11 @@
26806	28181
26807	28182	sqlite3_recover_run(p);
26808	28183	if( sqlite3_recover_errcode(p)!=SQLITE_OK ){
26809	28184	const char *zErr = sqlite3_recover_errmsg(p);
26810	28185	int errCode = sqlite3_recover_errcode(p);
26811		- eputf("sql error: %s (%d)\n", zErr, errCode);
	28186	+ sqlite3_fprintf(stderr,"sql error: %s (%d)\n", zErr, errCode);
26812	28187	}
26813	28188	rc = sqlite3_recover_finish(p);
26814	28189	return rc;
26815	28190	}
26816	28191	#endif /* SQLITE_SHELL_HAVE_RECOVER */
		@@ -26828,25 +28203,25 @@
26828	28203	i64 nError = 0;
26829	28204	const char *zErr = 0;
26830	28205	while( SQLITE_OK==sqlite3_intck_step(p) ){
26831	28206	const char *zMsg = sqlite3_intck_message(p);
26832	28207	if( zMsg ){
26833		- oputf("%s\n", zMsg);
	28208	+ sqlite3_fprintf(pState->out, "%s\n", zMsg);
26834	28209	nError++;
26835	28210	}
26836	28211	nStep++;
26837	28212	if( nStepPerUnlock && (nStep % nStepPerUnlock)==0 ){
26838	28213	sqlite3_intck_unlock(p);
26839	28214	}
26840	28215	}
26841	28216	rc = sqlite3_intck_error(p, &zErr);
26842	28217	if( zErr ){
26843		- eputf("%s\n", zErr);
	28218	+ sqlite3_fprintf(stderr,"%s\n", zErr);
26844	28219	}
26845	28220	sqlite3_intck_close(p);
26846	28221
26847		- oputf("%lld steps, %lld errors\n", nStep, nError);
	28222	+ sqlite3_fprintf(pState->out, "%lld steps, %lld errors\n", nStep, nError);
26848	28223	}
26849	28224
26850	28225	return rc;
26851	28226	}
26852	28227
		@@ -26865,11 +28240,11 @@
26865	28240	*/
26866	28241	#ifdef SHELL_DEBUG
26867	28242	#define rc_err_oom_die(rc) \
26868	28243	if( rc==SQLITE_NOMEM ) shell_check_oom(0); \
26869	28244	else if(!(rc==SQLITE_OK\|\|rc==SQLITE_DONE)) \
26870		- eputf("E:%d\n",rc), assert(0)
	28245	+ sqlite3_fprintf(stderr,"E:%d\n",rc), assert(0)
26871	28246	#else
26872	28247	static void rc_err_oom_die(int rc){
26873	28248	if( rc==SQLITE_NOMEM ) shell_check_oom(0);
26874	28249	assert(rc==SQLITE_OK\|\|rc==SQLITE_DONE);
26875	28250	}
		@@ -27023,12 +28398,12 @@
27023	28398	}else if( *pDb==0 ){
27024	28399	return 0;
27025	28400	}else{
27026	28401	/* Formulate the columns spec, close the DB, zero pDb. /
27027	28402	char *zColsSpec = 0;
27028		- int hasDupes = db_int(*pDb, zHasDupes);
27029		- int nDigits = (hasDupes)? db_int(*pDb, zColDigits) : 0;
	28403	+ int hasDupes = db_int(*pDb, "%s", zHasDupes);
	28404	+ int nDigits = (hasDupes)? db_int(*pDb, "%s", zColDigits) : 0;
27030	28405	if( hasDupes ){
27031	28406	#ifdef SHELL_COLUMN_RENAME_CLEAN
27032	28407	rc = sqlite3_exec(*pDb, zDedoctor, 0, 0, 0);
27033	28408	rc_err_oom_die(rc);
27034	28409	#endif
		@@ -27039,11 +28414,11 @@
27039	28414	sqlite3_bind_int(pStmt, 1, nDigits);
27040	28415	rc = sqlite3_step(pStmt);
27041	28416	sqlite3_finalize(pStmt);
27042	28417	if( rc!=SQLITE_DONE ) rc_err_oom_die(SQLITE_NOMEM);
27043	28418	}
27044		- assert(db_int(pDb, zHasDupes)==0); / Consider: remove this */
	28419	+ assert(db_int(pDb, "%s", zHasDupes)==0); / Consider: remove this */
27045	28420	rc = sqlite3_prepare_v2(*pDb, zCollectVar, -1, &pStmt, 0);
27046	28421	rc_err_oom_die(rc);
27047	28422	rc = sqlite3_step(pStmt);
27048	28423	if( rc==SQLITE_ROW ){
27049	28424	zColsSpec = sqlite3_mprintf("%s", sqlite3_column_text(pStmt, 0));
		@@ -27082,12 +28457,13 @@
27082	28457	shellPreparePrintf(p->db, &rc, &pStmt,
27083	28458	"SELECT 1 FROM sqlite_schema o WHERE "
27084	28459	"sql LIKE 'CREATE VIRTUAL TABLE%%' AND %s", zLike ? zLike : "true"
27085	28460	);
27086	28461	if( rc==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW ){
27087		- oputz("/* WARNING: "
27088		- "Script requires that SQLITE_DBCONFIG_DEFENSIVE be disabled */\n"
	28462	+ sqlite3_fputs("/* WARNING: "
	28463	+ "Script requires that SQLITE_DBCONFIG_DEFENSIVE be disabled */\n",
	28464	+ p->out
27089	28465	);
27090	28466	}
27091	28467	shellFinalize(&rc, pStmt);
27092	28468	return rc;
27093	28469	}
		@@ -27114,16 +28490,18 @@
27114	28490	return SQLITE_OK;
27115	28491	}
27116	28492	if( faultsim_state.iCnt ){
27117	28493	if( faultsim_state.iCnt>0 ) faultsim_state.iCnt--;
27118	28494	if( faultsim_state.eVerbose>=2 ){
27119		- oputf("FAULT-SIM id=%d no-fault (cnt=%d)\n", iArg, faultsim_state.iCnt);
	28495	+ sqlite3_fprintf(stdout,
	28496	+ "FAULT-SIM id=%d no-fault (cnt=%d)\n", iArg, faultsim_state.iCnt);
27120	28497	}
27121	28498	return SQLITE_OK;
27122	28499	}
27123	28500	if( faultsim_state.eVerbose>=1 ){
27124		- oputf("FAULT-SIM id=%d returns %d\n", iArg, faultsim_state.iErr);
	28501	+ sqlite3_fprintf(stdout,
	28502	+ "FAULT-SIM id=%d returns %d\n", iArg, faultsim_state.iErr);
27125	28503	}
27126	28504	faultsim_state.iCnt = faultsim_state.iInterval;
27127	28505	faultsim_state.nHit++;
27128	28506	if( faultsim_state.nRepeat>0 && faultsim_state.nRepeat<=faultsim_state.nHit ){
27129	28507	faultsim_state.iCnt = -1;
		@@ -27182,11 +28560,11 @@
27182	28560	clearTempFile(p);
27183	28561
27184	28562	#ifndef SQLITE_OMIT_AUTHORIZATION
27185	28563	if( c=='a' && cli_strncmp(azArg[0], "auth", n)==0 ){
27186	28564	if( nArg!=2 ){
27187		- eputz("Usage: .auth ON\|OFF\n");
	28565	+ sqlite3_fprintf(stderr, "Usage: .auth ON\|OFF\n");
27188	28566	rc = 1;
27189	28567	goto meta_command_exit;
27190	28568	}
27191	28569	open_db(p, 0);
27192	28570	if( booleanValue(azArg[1]) ){
		@@ -27229,32 +28607,32 @@
27229	28607	}else
27230	28608	if( cli_strcmp(z, "-async")==0 ){
27231	28609	bAsync = 1;
27232	28610	}else
27233	28611	{
27234		- eputf("unknown option: %s\n", azArg[j]);
	28612	+ sqlite3_fprintf(stderr,"unknown option: %s\n", azArg[j]);
27235	28613	return 1;
27236	28614	}
27237	28615	}else if( zDestFile==0 ){
27238	28616	zDestFile = azArg[j];
27239	28617	}else if( zDb==0 ){
27240	28618	zDb = zDestFile;
27241	28619	zDestFile = azArg[j];
27242	28620	}else{
27243		- eputz("Usage: .backup ?DB? ?OPTIONS? FILENAME\n");
	28621	+ sqlite3_fprintf(stderr, "Usage: .backup ?DB? ?OPTIONS? FILENAME\n");
27244	28622	return 1;
27245	28623	}
27246	28624	}
27247	28625	if( zDestFile==0 ){
27248		- eputz("missing FILENAME argument on .backup\n");
	28626	+ sqlite3_fprintf(stderr, "missing FILENAME argument on .backup\n");
27249	28627	return 1;
27250	28628	}
27251	28629	if( zDb==0 ) zDb = "main";
27252	28630	rc = sqlite3_open_v2(zDestFile, &pDest,
27253	28631	SQLITE_OPEN_READWRITE\|SQLITE_OPEN_CREATE, zVfs);
27254	28632	if( rc!=SQLITE_OK ){
27255		- eputf("Error: cannot open \"%s\"\n", zDestFile);
	28633	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zDestFile);
27256	28634	close_db(pDest);
27257	28635	return 1;
27258	28636	}
27259	28637	if( bAsync ){
27260	28638	sqlite3_exec(pDest, "PRAGMA synchronous=OFF; PRAGMA journal_mode=OFF;",
		@@ -27286,23 +28664,14 @@
27286	28664	eputz("Usage: .bail on\|off\n");
27287	28665	rc = 1;
27288	28666	}
27289	28667	}else
27290	28668
27291		- /* Undocumented. Legacy only. See "crnl" below */
	28669	+ /* Undocumented. Legacy only. See "crlf" below */
27292	28670	if( c=='b' && n>=3 && cli_strncmp(azArg[0], "binary", n)==0 ){
27293		- if( nArg==2 ){
27294		- if( booleanValue(azArg[1]) ){
27295		- setBinaryMode(p->out, 1);
27296		- }else{
27297		- setTextMode(p->out, 1);
27298		- }
27299		- }else{
27300		- eputz("The \".binary\" command is deprecated. Use \".crnl\" instead.\n"
27301		- "Usage: .binary on\|off\n");
27302		- rc = 1;
27303		- }
	28671	+ eputz("The \".binary\" command is deprecated.\n");
	28672	+ rc = 1;
27304	28673	}else
27305	28674
27306	28675	/* The undocumented ".breakpoint" command causes a call to the no-op
27307	28676	** routine named test_breakpoint().
27308	28677	*/
		@@ -27320,11 +28689,11 @@
27320	28689	sqlite3_free(z);
27321	28690	#else
27322	28691	rc = chdir(azArg[1]);
27323	28692	#endif
27324	28693	if( rc ){
27325		- eputf("Cannot change to directory \"%s\"\n", azArg[1]);
	28694	+ sqlite3_fprintf(stderr,"Cannot change to directory \"%s\"\n", azArg[1]);
27326	28695	rc = 1;
27327	28696	}
27328	28697	}else{
27329	28698	eputz("Usage: .cd DIRECTORY\n");
27330	28699	rc = 1;
		@@ -27353,15 +28722,16 @@
27353	28722	eputz("Usage: .check GLOB-PATTERN\n");
27354	28723	rc = 2;
27355	28724	}else if( (zRes = readFile("testcase-out.txt", 0))==0 ){
27356	28725	rc = 2;
27357	28726	}else if( testcase_glob(azArg[1],zRes)==0 ){
27358		- eputf("testcase-%s FAILED\n Expected: [%s]\n Got: [%s]\n",
	28727	+ sqlite3_fprintf(stderr,
	28728	+ "testcase-%s FAILED\n Expected: [%s]\n Got: [%s]\n",
27359	28729	p->zTestcase, azArg[1], zRes);
27360	28730	rc = 1;
27361	28731	}else{
27362		- oputf("testcase-%s ok\n", p->zTestcase);
	28732	+ sqlite3_fprintf(p->out, "testcase-%s ok\n", p->zTestcase);
27363	28733	p->nCheck++;
27364	28734	}
27365	28735	sqlite3_free(zRes);
27366	28736	}else
27367	28737	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
		@@ -27390,13 +28760,13 @@
27390	28760	zFile = "(memory)";
27391	28761	}else if( zFile[0]==0 ){
27392	28762	zFile = "(temporary-file)";
27393	28763	}
27394	28764	if( p->pAuxDb == &p->aAuxDb[i] ){
27395		- sputf(stdout, "ACTIVE %d: %s\n", i, zFile);
	28765	+ sqlite3_fprintf(stdout, "ACTIVE %d: %s\n", i, zFile);
27396	28766	}else if( p->aAuxDb[i].db!=0 ){
27397		- sputf(stdout, " %d: %s\n", i, zFile);
	28767	+ sqlite3_fprintf(stdout, " %d: %s\n", i, zFile);
27398	28768	}
27399	28769	}
27400	28770	}else if( nArg==2 && IsDigit(azArg[1][0]) && azArg[1][1]==0 ){
27401	28771	int i = azArg[1][0] - '0';
27402	28772	if( p->pAuxDb != &p->aAuxDb[i] && i>=0 && i<ArraySize(p->aAuxDb) ){
		@@ -27422,24 +28792,22 @@
27422	28792	eputz("Usage: .connection [close] [CONNECTION-NUMBER]\n");
27423	28793	rc = 1;
27424	28794	}
27425	28795	}else
27426	28796
27427		- if( c=='c' && n==4 && cli_strncmp(azArg[0], "crnl", n)==0 ){
	28797	+ if( c=='c' && n==4
	28798	+ && (cli_strncmp(azArg[0], "crlf", n)==0
	28799	+ \|\| cli_strncmp(azArg[0], "crnl",n)==0)
	28800	+ ){
27428	28801	if( nArg==2 ){
27429		- if( booleanValue(azArg[1]) ){
27430		- setTextMode(p->out, 1);
27431		- }else{
27432		- setBinaryMode(p->out, 1);
27433		- }
27434		- }else{
27435		-#if !defined(_WIN32) && !defined(WIN32)
27436		- eputz("The \".crnl\" is a no-op on non-Windows machines.\n");
	28802	+#ifdef _WIN32
	28803	+ p->crlfMode = booleanValue(azArg[1]);
	28804	+#else
	28805	+ p->crlfMode = 0;
27437	28806	#endif
27438		- eputz("Usage: .crnl on\|off\n");
27439		- rc = 1;
27440	28807	}
	28808	+ sqlite3_fprintf(stderr, "crlf is %s\n", p->crlfMode ? "ON" : "OFF");
27441	28809	}else
27442	28810
27443	28811	if( c=='d' && n>1 && cli_strncmp(azArg[0], "databases", n)==0 ){
27444	28812	char **azName = 0;
27445	28813	int nName = 0;
		@@ -27465,11 +28833,11 @@
27465	28833	sqlite3_finalize(pStmt);
27466	28834	for(i=0; i<nName; i++){
27467	28835	int eTxn = sqlite3_txn_state(p->db, azName[i*2]);
27468	28836	int bRdonly = sqlite3_db_readonly(p->db, azName[i*2]);
27469	28837	const char z = azName[i2+1];
27470		- oputf("%s: %s %s%s\n",
	28838	+ sqlite3_fprintf(p->out, "%s: %s %s%s\n",
27471	28839	azName[i*2], z && z[0] ? z : "\"\"", bRdonly ? "r/o" : "r/w",
27472	28840	eTxn==SQLITE_TXN_NONE ? "" :
27473	28841	eTxn==SQLITE_TXN_READ ? " read-txn" : " write-txn");
27474	28842	free(azName[i*2]);
27475	28843	free(azName[i*2+1]);
		@@ -27507,15 +28875,16 @@
27507	28875	if( nArg>1 && cli_strcmp(azArg[1], aDbConfig[ii].zName)!=0 ) continue;
27508	28876	if( nArg>=3 ){
27509	28877	sqlite3_db_config(p->db, aDbConfig[ii].op, booleanValue(azArg[2]), 0);
27510	28878	}
27511	28879	sqlite3_db_config(p->db, aDbConfig[ii].op, -1, &v);
27512		- oputf("%19s %s\n", aDbConfig[ii].zName, v ? "on" : "off");
	28880	+ sqlite3_fprintf(p->out, "%19s %s\n",
	28881	+ aDbConfig[ii].zName, v ? "on" : "off");
27513	28882	if( nArg>1 ) break;
27514	28883	}
27515	28884	if( nArg>1 && ii==ArraySize(aDbConfig) ){
27516		- eputf("Error: unknown dbconfig \"%s\"\n", azArg[1]);
	28885	+ sqlite3_fprintf(stderr,"Error: unknown dbconfig \"%s\"\n", azArg[1]);
27517	28886	eputz("Enter \".dbconfig\" with no arguments for a list\n");
27518	28887	}
27519	28888	}else
27520	28889
27521	28890	#if SQLITE_SHELL_HAVE_RECOVER
		@@ -27561,11 +28930,12 @@
27561	28930	}else
27562	28931	if( cli_strcmp(z,"nosys")==0 ){
27563	28932	ShellSetFlag(p, SHFLG_DumpNoSys);
27564	28933	}else
27565	28934	{
27566		- eputf("Unknown option \"%s\" on \".dump\"\n", azArg[i]);
	28935	+ sqlite3_fprintf(stderr,
	28936	+ "Unknown option \"%s\" on \".dump\"\n", azArg[i]);
27567	28937	rc = 1;
27568	28938	sqlite3_free(zLike);
27569	28939	goto meta_command_exit;
27570	28940	}
27571	28941	}else{
		@@ -27596,12 +28966,12 @@
27596	28966	outputDumpWarning(p, zLike);
27597	28967	if( (p->shellFlgs & SHFLG_DumpDataOnly)==0 ){
27598	28968	/* When playing back a "dump", the content might appear in an order
27599	28969	** which causes immediate foreign key constraints to be violated.
27600	28970	** So disable foreign-key constraint enforcement to prevent problems. */
27601		- oputz("PRAGMA foreign_keys=OFF;\n");
27602		- oputz("BEGIN TRANSACTION;\n");
	28971	+ sqlite3_fputs("PRAGMA foreign_keys=OFF;\n", p->out);
	28972	+ sqlite3_fputs("BEGIN TRANSACTION;\n", p->out);
27603	28973	}
27604	28974	p->writableSchema = 0;
27605	28975	p->showHeader = 0;
27606	28976	/* Set writable_schema=ON since doing so forces SQLite to initialize
27607	28977	** as much of the schema as it can even if the sqlite_schema table is
		@@ -27629,17 +28999,17 @@
27629	28999	run_table_dump_query(p, zSql);
27630	29000	sqlite3_free(zSql);
27631	29001	}
27632	29002	sqlite3_free(zLike);
27633	29003	if( p->writableSchema ){
27634		- oputz("PRAGMA writable_schema=OFF;\n");
	29004	+ sqlite3_fputs("PRAGMA writable_schema=OFF;\n", p->out);
27635	29005	p->writableSchema = 0;
27636	29006	}
27637	29007	sqlite3_exec(p->db, "PRAGMA writable_schema=OFF;", 0, 0, 0);
27638	29008	sqlite3_exec(p->db, "RELEASE dump;", 0, 0, 0);
27639	29009	if( (p->shellFlgs & SHFLG_DumpDataOnly)==0 ){
27640		- oputz(p->nErr?"ROLLBACK; -- due to errors\n":"COMMIT;\n");
	29010	+ sqlite3_fputs(p->nErr?"ROLLBACK; -- due to errors\n":"COMMIT;\n", p->out);
27641	29011	}
27642	29012	p->showHeader = savedShowHeader;
27643	29013	p->shellFlgs = savedShellFlags;
27644	29014	}else
27645	29015
		@@ -27649,10 +29019,14 @@
27649	29019	}else{
27650	29020	eputz("Usage: .echo on\|off\n");
27651	29021	rc = 1;
27652	29022	}
27653	29023	}else
	29024	+
	29025	+ if( c=='d' && n>=3 && cli_strncmp(azArg[0], "dbtotxt", n)==0 ){
	29026	+ rc = shell_dbtotxt_command(p, nArg, azArg);
	29027	+ }else
27654	29028
27655	29029	if( c=='e' && cli_strncmp(azArg[0], "eqp", n)==0 ){
27656	29030	if( nArg==2 ){
27657	29031	p->autoEQPtest = 0;
27658	29032	if( p->autoEQPtrace ){
		@@ -27715,11 +29089,12 @@
27715	29089	}else
27716	29090
27717	29091	#ifndef SQLITE_OMIT_VIRTUALTABLE
27718	29092	if( c=='e' && cli_strncmp(azArg[0], "expert", n)==0 ){
27719	29093	if( p->bSafeMode ){
27720		- eputf("Cannot run experimental commands such as \"%s\" in safe mode\n",
	29094	+ sqlite3_fprintf(stderr,
	29095	+ "Cannot run experimental commands such as \"%s\" in safe mode\n",
27721	29096	azArg[0]);
27722	29097	rc = 1;
27723	29098	}else{
27724	29099	open_db(p, 0);
27725	29100	expertDotCommand(p, azArg, nArg);
		@@ -27772,13 +29147,14 @@
27772	29147	if( zCmd[0]=='-' && zCmd[1] ) zCmd++;
27773	29148	}
27774	29149
27775	29150	/* --help lists all file-controls */
27776	29151	if( cli_strcmp(zCmd,"help")==0 ){
27777		- oputz("Available file-controls:\n");
	29152	+ sqlite3_fputs("Available file-controls:\n", p->out);
27778	29153	for(i=0; i<ArraySize(aCtrl); i++){
27779		- oputf(" .filectrl %s %s\n", aCtrl[i].zCtrlName, aCtrl[i].zUsage);
	29154	+ sqlite3_fprintf(p->out,
	29155	+ " .filectrl %s %s\n", aCtrl[i].zCtrlName, aCtrl[i].zUsage);
27780	29156	}
27781	29157	rc = 1;
27782	29158	goto meta_command_exit;
27783	29159	}
27784	29160
		@@ -27789,19 +29165,19 @@
27789	29165	if( cli_strncmp(zCmd, aCtrl[i].zCtrlName, n2)==0 ){
27790	29166	if( filectrl<0 ){
27791	29167	filectrl = aCtrl[i].ctrlCode;
27792	29168	iCtrl = i;
27793	29169	}else{
27794		- eputf("Error: ambiguous file-control: \"%s\"\n"
	29170	+ sqlite3_fprintf(stderr,"Error: ambiguous file-control: \"%s\"\n"
27795	29171	"Use \".filectrl --help\" for help\n", zCmd);
27796	29172	rc = 1;
27797	29173	goto meta_command_exit;
27798	29174	}
27799	29175	}
27800	29176	}
27801	29177	if( filectrl<0 ){
27802		- eputf("Error: unknown file-control: %s\n"
	29178	+ sqlite3_fprintf(stderr,"Error: unknown file-control: %s\n"
27803	29179	"Use \".filectrl --help\" for help\n", zCmd);
27804	29180	}else{
27805	29181	switch(filectrl){
27806	29182	case SQLITE_FCNTL_SIZE_LIMIT: {
27807	29183	if( nArg!=2 && nArg!=3 ) break;
		@@ -27841,11 +29217,11 @@
27841	29217	case SQLITE_FCNTL_TEMPFILENAME: {
27842	29218	char *z = 0;
27843	29219	if( nArg!=2 ) break;
27844	29220	sqlite3_file_control(p->db, zSchema, filectrl, &z);
27845	29221	if( z ){
27846		- oputf("%s\n", z);
	29222	+ sqlite3_fprintf(p->out, "%s\n", z);
27847	29223	sqlite3_free(z);
27848	29224	}
27849	29225	isOk = 2;
27850	29226	break;
27851	29227	}
		@@ -27855,23 +29231,24 @@
27855	29231	x = atoi(azArg[2]);
27856	29232	sqlite3_file_control(p->db, zSchema, filectrl, &x);
27857	29233	}
27858	29234	x = -1;
27859	29235	sqlite3_file_control(p->db, zSchema, filectrl, &x);
27860		- oputf("%d\n", x);
	29236	+ sqlite3_fprintf(p->out, "%d\n", x);
27861	29237	isOk = 2;
27862	29238	break;
27863	29239	}
27864	29240	}
27865	29241	}
27866	29242	if( isOk==0 && iCtrl>=0 ){
27867		- oputf("Usage: .filectrl %s %s\n", zCmd,aCtrl[iCtrl].zUsage);
	29243	+ sqlite3_fprintf(p->out, "Usage: .filectrl %s %s\n",
	29244	+ zCmd, aCtrl[iCtrl].zUsage);
27868	29245	rc = 1;
27869	29246	}else if( isOk==1 ){
27870	29247	char zBuf[100];
27871	29248	sqlite3_snprintf(sizeof(zBuf), zBuf, "%lld", iRes);
27872		- oputf("%s\n", zBuf);
	29249	+ sqlite3_fprintf(p->out, "%s\n", zBuf);
27873	29250	}
27874	29251	}else
27875	29252
27876	29253	if( c=='f' && cli_strncmp(azArg[0], "fullschema", n)==0 ){
27877	29254	ShellState data;
		@@ -27908,19 +29285,19 @@
27908	29285	doStats = sqlite3_step(pStmt)==SQLITE_ROW;
27909	29286	sqlite3_finalize(pStmt);
27910	29287	}
27911	29288	}
27912	29289	if( doStats==0 ){
27913		- oputz("/* No STAT tables available */\n");
	29290	+ sqlite3_fputs("/* No STAT tables available */\n", p->out);
27914	29291	}else{
27915		- oputz("ANALYZE sqlite_schema;\n");
	29292	+ sqlite3_fputs("ANALYZE sqlite_schema;\n", p->out);
27916	29293	data.cMode = data.mode = MODE_Insert;
27917	29294	data.zDestTable = "sqlite_stat1";
27918	29295	shell_exec(&data, "SELECT * FROM sqlite_stat1", 0);
27919	29296	data.zDestTable = "sqlite_stat4";
27920	29297	shell_exec(&data, "SELECT * FROM sqlite_stat4", 0);
27921		- oputz("ANALYZE sqlite_schema;\n");
	29298	+ sqlite3_fputs("ANALYZE sqlite_schema;\n", p->out);
27922	29299	}
27923	29300	}else
27924	29301
27925	29302	if( c=='h' && cli_strncmp(azArg[0], "headers", n)==0 ){
27926	29303	if( nArg==2 ){
		@@ -27934,11 +29311,11 @@
27934	29311
27935	29312	if( c=='h' && cli_strncmp(azArg[0], "help", n)==0 ){
27936	29313	if( nArg>=2 ){
27937	29314	n = showHelp(p->out, azArg[1]);
27938	29315	if( n==0 ){
27939		- oputf("Nothing matches '%s'\n", azArg[1]);
	29316	+ sqlite3_fprintf(p->out, "Nothing matches '%s'\n", azArg[1]);
27940	29317	}
27941	29318	}else{
27942	29319	showHelp(p->out, 0);
27943	29320	}
27944	29321	}else
		@@ -27977,11 +29354,11 @@
27977	29354	if( zFile==0 ){
27978	29355	zFile = z;
27979	29356	}else if( zTable==0 ){
27980	29357	zTable = z;
27981	29358	}else{
27982		- oputf("ERROR: extra argument: \"%s\". Usage:\n", z);
	29359	+ sqlite3_fprintf(p->out, "ERROR: extra argument: \"%s\". Usage:\n",z);
27983	29360	showHelp(p->out, "import");
27984	29361	goto meta_command_exit;
27985	29362	}
27986	29363	}else if( cli_strcmp(z,"-v")==0 ){
27987	29364	eVerbose++;
		@@ -27998,17 +29375,17 @@
27998	29375	sCtx.cColSep = ',';
27999	29376	sCtx.cRowSep = '\n';
28000	29377	xRead = csv_read_one_field;
28001	29378	useOutputMode = 0;
28002	29379	}else{
28003		- oputf("ERROR: unknown option: \"%s\". Usage:\n", z);
	29380	+ sqlite3_fprintf(p->out, "ERROR: unknown option: \"%s\". Usage:\n", z);
28004	29381	showHelp(p->out, "import");
28005	29382	goto meta_command_exit;
28006	29383	}
28007	29384	}
28008	29385	if( zTable==0 ){
28009		- oputf("ERROR: missing %s argument. Usage:\n",
	29386	+ sqlite3_fprintf(p->out, "ERROR: missing %s argument. Usage:\n",
28010	29387	zFile==0 ? "FILE" : "TABLE");
28011	29388	showHelp(p->out, "import");
28012	29389	goto meta_command_exit;
28013	29390	}
28014	29391	seenInterrupt = 0;
		@@ -28054,32 +29431,32 @@
28054	29431	if( sCtx.zFile[0]=='\|' ){
28055	29432	#ifdef SQLITE_OMIT_POPEN
28056	29433	eputz("Error: pipes are not supported in this OS\n");
28057	29434	goto meta_command_exit;
28058	29435	#else
28059		- sCtx.in = popen(sCtx.zFile+1, "r");
	29436	+ sCtx.in = sqlite3_popen(sCtx.zFile+1, "r");
28060	29437	sCtx.zFile = "<pipe>";
28061	29438	sCtx.xCloser = pclose;
28062	29439	#endif
28063	29440	}else{
28064		- sCtx.in = fopen(sCtx.zFile, "rb");
	29441	+ sCtx.in = sqlite3_fopen(sCtx.zFile, "rb");
28065	29442	sCtx.xCloser = fclose;
28066	29443	}
28067	29444	if( sCtx.in==0 ){
28068		- eputf("Error: cannot open \"%s\"\n", zFile);
	29445	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zFile);
28069	29446	goto meta_command_exit;
28070	29447	}
28071	29448	if( eVerbose>=2 \|\| (eVerbose>=1 && useOutputMode) ){
28072	29449	char zSep[2];
28073	29450	zSep[1] = 0;
28074	29451	zSep[0] = sCtx.cColSep;
28075		- oputz("Column separator ");
28076		- output_c_string(zSep);
28077		- oputz(", row separator ");
	29452	+ sqlite3_fputs("Column separator ", p->out);
	29453	+ output_c_string(p->out, zSep);
	29454	+ sqlite3_fputs(", row separator ", p->out);
28078	29455	zSep[0] = sCtx.cRowSep;
28079		- output_c_string(zSep);
28080		- oputz("\n");
	29456	+ output_c_string(p->out, zSep);
	29457	+ sqlite3_fputs("\n", p->out);
28081	29458	}
28082	29459	sCtx.z = sqlite3_malloc64(120);
28083	29460	if( sCtx.z==0 ){
28084	29461	import_cleanup(&sCtx);
28085	29462	shell_out_of_memory();
		@@ -28087,11 +29464,15 @@
28087	29464	/* Below, resources must be freed before exit. */
28088	29465	while( (nSkip--)>0 ){
28089	29466	while( xRead(&sCtx) && sCtx.cTerm==sCtx.cColSep ){}
28090	29467	}
28091	29468	import_append_char(&sCtx, 0); /* To ensure sCtx.z is allocated */
28092		- if( sqlite3_table_column_metadata(p->db, zSchema, zTable,0,0,0,0,0,0) ){
	29469	+ if( sqlite3_table_column_metadata(p->db, zSchema, zTable,0,0,0,0,0,0)
	29470	+ && 0==db_int(p->db, "SELECT count(*) FROM \"%w\".sqlite_schema"
	29471	+ " WHERE name=%Q AND type='view'",
	29472	+ zSchema ? zSchema : "main", zTable)
	29473	+ ){
28093	29474	/* Table does not exist. Create it. */
28094	29475	sqlite3 *dbCols = 0;
28095	29476	char *zRenames = 0;
28096	29477	char *zColDefs;
28097	29478	zCreate = sqlite3_mprintf("CREATE TABLE \"%w\".\"%w\"",
		@@ -28100,18 +29481,18 @@
28100	29481	zAutoColumn(sCtx.z, &dbCols, 0);
28101	29482	if( sCtx.cTerm!=sCtx.cColSep ) break;
28102	29483	}
28103	29484	zColDefs = zAutoColumn(0, &dbCols, &zRenames);
28104	29485	if( zRenames!=0 ){
28105		- sputf((stdin_is_interactive && p->in==stdin)? p->out : stderr,
	29486	+ sqlite3_fprintf((stdin_is_interactive && p->in==stdin)? p->out : stderr,
28106	29487	"Columns renamed during .import %s due to duplicates:\n"
28107	29488	"%s\n", sCtx.zFile, zRenames);
28108	29489	sqlite3_free(zRenames);
28109	29490	}
28110	29491	assert(dbCols==0);
28111	29492	if( zColDefs==0 ){
28112		- eputf("%s: empty file\n", sCtx.zFile);
	29493	+ sqlite3_fprintf(stderr,"%s: empty file\n", sCtx.zFile);
28113	29494	import_cleanup(&sCtx);
28114	29495	rc = 1;
28115	29496	sqlite3_free(zCreate);
28116	29497	goto meta_command_exit;
28117	29498	}
		@@ -28119,17 +29500,20 @@
28119	29500	if( zCreate==0 ){
28120	29501	import_cleanup(&sCtx);
28121	29502	shell_out_of_memory();
28122	29503	}
28123	29504	if( eVerbose>=1 ){
28124		- oputf("%s\n", zCreate);
	29505	+ sqlite3_fprintf(p->out, "%s\n", zCreate);
28125	29506	}
28126	29507	rc = sqlite3_exec(p->db, zCreate, 0, 0, 0);
	29508	+ if( rc ){
	29509	+ sqlite3_fprintf(stderr,
	29510	+ "%s failed:\n%s\n", zCreate, sqlite3_errmsg(p->db));
	29511	+ }
28127	29512	sqlite3_free(zCreate);
28128	29513	zCreate = 0;
28129	29514	if( rc ){
28130		- eputf("%s failed:\n%s\n", zCreate, sqlite3_errmsg(p->db));
28131	29515	import_cleanup(&sCtx);
28132	29516	rc = 1;
28133	29517	goto meta_command_exit;
28134	29518	}
28135	29519	}
		@@ -28180,11 +29564,11 @@
28180	29564	}
28181	29565	zSql[j++] = ')';
28182	29566	zSql[j] = 0;
28183	29567	assert( j<nByte );
28184	29568	if( eVerbose>=2 ){
28185		- oputf("Insert using: %s\n", zSql);
	29569	+ sqlite3_fprintf(p->out, "Insert using: %s\n", zSql);
28186	29570	}
28187	29571	rc = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0);
28188	29572	sqlite3_free(zSql);
28189	29573	zSql = 0;
28190	29574	if( rc ){
		@@ -28219,11 +29603,11 @@
28219	29603	if( z==0 && (xRead==csv_read_one_field) && i==nCol-1 && i>0 ){
28220	29604	z = "";
28221	29605	}
28222	29606	sqlite3_bind_text(pStmt, i+1, z, -1, SQLITE_TRANSIENT);
28223	29607	if( i<nCol-1 && sCtx.cTerm!=sCtx.cColSep ){
28224		- eputf("%s:%d: expected %d columns but found %d"
	29608	+ sqlite3_fprintf(stderr,"%s:%d: expected %d columns but found %d"
28225	29609	" - filling the rest with NULL\n",
28226	29610	sCtx.zFile, startLine, nCol, i+1);
28227	29611	i += 2;
28228	29612	while( i<=nCol ){ sqlite3_bind_null(pStmt, i); i++; }
28229	29613	}
		@@ -28231,18 +29615,19 @@
28231	29615	if( sCtx.cTerm==sCtx.cColSep ){
28232	29616	do{
28233	29617	xRead(&sCtx);
28234	29618	i++;
28235	29619	}while( sCtx.cTerm==sCtx.cColSep );
28236		- eputf("%s:%d: expected %d columns but found %d - extras ignored\n",
	29620	+ sqlite3_fprintf(stderr,
	29621	+ "%s:%d: expected %d columns but found %d - extras ignored\n",
28237	29622	sCtx.zFile, startLine, nCol, i);
28238	29623	}
28239	29624	if( i>=nCol ){
28240	29625	sqlite3_step(pStmt);
28241	29626	rc = sqlite3_reset(pStmt);
28242	29627	if( rc!=SQLITE_OK ){
28243		- eputf("%s:%d: INSERT failed: %s\n",
	29628	+ sqlite3_fprintf(stderr,"%s:%d: INSERT failed: %s\n",
28244	29629	sCtx.zFile, startLine, sqlite3_errmsg(p->db));
28245	29630	sCtx.nErr++;
28246	29631	}else{
28247	29632	sCtx.nRow++;
28248	29633	}
		@@ -28251,11 +29636,12 @@
28251	29636
28252	29637	import_cleanup(&sCtx);
28253	29638	sqlite3_finalize(pStmt);
28254	29639	if( needCommit ) sqlite3_exec(p->db, "COMMIT", 0, 0, 0);
28255	29640	if( eVerbose>0 ){
28256		- oputf("Added %d rows with %d errors using %d lines of input\n",
	29641	+ sqlite3_fprintf(p->out,
	29642	+ "Added %d rows with %d errors using %d lines of input\n",
28257	29643	sCtx.nRow, sCtx.nErr, sCtx.nLine-1);
28258	29644	}
28259	29645	}else
28260	29646	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
28261	29647
		@@ -28267,11 +29653,11 @@
28267	29653	int tnum = 0;
28268	29654	int isWO = 0; /* True if making an imposter of a WITHOUT ROWID table */
28269	29655	int lenPK = 0; /* Length of the PRIMARY KEY string for isWO tables */
28270	29656	int i;
28271	29657	if( !ShellHasFlag(p,SHFLG_TestingMode) ){
28272		- eputf(".%s unavailable without --unsafe-testing\n",
	29658	+ sqlite3_fprintf(stderr,".%s unavailable without --unsafe-testing\n",
28273	29659	"imposter");
28274	29660	rc = 1;
28275	29661	goto meta_command_exit;
28276	29662	}
28277	29663	if( !(nArg==3 \|\| (nArg==2 && sqlite3_stricmp(azArg[1],"off")==0)) ){
		@@ -28333,11 +29719,11 @@
28333	29719	zCollist = sqlite3_mprintf("%z,\"%w\"", zCollist, zCol);
28334	29720	}
28335	29721	}
28336	29722	sqlite3_finalize(pStmt);
28337	29723	if( i==0 \|\| tnum==0 ){
28338		- eputf("no such index: \"%s\"\n", azArg[1]);
	29724	+ sqlite3_fprintf(stderr,"no such index: \"%s\"\n", azArg[1]);
28339	29725	rc = 1;
28340	29726	sqlite3_free(zCollist);
28341	29727	goto meta_command_exit;
28342	29728	}
28343	29729	if( lenPK==0 ) lenPK = 100000;
		@@ -28348,18 +29734,20 @@
28348	29734	rc = sqlite3_test_control(SQLITE_TESTCTRL_IMPOSTER, p->db, "main", 1, tnum);
28349	29735	if( rc==SQLITE_OK ){
28350	29736	rc = sqlite3_exec(p->db, zSql, 0, 0, 0);
28351	29737	sqlite3_test_control(SQLITE_TESTCTRL_IMPOSTER, p->db, "main", 0, 0);
28352	29738	if( rc ){
28353		- eputf("Error in [%s]: %s\n", zSql, sqlite3_errmsg(p->db));
	29739	+ sqlite3_fprintf(stderr,
	29740	+ "Error in [%s]: %s\n", zSql, sqlite3_errmsg(p->db));
28354	29741	}else{
28355		- sputf(stdout, "%s;\n", zSql);
28356		- sputf(stdout, "WARNING: writing to an imposter table will corrupt"
	29742	+ sqlite3_fprintf(stdout, "%s;\n", zSql);
	29743	+ sqlite3_fprintf(stdout,
	29744	+ "WARNING: writing to an imposter table will corrupt"
28357	29745	" the \"%s\" %s!\n", azArg[1], isWO ? "table" : "index");
28358	29746	}
28359	29747	}else{
28360		- eputf("SQLITE_TESTCTRL_IMPOSTER returns %d\n", rc);
	29748	+ sqlite3_fprintf(stderr,"SQLITE_TESTCTRL_IMPOSTER returns %d\n", rc);
28361	29749	rc = 1;
28362	29750	}
28363	29751	sqlite3_free(zSql);
28364	29752	}else
28365	29753	#endif /* !defined(SQLITE_OMIT_TEST_CONTROL) */
		@@ -28369,11 +29757,11 @@
28369	29757	if( nArg==2 ){
28370	29758	iArg = integerValue(azArg[1]);
28371	29759	if( iArg==0 ) iArg = -1;
28372	29760	}
28373	29761	if( (nArg!=1 && nArg!=2) \|\| iArg<0 ){
28374		- eputf("%s","Usage: .intck STEPS_PER_UNLOCK\n");
	29762	+ sqlite3_fprintf(stderr,"%s","Usage: .intck STEPS_PER_UNLOCK\n");
28375	29763	rc = 1;
28376	29764	goto meta_command_exit;
28377	29765	}
28378	29766	open_db(p, 0);
28379	29767	rc = intckDatabaseCmd(p, iArg);
		@@ -28388,13 +29776,13 @@
28388	29776	sqlite3IoTrace = 0;
28389	29777	}else if( cli_strcmp(azArg[1], "-")==0 ){
28390	29778	sqlite3IoTrace = iotracePrintf;
28391	29779	iotrace = stdout;
28392	29780	}else{
28393		- iotrace = fopen(azArg[1], "w");
	29781	+ iotrace = sqlite3_fopen(azArg[1], "w");
28394	29782	if( iotrace==0 ){
28395		- eputf("Error: cannot open \"%s\"\n", azArg[1]);
	29783	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", azArg[1]);
28396	29784	sqlite3IoTrace = 0;
28397	29785	rc = 1;
28398	29786	}else{
28399	29787	sqlite3IoTrace = iotracePrintf;
28400	29788	}
		@@ -28422,11 +29810,11 @@
28422	29810	};
28423	29811	int i, n2;
28424	29812	open_db(p, 0);
28425	29813	if( nArg==1 ){
28426	29814	for(i=0; i<ArraySize(aLimit); i++){
28427		- sputf(stdout, "%20s %d\n", aLimit[i].zLimitName,
	29815	+ sqlite3_fprintf(stdout, "%20s %d\n", aLimit[i].zLimitName,
28428	29816	sqlite3_limit(p->db, aLimit[i].limitCode, -1));
28429	29817	}
28430	29818	}else if( nArg>3 ){
28431	29819	eputz("Usage: .limit NAME ?NEW-VALUE?\n");
28432	29820	rc = 1;
		@@ -28437,28 +29825,28 @@
28437	29825	for(i=0; i<ArraySize(aLimit); i++){
28438	29826	if( sqlite3_strnicmp(aLimit[i].zLimitName, azArg[1], n2)==0 ){
28439	29827	if( iLimit<0 ){
28440	29828	iLimit = i;
28441	29829	}else{
28442		- eputf("ambiguous limit: \"%s\"\n", azArg[1]);
	29830	+ sqlite3_fprintf(stderr,"ambiguous limit: \"%s\"\n", azArg[1]);
28443	29831	rc = 1;
28444	29832	goto meta_command_exit;
28445	29833	}
28446	29834	}
28447	29835	}
28448	29836	if( iLimit<0 ){
28449		- eputf("unknown limit: \"%s\"\n"
	29837	+ sqlite3_fprintf(stderr,"unknown limit: \"%s\"\n"
28450	29838	"enter \".limits\" with no arguments for a list.\n",
28451	29839	azArg[1]);
28452	29840	rc = 1;
28453	29841	goto meta_command_exit;
28454	29842	}
28455	29843	if( nArg==3 ){
28456	29844	sqlite3_limit(p->db, aLimit[iLimit].limitCode,
28457	29845	(int)integerValue(azArg[2]));
28458	29846	}
28459		- sputf(stdout, "%20s %d\n", aLimit[iLimit].zLimitName,
	29847	+ sqlite3_fprintf(stdout, "%20s %d\n", aLimit[iLimit].zLimitName,
28460	29848	sqlite3_limit(p->db, aLimit[iLimit].limitCode, -1));
28461	29849	}
28462	29850	}else
28463	29851
28464	29852	if( c=='l' && n>2 && cli_strncmp(azArg[0], "lint", n)==0 ){
		@@ -28503,11 +29891,11 @@
28503	29891	" than \"on\" or \"off\"\n");
28504	29892	zFile = "off";
28505	29893	}
28506	29894	output_file_close(p->pLog);
28507	29895	if( cli_strcmp(zFile,"on")==0 ) zFile = "stdout";
28508		- p->pLog = output_file_open(zFile, 0);
	29896	+ p->pLog = output_file_open(zFile);
28509	29897	}
28510	29898	}else
28511	29899
28512	29900	if( c=='m' && cli_strncmp(azArg[0], "mode", n)==0 ){
28513	29901	const char *zMode = 0;
		@@ -28537,35 +29925,37 @@
28537	29925	cmOpts = cmo;
28538	29926	}
28539	29927	}else if( zTabname==0 ){
28540	29928	zTabname = z;
28541	29929	}else if( z[0]=='-' ){
28542		- eputf("unknown option: %s\n", z);
	29930	+ sqlite3_fprintf(stderr,"unknown option: %s\n", z);
28543	29931	eputz("options:\n"
28544	29932	" --noquote\n"
28545	29933	" --quote\n"
28546	29934	" --wordwrap on/off\n"
28547	29935	" --wrap N\n"
28548	29936	" --ww\n");
28549	29937	rc = 1;
28550	29938	goto meta_command_exit;
28551	29939	}else{
28552		- eputf("extra argument: \"%s\"\n", z);
	29940	+ sqlite3_fprintf(stderr,"extra argument: \"%s\"\n", z);
28553	29941	rc = 1;
28554	29942	goto meta_command_exit;
28555	29943	}
28556	29944	}
28557	29945	if( zMode==0 ){
28558	29946	if( p->mode==MODE_Column
28559	29947	\|\| (p->mode>=MODE_Markdown && p->mode<=MODE_Box)
28560	29948	){
28561		- oputf("current output mode: %s --wrap %d --wordwrap %s --%squote\n",
	29949	+ sqlite3_fprintf(p->out,
	29950	+ "current output mode: %s --wrap %d --wordwrap %s --%squote\n",
28562	29951	modeDescr[p->mode], p->cmOpts.iWrap,
28563	29952	p->cmOpts.bWordWrap ? "on" : "off",
28564	29953	p->cmOpts.bQuote ? "" : "no");
28565	29954	}else{
28566		- oputf("current output mode: %s\n", modeDescr[p->mode]);
	29955	+ sqlite3_fprintf(p->out,
	29956	+ "current output mode: %s\n", modeDescr[p->mode]);
28567	29957	}
28568	29958	zMode = modeDescr[p->mode];
28569	29959	}
28570	29960	n2 = strlen30(zMode);
28571	29961	if( cli_strncmp(zMode,"lines",n2)==0 ){
		@@ -28634,11 +30024,11 @@
28634	30024	if( c=='n' && cli_strcmp(azArg[0], "nonce")==0 ){
28635	30025	if( nArg!=2 ){
28636	30026	eputz("Usage: .nonce NONCE\n");
28637	30027	rc = 1;
28638	30028	}else if( p->zNonce==0 \|\| cli_strcmp(azArg[1],p->zNonce)!=0 ){
28639		- eputf("line %d: incorrect nonce: \"%s\"\n",
	30029	+ sqlite3_fprintf(stderr,"line %d: incorrect nonce: \"%s\"\n",
28640	30030	p->lineno, azArg[1]);
28641	30031	exit(1);
28642	30032	}else{
28643	30033	p->bSafeMode = 0;
28644	30034	return 0; /* Return immediately to bypass the safe mode reset
		@@ -28689,15 +30079,15 @@
28689	30079	p->szMax = integerValue(azArg[++iName]);
28690	30080	#endif /* SQLITE_OMIT_DESERIALIZE */
28691	30081	}else
28692	30082	#endif /* !SQLITE_SHELL_FIDDLE */
28693	30083	if( z[0]=='-' ){
28694		- eputf("unknown option: %s\n", z);
	30084	+ sqlite3_fprintf(stderr,"unknown option: %s\n", z);
28695	30085	rc = 1;
28696	30086	goto meta_command_exit;
28697	30087	}else if( zFN ){
28698		- eputf("extra argument: \"%s\"\n", z);
	30088	+ sqlite3_fprintf(stderr,"extra argument: \"%s\"\n", z);
28699	30089	rc = 1;
28700	30090	goto meta_command_exit;
28701	30091	}else{
28702	30092	zFN = z;
28703	30093	}
		@@ -28735,11 +30125,11 @@
28735	30125	zNewFilename = 0;
28736	30126	}
28737	30127	p->pAuxDb->zDbFilename = zNewFilename;
28738	30128	open_db(p, OPEN_DB_KEEPALIVE);
28739	30129	if( p->db==0 ){
28740		- eputf("Error: cannot open '%s'\n", zNewFilename);
	30130	+ sqlite3_fprintf(stderr,"Error: cannot open '%s'\n", zNewFilename);
28741	30131	sqlite3_free(zNewFilename);
28742	30132	}else{
28743	30133	p->pAuxDb->zFreeOnClose = zNewFilename;
28744	30134	}
28745	30135	}
		@@ -28753,22 +30143,26 @@
28753	30143	#ifndef SQLITE_SHELL_FIDDLE
28754	30144	if( (c=='o'
28755	30145	&& (cli_strncmp(azArg[0], "output", n)==0
28756	30146	\|\| cli_strncmp(azArg[0], "once", n)==0))
28757	30147	\|\| (c=='e' && n==5 && cli_strcmp(azArg[0],"excel")==0)
	30148	+ \|\| (c=='w' && n==3 && cli_strcmp(azArg[0],"www")==0)
28758	30149	){
28759	30150	char *zFile = 0;
28760		- int bTxtMode = 0;
28761	30151	int i;
28762	30152	int eMode = 0;
28763		- int bOnce = 0; /* 0: .output, 1: .once, 2: .excel */
28764		- static const char *zBomUtf8 = "\xef\xbb\xbf";
	30153	+ int bOnce = 0; /* 0: .output, 1: .once, 2: .excel/.www */
	30154	+ int bPlain = 0; /* --plain option */
	30155	+ static const char *zBomUtf8 = "\357\273\277";
28765	30156	const char *zBom = 0;
28766	30157
28767	30158	failIfSafeMode(p, "cannot run .%s in safe mode", azArg[0]);
28768	30159	if( c=='e' ){
28769	30160	eMode = 'x';
	30161	+ bOnce = 2;
	30162	+ }else if( c=='w' ){
	30163	+ eMode = 'w';
28770	30164	bOnce = 2;
28771	30165	}else if( cli_strncmp(azArg[0],"once",n)==0 ){
28772	30166	bOnce = 1;
28773	30167	}
28774	30168	for(i=1; i<nArg; i++){
		@@ -28775,28 +30169,34 @@
28775	30169	char *z = azArg[i];
28776	30170	if( z[0]=='-' ){
28777	30171	if( z[1]=='-' ) z++;
28778	30172	if( cli_strcmp(z,"-bom")==0 ){
28779	30173	zBom = zBomUtf8;
28780		- }else if( c!='e' && cli_strcmp(z,"-x")==0 ){
	30174	+ }else if( cli_strcmp(z,"-plain")==0 ){
	30175	+ bPlain = 1;
	30176	+ }else if( c=='o' && cli_strcmp(z,"-x")==0 ){
28781	30177	eMode = 'x'; /* spreadsheet */
28782		- }else if( c!='e' && cli_strcmp(z,"-e")==0 ){
	30178	+ }else if( c=='o' && cli_strcmp(z,"-e")==0 ){
28783	30179	eMode = 'e'; /* text editor */
	30180	+ }else if( c=='o' && cli_strcmp(z,"-w")==0 ){
	30181	+ eMode = 'w'; /* Web browser */
28784	30182	}else{
28785		- oputf("ERROR: unknown option: \"%s\". Usage:\n", azArg[i]);
	30183	+ sqlite3_fprintf(p->out,
	30184	+ "ERROR: unknown option: \"%s\". Usage:\n", azArg[i]);
28786	30185	showHelp(p->out, azArg[0]);
28787	30186	rc = 1;
28788	30187	goto meta_command_exit;
28789	30188	}
28790		- }else if( zFile==0 && eMode!='e' && eMode!='x' ){
	30189	+ }else if( zFile==0 && eMode==0 ){
28791	30190	zFile = sqlite3_mprintf("%s", z);
28792	30191	if( zFile && zFile[0]=='\|' ){
28793	30192	while( i+1<nArg ) zFile = sqlite3_mprintf("%z %s", zFile, azArg[++i]);
28794	30193	break;
28795	30194	}
28796	30195	}else{
28797		- oputf("ERROR: extra parameter: \"%s\". Usage:\n", azArg[i]);
	30196	+ sqlite3_fprintf(p->out,
	30197	+ "ERROR: extra parameter: \"%s\". Usage:\n", azArg[i]);
28798	30198	showHelp(p->out, azArg[0]);
28799	30199	rc = 1;
28800	30200	sqlite3_free(zFile);
28801	30201	goto meta_command_exit;
28802	30202	}
		@@ -28809,24 +30209,31 @@
28809	30209	}else{
28810	30210	p->outCount = 0;
28811	30211	}
28812	30212	output_reset(p);
28813	30213	#ifndef SQLITE_NOHAVE_SYSTEM
28814		- if( eMode=='e' \|\| eMode=='x' ){
	30214	+ if( eMode=='e' \|\| eMode=='x' \|\| eMode=='w' ){
28815	30215	p->doXdgOpen = 1;
28816	30216	outputModePush(p);
28817	30217	if( eMode=='x' ){
28818	30218	/* spreadsheet mode. Output as CSV. */
28819	30219	newTempFile(p, "csv");
28820	30220	ShellClearFlag(p, SHFLG_Echo);
28821	30221	p->mode = MODE_Csv;
28822	30222	sqlite3_snprintf(sizeof(p->colSeparator), p->colSeparator, SEP_Comma);
28823	30223	sqlite3_snprintf(sizeof(p->rowSeparator), p->rowSeparator, SEP_CrLf);
	30224	+#ifdef _WIN32
	30225	+ zBom = zBomUtf8; /* Always include the BOM on Windows, as Excel does
	30226	+ ** not work without it. */
	30227	+#endif
	30228	+ }else if( eMode=='w' ){
	30229	+ /* web-browser mode. */
	30230	+ newTempFile(p, "html");
	30231	+ if( !bPlain ) p->mode = MODE_Www;
28824	30232	}else{
28825	30233	/* text editor mode */
28826	30234	newTempFile(p, "txt");
28827		- bTxtMode = 1;
28828	30235	}
28829	30236	sqlite3_free(zFile);
28830	30237	zFile = sqlite3_mprintf("%s", p->zTempFile);
28831	30238	}
28832	30239	#endif /* SQLITE_NOHAVE_SYSTEM */
		@@ -28835,30 +30242,36 @@
28835	30242	#ifdef SQLITE_OMIT_POPEN
28836	30243	eputz("Error: pipes are not supported in this OS\n");
28837	30244	rc = 1;
28838	30245	output_redir(p, stdout);
28839	30246	#else
28840		- FILE *pfPipe = popen(zFile + 1, "w");
	30247	+ FILE *pfPipe = sqlite3_popen(zFile + 1, "w");
28841	30248	if( pfPipe==0 ){
28842		- eputf("Error: cannot open pipe \"%s\"\n", zFile + 1);
	30249	+ sqlite3_fprintf(stderr,"Error: cannot open pipe \"%s\"\n", zFile + 1);
28843	30250	rc = 1;
28844	30251	}else{
28845	30252	output_redir(p, pfPipe);
28846		- if( zBom ) oputz(zBom);
	30253	+ if( zBom ) sqlite3_fputs(zBom, pfPipe);
28847	30254	sqlite3_snprintf(sizeof(p->outfile), p->outfile, "%s", zFile);
28848	30255	}
28849	30256	#endif
28850	30257	}else{
28851		- FILE *pfFile = output_file_open(zFile, bTxtMode);
	30258	+ FILE *pfFile = output_file_open(zFile);
28852	30259	if( pfFile==0 ){
28853	30260	if( cli_strcmp(zFile,"off")!=0 ){
28854		- eputf("Error: cannot write to \"%s\"\n", zFile);
	30261	+ sqlite3_fprintf(stderr,"Error: cannot write to \"%s\"\n", zFile);
28855	30262	}
28856	30263	rc = 1;
28857	30264	} else {
28858	30265	output_redir(p, pfFile);
28859		- if( zBom ) oputz(zBom);
	30266	+ if( zBom ) sqlite3_fputs(zBom, pfFile);
	30267	+ if( bPlain && eMode=='w' ){
	30268	+ sqlite3_fputs(
	30269	+ "<!DOCTYPE html>\n<BODY>\n<PLAINTEXT>\n",
	30270	+ pfFile
	30271	+ );
	30272	+ }
28860	30273	sqlite3_snprintf(sizeof(p->outfile), p->outfile, "%s", zFile);
28861	30274	}
28862	30275	}
28863	30276	sqlite3_free(zFile);
28864	30277	}else
		@@ -28895,11 +30308,12 @@
28895	30308	if( len ){
28896	30309	rx = sqlite3_prepare_v2(p->db,
28897	30310	"SELECT key, quote(value) "
28898	30311	"FROM temp.sqlite_parameters;", -1, &pStmt, 0);
28899	30312	while( rx==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW ){
28900		- oputf("%-*s %s\n", len, sqlite3_column_text(pStmt,0),
	30313	+ sqlite3_fprintf(p->out,
	30314	+ "%-*s %s\n", len, sqlite3_column_text(pStmt,0),
28901	30315	sqlite3_column_text(pStmt,1));
28902	30316	}
28903	30317	sqlite3_finalize(pStmt);
28904	30318	}
28905	30319	}else
		@@ -28940,11 +30354,11 @@
28940	30354	"VALUES(%Q,%Q);", zKey, zValue);
28941	30355	shell_check_oom(zSql);
28942	30356	rx = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0);
28943	30357	sqlite3_free(zSql);
28944	30358	if( rx!=SQLITE_OK ){
28945		- oputf("Error: %s\n", sqlite3_errmsg(p->db));
	30359	+ sqlite3_fprintf(p->out, "Error: %s\n", sqlite3_errmsg(p->db));
28946	30360	sqlite3_finalize(pStmt);
28947	30361	pStmt = 0;
28948	30362	rc = 1;
28949	30363	}
28950	30364	}
		@@ -28969,14 +30383,14 @@
28969	30383	}else
28970	30384
28971	30385	if( c=='p' && n>=3 && cli_strncmp(azArg[0], "print", n)==0 ){
28972	30386	int i;
28973	30387	for(i=1; i<nArg; i++){
28974		- if( i>1 ) oputz(" ");
28975		- oputz(azArg[i]);
	30388	+ if( i>1 ) sqlite3_fputs(" ", p->out);
	30389	+ sqlite3_fputs(azArg[i], p->out);
28976	30390	}
28977		- oputz("\n");
	30391	+ sqlite3_fputs("\n", p->out);
28978	30392	}else
28979	30393
28980	30394	#ifndef SQLITE_OMIT_PROGRESS_CALLBACK
28981	30395	if( c=='p' && n>=3 && cli_strncmp(azArg[0], "progress", n)==0 ){
28982	30396	int i;
		@@ -29009,11 +30423,11 @@
29009	30423	}else{
29010	30424	p->mxProgress = (int)integerValue(azArg[++i]);
29011	30425	}
29012	30426	continue;
29013	30427	}
29014		- eputf("Error: unknown option: \"%s\"\n", azArg[i]);
	30428	+ sqlite3_fprintf(stderr,"Error: unknown option: \"%s\"\n", azArg[i]);
29015	30429	rc = 1;
29016	30430	goto meta_command_exit;
29017	30431	}else{
29018	30432	nn = (int)integerValue(z);
29019	30433	}
		@@ -29050,23 +30464,22 @@
29050	30464	}
29051	30465	if( azArg[1][0]=='\|' ){
29052	30466	#ifdef SQLITE_OMIT_POPEN
29053	30467	eputz("Error: pipes are not supported in this OS\n");
29054	30468	rc = 1;
29055		- p->out = stdout;
29056	30469	#else
29057		- p->in = popen(azArg[1]+1, "r");
	30470	+ p->in = sqlite3_popen(azArg[1]+1, "r");
29058	30471	if( p->in==0 ){
29059		- eputf("Error: cannot open \"%s\"\n", azArg[1]);
	30472	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", azArg[1]);
29060	30473	rc = 1;
29061	30474	}else{
29062	30475	rc = process_input(p);
29063	30476	pclose(p->in);
29064	30477	}
29065	30478	#endif
29066	30479	}else if( (p->in = openChrSource(azArg[1]))==0 ){
29067		- eputf("Error: cannot open \"%s\"\n", azArg[1]);
	30480	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", azArg[1]);
29068	30481	rc = 1;
29069	30482	}else{
29070	30483	rc = process_input(p);
29071	30484	fclose(p->in);
29072	30485	}
		@@ -29095,11 +30508,11 @@
29095	30508	rc = 1;
29096	30509	goto meta_command_exit;
29097	30510	}
29098	30511	rc = sqlite3_open(zSrcFile, &pSrc);
29099	30512	if( rc!=SQLITE_OK ){
29100		- eputf("Error: cannot open \"%s\"\n", zSrcFile);
	30513	+ sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zSrcFile);
29101	30514	close_db(pSrc);
29102	30515	return 1;
29103	30516	}
29104	30517	open_db(p, 0);
29105	30518	pBackup = sqlite3_backup_init(p->db, zDb, pSrc, "main");
		@@ -29127,11 +30540,14 @@
29127	30540	}
29128	30541	close_db(pSrc);
29129	30542	}else
29130	30543	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
29131	30544
29132		- if( c=='s' && cli_strncmp(azArg[0], "scanstats", n)==0 ){
	30545	+ if( c=='s' &&
	30546	+ (cli_strncmp(azArg[0], "scanstats", n)==0 \|\|
	30547	+ cli_strncmp(azArg[0], "scanstatus", n)==0)
	30548	+ ){
29133	30549	if( nArg==2 ){
29134	30550	if( cli_strcmp(azArg[1], "vm")==0 ){
29135	30551	p->scanstatsOn = 3;
29136	30552	}else
29137	30553	if( cli_strcmp(azArg[1], "est")==0 ){
		@@ -29178,11 +30594,11 @@
29178	30594	}else if( optionMatch(azArg[ii],"debug") ){
29179	30595	bDebug = 1;
29180	30596	}else if( optionMatch(azArg[ii],"nosys") ){
29181	30597	bNoSystemTabs = 1;
29182	30598	}else if( azArg[ii][0]=='-' ){
29183		- eputf("Unknown option: \"%s\"\n", azArg[ii]);
	30599	+ sqlite3_fprintf(stderr,"Unknown option: \"%s\"\n", azArg[ii]);
29184	30600	rc = 1;
29185	30601	goto meta_command_exit;
29186	30602	}else if( zName==0 ){
29187	30603	zName = azArg[ii];
29188	30604	}else{
		@@ -29279,11 +30695,11 @@
29279	30695	appendText(&sSelect, "name NOT LIKE 'sqlite_%%' AND ", 0);
29280	30696	}
29281	30697	appendText(&sSelect, "sql IS NOT NULL"
29282	30698	" ORDER BY snum, rowid", 0);
29283	30699	if( bDebug ){
29284		- oputf("SQL: %s;\n", sSelect.z);
	30700	+ sqlite3_fprintf(p->out, "SQL: %s;\n", sSelect.z);
29285	30701	}else{
29286	30702	rc = sqlite3_exec(p->db, sSelect.z, callback, &data, &zErrMsg);
29287	30703	}
29288	30704	freeText(&sSelect);
29289	30705	}
		@@ -29340,11 +30756,12 @@
29340	30756	session_not_open:
29341	30757	eputz("ERROR: No sessions are open\n");
29342	30758	}else{
29343	30759	rc = sqlite3session_attach(pSession->p, azCmd[1]);
29344	30760	if( rc ){
29345		- eputf("ERROR: sqlite3session_attach() returns %d\n",rc);
	30761	+ sqlite3_fprintf(stderr,
	30762	+ "ERROR: sqlite3session_attach() returns %d\n",rc);
29346	30763	rc = 0;
29347	30764	}
29348	30765	}
29349	30766	}else
29350	30767
		@@ -29357,13 +30774,13 @@
29357	30774	){
29358	30775	FILE *out = 0;
29359	30776	failIfSafeMode(p, "cannot run \".session %s\" in safe mode", azCmd[0]);
29360	30777	if( nCmd!=2 ) goto session_syntax_error;
29361	30778	if( pSession->p==0 ) goto session_not_open;
29362		- out = fopen(azCmd[1], "wb");
	30779	+ out = sqlite3_fopen(azCmd[1], "wb");
29363	30780	if( out==0 ){
29364		- eputf("ERROR: cannot open \"%s\" for writing\n",
	30781	+ sqlite3_fprintf(stderr,"ERROR: cannot open \"%s\" for writing\n",
29365	30782	azCmd[1]);
29366	30783	}else{
29367	30784	int szChng;
29368	30785	void *pChng;
29369	30786	if( azCmd[0][0]=='c' ){
		@@ -29370,16 +30787,17 @@
29370	30787	rc = sqlite3session_changeset(pSession->p, &szChng, &pChng);
29371	30788	}else{
29372	30789	rc = sqlite3session_patchset(pSession->p, &szChng, &pChng);
29373	30790	}
29374	30791	if( rc ){
29375		- sputf(stdout, "Error: error code %d\n", rc);
	30792	+ sqlite3_fprintf(stdout, "Error: error code %d\n", rc);
29376	30793	rc = 0;
29377	30794	}
29378	30795	if( pChng
29379	30796	&& fwrite(pChng, szChng, 1, out)!=1 ){
29380		- eputf("ERROR: Failed to write entire %d-byte output\n", szChng);
	30797	+ sqlite3_fprintf(stderr,
	30798	+ "ERROR: Failed to write entire %d-byte output\n", szChng);
29381	30799	}
29382	30800	sqlite3_free(pChng);
29383	30801	fclose(out);
29384	30802	}
29385	30803	}else
		@@ -29402,11 +30820,12 @@
29402	30820	int ii;
29403	30821	if( nCmd>2 ) goto session_syntax_error;
29404	30822	ii = nCmd==1 ? -1 : booleanValue(azCmd[1]);
29405	30823	if( pAuxDb->nSession ){
29406	30824	ii = sqlite3session_enable(pSession->p, ii);
29407		- oputf("session %s enable flag = %d\n", pSession->zName, ii);
	30825	+ sqlite3_fprintf(p->out,
	30826	+ "session %s enable flag = %d\n", pSession->zName, ii);
29408	30827	}
29409	30828	}else
29410	30829
29411	30830	/* .session filter GLOB ....
29412	30831	** Set a list of GLOB patterns of table names to be excluded.
		@@ -29437,11 +30856,12 @@
29437	30856	int ii;
29438	30857	if( nCmd>2 ) goto session_syntax_error;
29439	30858	ii = nCmd==1 ? -1 : booleanValue(azCmd[1]);
29440	30859	if( pAuxDb->nSession ){
29441	30860	ii = sqlite3session_indirect(pSession->p, ii);
29442		- oputf("session %s indirect flag = %d\n", pSession->zName, ii);
	30861	+ sqlite3_fprintf(p->out,
	30862	+ "session %s indirect flag = %d\n", pSession->zName, ii);
29443	30863	}
29444	30864	}else
29445	30865
29446	30866	/* .session isempty
29447	30867	** Determine if the session is empty
		@@ -29449,20 +30869,21 @@
29449	30869	if( cli_strcmp(azCmd[0], "isempty")==0 ){
29450	30870	int ii;
29451	30871	if( nCmd!=1 ) goto session_syntax_error;
29452	30872	if( pAuxDb->nSession ){
29453	30873	ii = sqlite3session_isempty(pSession->p);
29454		- oputf("session %s isempty flag = %d\n", pSession->zName, ii);
	30874	+ sqlite3_fprintf(p->out,
	30875	+ "session %s isempty flag = %d\n", pSession->zName, ii);
29455	30876	}
29456	30877	}else
29457	30878
29458	30879	/* .session list
29459	30880	** List all currently open sessions
29460	30881	*/
29461	30882	if( cli_strcmp(azCmd[0],"list")==0 ){
29462	30883	for(i=0; i<pAuxDb->nSession; i++){
29463		- oputf("%d %s\n", i, pAuxDb->aSession[i].zName);
	30884	+ sqlite3_fprintf(p->out, "%d %s\n", i, pAuxDb->aSession[i].zName);
29464	30885	}
29465	30886	}else
29466	30887
29467	30888	/* .session open DB NAME
29468	30889	** Open a new session called NAME on the attached database DB.
		@@ -29473,22 +30894,23 @@
29473	30894	if( nCmd!=3 ) goto session_syntax_error;
29474	30895	zName = azCmd[2];
29475	30896	if( zName[0]==0 ) goto session_syntax_error;
29476	30897	for(i=0; i<pAuxDb->nSession; i++){
29477	30898	if( cli_strcmp(pAuxDb->aSession[i].zName,zName)==0 ){
29478		- eputf("Session \"%s\" already exists\n", zName);
	30899	+ sqlite3_fprintf(stderr,"Session \"%s\" already exists\n", zName);
29479	30900	goto meta_command_exit;
29480	30901	}
29481	30902	}
29482	30903	if( pAuxDb->nSession>=ArraySize(pAuxDb->aSession) ){
29483		- eputf("Maximum of %d sessions\n", ArraySize(pAuxDb->aSession));
	30904	+ sqlite3_fprintf(stderr,
	30905	+ "Maximum of %d sessions\n", ArraySize(pAuxDb->aSession));
29484	30906	goto meta_command_exit;
29485	30907	}
29486	30908	pSession = &pAuxDb->aSession[pAuxDb->nSession];
29487	30909	rc = sqlite3session_create(p->db, azCmd[1], &pSession->p);
29488	30910	if( rc ){
29489		- eputf("Cannot open session: error code=%d\n", rc);
	30911	+ sqlite3_fprintf(stderr,"Cannot open session: error code=%d\n", rc);
29490	30912	rc = 0;
29491	30913	goto meta_command_exit;
29492	30914	}
29493	30915	pSession->nFilter = 0;
29494	30916	sqlite3session_table_filter(pSession->p, session_filter, pSession);
		@@ -29508,20 +30930,20 @@
29508	30930	if( c=='s' && n>=10 && cli_strncmp(azArg[0], "selftest-", 9)==0 ){
29509	30931	if( cli_strncmp(azArg[0]+9, "boolean", n-9)==0 ){
29510	30932	int i, v;
29511	30933	for(i=1; i<nArg; i++){
29512	30934	v = booleanValue(azArg[i]);
29513		- oputf("%s: %d 0x%x\n", azArg[i], v, v);
	30935	+ sqlite3_fprintf(p->out, "%s: %d 0x%x\n", azArg[i], v, v);
29514	30936	}
29515	30937	}
29516	30938	if( cli_strncmp(azArg[0]+9, "integer", n-9)==0 ){
29517	30939	int i; sqlite3_int64 v;
29518	30940	for(i=1; i<nArg; i++){
29519	30941	char zBuf[200];
29520	30942	v = integerValue(azArg[i]);
29521	30943	sqlite3_snprintf(sizeof(zBuf),zBuf,"%s: %lld 0x%llx\n", azArg[i],v,v);
29522		- oputz(zBuf);
	30944	+ sqlite3_fputs(zBuf, p->out);
29523	30945	}
29524	30946	}
29525	30947	}else
29526	30948	#endif
29527	30949
		@@ -29544,12 +30966,13 @@
29544	30966	}else
29545	30967	if( cli_strcmp(z,"-v")==0 ){
29546	30968	bVerbose++;
29547	30969	}else
29548	30970	{
29549		- eputf("Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
29550		- eputz("Should be one of: --init -v\n");
	30971	+ sqlite3_fprintf(stderr,
	30972	+ "Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
	30973	+ sqlite3_fputs("Should be one of: --init -v\n", stderr);
29551	30974	rc = 1;
29552	30975	goto meta_command_exit;
29553	30976	}
29554	30977	}
29555	30978	if( sqlite3_table_column_metadata(p->db,"main","selftest",0,0,0,0,0,0)
		@@ -29590,46 +31013,47 @@
29590	31013	if( zOp==0 ) continue;
29591	31014	if( zSql==0 ) continue;
29592	31015	if( zAns==0 ) continue;
29593	31016	k = 0;
29594	31017	if( bVerbose>0 ){
29595		- sputf(stdout, "%d: %s %s\n", tno, zOp, zSql);
	31018	+ sqlite3_fprintf(stdout, "%d: %s %s\n", tno, zOp, zSql);
29596	31019	}
29597	31020	if( cli_strcmp(zOp,"memo")==0 ){
29598		- oputf("%s\n", zSql);
	31021	+ sqlite3_fprintf(p->out, "%s\n", zSql);
29599	31022	}else
29600	31023	if( cli_strcmp(zOp,"run")==0 ){
29601	31024	char *zErrMsg = 0;
29602	31025	str.n = 0;
29603	31026	str.z[0] = 0;
29604	31027	rc = sqlite3_exec(p->db, zSql, captureOutputCallback, &str, &zErrMsg);
29605	31028	nTest++;
29606	31029	if( bVerbose ){
29607		- oputf("Result: %s\n", str.z);
	31030	+ sqlite3_fprintf(p->out, "Result: %s\n", str.z);
29608	31031	}
29609	31032	if( rc \|\| zErrMsg ){
29610	31033	nErr++;
29611	31034	rc = 1;
29612		- oputf("%d: error-code-%d: %s\n", tno, rc, zErrMsg);
	31035	+ sqlite3_fprintf(p->out, "%d: error-code-%d: %s\n", tno, rc,zErrMsg);
29613	31036	sqlite3_free(zErrMsg);
29614	31037	}else if( cli_strcmp(zAns,str.z)!=0 ){
29615	31038	nErr++;
29616	31039	rc = 1;
29617		- oputf("%d: Expected: [%s]\n", tno, zAns);
29618		- oputf("%d: Got: [%s]\n", tno, str.z);
	31040	+ sqlite3_fprintf(p->out, "%d: Expected: [%s]\n", tno, zAns);
	31041	+ sqlite3_fprintf(p->out, "%d: Got: [%s]\n", tno, str.z);
29619	31042	}
29620	31043	}
29621	31044	else{
29622		- eputf("Unknown operation \"%s\" on selftest line %d\n", zOp, tno);
	31045	+ sqlite3_fprintf(stderr,
	31046	+ "Unknown operation \"%s\" on selftest line %d\n", zOp, tno);
29623	31047	rc = 1;
29624	31048	break;
29625	31049	}
29626	31050	} /* End loop over rows of content from SELFTEST */
29627	31051	sqlite3_finalize(pStmt);
29628	31052	} /* End loop over k */
29629	31053	freeText(&str);
29630		- oputf("%d errors out of %d tests\n", nErr, nTest);
	31054	+ sqlite3_fprintf(p->out, "%d errors out of %d tests\n", nErr, nTest);
29631	31055	}else
29632	31056
29633	31057	if( c=='s' && cli_strncmp(azArg[0], "separator", n)==0 ){
29634	31058	if( nArg<2 \|\| nArg>3 ){
29635	31059	eputz("Usage: .separator COL ?ROW?\n");
		@@ -29673,11 +31097,12 @@
29673	31097	}else
29674	31098	if( cli_strcmp(z,"debug")==0 ){
29675	31099	bDebug = 1;
29676	31100	}else
29677	31101	{
29678		- eputf("Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
	31102	+ sqlite3_fprintf(stderr,
	31103	+ "Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
29679	31104	showHelp(p->out, azArg[0]);
29680	31105	rc = 1;
29681	31106	goto meta_command_exit;
29682	31107	}
29683	31108	}else if( zLike ){
		@@ -29751,11 +31176,11 @@
29751	31176	}
29752	31177	shell_check_oom(zSql);
29753	31178	freeText(&sQuery);
29754	31179	freeText(&sSql);
29755	31180	if( bDebug ){
29756		- oputf("%s\n", zSql);
	31181	+ sqlite3_fprintf(p->out, "%s\n", zSql);
29757	31182	}else{
29758	31183	shell_exec(p, zSql, 0);
29759	31184	}
29760	31185	#if !defined(SQLITE_OMIT_SCHEMA_PRAGMAS) && !defined(SQLITE_OMIT_VIRTUALTABLE)
29761	31186	{
		@@ -29781,11 +31206,11 @@
29781	31206	"\|\|group_concat('CAST(CAST('\|\|cname\|\|' AS BLOB) AS TEXT)<>'\|\|cname\n"
29782	31207	"\|\| ' AND typeof('\|\|cname\|\|')=''text'' ',\n"
29783	31208	"' OR ') as query, tname from tabcols group by tname)"
29784	31209	, zRevText);
29785	31210	shell_check_oom(zRevText);
29786		- if( bDebug ) oputf("%s\n", zRevText);
	31211	+ if( bDebug ) sqlite3_fprintf(p->out, "%s\n", zRevText);
29787	31212	lrc = sqlite3_prepare_v2(p->db, zRevText, -1, &pStmt, 0);
29788	31213	if( lrc!=SQLITE_OK ){
29789	31214	/* assert(lrc==SQLITE_NOMEM); // might also be SQLITE_ERROR if the
29790	31215	** user does cruel and unnatural things like ".limit expr_depth 0". */
29791	31216	rc = 1;
		@@ -29794,19 +31219,20 @@
29794	31219	lrc = SQLITE_ROW==sqlite3_step(pStmt);
29795	31220	if( lrc ){
29796	31221	const char zGenQuery = (char)sqlite3_column_text(pStmt,0);
29797	31222	sqlite3_stmt *pCheckStmt;
29798	31223	lrc = sqlite3_prepare_v2(p->db, zGenQuery, -1, &pCheckStmt, 0);
29799		- if( bDebug ) oputf("%s\n", zGenQuery);
	31224	+ if( bDebug ) sqlite3_fprintf(p->out, "%s\n", zGenQuery);
29800	31225	if( lrc!=SQLITE_OK ){
29801	31226	rc = 1;
29802	31227	}else{
29803	31228	if( SQLITE_ROW==sqlite3_step(pCheckStmt) ){
29804	31229	double countIrreversible = sqlite3_column_double(pCheckStmt, 0);
29805	31230	if( countIrreversible>0 ){
29806	31231	int sz = (int)(countIrreversible + 0.5);
29807		- eputf("Digest includes %d invalidly encoded text field%s.\n",
	31232	+ sqlite3_fprintf(stderr,
	31233	+ "Digest includes %d invalidly encoded text field%s.\n",
29808	31234	sz, (sz>1)? "s": "");
29809	31235	}
29810	31236	}
29811	31237	sqlite3_finalize(pCheckStmt);
29812	31238	}
		@@ -29836,15 +31262,15 @@
29836	31262	zCmd = sqlite3_mprintf(strchr(azArg[1],' ')==0?"%s":"\"%s\"", azArg[1]);
29837	31263	for(i=2; i<nArg && zCmd!=0; i++){
29838	31264	zCmd = sqlite3_mprintf(strchr(azArg[i],' ')==0?"%z %s":"%z \"%s\"",
29839	31265	zCmd, azArg[i]);
29840	31266	}
29841		- consoleRestore();
	31267	+ /consoleRestore();/
29842	31268	x = zCmd!=0 ? system(zCmd) : 1;
29843		- consoleRenewSetup();
	31269	+ /consoleRenewSetup();/
29844	31270	sqlite3_free(zCmd);
29845		- if( x ) eputf("System command returns %d\n", x);
	31271	+ if( x ) sqlite3_fprintf(stderr,"System command returns %d\n", x);
29846	31272	}else
29847	31273	#endif /* !defined(SQLITE_NOHAVE_SYSTEM) && !defined(SQLITE_SHELL_FIDDLE) */
29848	31274
29849	31275	if( c=='s' && cli_strncmp(azArg[0], "show", n)==0 ){
29850	31276	static const char *azBool[] = { "off", "on", "trigger", "full"};
		@@ -29853,50 +31279,52 @@
29853	31279	if( nArg!=1 ){
29854	31280	eputz("Usage: .show\n");
29855	31281	rc = 1;
29856	31282	goto meta_command_exit;
29857	31283	}
29858		- oputf("%12.12s: %s\n","echo",
	31284	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","echo",
29859	31285	azBool[ShellHasFlag(p, SHFLG_Echo)]);
29860		- oputf("%12.12s: %s\n","eqp", azBool[p->autoEQP&3]);
29861		- oputf("%12.12s: %s\n","explain",
	31286	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","eqp", azBool[p->autoEQP&3]);
	31287	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","explain",
29862	31288	p->mode==MODE_Explain ? "on" : p->autoExplain ? "auto" : "off");
29863		- oputf("%12.12s: %s\n","headers", azBool[p->showHeader!=0]);
	31289	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","headers",
	31290	+ azBool[p->showHeader!=0]);
29864	31291	if( p->mode==MODE_Column
29865	31292	\|\| (p->mode>=MODE_Markdown && p->mode<=MODE_Box)
29866	31293	){
29867		- oputf("%12.12s: %s --wrap %d --wordwrap %s --%squote\n", "mode",
	31294	+ sqlite3_fprintf(p->out,
	31295	+ "%12.12s: %s --wrap %d --wordwrap %s --%squote\n", "mode",
29868	31296	modeDescr[p->mode], p->cmOpts.iWrap,
29869	31297	p->cmOpts.bWordWrap ? "on" : "off",
29870	31298	p->cmOpts.bQuote ? "" : "no");
29871	31299	}else{
29872		- oputf("%12.12s: %s\n","mode", modeDescr[p->mode]);
	31300	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","mode", modeDescr[p->mode]);
29873	31301	}
29874		- oputf("%12.12s: ", "nullvalue");
29875		- output_c_string(p->nullValue);
29876		- oputz("\n");
29877		- oputf("%12.12s: %s\n","output",
	31302	+ sqlite3_fprintf(p->out, "%12.12s: ", "nullvalue");
	31303	+ output_c_string(p->out, p->nullValue);
	31304	+ sqlite3_fputs("\n", p->out);
	31305	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","output",
29878	31306	strlen30(p->outfile) ? p->outfile : "stdout");
29879		- oputf("%12.12s: ", "colseparator");
29880		- output_c_string(p->colSeparator);
29881		- oputz("\n");
29882		- oputf("%12.12s: ", "rowseparator");
29883		- output_c_string(p->rowSeparator);
29884		- oputz("\n");
	31307	+ sqlite3_fprintf(p->out, "%12.12s: ", "colseparator");
	31308	+ output_c_string(p->out, p->colSeparator);
	31309	+ sqlite3_fputs("\n", p->out);
	31310	+ sqlite3_fprintf(p->out, "%12.12s: ", "rowseparator");
	31311	+ output_c_string(p->out, p->rowSeparator);
	31312	+ sqlite3_fputs("\n", p->out);
29885	31313	switch( p->statsOn ){
29886	31314	case 0: zOut = "off"; break;
29887	31315	default: zOut = "on"; break;
29888	31316	case 2: zOut = "stmt"; break;
29889	31317	case 3: zOut = "vmstep"; break;
29890	31318	}
29891		- oputf("%12.12s: %s\n","stats", zOut);
29892		- oputf("%12.12s: ", "width");
	31319	+ sqlite3_fprintf(p->out, "%12.12s: %s\n","stats", zOut);
	31320	+ sqlite3_fprintf(p->out, "%12.12s: ", "width");
29893	31321	for (i=0;i<p->nWidth;i++) {
29894		- oputf("%d ", p->colWidth[i]);
	31322	+ sqlite3_fprintf(p->out, "%d ", p->colWidth[i]);
29895	31323	}
29896		- oputz("\n");
29897		- oputf("%12.12s: %s\n", "filename",
	31324	+ sqlite3_fputs("\n", p->out);
	31325	+ sqlite3_fprintf(p->out, "%12.12s: %s\n", "filename",
29898	31326	p->pAuxDb->zDbFilename ? p->pAuxDb->zDbFilename : "");
29899	31327	}else
29900	31328
29901	31329	if( c=='s' && cli_strncmp(azArg[0], "stats", n)==0 ){
29902	31330	if( nArg==2 ){
		@@ -30010,13 +31438,14 @@
30010	31438	if( nPrintCol<1 ) nPrintCol = 1;
30011	31439	nPrintRow = (nRow + nPrintCol - 1)/nPrintCol;
30012	31440	for(i=0; i<nPrintRow; i++){
30013	31441	for(j=i; j<nRow; j+=nPrintRow){
30014	31442	char *zSp = j<nPrintRow ? "" : " ";
30015		- oputf("%s%-*s", zSp, maxlen, azResult[j] ? azResult[j]:"");
	31443	+ sqlite3_fprintf(p->out,
	31444	+ "%s%-*s", zSp, maxlen, azResult[j] ? azResult[j]:"");
30016	31445	}
30017		- oputz("\n");
	31446	+ sqlite3_fputs("\n", p->out);
30018	31447	}
30019	31448	}
30020	31449
30021	31450	for(ii=0; ii<nRow; ii++) sqlite3_free(azResult[ii]);
30022	31451	sqlite3_free(azResult);
		@@ -30024,11 +31453,11 @@
30024	31453
30025	31454	#ifndef SQLITE_SHELL_FIDDLE
30026	31455	/* Begin redirecting output to the file "testcase-out.txt" */
30027	31456	if( c=='t' && cli_strcmp(azArg[0],"testcase")==0 ){
30028	31457	output_reset(p);
30029		- p->out = output_file_open("testcase-out.txt", 0);
	31458	+ p->out = output_file_open("testcase-out.txt");
30030	31459	if( p->out==0 ){
30031	31460	eputz("Error: cannot open 'testcase-out.txt'\n");
30032	31461	}
30033	31462	if( nArg>=2 ){
30034	31463	sqlite3_snprintf(sizeof(p->zTestcase), p->zTestcase, "%s", azArg[1]);
		@@ -30068,11 +31497,10 @@
30068	31497	{"prng_save", SQLITE_TESTCTRL_PRNG_SAVE, 0, "" },
30069	31498	{"prng_seed", SQLITE_TESTCTRL_PRNG_SEED, 0, "SEED ?db?" },
30070	31499	{"seek_count", SQLITE_TESTCTRL_SEEK_COUNT, 0, "" },
30071	31500	{"sorter_mmap", SQLITE_TESTCTRL_SORTER_MMAP, 0, "NMAX" },
30072	31501	{"tune", SQLITE_TESTCTRL_TUNE, 1, "ID VALUE" },
30073		- {"uselongdouble", SQLITE_TESTCTRL_USELONGDOUBLE,0,"?BOOLEAN\|\"default\"?"},
30074	31502	};
30075	31503	int testctrl = -1;
30076	31504	int iCtrl = -1;
30077	31505	int rc2 = 0; /* 0: usage. 1: %d 2: %x 3: no-output */
30078	31506	int isOk = 0;
		@@ -30088,14 +31516,14 @@
30088	31516	if( zCmd[0]=='-' && zCmd[1] ) zCmd++;
30089	31517	}
30090	31518
30091	31519	/* --help lists all test-controls */
30092	31520	if( cli_strcmp(zCmd,"help")==0 ){
30093		- oputz("Available test-controls:\n");
	31521	+ sqlite3_fputs("Available test-controls:\n", p->out);
30094	31522	for(i=0; i<ArraySize(aCtrl); i++){
30095	31523	if( aCtrl[i].unSafe && !ShellHasFlag(p,SHFLG_TestingMode) ) continue;
30096		- oputf(" .testctrl %s %s\n",
	31524	+ sqlite3_fprintf(p->out, " .testctrl %s %s\n",
30097	31525	aCtrl[i].zCtrlName, aCtrl[i].zUsage);
30098	31526	}
30099	31527	rc = 1;
30100	31528	goto meta_command_exit;
30101	31529	}
		@@ -30108,19 +31536,19 @@
30108	31536	if( cli_strncmp(zCmd, aCtrl[i].zCtrlName, n2)==0 ){
30109	31537	if( testctrl<0 ){
30110	31538	testctrl = aCtrl[i].ctrlCode;
30111	31539	iCtrl = i;
30112	31540	}else{
30113		- eputf("Error: ambiguous test-control: \"%s\"\n"
	31541	+ sqlite3_fprintf(stderr,"Error: ambiguous test-control: \"%s\"\n"
30114	31542	"Use \".testctrl --help\" for help\n", zCmd);
30115	31543	rc = 1;
30116	31544	goto meta_command_exit;
30117	31545	}
30118	31546	}
30119	31547	}
30120	31548	if( testctrl<0 ){
30121		- eputf("Error: unknown test-control: %s\n"
	31549	+ sqlite3_fprintf(stderr,"Error: unknown test-control: %s\n"
30122	31550	"Use \".testctrl --help\" for help\n", zCmd);
30123	31551	}else{
30124	31552	switch(testctrl){
30125	31553
30126	31554	/* Special processing for .testctrl opt MASK ...
		@@ -30170,11 +31598,13 @@
30170	31598	{ 0x10000000, 1, "OrderBySubq" },
30171	31599	{ 0xffffffff, 0, "All" },
30172	31600	};
30173	31601	unsigned int curOpt;
30174	31602	unsigned int newOpt;
	31603	+ unsigned int m;
30175	31604	int ii;
	31605	+ int nOff;
30176	31606	sqlite3_test_control(SQLITE_TESTCTRL_GETOPT, p->db, &curOpt);
30177	31607	newOpt = curOpt;
30178	31608	for(ii=2; ii<nArg; ii++){
30179	31609	const char *z = azArg[ii];
30180	31610	int useLabel = 0;
		@@ -30192,16 +31622,17 @@
30192	31622	int jj;
30193	31623	for(jj=0; jj<ArraySize(aLabel); jj++){
30194	31624	if( sqlite3_stricmp(zLabel, aLabel[jj].zLabel)==0 ) break;
30195	31625	}
30196	31626	if( jj>=ArraySize(aLabel) ){
30197		- eputf("Error: no such optimization: \"%s\"\n", zLabel);
30198		- eputz("Should be one of:");
	31627	+ sqlite3_fprintf(stderr,
	31628	+ "Error: no such optimization: \"%s\"\n", zLabel);
	31629	+ sqlite3_fputs("Should be one of:", stderr);
30199	31630	for(jj=0; jj<ArraySize(aLabel); jj++){
30200		- eputf(" %s", aLabel[jj].zLabel);
	31631	+ sqlite3_fprintf(stderr," %s", aLabel[jj].zLabel);
30201	31632	}
30202		- eputz("\n");
	31633	+ sqlite3_fputs("\n", stderr);
30203	31634	rc = 1;
30204	31635	goto meta_command_exit;
30205	31636	}
30206	31637	if( useLabel=='+' ){
30207	31638	newOpt &= ~aLabel[jj].mask;
		@@ -30210,28 +31641,32 @@
30210	31641	}
30211	31642	}
30212	31643	}
30213	31644	if( curOpt!=newOpt ){
30214	31645	sqlite3_test_control(SQLITE_TESTCTRL_OPTIMIZATIONS,p->db,newOpt);
30215		- }else if( nArg<3 ){
30216		- curOpt = ~newOpt;
	31646	+ }
	31647	+ for(ii=nOff=0, m=1; ii<32; ii++, m <<= 1){
	31648	+ if( m & newOpt ) nOff++;
30217	31649	}
30218		- if( newOpt==0 ){
30219		- oputz("+All\n");
30220		- }else if( newOpt==0xffffffff ){
30221		- oputz("-All\n");
	31650	+ if( nOff<12 ){
	31651	+ sqlite3_fputs("+All", p->out);
	31652	+ for(ii=0; ii<ArraySize(aLabel); ii++){
	31653	+ if( !aLabel[ii].bDsply ) continue;
	31654	+ if( (newOpt & aLabel[ii].mask)!=0 ){
	31655	+ sqlite3_fprintf(p->out, " -%s", aLabel[ii].zLabel);
	31656	+ }
	31657	+ }
30222	31658	}else{
30223		- int jj;
30224		- for(jj=0; jj<ArraySize(aLabel); jj++){
30225		- unsigned int m = aLabel[jj].mask;
30226		- if( !aLabel[jj].bDsply ) continue;
30227		- if( (curOpt&m)!=(newOpt&m) ){
30228		- oputf("%c%s\n", (newOpt & m)==0 ? '+' : '-',
30229		- aLabel[jj].zLabel);
	31659	+ sqlite3_fputs("-All", p->out);
	31660	+ for(ii=0; ii<ArraySize(aLabel); ii++){
	31661	+ if( !aLabel[ii].bDsply ) continue;
	31662	+ if( (newOpt & aLabel[ii].mask)==0 ){
	31663	+ sqlite3_fprintf(p->out, " +%s", aLabel[ii].zLabel);
30230	31664	}
30231	31665	}
30232	31666	}
	31667	+ sqlite3_fputs("\n", p->out);
30233	31668	rc2 = isOk = 3;
30234	31669	break;
30235	31670	}
30236	31671
30237	31672	/* sqlite3_test_control(int, db, int) */
		@@ -30267,11 +31702,11 @@
30267	31702	if( nArg==3 \|\| nArg==4 ){
30268	31703	int ii = (int)integerValue(azArg[2]);
30269	31704	sqlite3 *db;
30270	31705	if( ii==0 && cli_strcmp(azArg[2],"random")==0 ){
30271	31706	sqlite3_randomness(sizeof(ii),&ii);
30272		- sputf(stdout, "-- random seed: %d\n", ii);
	31707	+ sqlite3_fprintf(stdout, "-- random seed: %d\n", ii);
30273	31708	}
30274	31709	if( nArg==3 ){
30275	31710	db = 0;
30276	31711	}else{
30277	31712	db = p->db;
		@@ -30301,25 +31736,10 @@
30301	31736	rc2 = sqlite3_test_control(testctrl, opt);
30302	31737	isOk = 3;
30303	31738	}
30304	31739	break;
30305	31740
30306		- /* sqlite3_test_control(int, int) */
30307		- case SQLITE_TESTCTRL_USELONGDOUBLE: {
30308		- int opt = -1;
30309		- if( nArg==3 ){
30310		- if( cli_strcmp(azArg[2],"default")==0 ){
30311		- opt = 2;
30312		- }else{
30313		- opt = booleanValue(azArg[2]);
30314		- }
30315		- }
30316		- rc2 = sqlite3_test_control(testctrl, opt);
30317		- isOk = 1;
30318		- break;
30319		- }
30320		-
30321	31741	/* sqlite3_test_control(sqlite3) /
30322	31742	case SQLITE_TESTCTRL_INTERNAL_FUNCTIONS:
30323	31743	rc2 = sqlite3_test_control(testctrl, p->db);
30324	31744	isOk = 3;
30325	31745	break;
		@@ -30335,11 +31755,11 @@
30335	31755	break;
30336	31756
30337	31757	case SQLITE_TESTCTRL_SEEK_COUNT: {
30338	31758	u64 x = 0;
30339	31759	rc2 = sqlite3_test_control(testctrl, p->db, &x);
30340		- oputf("%llu\n", x);
	31760	+ sqlite3_fprintf(p->out, "%llu\n", x);
30341	31761	isOk = 3;
30342	31762	break;
30343	31763	}
30344	31764
30345	31765	#ifdef YYCOVERAGE
		@@ -30366,15 +31786,15 @@
30366	31786	int id = 1;
30367	31787	while(1){
30368	31788	int val = 0;
30369	31789	rc2 = sqlite3_test_control(testctrl, -id, &val);
30370	31790	if( rc2!=SQLITE_OK ) break;
30371		- if( id>1 ) oputz(" ");
30372		- oputf("%d: %d", id, val);
	31791	+ if( id>1 ) sqlite3_fputs(" ", p->out);
	31792	+ sqlite3_fprintf(p->out, "%d: %d", id, val);
30373	31793	id++;
30374	31794	}
30375		- if( id>1 ) oputz("\n");
	31795	+ if( id>1 ) sqlite3_fputs("\n", p->out);
30376	31796	isOk = 3;
30377	31797	}
30378	31798	break;
30379	31799	}
30380	31800	#endif
		@@ -30412,18 +31832,26 @@
30412	31832	}else if( cli_strcmp(z,"reset")==0 ){
30413	31833	faultsim_state.iCnt = faultsim_state.nSkip;
30414	31834	faultsim_state.nHit = 0;
30415	31835	sqlite3_test_control(testctrl, faultsim_callback);
30416	31836	}else if( cli_strcmp(z,"status")==0 ){
30417		- oputf("faultsim.iId: %d\n", faultsim_state.iId);
30418		- oputf("faultsim.iErr: %d\n", faultsim_state.iErr);
30419		- oputf("faultsim.iCnt: %d\n", faultsim_state.iCnt);
30420		- oputf("faultsim.nHit: %d\n", faultsim_state.nHit);
30421		- oputf("faultsim.iInterval: %d\n", faultsim_state.iInterval);
30422		- oputf("faultsim.eVerbose: %d\n", faultsim_state.eVerbose);
30423		- oputf("faultsim.nRepeat: %d\n", faultsim_state.nRepeat);
30424		- oputf("faultsim.nSkip: %d\n", faultsim_state.nSkip);
	31837	+ sqlite3_fprintf(p->out, "faultsim.iId: %d\n",
	31838	+ faultsim_state.iId);
	31839	+ sqlite3_fprintf(p->out, "faultsim.iErr: %d\n",
	31840	+ faultsim_state.iErr);
	31841	+ sqlite3_fprintf(p->out, "faultsim.iCnt: %d\n",
	31842	+ faultsim_state.iCnt);
	31843	+ sqlite3_fprintf(p->out, "faultsim.nHit: %d\n",
	31844	+ faultsim_state.nHit);
	31845	+ sqlite3_fprintf(p->out, "faultsim.iInterval: %d\n",
	31846	+ faultsim_state.iInterval);
	31847	+ sqlite3_fprintf(p->out, "faultsim.eVerbose: %d\n",
	31848	+ faultsim_state.eVerbose);
	31849	+ sqlite3_fprintf(p->out, "faultsim.nRepeat: %d\n",
	31850	+ faultsim_state.nRepeat);
	31851	+ sqlite3_fprintf(p->out, "faultsim.nSkip: %d\n",
	31852	+ faultsim_state.nSkip);
30425	31853	}else if( cli_strcmp(z,"-v")==0 ){
30426	31854	if( faultsim_state.eVerbose<2 ) faultsim_state.eVerbose++;
30427	31855	}else if( cli_strcmp(z,"-q")==0 ){
30428	31856	if( faultsim_state.eVerbose>0 ) faultsim_state.eVerbose--;
30429	31857	}else if( cli_strcmp(z,"-id")==0 && kk+1<nArg ){
		@@ -30437,19 +31865,20 @@
30437	31865	}else if( cli_strcmp(z,"-skip")==0 && kk+1<nArg ){
30438	31866	faultsim_state.nSkip = atoi(azArg[++kk]);
30439	31867	}else if( cli_strcmp(z,"-?")==0 \|\| sqlite3_strglob("help",z)==0){
30440	31868	bShowHelp = 1;
30441	31869	}else{
30442		- eputf("Unrecognized fault_install argument: \"%s\"\n",
	31870	+ sqlite3_fprintf(stderr,
	31871	+ "Unrecognized fault_install argument: \"%s\"\n",
30443	31872	azArg[kk]);
30444	31873	rc = 1;
30445	31874	bShowHelp = 1;
30446	31875	break;
30447	31876	}
30448	31877	}
30449	31878	if( bShowHelp ){
30450		- oputz(
	31879	+ sqlite3_fputs(
30451	31880	"Usage: .testctrl fault_install ARGS\n"
30452	31881	"Possible arguments:\n"
30453	31882	" off Disable faultsim\n"
30454	31883	" on Activate faultsim\n"
30455	31884	" reset Reset the trigger counter\n"
		@@ -30459,23 +31888,25 @@
30459	31888	" --errcode N When triggered, return N as error code\n"
30460	31889	" --id ID Trigger only for the ID specified\n"
30461	31890	" --interval N Trigger only after every N-th call\n"
30462	31891	" --repeat N Turn off after N hits. 0 means never\n"
30463	31892	" --skip N Skip the first N encounters\n"
	31893	+ ,p->out
30464	31894	);
30465	31895	}
30466	31896	break;
30467	31897	}
30468	31898	}
30469	31899	}
30470	31900	if( isOk==0 && iCtrl>=0 ){
30471		- oputf("Usage: .testctrl %s %s\n", zCmd,aCtrl[iCtrl].zUsage);
	31901	+ sqlite3_fprintf(p->out,
	31902	+ "Usage: .testctrl %s %s\n", zCmd,aCtrl[iCtrl].zUsage);
30472	31903	rc = 1;
30473	31904	}else if( isOk==1 ){
30474		- oputf("%d\n", rc2);
	31905	+ sqlite3_fprintf(p->out, "%d\n", rc2);
30475	31906	}else if( isOk==2 ){
30476		- oputf("0x%08x\n", rc2);
	31907	+ sqlite3_fprintf(p->out, "0x%08x\n", rc2);
30477	31908	}
30478	31909	}else
30479	31910	#endif /* !defined(SQLITE_UNTESTABLE) */
30480	31911
30481	31912	if( c=='t' && n>4 && cli_strncmp(azArg[0], "timeout", n)==0 ){
		@@ -30526,17 +31957,17 @@
30526	31957	}
30527	31958	else if( optionMatch(z, "close") ){
30528	31959	mType \|= SQLITE_TRACE_CLOSE;
30529	31960	}
30530	31961	else {
30531		- eputf("Unknown option \"%s\" on \".trace\"\n", z);
	31962	+ sqlite3_fprintf(stderr,"Unknown option \"%s\" on \".trace\"\n", z);
30532	31963	rc = 1;
30533	31964	goto meta_command_exit;
30534	31965	}
30535	31966	}else{
30536	31967	output_file_close(p->traceOut);
30537		- p->traceOut = output_file_open(z, 0);
	31968	+ p->traceOut = output_file_open(z);
30538	31969	}
30539	31970	}
30540	31971	if( p->traceOut==0 ){
30541	31972	sqlite3_trace_v2(p->db, 0, 0, 0);
30542	31973	}else{
		@@ -30569,103 +32000,40 @@
30569	32000	}
30570	32001	}
30571	32002	}else
30572	32003	#endif
30573	32004
30574		-#if SQLITE_USER_AUTHENTICATION
30575		- if( c=='u' && cli_strncmp(azArg[0], "user", n)==0 ){
30576		- if( nArg<2 ){
30577		- eputz("Usage: .user SUBCOMMAND ...\n");
30578		- rc = 1;
30579		- goto meta_command_exit;
30580		- }
30581		- open_db(p, 0);
30582		- if( cli_strcmp(azArg[1],"login")==0 ){
30583		- if( nArg!=4 ){
30584		- eputz("Usage: .user login USER PASSWORD\n");
30585		- rc = 1;
30586		- goto meta_command_exit;
30587		- }
30588		- rc = sqlite3_user_authenticate(p->db, azArg[2], azArg[3],
30589		- strlen30(azArg[3]));
30590		- if( rc ){
30591		- eputf("Authentication failed for user %s\n", azArg[2]);
30592		- rc = 1;
30593		- }
30594		- }else if( cli_strcmp(azArg[1],"add")==0 ){
30595		- if( nArg!=5 ){
30596		- eputz("Usage: .user add USER PASSWORD ISADMIN\n");
30597		- rc = 1;
30598		- goto meta_command_exit;
30599		- }
30600		- rc = sqlite3_user_add(p->db, azArg[2], azArg[3], strlen30(azArg[3]),
30601		- booleanValue(azArg[4]));
30602		- if( rc ){
30603		- eputf("User-Add failed: %d\n", rc);
30604		- rc = 1;
30605		- }
30606		- }else if( cli_strcmp(azArg[1],"edit")==0 ){
30607		- if( nArg!=5 ){
30608		- eputz("Usage: .user edit USER PASSWORD ISADMIN\n");
30609		- rc = 1;
30610		- goto meta_command_exit;
30611		- }
30612		- rc = sqlite3_user_change(p->db, azArg[2], azArg[3], strlen30(azArg[3]),
30613		- booleanValue(azArg[4]));
30614		- if( rc ){
30615		- eputf("User-Edit failed: %d\n", rc);
30616		- rc = 1;
30617		- }
30618		- }else if( cli_strcmp(azArg[1],"delete")==0 ){
30619		- if( nArg!=3 ){
30620		- eputz("Usage: .user delete USER\n");
30621		- rc = 1;
30622		- goto meta_command_exit;
30623		- }
30624		- rc = sqlite3_user_delete(p->db, azArg[2]);
30625		- if( rc ){
30626		- eputf("User-Delete failed: %d\n", rc);
30627		- rc = 1;
30628		- }
30629		- }else{
30630		- eputz("Usage: .user login\|add\|edit\|delete ...\n");
30631		- rc = 1;
30632		- goto meta_command_exit;
30633		- }
30634		- }else
30635		-#endif /* SQLITE_USER_AUTHENTICATION */
30636		-
30637	32005	if( c=='v' && cli_strncmp(azArg[0], "version", n)==0 ){
30638	32006	char zPtrSz = sizeof(void)==8 ? "64-bit" : "32-bit";
30639		- oputf("SQLite %s %s\n" /extra-version-info/,
	32007	+ sqlite3_fprintf(p->out, "SQLite %s %s\n" /extra-version-info/,
30640	32008	sqlite3_libversion(), sqlite3_sourceid());
30641	32009	#if SQLITE_HAVE_ZLIB
30642		- oputf("zlib version %s\n", zlibVersion());
	32010	+ sqlite3_fprintf(p->out, "zlib version %s\n", zlibVersion());
30643	32011	#endif
30644	32012	#define CTIMEOPT_VAL_(opt) #opt
30645	32013	#define CTIMEOPT_VAL(opt) CTIMEOPT_VAL_(opt)
30646	32014	#if defined(__clang__) && defined(__clang_major__)
30647		- oputf("clang-" CTIMEOPT_VAL(__clang_major__) "."
	32015	+ sqlite3_fprintf(p->out, "clang-" CTIMEOPT_VAL(__clang_major__) "."
30648	32016	CTIMEOPT_VAL(__clang_minor__) "."
30649	32017	CTIMEOPT_VAL(__clang_patchlevel__) " (%s)\n", zPtrSz);
30650	32018	#elif defined(_MSC_VER)
30651		- oputf("msvc-" CTIMEOPT_VAL(_MSC_VER) " (%s)\n", zPtrSz);
	32019	+ sqlite3_fprintf(p->out, "msvc-" CTIMEOPT_VAL(_MSC_VER) " (%s)\n", zPtrSz);
30652	32020	#elif defined(__GNUC__) && defined(__VERSION__)
30653		- oputf("gcc-" __VERSION__ " (%s)\n", zPtrSz);
	32021	+ sqlite3_fprintf(p->out, "gcc-" __VERSION__ " (%s)\n", zPtrSz);
30654	32022	#endif
30655	32023	}else
30656	32024
30657	32025	if( c=='v' && cli_strncmp(azArg[0], "vfsinfo", n)==0 ){
30658	32026	const char *zDbName = nArg==2 ? azArg[1] : "main";
30659	32027	sqlite3_vfs *pVfs = 0;
30660	32028	if( p->db ){
30661	32029	sqlite3_file_control(p->db, zDbName, SQLITE_FCNTL_VFS_POINTER, &pVfs);
30662	32030	if( pVfs ){
30663		- oputf("vfs.zName = \"%s\"\n", pVfs->zName);
30664		- oputf("vfs.iVersion = %d\n", pVfs->iVersion);
30665		- oputf("vfs.szOsFile = %d\n", pVfs->szOsFile);
30666		- oputf("vfs.mxPathname = %d\n", pVfs->mxPathname);
	32031	+ sqlite3_fprintf(p->out, "vfs.zName = \"%s\"\n", pVfs->zName);
	32032	+ sqlite3_fprintf(p->out, "vfs.iVersion = %d\n", pVfs->iVersion);
	32033	+ sqlite3_fprintf(p->out, "vfs.szOsFile = %d\n", pVfs->szOsFile);
	32034	+ sqlite3_fprintf(p->out, "vfs.mxPathname = %d\n", pVfs->mxPathname);
30667	32035	}
30668	32036	}
30669	32037	}else
30670	32038
30671	32039	if( c=='v' && cli_strncmp(azArg[0], "vfslist", n)==0 ){
		@@ -30673,17 +32041,17 @@
30673	32041	sqlite3_vfs *pCurrent = 0;
30674	32042	if( p->db ){
30675	32043	sqlite3_file_control(p->db, "main", SQLITE_FCNTL_VFS_POINTER, &pCurrent);
30676	32044	}
30677	32045	for(pVfs=sqlite3_vfs_find(0); pVfs; pVfs=pVfs->pNext){
30678		- oputf("vfs.zName = \"%s\"%s\n", pVfs->zName,
	32046	+ sqlite3_fprintf(p->out, "vfs.zName = \"%s\"%s\n", pVfs->zName,
30679	32047	pVfs==pCurrent ? " <--- CURRENT" : "");
30680		- oputf("vfs.iVersion = %d\n", pVfs->iVersion);
30681		- oputf("vfs.szOsFile = %d\n", pVfs->szOsFile);
30682		- oputf("vfs.mxPathname = %d\n", pVfs->mxPathname);
	32048	+ sqlite3_fprintf(p->out, "vfs.iVersion = %d\n", pVfs->iVersion);
	32049	+ sqlite3_fprintf(p->out, "vfs.szOsFile = %d\n", pVfs->szOsFile);
	32050	+ sqlite3_fprintf(p->out, "vfs.mxPathname = %d\n", pVfs->mxPathname);
30683	32051	if( pVfs->pNext ){
30684		- oputz("-----------------------------------\n");
	32052	+ sqlite3_fputs("-----------------------------------\n", p->out);
30685	32053	}
30686	32054	}
30687	32055	}else
30688	32056
30689	32057	if( c=='v' && cli_strncmp(azArg[0], "vfsname", n)==0 ){
		@@ -30690,11 +32058,11 @@
30690	32058	const char *zDbName = nArg==2 ? azArg[1] : "main";
30691	32059	char *zVfsName = 0;
30692	32060	if( p->db ){
30693	32061	sqlite3_file_control(p->db, zDbName, SQLITE_FCNTL_VFSNAME, &zVfsName);
30694	32062	if( zVfsName ){
30695		- oputf("%s\n", zVfsName);
	32063	+ sqlite3_fprintf(p->out, "%s\n", zVfsName);
30696	32064	sqlite3_free(zVfsName);
30697	32065	}
30698	32066	}
30699	32067	}else
30700	32068
		@@ -30714,11 +32082,11 @@
30714	32082	p->colWidth[j-1] = (int)integerValue(azArg[j]);
30715	32083	}
30716	32084	}else
30717	32085
30718	32086	{
30719		- eputf("Error: unknown command or invalid arguments: "
	32087	+ sqlite3_fprintf(stderr,"Error: unknown command or invalid arguments: "
30720	32088	" \"%s\". Enter \".help\" for help\n", azArg[0]);
30721	32089	rc = 1;
30722	32090	}
30723	32091
30724	32092	meta_command_exit:
		@@ -30755,11 +32123,10 @@
30755	32123	SCAN_TRACKER_REFTYPE pst){
30756	32124	char cin;
30757	32125	char cWait = (char)qss; /* intentional narrowing loss */
30758	32126	if( cWait==0 ){
30759	32127	PlainScan:
30760		- assert( cWait==0 );
30761	32128	while( (cin = *zLine++)!=0 ){
30762	32129	if( IsSpace(cin) )
30763	32130	continue;
30764	32131	switch (cin){
30765	32132	case '-':
		@@ -30807,11 +32174,10 @@
30807	32174	switch( cWait ){
30808	32175	case '*':
30809	32176	if( *zLine != '/' )
30810	32177	continue;
30811	32178	++zLine;
30812		- cWait = 0;
30813	32179	CONTINUE_PROMPT_AWAITC(pst, 0);
30814	32180	qss = QSS_SETV(qss, 0);
30815	32181	goto PlainScan;
30816	32182	case '`': case '\'': case '"':
30817	32183	if(*zLine==cWait){
		@@ -30819,11 +32185,10 @@
30819	32185	++zLine;
30820	32186	continue;
30821	32187	}
30822	32188	deliberate_fall_through;
30823	32189	case ']':
30824		- cWait = 0;
30825	32190	CONTINUE_PROMPT_AWAITC(pst, 0);
30826	32191	qss = QSS_SETV(qss, 0);
30827	32192	goto PlainScan;
30828	32193	default: assert(0);
30829	32194	}
		@@ -30966,11 +32331,11 @@
30966	32331	open_db(p, 0);
30967	32332	if( ShellHasFlag(p,SHFLG_Backslash) ) resolve_backslashes(zSql);
30968	32333	if( p->flgProgress & SHELL_PROGRESS_RESET ) p->nProgress = 0;
30969	32334	BEGIN_TIMER;
30970	32335	rc = shell_exec(p, zSql, &zErrMsg);
30971		- END_TIMER;
	32336	+ END_TIMER(p->out);
30972	32337	if( rc \|\| zErrMsg ){
30973	32338	char zPrefix[100];
30974	32339	const char *zErrorTail;
30975	32340	const char *zErrorType;
30976	32341	if( zErrMsg==0 ){
		@@ -30990,28 +32355,31 @@
30990	32355	sqlite3_snprintf(sizeof(zPrefix), zPrefix,
30991	32356	"%s near line %d:", zErrorType, startline);
30992	32357	}else{
30993	32358	sqlite3_snprintf(sizeof(zPrefix), zPrefix, "%s:", zErrorType);
30994	32359	}
30995		- eputf("%s %s\n", zPrefix, zErrorTail);
	32360	+ sqlite3_fprintf(stderr,"%s %s\n", zPrefix, zErrorTail);
30996	32361	sqlite3_free(zErrMsg);
30997	32362	zErrMsg = 0;
30998	32363	return 1;
30999	32364	}else if( ShellHasFlag(p, SHFLG_CountChanges) ){
31000	32365	char zLineBuf[2000];
31001	32366	sqlite3_snprintf(sizeof(zLineBuf), zLineBuf,
31002	32367	"changes: %lld total_changes: %lld",
31003	32368	sqlite3_changes64(p->db), sqlite3_total_changes64(p->db));
31004		- oputf("%s\n", zLineBuf);
	32369	+ sqlite3_fprintf(p->out, "%s\n", zLineBuf);
31005	32370	}
31006	32371
31007	32372	if( doAutoDetectRestore(p, zSql) ) return 1;
31008	32373	return 0;
31009	32374	}
31010	32375
31011	32376	static void echo_group_input(ShellState p, const char zDo){
31012		- if( ShellHasFlag(p, SHFLG_Echo) ) oputf("%s\n", zDo);
	32377	+ if( ShellHasFlag(p, SHFLG_Echo) ){
	32378	+ sqlite3_fprintf(p->out, "%s\n", zDo);
	32379	+ fflush(p->out);
	32380	+ }
31013	32381	}
31014	32382
31015	32383	#ifdef SQLITE_SHELL_FIDDLE
31016	32384	/*
31017	32385	** Alternate one_input_line() impl for wasm mode. This is not in the primary
		@@ -31065,11 +32433,11 @@
31065	32433	i64 startline = 0; /* Line number for start of current input */
31066	32434	QuickScanState qss = QSS_Start; /* Accumulated line status (so far) */
31067	32435
31068	32436	if( p->inputNesting==MAX_INPUT_NESTING ){
31069	32437	/* This will be more informative in a later version. */
31070		- eputf("Input nesting limit (%d) reached at line %d."
	32438	+ sqlite3_fprintf(stderr,"Input nesting limit (%d) reached at line %d."
31071	32439	" Check recursion.\n", MAX_INPUT_NESTING, p->lineno);
31072	32440	return 1;
31073	32441	}
31074	32442	++p->inputNesting;
31075	32443	p->lineno = 0;
		@@ -31077,11 +32445,11 @@
31077	32445	while( errCnt==0 \|\| !bail_on_error \|\| (p->in==0 && stdin_is_interactive) ){
31078	32446	fflush(p->out);
31079	32447	zLine = one_input_line(p->in, zLine, nSql>0);
31080	32448	if( zLine==0 ){
31081	32449	/* End of input */
31082		- if( p->in==0 && stdin_is_interactive ) oputz("\n");
	32450	+ if( p->in==0 && stdin_is_interactive ) sqlite3_fputs("\n", p->out);
31083	32451	break;
31084	32452	}
31085	32453	if( seenInterrupt ){
31086	32454	if( p->in!=0 ) break;
31087	32455	seenInterrupt = 0;
		@@ -31297,19 +32665,19 @@
31297	32665	}
31298	32666	zBuf = sqlite3_mprintf("%s/.sqliterc",home_dir);
31299	32667	shell_check_oom(zBuf);
31300	32668	sqliterc = zBuf;
31301	32669	}
31302		- p->in = fopen(sqliterc,"rb");
	32670	+ p->in = sqlite3_fopen(sqliterc,"rb");
31303	32671	if( p->in ){
31304	32672	if( stdin_is_interactive ){
31305		- eputf("-- Loading resources from %s\n", sqliterc);
	32673	+ sqlite3_fprintf(stderr,"-- Loading resources from %s\n", sqliterc);
31306	32674	}
31307	32675	if( process_input(p) && bail_on_error ) exit(1);
31308	32676	fclose(p->in);
31309	32677	}else if( sqliterc_override!=0 ){
31310		- eputf("cannot open: \"%s\"\n", sqliterc);
	32678	+ sqlite3_fprintf(stderr,"cannot open: \"%s\"\n", sqliterc);
31311	32679	if( bail_on_error ) exit(1);
31312	32680	}
31313	32681	p->in = inSaved;
31314	32682	p->lineno = savedLineno;
31315	32683	sqlite3_free(zBuf);
		@@ -31374,23 +32742,21 @@
31374	32742	" -table set output mode to 'table'\n"
31375	32743	" -tabs set output mode to 'tabs'\n"
31376	32744	" -unsafe-testing allow unsafe commands and modes for testing\n"
31377	32745	" -version show SQLite version\n"
31378	32746	" -vfs NAME use NAME as the default VFS\n"
31379		-#ifdef SQLITE_ENABLE_VFSTRACE
31380	32747	" -vfstrace enable tracing of all VFS calls\n"
31381		-#endif
31382	32748	#ifdef SQLITE_HAVE_ZLIB
31383	32749	" -zip open the file as a ZIP Archive\n"
31384	32750	#endif
31385	32751	;
31386	32752	static void usage(int showDetail){
31387		- eputf("Usage: %s [OPTIONS] [FILENAME [SQL]]\n"
	32753	+ sqlite3_fprintf(stderr,"Usage: %s [OPTIONS] [FILENAME [SQL]]\n"
31388	32754	"FILENAME is the name of an SQLite database. A new database is created\n"
31389	32755	"if the file does not previously exist. Defaults to :memory:.\n", Argv0);
31390	32756	if( showDetail ){
31391		- eputf("OPTIONS include:\n%s", zOptions);
	32757	+ sqlite3_fprintf(stderr,"OPTIONS include:\n%s", zOptions);
31392	32758	}else{
31393	32759	eputz("Use the -help option for additional information\n");
31394	32760	}
31395	32761	exit(0);
31396	32762	}
		@@ -31411,10 +32777,13 @@
31411	32777	*/
31412	32778	static void main_init(ShellState *data) {
31413	32779	memset(data, 0, sizeof(*data));
31414	32780	data->normalMode = data->cMode = data->mode = MODE_List;
31415	32781	data->autoExplain = 1;
	32782	+#ifdef _WIN32
	32783	+ data->crlfMode = 1;
	32784	+#endif
31416	32785	data->pAuxDb = &data->aAuxDb[0];
31417	32786	memcpy(data->colSeparator,SEP_Column, 2);
31418	32787	memcpy(data->rowSeparator,SEP_Row, 2);
31419	32788	data->showHeader = 0;
31420	32789	data->shellFlgs = SHFLG_Lookaside;
		@@ -31446,29 +32815,39 @@
31446	32815	SetConsoleTextAttribute(out, defaultScreenInfo.wAttributes);
31447	32816	#endif
31448	32817	}
31449	32818	#else
31450	32819	static void printBold(const char *zText){
31451		- sputf(stdout, "\033[1m%s\033[0m", zText);
	32820	+ sqlite3_fprintf(stdout, "\033[1m%s\033[0m", zText);
31452	32821	}
31453	32822	#endif
31454	32823
31455	32824	/*
31456	32825	** Get the argument to an --option. Throw an error and die if no argument
31457	32826	** is available.
31458	32827	*/
31459	32828	static char cmdline_option_value(int argc, char *argv, int i){
31460	32829	if( i==argc ){
31461		- eputf("%s: Error: missing argument to %s\n", argv[0], argv[argc-1]);
	32830	+ sqlite3_fprintf(stderr,
	32831	+ "%s: Error: missing argument to %s\n", argv[0], argv[argc-1]);
31462	32832	exit(1);
31463	32833	}
31464	32834	return argv[i];
31465	32835	}
31466	32836
31467	32837	static void sayAbnormalExit(void){
31468	32838	if( seenInterrupt ) eputz("Program interrupted.\n");
31469	32839	}
	32840	+
	32841	+/* Routine to output from vfstrace
	32842	+*/
	32843	+static int vfstraceOut(const char z, void pArg){
	32844	+ ShellState p = (ShellState)pArg;
	32845	+ sqlite3_fputs(z, p->out);
	32846	+ fflush(p->out);
	32847	+ return 1;
	32848	+}
31470	32849
31471	32850	#ifndef SQLITE_SHELL_IS_UTF8
31472	32851	# if (defined(_WIN32) \|\| defined(WIN32)) \
31473	32852	&& (defined(_MSC_VER) \|\| (defined(UNICODE) && defined(__GNUC__)))
31474	32853	# define SQLITE_SHELL_IS_UTF8 (0)
		@@ -31493,19 +32872,19 @@
31493	32872	char *zErrMsg = 0;
31494	32873	#ifdef SQLITE_SHELL_FIDDLE
31495	32874	# define data shellState
31496	32875	#else
31497	32876	ShellState data;
31498		- StreamsAreConsole consStreams = SAC_NoConsole;
31499	32877	#endif
31500	32878	const char *zInitFile = 0;
31501	32879	int i;
31502	32880	int rc = 0;
31503	32881	int warnInmemoryDb = 0;
31504	32882	int readStdin = 1;
31505	32883	int nCmd = 0;
31506	32884	int nOptsEnd = argc;
	32885	+ int bEnableVfstrace = 0;
31507	32886	char **azCmd = 0;
31508	32887	const char zVfs = 0; / Value of -vfs command-line option */
31509	32888	#if !SQLITE_SHELL_IS_UTF8
31510	32889	char **argvToFree = 0;
31511	32890	int argcToFree = 0;
		@@ -31515,25 +32894,29 @@
31515	32894	#ifdef SQLITE_SHELL_FIDDLE
31516	32895	stdin_is_interactive = 0;
31517	32896	stdout_is_console = 1;
31518	32897	data.wasm.zDefaultDbName = "/fiddle.sqlite3";
31519	32898	#else
31520		- consStreams = consoleClassifySetup(stdin, stdout, stderr);
31521		- stdin_is_interactive = (consStreams & SAC_InConsole)!=0;
31522		- stdout_is_console = (consStreams & SAC_OutConsole)!=0;
31523		- atexit(consoleRestore);
	32899	+ stdin_is_interactive = isatty(0);
	32900	+ stdout_is_console = isatty(1);
31524	32901	#endif
31525	32902	atexit(sayAbnormalExit);
31526	32903	#ifdef SQLITE_DEBUG
31527	32904	mem_main_enter = sqlite3_memory_used();
31528	32905	#endif
31529	32906	#if !defined(_WIN32_WCE)
31530	32907	if( getenv("SQLITE_DEBUG_BREAK") ){
31531	32908	if( isatty(0) && isatty(2) ){
31532		- eputf("attach debugger to process %d and press any key to continue.\n",
	32909	+ char zLine[100];
	32910	+ sqlite3_fprintf(stderr,
	32911	+ "attach debugger to process %d and press ENTER to continue...",
31533	32912	GETPID());
31534		- fgetc(stdin);
	32913	+ if( sqlite3_fgets(zLine, sizeof(zLine), stdin)!=0
	32914	+ && cli_strcmp(zLine,"stop")==0
	32915	+ ){
	32916	+ exit(1);
	32917	+ }
31535	32918	}else{
31536	32919	#if defined(_WIN32) \|\| defined(WIN32)
31537	32920	#if SQLITE_OS_WINRT
31538	32921	__debugbreak();
31539	32922	#else
		@@ -31554,11 +32937,12 @@
31554	32937	}
31555	32938	#endif
31556	32939
31557	32940	#if USE_SYSTEM_SQLITE+0!=1
31558	32941	if( cli_strncmp(sqlite3_sourceid(),SQLITE_SOURCE_ID,60)!=0 ){
31559		- eputf("SQLite header and source version mismatch\n%s\n%s\n",
	32942	+ sqlite3_fprintf(stderr,
	32943	+ "SQLite header and source version mismatch\n%s\n%s\n",
31560	32944	sqlite3_sourceid(), SQLITE_SOURCE_ID);
31561	32945	exit(1);
31562	32946	}
31563	32947	#endif
31564	32948	main_init(&data);
		@@ -31696,21 +33080,12 @@
31696	33080	switch( n ){
31697	33081	case 0: sqlite3_config(SQLITE_CONFIG_SINGLETHREAD); break;
31698	33082	case 2: sqlite3_config(SQLITE_CONFIG_MULTITHREAD); break;
31699	33083	default: sqlite3_config(SQLITE_CONFIG_SERIALIZED); break;
31700	33084	}
31701		-#ifdef SQLITE_ENABLE_VFSTRACE
31702	33085	}else if( cli_strcmp(z,"-vfstrace")==0 ){
31703		- extern int vfstrace_register(
31704		- const char *zTraceName,
31705		- const char *zOldVfsName,
31706		- int (xOut)(const char,void*),
31707		- void *pOutArg,
31708		- int makeDefault
31709		- );
31710		- vfstrace_register("trace",0,(int()(const char,void*))fputs,stderr,1);
31711		-#endif
	33086	+ bEnableVfstrace = 1;
31712	33087	#ifdef SQLITE_ENABLE_MULTIPLEX
31713	33088	}else if( cli_strcmp(z,"-multiplex")==0 ){
31714	33089	extern int sqlite3_multiplex_initialize(const char*,int);
31715	33090	sqlite3_multiplex_initialize(0, 1);
31716	33091	#endif
		@@ -31762,11 +33137,11 @@
31762	33137	}else if( cli_strcmp(z,"-safe")==0 ){
31763	33138	/* no-op - catch this on the second pass */
31764	33139	}
31765	33140	}
31766	33141	#ifndef SQLITE_SHELL_FIDDLE
31767		- verify_uninitialized();
	33142	+ if( !bEnableVfstrace ) verify_uninitialized();
31768	33143	#endif
31769	33144
31770	33145
31771	33146	#ifdef SQLITE_SHELL_INIT_PROC
31772	33147	{
		@@ -31786,25 +33161,29 @@
31786	33161	if( zVfs ){
31787	33162	sqlite3_vfs *pVfs = sqlite3_vfs_find(zVfs);
31788	33163	if( pVfs ){
31789	33164	sqlite3_vfs_register(pVfs, 1);
31790	33165	}else{
31791		- eputf("no such VFS: \"%s\"\n", zVfs);
	33166	+ sqlite3_fprintf(stderr,"no such VFS: \"%s\"\n", zVfs);
31792	33167	exit(1);
31793	33168	}
31794	33169	}
31795	33170
31796	33171	if( data.pAuxDb->zDbFilename==0 ){
31797	33172	#ifndef SQLITE_OMIT_MEMORYDB
31798	33173	data.pAuxDb->zDbFilename = ":memory:";
31799	33174	warnInmemoryDb = argc==1;
31800	33175	#else
31801		- eputf("%s: Error: no database filename specified\n", Argv0);
	33176	+ sqlite3_fprintf(stderr,
	33177	+ "%s: Error: no database filename specified\n", Argv0);
31802	33178	return 1;
31803	33179	#endif
31804	33180	}
31805	33181	data.out = stdout;
	33182	+ if( bEnableVfstrace ){
	33183	+ vfstrace_register("trace",0,vfstraceOut, &data, 1);
	33184	+ }
31806	33185	#ifndef SQLITE_SHELL_FIDDLE
31807	33186	sqlite3_appendvfs_init(0,0,0);
31808	33187	#endif
31809	33188
31810	33189	/* Go ahead and open the database file if it already exists. If the
		@@ -31913,11 +33292,11 @@
31913	33292	*/
31914	33293	ShellSetFlag(&data, SHFLG_Backslash);
31915	33294	}else if( cli_strcmp(z,"-bail")==0 ){
31916	33295	/* No-op. The bail_on_error flag should already be set. */
31917	33296	}else if( cli_strcmp(z,"-version")==0 ){
31918		- sputf(stdout, "%s %s (%d-bit)\n",
	33297	+ sqlite3_fprintf(stdout, "%s %s (%d-bit)\n",
31919	33298	sqlite3_libversion(), sqlite3_sourceid(), 8(int)sizeof(char));
31920	33299	return 0;
31921	33300	}else if( cli_strcmp(z,"-interactive")==0 ){
31922	33301	/* Need to check for interactive override here to so that it can
31923	33302	** affect console setup (for Windows only) and testing thereof.
		@@ -31951,14 +33330,12 @@
31951	33330	}else if( cli_strcmp(z,"-sorterref")==0 ){
31952	33331	i++;
31953	33332	#endif
31954	33333	}else if( cli_strcmp(z,"-vfs")==0 ){
31955	33334	i++;
31956		-#ifdef SQLITE_ENABLE_VFSTRACE
31957	33335	}else if( cli_strcmp(z,"-vfstrace")==0 ){
31958	33336	i++;
31959		-#endif
31960	33337	#ifdef SQLITE_ENABLE_MULTIPLEX
31961	33338	}else if( cli_strcmp(z,"-multiplex")==0 ){
31962	33339	i++;
31963	33340	#endif
31964	33341	}else if( cli_strcmp(z,"-help")==0 ){
		@@ -31978,18 +33355,18 @@
31978	33355	rc = shell_exec(&data, z, &zErrMsg);
31979	33356	if( zErrMsg!=0 ){
31980	33357	shellEmitError(zErrMsg);
31981	33358	if( bail_on_error ) return rc!=0 ? rc : 1;
31982	33359	}else if( rc!=0 ){
31983		- eputf("Error: unable to process SQL \"%s\"\n", z);
	33360	+ sqlite3_fprintf(stderr,"Error: unable to process SQL \"%s\"\n", z);
31984	33361	if( bail_on_error ) return rc;
31985	33362	}
31986	33363	}
31987	33364	#if !defined(SQLITE_OMIT_VIRTUALTABLE) && defined(SQLITE_HAVE_ZLIB)
31988	33365	}else if( cli_strncmp(z, "-A", 2)==0 ){
31989	33366	if( nCmd>0 ){
31990		- eputf("Error: cannot mix regular SQL or dot-commands"
	33367	+ sqlite3_fprintf(stderr,"Error: cannot mix regular SQL or dot-commands"
31991	33368	" with \"%s\"\n", z);
31992	33369	return 1;
31993	33370	}
31994	33371	open_db(&data, OPEN_DB_ZIPFILE);
31995	33372	if( z[2] ){
		@@ -32004,11 +33381,11 @@
32004	33381	}else if( cli_strcmp(z,"-safe")==0 ){
32005	33382	data.bSafeMode = data.bSafeModePersist = 1;
32006	33383	}else if( cli_strcmp(z,"-unsafe-testing")==0 ){
32007	33384	/* Acted upon in first pass. */
32008	33385	}else{
32009		- eputf("%s: Error: unknown option: %s\n", Argv0, z);
	33386	+ sqlite3_fprintf(stderr,"%s: Error: unknown option: %s\n", Argv0, z);
32010	33387	eputz("Use -help for a list of options.\n");
32011	33388	return 1;
32012	33389	}
32013	33390	data.cMode = data.mode;
32014	33391	}
		@@ -32031,11 +33408,12 @@
32031	33408	rc = shell_exec(&data, azCmd[i], &zErrMsg);
32032	33409	if( zErrMsg \|\| rc ){
32033	33410	if( zErrMsg!=0 ){
32034	33411	shellEmitError(zErrMsg);
32035	33412	}else{
32036		- eputf("Error: unable to process SQL: %s\n", azCmd[i]);
	33413	+ sqlite3_fprintf(stderr,
	33414	+ "Error: unable to process SQL: %s\n", azCmd[i]);
32037	33415	}
32038	33416	sqlite3_free(zErrMsg);
32039	33417	if( rc==0 ) rc = 1;
32040	33418	goto shell_main_exit;
32041	33419	}
		@@ -32046,18 +33424,14 @@
32046	33424	*/
32047	33425	if( stdin_is_interactive ){
32048	33426	char *zHome;
32049	33427	char *zHistory;
32050	33428	int nHistory;
32051		-#if CIO_WIN_WC_XLATE
32052		-# define SHELL_CIO_CHAR_SET (stdout_is_console? " (UTF-16 console I/O)" : "")
32053		-#else
32054		-# define SHELL_CIO_CHAR_SET ""
32055		-#endif
32056		- sputf(stdout, "SQLite version %s %.19s%s\n" /extra-version-info/
	33429	+ sqlite3_fprintf(stdout,
	33430	+ "SQLite version %s %.19s\n" /extra-version-info/
32057	33431	"Enter \".help\" for usage hints.\n",
32058		- sqlite3_libversion(), sqlite3_sourceid(), SHELL_CIO_CHAR_SET);
	33432	+ sqlite3_libversion(), sqlite3_sourceid());
32059	33433	if( warnInmemoryDb ){
32060	33434	sputz(stdout, "Connected to a ");
32061	33435	printBold("transient in-memory database");
32062	33436	sputz(stdout, ".\nUse \".open FILENAME\" to reopen on a"
32063	33437	" persistent database.\n");
		@@ -32070,13 +33444,15 @@
32070	33444	if( (zHistory = malloc(nHistory))!=0 ){
32071	33445	sqlite3_snprintf(nHistory, zHistory,"%s/.sqlite_history", zHome);
32072	33446	}
32073	33447	}
32074	33448	if( zHistory ){ shell_read_history(zHistory); }
32075		-#if HAVE_READLINE \|\| HAVE_EDITLINE
	33449	+#if (HAVE_READLINE \|\| HAVE_EDITLINE) && !defined(SQLITE_OMIT_READLINE_COMPLETION)
32076	33450	rl_attempted_completion_function = readline_completion;
32077		-#elif HAVE_LINENOISE
	33451	+#elif HAVE_LINENOISE==1
	33452	+ linenoiseSetCompletionCallback(linenoise_completion);
	33453	+#elif HAVE_LINENOISE==2
32078	33454	linenoiseSetCompletionCallback(linenoise_completion, NULL);
32079	33455	#endif
32080	33456	data.in = 0;
32081	33457	rc = process_input(&data);
32082	33458	if( zHistory ){
		@@ -32122,13 +33498,16 @@
32122	33498	free(data.colWidth);
32123	33499	free(data.zNonce);
32124	33500	/* Clear the global data structure so that valgrind will detect memory
32125	33501	** leaks */
32126	33502	memset(&data, 0, sizeof(data));
	33503	+ if( bEnableVfstrace ){
	33504	+ vfstrace_unregister("trace");
	33505	+ }
32127	33506	#ifdef SQLITE_DEBUG
32128	33507	if( sqlite3_memory_used()>mem_main_enter ){
32129		- eputf("Memory leaked: %u bytes\n",
	33508	+ sqlite3_fprintf(stderr,"Memory leaked: %u bytes\n",
32130	33509	(unsigned int)(sqlite3_memory_used()-mem_main_enter));
32131	33510	}
32132	33511	#endif
32133	33512	#else /* SQLITE_SHELL_FIDDLE... */
32134	33513	shell_main_exit:
		@@ -32164,11 +33543,11 @@
32164	33543	return pVfs;
32165	33544	}
32166	33545
32167	33546	/* Only for emcc experimentation purposes. */
32168	33547	sqlite3 * fiddle_db_arg(sqlite3 *arg){
32169		- oputf("fiddle_db_arg(%p)\n", (const void*)arg);
	33548	+ sqlite3_fprintf(stdout, "fiddle_db_arg(%p)\n", (const void*)arg);
32170	33549	return arg;
32171	33550	}
32172	33551
32173	33552	/*
32174	33553	** Intended to be called via a SharedWorker() while a separate
		@@ -32201,11 +33580,11 @@
32201	33580	while( sqlite3_txn_state(globalDb,0)>0 ){
32202	33581	/*
32203	33582	** Resolve problem reported in
32204	33583	** https://sqlite.org/forum/forumpost/0b41a25d65
32205	33584	*/
32206		- oputz("Rolling back in-progress transaction.\n");
	33585	+ sqlite3_fputs("Rolling back in-progress transaction.\n", stdout);
32207	33586	sqlite3_exec(globalDb,"ROLLBACK", 0, 0, 0);
32208	33587	}
32209	33588	rc = sqlite3_db_config(globalDb, SQLITE_DBCONFIG_RESET_DATABASE, 1, 0);
32210	33589	if( 0==rc ) sqlite3_exec(globalDb, "VACUUM", 0, 0, 0);
32211	33590	sqlite3_db_config(globalDb, SQLITE_DBCONFIG_RESET_DATABASE, 0, 0);
32212	33591

	--- extsrc/shell.c
	+++ extsrc/shell.c
	@@ -120,13 +120,10 @@
120	#include <math.h>
121	#include "sqlite3.h"
122	typedef sqlite3_int64 i64;
123	typedef sqlite3_uint64 u64;
124	typedef unsigned char u8;
125	#if SQLITE_USER_AUTHENTICATION
126	# include "sqlite3userauth.h"
127	#endif
128	#include <ctype.h>
129	#include <stdarg.h>
130
131	#if !defined(_WIN32) && !defined(WIN32)
132	# include <signal.h>
	@@ -208,12 +205,10 @@
208	# define unlink _unlink
209	# endif
210	# ifndef strdup
211	# define strdup _strdup
212	# endif
213	# undef popen
214	# define popen _popen
215	# undef pclose
216	# define pclose _pclose
217	# endif
218	#else
219	/* Make sure isatty() has a prototype. */
	@@ -253,10 +248,370 @@
253	/* string conversion routines only needed on Win32 */
254	extern char *sqlite3_win32_unicode_to_utf8(LPCWSTR);
255	extern LPWSTR sqlite3_win32_utf8_to_unicode(const char *zText);
256	#endif
257








































































































































































































































































































































































258	/* Use console I/O package as a direct INCLUDE. */
259	#define SQLITE_INTERNAL_LINKAGE static
260
261	#ifdef SQLITE_SHELL_FIDDLE
262	/* Deselect most features from the console I/O package for Fiddle. */
	@@ -264,1118 +619,13 @@
264	# define SQLITE_CIO_NO_CLASSIFY
265	# define SQLITE_CIO_NO_TRANSLATE
266	# define SQLITE_CIO_NO_SETMODE
267	# define SQLITE_CIO_NO_FLUSH
268	#endif
269	/*********************** Begin ../ext/consio/console_io.h ****************/
270	/*
271	** 2023 November 1
272	**
273	** The author disclaims copyright to this source code. In place of
274	** a legal notice, here is a blessing:
275	**
276	** May you do good and not evil.
277	** May you find forgiveness for yourself and forgive others.
278	** May you share freely, never taking more than you give.
279	**
280	********************************************************************************
281	** This file exposes various interfaces used for console and other I/O
282	** by the SQLite project command-line tools. These interfaces are used
283	** at either source conglomeration time, compilation time, or run time.
284	** This source provides for either inclusion into conglomerated,
285	** "single-source" forms or separate compilation then linking.
286	**
287	** Platform dependencies are "hidden" here by various stratagems so
288	** that, provided certain conditions are met, the programs using this
289	** source or object code compiled from it need no explicit conditional
290	** compilation in their source for their console and stream I/O.
291	**
292	** The symbols and functionality exposed here are not a public API.
293	** This code may change in tandem with other project code as needed.
294	**
295	** When this .h file and its companion .c are directly incorporated into
296	** a source conglomeration (such as shell.c), the preprocessor symbol
297	** CIO_WIN_WC_XLATE is defined as 0 or 1, reflecting whether console I/O
298	** translation for Windows is effected for the build.
299	*/
300	#define HAVE_CONSOLE_IO_H 1
301	#ifndef SQLITE_INTERNAL_LINKAGE
302	# define SQLITE_INTERNAL_LINKAGE extern /* external to translation unit */
303	# include <stdio.h>
304	#else
305	# define SHELL_NO_SYSINC /* Better yet, modify mkshellc.tcl for this. */
306	#endif
307
308	#ifndef SQLITE3_H
309	/* # include "sqlite3.h" */
310	#endif
311
312	#ifndef SQLITE_CIO_NO_CLASSIFY
313
314	/* Define enum for use with following function. */
315	typedef enum StreamsAreConsole {
316	SAC_NoConsole = 0,
317	SAC_InConsole = 1, SAC_OutConsole = 2, SAC_ErrConsole = 4,
318	SAC_AnyConsole = 0x7
319	} StreamsAreConsole;
320
321	/*
322	** Classify the three standard I/O streams according to whether
323	** they are connected to a console attached to the process.
324	**
325	** Returns the bit-wise OR of SAC_{In,Out,Err}Console values,
326	** or SAC_NoConsole if none of the streams reaches a console.
327	**
328	** This function should be called before any I/O is done with
329	** the given streams. As a side-effect, the given inputs are
330	** recorded so that later I/O operations on them may be done
331	** differently than the C library FILE* I/O would be done,
332	** iff the stream is used for the I/O functions that follow,
333	** and to support the ones that use an implicit stream.
334	**
335	** On some platforms, stream or console mode alteration (aka
336	** "Setup") may be made which is undone by consoleRestore().
337	*/
338	SQLITE_INTERNAL_LINKAGE StreamsAreConsole
339	consoleClassifySetup( FILE pfIn, FILE pfOut, FILE *pfErr );
340	/* A usual call for convenience: */
341	#define SQLITE_STD_CONSOLE_INIT() consoleClassifySetup(stdin,stdout,stderr)
342
343	/*
344	** After an initial call to consoleClassifySetup(...), renew
345	** the same setup it effected. (A call not after is an error.)
346	** This will restore state altered by consoleRestore();
347	**
348	** Applications which run an inferior (child) process which
349	** inherits the same I/O streams may call this function after
350	** such a process exits to guard against console mode changes.
351	*/
352	SQLITE_INTERNAL_LINKAGE void consoleRenewSetup(void);
353
354	/*
355	** Undo any side-effects left by consoleClassifySetup(...).
356	**
357	** This should be called after consoleClassifySetup() and
358	** before the process terminates normally. It is suitable
359	** for use with the atexit() C library procedure. After
360	** this call, no console I/O should be done until one of
361	** console{Classify or Renew}Setup(...) is called again.
362	**
363	** Applications which run an inferior (child) process that
364	** inherits the same I/O streams might call this procedure
365	** before so that said process will have a console setup
366	** however users have configured it or come to expect.
367	*/
368	SQLITE_INTERNAL_LINKAGE void SQLITE_CDECL consoleRestore( void );
369
370	#else /* defined(SQLITE_CIO_NO_CLASSIFY) */
371	# define consoleClassifySetup(i,o,e)
372	# define consoleRenewSetup()
373	# define consoleRestore()
374	#endif /* defined(SQLITE_CIO_NO_CLASSIFY) */
375
376	#ifndef SQLITE_CIO_NO_REDIRECT
377	/*
378	** Set stream to be used for the functions below which write
379	** to "the designated X stream", where X is Output or Error.
380	** Returns the previous value.
381	**
382	** Alternatively, pass the special value, invalidFileStream,
383	** to get the designated stream value without setting it.
384	**
385	** Before the designated streams are set, they default to
386	** those passed to consoleClassifySetup(...), and before
387	** that is called they default to stdout and stderr.
388	**
389	** It is error to close a stream so designated, then, without
390	** designating another, use the corresponding {o,e}Emit(...).
391	*/
392	SQLITE_INTERNAL_LINKAGE FILE *invalidFileStream;
393	SQLITE_INTERNAL_LINKAGE FILE setOutputStream(FILE pf);
394	# ifdef CONSIO_SET_ERROR_STREAM
395	SQLITE_INTERNAL_LINKAGE FILE setErrorStream(FILE pf);
396	# endif
397	#else
398	# define setOutputStream(pf)
399	# define setErrorStream(pf)
400	#endif /* !defined(SQLITE_CIO_NO_REDIRECT) */
401
402	#ifndef SQLITE_CIO_NO_TRANSLATE
403	/*
404	** Emit output like fprintf(). If the output is going to the
405	** console and translation from UTF-8 is necessary, perform
406	** the needed translation. Otherwise, write formatted output
407	** to the provided stream almost as-is, possibly with newline
408	** translation as specified by set{Binary,Text}Mode().
409	*/
410	SQLITE_INTERNAL_LINKAGE int fPrintfUtf8(FILE pfO, const char zFormat, ...);
411	/* Like fPrintfUtf8 except stream is always the designated output. */
412	SQLITE_INTERNAL_LINKAGE int oPrintfUtf8(const char *zFormat, ...);
413	/* Like fPrintfUtf8 except stream is always the designated error. */
414	SQLITE_INTERNAL_LINKAGE int ePrintfUtf8(const char *zFormat, ...);
415
416	/*
417	** Emit output like fputs(). If the output is going to the
418	** console and translation from UTF-8 is necessary, perform
419	** the needed translation. Otherwise, write given text to the
420	** provided stream almost as-is, possibly with newline
421	** translation as specified by set{Binary,Text}Mode().
422	*/
423	SQLITE_INTERNAL_LINKAGE int fPutsUtf8(const char z, FILE pfO);
424	/* Like fPutsUtf8 except stream is always the designated output. */
425	SQLITE_INTERNAL_LINKAGE int oPutsUtf8(const char *z);
426	/* Like fPutsUtf8 except stream is always the designated error. */
427	SQLITE_INTERNAL_LINKAGE int ePutsUtf8(const char *z);
428
429	/*
430	** Emit output like fPutsUtf8(), except that the length of the
431	** accepted char or character sequence is limited by nAccept.
432	**
433	** Returns the number of accepted char values.
434	*/
435	#ifdef CONSIO_SPUTB
436	SQLITE_INTERNAL_LINKAGE int
437	fPutbUtf8(FILE pfOut, const char cBuf, int nAccept);
438	/* Like fPutbUtf8 except stream is always the designated output. */
439	#endif
440	SQLITE_INTERNAL_LINKAGE int
441	oPutbUtf8(const char *cBuf, int nAccept);
442	/* Like fPutbUtf8 except stream is always the designated error. */
443	#ifdef CONSIO_EPUTB
444	SQLITE_INTERNAL_LINKAGE int
445	ePutbUtf8(const char *cBuf, int nAccept);
446	#endif
447
448	/*
449	** Flush the given output stream. Return non-zero for success, else 0.
450	*/
451	#if !defined(SQLITE_CIO_NO_FLUSH) && !defined(SQLITE_CIO_NO_SETMODE)
452	SQLITE_INTERNAL_LINKAGE int
453	fFlushBuffer(FILE *pfOut);
454	#endif
455
456	/*
457	** Collect input like fgets(...) with special provisions for input
458	** from the console on such platforms as require same. Newline
459	** translation may be done as set by set{Binary,Text}Mode().
460	** As a convenience, pfIn==NULL is treated as stdin.
461	*/
462	SQLITE_INTERNAL_LINKAGE char* fGetsUtf8(char cBuf, int ncMax, FILE pfIn);
463	/* Like fGetsUtf8 except stream is always the designated input. */
464	/* SQLITE_INTERNAL_LINKAGE char* iGetsUtf8(char cBuf, int ncMax); /
465
466	#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
467
468	#ifndef SQLITE_CIO_NO_SETMODE
469	/*
470	** Set given stream for binary mode, where newline translation is
471	** not done, or for text mode where, for some platforms, newlines
472	** are translated to the platform's conventional char sequence.
473	** If bFlush true, flush the stream.
474	**
475	** An additional side-effect is that if the stream is one passed
476	** to consoleClassifySetup() as an output, it is flushed first.
477	**
478	** Note that binary/text mode has no effect on console I/O
479	** translation. On all platforms, newline to the console starts
480	** a new line and CR,LF chars from the console become a newline.
481	*/
482	SQLITE_INTERNAL_LINKAGE void setBinaryMode(FILE *, short bFlush);
483	SQLITE_INTERNAL_LINKAGE void setTextMode(FILE *, short bFlush);
484	#endif
485
486	#ifdef SQLITE_CIO_PROMPTED_IN
487	typedef struct Prompts {
488	int numPrompts;
489	const char **azPrompts;
490	} Prompts;
491
492	/*
493	** Macros for use of a line editor.
494	**
495	** The following macros define operations involving use of a
496	** line-editing library or simple console interaction.
497	** A "T" argument is a text (char *) buffer or filename.
498	** A "N" argument is an integer.
499	**
500	** SHELL_ADD_HISTORY(T) // Record text as line(s) of history.
501	** SHELL_READ_HISTORY(T) // Read history from file named by T.
502	** SHELL_WRITE_HISTORY(T) // Write history to file named by T.
503	** SHELL_STIFLE_HISTORY(N) // Limit history to N entries.
504	**
505	** A console program which does interactive console input is
506	** expected to call:
507	** SHELL_READ_HISTORY(T) before collecting such input;
508	** SHELL_ADD_HISTORY(T) as record-worthy input is taken;
509	** SHELL_STIFLE_HISTORY(N) after console input ceases; then
510	** SHELL_WRITE_HISTORY(T) before the program exits.
511	*/
512
513	/*
514	** Retrieve a single line of input text from an input stream.
515	**
516	** If pfIn is the input stream passed to consoleClassifySetup(),
517	** and azPrompt is not NULL, then a prompt is issued before the
518	** line is collected, as selected by the isContinuation flag.
519	** Array azPrompt[{0,1}] holds the {main,continuation} prompt.
520	**
521	** If zBufPrior is not NULL then it is a buffer from a prior
522	** call to this routine that can be reused, or will be freed.
523	**
524	** The result is stored in space obtained from malloc() and
525	** must either be freed by the caller or else passed back to
526	** this function as zBufPrior for reuse.
527	**
528	** This function may call upon services of a line-editing
529	** library to interactively collect line edited input.
530	*/
531	SQLITE_INTERNAL_LINKAGE char *
532	shellGetLine(FILE pfIn, char zBufPrior, int nLen,
533	short isContinuation, Prompts azPrompt);
534	#endif /* defined(SQLITE_CIO_PROMPTED_IN) */
535	/*
536	** TBD: Define an interface for application(s) to generate
537	** completion candidates for use by the line-editor.
538	**
539	** This may be premature; the CLI is the only application
540	** that does this. Yet, getting line-editing melded into
541	** console I/O is desirable because a line-editing library
542	** may have to establish console operating mode, possibly
543	** in a way that interferes with the above functionality.
544	*/
545
546	#if !(defined(SQLITE_CIO_NO_UTF8SCAN)&&defined(SQLITE_CIO_NO_TRANSLATE))
547	/* Skip over as much z[] input char sequence as is valid UTF-8,
548	** limited per nAccept char's or whole characters and containing
549	** no char cn such that ((1<<cn) & ccm)!=0. On return, the
550	** sequence z:return (inclusive:exclusive) is validated UTF-8.
551	** Limit: nAccept>=0 => char count, nAccept<0 => character
552	*/
553	SQLITE_INTERNAL_LINKAGE const char*
554	zSkipValidUtf8(const char *z, int nAccept, long ccm);
555
556	#endif
557
558	/*********************** End ../ext/consio/console_io.h ******************/
559	/*********************** Begin ../ext/consio/console_io.c ****************/
560	/*
561	** 2023 November 4
562	**
563	** The author disclaims copyright to this source code. In place of
564	** a legal notice, here is a blessing:
565	**
566	** May you do good and not evil.
567	** May you find forgiveness for yourself and forgive others.
568	** May you share freely, never taking more than you give.
569	**
570	********************************************************************************
571	** This file implements various interfaces used for console and stream I/O
572	** by the SQLite project command-line tools, as explained in console_io.h .
573	** Functions prefixed by "SQLITE_INTERNAL_LINKAGE" behave as described there.
574	*/
575
576	#ifndef SQLITE_CDECL
577	# define SQLITE_CDECL
578	#endif
579
580	#ifndef SHELL_NO_SYSINC
581	# include <stdarg.h>
582	# include <string.h>
583	# include <stdlib.h>
584	# include <limits.h>
585	# include <assert.h>
586	/* # include "sqlite3.h" */
587	#endif
588	#ifndef HAVE_CONSOLE_IO_H
589	# include "console_io.h"
590	#endif
591	#if defined(_MSC_VER)
592	# pragma warning(disable : 4204)
593	#endif
594
595	#ifndef SQLITE_CIO_NO_TRANSLATE
596	# if (defined(_WIN32) \|\| defined(WIN32)) && !SQLITE_OS_WINRT
597	# ifndef SHELL_NO_SYSINC
598	# include <io.h>
599	# include <fcntl.h>
600	# undef WIN32_LEAN_AND_MEAN
601	# define WIN32_LEAN_AND_MEAN
602	# include <windows.h>
603	# endif
604	# define CIO_WIN_WC_XLATE 1 /* Use WCHAR Windows APIs for console I/O */
605	# else
606	# ifndef SHELL_NO_SYSINC
607	# include <unistd.h>
608	# endif
609	# define CIO_WIN_WC_XLATE 0 /* Use plain C library stream I/O at console */
610	# endif
611	#else
612	# define CIO_WIN_WC_XLATE 0 /* Not exposing translation routines at all */
613	#endif
614
615	#if CIO_WIN_WC_XLATE
616	static HANDLE handleOfFile(FILE *pf){
617	int fileDesc = _fileno(pf);
618	union { intptr_t osfh; HANDLE fh; } fid = {
619	(fileDesc>=0)? _get_osfhandle(fileDesc) : (intptr_t)INVALID_HANDLE_VALUE
620	};
621	return fid.fh;
622	}
623	#endif
624
625	#ifndef SQLITE_CIO_NO_TRANSLATE
626	typedef struct PerStreamTags {
627	# if CIO_WIN_WC_XLATE
628	HANDLE hx;
629	DWORD consMode;
630	char acIncomplete[4];
631	# else
632	short reachesConsole;
633	# endif
634	FILE *pf;
635	} PerStreamTags;
636
637	/* Define NULL-like value for things which can validly be 0. */
638	# define SHELL_INVALID_FILE_PTR ((FILE *)~0)
639	# if CIO_WIN_WC_XLATE
640	# define SHELL_INVALID_CONS_MODE 0xFFFF0000
641	# endif
642
643	# if CIO_WIN_WC_XLATE
644	# define PST_INITIALIZER { INVALID_HANDLE_VALUE, SHELL_INVALID_CONS_MODE, \
645	{0,0,0,0}, SHELL_INVALID_FILE_PTR }
646	# else
647	# define PST_INITIALIZER { 0, SHELL_INVALID_FILE_PTR }
648	# endif
649
650	/* Quickly say whether a known output is going to the console. */
651	# if CIO_WIN_WC_XLATE
652	static short pstReachesConsole(PerStreamTags *ppst){
653	return (ppst->hx != INVALID_HANDLE_VALUE);
654	}
655	# else
656	# define pstReachesConsole(ppst) 0
657	# endif
658
659	# if CIO_WIN_WC_XLATE
660	static void restoreConsoleArb(PerStreamTags *ppst){
661	if( pstReachesConsole(ppst) ) SetConsoleMode(ppst->hx, ppst->consMode);
662	}
663	# else
664	# define restoreConsoleArb(ppst)
665	# endif
666
667	/* Say whether FILE* appears to be a console, collect associated info. */
668	static short streamOfConsole(FILE pf, / out / PerStreamTags ppst){
669	# if CIO_WIN_WC_XLATE
670	short rv = 0;
671	DWORD dwCM = SHELL_INVALID_CONS_MODE;
672	HANDLE fh = handleOfFile(pf);
673	ppst->pf = pf;
674	if( INVALID_HANDLE_VALUE != fh ){
675	rv = (GetFileType(fh) == FILE_TYPE_CHAR && GetConsoleMode(fh,&dwCM));
676	}
677	ppst->hx = (rv)? fh : INVALID_HANDLE_VALUE;
678	ppst->consMode = dwCM;
679	return rv;
680	# else
681	ppst->pf = pf;
682	ppst->reachesConsole = ( (short)isatty(fileno(pf)) );
683	return ppst->reachesConsole;
684	# endif
685	}
686
687	# ifndef ENABLE_VIRTUAL_TERMINAL_PROCESSING
688	# define ENABLE_VIRTUAL_TERMINAL_PROCESSING (0x4)
689	# endif
690
691	# if CIO_WIN_WC_XLATE
692	/* Define console modes for use with the Windows Console API. */
693	# define SHELL_CONI_MODE \
694	(ENABLE_ECHO_INPUT \| ENABLE_INSERT_MODE \| ENABLE_LINE_INPUT \| 0x80 \
695	\| ENABLE_QUICK_EDIT_MODE \| ENABLE_EXTENDED_FLAGS \| ENABLE_PROCESSED_INPUT)
696	# define SHELL_CONO_MODE (ENABLE_PROCESSED_OUTPUT \| ENABLE_WRAP_AT_EOL_OUTPUT \
697	\| ENABLE_VIRTUAL_TERMINAL_PROCESSING)
698	# endif
699
700	typedef struct ConsoleInfo {
701	PerStreamTags pstSetup[3];
702	PerStreamTags pstDesignated[3];
703	StreamsAreConsole sacSetup;
704	} ConsoleInfo;
705
706	static short isValidStreamInfo(PerStreamTags *ppst){
707	return (ppst->pf != SHELL_INVALID_FILE_PTR);
708	}
709
710	static ConsoleInfo consoleInfo = {
711	{ /* pstSetup */ PST_INITIALIZER, PST_INITIALIZER, PST_INITIALIZER },
712	{ /* pstDesignated[] */ PST_INITIALIZER, PST_INITIALIZER, PST_INITIALIZER },
713	SAC_NoConsole /* sacSetup */
714	};
715
716	SQLITE_INTERNAL_LINKAGE FILE* invalidFileStream = (FILE *)~0;
717
718	# if CIO_WIN_WC_XLATE
719	static void maybeSetupAsConsole(PerStreamTags *ppst, short odir){
720	if( pstReachesConsole(ppst) ){
721	DWORD cm = odir? SHELL_CONO_MODE : SHELL_CONI_MODE;
722	SetConsoleMode(ppst->hx, cm);
723	}
724	}
725	# else
726	# define maybeSetupAsConsole(ppst,odir)
727	# endif
728
729	SQLITE_INTERNAL_LINKAGE void consoleRenewSetup(void){
730	# if CIO_WIN_WC_XLATE
731	int ix = 0;
732	while( ix < 6 ){
733	PerStreamTags *ppst = (ix<3)?
734	&consoleInfo.pstSetup[ix] : &consoleInfo.pstDesignated[ix-3];
735	maybeSetupAsConsole(ppst, (ix % 3)>0);
736	++ix;
737	}
738	# endif
739	}
740
741	SQLITE_INTERNAL_LINKAGE StreamsAreConsole
742	consoleClassifySetup( FILE pfIn, FILE pfOut, FILE *pfErr ){
743	StreamsAreConsole rv = SAC_NoConsole;
744	FILE* apf[3] = { pfIn, pfOut, pfErr };
745	int ix;
746	for( ix = 2; ix >= 0; --ix ){
747	PerStreamTags *ppst = &consoleInfo.pstSetup[ix];
748	if( streamOfConsole(apf[ix], ppst) ){
749	rv \|= (SAC_InConsole<<ix);
750	}
751	consoleInfo.pstDesignated[ix] = *ppst;
752	if( ix > 0 ) fflush(apf[ix]);
753	}
754	consoleInfo.sacSetup = rv;
755	consoleRenewSetup();
756	return rv;
757	}
758
759	SQLITE_INTERNAL_LINKAGE void SQLITE_CDECL consoleRestore( void ){
760	# if CIO_WIN_WC_XLATE
761	static ConsoleInfo *pci = &consoleInfo;
762	if( pci->sacSetup ){
763	int ix;
764	for( ix=0; ix<3; ++ix ){
765	if( pci->sacSetup & (SAC_InConsole<<ix) ){
766	PerStreamTags *ppst = &pci->pstSetup[ix];
767	SetConsoleMode(ppst->hx, ppst->consMode);
768	}
769	}
770	}
771	# endif
772	}
773	#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
774
775	#ifdef SQLITE_CIO_INPUT_REDIR
776	/* Say whether given FILE* is among those known, via either
777	** consoleClassifySetup() or set{Output,Error}Stream, as
778	** readable, and return an associated PerStreamTags pointer
779	** if so. Otherwise, return 0.
780	*/
781	static PerStreamTags * isKnownReadable(FILE *pf){
782	static PerStreamTags *apst[] = {
783	&consoleInfo.pstDesignated[0], &consoleInfo.pstSetup[0], 0
784	};
785	int ix = 0;
786	do {
787	if( apst[ix]->pf == pf ) break;
788	} while( apst[++ix] != 0 );
789	return apst[ix];
790	}
791	#endif
792
793	#ifndef SQLITE_CIO_NO_TRANSLATE
794	/* Say whether given FILE* is among those known, via either
795	** consoleClassifySetup() or set{Output,Error}Stream, as
796	** writable, and return an associated PerStreamTags pointer
797	** if so. Otherwise, return 0.
798	*/
799	static PerStreamTags * isKnownWritable(FILE *pf){
800	static PerStreamTags *apst[] = {
801	&consoleInfo.pstDesignated[1], &consoleInfo.pstDesignated[2],
802	&consoleInfo.pstSetup[1], &consoleInfo.pstSetup[2], 0
803	};
804	int ix = 0;
805	do {
806	if( apst[ix]->pf == pf ) break;
807	} while( apst[++ix] != 0 );
808	return apst[ix];
809	}
810
811	static FILE designateEmitStream(FILE pf, unsigned chix){
812	FILE *rv = consoleInfo.pstDesignated[chix].pf;
813	if( pf == invalidFileStream ) return rv;
814	else{
815	/* Setting a possibly new output stream. */
816	PerStreamTags *ppst = isKnownWritable(pf);
817	if( ppst != 0 ){
818	PerStreamTags pst = *ppst;
819	consoleInfo.pstDesignated[chix] = pst;
820	}else streamOfConsole(pf, &consoleInfo.pstDesignated[chix]);
821	}
822	return rv;
823	}
824
825	SQLITE_INTERNAL_LINKAGE FILE setOutputStream(FILE pf){
826	return designateEmitStream(pf, 1);
827	}
828	# ifdef CONSIO_SET_ERROR_STREAM
829	SQLITE_INTERNAL_LINKAGE FILE setErrorStream(FILE pf){
830	return designateEmitStream(pf, 2);
831	}
832	# endif
833	#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
834
835	#ifndef SQLITE_CIO_NO_SETMODE
836	# if CIO_WIN_WC_XLATE
837	static void setModeFlushQ(FILE *pf, short bFlush, int mode){
838	if( bFlush ) fflush(pf);
839	_setmode(_fileno(pf), mode);
840	}
841	# else
842	# define setModeFlushQ(f, b, m) if(b) fflush(f)
843	# endif
844
845	SQLITE_INTERNAL_LINKAGE void setBinaryMode(FILE *pf, short bFlush){
846	setModeFlushQ(pf, bFlush, _O_BINARY);
847	}
848	SQLITE_INTERNAL_LINKAGE void setTextMode(FILE *pf, short bFlush){
849	setModeFlushQ(pf, bFlush, _O_TEXT);
850	}
851	# undef setModeFlushQ
852
853	#else /* defined(SQLITE_CIO_NO_SETMODE) */
854	# define setBinaryMode(f, bFlush) do{ if((bFlush)) fflush(f); }while(0)
855	# define setTextMode(f, bFlush) do{ if((bFlush)) fflush(f); }while(0)
856	#endif /* defined(SQLITE_CIO_NO_SETMODE) */
857
858	#ifndef SQLITE_CIO_NO_TRANSLATE
859	# if CIO_WIN_WC_XLATE
860	/* Write buffer cBuf as output to stream known to reach console,
861	** limited to ncTake char's. Return ncTake on success, else 0. */
862	static int conZstrEmit(PerStreamTags ppst, const char z, int ncTake){
863	int rv = 0;
864	if( z!=NULL ){
865	int nwc = MultiByteToWideChar(CP_UTF8,0, z,ncTake, 0,0);
866	if( nwc > 0 ){
867	WCHAR zw = sqlite3_malloc64(nwcsizeof(WCHAR));
868	if( zw!=NULL ){
869	nwc = MultiByteToWideChar(CP_UTF8,0, z,ncTake, zw,nwc);
870	if( nwc > 0 ){
871	/* Translation from UTF-8 to UTF-16, then WCHARs out. */
872	if( WriteConsoleW(ppst->hx, zw,nwc, 0, NULL) ){
873	rv = ncTake;
874	}
875	}
876	sqlite3_free(zw);
877	}
878	}
879	}
880	return rv;
881	}
882
883	/* For {f,o,e}PrintfUtf8() when stream is known to reach console. */
884	static int conioVmPrintf(PerStreamTags ppst, const char zFormat, va_list ap){
885	char *z = sqlite3_vmprintf(zFormat, ap);
886	if( z ){
887	int rv = conZstrEmit(ppst, z, (int)strlen(z));
888	sqlite3_free(z);
889	return rv;
890	}else return 0;
891	}
892	# endif /* CIO_WIN_WC_XLATE */
893
894	# ifdef CONSIO_GET_EMIT_STREAM
895	static PerStreamTags * getDesignatedEmitStream(FILE *pf, unsigned chix,
896	PerStreamTags *ppst){
897	PerStreamTags *rv = isKnownWritable(pf);
898	short isValid = (rv!=0)? isValidStreamInfo(rv) : 0;
899	if( rv != 0 && isValid ) return rv;
900	streamOfConsole(pf, ppst);
901	return ppst;
902	}
903	# endif
904
905	/* Get stream info, either for designated output or error stream when
906	** chix equals 1 or 2, or for an arbitrary stream when chix == 0.
907	** In either case, ppst references a caller-owned PerStreamTags
908	** struct which may be filled in if none of the known writable
909	** streams is being held by consoleInfo. The ppf parameter is a
910	** byref output when chix!=0 and a byref input when chix==0.
911	*/
912	static PerStreamTags *
913	getEmitStreamInfo(unsigned chix, PerStreamTags *ppst,
914	/* in/out / FILE *ppf){
915	PerStreamTags *ppstTry;
916	FILE *pfEmit;
917	if( chix > 0 ){
918	ppstTry = &consoleInfo.pstDesignated[chix];
919	if( !isValidStreamInfo(ppstTry) ){
920	ppstTry = &consoleInfo.pstSetup[chix];
921	pfEmit = ppst->pf;
922	}else pfEmit = ppstTry->pf;
923	if( !isValidStreamInfo(ppstTry) ){
924	pfEmit = (chix > 1)? stderr : stdout;
925	ppstTry = ppst;
926	streamOfConsole(pfEmit, ppstTry);
927	}
928	*ppf = pfEmit;
929	}else{
930	ppstTry = isKnownWritable(*ppf);
931	if( ppstTry != 0 ) return ppstTry;
932	streamOfConsole(*ppf, ppst);
933	return ppst;
934	}
935	return ppstTry;
936	}
937
938	SQLITE_INTERNAL_LINKAGE int oPrintfUtf8(const char *zFormat, ...){
939	va_list ap;
940	int rv;
941	FILE *pfOut;
942	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
943	# if CIO_WIN_WC_XLATE
944	PerStreamTags *ppst = getEmitStreamInfo(1, &pst, &pfOut);
945	# else
946	getEmitStreamInfo(1, &pst, &pfOut);
947	# endif
948	assert(zFormat!=0);
949	va_start(ap, zFormat);
950	# if CIO_WIN_WC_XLATE
951	if( pstReachesConsole(ppst) ){
952	rv = conioVmPrintf(ppst, zFormat, ap);
953	}else{
954	# endif
955	rv = vfprintf(pfOut, zFormat, ap);
956	# if CIO_WIN_WC_XLATE
957	}
958	# endif
959	va_end(ap);
960	return rv;
961	}
962
963	SQLITE_INTERNAL_LINKAGE int ePrintfUtf8(const char *zFormat, ...){
964	va_list ap;
965	int rv;
966	FILE *pfErr;
967	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
968	# if CIO_WIN_WC_XLATE
969	PerStreamTags *ppst = getEmitStreamInfo(2, &pst, &pfErr);
970	# else
971	getEmitStreamInfo(2, &pst, &pfErr);
972	# endif
973	assert(zFormat!=0);
974	va_start(ap, zFormat);
975	# if CIO_WIN_WC_XLATE
976	if( pstReachesConsole(ppst) ){
977	rv = conioVmPrintf(ppst, zFormat, ap);
978	}else{
979	# endif
980	rv = vfprintf(pfErr, zFormat, ap);
981	# if CIO_WIN_WC_XLATE
982	}
983	# endif
984	va_end(ap);
985	return rv;
986	}
987
988	SQLITE_INTERNAL_LINKAGE int fPrintfUtf8(FILE pfO, const char zFormat, ...){
989	va_list ap;
990	int rv;
991	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
992	# if CIO_WIN_WC_XLATE
993	PerStreamTags *ppst = getEmitStreamInfo(0, &pst, &pfO);
994	# else
995	getEmitStreamInfo(0, &pst, &pfO);
996	# endif
997	assert(zFormat!=0);
998	va_start(ap, zFormat);
999	# if CIO_WIN_WC_XLATE
1000	if( pstReachesConsole(ppst) ){
1001	maybeSetupAsConsole(ppst, 1);
1002	rv = conioVmPrintf(ppst, zFormat, ap);
1003	if( 0 == isKnownWritable(ppst->pf) ) restoreConsoleArb(ppst);
1004	}else{
1005	# endif
1006	rv = vfprintf(pfO, zFormat, ap);
1007	# if CIO_WIN_WC_XLATE
1008	}
1009	# endif
1010	va_end(ap);
1011	return rv;
1012	}
1013
1014	SQLITE_INTERNAL_LINKAGE int fPutsUtf8(const char z, FILE pfO){
1015	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1016	# if CIO_WIN_WC_XLATE
1017	PerStreamTags *ppst = getEmitStreamInfo(0, &pst, &pfO);
1018	# else
1019	getEmitStreamInfo(0, &pst, &pfO);
1020	# endif
1021	assert(z!=0);
1022	# if CIO_WIN_WC_XLATE
1023	if( pstReachesConsole(ppst) ){
1024	int rv;
1025	maybeSetupAsConsole(ppst, 1);
1026	rv = conZstrEmit(ppst, z, (int)strlen(z));
1027	if( 0 == isKnownWritable(ppst->pf) ) restoreConsoleArb(ppst);
1028	return rv;
1029	}else {
1030	# endif
1031	return (fputs(z, pfO)<0)? 0 : (int)strlen(z);
1032	# if CIO_WIN_WC_XLATE
1033	}
1034	# endif
1035	}
1036
1037	SQLITE_INTERNAL_LINKAGE int ePutsUtf8(const char *z){
1038	FILE *pfErr;
1039	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1040	# if CIO_WIN_WC_XLATE
1041	PerStreamTags *ppst = getEmitStreamInfo(2, &pst, &pfErr);
1042	# else
1043	getEmitStreamInfo(2, &pst, &pfErr);
1044	# endif
1045	assert(z!=0);
1046	# if CIO_WIN_WC_XLATE
1047	if( pstReachesConsole(ppst) ) return conZstrEmit(ppst, z, (int)strlen(z));
1048	else {
1049	# endif
1050	return (fputs(z, pfErr)<0)? 0 : (int)strlen(z);
1051	# if CIO_WIN_WC_XLATE
1052	}
1053	# endif
1054	}
1055
1056	SQLITE_INTERNAL_LINKAGE int oPutsUtf8(const char *z){
1057	FILE *pfOut;
1058	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1059	# if CIO_WIN_WC_XLATE
1060	PerStreamTags *ppst = getEmitStreamInfo(1, &pst, &pfOut);
1061	# else
1062	getEmitStreamInfo(1, &pst, &pfOut);
1063	# endif
1064	assert(z!=0);
1065	# if CIO_WIN_WC_XLATE
1066	if( pstReachesConsole(ppst) ) return conZstrEmit(ppst, z, (int)strlen(z));
1067	else {
1068	# endif
1069	return (fputs(z, pfOut)<0)? 0 : (int)strlen(z);
1070	# if CIO_WIN_WC_XLATE
1071	}
1072	# endif
1073	}
1074
1075	#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
1076
1077	#if !(defined(SQLITE_CIO_NO_UTF8SCAN) && defined(SQLITE_CIO_NO_TRANSLATE))
1078	/* Skip over as much z[] input char sequence as is valid UTF-8,
1079	** limited per nAccept char's or whole characters and containing
1080	** no char cn such that ((1<<cn) & ccm)!=0. On return, the
1081	** sequence z:return (inclusive:exclusive) is validated UTF-8.
1082	** Limit: nAccept>=0 => char count, nAccept<0 => character
1083	*/
1084	SQLITE_INTERNAL_LINKAGE const char*
1085	zSkipValidUtf8(const char *z, int nAccept, long ccm){
1086	int ng = (nAccept<0)? -nAccept : 0;
1087	const char *pcLimit = (nAccept>=0)? z+nAccept : 0;
1088	assert(z!=0);
1089	while( (pcLimit)? (z<pcLimit) : (ng-- != 0) ){
1090	char c = *z;
1091	if( (c & 0x80) == 0 ){
1092	if( ccm != 0L && c < 0x20 && ((1L<<c) & ccm) != 0 ) return z;
1093	++z; /* ASCII */
1094	}else if( (c & 0xC0) != 0xC0 ) return z; /* not a lead byte */
1095	else{
1096	const char zt = z+1; / Got lead byte, look at trail bytes.*/
1097	do{
1098	if( pcLimit && zt >= pcLimit ) return z;
1099	else{
1100	char ct = *zt++;
1101	if( ct==0 \|\| (zt-z)>4 \|\| (ct & 0xC0)!=0x80 ){
1102	/* Trailing bytes are too few, too many, or invalid. */
1103	return z;
1104	}
1105	}
1106	} while( ((c <<= 1) & 0x40) == 0x40 ); /* Eat lead byte's count. */
1107	z = zt;
1108	}
1109	}
1110	return z;
1111	}
1112	#endif /!(defined(SQLITE_CIO_NO_UTF8SCAN)&&defined(SQLITE_CIO_NO_TRANSLATE))/
1113
1114	#ifndef SQLITE_CIO_NO_TRANSLATE
1115	# ifdef CONSIO_SPUTB
1116	SQLITE_INTERNAL_LINKAGE int
1117	fPutbUtf8(FILE pfO, const char cBuf, int nAccept){
1118	assert(pfO!=0);
1119	# if CIO_WIN_WC_XLATE
1120	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1121	PerStreamTags *ppst = getEmitStreamInfo(0, &pst, &pfO);
1122	if( pstReachesConsole(ppst) ){
1123	int rv;
1124	maybeSetupAsConsole(ppst, 1);
1125	rv = conZstrEmit(ppst, cBuf, nAccept);
1126	if( 0 == isKnownWritable(ppst->pf) ) restoreConsoleArb(ppst);
1127	return rv;
1128	}else {
1129	# endif
1130	return (int)fwrite(cBuf, 1, nAccept, pfO);
1131	# if CIO_WIN_WC_XLATE
1132	}
1133	# endif
1134	}
1135	# endif
1136
1137	SQLITE_INTERNAL_LINKAGE int
1138	oPutbUtf8(const char *cBuf, int nAccept){
1139	FILE *pfOut;
1140	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1141	# if CIO_WIN_WC_XLATE
1142	PerStreamTags *ppst = getEmitStreamInfo(1, &pst, &pfOut);
1143	# else
1144	getEmitStreamInfo(1, &pst, &pfOut);
1145	# endif
1146	# if CIO_WIN_WC_XLATE
1147	if( pstReachesConsole(ppst) ){
1148	return conZstrEmit(ppst, cBuf, nAccept);
1149	}else {
1150	# endif
1151	return (int)fwrite(cBuf, 1, nAccept, pfOut);
1152	# if CIO_WIN_WC_XLATE
1153	}
1154	# endif
1155	}
1156
1157	/*
1158	** Flush the given output stream. Return non-zero for success, else 0.
1159	*/
1160	#if !defined(SQLITE_CIO_NO_FLUSH) && !defined(SQLITE_CIO_NO_SETMODE)
1161	SQLITE_INTERNAL_LINKAGE int
1162	fFlushBuffer(FILE *pfOut){
1163	# if CIO_WIN_WC_XLATE && !defined(SHELL_OMIT_FIO_DUPE)
1164	return FlushFileBuffers(handleOfFile(pfOut))? 1 : 0;
1165	# else
1166	return fflush(pfOut);
1167	# endif
1168	}
1169	#endif
1170
1171	#if CIO_WIN_WC_XLATE \
1172	&& !defined(SHELL_OMIT_FIO_DUPE) \
1173	&& defined(SQLITE_USE_ONLY_WIN32)
1174	static struct FileAltIds {
1175	int fd;
1176	HANDLE fh;
1177	} altIdsOfFile(FILE *pf){
1178	struct FileAltIds rv = { _fileno(pf) };
1179	union { intptr_t osfh; HANDLE fh; } fid = {
1180	(rv.fd>=0)? _get_osfhandle(rv.fd) : (intptr_t)INVALID_HANDLE_VALUE
1181	};
1182	rv.fh = fid.fh;
1183	return rv;
1184	}
1185
1186	SQLITE_INTERNAL_LINKAGE size_t
1187	cfWrite(const void buf, size_t osz, size_t ocnt, FILE pf){
1188	size_t rv = 0;
1189	struct FileAltIds fai = altIdsOfFile(pf);
1190	int fmode = _setmode(fai.fd, _O_BINARY);
1191	_setmode(fai.fd, fmode);
1192	while( rv < ocnt ){
1193	size_t nbo = osz;
1194	while( nbo > 0 ){
1195	DWORD dwno = (nbo>(1L<<24))? 1L<<24 : (DWORD)nbo;
1196	BOOL wrc = TRUE;
1197	BOOL genCR = (fmode & _O_TEXT)!=0;
1198	if( genCR ){
1199	const char pnl = (const char)memchr(buf, '\n', nbo);
1200	if( pnl ) nbo = pnl - (const char*)buf;
1201	else genCR = 0;
1202	}
1203	if( dwno>0 ) wrc = WriteFile(fai.fh, buf, dwno, 0,0);
1204	if( genCR && wrc ){
1205	wrc = WriteFile(fai.fh, "\r\n", 2, 0,0);
1206	++dwno; /* Skip over the LF */
1207	}
1208	if( !wrc ) return rv;
1209	buf = (const char*)buf + dwno;
1210	nbo += dwno;
1211	}
1212	++rv;
1213	}
1214	return rv;
1215	}
1216
1217	SQLITE_INTERNAL_LINKAGE char *
1218	cfGets(char cBuf, int n, FILE pf){
1219	int nci = 0;
1220	struct FileAltIds fai = altIdsOfFile(pf);
1221	int fmode = _setmode(fai.fd, _O_BINARY);
1222	BOOL eatCR = (fmode & _O_TEXT)!=0;
1223	_setmode(fai.fd, fmode);
1224	while( nci < n-1 ){
1225	DWORD nr;
1226	if( !ReadFile(fai.fh, cBuf+nci, 1, &nr, 0) \|\| nr==0 ) break;
1227	if( nr>0 && (!eatCR \|\| cBuf[nci]!='\r') ) nci += nr;
1228	}
1229	if( nci < n ) cBuf[nci] = 0;
1230	return (nci>0)? cBuf : 0;
1231	}
1232	# else
1233	# define cfWrite(b,os,no,f) fwrite(b,os,no,f)
1234	# define cfGets(b,n,f) fgets(b,n,f)
1235	# endif
1236
1237	# ifdef CONSIO_EPUTB
1238	SQLITE_INTERNAL_LINKAGE int
1239	ePutbUtf8(const char *cBuf, int nAccept){
1240	FILE *pfErr;
1241	PerStreamTags pst = PST_INITIALIZER; /* for unknown streams */
1242	PerStreamTags *ppst = getEmitStreamInfo(2, &pst, &pfErr);
1243	# if CIO_WIN_WC_XLATE
1244	if( pstReachesConsole(ppst) ){
1245	return conZstrEmit(ppst, cBuf, nAccept);
1246	}else {
1247	# endif
1248	return (int)cfWrite(cBuf, 1, nAccept, pfErr);
1249	# if CIO_WIN_WC_XLATE
1250	}
1251	# endif
1252	}
1253	# endif /* defined(CONSIO_EPUTB) */
1254
1255	SQLITE_INTERNAL_LINKAGE char* fGetsUtf8(char cBuf, int ncMax, FILE pfIn){
1256	if( pfIn==0 ) pfIn = stdin;
1257	# if CIO_WIN_WC_XLATE
1258	if( pfIn == consoleInfo.pstSetup[0].pf
1259	&& (consoleInfo.sacSetup & SAC_InConsole)!=0 ){
1260	# if CIO_WIN_WC_XLATE==1
1261	# define SHELL_GULP 150 /* Count of WCHARS to be gulped at a time */
1262	WCHAR wcBuf[SHELL_GULP+1];
1263	int lend = 0, noc = 0;
1264	if( ncMax > 0 ) cBuf[0] = 0;
1265	while( noc < ncMax-8-1 && !lend ){
1266	/* There is room for at least 2 more characters and a 0-terminator. */
1267	int na = (ncMax > SHELL_GULP*4+1 + noc)? SHELL_GULP : (ncMax-1 - noc)/4;
1268	# undef SHELL_GULP
1269	DWORD nbr = 0;
1270	BOOL bRC = ReadConsoleW(consoleInfo.pstSetup[0].hx, wcBuf, na, &nbr, 0);
1271	if( bRC && nbr>0 && (wcBuf[nbr-1]&0xF800)==0xD800 ){
1272	/* Last WHAR read is first of a UTF-16 surrogate pair. Grab its mate. */
1273	DWORD nbrx;
1274	bRC &= ReadConsoleW(consoleInfo.pstSetup[0].hx, wcBuf+nbr, 1, &nbrx, 0);
1275	if( bRC ) nbr += nbrx;
1276	}
1277	if( !bRC \|\| (noc==0 && nbr==0) ) return 0;
1278	if( nbr > 0 ){
1279	int nmb = WideCharToMultiByte(CP_UTF8, 0, wcBuf,nbr,0,0,0,0);
1280	if( nmb != 0 && noc+nmb <= ncMax ){
1281	int iseg = noc;
1282	nmb = WideCharToMultiByte(CP_UTF8, 0, wcBuf,nbr,cBuf+noc,nmb,0,0);
1283	noc += nmb;
1284	/* Fixup line-ends as coded by Windows for CR (or "Enter".)
1285	** This is done without regard for any setMode{Text,Binary}()
1286	** call that might have been done on the interactive input.
1287	*/
1288	if( noc > 0 ){
1289	if( cBuf[noc-1]=='\n' ){
1290	lend = 1;
1291	if( noc > 1 && cBuf[noc-2]=='\r' ) cBuf[--noc-1] = '\n';
1292	}
1293	}
1294	/* Check for ^Z (anywhere in line) too, to act as EOF. */
1295	while( iseg < noc ){
1296	if( cBuf[iseg]=='\x1a' ){
1297	noc = iseg; /* Chop ^Z and anything following. */
1298	lend = 1; /* Counts as end of line too. */
1299	break;
1300	}
1301	++iseg;
1302	}
1303	}else break; /* Drop apparent garbage in. (Could assert.) */
1304	}else break;
1305	}
1306	/* If got nothing, (after ^Z chop), must be at end-of-file. */
1307	if( noc > 0 ){
1308	cBuf[noc] = 0;
1309	return cBuf;
1310	}else return 0;
1311	# endif
1312	}else{
1313	# endif
1314	return cfGets(cBuf, ncMax, pfIn);
1315	# if CIO_WIN_WC_XLATE
1316	}
1317	# endif
1318	}
1319	#endif /* !defined(SQLITE_CIO_NO_TRANSLATE) */
1320
1321	#if defined(_MSC_VER)
1322	# pragma warning(default : 4204)
1323	#endif
1324
1325	#undef SHELL_INVALID_FILE_PTR
1326
1327	/*********************** End ../ext/consio/console_io.c ******************/
1328
1329	#ifndef SQLITE_SHELL_FIDDLE
1330
1331	/* From here onward, fgets() is redirected to the console_io library. */
1332	# define fgets(b,n,f) fGetsUtf8(b,n,f)
1333	/*
1334	* Define macros for emitting output text in various ways:
1335	* sputz(s, z) => emit 0-terminated string z to given stream s
1336	* sputf(s, f, ...) => emit varargs per format f to given stream s
1337	* oputz(z) => emit 0-terminated string z to default stream
1338	* oputf(f, ...) => emit varargs per format f to default stream
1339	* eputz(z) => emit 0-terminated string z to error stream
1340	* eputf(f, ...) => emit varargs per format f to error stream
1341	* oputb(b, n) => emit char buffer b[0..n-1] to default stream
1342	*
1343	* Note that the default stream is whatever has been last set via:
1344	* setOutputStream(FILE *pf)
1345	* This is normally the stream that CLI normal output goes to.
1346	* For the stand-alone CLI, it is stdout with no .output redirect.
1347	*
1348	* The ?putz(z) forms are required for the Fiddle builds for string literal
1349	* output, in aid of enforcing format string to argument correspondence.
1350	*/
1351	# define sputz(s,z) fPutsUtf8(z,s)
1352	# define sputf fPrintfUtf8
1353	# define oputz(z) oPutsUtf8(z)
1354	# define oputf oPrintfUtf8
1355	# define eputz(z) ePutsUtf8(z)
1356	# define eputf ePrintfUtf8
1357	# define oputb(buf,na) oPutbUtf8(buf,na)
1358	# define fflush(s) fFlushBuffer(s);
1359
1360	#else
1361	/* For Fiddle, all console handling and emit redirection is omitted. */
1362	/* These next 3 macros are for emitting formatted output. When complaints
1363	* from the WASM build are issued for non-formatted output, when a mere
1364	* string literal is to be emitted, the ?putz(z) forms should be used.
1365	* (This permits compile-time checking of format string / argument mismatch.)
1366	*/
1367	# define oputf(fmt, ...) printf(fmt,__VA_ARGS__)
1368	# define eputf(fmt, ...) fprintf(stderr,fmt,__VA_ARGS__)
1369	# define sputf(fp,fmt, ...) fprintf(fp,fmt,__VA_ARGS__)
1370	/* These next 3 macros are for emitting simple string literals. */
1371	# define oputz(z) fputs(z,stdout)
1372	# define eputz(z) fputs(z,stderr)
1373	# define sputz(fp,z) fputs(z,fp)
1374	# define oputb(buf,na) fwrite(buf,1,na,stdout)
1375	# undef fflush
1376	#endif
1377
1378	/* True if the timer is enabled */
1379	static int enableTimer = 0;
1380
1381	/* A version of strcmp() that works with NULL values */
	@@ -1416,10 +666,11 @@
1416	struct timeval ru_utime; /* user CPU time used */
1417	struct timeval ru_stime; /* system CPU time used */
1418	};
1419	#define getrusage(A,B) memset(B,0,sizeof(*B))
1420	#endif

1421
1422	/* Saved resource information for the beginning of an operation */
1423	static struct rusage sBegin; /* CPU time at start */
1424	static sqlite3_int64 iBegin; /* Wall-clock time at start */
1425
	@@ -1440,24 +691,24 @@
1440	}
1441
1442	/*
1443	** Print the timing results.
1444	*/
1445	static void endTimer(void){
1446	if( enableTimer ){
1447	sqlite3_int64 iEnd = timeOfDay();
1448	struct rusage sEnd;
1449	getrusage(RUSAGE_SELF, &sEnd);
1450	sputf(stdout, "Run Time: real %.3f user %f sys %f\n",
1451	(iEnd - iBegin)*0.001,
1452	timeDiff(&sBegin.ru_utime, &sEnd.ru_utime),
1453	timeDiff(&sBegin.ru_stime, &sEnd.ru_stime));
1454	}
1455	}
1456
1457	#define BEGIN_TIMER beginTimer()
1458	#define END_TIMER endTimer()
1459	#define HAS_TIMER 1
1460
1461	#elif (defined(_WIN32) \|\| defined(WIN32))
1462
1463	/* Saved resource information for the beginning of an operation */
	@@ -1519,29 +770,29 @@
1519	}
1520
1521	/*
1522	** Print the timing results.
1523	*/
1524	static void endTimer(void){
1525	if( enableTimer && getProcessTimesAddr){
1526	FILETIME ftCreation, ftExit, ftKernelEnd, ftUserEnd;
1527	sqlite3_int64 ftWallEnd = timeOfDay();
1528	getProcessTimesAddr(hProcess,&ftCreation,&ftExit,&ftKernelEnd,&ftUserEnd);
1529	sputf(stdout, "Run Time: real %.3f user %f sys %f\n",
1530	(ftWallEnd - ftWallBegin)*0.001,
1531	timeDiff(&ftUserBegin, &ftUserEnd),
1532	timeDiff(&ftKernelBegin, &ftKernelEnd));
1533	}
1534	}
1535
1536	#define BEGIN_TIMER beginTimer()
1537	#define END_TIMER endTimer()
1538	#define HAS_TIMER hasTimer()
1539
1540	#else
1541	#define BEGIN_TIMER
1542	#define END_TIMER
1543	#define HAS_TIMER 0
1544	#endif
1545
1546	/*
1547	** Used to prevent warnings about unused parameters
	@@ -1738,41 +989,216 @@
1738	char *z;
1739	if( iotrace==0 ) return;
1740	va_start(ap, zFormat);
1741	z = sqlite3_vmprintf(zFormat, ap);
1742	va_end(ap);
1743	sputf(iotrace, "%s", z);
1744	sqlite3_free(z);
1745	}
1746	#endif
1747







































































1748	/*
1749	** Output string zUtf to Out stream as w characters. If w is negative,


























































































1750	** then right-justify the text. W is the width in UTF-8 characters, not
1751	** in bytes. This is different from the %.s specification in printf
1752	** since with %.s the width is measured in bytes, not characters.




1753	*/
1754	static void utf8_width_print(int w, const char *zUtf){
1755	int i;
1756	int n;


1757	int aw = w<0 ? -w : w;
1758	if( zUtf==0 ) zUtf = "";
1759	for(i=n=0; zUtf[i]; i++){
1760	if( (zUtf[i]&0xc0)!=0x80 ){











1761	n++;
1762	if( n==aw ){
1763	do{ i++; }while( (zUtf[i]&0xc0)==0x80 );
1764	break;
1765	}
1766	}
1767	}
1768	if( n>=aw ){
1769	oputf("%.*s", i, zUtf);
1770	}else if( w<0 ){
1771	oputf("%*s%s", aw-n, "", zUtf);
1772	}else{
1773	oputf("%s%*s", zUtf, aw-n, "");
1774	}
1775	}
1776
1777
1778	/*
	@@ -1834,11 +1260,11 @@
1834	struct __stat64 x = {0};
1835	# define STAT_CHR_SRC(mode) ((mode & (_S_IFCHR\|_S_IFIFO\|_S_IFREG))!=0)
1836	/* On Windows, open first, then check the stream nature. This order
1837	** is necessary because _stat() and sibs, when checking a named pipe,
1838	** effectively break the pipe as its supplier sees it. */
1839	FILE *rv = fopen(zFile, "rb");
1840	if( rv==0 ) return 0;
1841	if( _fstat64(_fileno(rv), &x) != 0
1842	\|\| !STAT_CHR_SRC(x.st_mode)){
1843	fclose(rv);
1844	rv = 0;
	@@ -1848,11 +1274,11 @@
1848	struct stat x = {0};
1849	int rc = stat(zFile, &x);
1850	# define STAT_CHR_SRC(mode) (S_ISREG(mode)\|\|S_ISFIFO(mode)\|\|S_ISCHR(mode))
1851	if( rc!=0 ) return 0;
1852	if( STAT_CHR_SRC(x.st_mode) ){
1853	return fopen(zFile, "rb");
1854	}else{
1855	return 0;
1856	}
1857	#endif
1858	#undef STAT_CHR_SRC
	@@ -1875,11 +1301,11 @@
1875	if( n+100>nLine ){
1876	nLine = nLine*2 + 100;
1877	zLine = realloc(zLine, nLine);
1878	shell_check_oom(zLine);
1879	}
1880	if( fgets(&zLine[n], nLine - n, in)==0 ){
1881	if( n==0 ){
1882	free(zLine);
1883	return 0;
1884	}
1885	zLine[n] = 0;
	@@ -2941,11 +2367,11 @@
2941	**
2942	******************************************************************************
2943	**
2944	** This SQLite extension implements functions that compute SHA3 hashes
2945	** in the way described by the (U.S.) NIST FIPS 202 SHA-3 Standard.
2946	** Two SQL functions are implemented:
2947	**
2948	** sha3(X,SIZE)
2949	** sha3_agg(Y,SIZE)
2950	** sha3_query(Z,SIZE)
2951	**
	@@ -3783,10 +3209,422 @@
3783	}
3784	return rc;
3785	}
3786
3787	/*********************** End ../ext/misc/shathree.c ******************/




























































































































































































































































































































































































































3788	/*********************** Begin ../ext/misc/uint.c ****************/
3789	/*
3790	** 2020-04-14
3791	**
3792	** The author disclaims copyright to this source code. In place of
	@@ -5075,11 +4913,11 @@
5075	}else if( !p->bSorted \|\| y>=p->a[p->nUsed-1] ){
5076	p->a[p->nUsed++] = y;
5077	}else if( p->bKeepSorted ){
5078	int i;
5079	i = percentBinarySearch(p, y, 0);
5080	if( i<p->nUsed ){
5081	memmove(&p->a[i+1], &p->a[i], (p->nUsed-i)*sizeof(p->a[0]));
5082	}
5083	p->a[i] = y;
5084	p->nUsed++;
5085	}else{
	@@ -5191,11 +5029,11 @@
5191
5192	/* Find and remove the row */
5193	i = percentBinarySearch(p, y, 1);
5194	if( i>=0 ){
5195	p->nUsed--;
5196	if( i<p->nUsed ){
5197	memmove(&p->a[i], &p->a[i+1], (p->nUsed - i)*sizeof(p->a[0]));
5198	}
5199	}
5200	}
5201
	@@ -5251,15 +5089,15 @@
5251	sqlite3 *db,
5252	char **pzErrMsg,
5253	const sqlite3_api_routines *pApi
5254	){
5255	int rc = SQLITE_OK;
5256	int i;
5257	#if defined(SQLITE3_H) \|\| defined(SQLITE_STATIC_PERCENTILE)
5258	(void)pApi; /* Unused parameter */
5259	#else
5260	SQLITE_EXTENSION_INIT2(pApi);


5261	#endif
5262	(void)pzErrMsg; /* Unused parameter */
5263	for(i=0; i<sizeof(aPercentFunc)/sizeof(aPercentFunc[0]); i++){
5264	rc = sqlite3_create_window_function(db,
5265	aPercentFunc[i].zName,
	@@ -7012,50 +6850,58 @@
7012	aIdx[4] = i;
7013	idxNum \|= 0x40;
7014	}
7015	continue;
7016	}
7017	if( pConstraint->iColumn==SERIES_COLUMN_VALUE ){
7018	switch( op ){
7019	case SQLITE_INDEX_CONSTRAINT_EQ:
7020	case SQLITE_INDEX_CONSTRAINT_IS: {
7021	idxNum \|= 0x0080;
7022	idxNum &= ~0x3300;
7023	aIdx[5] = i;
7024	aIdx[6] = -1;
7025	bStartSeen = 1;
7026	break;
7027	}
7028	case SQLITE_INDEX_CONSTRAINT_GE: {
7029	if( idxNum & 0x0080 ) break;
7030	idxNum \|= 0x0100;
7031	idxNum &= ~0x0200;
7032	aIdx[5] = i;
7033	bStartSeen = 1;
7034	break;
7035	}
7036	case SQLITE_INDEX_CONSTRAINT_GT: {
7037	if( idxNum & 0x0080 ) break;
7038	idxNum \|= 0x0200;
7039	idxNum &= ~0x0100;
7040	aIdx[5] = i;
7041	bStartSeen = 1;
7042	break;
7043	}
7044	case SQLITE_INDEX_CONSTRAINT_LE: {
7045	if( idxNum & 0x0080 ) break;
7046	idxNum \|= 0x1000;
7047	idxNum &= ~0x2000;
7048	aIdx[6] = i;
7049	break;
7050	}
7051	case SQLITE_INDEX_CONSTRAINT_LT: {
7052	if( idxNum & 0x0080 ) break;
7053	idxNum \|= 0x2000;
7054	idxNum &= ~0x1000;
7055	aIdx[6] = i;
7056	break;








7057	}
7058	}
7059	continue;
7060	}
7061	iCol = pConstraint->iColumn - SERIES_COLUMN_START;
	@@ -8115,11 +7961,11 @@
8115	** If the optional MTIME argument is present, then it is interpreted
8116	** as an integer - the number of seconds since the unix epoch. The
8117	** modification-time of the target file is set to this value before
8118	** returning.
8119	**
8120	** If three or more arguments are passed to this function and an
8121	** error is encountered, an exception is raised.
8122	**
8123	** READFILE(FILE):
8124	**
8125	** Read and return the contents of file FILE (type blob) from disk.
	@@ -8185,10 +8031,17 @@
8185	# define lstat(path,buf) stat(path,buf)
8186	#endif
8187	#include <time.h>
8188	#include <errno.h>
8189







8190
8191	/*
8192	** Structure of the fsdir() table-valued function
8193	*/
8194	/* 0 1 2 3 4 5 */
	@@ -8217,11 +8070,11 @@
8217	sqlite3_int64 nIn;
8218	void *pBuf;
8219	sqlite3 *db;
8220	int mxBlob;
8221
8222	in = fopen(zName, "rb");
8223	if( in==0 ){
8224	/* File does not exist or is unreadable. Leave the result set to NULL. */
8225	return;
8226	}
8227	fseek(in, 0, SEEK_END);
	@@ -8472,11 +8325,11 @@
8472	}
8473	}else{
8474	sqlite3_int64 nWrite = 0;
8475	const char *z;
8476	int rc = 0;
8477	FILE *out = fopen(zFile, "wb");
8478	if( out==0 ) return 1;
8479	z = (const char*)sqlite3_value_blob(pData);
8480	if( z ){
8481	sqlite3_int64 n = fwrite(z, 1, sqlite3_value_bytes(pData), out);
8482	nWrite = sqlite3_value_bytes(pData);
	@@ -10333,10 +10186,18 @@
10333	#ifndef SQLITE_NO_STDINT
10334	# include <stdint.h>
10335	#endif
10336
10337	#include <zlib.h>








10338
10339	#ifndef SQLITE_OMIT_VIRTUALTABLE
10340
10341	#ifndef SQLITE_AMALGAMATION
10342
	@@ -11590,11 +11451,11 @@
11590	}else{
11591	zFile = (const char*)sqlite3_value_text(argv[0]);
11592	}
11593
11594	if( 0==pTab->pWriteFd && 0==bInMemory ){
11595	pCsr->pFile = zFile ? fopen(zFile, "rb") : 0;
11596	if( pCsr->pFile==0 ){
11597	zipfileCursorErr(pCsr, "cannot open file: %s", zFile);
11598	rc = SQLITE_ERROR;
11599	}else{
11600	rc = zipfileReadEOCD(pTab, 0, 0, pCsr->pFile, &pCsr->eocd);
	@@ -11780,11 +11641,11 @@
11780
11781	/* Open a write fd on the file. Also load the entire central directory
11782	** structure into memory. During the transaction any new file data is
11783	** appended to the archive file, but the central directory is accumulated
11784	** in main-memory until the transaction is committed. */
11785	pTab->pWriteFd = fopen(pTab->zFile, "ab+");
11786	if( pTab->pWriteFd==0 ){
11787	pTab->base.zErrMsg = sqlite3_mprintf(
11788	"zipfile: failed to open file %s for writing", pTab->zFile
11789	);
11790	rc = SQLITE_ERROR;
	@@ -14233,10 +14094,70 @@
14233	}
14234
14235	return rc;
14236	}
14237




























































14238
14239	static int idxCreateVtabSchema(sqlite3expert p, char *pzErrmsg){
14240	int rc = idxRegisterVtab(p);
14241	sqlite3_stmt *pSchema = 0;
14242
	@@ -14244,30 +14165,39 @@
14244	**
14245	** 1) Add an entry to the p->pTable list, and
14246	** 2) Create the equivalent virtual table in dbv.
14247	*/
14248	rc = idxPrepareStmt(p->db, &pSchema, pzErrmsg,
14249	"SELECT type, name, sql, 1 FROM sqlite_schema "
14250	"WHERE type IN ('table','view') AND name NOT LIKE 'sqlite_%%' "



14251	" UNION ALL "
14252	"SELECT type, name, sql, 2 FROM sqlite_schema "
14253	"WHERE type = 'trigger'"
14254	" AND tbl_name IN(SELECT name FROM sqlite_schema WHERE type = 'view') "
14255	"ORDER BY 4, 1"
14256	);
14257	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSchema) ){
14258	const char zType = (const char)sqlite3_column_text(pSchema, 0);
14259	const char zName = (const char)sqlite3_column_text(pSchema, 1);
14260	const char zSql = (const char)sqlite3_column_text(pSchema, 2);


14261
14262	if( zType==0 \|\| zName==0 ) continue;
14263	if( zType[0]=='v' \|\| zType[1]=='r' ){
14264	if( zSql ) rc = sqlite3_exec(p->dbv, zSql, 0, 0, pzErrmsg);




14265	}else{
14266	IdxTable *pTab;
14267	rc = idxGetTableInfo(p->db, zName, &pTab, pzErrmsg);
14268	if( rc==SQLITE_OK ){
14269	int i;
14270	char *zInner = 0;
14271	char *zOuter = 0;
14272	pTab->pNext = p->pTable;
14273	p->pTable = pTab;
	@@ -14464,10 +14394,16 @@
14464	sqlite3_bind_text(pIndexXInfo, 1, zIdx, -1, SQLITE_STATIC);
14465	while( SQLITE_OK==rc && SQLITE_ROW==sqlite3_step(pIndexXInfo) ){
14466	const char *zComma = zCols==0 ? "" : ", ";
14467	const char zName = (const char)sqlite3_column_text(pIndexXInfo, 0);
14468	const char zColl = (const char)sqlite3_column_text(pIndexXInfo, 1);






14469	zCols = idxAppendText(&rc, zCols,
14470	"%sx.%Q IS sqlite_expert_rem(%d, x.%Q) COLLATE %s",
14471	zComma, zName, nCol, zName, zColl
14472	);
14473	zOrder = idxAppendText(&rc, zOrder, "%s%d", zComma, ++nCol);
	@@ -14792,16 +14728,22 @@
14792
14793	/* Copy the entire schema of database [db] into [dbm]. */
14794	if( rc==SQLITE_OK ){
14795	sqlite3_stmt *pSql = 0;
14796	rc = idxPrintfPrepareStmt(pNew->db, &pSql, pzErrmsg,
14797	"SELECT sql FROM sqlite_schema WHERE name NOT LIKE 'sqlite_%%'"
14798	" AND sql NOT LIKE 'CREATE VIRTUAL %%' ORDER BY rowid"

14799	);
14800	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
14801	const char zSql = (const char)sqlite3_column_text(pSql, 0);
14802	if( zSql ) rc = sqlite3_exec(pNew->dbm, zSql, 0, 0, pzErrmsg);





14803	}
14804	idxFinalize(&rc, pSql);
14805	}
14806
14807	/* Create the vtab schema */
	@@ -16211,10 +16153,1197 @@
16211	}
16212	return rc;
16213	}
16214
16215	/*********************** End ../ext/misc/stmtrand.c ******************/



































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































16216
16217	#if !defined(SQLITE_OMIT_VIRTUALTABLE) && defined(SQLITE_ENABLE_DBPAGE_VTAB)
16218	#define SQLITE_SHELL_HAVE_RECOVER 1
16219	#else
16220	#define SQLITE_SHELL_HAVE_RECOVER 0
	@@ -18241,11 +19370,11 @@
18241	}while( strstr(z,zBuf)!=0 );
18242	return zBuf;
18243	}
18244
18245	/*
18246	** Implementation of scalar SQL function "escape_crnl". The argument passed to
18247	** this function is the output of built-in function quote(). If the first
18248	** character of the input is "'", indicating that the value passed to quote()
18249	** was a text value, then this function searches the input for "\n" and "\r"
18250	** characters and adds a wrapper similar to the following:
18251	**
	@@ -18252,11 +19381,11 @@
18252	** replace(replace(<input>, '\n', char(10), '\r', char(13));
18253	**
18254	** Or, if the first character of the input is not "'", then a copy of the input
18255	** is returned.
18256	*/
18257	static void recoverEscapeCrnl(
18258	sqlite3_context *context,
18259	int argc,
18260	sqlite3_value **argv
18261	){
18262	const char zText = (const char)sqlite3_value_text(argv[0]);
	@@ -18467,11 +19596,11 @@
18467	void (xFunc)(sqlite3_context,int,sqlite3_value **);
18468	} aFunc[] = {
18469	{ "getpage", 1, recoverGetPage },
18470	{ "page_is_used", 1, recoverPageIsUsed },
18471	{ "read_i32", 2, recoverReadI32 },
18472	{ "escape_crnl", 1, recoverEscapeCrnl },
18473	};
18474
18475	const int flags = SQLITE_OPEN_URI\|SQLITE_OPEN_CREATE\|SQLITE_OPEN_READWRITE;
18476	sqlite3 db = 0; / New database handle */
18477	int ii; /* For iterating through aFunc[] */
	@@ -18820,11 +19949,11 @@
18820	assert( pTab->aCol[ii].iField>=0 && pTab->aCol[ii].iBind>=1 );
18821	zSql = recoverMPrintf(p, "%z%s%Q", zSql, zSep, pTab->aCol[ii].zCol);
18822
18823	if( bSql ){
18824	zBind = recoverMPrintf(p,
18825	"%z%sescape_crnl(quote(?%d))", zBind, zSqlSep, pTab->aCol[ii].iBind
18826	);
18827	zSqlSep = "\|\|', '\|\|";
18828	}else{
18829	zBind = recoverMPrintf(p, "%z%s?%d", zBind, zSep, pTab->aCol[ii].iBind);
18830	}
	@@ -19322,10 +20451,12 @@
19322	apVal[iField] = sqlite3_value_dup( pVal );
19323	if( apVal[iField]==0 ){
19324	recoverError(p, SQLITE_NOMEM, 0);
19325	}
19326	p1->nVal = iField+1;


19327	}
19328	p1->iPrevCell = iCell;
19329	p1->iPrevPage = iPage;
19330	}
19331	}else{
	@@ -20437,10 +21568,11 @@
20437	u8 nEqpLevel; /* Depth of the EQP output graph */
20438	u8 eTraceType; /* SHELL_TRACE_* value for type of trace */
20439	u8 bSafeMode; /* True to prohibit unsafe operations */
20440	u8 bSafeModePersist; /* The long-term value of bSafeMode */
20441	u8 eRestoreState; /* See comments above doAutoDetectRestore() */

20442	ColModeOpts cmOpts; /* Option values affecting columnar mode output */
20443	unsigned statsOn; /* True to display memory stats before each finalize */
20444	unsigned mEqpLines; /* Mask of vertical lines in the EQP output graph */
20445	int inputNesting; /* Track nesting level of .read and other redirects */
20446	int outCount; /* Revert to stdout when reaching zero */
	@@ -20582,10 +21714,11 @@
20582	#define MODE_Table 15 /* MySQL-style table formatting */
20583	#define MODE_Box 16 /* Unicode box-drawing characters */
20584	#define MODE_Count 17 /* Output only a count of the rows of output */
20585	#define MODE_Off 18 /* No query output shown */
20586	#define MODE_ScanExp 19 /* Like MODE_Explain, but for ".scanstats vm" */

20587
20588	static const char *modeDescr[] = {
20589	"line",
20590	"column",
20591	"list",
	@@ -20602,11 +21735,13 @@
20602	"json",
20603	"markdown",
20604	"table",
20605	"box",
20606	"count",
20607	"off"


20608	};
20609
20610	/*
20611	** These are the column/row/line separators used by the various
20612	** import/export modes.
	@@ -20630,11 +21765,11 @@
20630	** A callback for the sqlite3_log() interface.
20631	*/
20632	static void shellLog(void pArg, int iErrCode, const char zMsg){
20633	ShellState p = (ShellState)pArg;
20634	if( p->pLog==0 ) return;
20635	sputf(p->pLog, "(%d) %s\n", iErrCode, zMsg);
20636	fflush(p->pLog);
20637	}
20638
20639	/*
20640	** SQL function: shell_putsnl(X)
	@@ -20645,13 +21780,13 @@
20645	static void shellPutsFunc(
20646	sqlite3_context *pCtx,
20647	int nVal,
20648	sqlite3_value **apVal
20649	){
20650	/* Unused: (ShellState)sqlite3_user_data(pCtx); /
20651	(void)nVal;
20652	oputf("%s\n", sqlite3_value_text(apVal[0]));
20653	sqlite3_result_value(pCtx, apVal[0]);
20654	}
20655
20656	/*
20657	** If in safe mode, print an error message described by the arguments
	@@ -20666,11 +21801,11 @@
20666	va_list ap;
20667	char *zMsg;
20668	va_start(ap, zErrMsg);
20669	zMsg = sqlite3_vmprintf(zErrMsg, ap);
20670	va_end(ap);
20671	eputf("line %d: %s\n", p->lineno, zMsg);
20672	exit(1);
20673	}
20674	}
20675
20676	/*
	@@ -20699,11 +21834,11 @@
20699	char *zTempFile = 0;
20700	sqlite3 *db;
20701	char *zCmd = 0;
20702	int bBin;
20703	int rc;
20704	int hasCRNL = 0;
20705	FILE *f = 0;
20706	sqlite3_int64 sz;
20707	sqlite3_int64 x;
20708	unsigned char *p = 0;
20709
	@@ -20733,11 +21868,11 @@
20733	}
20734	}
20735	bBin = sqlite3_value_type(argv[0])==SQLITE_BLOB;
20736	/* When writing the file to be edited, do \n to \r\n conversions on systems
20737	** that want \r\n line endings */
20738	f = fopen(zTempFile, bBin ? "wb" : "w");
20739	if( f==0 ){
20740	sqlite3_result_error(context, "edit() cannot open temp file", -1);
20741	goto edit_func_end;
20742	}
20743	sz = sqlite3_value_bytes(argv[0]);
	@@ -20744,11 +21879,11 @@
20744	if( bBin ){
20745	x = fwrite(sqlite3_value_blob(argv[0]), 1, (size_t)sz, f);
20746	}else{
20747	const char z = (const char)sqlite3_value_text(argv[0]);
20748	/* Remember whether or not the value originally contained \r\n */
20749	if( z && strstr(z,"\r\n")!=0 ) hasCRNL = 1;
20750	x = fwrite(sqlite3_value_text(argv[0]), 1, (size_t)sz, f);
20751	}
20752	fclose(f);
20753	f = 0;
20754	if( x!=sz ){
	@@ -20764,11 +21899,11 @@
20764	sqlite3_free(zCmd);
20765	if( rc ){
20766	sqlite3_result_error(context, "EDITOR returned non-zero", -1);
20767	goto edit_func_end;
20768	}
20769	f = fopen(zTempFile, "rb");
20770	if( f==0 ){
20771	sqlite3_result_error(context,
20772	"edit() cannot reopen temp file after edit", -1);
20773	goto edit_func_end;
20774	}
	@@ -20789,11 +21924,11 @@
20789	}
20790	if( bBin ){
20791	sqlite3_result_blob64(context, p, sz, sqlite3_free);
20792	}else{
20793	sqlite3_int64 i, j;
20794	if( hasCRNL ){
20795	/* If the original contains \r\n then do no conversions back to \n */
20796	}else{
20797	/* If the file did not originally contain \r\n then convert any new
20798	** \r\n back into \n */
20799	p[sz] = 0;
	@@ -20830,15 +21965,30 @@
20830	p->mode = p->modePrior;
20831	p->shellFlgs = p->priorShFlgs;
20832	memcpy(p->colSeparator, p->colSepPrior, sizeof(p->colSeparator));
20833	memcpy(p->rowSeparator, p->rowSepPrior, sizeof(p->rowSeparator));
20834	}















20835
20836	/*
20837	** Output the given string as a hex-encoded blob (eg. X'1234' )
20838	*/
20839	static void output_hex_blob(const void *pBlob, int nBlob){
20840	int i;
20841	unsigned char aBlob = (unsigned char)pBlob;
20842
20843	char zStr = sqlite3_malloc(nBlob2 + 1);
20844	shell_check_oom(zStr);
	@@ -20851,11 +22001,11 @@
20851	zStr[i*2] = aHex[ (aBlob[i] >> 4) ];
20852	zStr[i*2+1] = aHex[ (aBlob[i] & 0x0F) ];
20853	}
20854	zStr[i*2] = '\0';
20855
20856	oputf("X'%s'", zStr);
20857	sqlite3_free(zStr);
20858	}
20859
20860	/*
20861	** Find a string that is not found anywhere in z[]. Return a pointer
	@@ -20881,46 +22031,40 @@
20881	/*
20882	** Output the given string as a quoted string using SQL quoting conventions.
20883	**
20884	** See also: output_quoted_escaped_string()
20885	*/
20886	static void output_quoted_string(const char *z){
20887	int i;
20888	char c;
20889	#ifndef SQLITE_SHELL_FIDDLE
20890	FILE *pfO = setOutputStream(invalidFileStream);
20891	setBinaryMode(pfO, 1);
20892	#endif
20893	if( z==0 ) return;
20894	for(i=0; (c = z[i])!=0 && c!='\''; i++){}
20895	if( c==0 ){
20896	oputf("'%s'",z);
20897	}else{
20898	oputz("'");
20899	while( *z ){
20900	for(i=0; (c = z[i])!=0 && c!='\''; i++){}
20901	if( c=='\'' ) i++;
20902	if( i ){
20903	oputf("%.*s", i, z);
20904	z += i;
20905	}
20906	if( c=='\'' ){
20907	oputz("'");
20908	continue;
20909	}
20910	if( c==0 ){
20911	break;
20912	}
20913	z++;
20914	}
20915	oputz("'");
20916	}
20917	#ifndef SQLITE_SHELL_FIDDLE
20918	setTextMode(pfO, 1);
20919	#else
20920	setTextMode(stdout, 1);
20921	#endif
20922	}
20923
20924	/*
20925	** Output the given string as a quoted string using SQL quoting conventions.
20926	** Additionallly , escape the "\n" and "\r" characters so that they do not
	@@ -20928,20 +22072,18 @@
20928	** systems.
20929	**
20930	** This is like output_quoted_string() but with the addition of the \r\n
20931	** escape mechanism.
20932	*/
20933	static void output_quoted_escaped_string(const char *z){
20934	int i;
20935	char c;
20936	#ifndef SQLITE_SHELL_FIDDLE
20937	FILE *pfO = setOutputStream(invalidFileStream);
20938	setBinaryMode(pfO, 1);
20939	#endif
20940	for(i=0; (c = z[i])!=0 && c!='\'' && c!='\n' && c!='\r'; i++){}
20941	if( c==0 ){
20942	oputf("'%s'",z);
20943	}else{
20944	const char *zNL = 0;
20945	const char *zCR = 0;
20946	int nNL = 0;
20947	int nCR = 0;
	@@ -20949,52 +22091,48 @@
20949	for(i=0; z[i]; i++){
20950	if( z[i]=='\n' ) nNL++;
20951	if( z[i]=='\r' ) nCR++;
20952	}
20953	if( nNL ){
20954	oputz("replace(");
20955	zNL = unused_string(z, "\\n", "\\012", zBuf1);
20956	}
20957	if( nCR ){
20958	oputz("replace(");
20959	zCR = unused_string(z, "\\r", "\\015", zBuf2);
20960	}
20961	oputz("'");
20962	while( *z ){
20963	for(i=0; (c = z[i])!=0 && c!='\n' && c!='\r' && c!='\''; i++){}
20964	if( c=='\'' ) i++;
20965	if( i ){
20966	oputf("%.*s", i, z);
20967	z += i;
20968	}
20969	if( c=='\'' ){
20970	oputz("'");
20971	continue;
20972	}
20973	if( c==0 ){
20974	break;
20975	}
20976	z++;
20977	if( c=='\n' ){
20978	oputz(zNL);
20979	continue;
20980	}
20981	oputz(zCR);
20982	}
20983	oputz("'");
20984	if( nCR ){
20985	oputf(",'%s',char(13))", zCR);
20986	}
20987	if( nNL ){
20988	oputf(",'%s',char(10))", zNL);
20989	}
20990	}
20991	#ifndef SQLITE_SHELL_FIDDLE
20992	setTextMode(pfO, 1);
20993	#else
20994	setTextMode(stdout, 1);
20995	#endif
20996	}
20997
20998	/*
20999	** Find earliest of chars within s specified in zAny.
21000	** With ns == ~0, is like strpbrk(s,zAny) and s must be 0-terminated.
	@@ -21010,26 +22148,64 @@
21010	}
21011	++zAny;
21012	}
21013	return pcFirst;
21014	}




































21015	/*
21016	** Output the given string as a quoted according to C or TCL quoting rules.
21017	*/
21018	static void output_c_string(const char *z){
21019	char c;
21020	static const char *zq = "\"";
21021	static long ctrlMask = ~0L;
21022	static const char zDQBSRO = "\"\\\x7f"; / double-quote, backslash, rubout */
21023	char ace[3] = "\\?";
21024	char cbsSay;
21025	oputz(zq);
21026	while( *z!=0 ){
21027	const char *pcDQBSRO = anyOfInStr(z, zDQBSRO, ~(size_t)0);
21028	const char *pcPast = zSkipValidUtf8(z, INT_MAX, ctrlMask);
21029	const char *pcEnd = (pcDQBSRO && pcDQBSRO < pcPast)? pcDQBSRO : pcPast;
21030	if( pcEnd > z ) oputb(z, (int)(pcEnd-z));


21031	if( (c = *pcEnd)==0 ) break;
21032	++pcEnd;
21033	switch( c ){
21034	case '\\': case '"':
21035	cbsSay = (char)c;
	@@ -21040,47 +22216,47 @@
21040	case '\f': cbsSay = 'f'; break;
21041	default: cbsSay = 0; break;
21042	}
21043	if( cbsSay ){
21044	ace[1] = cbsSay;
21045	oputz(ace);
21046	}else if( !isprint(c&0xff) ){
21047	oputf("\\%03o", c&0xff);
21048	}else{
21049	ace[1] = (char)c;
21050	oputz(ace+1);
21051	}
21052	z = pcEnd;
21053	}
21054	oputz(zq);
21055	}
21056
21057	/*
21058	** Output the given string as a quoted according to JSON quoting rules.
21059	*/
21060	static void output_json_string(const char *z, i64 n){
21061	char c;
21062	static const char *zq = "\"";
21063	static long ctrlMask = ~0L;
21064	static const char *zDQBS = "\"\\";
21065	const char *pcLimit;
21066	char ace[3] = "\\?";
21067	char cbsSay;
21068
21069	if( z==0 ) z = "";
21070	pcLimit = z + ((n<0)? strlen(z) : (size_t)n);
21071	oputz(zq);
21072	while( z < pcLimit ){
21073	const char *pcDQBS = anyOfInStr(z, zDQBS, pcLimit-z);
21074	const char *pcPast = zSkipValidUtf8(z, (int)(pcLimit-z), ctrlMask);
21075	const char *pcEnd = (pcDQBS && pcDQBS < pcPast)? pcDQBS : pcPast;
21076	if( pcEnd > z ){
21077	oputb(z, (int)(pcEnd-z));
21078	z = pcEnd;
21079	}
21080	if( z >= pcLimit ) break;
21081	c = *(z++);
21082	switch( c ){
21083	case '"': case '\\':
21084	cbsSay = (char)c;
21085	break;
21086	case '\b': cbsSay = 'b'; break;
	@@ -21090,26 +22266,26 @@
21090	case '\t': cbsSay = 't'; break;
21091	default: cbsSay = 0; break;
21092	}
21093	if( cbsSay ){
21094	ace[1] = cbsSay;
21095	oputz(ace);
21096	}else if( c<=0x1f ){
21097	oputf("u%04x", c);
21098	}else{
21099	ace[1] = (char)c;
21100	oputz(ace+1);
21101	}
21102	}
21103	oputz(zq);
21104	}
21105
21106	/*
21107	** Output the given string with characters that are special to
21108	** HTML escaped.
21109	*/
21110	static void output_html_string(const char *z){
21111	int i;
21112	if( z==0 ) z = "";
21113	while( *z ){
21114	for(i=0; z[i]
21115	&& z[i]!='<'
	@@ -21117,22 +22293,22 @@
21117	&& z[i]!='>'
21118	&& z[i]!='\"'
21119	&& z[i]!='\'';
21120	i++){}
21121	if( i>0 ){
21122	oputf("%.*s",i,z);
21123	}
21124	if( z[i]=='<' ){
21125	oputz("<");
21126	}else if( z[i]=='&' ){
21127	oputz("&");
21128	}else if( z[i]=='>' ){
21129	oputz(">");
21130	}else if( z[i]=='\"' ){
21131	oputz(""");
21132	}else if( z[i]=='\'' ){
21133	oputz("'");
21134	}else{
21135	break;
21136	}
21137	z += i + 1;
21138	}
	@@ -21167,11 +22343,11 @@
21167	** the null value. Strings are quoted if necessary. The separator
21168	** is only issued if bSep is true.
21169	*/
21170	static void output_csv(ShellState p, const char z, int bSep){
21171	if( z==0 ){
21172	oputf("%s",p->nullValue);
21173	}else{
21174	unsigned i;
21175	for(i=0; z[i]; i++){
21176	if( needCsvQuote[((unsigned char*)z)[i]] ){
21177	i = 0;
	@@ -21179,18 +22355,18 @@
21179	}
21180	}
21181	if( i==0 \|\| strstr(z, p->colSeparator)!=0 ){
21182	char *zQuoted = sqlite3_mprintf("\"%w\"", z);
21183	shell_check_oom(zQuoted);
21184	oputz(zQuoted);
21185	sqlite3_free(zQuoted);
21186	}else{
21187	oputz(z);
21188	}
21189	}
21190	if( bSep ){
21191	oputz(p->colSeparator);
21192	}
21193	}
21194
21195	/*
21196	** This routine runs when the user presses Ctrl-C
	@@ -21294,20 +22470,20 @@
21294	const char *az[4];
21295	az[0] = zA1;
21296	az[1] = zA2;
21297	az[2] = zA3;
21298	az[3] = zA4;
21299	oputf("authorizer: %s", azAction[op]);
21300	for(i=0; i<4; i++){
21301	oputz(" ");
21302	if( az[i] ){
21303	output_c_string(az[i]);
21304	}else{
21305	oputz("NULL");
21306	}
21307	}
21308	oputz("\n");
21309	if( p->bSafeMode ) (void)safeModeAuth(pClientData, op, zA1, zA2, zA3, zA4);
21310	return SQLITE_OK;
21311	}
21312	#endif
21313
	@@ -21319,11 +22495,11 @@
21319	**
21320	** If the schema statement in z[] contains a start-of-comment and if
21321	** sqlite3_complete() returns false, try to terminate the comment before
21322	** printing the result. https://sqlite.org/forum/forumpost/d7be961c5c
21323	*/
21324	static void printSchemaLine(const char z, const char zTail){
21325	char *zToFree = 0;
21326	if( z==0 ) return;
21327	if( zTail==0 ) return;
21328	if( zTail[0]==';' && (strstr(z, "/*")!=0 \|\| strstr(z,"--")!=0) ){
21329	const char *zOrig = z;
	@@ -21341,20 +22517,20 @@
21341	}
21342	sqlite3_free(zNew);
21343	}
21344	}
21345	if( sqlite3_strglob("CREATE TABLE ['\"]*", z)==0 ){
21346	oputf("CREATE TABLE IF NOT EXISTS %s%s", z+13, zTail);
21347	}else{
21348	oputf("%s%s", z, zTail);
21349	}
21350	sqlite3_free(zToFree);
21351	}
21352	static void printSchemaLineN(char z, int n, const char zTail){
21353	char c = z[n];
21354	z[n] = 0;
21355	printSchemaLine(z, zTail);
21356	z[n] = c;
21357	}
21358
21359	/*
21360	** Return true if string z[] has nothing but whitespace and comments to the
	@@ -21378,11 +22554,11 @@
21378	EQPGraphRow *pNew;
21379	i64 nText;
21380	if( zText==0 ) return;
21381	nText = strlen(zText);
21382	if( p->autoEQPtest ){
21383	oputf("%d,%d,%s\n", iEqpId, p2, zText);
21384	}
21385	pNew = sqlite3_malloc64( sizeof(*pNew) + nText );
21386	shell_check_oom(pNew);
21387	pNew->iEqpId = iEqpId;
21388	pNew->iParentId = p2;
	@@ -21426,11 +22602,12 @@
21426	i64 n = strlen(p->sGraph.zPrefix);
21427	char *z;
21428	for(pRow = eqp_next_row(p, iEqpId, 0); pRow; pRow = pNext){
21429	pNext = eqp_next_row(p, iEqpId, pRow);
21430	z = pRow->zText;
21431	oputf("%s%s%s\n", p->sGraph.zPrefix, pNext ? "\|--" : "`--", z);

21432	if( n<(i64)sizeof(p->sGraph.zPrefix)-7 ){
21433	memcpy(&p->sGraph.zPrefix[n], pNext ? "\| " : " ", 4);
21434	eqp_render_level(p, pRow->iEqpId);
21435	p->sGraph.zPrefix[n] = 0;
21436	}
	@@ -21446,17 +22623,17 @@
21446	if( pRow->zText[0]=='-' ){
21447	if( pRow->pNext==0 ){
21448	eqp_reset(p);
21449	return;
21450	}
21451	oputf("%s\n", pRow->zText+3);
21452	p->sGraph.pRow = pRow->pNext;
21453	sqlite3_free(pRow);
21454	}else if( nCycle>0 ){
21455	oputf("QUERY PLAN (cycles=%lld [100%%])\n", nCycle);
21456	}else{
21457	oputz("QUERY PLAN\n");
21458	}
21459	p->sGraph.zPrefix[0] = 0;
21460	eqp_render_level(p, 0);
21461	eqp_reset(p);
21462	}
	@@ -21468,33 +22645,33 @@
21468	*/
21469	static int progress_handler(void *pClientData) {
21470	ShellState p = (ShellState)pClientData;
21471	p->nProgress++;
21472	if( p->nProgress>=p->mxProgress && p->mxProgress>0 ){
21473	oputf("Progress limit reached (%u)\n", p->nProgress);
21474	if( p->flgProgress & SHELL_PROGRESS_RESET ) p->nProgress = 0;
21475	if( p->flgProgress & SHELL_PROGRESS_ONCE ) p->mxProgress = 0;
21476	return 1;
21477	}
21478	if( (p->flgProgress & SHELL_PROGRESS_QUIET)==0 ){
21479	oputf("Progress %u\n", p->nProgress);
21480	}
21481	return 0;
21482	}
21483	#endif /* SQLITE_OMIT_PROGRESS_CALLBACK */
21484
21485	/*
21486	** Print N dashes
21487	*/
21488	static void print_dashes(int N){
21489	const char zDash[] = "--------------------------------------------------";
21490	const int nDash = sizeof(zDash) - 1;
21491	while( N>nDash ){
21492	oputz(zDash);
21493	N -= nDash;
21494	}
21495	oputf("%.*s", N, zDash);
21496	}
21497
21498	/*
21499	** Print a markdown or table-style row separator using ascii-art
21500	*/
	@@ -21503,19 +22680,19 @@
21503	int nArg,
21504	const char *zSep
21505	){
21506	int i;
21507	if( nArg>0 ){
21508	oputz(zSep);
21509	print_dashes(p->actualWidth[0]+2);
21510	for(i=1; i<nArg; i++){
21511	oputz(zSep);
21512	print_dashes(p->actualWidth[i]+2);
21513	}
21514	oputz(zSep);
21515	}
21516	oputz("\n");
21517	}
21518
21519	/*
21520	** This is the callback routine that the shell
21521	** invokes for each row of a query result.
	@@ -21541,13 +22718,13 @@
21541	if( azArg==0 ) break;
21542	for(i=0; i<nArg; i++){
21543	int len = strlen30(azCol[i] ? azCol[i] : "");
21544	if( len>w ) w = len;
21545	}
21546	if( p->cnt++>0 ) oputz(p->rowSeparator);
21547	for(i=0; i<nArg; i++){
21548	oputf("%*s = %s%s", w, azCol[i],
21549	azArg[i] ? azArg[i] : p->nullValue, p->rowSeparator);
21550	}
21551	break;
21552	}
21553	case MODE_ScanExp:
	@@ -21571,16 +22748,16 @@
21571	if( nArg>nWidth ) nArg = nWidth;
21572
21573	/* If this is the first row seen, print out the headers */
21574	if( p->cnt++==0 ){
21575	for(i=0; i<nArg; i++){
21576	utf8_width_print(aWidth[i], azCol[ aMap[i] ]);
21577	oputz(i==nArg-1 ? "\n" : " ");
21578	}
21579	for(i=0; i<nArg; i++){
21580	print_dashes(aWidth[i]);
21581	oputz(i==nArg-1 ? "\n" : " ");
21582	}
21583	}
21584
21585	/* If there is no data, exit early. */
21586	if( azArg==0 ) break;
	@@ -21594,21 +22771,21 @@
21594	w = strlenChar(zVal);
21595	zSep = " ";
21596	}
21597	if( i==iIndent && p->aiIndent && p->pStmt ){
21598	if( p->iIndent<p->nIndent ){
21599	oputf("%*.s", p->aiIndent[p->iIndent], "");
21600	}
21601	p->iIndent++;
21602	}
21603	utf8_width_print(w, zVal ? zVal : p->nullValue);
21604	oputz(i==nArg-1 ? "\n" : zSep);
21605	}
21606	break;
21607	}
21608	case MODE_Semi: { /* .schema and .fullschema output */
21609	printSchemaLine(azArg[0], ";\n");
21610	break;
21611	}
21612	case MODE_Pretty: { /* .schema and .fullschema with --indent */
21613	char *z;
21614	int j;
	@@ -21619,11 +22796,11 @@
21619	assert( nArg==1 );
21620	if( azArg[0]==0 ) break;
21621	if( sqlite3_strlike("CREATE VIEW%", azArg[0], 0)==0
21622	\|\| sqlite3_strlike("CREATE TRIG%", azArg[0], 0)==0
21623	){
21624	oputf("%s;\n", azArg[0]);
21625	break;
21626	}
21627	z = sqlite3_mprintf("%s", azArg[0]);
21628	shell_check_oom(z);
21629	j = 0;
	@@ -21652,255 +22829,265 @@
21652	}else if( c=='(' ){
21653	nParen++;
21654	}else if( c==')' ){
21655	nParen--;
21656	if( nLine>0 && nParen==0 && j>0 ){
21657	printSchemaLineN(z, j, "\n");
21658	j = 0;
21659	}
21660	}
21661	z[j++] = c;
21662	if( nParen==1 && cEnd==0
21663	&& (c=='(' \|\| c=='\n' \|\| (c==',' && !wsToEol(z+i+1)))
21664	){
21665	if( c=='\n' ) j--;
21666	printSchemaLineN(z, j, "\n ");
21667	j = 0;
21668	nLine++;
21669	while( IsSpace(z[i+1]) ){ i++; }
21670	}
21671	}
21672	z[j] = 0;
21673	}
21674	printSchemaLine(z, ";\n");
21675	sqlite3_free(z);
21676	break;
21677	}
21678	case MODE_List: {
21679	if( p->cnt++==0 && p->showHeader ){
21680	for(i=0; i<nArg; i++){
21681	oputf("%s%s",azCol[i], i==nArg-1 ? p->rowSeparator : p->colSeparator);

21682	}
21683	}
21684	if( azArg==0 ) break;
21685	for(i=0; i<nArg; i++){
21686	char *z = azArg[i];
21687	if( z==0 ) z = p->nullValue;
21688	oputz(z);
21689	oputz((i<nArg-1)? p->colSeparator : p->rowSeparator);
21690	}
21691	break;
21692	}

21693	case MODE_Html: {
21694	if( p->cnt++==0 && p->showHeader ){
21695	oputz("<TR>");







21696	for(i=0; i<nArg; i++){
21697	oputz("<TH>");
21698	output_html_string(azCol[i]);
21699	oputz("</TH>\n");
21700	}
21701	oputz("</TR>\n");
21702	}

21703	if( azArg==0 ) break;
21704	oputz("<TR>");
21705	for(i=0; i<nArg; i++){
21706	oputz("<TD>");
21707	output_html_string(azArg[i] ? azArg[i] : p->nullValue);
21708	oputz("</TD>\n");
21709	}
21710	oputz("</TR>\n");
21711	break;
21712	}
21713	case MODE_Tcl: {
21714	if( p->cnt++==0 && p->showHeader ){
21715	for(i=0; i<nArg; i++){
21716	output_c_string(azCol[i] ? azCol[i] : "");
21717	if(i<nArg-1) oputz(p->colSeparator);
21718	}
21719	oputz(p->rowSeparator);
21720	}
21721	if( azArg==0 ) break;
21722	for(i=0; i<nArg; i++){
21723	output_c_string(azArg[i] ? azArg[i] : p->nullValue);
21724	if(i<nArg-1) oputz(p->colSeparator);
21725	}
21726	oputz(p->rowSeparator);
21727	break;
21728	}
21729	case MODE_Csv: {
21730	setBinaryMode(p->out, 1);
21731	if( p->cnt++==0 && p->showHeader ){
21732	for(i=0; i<nArg; i++){
21733	output_csv(p, azCol[i] ? azCol[i] : "", i<nArg-1);
21734	}
21735	oputz(p->rowSeparator);
21736	}
21737	if( nArg>0 ){
21738	for(i=0; i<nArg; i++){
21739	output_csv(p, azArg[i], i<nArg-1);
21740	}
21741	oputz(p->rowSeparator);
21742	}
21743	setTextMode(p->out, 1);
21744	break;
21745	}
21746	case MODE_Insert: {
21747	if( azArg==0 ) break;
21748	oputf("INSERT INTO %s",p->zDestTable);
21749	if( p->showHeader ){
21750	oputz("(");
21751	for(i=0; i<nArg; i++){
21752	if( i>0 ) oputz(",");
21753	if( quoteChar(azCol[i]) ){
21754	char *z = sqlite3_mprintf("\"%w\"", azCol[i]);
21755	shell_check_oom(z);
21756	oputz(z);
21757	sqlite3_free(z);
21758	}else{
21759	oputf("%s", azCol[i]);
21760	}
21761	}
21762	oputz(")");
21763	}
21764	p->cnt++;
21765	for(i=0; i<nArg; i++){
21766	oputz(i>0 ? "," : " VALUES(");
21767	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
21768	oputz("NULL");
21769	}else if( aiType && aiType[i]==SQLITE_TEXT ){
21770	if( ShellHasFlag(p, SHFLG_Newlines) ){
21771	output_quoted_string(azArg[i]);
21772	}else{
21773	output_quoted_escaped_string(azArg[i]);
21774	}
21775	}else if( aiType && aiType[i]==SQLITE_INTEGER ){
21776	oputz(azArg[i]);
21777	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
21778	char z[50];
21779	double r = sqlite3_column_double(p->pStmt, i);
21780	sqlite3_uint64 ur;
21781	memcpy(&ur,&r,sizeof(r));
21782	if( ur==0x7ff0000000000000LL ){
21783	oputz("9.0e+999");
21784	}else if( ur==0xfff0000000000000LL ){
21785	oputz("-9.0e+999");
21786	}else{
21787	sqlite3_int64 ir = (sqlite3_int64)r;
21788	if( r==(double)ir ){
21789	sqlite3_snprintf(50,z,"%lld.0", ir);
21790	}else{
21791	sqlite3_snprintf(50,z,"%!.20g", r);
21792	}
21793	oputz(z);
21794	}
21795	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
21796	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
21797	int nBlob = sqlite3_column_bytes(p->pStmt, i);
21798	output_hex_blob(pBlob, nBlob);
21799	}else if( isNumber(azArg[i], 0) ){
21800	oputz(azArg[i]);
21801	}else if( ShellHasFlag(p, SHFLG_Newlines) ){
21802	output_quoted_string(azArg[i]);
21803	}else{
21804	output_quoted_escaped_string(azArg[i]);
21805	}
21806	}
21807	oputz(");\n");
21808	break;
21809	}
21810	case MODE_Json: {
21811	if( azArg==0 ) break;
21812	if( p->cnt==0 ){
21813	fputs("[{", p->out);
21814	}else{
21815	fputs(",\n{", p->out);
21816	}
21817	p->cnt++;
21818	for(i=0; i<nArg; i++){
21819	output_json_string(azCol[i], -1);
21820	oputz(":");
21821	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
21822	oputz("null");
21823	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
21824	char z[50];
21825	double r = sqlite3_column_double(p->pStmt, i);
21826	sqlite3_uint64 ur;
21827	memcpy(&ur,&r,sizeof(r));
21828	if( ur==0x7ff0000000000000LL ){
21829	oputz("9.0e+999");
21830	}else if( ur==0xfff0000000000000LL ){
21831	oputz("-9.0e+999");
21832	}else{
21833	sqlite3_snprintf(50,z,"%!.20g", r);
21834	oputz(z);
21835	}
21836	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
21837	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
21838	int nBlob = sqlite3_column_bytes(p->pStmt, i);
21839	output_json_string(pBlob, nBlob);
21840	}else if( aiType && aiType[i]==SQLITE_TEXT ){
21841	output_json_string(azArg[i], -1);
21842	}else{
21843	oputz(azArg[i]);
21844	}
21845	if( i<nArg-1 ){
21846	oputz(",");
21847	}
21848	}
21849	oputz("}");
21850	break;
21851	}
21852	case MODE_Quote: {
21853	if( azArg==0 ) break;
21854	if( p->cnt==0 && p->showHeader ){
21855	for(i=0; i<nArg; i++){
21856	if( i>0 ) fputs(p->colSeparator, p->out);
21857	output_quoted_string(azCol[i]);
21858	}
21859	fputs(p->rowSeparator, p->out);
21860	}
21861	p->cnt++;
21862	for(i=0; i<nArg; i++){
21863	if( i>0 ) fputs(p->colSeparator, p->out);
21864	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
21865	oputz("NULL");
21866	}else if( aiType && aiType[i]==SQLITE_TEXT ){
21867	output_quoted_string(azArg[i]);
21868	}else if( aiType && aiType[i]==SQLITE_INTEGER ){
21869	oputz(azArg[i]);
21870	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
21871	char z[50];
21872	double r = sqlite3_column_double(p->pStmt, i);
21873	sqlite3_snprintf(50,z,"%!.20g", r);
21874	oputz(z);
21875	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
21876	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
21877	int nBlob = sqlite3_column_bytes(p->pStmt, i);
21878	output_hex_blob(pBlob, nBlob);
21879	}else if( isNumber(azArg[i], 0) ){
21880	oputz(azArg[i]);
21881	}else{
21882	output_quoted_string(azArg[i]);
21883	}
21884	}
21885	fputs(p->rowSeparator, p->out);
21886	break;
21887	}
21888	case MODE_Ascii: {
21889	if( p->cnt++==0 && p->showHeader ){
21890	for(i=0; i<nArg; i++){
21891	if( i>0 ) oputz(p->colSeparator);
21892	oputz(azCol[i] ? azCol[i] : "");
21893	}
21894	oputz(p->rowSeparator);
21895	}
21896	if( azArg==0 ) break;
21897	for(i=0; i<nArg; i++){
21898	if( i>0 ) oputz(p->colSeparator);
21899	oputz(azArg[i] ? azArg[i] : p->nullValue);
21900	}
21901	oputz(p->rowSeparator);
21902	break;
21903	}
21904	case MODE_EQP: {
21905	eqp_append(p, atoi(azArg[0]), atoi(azArg[1]), azArg[3]);
21906	break;
	@@ -21975,11 +23162,11 @@
21975	"INSERT INTO selftest(tno,op,cmd,ans)"
21976	" SELECT rowid*10,op,cmd,ans FROM [_shell$self];\n"
21977	"DROP TABLE [_shell$self];"
21978	,0,0,&zErrMsg);
21979	if( zErrMsg ){
21980	eputf("SELFTEST initialization failure: %s\n", zErrMsg);
21981	sqlite3_free(zErrMsg);
21982	}
21983	sqlite3_exec(p->db, "RELEASE selftest_init",0,0,0);
21984	}
21985
	@@ -22078,36 +23265,37 @@
22078	int i;
22079	const char *z;
22080	rc = sqlite3_prepare_v2(p->db, zSelect, -1, &pSelect, 0);
22081	if( rc!=SQLITE_OK \|\| !pSelect ){
22082	char *zContext = shell_error_context(zSelect, p->db);
22083	oputf("/** ERROR: (%d) %s ***/\n%s",
22084	rc, sqlite3_errmsg(p->db), zContext);
22085	sqlite3_free(zContext);
22086	if( (rc&0xff)!=SQLITE_CORRUPT ) p->nErr++;
22087	return rc;
22088	}
22089	rc = sqlite3_step(pSelect);
22090	nResult = sqlite3_column_count(pSelect);
22091	while( rc==SQLITE_ROW ){
22092	z = (const char*)sqlite3_column_text(pSelect, 0);
22093	oputf("%s", z);
22094	for(i=1; i<nResult; i++){
22095	oputf(",%s", sqlite3_column_text(pSelect, i));
22096	}
22097	if( z==0 ) z = "";
22098	while( z[0] && (z[0]!='-' \|\| z[1]!='-') ) z++;
22099	if( z[0] ){
22100	oputz("\n;\n");
22101	}else{
22102	oputz(";\n");
22103	}
22104	rc = sqlite3_step(pSelect);
22105	}
22106	rc = sqlite3_finalize(pSelect);
22107	if( rc!=SQLITE_OK ){
22108	oputf("/** ERROR: (%d) %s ***/\n", rc, sqlite3_errmsg(p->db));

22109	if( (rc&0xff)!=SQLITE_CORRUPT ) p->nErr++;
22110	}
22111	return rc;
22112	}
22113
	@@ -22139,17 +23327,17 @@
22139
22140	#ifdef __linux__
22141	/*
22142	** Attempt to display I/O stats on Linux using /proc/PID/io
22143	*/
22144	static void displayLinuxIoStats(void){
22145	FILE *in;
22146	char z[200];
22147	sqlite3_snprintf(sizeof(z), z, "/proc/%d/io", getpid());
22148	in = fopen(z, "rb");
22149	if( in==0 ) return;
22150	while( fgets(z, sizeof(z), in)!=0 ){
22151	static const struct {
22152	const char *zPattern;
22153	const char *zDesc;
22154	} aTrans[] = {
22155	{ "rchar: ", "Bytes received by read():" },
	@@ -22162,11 +23350,11 @@
22162	};
22163	int i;
22164	for(i=0; i<ArraySize(aTrans); i++){
22165	int n = strlen30(aTrans[i].zPattern);
22166	if( cli_strncmp(aTrans[i].zPattern, z, n)==0 ){
22167	oputf("%-36s %s", aTrans[i].zDesc, &z[n]);
22168	break;
22169	}
22170	}
22171	}
22172	fclose(in);
	@@ -22175,10 +23363,11 @@
22175
22176	/*
22177	** Display a single line of status using 64-bit values.
22178	*/
22179	static void displayStatLine(

22180	char zLabel, / Label for this one line */
22181	char zFormat, / Format for the result */
22182	int iStatusCtrl, /* Which status to display */
22183	int bReset /* True to reset the stats */
22184	){
	@@ -22193,11 +23382,11 @@
22193	if( nPercent>1 ){
22194	sqlite3_snprintf(sizeof(zLine), zLine, zFormat, iCur, iHiwtr);
22195	}else{
22196	sqlite3_snprintf(sizeof(zLine), zLine, zFormat, iHiwtr);
22197	}
22198	oputf("%-36s %s\n", zLabel, zLine);
22199	}
22200
22201	/*
22202	** Display memory stats.
22203	*/
	@@ -22206,130 +23395,152 @@
22206	ShellState pArg, / Pointer to ShellState */
22207	int bReset /* True to reset the stats */
22208	){
22209	int iCur;
22210	int iHiwtr;

22211	if( pArg==0 \|\| pArg->out==0 ) return 0;

22212
22213	if( pArg->pStmt && pArg->statsOn==2 ){
22214	int nCol, i, x;
22215	sqlite3_stmt *pStmt = pArg->pStmt;
22216	char z[100];
22217	nCol = sqlite3_column_count(pStmt);
22218	oputf("%-36s %d\n", "Number of output columns:", nCol);
22219	for(i=0; i<nCol; i++){
22220	sqlite3_snprintf(sizeof(z),z,"Column %d %nname:", i, &x);
22221	oputf("%-36s %s\n", z, sqlite3_column_name(pStmt,i));
22222	#ifndef SQLITE_OMIT_DECLTYPE
22223	sqlite3_snprintf(30, z+x, "declared type:");
22224	oputf("%-36s %s\n", z, sqlite3_column_decltype(pStmt, i));
22225	#endif
22226	#ifdef SQLITE_ENABLE_COLUMN_METADATA
22227	sqlite3_snprintf(30, z+x, "database name:");
22228	oputf("%-36s %s\n", z, sqlite3_column_database_name(pStmt,i));

22229	sqlite3_snprintf(30, z+x, "table name:");
22230	oputf("%-36s %s\n", z, sqlite3_column_table_name(pStmt,i));
22231	sqlite3_snprintf(30, z+x, "origin name:");
22232	oputf("%-36s %s\n", z, sqlite3_column_origin_name(pStmt,i));
22233	#endif
22234	}
22235	}
22236
22237	if( pArg->statsOn==3 ){
22238	if( pArg->pStmt ){
22239	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_VM_STEP,bReset);
22240	oputf("VM-steps: %d\n", iCur);
22241	}
22242	return 0;
22243	}
22244
22245	displayStatLine("Memory Used:",
22246	"%lld (max %lld) bytes", SQLITE_STATUS_MEMORY_USED, bReset);
22247	displayStatLine("Number of Outstanding Allocations:",
22248	"%lld (max %lld)", SQLITE_STATUS_MALLOC_COUNT, bReset);
22249	if( pArg->shellFlgs & SHFLG_Pagecache ){
22250	displayStatLine("Number of Pcache Pages Used:",
22251	"%lld (max %lld) pages", SQLITE_STATUS_PAGECACHE_USED, bReset);
22252	}
22253	displayStatLine("Number of Pcache Overflow Bytes:",
22254	"%lld (max %lld) bytes", SQLITE_STATUS_PAGECACHE_OVERFLOW, bReset);
22255	displayStatLine("Largest Allocation:",
22256	"%lld bytes", SQLITE_STATUS_MALLOC_SIZE, bReset);
22257	displayStatLine("Largest Pcache Allocation:",
22258	"%lld bytes", SQLITE_STATUS_PAGECACHE_SIZE, bReset);
22259	#ifdef YYTRACKMAXSTACKDEPTH
22260	displayStatLine("Deepest Parser Stack:",
22261	"%lld (max %lld)", SQLITE_STATUS_PARSER_STACK, bReset);
22262	#endif
22263
22264	if( db ){
22265	if( pArg->shellFlgs & SHFLG_Lookaside ){
22266	iHiwtr = iCur = -1;
22267	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_USED,
22268	&iCur, &iHiwtr, bReset);
22269	oputf("Lookaside Slots Used: %d (max %d)\n", iCur, iHiwtr);

22270	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_HIT,
22271	&iCur, &iHiwtr, bReset);
22272	oputf("Successful lookaside attempts: %d\n", iHiwtr);

22273	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_MISS_SIZE,
22274	&iCur, &iHiwtr, bReset);
22275	oputf("Lookaside failures due to size: %d\n", iHiwtr);

22276	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_MISS_FULL,
22277	&iCur, &iHiwtr, bReset);
22278	oputf("Lookaside failures due to OOM: %d\n", iHiwtr);

22279	}
22280	iHiwtr = iCur = -1;
22281	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_USED, &iCur, &iHiwtr, bReset);
22282	oputf("Pager Heap Usage: %d bytes\n", iCur);

22283	iHiwtr = iCur = -1;
22284	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_HIT, &iCur, &iHiwtr, 1);
22285	oputf("Page cache hits: %d\n", iCur);

22286	iHiwtr = iCur = -1;
22287	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_MISS, &iCur, &iHiwtr, 1);
22288	oputf("Page cache misses: %d\n", iCur);

22289	iHiwtr = iCur = -1;
22290	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_WRITE, &iCur, &iHiwtr, 1);
22291	oputf("Page cache writes: %d\n", iCur);

22292	iHiwtr = iCur = -1;
22293	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_SPILL, &iCur, &iHiwtr, 1);
22294	oputf("Page cache spills: %d\n", iCur);

22295	iHiwtr = iCur = -1;
22296	sqlite3_db_status(db, SQLITE_DBSTATUS_SCHEMA_USED, &iCur, &iHiwtr, bReset);
22297	oputf("Schema Heap Usage: %d bytes\n", iCur);

22298	iHiwtr = iCur = -1;
22299	sqlite3_db_status(db, SQLITE_DBSTATUS_STMT_USED, &iCur, &iHiwtr, bReset);
22300	oputf("Statement Heap/Lookaside Usage: %d bytes\n", iCur);

22301	}
22302
22303	if( pArg->pStmt ){
22304	int iHit, iMiss;
22305	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FULLSCAN_STEP,
22306	bReset);
22307	oputf("Fullscan Steps: %d\n", iCur);

22308	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_SORT, bReset);
22309	oputf("Sort Operations: %d\n", iCur);

22310	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_AUTOINDEX,bReset);
22311	oputf("Autoindex Inserts: %d\n", iCur);

22312	iHit = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FILTER_HIT,
22313	bReset);
22314	iMiss = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FILTER_MISS,
22315	bReset);
22316	if( iHit \|\| iMiss ){
22317	oputf("Bloom filter bypass taken: %d/%d\n", iHit, iHit+iMiss);

22318	}
22319	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_VM_STEP, bReset);
22320	oputf("Virtual Machine Steps: %d\n", iCur);

22321	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_REPREPARE,bReset);
22322	oputf("Reprepare operations: %d\n", iCur);

22323	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_RUN, bReset);
22324	oputf("Number of times run: %d\n", iCur);

22325	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_MEMUSED, bReset);
22326	oputf("Memory used by prepared stmt: %d\n", iCur);

22327	}
22328
22329	#ifdef __linux__
22330	displayLinuxIoStats();
22331	#endif
22332
22333	/* Do not remove this machine readable comment: extra-stats-output-here */
22334
22335	return 0;
	@@ -22721,21 +23932,21 @@
22721	#define BOX_1234 "\342\224\274" /* U+253c -\|- */
22722
22723	/* Draw horizontal line N characters long using unicode box
22724	** characters
22725	*/
22726	static void print_box_line(int N){
22727	const char zDash[] =
22728	BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24
22729	BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24;
22730	const int nDash = sizeof(zDash) - 1;
22731	N *= 3;
22732	while( N>nDash ){
22733	oputz(zDash);
22734	N -= nDash;
22735	}
22736	oputf("%.*s", N, zDash);
22737	}
22738
22739	/*
22740	** Draw a horizontal separator for a MODE_Box table.
22741	*/
	@@ -22746,19 +23957,19 @@
22746	const char *zSep2,
22747	const char *zSep3
22748	){
22749	int i;
22750	if( nArg>0 ){
22751	oputz(zSep1);
22752	print_box_line(p->actualWidth[0]+2);
22753	for(i=1; i<nArg; i++){
22754	oputz(zSep2);
22755	print_box_line(p->actualWidth[i]+2);
22756	}
22757	oputz(zSep3);
22758	}
22759	oputz("\n");
22760	}
22761
22762	/*
22763	** z[] is a line of text that is to be displayed the .mode box or table or
22764	** similar tabular formats. z[] might contain control characters such
	@@ -22788,16 +23999,26 @@
22788	}
22789	if( mxWidth<0 ) mxWidth = -mxWidth;
22790	if( mxWidth==0 ) mxWidth = 1000000;
22791	i = j = n = 0;
22792	while( n<mxWidth ){
22793	if( z[i]>=' ' ){









22794	n++;
22795	do{ i++; j++; }while( (z[i]&0xc0)==0x80 );

22796	continue;
22797	}
22798	if( z[i]=='\t' ){
22799	do{
22800	n++;
22801	j++;
22802	}while( (n&7)!=0 && n<mxWidth );
22803	i++;
	@@ -22835,13 +24056,21 @@
22835	}
22836	zOut = malloc( j+1 );
22837	shell_check_oom(zOut);
22838	i = j = n = 0;
22839	while( i<k ){
22840	if( z[i]>=' ' ){








22841	n++;
22842	do{ zOut[j++] = z[i++]; }while( (z[i]&0xc0)==0x80 );
22843	continue;
22844	}
22845	if( z[i]=='\t' ){
22846	do{
22847	n++;
	@@ -23017,97 +24246,99 @@
23017	rowSep = "\n";
23018	if( p->showHeader ){
23019	for(i=0; i<nColumn; i++){
23020	w = p->actualWidth[i];
23021	if( p->colWidth[i]<0 ) w = -w;
23022	utf8_width_print(w, azData[i]);
23023	fputs(i==nColumn-1?"\n":" ", p->out);
23024	}
23025	for(i=0; i<nColumn; i++){
23026	print_dashes(p->actualWidth[i]);
23027	fputs(i==nColumn-1?"\n":" ", p->out);
23028	}
23029	}
23030	break;
23031	}
23032	case MODE_Table: {
23033	colSep = " \| ";
23034	rowSep = " \|\n";
23035	print_row_separator(p, nColumn, "+");
23036	fputs("\| ", p->out);
23037	for(i=0; i<nColumn; i++){
23038	w = p->actualWidth[i];
23039	n = strlenChar(azData[i]);
23040	oputf("%s%s%s", (w-n)/2, "", azData[i], (w-n+1)/2, "");
23041	oputz(i==nColumn-1?" \|\n":" \| ");

23042	}
23043	print_row_separator(p, nColumn, "+");
23044	break;
23045	}
23046	case MODE_Markdown: {
23047	colSep = " \| ";
23048	rowSep = " \|\n";
23049	fputs("\| ", p->out);
23050	for(i=0; i<nColumn; i++){
23051	w = p->actualWidth[i];
23052	n = strlenChar(azData[i]);
23053	oputf("%s%s%s", (w-n)/2, "", azData[i], (w-n+1)/2, "");
23054	oputz(i==nColumn-1?" \|\n":" \| ");

23055	}
23056	print_row_separator(p, nColumn, "\|");
23057	break;
23058	}
23059	case MODE_Box: {
23060	colSep = " " BOX_13 " ";
23061	rowSep = " " BOX_13 "\n";
23062	print_box_row_separator(p, nColumn, BOX_23, BOX_234, BOX_34);
23063	oputz(BOX_13 " ");
23064	for(i=0; i<nColumn; i++){
23065	w = p->actualWidth[i];
23066	n = strlenChar(azData[i]);
23067	oputf("%s%s%s%s",
23068	(w-n)/2, "", azData[i], (w-n+1)/2, "",
23069	i==nColumn-1?" "BOX_13"\n":" "BOX_13" ");
23070	}
23071	print_box_row_separator(p, nColumn, BOX_123, BOX_1234, BOX_134);
23072	break;
23073	}
23074	}
23075	for(i=nColumn, j=0; i<nTotal; i++, j++){
23076	if( j==0 && p->cMode!=MODE_Column ){
23077	oputz(p->cMode==MODE_Box?BOX_13" ":"\| ");
23078	}
23079	z = azData[i];
23080	if( z==0 ) z = p->nullValue;
23081	w = p->actualWidth[j];
23082	if( p->colWidth[j]<0 ) w = -w;
23083	utf8_width_print(w, z);
23084	if( j==nColumn-1 ){
23085	oputz(rowSep);
23086	if( bMultiLineRowExists && abRowDiv[i/nColumn-1] && i+1<nTotal ){
23087	if( p->cMode==MODE_Table ){
23088	print_row_separator(p, nColumn, "+");
23089	}else if( p->cMode==MODE_Box ){
23090	print_box_row_separator(p, nColumn, BOX_123, BOX_1234, BOX_134);
23091	}else if( p->cMode==MODE_Column ){
23092	oputz("\n");
23093	}
23094	}
23095	j = -1;
23096	if( seenInterrupt ) goto columnar_end;
23097	}else{
23098	oputz(colSep);
23099	}
23100	}
23101	if( p->cMode==MODE_Table ){
23102	print_row_separator(p, nColumn, "+");
23103	}else if( p->cMode==MODE_Box ){
23104	print_box_row_separator(p, nColumn, BOX_12, BOX_124, BOX_14);
23105	}
23106	columnar_end:
23107	if( seenInterrupt ){
23108	oputz("Interrupt\n");
23109	}
23110	nData = (nRow+1)*nColumn;
23111	for(i=0; i<nData; i++){
23112	z = azData[i];
23113	if( z!=zEmpty && z!=zShowNull ) free(azData[i]);
	@@ -23190,11 +24421,13 @@
23190	}
23191	}
23192	} while( SQLITE_ROW == rc );
23193	sqlite3_free(pData);
23194	if( pArg->cMode==MODE_Json ){
23195	fputs("]\n", pArg->out);


23196	}else if( pArg->cMode==MODE_Count ){
23197	char zBuf[200];
23198	sqlite3_snprintf(sizeof(zBuf), zBuf, "%llu row%s\n",
23199	nRow, nRow!=1 ? "s" : "");
23200	printf("%s", zBuf);
	@@ -23239,10 +24472,11 @@
23239	int bCancel,
23240	char **pzErr
23241	){
23242	int rc = SQLITE_OK;
23243	sqlite3expert *p = pState->expert.pExpert;

23244	assert( p );
23245	assert( bCancel \|\| pzErr==0 \|\| *pzErr==0 );
23246	if( bCancel==0 ){
23247	int bVerbose = pState->expert.bVerbose;
23248
	@@ -23251,24 +24485,25 @@
23251	int nQuery = sqlite3_expert_count(p);
23252	int i;
23253
23254	if( bVerbose ){
23255	const char *zCand = sqlite3_expert_report(p,0,EXPERT_REPORT_CANDIDATES);
23256	oputz("-- Candidates -----------------------------\n");
23257	oputf("%s\n", zCand);
23258	}
23259	for(i=0; i<nQuery; i++){
23260	const char *zSql = sqlite3_expert_report(p, i, EXPERT_REPORT_SQL);
23261	const char *zIdx = sqlite3_expert_report(p, i, EXPERT_REPORT_INDEXES);
23262	const char *zEQP = sqlite3_expert_report(p, i, EXPERT_REPORT_PLAN);
23263	if( zIdx==0 ) zIdx = "(no new indexes)\n";
23264	if( bVerbose ){
23265	oputf("-- Query %d --------------------------------\n",i+1);
23266	oputf("%s\n\n", zSql);


23267	}
23268	oputf("%s\n", zIdx);
23269	oputf("%s\n", zEQP);
23270	}
23271	}
23272	}
23273	sqlite3_expert_destroy(p);
23274	pState->expert.pExpert = 0;
	@@ -23299,30 +24534,31 @@
23299	if( n>=2 && 0==cli_strncmp(z, "-verbose", n) ){
23300	pState->expert.bVerbose = 1;
23301	}
23302	else if( n>=2 && 0==cli_strncmp(z, "-sample", n) ){
23303	if( i==(nArg-1) ){
23304	eputf("option requires an argument: %s\n", z);
23305	rc = SQLITE_ERROR;
23306	}else{
23307	iSample = (int)integerValue(azArg[++i]);
23308	if( iSample<0 \|\| iSample>100 ){
23309	eputf("value out of range: %s\n", azArg[i]);
23310	rc = SQLITE_ERROR;
23311	}
23312	}
23313	}
23314	else{
23315	eputf("unknown option: %s\n", z);
23316	rc = SQLITE_ERROR;
23317	}
23318	}
23319
23320	if( rc==SQLITE_OK ){
23321	pState->expert.pExpert = sqlite3_expert_new(pState->db, &zErr);
23322	if( pState->expert.pExpert==0 ){
23323	eputf("sqlite3_expert_new: %s\n", zErr ? zErr : "out of memory");

23324	rc = SQLITE_ERROR;
23325	}else{
23326	sqlite3_expert_config(
23327	pState->expert.pExpert, EXPERT_CONFIG_SAMPLE, iSample
23328	);
	@@ -23647,33 +24883,33 @@
23647	if( zType==0 ) return 0;
23648	dataOnly = (p->shellFlgs & SHFLG_DumpDataOnly)!=0;
23649	noSys = (p->shellFlgs & SHFLG_DumpNoSys)!=0;
23650
23651	if( cli_strcmp(zTable, "sqlite_sequence")==0 && !noSys ){
23652	if( !dataOnly ) oputz("DELETE FROM sqlite_sequence;\n");
23653	}else if( sqlite3_strglob("sqlite_stat?", zTable)==0 && !noSys ){
23654	if( !dataOnly ) oputz("ANALYZE sqlite_schema;\n");
23655	}else if( cli_strncmp(zTable, "sqlite_", 7)==0 ){
23656	return 0;
23657	}else if( dataOnly ){
23658	/* no-op */
23659	}else if( cli_strncmp(zSql, "CREATE VIRTUAL TABLE", 20)==0 ){
23660	char *zIns;
23661	if( !p->writableSchema ){
23662	oputz("PRAGMA writable_schema=ON;\n");
23663	p->writableSchema = 1;
23664	}
23665	zIns = sqlite3_mprintf(
23666	"INSERT INTO sqlite_schema(type,name,tbl_name,rootpage,sql)"
23667	"VALUES('table','%q','%q',0,'%q');",
23668	zTable, zTable, zSql);
23669	shell_check_oom(zIns);
23670	oputf("%s\n", zIns);
23671	sqlite3_free(zIns);
23672	return 0;
23673	}else{
23674	printSchemaLine(zSql, ";\n");
23675	}
23676
23677	if( cli_strcmp(zType, "table")==0 ){
23678	ShellText sSelect;
23679	ShellText sTable;
	@@ -23727,11 +24963,11 @@
23727	savedMode = p->mode;
23728	p->zDestTable = sTable.z;
23729	p->mode = p->cMode = MODE_Insert;
23730	rc = shell_exec(p, sSelect.z, 0);
23731	if( (rc&0xff)==SQLITE_CORRUPT ){
23732	oputz("/**** CORRUPTION ERROR *****/\n");
23733	toggleSelectOrder(p->db);
23734	shell_exec(p, sSelect.z, 0);
23735	toggleSelectOrder(p->db);
23736	}
23737	p->zDestTable = savedDestTable;
	@@ -23758,28 +24994,28 @@
23758	char *zErr = 0;
23759	rc = sqlite3_exec(p->db, zQuery, dump_callback, p, &zErr);
23760	if( rc==SQLITE_CORRUPT ){
23761	char *zQ2;
23762	int len = strlen30(zQuery);
23763	oputz("/**** CORRUPTION ERROR *****/\n");
23764	if( zErr ){
23765	oputf("/**** %s ****/\n", zErr);
23766	sqlite3_free(zErr);
23767	zErr = 0;
23768	}
23769	zQ2 = malloc( len+100 );
23770	if( zQ2==0 ) return rc;
23771	sqlite3_snprintf(len+100, zQ2, "%s ORDER BY rowid DESC", zQuery);
23772	rc = sqlite3_exec(p->db, zQ2, dump_callback, p, &zErr);
23773	if( rc ){
23774	oputf("/**** ERROR: %s ****/\n", zErr);
23775	}else{
23776	rc = SQLITE_CORRUPT;
23777	}
23778	sqlite3_free(zErr);
23779	free(zQ2);
23780	}

23781	return rc;
23782	}
23783
23784	/*
23785	** Text of help messages.
	@@ -23832,18 +25068,17 @@
23832	#ifndef SQLITE_SHELL_FIDDLE
23833	".check GLOB Fail if output since .testcase does not match",
23834	".clone NEWDB Clone data into NEWDB from the existing database",
23835	#endif
23836	".connection [close] [#] Open or close an auxiliary database connection",
23837	#if defined(_WIN32) \|\| defined(WIN32)
23838	".crnl on\|off Translate \\n to \\r\\n. Default ON",
23839	#endif
23840	".databases List names and files of attached databases",
23841	".dbconfig ?op? ?val? List or change sqlite3_db_config() options",
23842	#if SQLITE_SHELL_HAVE_RECOVER
23843	".dbinfo ?DB? Show status information about the database",
23844	#endif

23845	".dump ?OBJECTS? Render database content as SQL",
23846	" Options:",
23847	" --data-only Output only INSERT statements",
23848	" --newlines Allow unescaped newline characters in output",
23849	" --nosys Omit system tables (ex: \"sqlite_stat1\")",
	@@ -23940,13 +25175,15 @@
23940	#endif
23941	".nullvalue STRING Use STRING in place of NULL values",
23942	#ifndef SQLITE_SHELL_FIDDLE
23943	".once ?OPTIONS? ?FILE? Output for the next SQL command only to FILE",
23944	" If FILE begins with '\|' then open as a pipe",
23945	" --bom Put a UTF8 byte-order mark at the beginning",
23946	" -e Send output to the system text editor",
23947	" -x Send output as CSV to a spreadsheet (same as \".excel\")",


23948	/* Note that .open is (partially) available in WASM builds but is
23949	** currently only intended to be used by the fiddle tool, not
23950	** end users, so is "undocumented." */
23951	".open ?OPTIONS? ?FILE? Close existing database and reopen FILE",
23952	" Options:",
	@@ -23965,10 +25202,12 @@
23965	".output ?FILE? Send output to FILE or stdout if FILE is omitted",
23966	" If FILE begins with '\|' then open it as a pipe.",
23967	" Options:",
23968	" --bom Prefix output with a UTF8 byte-order mark",
23969	" -e Send output to the system text editor",


23970	" -x Send output as CSV to a spreadsheet",
23971	#endif
23972	".parameter CMD ... Manage SQL parameter bindings",
23973	" clear Erase all bindings",
23974	" init Initialize the TEMP table that holds bindings",
	@@ -24078,10 +25317,14 @@
24078	".vfsinfo ?AUX? Information about the top-level VFS",
24079	".vfslist List all available VFSes",
24080	".vfsname ?AUX? Print the name of the VFS stack",
24081	".width NUM1 NUM2 ... Set minimum column widths for columnar output",
24082	" Negative values right-justify",




24083	};
24084
24085	/*
24086	** Output help text.
24087	**
	@@ -24126,24 +25369,24 @@
24126	hh &= ~HH_Summary;
24127	break;
24128	}
24129	if( ((hw^hh)&HH_Undoc)==0 ){
24130	if( (hh&HH_Summary)!=0 ){
24131	sputf(out, ".%s\n", azHelp[i]+1);
24132	++n;
24133	}else if( (hw&HW_SummaryOnly)==0 ){
24134	sputf(out, "%s\n", azHelp[i]);
24135	}
24136	}
24137	}
24138	}else{
24139	/* Seek documented commands for which zPattern is an exact prefix */
24140	zPat = sqlite3_mprintf(".%s*", zPattern);
24141	shell_check_oom(zPat);
24142	for(i=0; i<ArraySize(azHelp); i++){
24143	if( sqlite3_strglob(zPat, azHelp[i])==0 ){
24144	sputf(out, "%s\n", azHelp[i]);
24145	j = i+1;
24146	n++;
24147	}
24148	}
24149	sqlite3_free(zPat);
	@@ -24150,11 +25393,11 @@
24150	if( n ){
24151	if( n==1 ){
24152	/* when zPattern is a prefix of exactly one command, then include
24153	** the details of that command, which should begin at offset j */
24154	while( j<ArraySize(azHelp)-1 && azHelp[j][0]==' ' ){
24155	sputf(out, "%s\n", azHelp[j]);
24156	j++;
24157	}
24158	}
24159	return n;
24160	}
	@@ -24167,14 +25410,14 @@
24167	while( i<ArraySize(azHelp)-1 && azHelp[i+1][0]==' ' ) ++i;
24168	continue;
24169	}
24170	if( azHelp[i][0]=='.' ) j = i;
24171	if( sqlite3_strlike(zPat, azHelp[i], 0)==0 ){
24172	sputf(out, "%s\n", azHelp[j]);
24173	while( j<ArraySize(azHelp)-1 && azHelp[j+1][0]==' ' ){
24174	j++;
24175	sputf(out, "%s\n", azHelp[j]);
24176	}
24177	i = j;
24178	n++;
24179	}
24180	}
	@@ -24200,35 +25443,35 @@
24200	**
24201	** NULL is returned if any error is encountered. The final value of *pnByte
24202	** is undefined in this case.
24203	*/
24204	static char readFile(const char zName, int *pnByte){
24205	FILE *in = fopen(zName, "rb");
24206	long nIn;
24207	size_t nRead;
24208	char *pBuf;
24209	int rc;
24210	if( in==0 ) return 0;
24211	rc = fseek(in, 0, SEEK_END);
24212	if( rc!=0 ){
24213	eputf("Error: '%s' not seekable\n", zName);
24214	fclose(in);
24215	return 0;
24216	}
24217	nIn = ftell(in);
24218	rewind(in);
24219	pBuf = sqlite3_malloc64( nIn+1 );
24220	if( pBuf==0 ){
24221	eputz("Error: out of memory\n");
24222	fclose(in);
24223	return 0;
24224	}
24225	nRead = fread(pBuf, nIn, 1, in);
24226	fclose(in);
24227	if( nRead!=1 ){
24228	sqlite3_free(pBuf);
24229	eputf("Error: cannot read '%s'\n", zName);
24230	return 0;
24231	}
24232	pBuf[nIn] = 0;
24233	if( pnByte ) *pnByte = nIn;
24234	return pBuf;
	@@ -24290,11 +25533,11 @@
24290	** archive and the dfltZip flag is true, then assume it is a ZIP archive.
24291	** Otherwise, assume an ordinary database regardless of the filename if
24292	** the type cannot be determined from content.
24293	*/
24294	int deduceDatabaseType(const char *zName, int dfltZip){
24295	FILE *f = fopen(zName, "rb");
24296	size_t n;
24297	int rc = SHELL_OPEN_UNSPEC;
24298	char zBuf[100];
24299	if( f==0 ){
24300	if( dfltZip && sqlite3_strlike("%.zip",zName,0)==0 ){
	@@ -24343,13 +25586,13 @@
24343	FILE *in;
24344	const char *zDbFilename = p->pAuxDb->zDbFilename;
24345	unsigned int x[16];
24346	char zLine[1000];
24347	if( zDbFilename ){
24348	in = fopen(zDbFilename, "r");
24349	if( in==0 ){
24350	eputf("cannot open \"%s\" for reading\n", zDbFilename);
24351	return 0;
24352	}
24353	nLine = 0;
24354	}else{
24355	in = p->in;
	@@ -24356,24 +25599,24 @@
24356	nLine = p->lineno;
24357	if( in==0 ) in = stdin;
24358	}
24359	*pnData = 0;
24360	nLine++;
24361	if( fgets(zLine, sizeof(zLine), in)==0 ) goto readHexDb_error;
24362	rc = sscanf(zLine, "\| size %d pagesize %d", &n, &pgsz);
24363	if( rc!=2 ) goto readHexDb_error;
24364	if( n<0 ) goto readHexDb_error;
24365	if( pgsz<512 \|\| pgsz>65536 \|\| (pgsz&(pgsz-1))!=0 ) goto readHexDb_error;
24366	n = (n+pgsz-1)&~(pgsz-1); /* Round n up to the next multiple of pgsz */
24367	a = sqlite3_malloc( n ? n : 1 );
24368	shell_check_oom(a);
24369	memset(a, 0, n);
24370	if( pgsz<512 \|\| pgsz>65536 \|\| (pgsz & (pgsz-1))!=0 ){
24371	eputz("invalid pagesize\n");
24372	goto readHexDb_error;
24373	}
24374	for(nLine++; fgets(zLine, sizeof(zLine), in)!=0; nLine++){
24375	rc = sscanf(zLine, "\| page %d offset %d", &j, &k);
24376	if( rc==2 ){
24377	iOffset = k;
24378	continue;
24379	}
	@@ -24401,18 +25644,18 @@
24401
24402	readHexDb_error:
24403	if( in!=p->in ){
24404	fclose(in);
24405	}else{
24406	while( fgets(zLine, sizeof(zLine), p->in)!=0 ){
24407	nLine++;
24408	if(cli_strncmp(zLine, "\| end ", 6)==0 ) break;
24409	}
24410	p->lineno = nLine;
24411	}
24412	sqlite3_free(a);
24413	eputf("Error on line %d of --hexdb input\n", nLine);
24414	return 0;
24415	}
24416	#endif /* SQLITE_OMIT_DESERIALIZE */
24417
24418	/*
	@@ -24483,22 +25726,24 @@
24483	SQLITE_OPEN_READWRITE\|SQLITE_OPEN_CREATE\|p->openFlags, 0);
24484	break;
24485	}
24486	}
24487	if( p->db==0 \|\| SQLITE_OK!=sqlite3_errcode(p->db) ){
24488	eputf("Error: unable to open database \"%s\": %s\n",
24489	zDbFilename, sqlite3_errmsg(p->db));
24490	if( (openFlags & OPEN_DB_KEEPALIVE)==0 ){
24491	exit(1);
24492	}
24493	sqlite3_close(p->db);
24494	sqlite3_open(":memory:", &p->db);
24495	if( p->db==0 \|\| SQLITE_OK!=sqlite3_errcode(p->db) ){
24496	eputz("Also: unable to open substitute in-memory database.\n");

24497	exit(1);
24498	}else{
24499	eputf("Notice: using substitute in-memory database instead of \"%s\"\n",

24500	zDbFilename);
24501	}
24502	}
24503	globalDb = p->db;
24504	sqlite3_db_config(p->db, SQLITE_DBCONFIG_STMT_SCANSTATUS, (int)0, (int*)0);
	@@ -24511,10 +25756,11 @@
24511	}
24512
24513	#ifndef SQLITE_OMIT_LOAD_EXTENSION
24514	sqlite3_enable_load_extension(p->db, 1);
24515	#endif

24516	sqlite3_shathree_init(p->db, 0, 0);
24517	sqlite3_uint_init(p->db, 0, 0);
24518	sqlite3_stmtrand_init(p->db, 0, 0);
24519	sqlite3_decimal_init(p->db, 0, 0);
24520	sqlite3_percentile_init(p->db, 0, 0);
	@@ -24605,11 +25851,11 @@
24605	}
24606	rc = sqlite3_deserialize(p->db, "main", aData, nData, nData,
24607	SQLITE_DESERIALIZE_RESIZEABLE \|
24608	SQLITE_DESERIALIZE_FREEONCLOSE);
24609	if( rc ){
24610	eputf("Error: sqlite3_deserialize() returns %d\n", rc);
24611	}
24612	if( p->szMax>0 ){
24613	sqlite3_file_control(p->db, "main", SQLITE_FCNTL_SIZE_LIMIT, &p->szMax);
24614	}
24615	}
	@@ -24629,15 +25875,17 @@
24629	** Attempt to close the database connection. Report errors.
24630	*/
24631	void close_db(sqlite3 *db){
24632	int rc = sqlite3_close(db);
24633	if( rc ){
24634	eputf("Error: sqlite3_close() returns %d: %s\n", rc, sqlite3_errmsg(db));

24635	}
24636	}
24637
24638	#if HAVE_READLINE \|\| HAVE_EDITLINE

24639	/*
24640	** Readline completion callbacks
24641	*/
24642	static char readline_completion_generator(const char text, int state){
24643	static sqlite3_stmt *pStmt = 0;
	@@ -24671,19 +25919,26 @@
24671	#elif HAVE_LINENOISE
24672	/*
24673	** Linenoise completion callback. Note that the 3rd argument is from
24674	** the "msteveb" version of linenoise, not the "antirez" version.
24675	*/
24676	static void linenoise_completion(const char zLine, linenoiseCompletions lc,
24677	void *pUserData){





24678	i64 nLine = strlen(zLine);
24679	i64 i, iStart;
24680	sqlite3_stmt *pStmt = 0;
24681	char *zSql;
24682	char zBuf[1000];
24683

24684	UNUSED_PARAMETER(pUserData);

24685	if( nLine>(i64)sizeof(zBuf)-30 ) return;
24686	if( zLine[0]=='.' \|\| zLine[0]=='#') return;
24687	for(i=nLine-1; i>=0 && (isalnum(zLine[i]) \|\| zLine[i]=='_'); i--){}
24688	if( i==nLine-1 ) return;
24689	iStart = i+1;
	@@ -24793,11 +26048,12 @@
24793	return 1;
24794	}
24795	if( sqlite3_stricmp(zArg, "off")==0 \|\| sqlite3_stricmp(zArg,"no")==0 ){
24796	return 0;
24797	}
24798	eputf("ERROR: Not a boolean value: \"%s\". Assuming \"no\".\n", zArg);

24799	return 0;
24800	}
24801
24802	/*
24803	** Set or clear a shell flag according to a boolean value.
	@@ -24820,22 +26076,22 @@
24820	/*
24821	** Try to open an output file. The names "stdout" and "stderr" are
24822	** recognized and do the right thing. NULL is returned if the output
24823	** filename is "off".
24824	*/
24825	static FILE output_file_open(const char zFile, int bTextMode){
24826	FILE *f;
24827	if( cli_strcmp(zFile,"stdout")==0 ){
24828	f = stdout;
24829	}else if( cli_strcmp(zFile, "stderr")==0 ){
24830	f = stderr;
24831	}else if( cli_strcmp(zFile, "off")==0 ){
24832	f = 0;
24833	}else{
24834	f = fopen(zFile, bTextMode ? "w" : "wb");
24835	if( f==0 ){
24836	eputf("Error: cannot open \"%s\"\n", zFile);
24837	}
24838	}
24839	return f;
24840	}
24841
	@@ -24884,16 +26140,17 @@
24884	if( nSql>1000000000 ) nSql = 1000000000;
24885	while( nSql>0 && zSql[nSql-1]==';' ){ nSql--; }
24886	switch( mType ){
24887	case SQLITE_TRACE_ROW:
24888	case SQLITE_TRACE_STMT: {
24889	sputf(p->traceOut, "%.*s;\n", (int)nSql, zSql);
24890	break;
24891	}
24892	case SQLITE_TRACE_PROFILE: {
24893	sqlite3_int64 nNanosec = pX ? (sqlite3_int64)pX : 0;
24894	sputf(p->traceOut, "%.*s; -- %lld ns\n", (int)nSql, zSql, nNanosec);

24895	break;
24896	}
24897	}
24898	return 0;
24899	}
	@@ -24996,14 +26253,15 @@
24996	do{ p->n--; }while( p->z[p->n]!=cQuote );
24997	p->cTerm = c;
24998	break;
24999	}
25000	if( pc==cQuote && c!='\r' ){
25001	eputf("%s:%d: unescaped %c character\n", p->zFile, p->nLine, cQuote);

25002	}
25003	if( c==EOF ){
25004	eputf("%s:%d: unterminated %c-quoted field\n",
25005	p->zFile, startLine, cQuote);
25006	p->cTerm = c;
25007	break;
25008	}
25009	import_append_char(p, c);
	@@ -25098,11 +26356,11 @@
25098
25099	zQuery = sqlite3_mprintf("SELECT * FROM \"%w\"", zTable);
25100	shell_check_oom(zQuery);
25101	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25102	if( rc ){
25103	eputf("Error %d: %s on [%s]\n",
25104	sqlite3_extended_errcode(p->db), sqlite3_errmsg(p->db), zQuery);
25105	goto end_data_xfer;
25106	}
25107	n = sqlite3_column_count(pQuery);
25108	zInsert = sqlite3_malloc64(200 + nTable + n*3);
	@@ -25115,11 +26373,11 @@
25115	i += 2;
25116	}
25117	memcpy(zInsert+i, ");", 3);
25118	rc = sqlite3_prepare_v2(newDb, zInsert, -1, &pInsert, 0);
25119	if( rc ){
25120	eputf("Error %d: %s on [%s]\n",
25121	sqlite3_extended_errcode(newDb), sqlite3_errmsg(newDb), zInsert);
25122	goto end_data_xfer;
25123	}
25124	for(k=0; k<2; k++){
25125	while( (rc = sqlite3_step(pQuery))==SQLITE_ROW ){
	@@ -25151,11 +26409,11 @@
25151	}
25152	}
25153	} /* End for */
25154	rc = sqlite3_step(pInsert);
25155	if( rc!=SQLITE_OK && rc!=SQLITE_ROW && rc!=SQLITE_DONE ){
25156	eputf("Error %d: %s\n",
25157	sqlite3_extended_errcode(newDb), sqlite3_errmsg(newDb));
25158	}
25159	sqlite3_reset(pInsert);
25160	cnt++;
25161	if( (cnt%spinRate)==0 ){
	@@ -25169,11 +26427,11 @@
25169	zQuery = sqlite3_mprintf("SELECT * FROM \"%w\" ORDER BY rowid DESC;",
25170	zTable);
25171	shell_check_oom(zQuery);
25172	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25173	if( rc ){
25174	eputf("Warning: cannot step \"%s\" backwards", zTable);
25175	break;
25176	}
25177	} /* End for(k=0...) */
25178
25179	end_data_xfer:
	@@ -25206,23 +26464,24 @@
25206	zQuery = sqlite3_mprintf("SELECT name, sql FROM sqlite_schema"
25207	" WHERE %s ORDER BY rowid ASC", zWhere);
25208	shell_check_oom(zQuery);
25209	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25210	if( rc ){
25211	eputf("Error: (%d) %s on [%s]\n", sqlite3_extended_errcode(p->db),

25212	sqlite3_errmsg(p->db), zQuery);
25213	goto end_schema_xfer;
25214	}
25215	while( (rc = sqlite3_step(pQuery))==SQLITE_ROW ){
25216	zName = sqlite3_column_text(pQuery, 0);
25217	zSql = sqlite3_column_text(pQuery, 1);
25218	if( zName==0 \|\| zSql==0 ) continue;
25219	if( sqlite3_stricmp((char*)zName, "sqlite_sequence")!=0 ){
25220	sputf(stdout, "%s... ", zName); fflush(stdout);
25221	sqlite3_exec(newDb, (const char*)zSql, 0, 0, &zErrMsg);
25222	if( zErrMsg ){
25223	eputf("Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
25224	sqlite3_free(zErrMsg);
25225	zErrMsg = 0;
25226	}
25227	}
25228	if( xForEach ){
	@@ -25236,23 +26495,23 @@
25236	zQuery = sqlite3_mprintf("SELECT name, sql FROM sqlite_schema"
25237	" WHERE %s ORDER BY rowid DESC", zWhere);
25238	shell_check_oom(zQuery);
25239	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
25240	if( rc ){
25241	eputf("Error: (%d) %s on [%s]\n",
25242	sqlite3_extended_errcode(p->db), sqlite3_errmsg(p->db), zQuery);
25243	goto end_schema_xfer;
25244	}
25245	while( sqlite3_step(pQuery)==SQLITE_ROW ){
25246	zName = sqlite3_column_text(pQuery, 0);
25247	zSql = sqlite3_column_text(pQuery, 1);
25248	if( zName==0 \|\| zSql==0 ) continue;
25249	if( sqlite3_stricmp((char*)zName, "sqlite_sequence")==0 ) continue;
25250	sputf(stdout, "%s... ", zName); fflush(stdout);
25251	sqlite3_exec(newDb, (const char*)zSql, 0, 0, &zErrMsg);
25252	if( zErrMsg ){
25253	eputf("Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
25254	sqlite3_free(zErrMsg);
25255	zErrMsg = 0;
25256	}
25257	if( xForEach ){
25258	xForEach(p, newDb, (const char*)zName);
	@@ -25272,16 +26531,17 @@
25272	*/
25273	static void tryToClone(ShellState p, const char zNewDb){
25274	int rc;
25275	sqlite3 *newDb = 0;
25276	if( access(zNewDb,0)==0 ){
25277	eputf("File \"%s\" already exists.\n", zNewDb);
25278	return;
25279	}
25280	rc = sqlite3_open(zNewDb, &newDb);
25281	if( rc ){
25282	eputf("Cannot create output database: %s\n", sqlite3_errmsg(newDb));

25283	}else{
25284	sqlite3_exec(p->db, "PRAGMA writable_schema=ON;", 0, 0, 0);
25285	sqlite3_exec(newDb, "BEGIN EXCLUSIVE;", 0, 0, 0);
25286	tryToCloneSchema(p, newDb, "type='table'", tryToCloneData);
25287	tryToCloneSchema(p, newDb, "type!='table'", 0);
	@@ -25294,14 +26554,22 @@
25294	#ifndef SQLITE_SHELL_FIDDLE
25295	/*
25296	** Change the output stream (file or pipe or console) to something else.
25297	*/
25298	static void output_redir(ShellState p, FILE pfNew){
25299	if( p->out != stdout ) eputz("Output already redirected.\n");
25300	else{

25301	p->out = pfNew;
25302	setOutputStream(pfNew);







25303	}
25304	}
25305
25306	/*
25307	** Change the output file back to stdout.
	@@ -25314,10 +26582,13 @@
25314	if( p->outfile[0]=='\|' ){
25315	#ifndef SQLITE_OMIT_POPEN
25316	pclose(p->out);
25317	#endif
25318	}else{



25319	output_file_close(p->out);
25320	#ifndef SQLITE_NOHAVE_SYSTEM
25321	if( p->doXdgOpen ){
25322	const char *zXdgOpenCmd =
25323	#if defined(_WIN32)
	@@ -25328,11 +26599,11 @@
25328	"xdg-open";
25329	#endif
25330	char *zCmd;
25331	zCmd = sqlite3_mprintf("%s %s", zXdgOpenCmd, p->zTempFile);
25332	if( system(zCmd) ){
25333	eputf("Failed: [%s]\n", zCmd);
25334	}else{
25335	/* Give the start/open/xdg-open command some time to get
25336	** going before we continue, and potential delete the
25337	** p->zTempFile data file out from under it */
25338	sqlite3_sleep(2000);
	@@ -25343,28 +26614,34 @@
25343	}
25344	#endif /* !defined(SQLITE_NOHAVE_SYSTEM) */
25345	}
25346	p->outfile[0] = 0;
25347	p->out = stdout;
25348	setOutputStream(stdout);
25349	}
25350	#else
25351	# define output_redir(SS,pfO)
25352	# define output_reset(SS)
25353	#endif
25354
25355	/*
25356	** Run an SQL command and return the single integer result.
25357	*/
25358	static int db_int(sqlite3 db, const char zSql){
25359	sqlite3_stmt *pStmt;
25360	int res = 0;
25361	sqlite3_prepare_v2(db, zSql, -1, &pStmt, 0);





25362	if( pStmt && sqlite3_step(pStmt)==SQLITE_ROW ){
25363	res = sqlite3_column_int(pStmt,0);
25364	}
25365	sqlite3_finalize(pStmt);

25366	return res;
25367	}
25368
25369	#if SQLITE_SHELL_HAVE_RECOVER
25370	/*
	@@ -25419,11 +26696,11 @@
25419	if( p->db==0 ) return 1;
25420	rc = sqlite3_prepare_v2(p->db,
25421	"SELECT data FROM sqlite_dbpage(?1) WHERE pgno=1",
25422	-1, &pStmt, 0);
25423	if( rc ){
25424	eputf("error: %s\n", sqlite3_errmsg(p->db));
25425	sqlite3_finalize(pStmt);
25426	return 1;
25427	}
25428	sqlite3_bind_text(pStmt, 1, zDb, -1, SQLITE_STATIC);
25429	if( sqlite3_step(pStmt)==SQLITE_ROW
	@@ -25432,58 +26709,149 @@
25432	const u8 *pb = sqlite3_column_blob(pStmt,0);
25433	shell_check_oom(pb);
25434	memcpy(aHdr, pb, 100);
25435	sqlite3_finalize(pStmt);
25436	}else{
25437	eputz("unable to read database header\n");
25438	sqlite3_finalize(pStmt);
25439	return 1;
25440	}
25441	i = get2byteInt(aHdr+16);
25442	if( i==1 ) i = 65536;
25443	oputf("%-20s %d\n", "database page size:", i);
25444	oputf("%-20s %d\n", "write format:", aHdr[18]);
25445	oputf("%-20s %d\n", "read format:", aHdr[19]);
25446	oputf("%-20s %d\n", "reserved bytes:", aHdr[20]);
25447	for(i=0; i<ArraySize(aField); i++){
25448	int ofst = aField[i].ofst;
25449	unsigned int val = get4byteInt(aHdr + ofst);
25450	oputf("%-20s %u", aField[i].zName, val);
25451	switch( ofst ){
25452	case 56: {
25453	if( val==1 ) oputz(" (utf8)");
25454	if( val==2 ) oputz(" (utf16le)");
25455	if( val==3 ) oputz(" (utf16be)");
25456	}
25457	}
25458	oputz("\n");
25459	}
25460	if( zDb==0 ){
25461	zSchemaTab = sqlite3_mprintf("main.sqlite_schema");
25462	}else if( cli_strcmp(zDb,"temp")==0 ){
25463	zSchemaTab = sqlite3_mprintf("%s", "sqlite_temp_schema");
25464	}else{
25465	zSchemaTab = sqlite3_mprintf("\"%w\".sqlite_schema", zDb);
25466	}
25467	for(i=0; i<ArraySize(aQuery); i++){
25468	char *zSql = sqlite3_mprintf(aQuery[i].zSql, zSchemaTab);
25469	int val = db_int(p->db, zSql);
25470	sqlite3_free(zSql);
25471	oputf("%-20s %d\n", aQuery[i].zName, val);
25472	}
25473	sqlite3_free(zSchemaTab);
25474	sqlite3_file_control(p->db, zDb, SQLITE_FCNTL_DATA_VERSION, &iDataVersion);
25475	oputf("%-20s %u\n", "data version", iDataVersion);
25476	return 0;
25477	}
25478	#endif /* SQLITE_SHELL_HAVE_RECOVER */





























































































25479
25480	/*
25481	** Print the given string as an error message.
25482	*/
25483	static void shellEmitError(const char *zErr){
25484	eputf("Error: %s\n", zErr);
25485	}
25486	/*
25487	** Print the current sqlite3_errmsg() value to stderr and return 1.
25488	*/
25489	static int shellDatabaseError(sqlite3 *db){
	@@ -25726,10 +27094,11 @@
25726	int bGroupByParent = 0; /* If -groupbyparent is present */
25727	int i; /* To iterate through azArg[] */
25728	const char zIndent = ""; / How much to indent CREATE INDEX by */
25729	int rc; /* Return code */
25730	sqlite3_stmt pSql = 0; / Compiled version of SQL statement below */

25731
25732	/*
25733	** This SELECT statement returns one row for each foreign key constraint
25734	** in the schema of the main database. The column values are:
25735	**
	@@ -25801,11 +27170,12 @@
25801	else if( n>1 && sqlite3_strnicmp("-groupbyparent", azArg[i], n)==0 ){
25802	bGroupByParent = 1;
25803	zIndent = " ";
25804	}
25805	else{
25806	eputf("Usage: %s %s ?-verbose? ?-groupbyparent?\n", azArg[0], azArg[1]);

25807	return SQLITE_ERROR;
25808	}
25809	}
25810
25811	/* Register the fkey_collate_clause() SQL function */
	@@ -25845,44 +27215,45 @@
25845	}
25846	rc = sqlite3_finalize(pExplain);
25847	if( rc!=SQLITE_OK ) break;
25848
25849	if( res<0 ){
25850	eputz("Error: internal error");
25851	break;
25852	}else{
25853	if( bGroupByParent
25854	&& (bVerbose \|\| res==0)
25855	&& (zPrev==0 \|\| sqlite3_stricmp(zParent, zPrev))
25856	){
25857	oputf("-- Parent table %s\n", zParent);
25858	sqlite3_free(zPrev);
25859	zPrev = sqlite3_mprintf("%s", zParent);
25860	}
25861
25862	if( res==0 ){
25863	oputf("%s%s --> %s\n", zIndent, zCI, zTarget);
25864	}else if( bVerbose ){
25865	oputf("%s/* no extra indexes required for %s -> %s */\n",

25866	zIndent, zFrom, zTarget
25867	);
25868	}
25869	}
25870	}
25871	sqlite3_free(zPrev);
25872
25873	if( rc!=SQLITE_OK ){
25874	eputf("%s\n", sqlite3_errmsg(db));
25875	}
25876
25877	rc2 = sqlite3_finalize(pSql);
25878	if( rc==SQLITE_OK && rc2!=SQLITE_OK ){
25879	rc = rc2;
25880	eputf("%s\n", sqlite3_errmsg(db));
25881	}
25882	}else{
25883	eputf("%s\n", sqlite3_errmsg(db));
25884	}
25885
25886	return rc;
25887	}
25888
	@@ -25898,13 +27269,13 @@
25898	n = (nArg>=2 ? strlen30(azArg[1]) : 0);
25899	if( n<1 \|\| sqlite3_strnicmp(azArg[1], "fkey-indexes", n) ) goto usage;
25900	return lintFkeyIndexes(pState, azArg, nArg);
25901
25902	usage:
25903	eputf("Usage %s sub-command ?switches...?\n", azArg[0]);
25904	eputz("Where sub-commands are:\n");
25905	eputz(" fkey-indexes\n");
25906	return SQLITE_ERROR;
25907	}
25908
25909	static void shellPrepare(
25910	sqlite3 *db,
	@@ -25914,11 +27285,12 @@
25914	){
25915	*ppStmt = 0;
25916	if( *pRc==SQLITE_OK ){
25917	int rc = sqlite3_prepare_v2(db, zSql, -1, ppStmt, 0);
25918	if( rc!=SQLITE_OK ){
25919	eputf("sql error: %s (%d)\n", sqlite3_errmsg(db), sqlite3_errcode(db));

25920	*pRc = rc;
25921	}
25922	}
25923	}
25924
	@@ -25958,11 +27330,11 @@
25958	if( pStmt ){
25959	sqlite3 *db = sqlite3_db_handle(pStmt);
25960	int rc = sqlite3_finalize(pStmt);
25961	if( *pRc==SQLITE_OK ){
25962	if( rc!=SQLITE_OK ){
25963	eputf("SQL error: %s\n", sqlite3_errmsg(db));
25964	}
25965	*pRc = rc;
25966	}
25967	}
25968	}
	@@ -25980,11 +27352,11 @@
25980	){
25981	int rc = sqlite3_reset(pStmt);
25982	if( *pRc==SQLITE_OK ){
25983	if( rc!=SQLITE_OK ){
25984	sqlite3 *db = sqlite3_db_handle(pStmt);
25985	eputf("SQL error: %s\n", sqlite3_errmsg(db));
25986	}
25987	*pRc = rc;
25988	}
25989	}
25990	#endif /* !defined SQLITE_OMIT_VIRTUALTABLE */
	@@ -26009,10 +27381,11 @@
26009	char zSrcTable; / "sqlar", "zipfile($file)" or "zip" */
26010	const char zFile; / --file argument, or NULL */
26011	const char zDir; / --directory argument, or NULL */
26012	char *azArg; / Array of command arguments */
26013	ShellState p; / Shell state */

26014	sqlite3 db; / Database containing the archive */
26015	};
26016
26017	/*
26018	** Print a usage message for the .ar command to stderr and return SQLITE_ERROR.
	@@ -26032,13 +27405,13 @@
26032	va_start(ap, zFmt);
26033	z = sqlite3_vmprintf(zFmt, ap);
26034	va_end(ap);
26035	shellEmitError(z);
26036	if( pAr->fromCmdLine ){
26037	eputz("Use \"-A\" for more help\n");
26038	}else{
26039	eputz("Use \".archive --help\" for more help\n");
26040	}
26041	sqlite3_free(z);
26042	return SQLITE_ERROR;
26043	}
26044
	@@ -26134,11 +27507,11 @@
26134	};
26135	int nSwitch = sizeof(aSwitch) / sizeof(struct ArSwitch);
26136	struct ArSwitch *pEnd = &aSwitch[nSwitch];
26137
26138	if( nArg<=1 ){
26139	eputz("Wrong number of arguments. Usage:\n");
26140	return arUsage(stderr);
26141	}else{
26142	char *z = azArg[1];
26143	if( z[0]!='-' ){
26144	/* Traditional style [tar] invocation */
	@@ -26240,11 +27613,11 @@
26240	}
26241	}
26242	}
26243	}
26244	if( pAr->eCmd==0 ){
26245	eputz("Required argument missing. Usage:\n");
26246	return arUsage(stderr);
26247	}
26248	return SQLITE_OK;
26249	}
26250
	@@ -26283,11 +27656,11 @@
26283	if( SQLITE_ROW==sqlite3_step(pTest) ){
26284	bOk = 1;
26285	}
26286	shellReset(&rc, pTest);
26287	if( rc==SQLITE_OK && bOk==0 ){
26288	eputf("not found in archive: %s\n", z);
26289	rc = SQLITE_ERROR;
26290	}
26291	}
26292	shellFinalize(&rc, pTest);
26293	}
	@@ -26350,19 +27723,19 @@
26350	arWhereClause(&rc, pAr, &zWhere);
26351
26352	shellPreparePrintf(pAr->db, &rc, &pSql, zSql, azCols[pAr->bVerbose],
26353	pAr->zSrcTable, zWhere);
26354	if( pAr->bDryRun ){
26355	oputf("%s\n", sqlite3_sql(pSql));
26356	}else{
26357	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
26358	if( pAr->bVerbose ){
26359	oputf("%s % 10d %s %s\n",
26360	sqlite3_column_text(pSql, 0), sqlite3_column_int(pSql, 1),
26361	sqlite3_column_text(pSql, 2),sqlite3_column_text(pSql, 3));
26362	}else{
26363	oputf("%s\n", sqlite3_column_text(pSql, 0));
26364	}
26365	}
26366	}
26367	shellFinalize(&rc, pSql);
26368	sqlite3_free(zWhere);
	@@ -26385,11 +27758,11 @@
26385	}
26386	if( rc==SQLITE_OK ){
26387	zSql = sqlite3_mprintf("DELETE FROM %s WHERE %s;",
26388	pAr->zSrcTable, zWhere);
26389	if( pAr->bDryRun ){
26390	oputf("%s\n", zSql);
26391	}else{
26392	char *zErr = 0;
26393	rc = sqlite3_exec(pAr->db, "SAVEPOINT ar;", 0, 0, 0);
26394	if( rc==SQLITE_OK ){
26395	rc = sqlite3_exec(pAr->db, zSql, 0, 0, &zErr);
	@@ -26398,11 +27771,11 @@
26398	}else{
26399	rc = sqlite3_exec(pAr->db, "RELEASE ar;", 0, 0, 0);
26400	}
26401	}
26402	if( zErr ){
26403	sputf(stdout, "ERROR: %s\n", zErr); /* stdout? */
26404	sqlite3_free(zErr);
26405	}
26406	}
26407	}
26408	sqlite3_free(zWhere);
	@@ -26462,15 +27835,15 @@
26462	** populating them changes the timestamp). */
26463	for(i=0; i<2; i++){
26464	j = sqlite3_bind_parameter_index(pSql, "$dirOnly");
26465	sqlite3_bind_int(pSql, j, i);
26466	if( pAr->bDryRun ){
26467	oputf("%s\n", sqlite3_sql(pSql));
26468	}else{
26469	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
26470	if( i==0 && pAr->bVerbose ){
26471	oputf("%s\n", sqlite3_column_text(pSql, 0));
26472	}
26473	}
26474	}
26475	shellReset(&rc, pSql);
26476	}
	@@ -26486,17 +27859,17 @@
26486	** Run the SQL statement in zSql. Or if doing a --dryrun, merely print it out.
26487	*/
26488	static int arExecSql(ArCommand pAr, const char zSql){
26489	int rc;
26490	if( pAr->bDryRun ){
26491	oputf("%s\n", zSql);
26492	rc = SQLITE_OK;
26493	}else{
26494	char *zErr = 0;
26495	rc = sqlite3_exec(pAr->db, zSql, 0, 0, &zErr);
26496	if( zErr ){
26497	sputf(stdout, "ERROR: %s\n", zErr);
26498	sqlite3_free(zErr);
26499	}
26500	}
26501	return rc;
26502	}
	@@ -26641,10 +28014,11 @@
26641	cmd.fromCmdLine = fromCmdLine;
26642	rc = arParseCommand(azArg, nArg, &cmd);
26643	if( rc==SQLITE_OK ){
26644	int eDbType = SHELL_OPEN_UNSPEC;
26645	cmd.p = pState;

26646	cmd.db = pState->db;
26647	if( cmd.zFile ){
26648	eDbType = deduceDatabaseType(cmd.zFile, 1);
26649	}else{
26650	eDbType = pState->openMode;
	@@ -26667,17 +28041,18 @@
26667	}else{
26668	flags = SQLITE_OPEN_READONLY;
26669	}
26670	cmd.db = 0;
26671	if( cmd.bDryRun ){
26672	oputf("-- open database '%s'%s\n", cmd.zFile,
26673	eDbType==SHELL_OPEN_APPENDVFS ? " using 'apndvfs'" : "");
26674	}
26675	rc = sqlite3_open_v2(cmd.zFile, &cmd.db, flags,
26676	eDbType==SHELL_OPEN_APPENDVFS ? "apndvfs" : 0);
26677	if( rc!=SQLITE_OK ){
26678	eputf("cannot open file: %s (%s)\n", cmd.zFile, sqlite3_errmsg(cmd.db));

26679	goto end_ar_command;
26680	}
26681	sqlite3_fileio_init(cmd.db, 0, 0);
26682	sqlite3_sqlar_init(cmd.db, 0, 0);
26683	sqlite3_create_function(cmd.db, "shell_putsnl", 1, SQLITE_UTF8, cmd.p,
	@@ -26686,11 +28061,11 @@
26686	}
26687	if( cmd.zSrcTable==0 && cmd.bZip==0 && cmd.eCmd!=AR_CMD_HELP ){
26688	if( cmd.eCmd!=AR_CMD_CREATE
26689	&& sqlite3_table_column_metadata(cmd.db,0,"sqlar","name",0,0,0,0,0)
26690	){
26691	eputz("database does not contain an 'sqlar' table\n");
26692	rc = SQLITE_ERROR;
26693	goto end_ar_command;
26694	}
26695	cmd.zSrcTable = sqlite3_mprintf("sqlar");
26696	}
	@@ -26744,11 +28119,11 @@
26744	** This function is used as a callback by the recover extension. Simply
26745	** print the supplied SQL statement to stdout.
26746	*/
26747	static int recoverSqlCb(void pCtx, const char zSql){
26748	ShellState pState = (ShellState)pCtx;
26749	sputf(pState->out, "%s;\n", zSql);
26750	return SQLITE_OK;
26751	}
26752
26753	/*
26754	** This function is called to recover data from the database. A script
	@@ -26787,11 +28162,11 @@
26787	}else
26788	if( n<=10 && memcmp("-no-rowids", z, n)==0 ){
26789	bRowids = 0;
26790	}
26791	else{
26792	eputf("unexpected option: %s\n", azArg[i]);
26793	showHelp(pState->out, azArg[0]);
26794	return 1;
26795	}
26796	}
26797
	@@ -26806,11 +28181,11 @@
26806
26807	sqlite3_recover_run(p);
26808	if( sqlite3_recover_errcode(p)!=SQLITE_OK ){
26809	const char *zErr = sqlite3_recover_errmsg(p);
26810	int errCode = sqlite3_recover_errcode(p);
26811	eputf("sql error: %s (%d)\n", zErr, errCode);
26812	}
26813	rc = sqlite3_recover_finish(p);
26814	return rc;
26815	}
26816	#endif /* SQLITE_SHELL_HAVE_RECOVER */
	@@ -26828,25 +28203,25 @@
26828	i64 nError = 0;
26829	const char *zErr = 0;
26830	while( SQLITE_OK==sqlite3_intck_step(p) ){
26831	const char *zMsg = sqlite3_intck_message(p);
26832	if( zMsg ){
26833	oputf("%s\n", zMsg);
26834	nError++;
26835	}
26836	nStep++;
26837	if( nStepPerUnlock && (nStep % nStepPerUnlock)==0 ){
26838	sqlite3_intck_unlock(p);
26839	}
26840	}
26841	rc = sqlite3_intck_error(p, &zErr);
26842	if( zErr ){
26843	eputf("%s\n", zErr);
26844	}
26845	sqlite3_intck_close(p);
26846
26847	oputf("%lld steps, %lld errors\n", nStep, nError);
26848	}
26849
26850	return rc;
26851	}
26852
	@@ -26865,11 +28240,11 @@
26865	*/
26866	#ifdef SHELL_DEBUG
26867	#define rc_err_oom_die(rc) \
26868	if( rc==SQLITE_NOMEM ) shell_check_oom(0); \
26869	else if(!(rc==SQLITE_OK\|\|rc==SQLITE_DONE)) \
26870	eputf("E:%d\n",rc), assert(0)
26871	#else
26872	static void rc_err_oom_die(int rc){
26873	if( rc==SQLITE_NOMEM ) shell_check_oom(0);
26874	assert(rc==SQLITE_OK\|\|rc==SQLITE_DONE);
26875	}
	@@ -27023,12 +28398,12 @@
27023	}else if( *pDb==0 ){
27024	return 0;
27025	}else{
27026	/* Formulate the columns spec, close the DB, zero pDb. /
27027	char *zColsSpec = 0;
27028	int hasDupes = db_int(*pDb, zHasDupes);
27029	int nDigits = (hasDupes)? db_int(*pDb, zColDigits) : 0;
27030	if( hasDupes ){
27031	#ifdef SHELL_COLUMN_RENAME_CLEAN
27032	rc = sqlite3_exec(*pDb, zDedoctor, 0, 0, 0);
27033	rc_err_oom_die(rc);
27034	#endif
	@@ -27039,11 +28414,11 @@
27039	sqlite3_bind_int(pStmt, 1, nDigits);
27040	rc = sqlite3_step(pStmt);
27041	sqlite3_finalize(pStmt);
27042	if( rc!=SQLITE_DONE ) rc_err_oom_die(SQLITE_NOMEM);
27043	}
27044	assert(db_int(pDb, zHasDupes)==0); / Consider: remove this */
27045	rc = sqlite3_prepare_v2(*pDb, zCollectVar, -1, &pStmt, 0);
27046	rc_err_oom_die(rc);
27047	rc = sqlite3_step(pStmt);
27048	if( rc==SQLITE_ROW ){
27049	zColsSpec = sqlite3_mprintf("%s", sqlite3_column_text(pStmt, 0));
	@@ -27082,12 +28457,13 @@
27082	shellPreparePrintf(p->db, &rc, &pStmt,
27083	"SELECT 1 FROM sqlite_schema o WHERE "
27084	"sql LIKE 'CREATE VIRTUAL TABLE%%' AND %s", zLike ? zLike : "true"
27085	);
27086	if( rc==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW ){
27087	oputz("/* WARNING: "
27088	"Script requires that SQLITE_DBCONFIG_DEFENSIVE be disabled */\n"

27089	);
27090	}
27091	shellFinalize(&rc, pStmt);
27092	return rc;
27093	}
	@@ -27114,16 +28490,18 @@
27114	return SQLITE_OK;
27115	}
27116	if( faultsim_state.iCnt ){
27117	if( faultsim_state.iCnt>0 ) faultsim_state.iCnt--;
27118	if( faultsim_state.eVerbose>=2 ){
27119	oputf("FAULT-SIM id=%d no-fault (cnt=%d)\n", iArg, faultsim_state.iCnt);

27120	}
27121	return SQLITE_OK;
27122	}
27123	if( faultsim_state.eVerbose>=1 ){
27124	oputf("FAULT-SIM id=%d returns %d\n", iArg, faultsim_state.iErr);

27125	}
27126	faultsim_state.iCnt = faultsim_state.iInterval;
27127	faultsim_state.nHit++;
27128	if( faultsim_state.nRepeat>0 && faultsim_state.nRepeat<=faultsim_state.nHit ){
27129	faultsim_state.iCnt = -1;
	@@ -27182,11 +28560,11 @@
27182	clearTempFile(p);
27183
27184	#ifndef SQLITE_OMIT_AUTHORIZATION
27185	if( c=='a' && cli_strncmp(azArg[0], "auth", n)==0 ){
27186	if( nArg!=2 ){
27187	eputz("Usage: .auth ON\|OFF\n");
27188	rc = 1;
27189	goto meta_command_exit;
27190	}
27191	open_db(p, 0);
27192	if( booleanValue(azArg[1]) ){
	@@ -27229,32 +28607,32 @@
27229	}else
27230	if( cli_strcmp(z, "-async")==0 ){
27231	bAsync = 1;
27232	}else
27233	{
27234	eputf("unknown option: %s\n", azArg[j]);
27235	return 1;
27236	}
27237	}else if( zDestFile==0 ){
27238	zDestFile = azArg[j];
27239	}else if( zDb==0 ){
27240	zDb = zDestFile;
27241	zDestFile = azArg[j];
27242	}else{
27243	eputz("Usage: .backup ?DB? ?OPTIONS? FILENAME\n");
27244	return 1;
27245	}
27246	}
27247	if( zDestFile==0 ){
27248	eputz("missing FILENAME argument on .backup\n");
27249	return 1;
27250	}
27251	if( zDb==0 ) zDb = "main";
27252	rc = sqlite3_open_v2(zDestFile, &pDest,
27253	SQLITE_OPEN_READWRITE\|SQLITE_OPEN_CREATE, zVfs);
27254	if( rc!=SQLITE_OK ){
27255	eputf("Error: cannot open \"%s\"\n", zDestFile);
27256	close_db(pDest);
27257	return 1;
27258	}
27259	if( bAsync ){
27260	sqlite3_exec(pDest, "PRAGMA synchronous=OFF; PRAGMA journal_mode=OFF;",
	@@ -27286,23 +28664,14 @@
27286	eputz("Usage: .bail on\|off\n");
27287	rc = 1;
27288	}
27289	}else
27290
27291	/* Undocumented. Legacy only. See "crnl" below */
27292	if( c=='b' && n>=3 && cli_strncmp(azArg[0], "binary", n)==0 ){
27293	if( nArg==2 ){
27294	if( booleanValue(azArg[1]) ){
27295	setBinaryMode(p->out, 1);
27296	}else{
27297	setTextMode(p->out, 1);
27298	}
27299	}else{
27300	eputz("The \".binary\" command is deprecated. Use \".crnl\" instead.\n"
27301	"Usage: .binary on\|off\n");
27302	rc = 1;
27303	}
27304	}else
27305
27306	/* The undocumented ".breakpoint" command causes a call to the no-op
27307	** routine named test_breakpoint().
27308	*/
	@@ -27320,11 +28689,11 @@
27320	sqlite3_free(z);
27321	#else
27322	rc = chdir(azArg[1]);
27323	#endif
27324	if( rc ){
27325	eputf("Cannot change to directory \"%s\"\n", azArg[1]);
27326	rc = 1;
27327	}
27328	}else{
27329	eputz("Usage: .cd DIRECTORY\n");
27330	rc = 1;
	@@ -27353,15 +28722,16 @@
27353	eputz("Usage: .check GLOB-PATTERN\n");
27354	rc = 2;
27355	}else if( (zRes = readFile("testcase-out.txt", 0))==0 ){
27356	rc = 2;
27357	}else if( testcase_glob(azArg[1],zRes)==0 ){
27358	eputf("testcase-%s FAILED\n Expected: [%s]\n Got: [%s]\n",

27359	p->zTestcase, azArg[1], zRes);
27360	rc = 1;
27361	}else{
27362	oputf("testcase-%s ok\n", p->zTestcase);
27363	p->nCheck++;
27364	}
27365	sqlite3_free(zRes);
27366	}else
27367	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
	@@ -27390,13 +28760,13 @@
27390	zFile = "(memory)";
27391	}else if( zFile[0]==0 ){
27392	zFile = "(temporary-file)";
27393	}
27394	if( p->pAuxDb == &p->aAuxDb[i] ){
27395	sputf(stdout, "ACTIVE %d: %s\n", i, zFile);
27396	}else if( p->aAuxDb[i].db!=0 ){
27397	sputf(stdout, " %d: %s\n", i, zFile);
27398	}
27399	}
27400	}else if( nArg==2 && IsDigit(azArg[1][0]) && azArg[1][1]==0 ){
27401	int i = azArg[1][0] - '0';
27402	if( p->pAuxDb != &p->aAuxDb[i] && i>=0 && i<ArraySize(p->aAuxDb) ){
	@@ -27422,24 +28792,22 @@
27422	eputz("Usage: .connection [close] [CONNECTION-NUMBER]\n");
27423	rc = 1;
27424	}
27425	}else
27426
27427	if( c=='c' && n==4 && cli_strncmp(azArg[0], "crnl", n)==0 ){



27428	if( nArg==2 ){
27429	if( booleanValue(azArg[1]) ){
27430	setTextMode(p->out, 1);
27431	}else{
27432	setBinaryMode(p->out, 1);
27433	}
27434	}else{
27435	#if !defined(_WIN32) && !defined(WIN32)
27436	eputz("The \".crnl\" is a no-op on non-Windows machines.\n");
27437	#endif
27438	eputz("Usage: .crnl on\|off\n");
27439	rc = 1;
27440	}

27441	}else
27442
27443	if( c=='d' && n>1 && cli_strncmp(azArg[0], "databases", n)==0 ){
27444	char **azName = 0;
27445	int nName = 0;
	@@ -27465,11 +28833,11 @@
27465	sqlite3_finalize(pStmt);
27466	for(i=0; i<nName; i++){
27467	int eTxn = sqlite3_txn_state(p->db, azName[i*2]);
27468	int bRdonly = sqlite3_db_readonly(p->db, azName[i*2]);
27469	const char z = azName[i2+1];
27470	oputf("%s: %s %s%s\n",
27471	azName[i*2], z && z[0] ? z : "\"\"", bRdonly ? "r/o" : "r/w",
27472	eTxn==SQLITE_TXN_NONE ? "" :
27473	eTxn==SQLITE_TXN_READ ? " read-txn" : " write-txn");
27474	free(azName[i*2]);
27475	free(azName[i*2+1]);
	@@ -27507,15 +28875,16 @@
27507	if( nArg>1 && cli_strcmp(azArg[1], aDbConfig[ii].zName)!=0 ) continue;
27508	if( nArg>=3 ){
27509	sqlite3_db_config(p->db, aDbConfig[ii].op, booleanValue(azArg[2]), 0);
27510	}
27511	sqlite3_db_config(p->db, aDbConfig[ii].op, -1, &v);
27512	oputf("%19s %s\n", aDbConfig[ii].zName, v ? "on" : "off");

27513	if( nArg>1 ) break;
27514	}
27515	if( nArg>1 && ii==ArraySize(aDbConfig) ){
27516	eputf("Error: unknown dbconfig \"%s\"\n", azArg[1]);
27517	eputz("Enter \".dbconfig\" with no arguments for a list\n");
27518	}
27519	}else
27520
27521	#if SQLITE_SHELL_HAVE_RECOVER
	@@ -27561,11 +28930,12 @@
27561	}else
27562	if( cli_strcmp(z,"nosys")==0 ){
27563	ShellSetFlag(p, SHFLG_DumpNoSys);
27564	}else
27565	{
27566	eputf("Unknown option \"%s\" on \".dump\"\n", azArg[i]);

27567	rc = 1;
27568	sqlite3_free(zLike);
27569	goto meta_command_exit;
27570	}
27571	}else{
	@@ -27596,12 +28966,12 @@
27596	outputDumpWarning(p, zLike);
27597	if( (p->shellFlgs & SHFLG_DumpDataOnly)==0 ){
27598	/* When playing back a "dump", the content might appear in an order
27599	** which causes immediate foreign key constraints to be violated.
27600	** So disable foreign-key constraint enforcement to prevent problems. */
27601	oputz("PRAGMA foreign_keys=OFF;\n");
27602	oputz("BEGIN TRANSACTION;\n");
27603	}
27604	p->writableSchema = 0;
27605	p->showHeader = 0;
27606	/* Set writable_schema=ON since doing so forces SQLite to initialize
27607	** as much of the schema as it can even if the sqlite_schema table is
	@@ -27629,17 +28999,17 @@
27629	run_table_dump_query(p, zSql);
27630	sqlite3_free(zSql);
27631	}
27632	sqlite3_free(zLike);
27633	if( p->writableSchema ){
27634	oputz("PRAGMA writable_schema=OFF;\n");
27635	p->writableSchema = 0;
27636	}
27637	sqlite3_exec(p->db, "PRAGMA writable_schema=OFF;", 0, 0, 0);
27638	sqlite3_exec(p->db, "RELEASE dump;", 0, 0, 0);
27639	if( (p->shellFlgs & SHFLG_DumpDataOnly)==0 ){
27640	oputz(p->nErr?"ROLLBACK; -- due to errors\n":"COMMIT;\n");
27641	}
27642	p->showHeader = savedShowHeader;
27643	p->shellFlgs = savedShellFlags;
27644	}else
27645
	@@ -27649,10 +29019,14 @@
27649	}else{
27650	eputz("Usage: .echo on\|off\n");
27651	rc = 1;
27652	}
27653	}else




27654
27655	if( c=='e' && cli_strncmp(azArg[0], "eqp", n)==0 ){
27656	if( nArg==2 ){
27657	p->autoEQPtest = 0;
27658	if( p->autoEQPtrace ){
	@@ -27715,11 +29089,12 @@
27715	}else
27716
27717	#ifndef SQLITE_OMIT_VIRTUALTABLE
27718	if( c=='e' && cli_strncmp(azArg[0], "expert", n)==0 ){
27719	if( p->bSafeMode ){
27720	eputf("Cannot run experimental commands such as \"%s\" in safe mode\n",

27721	azArg[0]);
27722	rc = 1;
27723	}else{
27724	open_db(p, 0);
27725	expertDotCommand(p, azArg, nArg);
	@@ -27772,13 +29147,14 @@
27772	if( zCmd[0]=='-' && zCmd[1] ) zCmd++;
27773	}
27774
27775	/* --help lists all file-controls */
27776	if( cli_strcmp(zCmd,"help")==0 ){
27777	oputz("Available file-controls:\n");
27778	for(i=0; i<ArraySize(aCtrl); i++){
27779	oputf(" .filectrl %s %s\n", aCtrl[i].zCtrlName, aCtrl[i].zUsage);

27780	}
27781	rc = 1;
27782	goto meta_command_exit;
27783	}
27784
	@@ -27789,19 +29165,19 @@
27789	if( cli_strncmp(zCmd, aCtrl[i].zCtrlName, n2)==0 ){
27790	if( filectrl<0 ){
27791	filectrl = aCtrl[i].ctrlCode;
27792	iCtrl = i;
27793	}else{
27794	eputf("Error: ambiguous file-control: \"%s\"\n"
27795	"Use \".filectrl --help\" for help\n", zCmd);
27796	rc = 1;
27797	goto meta_command_exit;
27798	}
27799	}
27800	}
27801	if( filectrl<0 ){
27802	eputf("Error: unknown file-control: %s\n"
27803	"Use \".filectrl --help\" for help\n", zCmd);
27804	}else{
27805	switch(filectrl){
27806	case SQLITE_FCNTL_SIZE_LIMIT: {
27807	if( nArg!=2 && nArg!=3 ) break;
	@@ -27841,11 +29217,11 @@
27841	case SQLITE_FCNTL_TEMPFILENAME: {
27842	char *z = 0;
27843	if( nArg!=2 ) break;
27844	sqlite3_file_control(p->db, zSchema, filectrl, &z);
27845	if( z ){
27846	oputf("%s\n", z);
27847	sqlite3_free(z);
27848	}
27849	isOk = 2;
27850	break;
27851	}
	@@ -27855,23 +29231,24 @@
27855	x = atoi(azArg[2]);
27856	sqlite3_file_control(p->db, zSchema, filectrl, &x);
27857	}
27858	x = -1;
27859	sqlite3_file_control(p->db, zSchema, filectrl, &x);
27860	oputf("%d\n", x);
27861	isOk = 2;
27862	break;
27863	}
27864	}
27865	}
27866	if( isOk==0 && iCtrl>=0 ){
27867	oputf("Usage: .filectrl %s %s\n", zCmd,aCtrl[iCtrl].zUsage);

27868	rc = 1;
27869	}else if( isOk==1 ){
27870	char zBuf[100];
27871	sqlite3_snprintf(sizeof(zBuf), zBuf, "%lld", iRes);
27872	oputf("%s\n", zBuf);
27873	}
27874	}else
27875
27876	if( c=='f' && cli_strncmp(azArg[0], "fullschema", n)==0 ){
27877	ShellState data;
	@@ -27908,19 +29285,19 @@
27908	doStats = sqlite3_step(pStmt)==SQLITE_ROW;
27909	sqlite3_finalize(pStmt);
27910	}
27911	}
27912	if( doStats==0 ){
27913	oputz("/* No STAT tables available */\n");
27914	}else{
27915	oputz("ANALYZE sqlite_schema;\n");
27916	data.cMode = data.mode = MODE_Insert;
27917	data.zDestTable = "sqlite_stat1";
27918	shell_exec(&data, "SELECT * FROM sqlite_stat1", 0);
27919	data.zDestTable = "sqlite_stat4";
27920	shell_exec(&data, "SELECT * FROM sqlite_stat4", 0);
27921	oputz("ANALYZE sqlite_schema;\n");
27922	}
27923	}else
27924
27925	if( c=='h' && cli_strncmp(azArg[0], "headers", n)==0 ){
27926	if( nArg==2 ){
	@@ -27934,11 +29311,11 @@
27934
27935	if( c=='h' && cli_strncmp(azArg[0], "help", n)==0 ){
27936	if( nArg>=2 ){
27937	n = showHelp(p->out, azArg[1]);
27938	if( n==0 ){
27939	oputf("Nothing matches '%s'\n", azArg[1]);
27940	}
27941	}else{
27942	showHelp(p->out, 0);
27943	}
27944	}else
	@@ -27977,11 +29354,11 @@
27977	if( zFile==0 ){
27978	zFile = z;
27979	}else if( zTable==0 ){
27980	zTable = z;
27981	}else{
27982	oputf("ERROR: extra argument: \"%s\". Usage:\n", z);
27983	showHelp(p->out, "import");
27984	goto meta_command_exit;
27985	}
27986	}else if( cli_strcmp(z,"-v")==0 ){
27987	eVerbose++;
	@@ -27998,17 +29375,17 @@
27998	sCtx.cColSep = ',';
27999	sCtx.cRowSep = '\n';
28000	xRead = csv_read_one_field;
28001	useOutputMode = 0;
28002	}else{
28003	oputf("ERROR: unknown option: \"%s\". Usage:\n", z);
28004	showHelp(p->out, "import");
28005	goto meta_command_exit;
28006	}
28007	}
28008	if( zTable==0 ){
28009	oputf("ERROR: missing %s argument. Usage:\n",
28010	zFile==0 ? "FILE" : "TABLE");
28011	showHelp(p->out, "import");
28012	goto meta_command_exit;
28013	}
28014	seenInterrupt = 0;
	@@ -28054,32 +29431,32 @@
28054	if( sCtx.zFile[0]=='\|' ){
28055	#ifdef SQLITE_OMIT_POPEN
28056	eputz("Error: pipes are not supported in this OS\n");
28057	goto meta_command_exit;
28058	#else
28059	sCtx.in = popen(sCtx.zFile+1, "r");
28060	sCtx.zFile = "<pipe>";
28061	sCtx.xCloser = pclose;
28062	#endif
28063	}else{
28064	sCtx.in = fopen(sCtx.zFile, "rb");
28065	sCtx.xCloser = fclose;
28066	}
28067	if( sCtx.in==0 ){
28068	eputf("Error: cannot open \"%s\"\n", zFile);
28069	goto meta_command_exit;
28070	}
28071	if( eVerbose>=2 \|\| (eVerbose>=1 && useOutputMode) ){
28072	char zSep[2];
28073	zSep[1] = 0;
28074	zSep[0] = sCtx.cColSep;
28075	oputz("Column separator ");
28076	output_c_string(zSep);
28077	oputz(", row separator ");
28078	zSep[0] = sCtx.cRowSep;
28079	output_c_string(zSep);
28080	oputz("\n");
28081	}
28082	sCtx.z = sqlite3_malloc64(120);
28083	if( sCtx.z==0 ){
28084	import_cleanup(&sCtx);
28085	shell_out_of_memory();
	@@ -28087,11 +29464,15 @@
28087	/* Below, resources must be freed before exit. */
28088	while( (nSkip--)>0 ){
28089	while( xRead(&sCtx) && sCtx.cTerm==sCtx.cColSep ){}
28090	}
28091	import_append_char(&sCtx, 0); /* To ensure sCtx.z is allocated */
28092	if( sqlite3_table_column_metadata(p->db, zSchema, zTable,0,0,0,0,0,0) ){




28093	/* Table does not exist. Create it. */
28094	sqlite3 *dbCols = 0;
28095	char *zRenames = 0;
28096	char *zColDefs;
28097	zCreate = sqlite3_mprintf("CREATE TABLE \"%w\".\"%w\"",
	@@ -28100,18 +29481,18 @@
28100	zAutoColumn(sCtx.z, &dbCols, 0);
28101	if( sCtx.cTerm!=sCtx.cColSep ) break;
28102	}
28103	zColDefs = zAutoColumn(0, &dbCols, &zRenames);
28104	if( zRenames!=0 ){
28105	sputf((stdin_is_interactive && p->in==stdin)? p->out : stderr,
28106	"Columns renamed during .import %s due to duplicates:\n"
28107	"%s\n", sCtx.zFile, zRenames);
28108	sqlite3_free(zRenames);
28109	}
28110	assert(dbCols==0);
28111	if( zColDefs==0 ){
28112	eputf("%s: empty file\n", sCtx.zFile);
28113	import_cleanup(&sCtx);
28114	rc = 1;
28115	sqlite3_free(zCreate);
28116	goto meta_command_exit;
28117	}
	@@ -28119,17 +29500,20 @@
28119	if( zCreate==0 ){
28120	import_cleanup(&sCtx);
28121	shell_out_of_memory();
28122	}
28123	if( eVerbose>=1 ){
28124	oputf("%s\n", zCreate);
28125	}
28126	rc = sqlite3_exec(p->db, zCreate, 0, 0, 0);




28127	sqlite3_free(zCreate);
28128	zCreate = 0;
28129	if( rc ){
28130	eputf("%s failed:\n%s\n", zCreate, sqlite3_errmsg(p->db));
28131	import_cleanup(&sCtx);
28132	rc = 1;
28133	goto meta_command_exit;
28134	}
28135	}
	@@ -28180,11 +29564,11 @@
28180	}
28181	zSql[j++] = ')';
28182	zSql[j] = 0;
28183	assert( j<nByte );
28184	if( eVerbose>=2 ){
28185	oputf("Insert using: %s\n", zSql);
28186	}
28187	rc = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0);
28188	sqlite3_free(zSql);
28189	zSql = 0;
28190	if( rc ){
	@@ -28219,11 +29603,11 @@
28219	if( z==0 && (xRead==csv_read_one_field) && i==nCol-1 && i>0 ){
28220	z = "";
28221	}
28222	sqlite3_bind_text(pStmt, i+1, z, -1, SQLITE_TRANSIENT);
28223	if( i<nCol-1 && sCtx.cTerm!=sCtx.cColSep ){
28224	eputf("%s:%d: expected %d columns but found %d"
28225	" - filling the rest with NULL\n",
28226	sCtx.zFile, startLine, nCol, i+1);
28227	i += 2;
28228	while( i<=nCol ){ sqlite3_bind_null(pStmt, i); i++; }
28229	}
	@@ -28231,18 +29615,19 @@
28231	if( sCtx.cTerm==sCtx.cColSep ){
28232	do{
28233	xRead(&sCtx);
28234	i++;
28235	}while( sCtx.cTerm==sCtx.cColSep );
28236	eputf("%s:%d: expected %d columns but found %d - extras ignored\n",

28237	sCtx.zFile, startLine, nCol, i);
28238	}
28239	if( i>=nCol ){
28240	sqlite3_step(pStmt);
28241	rc = sqlite3_reset(pStmt);
28242	if( rc!=SQLITE_OK ){
28243	eputf("%s:%d: INSERT failed: %s\n",
28244	sCtx.zFile, startLine, sqlite3_errmsg(p->db));
28245	sCtx.nErr++;
28246	}else{
28247	sCtx.nRow++;
28248	}
	@@ -28251,11 +29636,12 @@
28251
28252	import_cleanup(&sCtx);
28253	sqlite3_finalize(pStmt);
28254	if( needCommit ) sqlite3_exec(p->db, "COMMIT", 0, 0, 0);
28255	if( eVerbose>0 ){
28256	oputf("Added %d rows with %d errors using %d lines of input\n",

28257	sCtx.nRow, sCtx.nErr, sCtx.nLine-1);
28258	}
28259	}else
28260	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
28261
	@@ -28267,11 +29653,11 @@
28267	int tnum = 0;
28268	int isWO = 0; /* True if making an imposter of a WITHOUT ROWID table */
28269	int lenPK = 0; /* Length of the PRIMARY KEY string for isWO tables */
28270	int i;
28271	if( !ShellHasFlag(p,SHFLG_TestingMode) ){
28272	eputf(".%s unavailable without --unsafe-testing\n",
28273	"imposter");
28274	rc = 1;
28275	goto meta_command_exit;
28276	}
28277	if( !(nArg==3 \|\| (nArg==2 && sqlite3_stricmp(azArg[1],"off")==0)) ){
	@@ -28333,11 +29719,11 @@
28333	zCollist = sqlite3_mprintf("%z,\"%w\"", zCollist, zCol);
28334	}
28335	}
28336	sqlite3_finalize(pStmt);
28337	if( i==0 \|\| tnum==0 ){
28338	eputf("no such index: \"%s\"\n", azArg[1]);
28339	rc = 1;
28340	sqlite3_free(zCollist);
28341	goto meta_command_exit;
28342	}
28343	if( lenPK==0 ) lenPK = 100000;
	@@ -28348,18 +29734,20 @@
28348	rc = sqlite3_test_control(SQLITE_TESTCTRL_IMPOSTER, p->db, "main", 1, tnum);
28349	if( rc==SQLITE_OK ){
28350	rc = sqlite3_exec(p->db, zSql, 0, 0, 0);
28351	sqlite3_test_control(SQLITE_TESTCTRL_IMPOSTER, p->db, "main", 0, 0);
28352	if( rc ){
28353	eputf("Error in [%s]: %s\n", zSql, sqlite3_errmsg(p->db));

28354	}else{
28355	sputf(stdout, "%s;\n", zSql);
28356	sputf(stdout, "WARNING: writing to an imposter table will corrupt"

28357	" the \"%s\" %s!\n", azArg[1], isWO ? "table" : "index");
28358	}
28359	}else{
28360	eputf("SQLITE_TESTCTRL_IMPOSTER returns %d\n", rc);
28361	rc = 1;
28362	}
28363	sqlite3_free(zSql);
28364	}else
28365	#endif /* !defined(SQLITE_OMIT_TEST_CONTROL) */
	@@ -28369,11 +29757,11 @@
28369	if( nArg==2 ){
28370	iArg = integerValue(azArg[1]);
28371	if( iArg==0 ) iArg = -1;
28372	}
28373	if( (nArg!=1 && nArg!=2) \|\| iArg<0 ){
28374	eputf("%s","Usage: .intck STEPS_PER_UNLOCK\n");
28375	rc = 1;
28376	goto meta_command_exit;
28377	}
28378	open_db(p, 0);
28379	rc = intckDatabaseCmd(p, iArg);
	@@ -28388,13 +29776,13 @@
28388	sqlite3IoTrace = 0;
28389	}else if( cli_strcmp(azArg[1], "-")==0 ){
28390	sqlite3IoTrace = iotracePrintf;
28391	iotrace = stdout;
28392	}else{
28393	iotrace = fopen(azArg[1], "w");
28394	if( iotrace==0 ){
28395	eputf("Error: cannot open \"%s\"\n", azArg[1]);
28396	sqlite3IoTrace = 0;
28397	rc = 1;
28398	}else{
28399	sqlite3IoTrace = iotracePrintf;
28400	}
	@@ -28422,11 +29810,11 @@
28422	};
28423	int i, n2;
28424	open_db(p, 0);
28425	if( nArg==1 ){
28426	for(i=0; i<ArraySize(aLimit); i++){
28427	sputf(stdout, "%20s %d\n", aLimit[i].zLimitName,
28428	sqlite3_limit(p->db, aLimit[i].limitCode, -1));
28429	}
28430	}else if( nArg>3 ){
28431	eputz("Usage: .limit NAME ?NEW-VALUE?\n");
28432	rc = 1;
	@@ -28437,28 +29825,28 @@
28437	for(i=0; i<ArraySize(aLimit); i++){
28438	if( sqlite3_strnicmp(aLimit[i].zLimitName, azArg[1], n2)==0 ){
28439	if( iLimit<0 ){
28440	iLimit = i;
28441	}else{
28442	eputf("ambiguous limit: \"%s\"\n", azArg[1]);
28443	rc = 1;
28444	goto meta_command_exit;
28445	}
28446	}
28447	}
28448	if( iLimit<0 ){
28449	eputf("unknown limit: \"%s\"\n"
28450	"enter \".limits\" with no arguments for a list.\n",
28451	azArg[1]);
28452	rc = 1;
28453	goto meta_command_exit;
28454	}
28455	if( nArg==3 ){
28456	sqlite3_limit(p->db, aLimit[iLimit].limitCode,
28457	(int)integerValue(azArg[2]));
28458	}
28459	sputf(stdout, "%20s %d\n", aLimit[iLimit].zLimitName,
28460	sqlite3_limit(p->db, aLimit[iLimit].limitCode, -1));
28461	}
28462	}else
28463
28464	if( c=='l' && n>2 && cli_strncmp(azArg[0], "lint", n)==0 ){
	@@ -28503,11 +29891,11 @@
28503	" than \"on\" or \"off\"\n");
28504	zFile = "off";
28505	}
28506	output_file_close(p->pLog);
28507	if( cli_strcmp(zFile,"on")==0 ) zFile = "stdout";
28508	p->pLog = output_file_open(zFile, 0);
28509	}
28510	}else
28511
28512	if( c=='m' && cli_strncmp(azArg[0], "mode", n)==0 ){
28513	const char *zMode = 0;
	@@ -28537,35 +29925,37 @@
28537	cmOpts = cmo;
28538	}
28539	}else if( zTabname==0 ){
28540	zTabname = z;
28541	}else if( z[0]=='-' ){
28542	eputf("unknown option: %s\n", z);
28543	eputz("options:\n"
28544	" --noquote\n"
28545	" --quote\n"
28546	" --wordwrap on/off\n"
28547	" --wrap N\n"
28548	" --ww\n");
28549	rc = 1;
28550	goto meta_command_exit;
28551	}else{
28552	eputf("extra argument: \"%s\"\n", z);
28553	rc = 1;
28554	goto meta_command_exit;
28555	}
28556	}
28557	if( zMode==0 ){
28558	if( p->mode==MODE_Column
28559	\|\| (p->mode>=MODE_Markdown && p->mode<=MODE_Box)
28560	){
28561	oputf("current output mode: %s --wrap %d --wordwrap %s --%squote\n",

28562	modeDescr[p->mode], p->cmOpts.iWrap,
28563	p->cmOpts.bWordWrap ? "on" : "off",
28564	p->cmOpts.bQuote ? "" : "no");
28565	}else{
28566	oputf("current output mode: %s\n", modeDescr[p->mode]);

28567	}
28568	zMode = modeDescr[p->mode];
28569	}
28570	n2 = strlen30(zMode);
28571	if( cli_strncmp(zMode,"lines",n2)==0 ){
	@@ -28634,11 +30024,11 @@
28634	if( c=='n' && cli_strcmp(azArg[0], "nonce")==0 ){
28635	if( nArg!=2 ){
28636	eputz("Usage: .nonce NONCE\n");
28637	rc = 1;
28638	}else if( p->zNonce==0 \|\| cli_strcmp(azArg[1],p->zNonce)!=0 ){
28639	eputf("line %d: incorrect nonce: \"%s\"\n",
28640	p->lineno, azArg[1]);
28641	exit(1);
28642	}else{
28643	p->bSafeMode = 0;
28644	return 0; /* Return immediately to bypass the safe mode reset
	@@ -28689,15 +30079,15 @@
28689	p->szMax = integerValue(azArg[++iName]);
28690	#endif /* SQLITE_OMIT_DESERIALIZE */
28691	}else
28692	#endif /* !SQLITE_SHELL_FIDDLE */
28693	if( z[0]=='-' ){
28694	eputf("unknown option: %s\n", z);
28695	rc = 1;
28696	goto meta_command_exit;
28697	}else if( zFN ){
28698	eputf("extra argument: \"%s\"\n", z);
28699	rc = 1;
28700	goto meta_command_exit;
28701	}else{
28702	zFN = z;
28703	}
	@@ -28735,11 +30125,11 @@
28735	zNewFilename = 0;
28736	}
28737	p->pAuxDb->zDbFilename = zNewFilename;
28738	open_db(p, OPEN_DB_KEEPALIVE);
28739	if( p->db==0 ){
28740	eputf("Error: cannot open '%s'\n", zNewFilename);
28741	sqlite3_free(zNewFilename);
28742	}else{
28743	p->pAuxDb->zFreeOnClose = zNewFilename;
28744	}
28745	}
	@@ -28753,22 +30143,26 @@
28753	#ifndef SQLITE_SHELL_FIDDLE
28754	if( (c=='o'
28755	&& (cli_strncmp(azArg[0], "output", n)==0
28756	\|\| cli_strncmp(azArg[0], "once", n)==0))
28757	\|\| (c=='e' && n==5 && cli_strcmp(azArg[0],"excel")==0)

28758	){
28759	char *zFile = 0;
28760	int bTxtMode = 0;
28761	int i;
28762	int eMode = 0;
28763	int bOnce = 0; /* 0: .output, 1: .once, 2: .excel */
28764	static const char *zBomUtf8 = "\xef\xbb\xbf";

28765	const char *zBom = 0;
28766
28767	failIfSafeMode(p, "cannot run .%s in safe mode", azArg[0]);
28768	if( c=='e' ){
28769	eMode = 'x';



28770	bOnce = 2;
28771	}else if( cli_strncmp(azArg[0],"once",n)==0 ){
28772	bOnce = 1;
28773	}
28774	for(i=1; i<nArg; i++){
	@@ -28775,28 +30169,34 @@
28775	char *z = azArg[i];
28776	if( z[0]=='-' ){
28777	if( z[1]=='-' ) z++;
28778	if( cli_strcmp(z,"-bom")==0 ){
28779	zBom = zBomUtf8;
28780	}else if( c!='e' && cli_strcmp(z,"-x")==0 ){


28781	eMode = 'x'; /* spreadsheet */
28782	}else if( c!='e' && cli_strcmp(z,"-e")==0 ){
28783	eMode = 'e'; /* text editor */


28784	}else{
28785	oputf("ERROR: unknown option: \"%s\". Usage:\n", azArg[i]);

28786	showHelp(p->out, azArg[0]);
28787	rc = 1;
28788	goto meta_command_exit;
28789	}
28790	}else if( zFile==0 && eMode!='e' && eMode!='x' ){
28791	zFile = sqlite3_mprintf("%s", z);
28792	if( zFile && zFile[0]=='\|' ){
28793	while( i+1<nArg ) zFile = sqlite3_mprintf("%z %s", zFile, azArg[++i]);
28794	break;
28795	}
28796	}else{
28797	oputf("ERROR: extra parameter: \"%s\". Usage:\n", azArg[i]);

28798	showHelp(p->out, azArg[0]);
28799	rc = 1;
28800	sqlite3_free(zFile);
28801	goto meta_command_exit;
28802	}
	@@ -28809,24 +30209,31 @@
28809	}else{
28810	p->outCount = 0;
28811	}
28812	output_reset(p);
28813	#ifndef SQLITE_NOHAVE_SYSTEM
28814	if( eMode=='e' \|\| eMode=='x' ){
28815	p->doXdgOpen = 1;
28816	outputModePush(p);
28817	if( eMode=='x' ){
28818	/* spreadsheet mode. Output as CSV. */
28819	newTempFile(p, "csv");
28820	ShellClearFlag(p, SHFLG_Echo);
28821	p->mode = MODE_Csv;
28822	sqlite3_snprintf(sizeof(p->colSeparator), p->colSeparator, SEP_Comma);
28823	sqlite3_snprintf(sizeof(p->rowSeparator), p->rowSeparator, SEP_CrLf);








28824	}else{
28825	/* text editor mode */
28826	newTempFile(p, "txt");
28827	bTxtMode = 1;
28828	}
28829	sqlite3_free(zFile);
28830	zFile = sqlite3_mprintf("%s", p->zTempFile);
28831	}
28832	#endif /* SQLITE_NOHAVE_SYSTEM */
	@@ -28835,30 +30242,36 @@
28835	#ifdef SQLITE_OMIT_POPEN
28836	eputz("Error: pipes are not supported in this OS\n");
28837	rc = 1;
28838	output_redir(p, stdout);
28839	#else
28840	FILE *pfPipe = popen(zFile + 1, "w");
28841	if( pfPipe==0 ){
28842	eputf("Error: cannot open pipe \"%s\"\n", zFile + 1);
28843	rc = 1;
28844	}else{
28845	output_redir(p, pfPipe);
28846	if( zBom ) oputz(zBom);
28847	sqlite3_snprintf(sizeof(p->outfile), p->outfile, "%s", zFile);
28848	}
28849	#endif
28850	}else{
28851	FILE *pfFile = output_file_open(zFile, bTxtMode);
28852	if( pfFile==0 ){
28853	if( cli_strcmp(zFile,"off")!=0 ){
28854	eputf("Error: cannot write to \"%s\"\n", zFile);
28855	}
28856	rc = 1;
28857	} else {
28858	output_redir(p, pfFile);
28859	if( zBom ) oputz(zBom);






28860	sqlite3_snprintf(sizeof(p->outfile), p->outfile, "%s", zFile);
28861	}
28862	}
28863	sqlite3_free(zFile);
28864	}else
	@@ -28895,11 +30308,12 @@
28895	if( len ){
28896	rx = sqlite3_prepare_v2(p->db,
28897	"SELECT key, quote(value) "
28898	"FROM temp.sqlite_parameters;", -1, &pStmt, 0);
28899	while( rx==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW ){
28900	oputf("%-*s %s\n", len, sqlite3_column_text(pStmt,0),

28901	sqlite3_column_text(pStmt,1));
28902	}
28903	sqlite3_finalize(pStmt);
28904	}
28905	}else
	@@ -28940,11 +30354,11 @@
28940	"VALUES(%Q,%Q);", zKey, zValue);
28941	shell_check_oom(zSql);
28942	rx = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0);
28943	sqlite3_free(zSql);
28944	if( rx!=SQLITE_OK ){
28945	oputf("Error: %s\n", sqlite3_errmsg(p->db));
28946	sqlite3_finalize(pStmt);
28947	pStmt = 0;
28948	rc = 1;
28949	}
28950	}
	@@ -28969,14 +30383,14 @@
28969	}else
28970
28971	if( c=='p' && n>=3 && cli_strncmp(azArg[0], "print", n)==0 ){
28972	int i;
28973	for(i=1; i<nArg; i++){
28974	if( i>1 ) oputz(" ");
28975	oputz(azArg[i]);
28976	}
28977	oputz("\n");
28978	}else
28979
28980	#ifndef SQLITE_OMIT_PROGRESS_CALLBACK
28981	if( c=='p' && n>=3 && cli_strncmp(azArg[0], "progress", n)==0 ){
28982	int i;
	@@ -29009,11 +30423,11 @@
29009	}else{
29010	p->mxProgress = (int)integerValue(azArg[++i]);
29011	}
29012	continue;
29013	}
29014	eputf("Error: unknown option: \"%s\"\n", azArg[i]);
29015	rc = 1;
29016	goto meta_command_exit;
29017	}else{
29018	nn = (int)integerValue(z);
29019	}
	@@ -29050,23 +30464,22 @@
29050	}
29051	if( azArg[1][0]=='\|' ){
29052	#ifdef SQLITE_OMIT_POPEN
29053	eputz("Error: pipes are not supported in this OS\n");
29054	rc = 1;
29055	p->out = stdout;
29056	#else
29057	p->in = popen(azArg[1]+1, "r");
29058	if( p->in==0 ){
29059	eputf("Error: cannot open \"%s\"\n", azArg[1]);
29060	rc = 1;
29061	}else{
29062	rc = process_input(p);
29063	pclose(p->in);
29064	}
29065	#endif
29066	}else if( (p->in = openChrSource(azArg[1]))==0 ){
29067	eputf("Error: cannot open \"%s\"\n", azArg[1]);
29068	rc = 1;
29069	}else{
29070	rc = process_input(p);
29071	fclose(p->in);
29072	}
	@@ -29095,11 +30508,11 @@
29095	rc = 1;
29096	goto meta_command_exit;
29097	}
29098	rc = sqlite3_open(zSrcFile, &pSrc);
29099	if( rc!=SQLITE_OK ){
29100	eputf("Error: cannot open \"%s\"\n", zSrcFile);
29101	close_db(pSrc);
29102	return 1;
29103	}
29104	open_db(p, 0);
29105	pBackup = sqlite3_backup_init(p->db, zDb, pSrc, "main");
	@@ -29127,11 +30540,14 @@
29127	}
29128	close_db(pSrc);
29129	}else
29130	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
29131
29132	if( c=='s' && cli_strncmp(azArg[0], "scanstats", n)==0 ){



29133	if( nArg==2 ){
29134	if( cli_strcmp(azArg[1], "vm")==0 ){
29135	p->scanstatsOn = 3;
29136	}else
29137	if( cli_strcmp(azArg[1], "est")==0 ){
	@@ -29178,11 +30594,11 @@
29178	}else if( optionMatch(azArg[ii],"debug") ){
29179	bDebug = 1;
29180	}else if( optionMatch(azArg[ii],"nosys") ){
29181	bNoSystemTabs = 1;
29182	}else if( azArg[ii][0]=='-' ){
29183	eputf("Unknown option: \"%s\"\n", azArg[ii]);
29184	rc = 1;
29185	goto meta_command_exit;
29186	}else if( zName==0 ){
29187	zName = azArg[ii];
29188	}else{
	@@ -29279,11 +30695,11 @@
29279	appendText(&sSelect, "name NOT LIKE 'sqlite_%%' AND ", 0);
29280	}
29281	appendText(&sSelect, "sql IS NOT NULL"
29282	" ORDER BY snum, rowid", 0);
29283	if( bDebug ){
29284	oputf("SQL: %s;\n", sSelect.z);
29285	}else{
29286	rc = sqlite3_exec(p->db, sSelect.z, callback, &data, &zErrMsg);
29287	}
29288	freeText(&sSelect);
29289	}
	@@ -29340,11 +30756,12 @@
29340	session_not_open:
29341	eputz("ERROR: No sessions are open\n");
29342	}else{
29343	rc = sqlite3session_attach(pSession->p, azCmd[1]);
29344	if( rc ){
29345	eputf("ERROR: sqlite3session_attach() returns %d\n",rc);

29346	rc = 0;
29347	}
29348	}
29349	}else
29350
	@@ -29357,13 +30774,13 @@
29357	){
29358	FILE *out = 0;
29359	failIfSafeMode(p, "cannot run \".session %s\" in safe mode", azCmd[0]);
29360	if( nCmd!=2 ) goto session_syntax_error;
29361	if( pSession->p==0 ) goto session_not_open;
29362	out = fopen(azCmd[1], "wb");
29363	if( out==0 ){
29364	eputf("ERROR: cannot open \"%s\" for writing\n",
29365	azCmd[1]);
29366	}else{
29367	int szChng;
29368	void *pChng;
29369	if( azCmd[0][0]=='c' ){
	@@ -29370,16 +30787,17 @@
29370	rc = sqlite3session_changeset(pSession->p, &szChng, &pChng);
29371	}else{
29372	rc = sqlite3session_patchset(pSession->p, &szChng, &pChng);
29373	}
29374	if( rc ){
29375	sputf(stdout, "Error: error code %d\n", rc);
29376	rc = 0;
29377	}
29378	if( pChng
29379	&& fwrite(pChng, szChng, 1, out)!=1 ){
29380	eputf("ERROR: Failed to write entire %d-byte output\n", szChng);

29381	}
29382	sqlite3_free(pChng);
29383	fclose(out);
29384	}
29385	}else
	@@ -29402,11 +30820,12 @@
29402	int ii;
29403	if( nCmd>2 ) goto session_syntax_error;
29404	ii = nCmd==1 ? -1 : booleanValue(azCmd[1]);
29405	if( pAuxDb->nSession ){
29406	ii = sqlite3session_enable(pSession->p, ii);
29407	oputf("session %s enable flag = %d\n", pSession->zName, ii);

29408	}
29409	}else
29410
29411	/* .session filter GLOB ....
29412	** Set a list of GLOB patterns of table names to be excluded.
	@@ -29437,11 +30856,12 @@
29437	int ii;
29438	if( nCmd>2 ) goto session_syntax_error;
29439	ii = nCmd==1 ? -1 : booleanValue(azCmd[1]);
29440	if( pAuxDb->nSession ){
29441	ii = sqlite3session_indirect(pSession->p, ii);
29442	oputf("session %s indirect flag = %d\n", pSession->zName, ii);

29443	}
29444	}else
29445
29446	/* .session isempty
29447	** Determine if the session is empty
	@@ -29449,20 +30869,21 @@
29449	if( cli_strcmp(azCmd[0], "isempty")==0 ){
29450	int ii;
29451	if( nCmd!=1 ) goto session_syntax_error;
29452	if( pAuxDb->nSession ){
29453	ii = sqlite3session_isempty(pSession->p);
29454	oputf("session %s isempty flag = %d\n", pSession->zName, ii);

29455	}
29456	}else
29457
29458	/* .session list
29459	** List all currently open sessions
29460	*/
29461	if( cli_strcmp(azCmd[0],"list")==0 ){
29462	for(i=0; i<pAuxDb->nSession; i++){
29463	oputf("%d %s\n", i, pAuxDb->aSession[i].zName);
29464	}
29465	}else
29466
29467	/* .session open DB NAME
29468	** Open a new session called NAME on the attached database DB.
	@@ -29473,22 +30894,23 @@
29473	if( nCmd!=3 ) goto session_syntax_error;
29474	zName = azCmd[2];
29475	if( zName[0]==0 ) goto session_syntax_error;
29476	for(i=0; i<pAuxDb->nSession; i++){
29477	if( cli_strcmp(pAuxDb->aSession[i].zName,zName)==0 ){
29478	eputf("Session \"%s\" already exists\n", zName);
29479	goto meta_command_exit;
29480	}
29481	}
29482	if( pAuxDb->nSession>=ArraySize(pAuxDb->aSession) ){
29483	eputf("Maximum of %d sessions\n", ArraySize(pAuxDb->aSession));

29484	goto meta_command_exit;
29485	}
29486	pSession = &pAuxDb->aSession[pAuxDb->nSession];
29487	rc = sqlite3session_create(p->db, azCmd[1], &pSession->p);
29488	if( rc ){
29489	eputf("Cannot open session: error code=%d\n", rc);
29490	rc = 0;
29491	goto meta_command_exit;
29492	}
29493	pSession->nFilter = 0;
29494	sqlite3session_table_filter(pSession->p, session_filter, pSession);
	@@ -29508,20 +30930,20 @@
29508	if( c=='s' && n>=10 && cli_strncmp(azArg[0], "selftest-", 9)==0 ){
29509	if( cli_strncmp(azArg[0]+9, "boolean", n-9)==0 ){
29510	int i, v;
29511	for(i=1; i<nArg; i++){
29512	v = booleanValue(azArg[i]);
29513	oputf("%s: %d 0x%x\n", azArg[i], v, v);
29514	}
29515	}
29516	if( cli_strncmp(azArg[0]+9, "integer", n-9)==0 ){
29517	int i; sqlite3_int64 v;
29518	for(i=1; i<nArg; i++){
29519	char zBuf[200];
29520	v = integerValue(azArg[i]);
29521	sqlite3_snprintf(sizeof(zBuf),zBuf,"%s: %lld 0x%llx\n", azArg[i],v,v);
29522	oputz(zBuf);
29523	}
29524	}
29525	}else
29526	#endif
29527
	@@ -29544,12 +30966,13 @@
29544	}else
29545	if( cli_strcmp(z,"-v")==0 ){
29546	bVerbose++;
29547	}else
29548	{
29549	eputf("Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
29550	eputz("Should be one of: --init -v\n");

29551	rc = 1;
29552	goto meta_command_exit;
29553	}
29554	}
29555	if( sqlite3_table_column_metadata(p->db,"main","selftest",0,0,0,0,0,0)
	@@ -29590,46 +31013,47 @@
29590	if( zOp==0 ) continue;
29591	if( zSql==0 ) continue;
29592	if( zAns==0 ) continue;
29593	k = 0;
29594	if( bVerbose>0 ){
29595	sputf(stdout, "%d: %s %s\n", tno, zOp, zSql);
29596	}
29597	if( cli_strcmp(zOp,"memo")==0 ){
29598	oputf("%s\n", zSql);
29599	}else
29600	if( cli_strcmp(zOp,"run")==0 ){
29601	char *zErrMsg = 0;
29602	str.n = 0;
29603	str.z[0] = 0;
29604	rc = sqlite3_exec(p->db, zSql, captureOutputCallback, &str, &zErrMsg);
29605	nTest++;
29606	if( bVerbose ){
29607	oputf("Result: %s\n", str.z);
29608	}
29609	if( rc \|\| zErrMsg ){
29610	nErr++;
29611	rc = 1;
29612	oputf("%d: error-code-%d: %s\n", tno, rc, zErrMsg);
29613	sqlite3_free(zErrMsg);
29614	}else if( cli_strcmp(zAns,str.z)!=0 ){
29615	nErr++;
29616	rc = 1;
29617	oputf("%d: Expected: [%s]\n", tno, zAns);
29618	oputf("%d: Got: [%s]\n", tno, str.z);
29619	}
29620	}
29621	else{
29622	eputf("Unknown operation \"%s\" on selftest line %d\n", zOp, tno);

29623	rc = 1;
29624	break;
29625	}
29626	} /* End loop over rows of content from SELFTEST */
29627	sqlite3_finalize(pStmt);
29628	} /* End loop over k */
29629	freeText(&str);
29630	oputf("%d errors out of %d tests\n", nErr, nTest);
29631	}else
29632
29633	if( c=='s' && cli_strncmp(azArg[0], "separator", n)==0 ){
29634	if( nArg<2 \|\| nArg>3 ){
29635	eputz("Usage: .separator COL ?ROW?\n");
	@@ -29673,11 +31097,12 @@
29673	}else
29674	if( cli_strcmp(z,"debug")==0 ){
29675	bDebug = 1;
29676	}else
29677	{
29678	eputf("Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);

29679	showHelp(p->out, azArg[0]);
29680	rc = 1;
29681	goto meta_command_exit;
29682	}
29683	}else if( zLike ){
	@@ -29751,11 +31176,11 @@
29751	}
29752	shell_check_oom(zSql);
29753	freeText(&sQuery);
29754	freeText(&sSql);
29755	if( bDebug ){
29756	oputf("%s\n", zSql);
29757	}else{
29758	shell_exec(p, zSql, 0);
29759	}
29760	#if !defined(SQLITE_OMIT_SCHEMA_PRAGMAS) && !defined(SQLITE_OMIT_VIRTUALTABLE)
29761	{
	@@ -29781,11 +31206,11 @@
29781	"\|\|group_concat('CAST(CAST('\|\|cname\|\|' AS BLOB) AS TEXT)<>'\|\|cname\n"
29782	"\|\| ' AND typeof('\|\|cname\|\|')=''text'' ',\n"
29783	"' OR ') as query, tname from tabcols group by tname)"
29784	, zRevText);
29785	shell_check_oom(zRevText);
29786	if( bDebug ) oputf("%s\n", zRevText);
29787	lrc = sqlite3_prepare_v2(p->db, zRevText, -1, &pStmt, 0);
29788	if( lrc!=SQLITE_OK ){
29789	/* assert(lrc==SQLITE_NOMEM); // might also be SQLITE_ERROR if the
29790	** user does cruel and unnatural things like ".limit expr_depth 0". */
29791	rc = 1;
	@@ -29794,19 +31219,20 @@
29794	lrc = SQLITE_ROW==sqlite3_step(pStmt);
29795	if( lrc ){
29796	const char zGenQuery = (char)sqlite3_column_text(pStmt,0);
29797	sqlite3_stmt *pCheckStmt;
29798	lrc = sqlite3_prepare_v2(p->db, zGenQuery, -1, &pCheckStmt, 0);
29799	if( bDebug ) oputf("%s\n", zGenQuery);
29800	if( lrc!=SQLITE_OK ){
29801	rc = 1;
29802	}else{
29803	if( SQLITE_ROW==sqlite3_step(pCheckStmt) ){
29804	double countIrreversible = sqlite3_column_double(pCheckStmt, 0);
29805	if( countIrreversible>0 ){
29806	int sz = (int)(countIrreversible + 0.5);
29807	eputf("Digest includes %d invalidly encoded text field%s.\n",

29808	sz, (sz>1)? "s": "");
29809	}
29810	}
29811	sqlite3_finalize(pCheckStmt);
29812	}
	@@ -29836,15 +31262,15 @@
29836	zCmd = sqlite3_mprintf(strchr(azArg[1],' ')==0?"%s":"\"%s\"", azArg[1]);
29837	for(i=2; i<nArg && zCmd!=0; i++){
29838	zCmd = sqlite3_mprintf(strchr(azArg[i],' ')==0?"%z %s":"%z \"%s\"",
29839	zCmd, azArg[i]);
29840	}
29841	consoleRestore();
29842	x = zCmd!=0 ? system(zCmd) : 1;
29843	consoleRenewSetup();
29844	sqlite3_free(zCmd);
29845	if( x ) eputf("System command returns %d\n", x);
29846	}else
29847	#endif /* !defined(SQLITE_NOHAVE_SYSTEM) && !defined(SQLITE_SHELL_FIDDLE) */
29848
29849	if( c=='s' && cli_strncmp(azArg[0], "show", n)==0 ){
29850	static const char *azBool[] = { "off", "on", "trigger", "full"};
	@@ -29853,50 +31279,52 @@
29853	if( nArg!=1 ){
29854	eputz("Usage: .show\n");
29855	rc = 1;
29856	goto meta_command_exit;
29857	}
29858	oputf("%12.12s: %s\n","echo",
29859	azBool[ShellHasFlag(p, SHFLG_Echo)]);
29860	oputf("%12.12s: %s\n","eqp", azBool[p->autoEQP&3]);
29861	oputf("%12.12s: %s\n","explain",
29862	p->mode==MODE_Explain ? "on" : p->autoExplain ? "auto" : "off");
29863	oputf("%12.12s: %s\n","headers", azBool[p->showHeader!=0]);

29864	if( p->mode==MODE_Column
29865	\|\| (p->mode>=MODE_Markdown && p->mode<=MODE_Box)
29866	){
29867	oputf("%12.12s: %s --wrap %d --wordwrap %s --%squote\n", "mode",

29868	modeDescr[p->mode], p->cmOpts.iWrap,
29869	p->cmOpts.bWordWrap ? "on" : "off",
29870	p->cmOpts.bQuote ? "" : "no");
29871	}else{
29872	oputf("%12.12s: %s\n","mode", modeDescr[p->mode]);
29873	}
29874	oputf("%12.12s: ", "nullvalue");
29875	output_c_string(p->nullValue);
29876	oputz("\n");
29877	oputf("%12.12s: %s\n","output",
29878	strlen30(p->outfile) ? p->outfile : "stdout");
29879	oputf("%12.12s: ", "colseparator");
29880	output_c_string(p->colSeparator);
29881	oputz("\n");
29882	oputf("%12.12s: ", "rowseparator");
29883	output_c_string(p->rowSeparator);
29884	oputz("\n");
29885	switch( p->statsOn ){
29886	case 0: zOut = "off"; break;
29887	default: zOut = "on"; break;
29888	case 2: zOut = "stmt"; break;
29889	case 3: zOut = "vmstep"; break;
29890	}
29891	oputf("%12.12s: %s\n","stats", zOut);
29892	oputf("%12.12s: ", "width");
29893	for (i=0;i<p->nWidth;i++) {
29894	oputf("%d ", p->colWidth[i]);
29895	}
29896	oputz("\n");
29897	oputf("%12.12s: %s\n", "filename",
29898	p->pAuxDb->zDbFilename ? p->pAuxDb->zDbFilename : "");
29899	}else
29900
29901	if( c=='s' && cli_strncmp(azArg[0], "stats", n)==0 ){
29902	if( nArg==2 ){
	@@ -30010,13 +31438,14 @@
30010	if( nPrintCol<1 ) nPrintCol = 1;
30011	nPrintRow = (nRow + nPrintCol - 1)/nPrintCol;
30012	for(i=0; i<nPrintRow; i++){
30013	for(j=i; j<nRow; j+=nPrintRow){
30014	char *zSp = j<nPrintRow ? "" : " ";
30015	oputf("%s%-*s", zSp, maxlen, azResult[j] ? azResult[j]:"");

30016	}
30017	oputz("\n");
30018	}
30019	}
30020
30021	for(ii=0; ii<nRow; ii++) sqlite3_free(azResult[ii]);
30022	sqlite3_free(azResult);
	@@ -30024,11 +31453,11 @@
30024
30025	#ifndef SQLITE_SHELL_FIDDLE
30026	/* Begin redirecting output to the file "testcase-out.txt" */
30027	if( c=='t' && cli_strcmp(azArg[0],"testcase")==0 ){
30028	output_reset(p);
30029	p->out = output_file_open("testcase-out.txt", 0);
30030	if( p->out==0 ){
30031	eputz("Error: cannot open 'testcase-out.txt'\n");
30032	}
30033	if( nArg>=2 ){
30034	sqlite3_snprintf(sizeof(p->zTestcase), p->zTestcase, "%s", azArg[1]);
	@@ -30068,11 +31497,10 @@
30068	{"prng_save", SQLITE_TESTCTRL_PRNG_SAVE, 0, "" },
30069	{"prng_seed", SQLITE_TESTCTRL_PRNG_SEED, 0, "SEED ?db?" },
30070	{"seek_count", SQLITE_TESTCTRL_SEEK_COUNT, 0, "" },
30071	{"sorter_mmap", SQLITE_TESTCTRL_SORTER_MMAP, 0, "NMAX" },
30072	{"tune", SQLITE_TESTCTRL_TUNE, 1, "ID VALUE" },
30073	{"uselongdouble", SQLITE_TESTCTRL_USELONGDOUBLE,0,"?BOOLEAN\|\"default\"?"},
30074	};
30075	int testctrl = -1;
30076	int iCtrl = -1;
30077	int rc2 = 0; /* 0: usage. 1: %d 2: %x 3: no-output */
30078	int isOk = 0;
	@@ -30088,14 +31516,14 @@
30088	if( zCmd[0]=='-' && zCmd[1] ) zCmd++;
30089	}
30090
30091	/* --help lists all test-controls */
30092	if( cli_strcmp(zCmd,"help")==0 ){
30093	oputz("Available test-controls:\n");
30094	for(i=0; i<ArraySize(aCtrl); i++){
30095	if( aCtrl[i].unSafe && !ShellHasFlag(p,SHFLG_TestingMode) ) continue;
30096	oputf(" .testctrl %s %s\n",
30097	aCtrl[i].zCtrlName, aCtrl[i].zUsage);
30098	}
30099	rc = 1;
30100	goto meta_command_exit;
30101	}
	@@ -30108,19 +31536,19 @@
30108	if( cli_strncmp(zCmd, aCtrl[i].zCtrlName, n2)==0 ){
30109	if( testctrl<0 ){
30110	testctrl = aCtrl[i].ctrlCode;
30111	iCtrl = i;
30112	}else{
30113	eputf("Error: ambiguous test-control: \"%s\"\n"
30114	"Use \".testctrl --help\" for help\n", zCmd);
30115	rc = 1;
30116	goto meta_command_exit;
30117	}
30118	}
30119	}
30120	if( testctrl<0 ){
30121	eputf("Error: unknown test-control: %s\n"
30122	"Use \".testctrl --help\" for help\n", zCmd);
30123	}else{
30124	switch(testctrl){
30125
30126	/* Special processing for .testctrl opt MASK ...
	@@ -30170,11 +31598,13 @@
30170	{ 0x10000000, 1, "OrderBySubq" },
30171	{ 0xffffffff, 0, "All" },
30172	};
30173	unsigned int curOpt;
30174	unsigned int newOpt;

30175	int ii;

30176	sqlite3_test_control(SQLITE_TESTCTRL_GETOPT, p->db, &curOpt);
30177	newOpt = curOpt;
30178	for(ii=2; ii<nArg; ii++){
30179	const char *z = azArg[ii];
30180	int useLabel = 0;
	@@ -30192,16 +31622,17 @@
30192	int jj;
30193	for(jj=0; jj<ArraySize(aLabel); jj++){
30194	if( sqlite3_stricmp(zLabel, aLabel[jj].zLabel)==0 ) break;
30195	}
30196	if( jj>=ArraySize(aLabel) ){
30197	eputf("Error: no such optimization: \"%s\"\n", zLabel);
30198	eputz("Should be one of:");

30199	for(jj=0; jj<ArraySize(aLabel); jj++){
30200	eputf(" %s", aLabel[jj].zLabel);
30201	}
30202	eputz("\n");
30203	rc = 1;
30204	goto meta_command_exit;
30205	}
30206	if( useLabel=='+' ){
30207	newOpt &= ~aLabel[jj].mask;
	@@ -30210,28 +31641,32 @@
30210	}
30211	}
30212	}
30213	if( curOpt!=newOpt ){
30214	sqlite3_test_control(SQLITE_TESTCTRL_OPTIMIZATIONS,p->db,newOpt);
30215	}else if( nArg<3 ){
30216	curOpt = ~newOpt;

30217	}
30218	if( newOpt==0 ){
30219	oputz("+All\n");
30220	}else if( newOpt==0xffffffff ){
30221	oputz("-All\n");




30222	}else{
30223	int jj;
30224	for(jj=0; jj<ArraySize(aLabel); jj++){
30225	unsigned int m = aLabel[jj].mask;
30226	if( !aLabel[jj].bDsply ) continue;
30227	if( (curOpt&m)!=(newOpt&m) ){
30228	oputf("%c%s\n", (newOpt & m)==0 ? '+' : '-',
30229	aLabel[jj].zLabel);
30230	}
30231	}
30232	}

30233	rc2 = isOk = 3;
30234	break;
30235	}
30236
30237	/* sqlite3_test_control(int, db, int) */
	@@ -30267,11 +31702,11 @@
30267	if( nArg==3 \|\| nArg==4 ){
30268	int ii = (int)integerValue(azArg[2]);
30269	sqlite3 *db;
30270	if( ii==0 && cli_strcmp(azArg[2],"random")==0 ){
30271	sqlite3_randomness(sizeof(ii),&ii);
30272	sputf(stdout, "-- random seed: %d\n", ii);
30273	}
30274	if( nArg==3 ){
30275	db = 0;
30276	}else{
30277	db = p->db;
	@@ -30301,25 +31736,10 @@
30301	rc2 = sqlite3_test_control(testctrl, opt);
30302	isOk = 3;
30303	}
30304	break;
30305
30306	/* sqlite3_test_control(int, int) */
30307	case SQLITE_TESTCTRL_USELONGDOUBLE: {
30308	int opt = -1;
30309	if( nArg==3 ){
30310	if( cli_strcmp(azArg[2],"default")==0 ){
30311	opt = 2;
30312	}else{
30313	opt = booleanValue(azArg[2]);
30314	}
30315	}
30316	rc2 = sqlite3_test_control(testctrl, opt);
30317	isOk = 1;
30318	break;
30319	}
30320
30321	/* sqlite3_test_control(sqlite3) /
30322	case SQLITE_TESTCTRL_INTERNAL_FUNCTIONS:
30323	rc2 = sqlite3_test_control(testctrl, p->db);
30324	isOk = 3;
30325	break;
	@@ -30335,11 +31755,11 @@
30335	break;
30336
30337	case SQLITE_TESTCTRL_SEEK_COUNT: {
30338	u64 x = 0;
30339	rc2 = sqlite3_test_control(testctrl, p->db, &x);
30340	oputf("%llu\n", x);
30341	isOk = 3;
30342	break;
30343	}
30344
30345	#ifdef YYCOVERAGE
	@@ -30366,15 +31786,15 @@
30366	int id = 1;
30367	while(1){
30368	int val = 0;
30369	rc2 = sqlite3_test_control(testctrl, -id, &val);
30370	if( rc2!=SQLITE_OK ) break;
30371	if( id>1 ) oputz(" ");
30372	oputf("%d: %d", id, val);
30373	id++;
30374	}
30375	if( id>1 ) oputz("\n");
30376	isOk = 3;
30377	}
30378	break;
30379	}
30380	#endif
	@@ -30412,18 +31832,26 @@
30412	}else if( cli_strcmp(z,"reset")==0 ){
30413	faultsim_state.iCnt = faultsim_state.nSkip;
30414	faultsim_state.nHit = 0;
30415	sqlite3_test_control(testctrl, faultsim_callback);
30416	}else if( cli_strcmp(z,"status")==0 ){
30417	oputf("faultsim.iId: %d\n", faultsim_state.iId);
30418	oputf("faultsim.iErr: %d\n", faultsim_state.iErr);
30419	oputf("faultsim.iCnt: %d\n", faultsim_state.iCnt);
30420	oputf("faultsim.nHit: %d\n", faultsim_state.nHit);
30421	oputf("faultsim.iInterval: %d\n", faultsim_state.iInterval);
30422	oputf("faultsim.eVerbose: %d\n", faultsim_state.eVerbose);
30423	oputf("faultsim.nRepeat: %d\n", faultsim_state.nRepeat);
30424	oputf("faultsim.nSkip: %d\n", faultsim_state.nSkip);








30425	}else if( cli_strcmp(z,"-v")==0 ){
30426	if( faultsim_state.eVerbose<2 ) faultsim_state.eVerbose++;
30427	}else if( cli_strcmp(z,"-q")==0 ){
30428	if( faultsim_state.eVerbose>0 ) faultsim_state.eVerbose--;
30429	}else if( cli_strcmp(z,"-id")==0 && kk+1<nArg ){
	@@ -30437,19 +31865,20 @@
30437	}else if( cli_strcmp(z,"-skip")==0 && kk+1<nArg ){
30438	faultsim_state.nSkip = atoi(azArg[++kk]);
30439	}else if( cli_strcmp(z,"-?")==0 \|\| sqlite3_strglob("help",z)==0){
30440	bShowHelp = 1;
30441	}else{
30442	eputf("Unrecognized fault_install argument: \"%s\"\n",

30443	azArg[kk]);
30444	rc = 1;
30445	bShowHelp = 1;
30446	break;
30447	}
30448	}
30449	if( bShowHelp ){
30450	oputz(
30451	"Usage: .testctrl fault_install ARGS\n"
30452	"Possible arguments:\n"
30453	" off Disable faultsim\n"
30454	" on Activate faultsim\n"
30455	" reset Reset the trigger counter\n"
	@@ -30459,23 +31888,25 @@
30459	" --errcode N When triggered, return N as error code\n"
30460	" --id ID Trigger only for the ID specified\n"
30461	" --interval N Trigger only after every N-th call\n"
30462	" --repeat N Turn off after N hits. 0 means never\n"
30463	" --skip N Skip the first N encounters\n"

30464	);
30465	}
30466	break;
30467	}
30468	}
30469	}
30470	if( isOk==0 && iCtrl>=0 ){
30471	oputf("Usage: .testctrl %s %s\n", zCmd,aCtrl[iCtrl].zUsage);

30472	rc = 1;
30473	}else if( isOk==1 ){
30474	oputf("%d\n", rc2);
30475	}else if( isOk==2 ){
30476	oputf("0x%08x\n", rc2);
30477	}
30478	}else
30479	#endif /* !defined(SQLITE_UNTESTABLE) */
30480
30481	if( c=='t' && n>4 && cli_strncmp(azArg[0], "timeout", n)==0 ){
	@@ -30526,17 +31957,17 @@
30526	}
30527	else if( optionMatch(z, "close") ){
30528	mType \|= SQLITE_TRACE_CLOSE;
30529	}
30530	else {
30531	eputf("Unknown option \"%s\" on \".trace\"\n", z);
30532	rc = 1;
30533	goto meta_command_exit;
30534	}
30535	}else{
30536	output_file_close(p->traceOut);
30537	p->traceOut = output_file_open(z, 0);
30538	}
30539	}
30540	if( p->traceOut==0 ){
30541	sqlite3_trace_v2(p->db, 0, 0, 0);
30542	}else{
	@@ -30569,103 +32000,40 @@
30569	}
30570	}
30571	}else
30572	#endif
30573
30574	#if SQLITE_USER_AUTHENTICATION
30575	if( c=='u' && cli_strncmp(azArg[0], "user", n)==0 ){
30576	if( nArg<2 ){
30577	eputz("Usage: .user SUBCOMMAND ...\n");
30578	rc = 1;
30579	goto meta_command_exit;
30580	}
30581	open_db(p, 0);
30582	if( cli_strcmp(azArg[1],"login")==0 ){
30583	if( nArg!=4 ){
30584	eputz("Usage: .user login USER PASSWORD\n");
30585	rc = 1;
30586	goto meta_command_exit;
30587	}
30588	rc = sqlite3_user_authenticate(p->db, azArg[2], azArg[3],
30589	strlen30(azArg[3]));
30590	if( rc ){
30591	eputf("Authentication failed for user %s\n", azArg[2]);
30592	rc = 1;
30593	}
30594	}else if( cli_strcmp(azArg[1],"add")==0 ){
30595	if( nArg!=5 ){
30596	eputz("Usage: .user add USER PASSWORD ISADMIN\n");
30597	rc = 1;
30598	goto meta_command_exit;
30599	}
30600	rc = sqlite3_user_add(p->db, azArg[2], azArg[3], strlen30(azArg[3]),
30601	booleanValue(azArg[4]));
30602	if( rc ){
30603	eputf("User-Add failed: %d\n", rc);
30604	rc = 1;
30605	}
30606	}else if( cli_strcmp(azArg[1],"edit")==0 ){
30607	if( nArg!=5 ){
30608	eputz("Usage: .user edit USER PASSWORD ISADMIN\n");
30609	rc = 1;
30610	goto meta_command_exit;
30611	}
30612	rc = sqlite3_user_change(p->db, azArg[2], azArg[3], strlen30(azArg[3]),
30613	booleanValue(azArg[4]));
30614	if( rc ){
30615	eputf("User-Edit failed: %d\n", rc);
30616	rc = 1;
30617	}
30618	}else if( cli_strcmp(azArg[1],"delete")==0 ){
30619	if( nArg!=3 ){
30620	eputz("Usage: .user delete USER\n");
30621	rc = 1;
30622	goto meta_command_exit;
30623	}
30624	rc = sqlite3_user_delete(p->db, azArg[2]);
30625	if( rc ){
30626	eputf("User-Delete failed: %d\n", rc);
30627	rc = 1;
30628	}
30629	}else{
30630	eputz("Usage: .user login\|add\|edit\|delete ...\n");
30631	rc = 1;
30632	goto meta_command_exit;
30633	}
30634	}else
30635	#endif /* SQLITE_USER_AUTHENTICATION */
30636
30637	if( c=='v' && cli_strncmp(azArg[0], "version", n)==0 ){
30638	char zPtrSz = sizeof(void)==8 ? "64-bit" : "32-bit";
30639	oputf("SQLite %s %s\n" /extra-version-info/,
30640	sqlite3_libversion(), sqlite3_sourceid());
30641	#if SQLITE_HAVE_ZLIB
30642	oputf("zlib version %s\n", zlibVersion());
30643	#endif
30644	#define CTIMEOPT_VAL_(opt) #opt
30645	#define CTIMEOPT_VAL(opt) CTIMEOPT_VAL_(opt)
30646	#if defined(__clang__) && defined(__clang_major__)
30647	oputf("clang-" CTIMEOPT_VAL(__clang_major__) "."
30648	CTIMEOPT_VAL(__clang_minor__) "."
30649	CTIMEOPT_VAL(__clang_patchlevel__) " (%s)\n", zPtrSz);
30650	#elif defined(_MSC_VER)
30651	oputf("msvc-" CTIMEOPT_VAL(_MSC_VER) " (%s)\n", zPtrSz);
30652	#elif defined(__GNUC__) && defined(__VERSION__)
30653	oputf("gcc-" __VERSION__ " (%s)\n", zPtrSz);
30654	#endif
30655	}else
30656
30657	if( c=='v' && cli_strncmp(azArg[0], "vfsinfo", n)==0 ){
30658	const char *zDbName = nArg==2 ? azArg[1] : "main";
30659	sqlite3_vfs *pVfs = 0;
30660	if( p->db ){
30661	sqlite3_file_control(p->db, zDbName, SQLITE_FCNTL_VFS_POINTER, &pVfs);
30662	if( pVfs ){
30663	oputf("vfs.zName = \"%s\"\n", pVfs->zName);
30664	oputf("vfs.iVersion = %d\n", pVfs->iVersion);
30665	oputf("vfs.szOsFile = %d\n", pVfs->szOsFile);
30666	oputf("vfs.mxPathname = %d\n", pVfs->mxPathname);
30667	}
30668	}
30669	}else
30670
30671	if( c=='v' && cli_strncmp(azArg[0], "vfslist", n)==0 ){
	@@ -30673,17 +32041,17 @@
30673	sqlite3_vfs *pCurrent = 0;
30674	if( p->db ){
30675	sqlite3_file_control(p->db, "main", SQLITE_FCNTL_VFS_POINTER, &pCurrent);
30676	}
30677	for(pVfs=sqlite3_vfs_find(0); pVfs; pVfs=pVfs->pNext){
30678	oputf("vfs.zName = \"%s\"%s\n", pVfs->zName,
30679	pVfs==pCurrent ? " <--- CURRENT" : "");
30680	oputf("vfs.iVersion = %d\n", pVfs->iVersion);
30681	oputf("vfs.szOsFile = %d\n", pVfs->szOsFile);
30682	oputf("vfs.mxPathname = %d\n", pVfs->mxPathname);
30683	if( pVfs->pNext ){
30684	oputz("-----------------------------------\n");
30685	}
30686	}
30687	}else
30688
30689	if( c=='v' && cli_strncmp(azArg[0], "vfsname", n)==0 ){
	@@ -30690,11 +32058,11 @@
30690	const char *zDbName = nArg==2 ? azArg[1] : "main";
30691	char *zVfsName = 0;
30692	if( p->db ){
30693	sqlite3_file_control(p->db, zDbName, SQLITE_FCNTL_VFSNAME, &zVfsName);
30694	if( zVfsName ){
30695	oputf("%s\n", zVfsName);
30696	sqlite3_free(zVfsName);
30697	}
30698	}
30699	}else
30700
	@@ -30714,11 +32082,11 @@
30714	p->colWidth[j-1] = (int)integerValue(azArg[j]);
30715	}
30716	}else
30717
30718	{
30719	eputf("Error: unknown command or invalid arguments: "
30720	" \"%s\". Enter \".help\" for help\n", azArg[0]);
30721	rc = 1;
30722	}
30723
30724	meta_command_exit:
	@@ -30755,11 +32123,10 @@
30755	SCAN_TRACKER_REFTYPE pst){
30756	char cin;
30757	char cWait = (char)qss; /* intentional narrowing loss */
30758	if( cWait==0 ){
30759	PlainScan:
30760	assert( cWait==0 );
30761	while( (cin = *zLine++)!=0 ){
30762	if( IsSpace(cin) )
30763	continue;
30764	switch (cin){
30765	case '-':
	@@ -30807,11 +32174,10 @@
30807	switch( cWait ){
30808	case '*':
30809	if( *zLine != '/' )
30810	continue;
30811	++zLine;
30812	cWait = 0;
30813	CONTINUE_PROMPT_AWAITC(pst, 0);
30814	qss = QSS_SETV(qss, 0);
30815	goto PlainScan;
30816	case '`': case '\'': case '"':
30817	if(*zLine==cWait){
	@@ -30819,11 +32185,10 @@
30819	++zLine;
30820	continue;
30821	}
30822	deliberate_fall_through;
30823	case ']':
30824	cWait = 0;
30825	CONTINUE_PROMPT_AWAITC(pst, 0);
30826	qss = QSS_SETV(qss, 0);
30827	goto PlainScan;
30828	default: assert(0);
30829	}
	@@ -30966,11 +32331,11 @@
30966	open_db(p, 0);
30967	if( ShellHasFlag(p,SHFLG_Backslash) ) resolve_backslashes(zSql);
30968	if( p->flgProgress & SHELL_PROGRESS_RESET ) p->nProgress = 0;
30969	BEGIN_TIMER;
30970	rc = shell_exec(p, zSql, &zErrMsg);
30971	END_TIMER;
30972	if( rc \|\| zErrMsg ){
30973	char zPrefix[100];
30974	const char *zErrorTail;
30975	const char *zErrorType;
30976	if( zErrMsg==0 ){
	@@ -30990,28 +32355,31 @@
30990	sqlite3_snprintf(sizeof(zPrefix), zPrefix,
30991	"%s near line %d:", zErrorType, startline);
30992	}else{
30993	sqlite3_snprintf(sizeof(zPrefix), zPrefix, "%s:", zErrorType);
30994	}
30995	eputf("%s %s\n", zPrefix, zErrorTail);
30996	sqlite3_free(zErrMsg);
30997	zErrMsg = 0;
30998	return 1;
30999	}else if( ShellHasFlag(p, SHFLG_CountChanges) ){
31000	char zLineBuf[2000];
31001	sqlite3_snprintf(sizeof(zLineBuf), zLineBuf,
31002	"changes: %lld total_changes: %lld",
31003	sqlite3_changes64(p->db), sqlite3_total_changes64(p->db));
31004	oputf("%s\n", zLineBuf);
31005	}
31006
31007	if( doAutoDetectRestore(p, zSql) ) return 1;
31008	return 0;
31009	}
31010
31011	static void echo_group_input(ShellState p, const char zDo){
31012	if( ShellHasFlag(p, SHFLG_Echo) ) oputf("%s\n", zDo);



31013	}
31014
31015	#ifdef SQLITE_SHELL_FIDDLE
31016	/*
31017	** Alternate one_input_line() impl for wasm mode. This is not in the primary
	@@ -31065,11 +32433,11 @@
31065	i64 startline = 0; /* Line number for start of current input */
31066	QuickScanState qss = QSS_Start; /* Accumulated line status (so far) */
31067
31068	if( p->inputNesting==MAX_INPUT_NESTING ){
31069	/* This will be more informative in a later version. */
31070	eputf("Input nesting limit (%d) reached at line %d."
31071	" Check recursion.\n", MAX_INPUT_NESTING, p->lineno);
31072	return 1;
31073	}
31074	++p->inputNesting;
31075	p->lineno = 0;
	@@ -31077,11 +32445,11 @@
31077	while( errCnt==0 \|\| !bail_on_error \|\| (p->in==0 && stdin_is_interactive) ){
31078	fflush(p->out);
31079	zLine = one_input_line(p->in, zLine, nSql>0);
31080	if( zLine==0 ){
31081	/* End of input */
31082	if( p->in==0 && stdin_is_interactive ) oputz("\n");
31083	break;
31084	}
31085	if( seenInterrupt ){
31086	if( p->in!=0 ) break;
31087	seenInterrupt = 0;
	@@ -31297,19 +32665,19 @@
31297	}
31298	zBuf = sqlite3_mprintf("%s/.sqliterc",home_dir);
31299	shell_check_oom(zBuf);
31300	sqliterc = zBuf;
31301	}
31302	p->in = fopen(sqliterc,"rb");
31303	if( p->in ){
31304	if( stdin_is_interactive ){
31305	eputf("-- Loading resources from %s\n", sqliterc);
31306	}
31307	if( process_input(p) && bail_on_error ) exit(1);
31308	fclose(p->in);
31309	}else if( sqliterc_override!=0 ){
31310	eputf("cannot open: \"%s\"\n", sqliterc);
31311	if( bail_on_error ) exit(1);
31312	}
31313	p->in = inSaved;
31314	p->lineno = savedLineno;
31315	sqlite3_free(zBuf);
	@@ -31374,23 +32742,21 @@
31374	" -table set output mode to 'table'\n"
31375	" -tabs set output mode to 'tabs'\n"
31376	" -unsafe-testing allow unsafe commands and modes for testing\n"
31377	" -version show SQLite version\n"
31378	" -vfs NAME use NAME as the default VFS\n"
31379	#ifdef SQLITE_ENABLE_VFSTRACE
31380	" -vfstrace enable tracing of all VFS calls\n"
31381	#endif
31382	#ifdef SQLITE_HAVE_ZLIB
31383	" -zip open the file as a ZIP Archive\n"
31384	#endif
31385	;
31386	static void usage(int showDetail){
31387	eputf("Usage: %s [OPTIONS] [FILENAME [SQL]]\n"
31388	"FILENAME is the name of an SQLite database. A new database is created\n"
31389	"if the file does not previously exist. Defaults to :memory:.\n", Argv0);
31390	if( showDetail ){
31391	eputf("OPTIONS include:\n%s", zOptions);
31392	}else{
31393	eputz("Use the -help option for additional information\n");
31394	}
31395	exit(0);
31396	}
	@@ -31411,10 +32777,13 @@
31411	*/
31412	static void main_init(ShellState *data) {
31413	memset(data, 0, sizeof(*data));
31414	data->normalMode = data->cMode = data->mode = MODE_List;
31415	data->autoExplain = 1;



31416	data->pAuxDb = &data->aAuxDb[0];
31417	memcpy(data->colSeparator,SEP_Column, 2);
31418	memcpy(data->rowSeparator,SEP_Row, 2);
31419	data->showHeader = 0;
31420	data->shellFlgs = SHFLG_Lookaside;
	@@ -31446,29 +32815,39 @@
31446	SetConsoleTextAttribute(out, defaultScreenInfo.wAttributes);
31447	#endif
31448	}
31449	#else
31450	static void printBold(const char *zText){
31451	sputf(stdout, "\033[1m%s\033[0m", zText);
31452	}
31453	#endif
31454
31455	/*
31456	** Get the argument to an --option. Throw an error and die if no argument
31457	** is available.
31458	*/
31459	static char cmdline_option_value(int argc, char *argv, int i){
31460	if( i==argc ){
31461	eputf("%s: Error: missing argument to %s\n", argv[0], argv[argc-1]);

31462	exit(1);
31463	}
31464	return argv[i];
31465	}
31466
31467	static void sayAbnormalExit(void){
31468	if( seenInterrupt ) eputz("Program interrupted.\n");
31469	}









31470
31471	#ifndef SQLITE_SHELL_IS_UTF8
31472	# if (defined(_WIN32) \|\| defined(WIN32)) \
31473	&& (defined(_MSC_VER) \|\| (defined(UNICODE) && defined(__GNUC__)))
31474	# define SQLITE_SHELL_IS_UTF8 (0)
	@@ -31493,19 +32872,19 @@
31493	char *zErrMsg = 0;
31494	#ifdef SQLITE_SHELL_FIDDLE
31495	# define data shellState
31496	#else
31497	ShellState data;
31498	StreamsAreConsole consStreams = SAC_NoConsole;
31499	#endif
31500	const char *zInitFile = 0;
31501	int i;
31502	int rc = 0;
31503	int warnInmemoryDb = 0;
31504	int readStdin = 1;
31505	int nCmd = 0;
31506	int nOptsEnd = argc;

31507	char **azCmd = 0;
31508	const char zVfs = 0; / Value of -vfs command-line option */
31509	#if !SQLITE_SHELL_IS_UTF8
31510	char **argvToFree = 0;
31511	int argcToFree = 0;
	@@ -31515,25 +32894,29 @@
31515	#ifdef SQLITE_SHELL_FIDDLE
31516	stdin_is_interactive = 0;
31517	stdout_is_console = 1;
31518	data.wasm.zDefaultDbName = "/fiddle.sqlite3";
31519	#else
31520	consStreams = consoleClassifySetup(stdin, stdout, stderr);
31521	stdin_is_interactive = (consStreams & SAC_InConsole)!=0;
31522	stdout_is_console = (consStreams & SAC_OutConsole)!=0;
31523	atexit(consoleRestore);
31524	#endif
31525	atexit(sayAbnormalExit);
31526	#ifdef SQLITE_DEBUG
31527	mem_main_enter = sqlite3_memory_used();
31528	#endif
31529	#if !defined(_WIN32_WCE)
31530	if( getenv("SQLITE_DEBUG_BREAK") ){
31531	if( isatty(0) && isatty(2) ){
31532	eputf("attach debugger to process %d and press any key to continue.\n",


31533	GETPID());
31534	fgetc(stdin);




31535	}else{
31536	#if defined(_WIN32) \|\| defined(WIN32)
31537	#if SQLITE_OS_WINRT
31538	__debugbreak();
31539	#else
	@@ -31554,11 +32937,12 @@
31554	}
31555	#endif
31556
31557	#if USE_SYSTEM_SQLITE+0!=1
31558	if( cli_strncmp(sqlite3_sourceid(),SQLITE_SOURCE_ID,60)!=0 ){
31559	eputf("SQLite header and source version mismatch\n%s\n%s\n",

31560	sqlite3_sourceid(), SQLITE_SOURCE_ID);
31561	exit(1);
31562	}
31563	#endif
31564	main_init(&data);
	@@ -31696,21 +33080,12 @@
31696	switch( n ){
31697	case 0: sqlite3_config(SQLITE_CONFIG_SINGLETHREAD); break;
31698	case 2: sqlite3_config(SQLITE_CONFIG_MULTITHREAD); break;
31699	default: sqlite3_config(SQLITE_CONFIG_SERIALIZED); break;
31700	}
31701	#ifdef SQLITE_ENABLE_VFSTRACE
31702	}else if( cli_strcmp(z,"-vfstrace")==0 ){
31703	extern int vfstrace_register(
31704	const char *zTraceName,
31705	const char *zOldVfsName,
31706	int (xOut)(const char,void*),
31707	void *pOutArg,
31708	int makeDefault
31709	);
31710	vfstrace_register("trace",0,(int()(const char,void*))fputs,stderr,1);
31711	#endif
31712	#ifdef SQLITE_ENABLE_MULTIPLEX
31713	}else if( cli_strcmp(z,"-multiplex")==0 ){
31714	extern int sqlite3_multiplex_initialize(const char*,int);
31715	sqlite3_multiplex_initialize(0, 1);
31716	#endif
	@@ -31762,11 +33137,11 @@
31762	}else if( cli_strcmp(z,"-safe")==0 ){
31763	/* no-op - catch this on the second pass */
31764	}
31765	}
31766	#ifndef SQLITE_SHELL_FIDDLE
31767	verify_uninitialized();
31768	#endif
31769
31770
31771	#ifdef SQLITE_SHELL_INIT_PROC
31772	{
	@@ -31786,25 +33161,29 @@
31786	if( zVfs ){
31787	sqlite3_vfs *pVfs = sqlite3_vfs_find(zVfs);
31788	if( pVfs ){
31789	sqlite3_vfs_register(pVfs, 1);
31790	}else{
31791	eputf("no such VFS: \"%s\"\n", zVfs);
31792	exit(1);
31793	}
31794	}
31795
31796	if( data.pAuxDb->zDbFilename==0 ){
31797	#ifndef SQLITE_OMIT_MEMORYDB
31798	data.pAuxDb->zDbFilename = ":memory:";
31799	warnInmemoryDb = argc==1;
31800	#else
31801	eputf("%s: Error: no database filename specified\n", Argv0);

31802	return 1;
31803	#endif
31804	}
31805	data.out = stdout;



31806	#ifndef SQLITE_SHELL_FIDDLE
31807	sqlite3_appendvfs_init(0,0,0);
31808	#endif
31809
31810	/* Go ahead and open the database file if it already exists. If the
	@@ -31913,11 +33292,11 @@
31913	*/
31914	ShellSetFlag(&data, SHFLG_Backslash);
31915	}else if( cli_strcmp(z,"-bail")==0 ){
31916	/* No-op. The bail_on_error flag should already be set. */
31917	}else if( cli_strcmp(z,"-version")==0 ){
31918	sputf(stdout, "%s %s (%d-bit)\n",
31919	sqlite3_libversion(), sqlite3_sourceid(), 8(int)sizeof(char));
31920	return 0;
31921	}else if( cli_strcmp(z,"-interactive")==0 ){
31922	/* Need to check for interactive override here to so that it can
31923	** affect console setup (for Windows only) and testing thereof.
	@@ -31951,14 +33330,12 @@
31951	}else if( cli_strcmp(z,"-sorterref")==0 ){
31952	i++;
31953	#endif
31954	}else if( cli_strcmp(z,"-vfs")==0 ){
31955	i++;
31956	#ifdef SQLITE_ENABLE_VFSTRACE
31957	}else if( cli_strcmp(z,"-vfstrace")==0 ){
31958	i++;
31959	#endif
31960	#ifdef SQLITE_ENABLE_MULTIPLEX
31961	}else if( cli_strcmp(z,"-multiplex")==0 ){
31962	i++;
31963	#endif
31964	}else if( cli_strcmp(z,"-help")==0 ){
	@@ -31978,18 +33355,18 @@
31978	rc = shell_exec(&data, z, &zErrMsg);
31979	if( zErrMsg!=0 ){
31980	shellEmitError(zErrMsg);
31981	if( bail_on_error ) return rc!=0 ? rc : 1;
31982	}else if( rc!=0 ){
31983	eputf("Error: unable to process SQL \"%s\"\n", z);
31984	if( bail_on_error ) return rc;
31985	}
31986	}
31987	#if !defined(SQLITE_OMIT_VIRTUALTABLE) && defined(SQLITE_HAVE_ZLIB)
31988	}else if( cli_strncmp(z, "-A", 2)==0 ){
31989	if( nCmd>0 ){
31990	eputf("Error: cannot mix regular SQL or dot-commands"
31991	" with \"%s\"\n", z);
31992	return 1;
31993	}
31994	open_db(&data, OPEN_DB_ZIPFILE);
31995	if( z[2] ){
	@@ -32004,11 +33381,11 @@
32004	}else if( cli_strcmp(z,"-safe")==0 ){
32005	data.bSafeMode = data.bSafeModePersist = 1;
32006	}else if( cli_strcmp(z,"-unsafe-testing")==0 ){
32007	/* Acted upon in first pass. */
32008	}else{
32009	eputf("%s: Error: unknown option: %s\n", Argv0, z);
32010	eputz("Use -help for a list of options.\n");
32011	return 1;
32012	}
32013	data.cMode = data.mode;
32014	}
	@@ -32031,11 +33408,12 @@
32031	rc = shell_exec(&data, azCmd[i], &zErrMsg);
32032	if( zErrMsg \|\| rc ){
32033	if( zErrMsg!=0 ){
32034	shellEmitError(zErrMsg);
32035	}else{
32036	eputf("Error: unable to process SQL: %s\n", azCmd[i]);

32037	}
32038	sqlite3_free(zErrMsg);
32039	if( rc==0 ) rc = 1;
32040	goto shell_main_exit;
32041	}
	@@ -32046,18 +33424,14 @@
32046	*/
32047	if( stdin_is_interactive ){
32048	char *zHome;
32049	char *zHistory;
32050	int nHistory;
32051	#if CIO_WIN_WC_XLATE
32052	# define SHELL_CIO_CHAR_SET (stdout_is_console? " (UTF-16 console I/O)" : "")
32053	#else
32054	# define SHELL_CIO_CHAR_SET ""
32055	#endif
32056	sputf(stdout, "SQLite version %s %.19s%s\n" /extra-version-info/
32057	"Enter \".help\" for usage hints.\n",
32058	sqlite3_libversion(), sqlite3_sourceid(), SHELL_CIO_CHAR_SET);
32059	if( warnInmemoryDb ){
32060	sputz(stdout, "Connected to a ");
32061	printBold("transient in-memory database");
32062	sputz(stdout, ".\nUse \".open FILENAME\" to reopen on a"
32063	" persistent database.\n");
	@@ -32070,13 +33444,15 @@
32070	if( (zHistory = malloc(nHistory))!=0 ){
32071	sqlite3_snprintf(nHistory, zHistory,"%s/.sqlite_history", zHome);
32072	}
32073	}
32074	if( zHistory ){ shell_read_history(zHistory); }
32075	#if HAVE_READLINE \|\| HAVE_EDITLINE
32076	rl_attempted_completion_function = readline_completion;
32077	#elif HAVE_LINENOISE


32078	linenoiseSetCompletionCallback(linenoise_completion, NULL);
32079	#endif
32080	data.in = 0;
32081	rc = process_input(&data);
32082	if( zHistory ){
	@@ -32122,13 +33498,16 @@
32122	free(data.colWidth);
32123	free(data.zNonce);
32124	/* Clear the global data structure so that valgrind will detect memory
32125	** leaks */
32126	memset(&data, 0, sizeof(data));



32127	#ifdef SQLITE_DEBUG
32128	if( sqlite3_memory_used()>mem_main_enter ){
32129	eputf("Memory leaked: %u bytes\n",
32130	(unsigned int)(sqlite3_memory_used()-mem_main_enter));
32131	}
32132	#endif
32133	#else /* SQLITE_SHELL_FIDDLE... */
32134	shell_main_exit:
	@@ -32164,11 +33543,11 @@
32164	return pVfs;
32165	}
32166
32167	/* Only for emcc experimentation purposes. */
32168	sqlite3 * fiddle_db_arg(sqlite3 *arg){
32169	oputf("fiddle_db_arg(%p)\n", (const void*)arg);
32170	return arg;
32171	}
32172
32173	/*
32174	** Intended to be called via a SharedWorker() while a separate
	@@ -32201,11 +33580,11 @@
32201	while( sqlite3_txn_state(globalDb,0)>0 ){
32202	/*
32203	** Resolve problem reported in
32204	** https://sqlite.org/forum/forumpost/0b41a25d65
32205	*/
32206	oputz("Rolling back in-progress transaction.\n");
32207	sqlite3_exec(globalDb,"ROLLBACK", 0, 0, 0);
32208	}
32209	rc = sqlite3_db_config(globalDb, SQLITE_DBCONFIG_RESET_DATABASE, 1, 0);
32210	if( 0==rc ) sqlite3_exec(globalDb, "VACUUM", 0, 0, 0);
32211	sqlite3_db_config(globalDb, SQLITE_DBCONFIG_RESET_DATABASE, 0, 0);
32212

	--- extsrc/shell.c
	+++ extsrc/shell.c
	@@ -120,13 +120,10 @@
120	#include <math.h>
121	#include "sqlite3.h"
122	typedef sqlite3_int64 i64;
123	typedef sqlite3_uint64 u64;
124	typedef unsigned char u8;



125	#include <ctype.h>
126	#include <stdarg.h>
127
128	#if !defined(_WIN32) && !defined(WIN32)
129	# include <signal.h>
	@@ -208,12 +205,10 @@
205	# define unlink _unlink
206	# endif
207	# ifndef strdup
208	# define strdup _strdup
209	# endif


210	# undef pclose
211	# define pclose _pclose
212	# endif
213	#else
214	/* Make sure isatty() has a prototype. */
	@@ -253,10 +248,370 @@
248	/* string conversion routines only needed on Win32 */
249	extern char *sqlite3_win32_unicode_to_utf8(LPCWSTR);
250	extern LPWSTR sqlite3_win32_utf8_to_unicode(const char *zText);
251	#endif
252
253	/*********************** Begin ../ext/misc/sqlite3_stdio.h ****************/
254	/*
255	** 2024-09-24
256	**
257	** The author disclaims copyright to this source code. In place of
258	** a legal notice, here is a blessing:
259	**
260	** May you do good and not evil.
261	** May you find forgiveness for yourself and forgive others.
262	** May you share freely, never taking more than you give.
263	**
264	*************************************************************************
265	**
266	** This header file contains definitions of interfaces that provide
267	** cross-platform I/O for UTF-8 content.
268	**
269	** On most platforms, the interfaces definitions in this file are
270	** just #defines. For example sqlite3_fopen() is a macro that resolves
271	** to the standard fopen() in the C-library.
272	**
273	** But Windows does not have a standard C-library, at least not one that
274	** can handle UTF-8. So for windows build, the interfaces resolve to new
275	** C-language routines contained in the separate sqlite3_stdio.c source file.
276	**
277	** So on all non-Windows platforms, simply #include this header file and
278	** use the interfaces defined herein. Then to run your application on Windows,
279	** also link in the accompanying sqlite3_stdio.c source file when compiling
280	** to get compatible interfaces.
281	*/
282	#ifndef _SQLITE3_STDIO_H_
283	#define _SQLITE3_STDIO_H_ 1
284	#ifdef _WIN32
285	/** Definitions For Windows **/
286	#include <stdio.h>
287	#include <windows.h>
288
289	FILE sqlite3_fopen(const char zFilename, const char *zMode);
290	FILE sqlite3_popen(const char zCommand, const char *type);
291	char sqlite3_fgets(char s, int size, FILE *stream);
292	int sqlite3_fputs(const char s, FILE stream);
293	int sqlite3_fprintf(FILE stream, const char format, ...);
294	void sqlite3_fsetmode(FILE *stream, int mode);
295
296
297	#else
298	/** Definitions For All Other Platforms **/
299	#include <stdio.h>
300	#define sqlite3_fopen fopen
301	#define sqlite3_popen popen
302	#define sqlite3_fgets fgets
303	#define sqlite3_fputs fputs
304	#define sqlite3_fprintf fprintf
305	#define sqlite3_fsetmode(F,X) /no-op/
306
307	#endif
308	#endif /* _SQLITE3_STDIO_H_ */
309
310	/*********************** End ../ext/misc/sqlite3_stdio.h ******************/
311	/*********************** Begin ../ext/misc/sqlite3_stdio.c ****************/
312	/*
313	** 2024-09-24
314	**
315	** The author disclaims copyright to this source code. In place of
316	** a legal notice, here is a blessing:
317	**
318	** May you do good and not evil.
319	** May you find forgiveness for yourself and forgive others.
320	** May you share freely, never taking more than you give.
321	**
322	*************************************************************************
323	**
324	** Implementation of standard I/O interfaces for UTF-8 that are missing
325	** on Windows.
326	*/
327	#ifdef _WIN32 /* This file is a no-op on all platforms except Windows */
328	#ifndef _SQLITE3_STDIO_H_
329	/* #include "sqlite3_stdio.h" */
330	#endif
331	#undef WIN32_LEAN_AND_MEAN
332	#define WIN32_LEAN_AND_MEAN
333	#include <windows.h>
334	#include <stdlib.h>
335	#include <string.h>
336	#include <stdio.h>
337	#include <assert.h>
338	/* #include "sqlite3.h" */
339	#include <ctype.h>
340	#include <stdarg.h>
341	#include <io.h>
342	#include <fcntl.h>
343
344	/*
345	** If the SQLITE_U8TEXT_ONLY option is defined, then use O_U8TEXT
346	** when appropriate on all output. (Sometimes use O_BINARY when
347	** rendering ASCII text in cases where NL-to-CRLF expansion would
348	** not be correct.)
349	**
350	** If the SQLITE_U8TEXT_STDIO option is defined, then use O_U8TEXT
351	** when appropriate when writing to stdout or stderr. Use O_BINARY
352	** or O_TEXT (depending on things like the .mode and the .crlf setting
353	** in the CLI, or other context clues in other applications) for all
354	** other output channels.
355	**
356	** The default behavior, if neither of the above is defined is to
357	** use O_U8TEXT when writing to the Windows console (or anything
358	** else for which _isatty() returns true) and to use O_BINARY or O_TEXT
359	** for all other output channels.
360	*/
361	#if defined(SQLITE_U8TEXT_ONLY)
362	# define UseWtextForOutput(fd) 1
363	# define UseWtextForInput(fd) 1
364	# define IsConsole(fd) _isatty(_fileno(fd))
365	#elif defined(SQLITE_U8TEXT_STDIO)
366	# define UseWtextForOutput(fd) ((fd)==stdout \|\| (fd)==stderr)
367	# define UseWtextForInput(fd) ((fd)==stdin)
368	# define IsConsole(fd) _isatty(_fileno(fd))
369	#else
370	# define UseWtextForOutput(fd) _isatty(_fileno(fd))
371	# define UseWtextForInput(fd) _isatty(_fileno(fd))
372	# define IsConsole(fd) 1
373	#endif
374
375	/*
376	** Global variables determine if simulated O_BINARY mode is to be
377	** used for stdout or other, respectively. Simulated O_BINARY mode
378	** means the mode is usually O_BINARY, but switches to O_U8TEXT for
379	** unicode characters U+0080 or greater (any character that has a
380	** multi-byte representation in UTF-8). This is the only way we
381	** have found to render Unicode characters on a Windows console while
382	** at the same time avoiding undesirable \n to \r\n translation.
383	*/
384	static int simBinaryStdout = 0;
385	static int simBinaryOther = 0;
386
387
388	/*
389	** Determine if simulated binary mode should be used for output to fd
390	*/
391	static int UseBinaryWText(FILE *fd){
392	if( fd==stdout \|\| fd==stderr ){
393	return simBinaryStdout;
394	}else{
395	return simBinaryOther;
396	}
397	}
398
399
400	/*
401	** Work-alike for the fopen() routine from the standard C library.
402	*/
403	FILE sqlite3_fopen(const char zFilename, const char *zMode){
404	FILE *fp = 0;
405	wchar_t b1, b2;
406	int sz1, sz2;
407
408	sz1 = (int)strlen(zFilename);
409	sz2 = (int)strlen(zMode);
410	b1 = sqlite3_malloc( (sz1+1)*sizeof(b1[0]) );
411	b2 = sqlite3_malloc( (sz2+1)*sizeof(b1[0]) );
412	if( b1 && b2 ){
413	sz1 = MultiByteToWideChar(CP_UTF8, 0, zFilename, sz1, b1, sz1);
414	b1[sz1] = 0;
415	sz2 = MultiByteToWideChar(CP_UTF8, 0, zMode, sz2, b2, sz2);
416	b2[sz2] = 0;
417	fp = _wfopen(b1, b2);
418	}
419	sqlite3_free(b1);
420	sqlite3_free(b2);
421	simBinaryOther = 0;
422	return fp;
423	}
424
425
426	/*
427	** Work-alike for the popen() routine from the standard C library.
428	*/
429	FILE sqlite3_popen(const char zCommand, const char *zMode){
430	FILE *fp = 0;
431	wchar_t b1, b2;
432	int sz1, sz2;
433
434	sz1 = (int)strlen(zCommand);
435	sz2 = (int)strlen(zMode);
436	b1 = sqlite3_malloc( (sz1+1)*sizeof(b1[0]) );
437	b2 = sqlite3_malloc( (sz2+1)*sizeof(b1[0]) );
438	if( b1 && b2 ){
439	sz1 = MultiByteToWideChar(CP_UTF8, 0, zCommand, sz1, b1, sz1);
440	b1[sz1] = 0;
441	sz2 = MultiByteToWideChar(CP_UTF8, 0, zMode, sz2, b2, sz2);
442	b2[sz2] = 0;
443	fp = _wpopen(b1, b2);
444	}
445	sqlite3_free(b1);
446	sqlite3_free(b2);
447	return fp;
448	}
449
450	/*
451	** Work-alike for fgets() from the standard C library.
452	*/
453	char sqlite3_fgets(char buf, int sz, FILE *in){
454	if( UseWtextForInput(in) ){
455	/* When reading from the command-prompt in Windows, it is necessary
456	** to use _O_WTEXT input mode to read UTF-16 characters, then translate
457	** that into UTF-8. Otherwise, non-ASCII characters all get translated
458	** into '?'.
459	*/
460	wchar_t b1 = sqlite3_malloc( szsizeof(wchar_t) );
461	if( b1==0 ) return 0;
462	#ifndef SQLITE_USE_STDIO_FOR_CONSOLE
463	DWORD nRead = 0;
464	if( IsConsole(in)
465	&& ReadConsoleW(GetStdHandle(STD_INPUT_HANDLE), b1, sz, &nRead, 0)
466	){
467	b1[nRead] = 0;
468	}else
469	#endif
470	{
471	_setmode(_fileno(in), IsConsole(in) ? _O_WTEXT : _O_U8TEXT);
472	if( fgetws(b1, sz/4, in)==0 ){
473	sqlite3_free(b1);
474	return 0;
475	}
476	}
477	WideCharToMultiByte(CP_UTF8, 0, b1, -1, buf, sz, 0, 0);
478	sqlite3_free(b1);
479	return buf;
480	}else{
481	/* Reading from a file or other input source, just read bytes without
482	** any translation. */
483	return fgets(buf, sz, in);
484	}
485	}
486
487	/*
488	** Send ASCII text as O_BINARY. But for Unicode characters U+0080 and
489	** greater, switch to O_U8TEXT.
490	*/
491	static void piecemealOutput(wchar_t b1, int sz, FILE out){
492	int i;
493	wchar_t c;
494	while( sz>0 ){
495	for(i=0; i<sz && b1[i]>=0x80; i++){}
496	if( i>0 ){
497	c = b1[i];
498	b1[i] = 0;
499	fflush(out);
500	_setmode(_fileno(out), _O_U8TEXT);
501	fputws(b1, out);
502	fflush(out);
503	b1 += i;
504	b1[0] = c;
505	sz -= i;
506	}else{
507	fflush(out);
508	_setmode(_fileno(out), _O_TEXT);
509	_setmode(_fileno(out), _O_BINARY);
510	fwrite(&b1[0], 1, 1, out);
511	for(i=1; i<sz && b1[i]<0x80; i++){
512	fwrite(&b1[i], 1, 1, out);
513	}
514	fflush(out);
515	_setmode(_fileno(out), _O_U8TEXT);
516	b1 += i;
517	sz -= i;
518	}
519	}
520	}
521
522	/*
523	** Work-alike for fputs() from the standard C library.
524	*/
525	int sqlite3_fputs(const char z, FILE out){
526	if( !UseWtextForOutput(out) ){
527	/* Writing to a file or other destination, just write bytes without
528	** any translation. */
529	return fputs(z, out);
530	}else{
531	/* One must use UTF16 in order to get unicode support when writing
532	** to the console on Windows.
533	*/
534	int sz = (int)strlen(z);
535	wchar_t b1 = sqlite3_malloc( (sz+1)sizeof(wchar_t) );
536	if( b1==0 ) return 0;
537	sz = MultiByteToWideChar(CP_UTF8, 0, z, sz, b1, sz);
538	b1[sz] = 0;
539
540	#ifndef SQLITE_STDIO_FOR_CONSOLE
541	DWORD nWr = 0;
542	if( IsConsole(out)
543	&& WriteConsoleW(GetStdHandle(STD_OUTPUT_HANDLE),b1,sz,&nWr,0)
544	){
545	/* If writing to the console, then the WriteConsoleW() is all we
546	** need to do. */
547	}else
548	#endif
549	{
550	/* For non-console I/O, or if SQLITE_USE_STDIO_FOR_CONSOLE is defined
551	** then write using the standard library. */
552	_setmode(_fileno(out), _O_U8TEXT);
553	if( UseBinaryWText(out) ){
554	piecemealOutput(b1, sz, out);
555	}else{
556	fputws(b1, out);
557	}
558	}
559	sqlite3_free(b1);
560	return 0;
561	}
562	}
563
564
565	/*
566	** Work-alike for fprintf() from the standard C library.
567	*/
568	int sqlite3_fprintf(FILE out, const char zFormat, ...){
569	int rc;
570	if( UseWtextForOutput(out) ){
571	/* When writing to the command-prompt in Windows, it is necessary
572	** to use _O_WTEXT input mode and write UTF-16 characters.
573	*/
574	char *z;
575	va_list ap;
576
577	va_start(ap, zFormat);
578	z = sqlite3_vmprintf(zFormat, ap);
579	va_end(ap);
580	sqlite3_fputs(z, out);
581	rc = (int)strlen(z);
582	sqlite3_free(z);
583	}else{
584	/* Writing to a file or other destination, just write bytes without
585	** any translation. */
586	va_list ap;
587	va_start(ap, zFormat);
588	rc = vfprintf(out, zFormat, ap);
589	va_end(ap);
590	}
591	return rc;
592	}
593
594	/*
595	** Set the mode for an output stream. mode argument is typically _O_BINARY or
596	** _O_TEXT.
597	*/
598	void sqlite3_fsetmode(FILE *fp, int mode){
599	if( !UseWtextForOutput(fp) ){
600	fflush(fp);
601	_setmode(_fileno(fp), mode);
602	}else if( fp==stdout \|\| fp==stderr ){
603	simBinaryStdout = (mode==_O_BINARY);
604	}else{
605	simBinaryOther = (mode==_O_BINARY);
606	}
607	}
608
609	#endif /* defined(_WIN32) */
610
611	/*********************** End ../ext/misc/sqlite3_stdio.c ******************/
612
613	/* Use console I/O package as a direct INCLUDE. */
614	#define SQLITE_INTERNAL_LINKAGE static
615
616	#ifdef SQLITE_SHELL_FIDDLE
617	/* Deselect most features from the console I/O package for Fiddle. */
	@@ -264,1118 +619,13 @@
619	# define SQLITE_CIO_NO_CLASSIFY
620	# define SQLITE_CIO_NO_TRANSLATE
621	# define SQLITE_CIO_NO_SETMODE
622	# define SQLITE_CIO_NO_FLUSH
623	#endif
624
625	#define eputz(z) sqlite3_fputs(z,stderr)
626	#define sputz(fp,z) sqlite3_fputs(z,fp)

















































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































627
628	/* True if the timer is enabled */
629	static int enableTimer = 0;
630
631	/* A version of strcmp() that works with NULL values */
	@@ -1416,10 +666,11 @@
666	struct timeval ru_utime; /* user CPU time used */
667	struct timeval ru_stime; /* system CPU time used */
668	};
669	#define getrusage(A,B) memset(B,0,sizeof(*B))
670	#endif
671
672
673	/* Saved resource information for the beginning of an operation */
674	static struct rusage sBegin; /* CPU time at start */
675	static sqlite3_int64 iBegin; /* Wall-clock time at start */
676
	@@ -1440,24 +691,24 @@
691	}
692
693	/*
694	** Print the timing results.
695	*/
696	static void endTimer(FILE *out){
697	if( enableTimer ){
698	sqlite3_int64 iEnd = timeOfDay();
699	struct rusage sEnd;
700	getrusage(RUSAGE_SELF, &sEnd);
701	sqlite3_fprintf(out, "Run Time: real %.3f user %f sys %f\n",
702	(iEnd - iBegin)*0.001,
703	timeDiff(&sBegin.ru_utime, &sEnd.ru_utime),
704	timeDiff(&sBegin.ru_stime, &sEnd.ru_stime));
705	}
706	}
707
708	#define BEGIN_TIMER beginTimer()
709	#define END_TIMER(X) endTimer(X)
710	#define HAS_TIMER 1
711
712	#elif (defined(_WIN32) \|\| defined(WIN32))
713
714	/* Saved resource information for the beginning of an operation */
	@@ -1519,29 +770,29 @@
770	}
771
772	/*
773	** Print the timing results.
774	*/
775	static void endTimer(FILE *out){
776	if( enableTimer && getProcessTimesAddr){
777	FILETIME ftCreation, ftExit, ftKernelEnd, ftUserEnd;
778	sqlite3_int64 ftWallEnd = timeOfDay();
779	getProcessTimesAddr(hProcess,&ftCreation,&ftExit,&ftKernelEnd,&ftUserEnd);
780	sqlite3_fprintf(out, "Run Time: real %.3f user %f sys %f\n",
781	(ftWallEnd - ftWallBegin)*0.001,
782	timeDiff(&ftUserBegin, &ftUserEnd),
783	timeDiff(&ftKernelBegin, &ftKernelEnd));
784	}
785	}
786
787	#define BEGIN_TIMER beginTimer()
788	#define END_TIMER(X) endTimer(X)
789	#define HAS_TIMER hasTimer()
790
791	#else
792	#define BEGIN_TIMER
793	#define END_TIMER(X) /no-op/
794	#define HAS_TIMER 0
795	#endif
796
797	/*
798	** Used to prevent warnings about unused parameters
	@@ -1738,41 +989,216 @@
989	char *z;
990	if( iotrace==0 ) return;
991	va_start(ap, zFormat);
992	z = sqlite3_vmprintf(zFormat, ap);
993	va_end(ap);
994	sqlite3_fprintf(iotrace, "%s", z);
995	sqlite3_free(z);
996	}
997	#endif
998
999	/* Lookup table to estimate the number of columns consumed by a Unicode
1000	** character.
1001	*/
1002	static const struct {
1003	unsigned char w; /* Width of the character in columns */
1004	int iFirst; /* First character in a span having this width */
1005	} aUWidth[] = {
1006	/* {1, 0x00000}, */
1007	{0, 0x00300}, {1, 0x00370}, {0, 0x00483}, {1, 0x00487}, {0, 0x00488},
1008	{1, 0x0048a}, {0, 0x00591}, {1, 0x005be}, {0, 0x005bf}, {1, 0x005c0},
1009	{0, 0x005c1}, {1, 0x005c3}, {0, 0x005c4}, {1, 0x005c6}, {0, 0x005c7},
1010	{1, 0x005c8}, {0, 0x00600}, {1, 0x00604}, {0, 0x00610}, {1, 0x00616},
1011	{0, 0x0064b}, {1, 0x0065f}, {0, 0x00670}, {1, 0x00671}, {0, 0x006d6},
1012	{1, 0x006e5}, {0, 0x006e7}, {1, 0x006e9}, {0, 0x006ea}, {1, 0x006ee},
1013	{0, 0x0070f}, {1, 0x00710}, {0, 0x00711}, {1, 0x00712}, {0, 0x00730},
1014	{1, 0x0074b}, {0, 0x007a6}, {1, 0x007b1}, {0, 0x007eb}, {1, 0x007f4},
1015	{0, 0x00901}, {1, 0x00903}, {0, 0x0093c}, {1, 0x0093d}, {0, 0x00941},
1016	{1, 0x00949}, {0, 0x0094d}, {1, 0x0094e}, {0, 0x00951}, {1, 0x00955},
1017	{0, 0x00962}, {1, 0x00964}, {0, 0x00981}, {1, 0x00982}, {0, 0x009bc},
1018	{1, 0x009bd}, {0, 0x009c1}, {1, 0x009c5}, {0, 0x009cd}, {1, 0x009ce},
1019	{0, 0x009e2}, {1, 0x009e4}, {0, 0x00a01}, {1, 0x00a03}, {0, 0x00a3c},
1020	{1, 0x00a3d}, {0, 0x00a41}, {1, 0x00a43}, {0, 0x00a47}, {1, 0x00a49},
1021	{0, 0x00a4b}, {1, 0x00a4e}, {0, 0x00a70}, {1, 0x00a72}, {0, 0x00a81},
1022	{1, 0x00a83}, {0, 0x00abc}, {1, 0x00abd}, {0, 0x00ac1}, {1, 0x00ac6},
1023	{0, 0x00ac7}, {1, 0x00ac9}, {0, 0x00acd}, {1, 0x00ace}, {0, 0x00ae2},
1024	{1, 0x00ae4}, {0, 0x00b01}, {1, 0x00b02}, {0, 0x00b3c}, {1, 0x00b3d},
1025	{0, 0x00b3f}, {1, 0x00b40}, {0, 0x00b41}, {1, 0x00b44}, {0, 0x00b4d},
1026	{1, 0x00b4e}, {0, 0x00b56}, {1, 0x00b57}, {0, 0x00b82}, {1, 0x00b83},
1027	{0, 0x00bc0}, {1, 0x00bc1}, {0, 0x00bcd}, {1, 0x00bce}, {0, 0x00c3e},
1028	{1, 0x00c41}, {0, 0x00c46}, {1, 0x00c49}, {0, 0x00c4a}, {1, 0x00c4e},
1029	{0, 0x00c55}, {1, 0x00c57}, {0, 0x00cbc}, {1, 0x00cbd}, {0, 0x00cbf},
1030	{1, 0x00cc0}, {0, 0x00cc6}, {1, 0x00cc7}, {0, 0x00ccc}, {1, 0x00cce},
1031	{0, 0x00ce2}, {1, 0x00ce4}, {0, 0x00d41}, {1, 0x00d44}, {0, 0x00d4d},
1032	{1, 0x00d4e}, {0, 0x00dca}, {1, 0x00dcb}, {0, 0x00dd2}, {1, 0x00dd5},
1033	{0, 0x00dd6}, {1, 0x00dd7}, {0, 0x00e31}, {1, 0x00e32}, {0, 0x00e34},
1034	{1, 0x00e3b}, {0, 0x00e47}, {1, 0x00e4f}, {0, 0x00eb1}, {1, 0x00eb2},
1035	{0, 0x00eb4}, {1, 0x00eba}, {0, 0x00ebb}, {1, 0x00ebd}, {0, 0x00ec8},
1036	{1, 0x00ece}, {0, 0x00f18}, {1, 0x00f1a}, {0, 0x00f35}, {1, 0x00f36},
1037	{0, 0x00f37}, {1, 0x00f38}, {0, 0x00f39}, {1, 0x00f3a}, {0, 0x00f71},
1038	{1, 0x00f7f}, {0, 0x00f80}, {1, 0x00f85}, {0, 0x00f86}, {1, 0x00f88},
1039	{0, 0x00f90}, {1, 0x00f98}, {0, 0x00f99}, {1, 0x00fbd}, {0, 0x00fc6},
1040	{1, 0x00fc7}, {0, 0x0102d}, {1, 0x01031}, {0, 0x01032}, {1, 0x01033},
1041	{0, 0x01036}, {1, 0x01038}, {0, 0x01039}, {1, 0x0103a}, {0, 0x01058},
1042	{1, 0x0105a}, {2, 0x01100}, {0, 0x01160}, {1, 0x01200}, {0, 0x0135f},
1043	{1, 0x01360}, {0, 0x01712}, {1, 0x01715}, {0, 0x01732}, {1, 0x01735},
1044	{0, 0x01752}, {1, 0x01754}, {0, 0x01772}, {1, 0x01774}, {0, 0x017b4},
1045	{1, 0x017b6}, {0, 0x017b7}, {1, 0x017be}, {0, 0x017c6}, {1, 0x017c7},
1046	{0, 0x017c9}, {1, 0x017d4}, {0, 0x017dd}, {1, 0x017de}, {0, 0x0180b},
1047	{1, 0x0180e}, {0, 0x018a9}, {1, 0x018aa}, {0, 0x01920}, {1, 0x01923},
1048	{0, 0x01927}, {1, 0x01929}, {0, 0x01932}, {1, 0x01933}, {0, 0x01939},
1049	{1, 0x0193c}, {0, 0x01a17}, {1, 0x01a19}, {0, 0x01b00}, {1, 0x01b04},
1050	{0, 0x01b34}, {1, 0x01b35}, {0, 0x01b36}, {1, 0x01b3b}, {0, 0x01b3c},
1051	{1, 0x01b3d}, {0, 0x01b42}, {1, 0x01b43}, {0, 0x01b6b}, {1, 0x01b74},
1052	{0, 0x01dc0}, {1, 0x01dcb}, {0, 0x01dfe}, {1, 0x01e00}, {0, 0x0200b},
1053	{1, 0x02010}, {0, 0x0202a}, {1, 0x0202f}, {0, 0x02060}, {1, 0x02064},
1054	{0, 0x0206a}, {1, 0x02070}, {0, 0x020d0}, {1, 0x020f0}, {2, 0x02329},
1055	{1, 0x0232b}, {2, 0x02e80}, {0, 0x0302a}, {2, 0x03030}, {1, 0x0303f},
1056	{2, 0x03040}, {0, 0x03099}, {2, 0x0309b}, {1, 0x0a4d0}, {0, 0x0a806},
1057	{1, 0x0a807}, {0, 0x0a80b}, {1, 0x0a80c}, {0, 0x0a825}, {1, 0x0a827},
1058	{2, 0x0ac00}, {1, 0x0d7a4}, {2, 0x0f900}, {1, 0x0fb00}, {0, 0x0fb1e},
1059	{1, 0x0fb1f}, {0, 0x0fe00}, {2, 0x0fe10}, {1, 0x0fe1a}, {0, 0x0fe20},
1060	{1, 0x0fe24}, {2, 0x0fe30}, {1, 0x0fe70}, {0, 0x0feff}, {2, 0x0ff00},
1061	{1, 0x0ff61}, {2, 0x0ffe0}, {1, 0x0ffe7}, {0, 0x0fff9}, {1, 0x0fffc},
1062	{0, 0x10a01}, {1, 0x10a04}, {0, 0x10a05}, {1, 0x10a07}, {0, 0x10a0c},
1063	{1, 0x10a10}, {0, 0x10a38}, {1, 0x10a3b}, {0, 0x10a3f}, {1, 0x10a40},
1064	{0, 0x1d167}, {1, 0x1d16a}, {0, 0x1d173}, {1, 0x1d183}, {0, 0x1d185},
1065	{1, 0x1d18c}, {0, 0x1d1aa}, {1, 0x1d1ae}, {0, 0x1d242}, {1, 0x1d245},
1066	{2, 0x20000}, {1, 0x2fffe}, {2, 0x30000}, {1, 0x3fffe}, {0, 0xe0001},
1067	{1, 0xe0002}, {0, 0xe0020}, {1, 0xe0080}, {0, 0xe0100}, {1, 0xe01f0}
1068	};
1069
1070	/*
1071	** Return an estimate of the width, in columns, for the single Unicode
1072	** character c. For normal characters, the answer is always 1. But the
1073	** estimate might be 0 or 2 for zero-width and double-width characters.
1074	**
1075	** Different display devices display unicode using different widths. So
1076	** it is impossible to know that true display width with 100% accuracy.
1077	** Inaccuracies in the width estimates might cause columns to be misaligned.
1078	** Unfortunately, there is nothing we can do about that.
1079	*/
1080	int cli_wcwidth(int c){
1081	int iFirst, iLast;
1082
1083	/* Fast path for common characters */
1084	if( c<=0x300 ) return 1;
1085
1086	/* The general case */
1087	iFirst = 0;
1088	iLast = sizeof(aUWidth)/sizeof(aUWidth[0]) - 1;
1089	while( iFirst<iLast-1 ){
1090	int iMid = (iFirst+iLast)/2;
1091	int cMid = aUWidth[iMid].iFirst;
1092	if( cMid < c ){
1093	iFirst = iMid;
1094	}else if( cMid > c ){
1095	iLast = iMid - 1;
1096	}else{
1097	return aUWidth[iMid].w;
1098	}
1099	}
1100	if( aUWidth[iLast].iFirst > c ) return aUWidth[iFirst].w;
1101	return aUWidth[iLast].w;
1102	}
1103
1104	/*
1105	** Compute the value and length of a multi-byte UTF-8 character that
1106	** begins at z[0]. Return the length. Write the Unicode value into *pU.
1107	**
1108	** This routine only works for multi-byte UTF-8 characters.
1109	*/
1110	static int decodeUtf8(const unsigned char z, int pU){
1111	if( (z[0] & 0xe0)==0xc0 && (z[1] & 0xc0)==0x80 ){
1112	*pU = ((z[0] & 0x1f)<<6) \| (z[1] & 0x3f);
1113	return 2;
1114	}
1115	if( (z[0] & 0xf0)==0xe0 && (z[1] & 0xc0)==0x80 && (z[2] & 0xc0)==0x80 ){
1116	*pU = ((z[0] & 0x0f)<<12) \| ((z[1] & 0x3f)<<6) \| (z[2] & 0x3f);
1117	return 3;
1118	}
1119	if( (z[0] & 0xf8)==0xf0 && (z[1] & 0xc0)==0x80 && (z[2] & 0xc0)==0x80
1120	&& (z[3] & 0xc0)==0x80
1121	){
1122	*pU = ((z[0] & 0x0f)<<18) \| ((z[1] & 0x3f)<<12) \| ((z[2] & 0x3f))<<6
1123	\| (z[4] & 0x3f);
1124	return 4;
1125	}
1126	*pU = 0;
1127	return 1;
1128	}
1129
1130
1131	#if 0 /* NOT USED */
1132	/*
1133	** Return the width, in display columns, of a UTF-8 string.
1134	**
1135	** Each normal character counts as 1. Zero-width characters count
1136	** as zero, and double-width characters count as 2.
1137	*/
1138	int cli_wcswidth(const char *z){
1139	const unsigned char a = (const unsigned char)z;
1140	int n = 0;
1141	int i = 0;
1142	unsigned char c;
1143	while( (c = a[i])!=0 ){
1144	if( c>=0xc0 ){
1145	int u;
1146	int len = decodeUtf8(&a[i], &u);
1147	i += len;
1148	n += cli_wcwidth(u);
1149	}else if( c>=' ' ){
1150	n++;
1151	i++;
1152	}else{
1153	i++;
1154	}
1155	}
1156	return n;
1157	}
1158	#endif
1159
1160	/*
1161	** Output string zUtf to stdout as w characters. If w is negative,
1162	** then right-justify the text. W is the width in UTF-8 characters, not
1163	** in bytes. This is different from the %.s specification in printf
1164	** since with %.s the width is measured in bytes, not characters.
1165	**
1166	** Take into account zero-width and double-width Unicode characters.
1167	** In other words, a zero-width character does not count toward the
1168	** the w limit. A double-width character counts as two.
1169	*/
1170	static void utf8_width_print(FILE out, int w, const char zUtf){
1171	const unsigned char a = (const unsigned char)zUtf;
1172	unsigned char c;
1173	int i = 0;
1174	int n = 0;
1175	int aw = w<0 ? -w : w;
1176	if( zUtf==0 ) zUtf = "";
1177	while( (c = a[i])!=0 ){
1178	if( (c&0xc0)==0xc0 ){
1179	int u;
1180	int len = decodeUtf8(a+i, &u);
1181	int x = cli_wcwidth(u);
1182	if( x+n>aw ){
1183	break;
1184	}
1185	i += len;
1186	n += x;
1187	}else if( n>=aw ){
1188	break;
1189	}else{
1190	n++;
1191	i++;



1192	}
1193	}
1194	if( n>=aw ){
1195	sqlite3_fprintf(out, "%.*s", i, zUtf);
1196	}else if( w<0 ){
1197	sqlite3_fprintf(out, "%*s%s", aw-n, "", zUtf);
1198	}else{
1199	sqlite3_fprintf(out, "%s%*s", zUtf, aw-n, "");
1200	}
1201	}
1202
1203
1204	/*
	@@ -1834,11 +1260,11 @@
1260	struct __stat64 x = {0};
1261	# define STAT_CHR_SRC(mode) ((mode & (_S_IFCHR\|_S_IFIFO\|_S_IFREG))!=0)
1262	/* On Windows, open first, then check the stream nature. This order
1263	** is necessary because _stat() and sibs, when checking a named pipe,
1264	** effectively break the pipe as its supplier sees it. */
1265	FILE *rv = sqlite3_fopen(zFile, "rb");
1266	if( rv==0 ) return 0;
1267	if( _fstat64(_fileno(rv), &x) != 0
1268	\|\| !STAT_CHR_SRC(x.st_mode)){
1269	fclose(rv);
1270	rv = 0;
	@@ -1848,11 +1274,11 @@
1274	struct stat x = {0};
1275	int rc = stat(zFile, &x);
1276	# define STAT_CHR_SRC(mode) (S_ISREG(mode)\|\|S_ISFIFO(mode)\|\|S_ISCHR(mode))
1277	if( rc!=0 ) return 0;
1278	if( STAT_CHR_SRC(x.st_mode) ){
1279	return sqlite3_fopen(zFile, "rb");
1280	}else{
1281	return 0;
1282	}
1283	#endif
1284	#undef STAT_CHR_SRC
	@@ -1875,11 +1301,11 @@
1301	if( n+100>nLine ){
1302	nLine = nLine*2 + 100;
1303	zLine = realloc(zLine, nLine);
1304	shell_check_oom(zLine);
1305	}
1306	if( sqlite3_fgets(&zLine[n], nLine - n, in)==0 ){
1307	if( n==0 ){
1308	free(zLine);
1309	return 0;
1310	}
1311	zLine[n] = 0;
	@@ -2941,11 +2367,11 @@
2367	**
2368	******************************************************************************
2369	**
2370	** This SQLite extension implements functions that compute SHA3 hashes
2371	** in the way described by the (U.S.) NIST FIPS 202 SHA-3 Standard.
2372	** Three SQL functions are implemented:
2373	**
2374	** sha3(X,SIZE)
2375	** sha3_agg(Y,SIZE)
2376	** sha3_query(Z,SIZE)
2377	**
	@@ -3783,10 +3209,422 @@
3209	}
3210	return rc;
3211	}
3212
3213	/*********************** End ../ext/misc/shathree.c ******************/
3214	/*********************** Begin ../ext/misc/sha1.c ****************/
3215	/*
3216	** 2017-01-27
3217	**
3218	** The author disclaims copyright to this source code. In place of
3219	** a legal notice, here is a blessing:
3220	**
3221	** May you do good and not evil.
3222	** May you find forgiveness for yourself and forgive others.
3223	** May you share freely, never taking more than you give.
3224	**
3225	******************************************************************************
3226	**
3227	** This SQLite extension implements functions that compute SHA1 hashes.
3228	** Two SQL functions are implemented:
3229	**
3230	** sha1(X)
3231	** sha1_query(Y)
3232	**
3233	** The sha1(X) function computes the SHA1 hash of the input X, or NULL if
3234	** X is NULL.
3235	**
3236	** The sha1_query(Y) function evalutes all queries in the SQL statements of Y
3237	** and returns a hash of their results.
3238	*/
3239	/* #include "sqlite3ext.h" */
3240	SQLITE_EXTENSION_INIT1
3241	#include <assert.h>
3242	#include <string.h>
3243	#include <stdarg.h>
3244
3245	/******************************************************************************
3246	** The Hash Engine
3247	*/
3248	/* Context for the SHA1 hash */
3249	typedef struct SHA1Context SHA1Context;
3250	struct SHA1Context {
3251	unsigned int state[5];
3252	unsigned int count[2];
3253	unsigned char buffer[64];
3254	};
3255
3256	#define SHA_ROT(x,l,r) ((x) << (l) \| (x) >> (r))
3257	#define rol(x,k) SHA_ROT(x,k,32-(k))
3258	#define ror(x,k) SHA_ROT(x,32-(k),k)
3259
3260	#define blk0le(i) (block[i] = (ror(block[i],8)&0xFF00FF00) \
3261	\|(rol(block[i],8)&0x00FF00FF))
3262	#define blk0be(i) block[i]
3263	#define blk(i) (block[i&15] = rol(block[(i+13)&15]^block[(i+8)&15] \
3264	^block[(i+2)&15]^block[i&15],1))
3265
3266	/*
3267	* (R0+R1), R2, R3, R4 are the different operations (rounds) used in SHA1
3268	*
3269	* Rl0() for little-endian and Rb0() for big-endian. Endianness is
3270	* determined at run-time.
3271	*/
3272	#define Rl0(v,w,x,y,z,i) \
3273	z+=((w&(x^y))^y)+blk0le(i)+0x5A827999+rol(v,5);w=ror(w,2);
3274	#define Rb0(v,w,x,y,z,i) \
3275	z+=((w&(x^y))^y)+blk0be(i)+0x5A827999+rol(v,5);w=ror(w,2);
3276	#define R1(v,w,x,y,z,i) \
3277	z+=((w&(x^y))^y)+blk(i)+0x5A827999+rol(v,5);w=ror(w,2);
3278	#define R2(v,w,x,y,z,i) \
3279	z+=(w^x^y)+blk(i)+0x6ED9EBA1+rol(v,5);w=ror(w,2);
3280	#define R3(v,w,x,y,z,i) \
3281	z+=(((w\|x)&y)\|(w&x))+blk(i)+0x8F1BBCDC+rol(v,5);w=ror(w,2);
3282	#define R4(v,w,x,y,z,i) \
3283	z+=(w^x^y)+blk(i)+0xCA62C1D6+rol(v,5);w=ror(w,2);
3284
3285	/*
3286	* Hash a single 512-bit block. This is the core of the algorithm.
3287	*/
3288	static void SHA1Transform(unsigned int state[5], const unsigned char buffer[64]){
3289	unsigned int qq[5]; /* a, b, c, d, e; */
3290	static int one = 1;
3291	unsigned int block[16];
3292	memcpy(block, buffer, 64);
3293	memcpy(qq,state,5*sizeof(unsigned int));
3294
3295	#define a qq[0]
3296	#define b qq[1]
3297	#define c qq[2]
3298	#define d qq[3]
3299	#define e qq[4]
3300
3301	/* Copy p->state[] to working vars */
3302	/*
3303	a = state[0];
3304	b = state[1];
3305	c = state[2];
3306	d = state[3];
3307	e = state[4];
3308	*/
3309
3310	/* 4 rounds of 20 operations each. Loop unrolled. */
3311	if( 1 == (unsigned char)&one ){
3312	Rl0(a,b,c,d,e, 0); Rl0(e,a,b,c,d, 1); Rl0(d,e,a,b,c, 2); Rl0(c,d,e,a,b, 3);
3313	Rl0(b,c,d,e,a, 4); Rl0(a,b,c,d,e, 5); Rl0(e,a,b,c,d, 6); Rl0(d,e,a,b,c, 7);
3314	Rl0(c,d,e,a,b, 8); Rl0(b,c,d,e,a, 9); Rl0(a,b,c,d,e,10); Rl0(e,a,b,c,d,11);
3315	Rl0(d,e,a,b,c,12); Rl0(c,d,e,a,b,13); Rl0(b,c,d,e,a,14); Rl0(a,b,c,d,e,15);
3316	}else{
3317	Rb0(a,b,c,d,e, 0); Rb0(e,a,b,c,d, 1); Rb0(d,e,a,b,c, 2); Rb0(c,d,e,a,b, 3);
3318	Rb0(b,c,d,e,a, 4); Rb0(a,b,c,d,e, 5); Rb0(e,a,b,c,d, 6); Rb0(d,e,a,b,c, 7);
3319	Rb0(c,d,e,a,b, 8); Rb0(b,c,d,e,a, 9); Rb0(a,b,c,d,e,10); Rb0(e,a,b,c,d,11);
3320	Rb0(d,e,a,b,c,12); Rb0(c,d,e,a,b,13); Rb0(b,c,d,e,a,14); Rb0(a,b,c,d,e,15);
3321	}
3322	R1(e,a,b,c,d,16); R1(d,e,a,b,c,17); R1(c,d,e,a,b,18); R1(b,c,d,e,a,19);
3323	R2(a,b,c,d,e,20); R2(e,a,b,c,d,21); R2(d,e,a,b,c,22); R2(c,d,e,a,b,23);
3324	R2(b,c,d,e,a,24); R2(a,b,c,d,e,25); R2(e,a,b,c,d,26); R2(d,e,a,b,c,27);
3325	R2(c,d,e,a,b,28); R2(b,c,d,e,a,29); R2(a,b,c,d,e,30); R2(e,a,b,c,d,31);
3326	R2(d,e,a,b,c,32); R2(c,d,e,a,b,33); R2(b,c,d,e,a,34); R2(a,b,c,d,e,35);
3327	R2(e,a,b,c,d,36); R2(d,e,a,b,c,37); R2(c,d,e,a,b,38); R2(b,c,d,e,a,39);
3328	R3(a,b,c,d,e,40); R3(e,a,b,c,d,41); R3(d,e,a,b,c,42); R3(c,d,e,a,b,43);
3329	R3(b,c,d,e,a,44); R3(a,b,c,d,e,45); R3(e,a,b,c,d,46); R3(d,e,a,b,c,47);
3330	R3(c,d,e,a,b,48); R3(b,c,d,e,a,49); R3(a,b,c,d,e,50); R3(e,a,b,c,d,51);
3331	R3(d,e,a,b,c,52); R3(c,d,e,a,b,53); R3(b,c,d,e,a,54); R3(a,b,c,d,e,55);
3332	R3(e,a,b,c,d,56); R3(d,e,a,b,c,57); R3(c,d,e,a,b,58); R3(b,c,d,e,a,59);
3333	R4(a,b,c,d,e,60); R4(e,a,b,c,d,61); R4(d,e,a,b,c,62); R4(c,d,e,a,b,63);
3334	R4(b,c,d,e,a,64); R4(a,b,c,d,e,65); R4(e,a,b,c,d,66); R4(d,e,a,b,c,67);
3335	R4(c,d,e,a,b,68); R4(b,c,d,e,a,69); R4(a,b,c,d,e,70); R4(e,a,b,c,d,71);
3336	R4(d,e,a,b,c,72); R4(c,d,e,a,b,73); R4(b,c,d,e,a,74); R4(a,b,c,d,e,75);
3337	R4(e,a,b,c,d,76); R4(d,e,a,b,c,77); R4(c,d,e,a,b,78); R4(b,c,d,e,a,79);
3338
3339	/* Add the working vars back into context.state[] */
3340	state[0] += a;
3341	state[1] += b;
3342	state[2] += c;
3343	state[3] += d;
3344	state[4] += e;
3345
3346	#undef a
3347	#undef b
3348	#undef c
3349	#undef d
3350	#undef e
3351	}
3352
3353
3354	/* Initialize a SHA1 context */
3355	static void hash_init(SHA1Context *p){
3356	/* SHA1 initialization constants */
3357	p->state[0] = 0x67452301;
3358	p->state[1] = 0xEFCDAB89;
3359	p->state[2] = 0x98BADCFE;
3360	p->state[3] = 0x10325476;
3361	p->state[4] = 0xC3D2E1F0;
3362	p->count[0] = p->count[1] = 0;
3363	}
3364
3365	/* Add new content to the SHA1 hash */
3366	static void hash_step(
3367	SHA1Context p, / Add content to this context */
3368	const unsigned char data, / Data to be added */
3369	unsigned int len /* Number of bytes in data */
3370	){
3371	unsigned int i, j;
3372
3373	j = p->count[0];
3374	if( (p->count[0] += len << 3) < j ){
3375	p->count[1] += (len>>29)+1;
3376	}
3377	j = (j >> 3) & 63;
3378	if( (j + len) > 63 ){
3379	(void)memcpy(&p->buffer[j], data, (i = 64-j));
3380	SHA1Transform(p->state, p->buffer);
3381	for(; i + 63 < len; i += 64){
3382	SHA1Transform(p->state, &data[i]);
3383	}
3384	j = 0;
3385	}else{
3386	i = 0;
3387	}
3388	(void)memcpy(&p->buffer[j], &data[i], len - i);
3389	}
3390
3391	/* Compute a string using sqlite3_vsnprintf() and hash it */
3392	static void hash_step_vformat(
3393	SHA1Context p, / Add content to this context */
3394	const char *zFormat,
3395	...
3396	){
3397	va_list ap;
3398	int n;
3399	char zBuf[50];
3400	va_start(ap, zFormat);
3401	sqlite3_vsnprintf(sizeof(zBuf),zBuf,zFormat,ap);
3402	va_end(ap);
3403	n = (int)strlen(zBuf);
3404	hash_step(p, (unsigned char*)zBuf, n);
3405	}
3406
3407
3408	/* Add padding and compute the message digest. Render the
3409	** message digest as lower-case hexadecimal and put it into
3410	** zOut[]. zOut[] must be at least 41 bytes long. */
3411	static void hash_finish(
3412	SHA1Context p, / The SHA1 context to finish and render */
3413	char zOut, / Store hex or binary hash here */
3414	int bAsBinary /* 1 for binary hash, 0 for hex hash */
3415	){
3416	unsigned int i;
3417	unsigned char finalcount[8];
3418	unsigned char digest[20];
3419	static const char zEncode[] = "0123456789abcdef";
3420
3421	for (i = 0; i < 8; i++){
3422	finalcount[i] = (unsigned char)((p->count[(i >= 4 ? 0 : 1)]
3423	>> ((3-(i & 3)) * 8) ) & 255); /* Endian independent */
3424	}
3425	hash_step(p, (const unsigned char *)"\200", 1);
3426	while ((p->count[0] & 504) != 448){
3427	hash_step(p, (const unsigned char *)"\0", 1);
3428	}
3429	hash_step(p, finalcount, 8); /* Should cause a SHA1Transform() */
3430	for (i = 0; i < 20; i++){
3431	digest[i] = (unsigned char)((p->state[i>>2] >> ((3-(i & 3)) * 8) ) & 255);
3432	}
3433	if( bAsBinary ){
3434	memcpy(zOut, digest, 20);
3435	}else{
3436	for(i=0; i<20; i++){
3437	zOut[i*2] = zEncode[(digest[i]>>4)&0xf];
3438	zOut[i*2+1] = zEncode[digest[i] & 0xf];
3439	}
3440	zOut[i*2]= 0;
3441	}
3442	}
3443	/* End of the hashing logic
3444	*****************************************************************************/
3445
3446	/*
3447	** Implementation of the sha1(X) function.
3448	**
3449	** Return a lower-case hexadecimal rendering of the SHA1 hash of the
3450	** argument X. If X is a BLOB, it is hashed as is. For all other
3451	** types of input, X is converted into a UTF-8 string and the string
3452	** is hash without the trailing 0x00 terminator. The hash of a NULL
3453	** value is NULL.
3454	*/
3455	static void sha1Func(
3456	sqlite3_context *context,
3457	int argc,
3458	sqlite3_value **argv
3459	){
3460	SHA1Context cx;
3461	int eType = sqlite3_value_type(argv[0]);
3462	int nByte = sqlite3_value_bytes(argv[0]);
3463	char zOut[44];
3464
3465	assert( argc==1 );
3466	if( eType==SQLITE_NULL ) return;
3467	hash_init(&cx);
3468	if( eType==SQLITE_BLOB ){
3469	hash_step(&cx, sqlite3_value_blob(argv[0]), nByte);
3470	}else{
3471	hash_step(&cx, sqlite3_value_text(argv[0]), nByte);
3472	}
3473	if( sqlite3_user_data(context)!=0 ){
3474	hash_finish(&cx, zOut, 1);
3475	sqlite3_result_blob(context, zOut, 20, SQLITE_TRANSIENT);
3476	}else{
3477	hash_finish(&cx, zOut, 0);
3478	sqlite3_result_blob(context, zOut, 40, SQLITE_TRANSIENT);
3479	}
3480	}
3481
3482	/*
3483	** Implementation of the sha1_query(SQL) function.
3484	**
3485	** This function compiles and runs the SQL statement(s) given in the
3486	** argument. The results are hashed using SHA1 and that hash is returned.
3487	**
3488	** The original SQL text is included as part of the hash.
3489	**
3490	** The hash is not just a concatenation of the outputs. Each query
3491	** is delimited and each row and value within the query is delimited,
3492	** with all values being marked with their datatypes.
3493	*/
3494	static void sha1QueryFunc(
3495	sqlite3_context *context,
3496	int argc,
3497	sqlite3_value **argv
3498	){
3499	sqlite3 *db = sqlite3_context_db_handle(context);
3500	const char zSql = (const char)sqlite3_value_text(argv[0]);
3501	sqlite3_stmt *pStmt = 0;
3502	int nCol; /* Number of columns in the result set */
3503	int i; /* Loop counter */
3504	int rc;
3505	int n;
3506	const char *z;
3507	SHA1Context cx;
3508	char zOut[44];
3509
3510	assert( argc==1 );
3511	if( zSql==0 ) return;
3512	hash_init(&cx);
3513	while( zSql[0] ){
3514	rc = sqlite3_prepare_v2(db, zSql, -1, &pStmt, &zSql);
3515	if( rc ){
3516	char *zMsg = sqlite3_mprintf("error SQL statement [%s]: %s",
3517	zSql, sqlite3_errmsg(db));
3518	sqlite3_finalize(pStmt);
3519	sqlite3_result_error(context, zMsg, -1);
3520	sqlite3_free(zMsg);
3521	return;
3522	}
3523	if( !sqlite3_stmt_readonly(pStmt) ){
3524	char *zMsg = sqlite3_mprintf("non-query: [%s]", sqlite3_sql(pStmt));
3525	sqlite3_finalize(pStmt);
3526	sqlite3_result_error(context, zMsg, -1);
3527	sqlite3_free(zMsg);
3528	return;
3529	}
3530	nCol = sqlite3_column_count(pStmt);
3531	z = sqlite3_sql(pStmt);
3532	n = (int)strlen(z);
3533	hash_step_vformat(&cx,"S%d:",n);
3534	hash_step(&cx,(unsigned char*)z,n);
3535
3536	/* Compute a hash over the result of the query */
3537	while( SQLITE_ROW==sqlite3_step(pStmt) ){
3538	hash_step(&cx,(const unsigned char*)"R",1);
3539	for(i=0; i<nCol; i++){
3540	switch( sqlite3_column_type(pStmt,i) ){
3541	case SQLITE_NULL: {
3542	hash_step(&cx, (const unsigned char*)"N",1);
3543	break;
3544	}
3545	case SQLITE_INTEGER: {
3546	sqlite3_uint64 u;
3547	int j;
3548	unsigned char x[9];
3549	sqlite3_int64 v = sqlite3_column_int64(pStmt,i);
3550	memcpy(&u, &v, 8);
3551	for(j=8; j>=1; j--){
3552	x[j] = u & 0xff;
3553	u >>= 8;
3554	}
3555	x[0] = 'I';
3556	hash_step(&cx, x, 9);
3557	break;
3558	}
3559	case SQLITE_FLOAT: {
3560	sqlite3_uint64 u;
3561	int j;
3562	unsigned char x[9];
3563	double r = sqlite3_column_double(pStmt,i);
3564	memcpy(&u, &r, 8);
3565	for(j=8; j>=1; j--){
3566	x[j] = u & 0xff;
3567	u >>= 8;
3568	}
3569	x[0] = 'F';
3570	hash_step(&cx,x,9);
3571	break;
3572	}
3573	case SQLITE_TEXT: {
3574	int n2 = sqlite3_column_bytes(pStmt, i);
3575	const unsigned char *z2 = sqlite3_column_text(pStmt, i);
3576	hash_step_vformat(&cx,"T%d:",n2);
3577	hash_step(&cx, z2, n2);
3578	break;
3579	}
3580	case SQLITE_BLOB: {
3581	int n2 = sqlite3_column_bytes(pStmt, i);
3582	const unsigned char *z2 = sqlite3_column_blob(pStmt, i);
3583	hash_step_vformat(&cx,"B%d:",n2);
3584	hash_step(&cx, z2, n2);
3585	break;
3586	}
3587	}
3588	}
3589	}
3590	sqlite3_finalize(pStmt);
3591	}
3592	hash_finish(&cx, zOut, 0);
3593	sqlite3_result_text(context, zOut, 40, SQLITE_TRANSIENT);
3594	}
3595
3596
3597	#ifdef _WIN32
3598
3599	#endif
3600	int sqlite3_sha_init(
3601	sqlite3 *db,
3602	char **pzErrMsg,
3603	const sqlite3_api_routines *pApi
3604	){
3605	int rc = SQLITE_OK;
3606	static int one = 1;
3607	SQLITE_EXTENSION_INIT2(pApi);
3608	(void)pzErrMsg; /* Unused parameter */
3609	rc = sqlite3_create_function(db, "sha1", 1,
3610	SQLITE_UTF8 \| SQLITE_INNOCUOUS \| SQLITE_DETERMINISTIC,
3611	0, sha1Func, 0, 0);
3612	if( rc==SQLITE_OK ){
3613	rc = sqlite3_create_function(db, "sha1b", 1,
3614	SQLITE_UTF8 \| SQLITE_INNOCUOUS \| SQLITE_DETERMINISTIC,
3615	(void*)&one, sha1Func, 0, 0);
3616	}
3617	if( rc==SQLITE_OK ){
3618	rc = sqlite3_create_function(db, "sha1_query", 1,
3619	SQLITE_UTF8\|SQLITE_DIRECTONLY, 0,
3620	sha1QueryFunc, 0, 0);
3621	}
3622	return rc;
3623	}
3624
3625	/*********************** End ../ext/misc/sha1.c ******************/
3626	/*********************** Begin ../ext/misc/uint.c ****************/
3627	/*
3628	** 2020-04-14
3629	**
3630	** The author disclaims copyright to this source code. In place of
	@@ -5075,11 +4913,11 @@
4913	}else if( !p->bSorted \|\| y>=p->a[p->nUsed-1] ){
4914	p->a[p->nUsed++] = y;
4915	}else if( p->bKeepSorted ){
4916	int i;
4917	i = percentBinarySearch(p, y, 0);
4918	if( i<(int)p->nUsed ){
4919	memmove(&p->a[i+1], &p->a[i], (p->nUsed-i)*sizeof(p->a[0]));
4920	}
4921	p->a[i] = y;
4922	p->nUsed++;
4923	}else{
	@@ -5191,11 +5029,11 @@
5029
5030	/* Find and remove the row */
5031	i = percentBinarySearch(p, y, 1);
5032	if( i>=0 ){
5033	p->nUsed--;
5034	if( i<(int)p->nUsed ){
5035	memmove(&p->a[i], &p->a[i+1], (p->nUsed - i)*sizeof(p->a[0]));
5036	}
5037	}
5038	}
5039
	@@ -5251,15 +5089,15 @@
5089	sqlite3 *db,
5090	char **pzErrMsg,
5091	const sqlite3_api_routines *pApi
5092	){
5093	int rc = SQLITE_OK;
5094	unsigned int i;
5095	#ifdef SQLITE3EXT_H


5096	SQLITE_EXTENSION_INIT2(pApi);
5097	#else
5098	(void)pApi; /* Unused parameter */
5099	#endif
5100	(void)pzErrMsg; /* Unused parameter */
5101	for(i=0; i<sizeof(aPercentFunc)/sizeof(aPercentFunc[0]); i++){
5102	rc = sqlite3_create_window_function(db,
5103	aPercentFunc[i].zName,
	@@ -7012,50 +6850,58 @@
6850	aIdx[4] = i;
6851	idxNum \|= 0x40;
6852	}
6853	continue;
6854	}
6855	if( pConstraint->iColumn<SERIES_COLUMN_START ){
6856	if( pConstraint->iColumn==SERIES_COLUMN_VALUE && pConstraint->usable ){
6857	switch( op ){
6858	case SQLITE_INDEX_CONSTRAINT_EQ:
6859	case SQLITE_INDEX_CONSTRAINT_IS: {
6860	idxNum \|= 0x0080;
6861	idxNum &= ~0x3300;
6862	aIdx[5] = i;
6863	aIdx[6] = -1;
6864	#ifndef ZERO_ARGUMENT_GENERATE_SERIES
6865	bStartSeen = 1;
6866	#endif
6867	break;
6868	}
6869	case SQLITE_INDEX_CONSTRAINT_GE: {
6870	if( idxNum & 0x0080 ) break;
6871	idxNum \|= 0x0100;
6872	idxNum &= ~0x0200;
6873	aIdx[5] = i;
6874	#ifndef ZERO_ARGUMENT_GENERATE_SERIES
6875	bStartSeen = 1;
6876	#endif
6877	break;
6878	}
6879	case SQLITE_INDEX_CONSTRAINT_GT: {
6880	if( idxNum & 0x0080 ) break;
6881	idxNum \|= 0x0200;
6882	idxNum &= ~0x0100;
6883	aIdx[5] = i;
6884	#ifndef ZERO_ARGUMENT_GENERATE_SERIES
6885	bStartSeen = 1;
6886	#endif
6887	break;
6888	}
6889	case SQLITE_INDEX_CONSTRAINT_LE: {
6890	if( idxNum & 0x0080 ) break;
6891	idxNum \|= 0x1000;
6892	idxNum &= ~0x2000;
6893	aIdx[6] = i;
6894	break;
6895	}
6896	case SQLITE_INDEX_CONSTRAINT_LT: {
6897	if( idxNum & 0x0080 ) break;
6898	idxNum \|= 0x2000;
6899	idxNum &= ~0x1000;
6900	aIdx[6] = i;
6901	break;
6902	}
6903	}
6904	}
6905	continue;
6906	}
6907	iCol = pConstraint->iColumn - SERIES_COLUMN_START;
	@@ -8115,11 +7961,11 @@
7961	** If the optional MTIME argument is present, then it is interpreted
7962	** as an integer - the number of seconds since the unix epoch. The
7963	** modification-time of the target file is set to this value before
7964	** returning.
7965	**
7966	** If five or more arguments are passed to this function and an
7967	** error is encountered, an exception is raised.
7968	**
7969	** READFILE(FILE):
7970	**
7971	** Read and return the contents of file FILE (type blob) from disk.
	@@ -8185,10 +8031,17 @@
8031	# define lstat(path,buf) stat(path,buf)
8032	#endif
8033	#include <time.h>
8034	#include <errno.h>
8035
8036	/* When used as part of the CLI, the sqlite3_stdio.h module will have
8037	** been included before this one. In that case use the sqlite3_stdio.h
8038	** #defines. If not, create our own for fopen().
8039	*/
8040	#ifndef _SQLITE3_STDIO_H_
8041	# define sqlite3_fopen fopen
8042	#endif
8043
8044	/*
8045	** Structure of the fsdir() table-valued function
8046	*/
8047	/* 0 1 2 3 4 5 */
	@@ -8217,11 +8070,11 @@
8070	sqlite3_int64 nIn;
8071	void *pBuf;
8072	sqlite3 *db;
8073	int mxBlob;
8074
8075	in = sqlite3_fopen(zName, "rb");
8076	if( in==0 ){
8077	/* File does not exist or is unreadable. Leave the result set to NULL. */
8078	return;
8079	}
8080	fseek(in, 0, SEEK_END);
	@@ -8472,11 +8325,11 @@
8325	}
8326	}else{
8327	sqlite3_int64 nWrite = 0;
8328	const char *z;
8329	int rc = 0;
8330	FILE *out = sqlite3_fopen(zFile, "wb");
8331	if( out==0 ) return 1;
8332	z = (const char*)sqlite3_value_blob(pData);
8333	if( z ){
8334	sqlite3_int64 n = fwrite(z, 1, sqlite3_value_bytes(pData), out);
8335	nWrite = sqlite3_value_bytes(pData);
	@@ -10333,10 +10186,18 @@
10186	#ifndef SQLITE_NO_STDINT
10187	# include <stdint.h>
10188	#endif
10189
10190	#include <zlib.h>
10191
10192	/* When used as part of the CLI, the sqlite3_stdio.h module will have
10193	** been included before this one. In that case use the sqlite3_stdio.h
10194	** #defines. If not, create our own for fopen().
10195	*/
10196	#ifndef _SQLITE3_STDIO_H_
10197	# define sqlite3_fopen fopen
10198	#endif
10199
10200	#ifndef SQLITE_OMIT_VIRTUALTABLE
10201
10202	#ifndef SQLITE_AMALGAMATION
10203
	@@ -11590,11 +11451,11 @@
11451	}else{
11452	zFile = (const char*)sqlite3_value_text(argv[0]);
11453	}
11454
11455	if( 0==pTab->pWriteFd && 0==bInMemory ){
11456	pCsr->pFile = zFile ? sqlite3_fopen(zFile, "rb") : 0;
11457	if( pCsr->pFile==0 ){
11458	zipfileCursorErr(pCsr, "cannot open file: %s", zFile);
11459	rc = SQLITE_ERROR;
11460	}else{
11461	rc = zipfileReadEOCD(pTab, 0, 0, pCsr->pFile, &pCsr->eocd);
	@@ -11780,11 +11641,11 @@
11641
11642	/* Open a write fd on the file. Also load the entire central directory
11643	** structure into memory. During the transaction any new file data is
11644	** appended to the archive file, but the central directory is accumulated
11645	** in main-memory until the transaction is committed. */
11646	pTab->pWriteFd = sqlite3_fopen(pTab->zFile, "ab+");
11647	if( pTab->pWriteFd==0 ){
11648	pTab->base.zErrMsg = sqlite3_mprintf(
11649	"zipfile: failed to open file %s for writing", pTab->zFile
11650	);
11651	rc = SQLITE_ERROR;
	@@ -14233,10 +14094,70 @@
14094	}
14095
14096	return rc;
14097	}
14098
14099	/*
14100	** This function tests if the schema of the main database of database handle
14101	** db contains an object named zTab. Assuming no error occurs, output parameter
14102	** (*pbContains) is set to true if zTab exists, or false if it does not.
14103	**
14104	** Or, if an error occurs, an SQLite error code is returned. The final value
14105	** of (*pbContains) is undefined in this case.
14106	*/
14107	static int expertDbContainsObject(
14108	sqlite3 *db,
14109	const char *zTab,
14110	int pbContains / OUT: True if object exists */
14111	){
14112	const char *zSql = "SELECT 1 FROM sqlite_schema WHERE name = ?";
14113	sqlite3_stmt *pSql = 0;
14114	int rc = SQLITE_OK;
14115	int ret = 0;
14116
14117	rc = sqlite3_prepare_v2(db, zSql, -1, &pSql, 0);
14118	if( rc==SQLITE_OK ){
14119	sqlite3_bind_text(pSql, 1, zTab, -1, SQLITE_STATIC);
14120	if( SQLITE_ROW==sqlite3_step(pSql) ){
14121	ret = 1;
14122	}
14123	rc = sqlite3_finalize(pSql);
14124	}
14125
14126	*pbContains = ret;
14127	return rc;
14128	}
14129
14130	/*
14131	** Execute SQL command zSql using database handle db. If no error occurs,
14132	** set (*pzErr) to NULL and return SQLITE_OK.
14133	**
14134	** If an error does occur, return an SQLite error code and set (*pzErr) to
14135	** point to a buffer containing an English language error message. Except,
14136	** if the error message begins with "no such module:", then ignore the
14137	** error and return as if the SQL statement had succeeded.
14138	**
14139	** This is used to copy as much of the database schema as possible while
14140	** ignoring any errors related to missing virtual table modules.
14141	*/
14142	static int expertSchemaSql(sqlite3 db, const char zSql, char **pzErr){
14143	int rc = SQLITE_OK;
14144	char *zErr = 0;
14145
14146	rc = sqlite3_exec(db, zSql, 0, 0, &zErr);
14147	if( rc!=SQLITE_OK && zErr ){
14148	int nErr = STRLEN(zErr);
14149	if( nErr>=15 && memcmp(zErr, "no such module:", 15)==0 ){
14150	sqlite3_free(zErr);
14151	rc = SQLITE_OK;
14152	zErr = 0;
14153	}
14154	}
14155
14156	*pzErr = zErr;
14157	return rc;
14158	}
14159
14160	static int idxCreateVtabSchema(sqlite3expert p, char *pzErrmsg){
14161	int rc = idxRegisterVtab(p);
14162	sqlite3_stmt *pSchema = 0;
14163
	@@ -14244,30 +14165,39 @@
14165	**
14166	** 1) Add an entry to the p->pTable list, and
14167	** 2) Create the equivalent virtual table in dbv.
14168	*/
14169	rc = idxPrepareStmt(p->db, &pSchema, pzErrmsg,
14170	"SELECT type, name, sql, 1, "
14171	" substr(sql,1,14)=='create virtual' COLLATE nocase "
14172	"FROM sqlite_schema "
14173	"WHERE type IN ('table','view') AND "
14174	" substr(name,1,7)!='sqlite_' COLLATE nocase "
14175	" UNION ALL "
14176	"SELECT type, name, sql, 2, 0 FROM sqlite_schema "
14177	"WHERE type = 'trigger'"
14178	" AND tbl_name IN(SELECT name FROM sqlite_schema WHERE type = 'view') "
14179	"ORDER BY 4, 5 DESC, 1"
14180	);
14181	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSchema) ){
14182	const char zType = (const char)sqlite3_column_text(pSchema, 0);
14183	const char zName = (const char)sqlite3_column_text(pSchema, 1);
14184	const char zSql = (const char)sqlite3_column_text(pSchema, 2);
14185	int bVirtual = sqlite3_column_int(pSchema, 4);
14186	int bExists = 0;
14187
14188	if( zType==0 \|\| zName==0 ) continue;
14189	rc = expertDbContainsObject(p->dbv, zName, &bExists);
14190	if( rc \|\| bExists ) continue;
14191
14192	if( zType[0]=='v' \|\| zType[1]=='r' \|\| bVirtual ){
14193	/* A view. Or a trigger on a view. */
14194	if( zSql ) rc = expertSchemaSql(p->dbv, zSql, pzErrmsg);
14195	}else{
14196	IdxTable *pTab;
14197	rc = idxGetTableInfo(p->db, zName, &pTab, pzErrmsg);
14198	if( rc==SQLITE_OK && ALWAYS(pTab!=0) ){
14199	int i;
14200	char *zInner = 0;
14201	char *zOuter = 0;
14202	pTab->pNext = p->pTable;
14203	p->pTable = pTab;
	@@ -14464,10 +14394,16 @@
14394	sqlite3_bind_text(pIndexXInfo, 1, zIdx, -1, SQLITE_STATIC);
14395	while( SQLITE_OK==rc && SQLITE_ROW==sqlite3_step(pIndexXInfo) ){
14396	const char *zComma = zCols==0 ? "" : ", ";
14397	const char zName = (const char)sqlite3_column_text(pIndexXInfo, 0);
14398	const char zColl = (const char)sqlite3_column_text(pIndexXInfo, 1);
14399	if( zName==0 ){
14400	/* This index contains an expression. Ignore it. */
14401	sqlite3_free(zCols);
14402	sqlite3_free(zOrder);
14403	return sqlite3_reset(pIndexXInfo);
14404	}
14405	zCols = idxAppendText(&rc, zCols,
14406	"%sx.%Q IS sqlite_expert_rem(%d, x.%Q) COLLATE %s",
14407	zComma, zName, nCol, zName, zColl
14408	);
14409	zOrder = idxAppendText(&rc, zOrder, "%s%d", zComma, ++nCol);
	@@ -14792,16 +14728,22 @@
14728
14729	/* Copy the entire schema of database [db] into [dbm]. */
14730	if( rc==SQLITE_OK ){
14731	sqlite3_stmt *pSql = 0;
14732	rc = idxPrintfPrepareStmt(pNew->db, &pSql, pzErrmsg,
14733	"SELECT sql, name, substr(sql,1,14)=='create virtual' COLLATE nocase"
14734	" FROM sqlite_schema WHERE substr(name,1,7)!='sqlite_' COLLATE nocase"
14735	" ORDER BY 3 DESC, rowid"
14736	);
14737	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
14738	const char zSql = (const char)sqlite3_column_text(pSql, 0);
14739	const char zName = (const char)sqlite3_column_text(pSql, 1);
14740	int bExists = 0;
14741	rc = expertDbContainsObject(pNew->dbm, zName, &bExists);
14742	if( rc==SQLITE_OK && zSql && bExists==0 ){
14743	rc = expertSchemaSql(pNew->dbm, zSql, pzErrmsg);
14744	}
14745	}
14746	idxFinalize(&rc, pSql);
14747	}
14748
14749	/* Create the vtab schema */
	@@ -16211,10 +16153,1197 @@
16153	}
16154	return rc;
16155	}
16156
16157	/*********************** End ../ext/misc/stmtrand.c ******************/
16158	/*********************** Begin ../ext/misc/vfstrace.c ****************/
16159	/*
16160	** 2011 March 16
16161	**
16162	** The author disclaims copyright to this source code. In place of
16163	** a legal notice, here is a blessing:
16164	**
16165	** May you do good and not evil.
16166	** May you find forgiveness for yourself and forgive others.
16167	** May you share freely, never taking more than you give.
16168	**
16169	******************************************************************************
16170	**
16171	** This file contains code implements a VFS shim that writes diagnostic
16172	** output for each VFS call, similar to "strace".
16173	**
16174	** USAGE:
16175	**
16176	** This source file exports a single symbol which is the name of a
16177	** function:
16178	**
16179	** int vfstrace_register(
16180	** const char *zTraceName, // Name of the newly constructed VFS
16181	** const char *zOldVfsName, // Name of the underlying VFS
16182	** int (xOut)(const char,void*), // Output routine. ex: fputs
16183	** void *pOutArg, // 2nd argument to xOut. ex: stderr
16184	** int makeDefault // Make the new VFS the default
16185	** );
16186	**
16187	** Applications that want to trace their VFS usage must provide a callback
16188	** function with this prototype:
16189	**
16190	** int traceOutput(const char zMessage, void pAppData);
16191	**
16192	** This function will "output" the trace messages, where "output" can
16193	** mean different things to different applications. The traceOutput function
16194	** for the command-line shell (see shell.c) is "fputs" from the standard
16195	** library, which means that all trace output is written on the stream
16196	** specified by the second argument. In the case of the command-line shell
16197	** the second argument is stderr. Other applications might choose to output
16198	** trace information to a file, over a socket, or write it into a buffer.
16199	**
16200	** The vfstrace_register() function creates a new "shim" VFS named by
16201	** the zTraceName parameter. A "shim" VFS is an SQLite backend that does
16202	** not really perform the duties of a true backend, but simply filters or
16203	** interprets VFS calls before passing them off to another VFS which does
16204	** the actual work. In this case the other VFS - the one that does the
16205	** real work - is identified by the second parameter, zOldVfsName. If
16206	** the 2nd parameter is NULL then the default VFS is used. The common
16207	** case is for the 2nd parameter to be NULL.
16208	**
16209	** The third and fourth parameters are the pointer to the output function
16210	** and the second argument to the output function. For the SQLite
16211	** command-line shell, when the -vfstrace option is used, these parameters
16212	** are fputs and stderr, respectively.
16213	**
16214	** The fifth argument is true (non-zero) to cause the newly created VFS
16215	** to become the default VFS. The common case is for the fifth parameter
16216	** to be true.
16217	**
16218	** The call to vfstrace_register() simply creates the shim VFS that does
16219	** tracing. The application must also arrange to use the new VFS for
16220	** all database connections that are created and for which tracing is
16221	** desired. This can be done by specifying the trace VFS using URI filename
16222	** notation, or by specifying the trace VFS as the 4th parameter to
16223	** sqlite3_open_v2() or by making the trace VFS be the default (by setting
16224	** the 5th parameter of vfstrace_register() to 1).
16225	**
16226	**
16227	** ENABLING VFSTRACE IN A COMMAND-LINE SHELL
16228	**
16229	** The SQLite command line shell implemented by the shell.c source file
16230	** can be used with this module. To compile in -vfstrace support, first
16231	** gather this file (test_vfstrace.c), the shell source file (shell.c),
16232	** and the SQLite amalgamation source files (sqlite3.c, sqlite3.h) into
16233	** the working directory. Then compile using a command like the following:
16234	**
16235	** gcc -o sqlite3 -Os -I. -DSQLITE_ENABLE_VFSTRACE \
16236	** -DSQLITE_THREADSAFE=0 -DSQLITE_ENABLE_FTS3 -DSQLITE_ENABLE_RTREE \
16237	** -DHAVE_READLINE -DHAVE_USLEEP=1 \
16238	** shell.c test_vfstrace.c sqlite3.c -ldl -lreadline -lncurses
16239	**
16240	** The gcc command above works on Linux and provides (in addition to the
16241	** -vfstrace option) support for FTS3 and FTS4, RTREE, and command-line
16242	** editing using the readline library. The command-line shell does not
16243	** use threads so we added -DSQLITE_THREADSAFE=0 just to make the code
16244	** run a little faster. For compiling on a Mac, you'll probably need
16245	** to omit the -DHAVE_READLINE, the -lreadline, and the -lncurses options.
16246	** The compilation could be simplified to just this:
16247	**
16248	** gcc -DSQLITE_ENABLE_VFSTRACE \
16249	** shell.c test_vfstrace.c sqlite3.c -ldl -lpthread
16250	**
16251	** In this second example, all unnecessary options have been removed
16252	** Note that since the code is now threadsafe, we had to add the -lpthread
16253	** option to pull in the pthreads library.
16254	**
16255	** To cross-compile for windows using MinGW, a command like this might
16256	** work:
16257	**
16258	** /opt/mingw/bin/i386-mingw32msvc-gcc -o sqlite3.exe -Os -I \
16259	** -DSQLITE_THREADSAFE=0 -DSQLITE_ENABLE_VFSTRACE \
16260	** shell.c test_vfstrace.c sqlite3.c
16261	**
16262	** Similar compiler commands will work on different systems. The key
16263	** invariants are (1) you must have -DSQLITE_ENABLE_VFSTRACE so that
16264	** the shell.c source file will know to include the -vfstrace command-line
16265	** option and (2) you must compile and link the three source files
16266	** shell,c, test_vfstrace.c, and sqlite3.c.
16267	**
16268	** RUNTIME CONTROL OF VFSTRACE OUTPUT
16269	**
16270	** The application can use the "vfstrace" pragma to control which VFS
16271	** APIs are traced. To disable all output:
16272	**
16273	** PRAGMA vfstrace('-all');
16274	**
16275	** To enable all output (which is the default setting):
16276	**
16277	** PRAGMA vfstrace('+all');
16278	**
16279	** Individual APIs can be enabled or disabled by name, with or without
16280	** the initial "x" character. For example, to set up for tracing lock
16281	** primatives only:
16282	**
16283	** PRAGMA vfstrace('-all, +Lock,Unlock,ShmLock');
16284	**
16285	** The argument to the vfstrace pragma ignores capitalization and any
16286	** characters other than alphabetics, '+', and '-'.
16287	*/
16288	#include <stdlib.h>
16289	#include <string.h>
16290	/* #include "sqlite3.h" */
16291
16292	/*
16293	** An instance of this structure is attached to the each trace VFS to
16294	** provide auxiliary information.
16295	*/
16296	typedef struct vfstrace_info vfstrace_info;
16297	struct vfstrace_info {
16298	sqlite3_vfs pRootVfs; / The underlying real VFS */
16299	int (xOut)(const char, void); / Send output here */
16300	unsigned int mTrace; /* Mask of interfaces to trace */
16301	u8 bOn; /* Tracing on/off */
16302	void pOutArg; / First argument to xOut */
16303	const char zVfsName; / Name of this trace-VFS */
16304	sqlite3_vfs pTraceVfs; / Pointer back to the trace VFS */
16305	};
16306
16307	/*
16308	** The sqlite3_file object for the trace VFS
16309	*/
16310	typedef struct vfstrace_file vfstrace_file;
16311	struct vfstrace_file {
16312	sqlite3_file base; /* Base class. Must be first */
16313	vfstrace_info pInfo; / The trace-VFS to which this file belongs */
16314	const char zFName; / Base name of the file */
16315	sqlite3_file pReal; / The real underlying file */
16316	};
16317
16318	/*
16319	** Bit values for vfstrace_info.mTrace.
16320	*/
16321	#define VTR_CLOSE 0x00000001
16322	#define VTR_READ 0x00000002
16323	#define VTR_WRITE 0x00000004
16324	#define VTR_TRUNC 0x00000008
16325	#define VTR_SYNC 0x00000010
16326	#define VTR_FSIZE 0x00000020
16327	#define VTR_LOCK 0x00000040
16328	#define VTR_UNLOCK 0x00000080
16329	#define VTR_CRL 0x00000100
16330	#define VTR_FCTRL 0x00000200
16331	#define VTR_SECSZ 0x00000400
16332	#define VTR_DEVCHAR 0x00000800
16333	#define VTR_SHMLOCK 0x00001000
16334	#define VTR_SHMMAP 0x00002000
16335	#define VTR_SHMBAR 0x00004000
16336	#define VTR_SHMUNMAP 0x00008000
16337	#define VTR_OPEN 0x00010000
16338	#define VTR_DELETE 0x00020000
16339	#define VTR_ACCESS 0x00040000
16340	#define VTR_FULLPATH 0x00080000
16341	#define VTR_DLOPEN 0x00100000
16342	#define VTR_DLERR 0x00200000
16343	#define VTR_DLSYM 0x00400000
16344	#define VTR_DLCLOSE 0x00800000
16345	#define VTR_RAND 0x01000000
16346	#define VTR_SLEEP 0x02000000
16347	#define VTR_CURTIME 0x04000000
16348	#define VTR_LASTERR 0x08000000
16349
16350	/*
16351	** Method declarations for vfstrace_file.
16352	*/
16353	static int vfstraceClose(sqlite3_file*);
16354	static int vfstraceRead(sqlite3_file, void, int iAmt, sqlite3_int64 iOfst);
16355	static int vfstraceWrite(sqlite3_file,const void,int iAmt, sqlite3_int64);
16356	static int vfstraceTruncate(sqlite3_file*, sqlite3_int64 size);
16357	static int vfstraceSync(sqlite3_file*, int flags);
16358	static int vfstraceFileSize(sqlite3_file, sqlite3_int64 pSize);
16359	static int vfstraceLock(sqlite3_file*, int);
16360	static int vfstraceUnlock(sqlite3_file*, int);
16361	static int vfstraceCheckReservedLock(sqlite3_file, int );
16362	static int vfstraceFileControl(sqlite3_file, int op, void pArg);
16363	static int vfstraceSectorSize(sqlite3_file*);
16364	static int vfstraceDeviceCharacteristics(sqlite3_file*);
16365	static int vfstraceShmLock(sqlite3_file*,int,int,int);
16366	static int vfstraceShmMap(sqlite3_file,int,int,int, void volatile *);
16367	static void vfstraceShmBarrier(sqlite3_file*);
16368	static int vfstraceShmUnmap(sqlite3_file*,int);
16369
16370	/*
16371	** Method declarations for vfstrace_vfs.
16372	*/
16373	static int vfstraceOpen(sqlite3_vfs, const char , sqlite3_file, int , int );
16374	static int vfstraceDelete(sqlite3_vfs, const char zName, int syncDir);
16375	static int vfstraceAccess(sqlite3_vfs, const char zName, int flags, int *);
16376	static int vfstraceFullPathname(sqlite3_vfs, const char zName, int, char *);
16377	static void vfstraceDlOpen(sqlite3_vfs, const char *zFilename);
16378	static void vfstraceDlError(sqlite3_vfs, int nByte, char zErrMsg);
16379	static void (vfstraceDlSym(sqlite3_vfs,void, const char zSymbol))(void);
16380	static void vfstraceDlClose(sqlite3_vfs, void);
16381	static int vfstraceRandomness(sqlite3_vfs, int nByte, char zOut);
16382	static int vfstraceSleep(sqlite3_vfs*, int microseconds);
16383	static int vfstraceCurrentTime(sqlite3_vfs, double);
16384	static int vfstraceGetLastError(sqlite3_vfs, int, char);
16385	static int vfstraceCurrentTimeInt64(sqlite3_vfs, sqlite3_int64);
16386	static int vfstraceSetSystemCall(sqlite3_vfs,const char, sqlite3_syscall_ptr);
16387	static sqlite3_syscall_ptr vfstraceGetSystemCall(sqlite3_vfs, const char );
16388	static const char vfstraceNextSystemCall(sqlite3_vfs, const char *zName);
16389
16390	/*
16391	** Return a pointer to the tail of the pathname. Examples:
16392	**
16393	** /home/drh/xyzzy.txt -> xyzzy.txt
16394	** xyzzy.txt -> xyzzy.txt
16395	*/
16396	static const char fileTail(const char z){
16397	size_t i;
16398	if( z==0 ) return 0;
16399	i = strlen(z)-1;
16400	while( i>0 && z[i-1]!='/' ){ i--; }
16401	return &z[i];
16402	}
16403
16404	/*
16405	** Send trace output defined by zFormat and subsequent arguments.
16406	*/
16407	static void vfstrace_printf(
16408	vfstrace_info *pInfo,
16409	const char *zFormat,
16410	...
16411	){
16412	va_list ap;
16413	char *zMsg;
16414	if( pInfo->bOn ){
16415	va_start(ap, zFormat);
16416	zMsg = sqlite3_vmprintf(zFormat, ap);
16417	va_end(ap);
16418	pInfo->xOut(zMsg, pInfo->pOutArg);
16419	sqlite3_free(zMsg);
16420	}
16421	}
16422
16423	/*
16424	** Try to convert an error code into a symbolic name for that error code.
16425	*/
16426	static const char *vfstrace_errcode_name(int rc ){
16427	const char *zVal = 0;
16428	switch( rc ){
16429	case SQLITE_OK: zVal = "SQLITE_OK"; break;
16430	case SQLITE_INTERNAL: zVal = "SQLITE_INTERNAL"; break;
16431	case SQLITE_ERROR: zVal = "SQLITE_ERROR"; break;
16432	case SQLITE_PERM: zVal = "SQLITE_PERM"; break;
16433	case SQLITE_ABORT: zVal = "SQLITE_ABORT"; break;
16434	case SQLITE_BUSY: zVal = "SQLITE_BUSY"; break;
16435	case SQLITE_LOCKED: zVal = "SQLITE_LOCKED"; break;
16436	case SQLITE_NOMEM: zVal = "SQLITE_NOMEM"; break;
16437	case SQLITE_READONLY: zVal = "SQLITE_READONLY"; break;
16438	case SQLITE_INTERRUPT: zVal = "SQLITE_INTERRUPT"; break;
16439	case SQLITE_IOERR: zVal = "SQLITE_IOERR"; break;
16440	case SQLITE_CORRUPT: zVal = "SQLITE_CORRUPT"; break;
16441	case SQLITE_NOTFOUND: zVal = "SQLITE_NOTFOUND"; break;
16442	case SQLITE_FULL: zVal = "SQLITE_FULL"; break;
16443	case SQLITE_CANTOPEN: zVal = "SQLITE_CANTOPEN"; break;
16444	case SQLITE_PROTOCOL: zVal = "SQLITE_PROTOCOL"; break;
16445	case SQLITE_EMPTY: zVal = "SQLITE_EMPTY"; break;
16446	case SQLITE_SCHEMA: zVal = "SQLITE_SCHEMA"; break;
16447	case SQLITE_TOOBIG: zVal = "SQLITE_TOOBIG"; break;
16448	case SQLITE_CONSTRAINT: zVal = "SQLITE_CONSTRAINT"; break;
16449	case SQLITE_MISMATCH: zVal = "SQLITE_MISMATCH"; break;
16450	case SQLITE_MISUSE: zVal = "SQLITE_MISUSE"; break;
16451	case SQLITE_NOLFS: zVal = "SQLITE_NOLFS"; break;
16452	case SQLITE_IOERR_READ: zVal = "SQLITE_IOERR_READ"; break;
16453	case SQLITE_IOERR_SHORT_READ: zVal = "SQLITE_IOERR_SHORT_READ"; break;
16454	case SQLITE_IOERR_WRITE: zVal = "SQLITE_IOERR_WRITE"; break;
16455	case SQLITE_IOERR_FSYNC: zVal = "SQLITE_IOERR_FSYNC"; break;
16456	case SQLITE_IOERR_DIR_FSYNC: zVal = "SQLITE_IOERR_DIR_FSYNC"; break;
16457	case SQLITE_IOERR_TRUNCATE: zVal = "SQLITE_IOERR_TRUNCATE"; break;
16458	case SQLITE_IOERR_FSTAT: zVal = "SQLITE_IOERR_FSTAT"; break;
16459	case SQLITE_IOERR_UNLOCK: zVal = "SQLITE_IOERR_UNLOCK"; break;
16460	case SQLITE_IOERR_RDLOCK: zVal = "SQLITE_IOERR_RDLOCK"; break;
16461	case SQLITE_IOERR_DELETE: zVal = "SQLITE_IOERR_DELETE"; break;
16462	case SQLITE_IOERR_BLOCKED: zVal = "SQLITE_IOERR_BLOCKED"; break;
16463	case SQLITE_IOERR_NOMEM: zVal = "SQLITE_IOERR_NOMEM"; break;
16464	case SQLITE_IOERR_ACCESS: zVal = "SQLITE_IOERR_ACCESS"; break;
16465	case SQLITE_IOERR_CHECKRESERVEDLOCK:
16466	zVal = "SQLITE_IOERR_CHECKRESERVEDLOCK"; break;
16467	case SQLITE_IOERR_LOCK: zVal = "SQLITE_IOERR_LOCK"; break;
16468	case SQLITE_IOERR_CLOSE: zVal = "SQLITE_IOERR_CLOSE"; break;
16469	case SQLITE_IOERR_DIR_CLOSE: zVal = "SQLITE_IOERR_DIR_CLOSE"; break;
16470	case SQLITE_IOERR_SHMOPEN: zVal = "SQLITE_IOERR_SHMOPEN"; break;
16471	case SQLITE_IOERR_SHMSIZE: zVal = "SQLITE_IOERR_SHMSIZE"; break;
16472	case SQLITE_IOERR_SHMLOCK: zVal = "SQLITE_IOERR_SHMLOCK"; break;
16473	case SQLITE_IOERR_SHMMAP: zVal = "SQLITE_IOERR_SHMMAP"; break;
16474	case SQLITE_IOERR_SEEK: zVal = "SQLITE_IOERR_SEEK"; break;
16475	case SQLITE_IOERR_GETTEMPPATH: zVal = "SQLITE_IOERR_GETTEMPPATH"; break;
16476	case SQLITE_IOERR_CONVPATH: zVal = "SQLITE_IOERR_CONVPATH"; break;
16477	case SQLITE_READONLY_DBMOVED: zVal = "SQLITE_READONLY_DBMOVED"; break;
16478	case SQLITE_LOCKED_SHAREDCACHE: zVal = "SQLITE_LOCKED_SHAREDCACHE"; break;
16479	case SQLITE_BUSY_RECOVERY: zVal = "SQLITE_BUSY_RECOVERY"; break;
16480	case SQLITE_CANTOPEN_NOTEMPDIR: zVal = "SQLITE_CANTOPEN_NOTEMPDIR"; break;
16481	}
16482	return zVal;
16483	}
16484
16485	/*
16486	** Convert value rc into a string and print it using zFormat. zFormat
16487	** should have exactly one %s
16488	*/
16489	static void vfstrace_print_errcode(
16490	vfstrace_info *pInfo,
16491	const char *zFormat,
16492	int rc
16493	){
16494	const char *zVal;
16495	char zBuf[50];
16496	zVal = vfstrace_errcode_name(rc);
16497	if( zVal==0 ){
16498	zVal = vfstrace_errcode_name(rc&0xff);
16499	if( zVal ){
16500	sqlite3_snprintf(sizeof(zBuf), zBuf, "%s \| 0x%x", zVal, rc&0xffff00);
16501	}else{
16502	sqlite3_snprintf(sizeof(zBuf), zBuf, "%d (0x%x)", rc, rc);
16503	}
16504	zVal = zBuf;
16505	}
16506	vfstrace_printf(pInfo, zFormat, zVal);
16507	}
16508
16509	/*
16510	** Append to a buffer.
16511	*/
16512	static void strappend(char z, int pI, const char *zAppend){
16513	int i = *pI;
16514	while( zAppend[0] ){ z[i++] = *(zAppend++); }
16515	z[i] = 0;
16516	*pI = i;
16517	}
16518
16519	/*
16520	** Turn tracing output on or off according to mMask.
16521	*/
16522	static void vfstraceOnOff(vfstrace_info *pInfo, unsigned int mMask){
16523	pInfo->bOn = (pInfo->mTrace & mMask)!=0;
16524	}
16525
16526	/*
16527	** Close an vfstrace-file.
16528	*/
16529	static int vfstraceClose(sqlite3_file *pFile){
16530	vfstrace_file p = (vfstrace_file )pFile;
16531	vfstrace_info *pInfo = p->pInfo;
16532	int rc;
16533	vfstraceOnOff(pInfo, VTR_CLOSE);
16534	vfstrace_printf(pInfo, "%s.xClose(%s)", pInfo->zVfsName, p->zFName);
16535	rc = p->pReal->pMethods->xClose(p->pReal);
16536	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16537	if( rc==SQLITE_OK ){
16538	sqlite3_free((void*)p->base.pMethods);
16539	p->base.pMethods = 0;
16540	}
16541	return rc;
16542	}
16543
16544	/*
16545	** Read data from an vfstrace-file.
16546	*/
16547	static int vfstraceRead(
16548	sqlite3_file *pFile,
16549	void *zBuf,
16550	int iAmt,
16551	sqlite_int64 iOfst
16552	){
16553	vfstrace_file p = (vfstrace_file )pFile;
16554	vfstrace_info *pInfo = p->pInfo;
16555	int rc;
16556	vfstraceOnOff(pInfo, VTR_READ);
16557	vfstrace_printf(pInfo, "%s.xRead(%s,n=%d,ofst=%lld)",
16558	pInfo->zVfsName, p->zFName, iAmt, iOfst);
16559	rc = p->pReal->pMethods->xRead(p->pReal, zBuf, iAmt, iOfst);
16560	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16561	return rc;
16562	}
16563
16564	/*
16565	** Write data to an vfstrace-file.
16566	*/
16567	static int vfstraceWrite(
16568	sqlite3_file *pFile,
16569	const void *zBuf,
16570	int iAmt,
16571	sqlite_int64 iOfst
16572	){
16573	vfstrace_file p = (vfstrace_file )pFile;
16574	vfstrace_info *pInfo = p->pInfo;
16575	int rc;
16576	vfstraceOnOff(pInfo, VTR_WRITE);
16577	vfstrace_printf(pInfo, "%s.xWrite(%s,n=%d,ofst=%lld)",
16578	pInfo->zVfsName, p->zFName, iAmt, iOfst);
16579	rc = p->pReal->pMethods->xWrite(p->pReal, zBuf, iAmt, iOfst);
16580	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16581	return rc;
16582	}
16583
16584	/*
16585	** Truncate an vfstrace-file.
16586	*/
16587	static int vfstraceTruncate(sqlite3_file *pFile, sqlite_int64 size){
16588	vfstrace_file p = (vfstrace_file )pFile;
16589	vfstrace_info *pInfo = p->pInfo;
16590	int rc;
16591	vfstraceOnOff(pInfo, VTR_TRUNC);
16592	vfstrace_printf(pInfo, "%s.xTruncate(%s,%lld)", pInfo->zVfsName, p->zFName,
16593	size);
16594	rc = p->pReal->pMethods->xTruncate(p->pReal, size);
16595	vfstrace_printf(pInfo, " -> %d\n", rc);
16596	return rc;
16597	}
16598
16599	/*
16600	** Sync an vfstrace-file.
16601	*/
16602	static int vfstraceSync(sqlite3_file *pFile, int flags){
16603	vfstrace_file p = (vfstrace_file )pFile;
16604	vfstrace_info *pInfo = p->pInfo;
16605	int rc;
16606	int i;
16607	char zBuf[100];
16608	memcpy(zBuf, "\|0", 3);
16609	i = 0;
16610	if( flags & SQLITE_SYNC_FULL ) strappend(zBuf, &i, "\|FULL");
16611	else if( flags & SQLITE_SYNC_NORMAL ) strappend(zBuf, &i, "\|NORMAL");
16612	if( flags & SQLITE_SYNC_DATAONLY ) strappend(zBuf, &i, "\|DATAONLY");
16613	if( flags & ~(SQLITE_SYNC_FULL\|SQLITE_SYNC_DATAONLY) ){
16614	sqlite3_snprintf(sizeof(zBuf)-i, &zBuf[i], "\|0x%x", flags);
16615	}
16616	vfstraceOnOff(pInfo, VTR_SYNC);
16617	vfstrace_printf(pInfo, "%s.xSync(%s,%s)", pInfo->zVfsName, p->zFName,
16618	&zBuf[1]);
16619	rc = p->pReal->pMethods->xSync(p->pReal, flags);
16620	vfstrace_printf(pInfo, " -> %d\n", rc);
16621	return rc;
16622	}
16623
16624	/*
16625	** Return the current file-size of an vfstrace-file.
16626	*/
16627	static int vfstraceFileSize(sqlite3_file pFile, sqlite_int64 pSize){
16628	vfstrace_file p = (vfstrace_file )pFile;
16629	vfstrace_info *pInfo = p->pInfo;
16630	int rc;
16631	vfstraceOnOff(pInfo, VTR_FSIZE);
16632	vfstrace_printf(pInfo, "%s.xFileSize(%s)", pInfo->zVfsName, p->zFName);
16633	rc = p->pReal->pMethods->xFileSize(p->pReal, pSize);
16634	vfstrace_print_errcode(pInfo, " -> %s,", rc);
16635	vfstrace_printf(pInfo, " size=%lld\n", *pSize);
16636	return rc;
16637	}
16638
16639	/*
16640	** Return the name of a lock.
16641	*/
16642	static const char *lockName(int eLock){
16643	const char *azLockNames[] = {
16644	"NONE", "SHARED", "RESERVED", "PENDING", "EXCLUSIVE"
16645	};
16646	if( eLock<0 \|\| eLock>=(int)(sizeof(azLockNames)/sizeof(azLockNames[0])) ){
16647	return "???";
16648	}else{
16649	return azLockNames[eLock];
16650	}
16651	}
16652
16653	/*
16654	** Lock an vfstrace-file.
16655	*/
16656	static int vfstraceLock(sqlite3_file *pFile, int eLock){
16657	vfstrace_file p = (vfstrace_file )pFile;
16658	vfstrace_info *pInfo = p->pInfo;
16659	int rc;
16660	vfstraceOnOff(pInfo, VTR_LOCK);
16661	vfstrace_printf(pInfo, "%s.xLock(%s,%s)", pInfo->zVfsName, p->zFName,
16662	lockName(eLock));
16663	rc = p->pReal->pMethods->xLock(p->pReal, eLock);
16664	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16665	return rc;
16666	}
16667
16668	/*
16669	** Unlock an vfstrace-file.
16670	*/
16671	static int vfstraceUnlock(sqlite3_file *pFile, int eLock){
16672	vfstrace_file p = (vfstrace_file )pFile;
16673	vfstrace_info *pInfo = p->pInfo;
16674	int rc;
16675	vfstraceOnOff(pInfo, VTR_UNLOCK);
16676	vfstrace_printf(pInfo, "%s.xUnlock(%s,%s)", pInfo->zVfsName, p->zFName,
16677	lockName(eLock));
16678	rc = p->pReal->pMethods->xUnlock(p->pReal, eLock);
16679	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16680	return rc;
16681	}
16682
16683	/*
16684	** Check if another file-handle holds a RESERVED lock on an vfstrace-file.
16685	*/
16686	static int vfstraceCheckReservedLock(sqlite3_file pFile, int pResOut){
16687	vfstrace_file p = (vfstrace_file )pFile;
16688	vfstrace_info *pInfo = p->pInfo;
16689	int rc;
16690	vfstraceOnOff(pInfo, VTR_CRL);
16691	vfstrace_printf(pInfo, "%s.xCheckReservedLock(%s,%d)",
16692	pInfo->zVfsName, p->zFName);
16693	rc = p->pReal->pMethods->xCheckReservedLock(p->pReal, pResOut);
16694	vfstrace_print_errcode(pInfo, " -> %s", rc);
16695	vfstrace_printf(pInfo, ", out=%d\n", *pResOut);
16696	return rc;
16697	}
16698
16699	/*
16700	** File control method. For custom operations on an vfstrace-file.
16701	*/
16702	static int vfstraceFileControl(sqlite3_file pFile, int op, void pArg){
16703	vfstrace_file p = (vfstrace_file )pFile;
16704	vfstrace_info *pInfo = p->pInfo;
16705	int rc;
16706	char zBuf[100];
16707	char zBuf2[100];
16708	char *zOp;
16709	char *zRVal = 0;
16710	vfstraceOnOff(pInfo, VTR_FCTRL);
16711	switch( op ){
16712	case SQLITE_FCNTL_LOCKSTATE: zOp = "LOCKSTATE"; break;
16713	case SQLITE_GET_LOCKPROXYFILE: zOp = "GET_LOCKPROXYFILE"; break;
16714	case SQLITE_SET_LOCKPROXYFILE: zOp = "SET_LOCKPROXYFILE"; break;
16715	case SQLITE_LAST_ERRNO: zOp = "LAST_ERRNO"; break;
16716	case SQLITE_FCNTL_SIZE_HINT: {
16717	sqlite3_snprintf(sizeof(zBuf), zBuf, "SIZE_HINT,%lld",
16718	(sqlite3_int64)pArg);
16719	zOp = zBuf;
16720	break;
16721	}
16722	case SQLITE_FCNTL_CHUNK_SIZE: {
16723	sqlite3_snprintf(sizeof(zBuf), zBuf, "CHUNK_SIZE,%d", (int)pArg);
16724	zOp = zBuf;
16725	break;
16726	}
16727	case SQLITE_FCNTL_FILE_POINTER: zOp = "FILE_POINTER"; break;
16728	case SQLITE_FCNTL_WIN32_AV_RETRY: zOp = "WIN32_AV_RETRY"; break;
16729	case SQLITE_FCNTL_PERSIST_WAL: {
16730	sqlite3_snprintf(sizeof(zBuf), zBuf, "PERSIST_WAL,%d", (int)pArg);
16731	zOp = zBuf;
16732	break;
16733	}
16734	case SQLITE_FCNTL_OVERWRITE: zOp = "OVERWRITE"; break;
16735	case SQLITE_FCNTL_VFSNAME: zOp = "VFSNAME"; break;
16736	case SQLITE_FCNTL_POWERSAFE_OVERWRITE: zOp = "POWERSAFE_OVERWRITE"; break;
16737	case SQLITE_FCNTL_PRAGMA: {
16738	const char const a = (const charconst)pArg;
16739	if( a[1] && strcmp(a[1],"vfstrace")==0 && a[2] ){
16740	const u8 zArg = (const u8)a[2];
16741	if( zArg[0]>='0' && zArg[0]<=9 ){
16742	pInfo->mTrace = (sqlite3_uint64)strtoll(a[2], 0, 0);
16743	}else{
16744	static const struct {
16745	const char *z;
16746	unsigned int m;
16747	} aKw[] = {
16748	{ "all", 0xffffffff },
16749	{ "close", VTR_CLOSE },
16750	{ "read", VTR_READ },
16751	{ "write", VTR_WRITE },
16752	{ "truncate", VTR_TRUNC },
16753	{ "sync", VTR_SYNC },
16754	{ "filesize", VTR_FSIZE },
16755	{ "lock", VTR_LOCK },
16756	{ "unlock", VTR_UNLOCK },
16757	{ "checkreservedlock", VTR_CRL },
16758	{ "filecontrol", VTR_FCTRL },
16759	{ "sectorsize", VTR_SECSZ },
16760	{ "devicecharacteristics", VTR_DEVCHAR },
16761	{ "shmlock", VTR_SHMLOCK },
16762	{ "shmmap", VTR_SHMMAP },
16763	{ "shmummap", VTR_SHMUNMAP },
16764	{ "shmbarrier", VTR_SHMBAR },
16765	{ "open", VTR_OPEN },
16766	{ "delete", VTR_DELETE },
16767	{ "access", VTR_ACCESS },
16768	{ "fullpathname", VTR_FULLPATH },
16769	{ "dlopen", VTR_DLOPEN },
16770	{ "dlerror", VTR_DLERR },
16771	{ "dlsym", VTR_DLSYM },
16772	{ "dlclose", VTR_DLCLOSE },
16773	{ "randomness", VTR_RAND },
16774	{ "sleep", VTR_SLEEP },
16775	{ "currenttime", VTR_CURTIME },
16776	{ "currenttimeint64", VTR_CURTIME },
16777	{ "getlasterror", VTR_LASTERR },
16778	};
16779	int onOff = 1;
16780	while( zArg[0] ){
16781	int jj, n;
16782	while( zArg[0]!=0 && zArg[0]!='-' && zArg[0]!='+'
16783	&& !isalpha(zArg[0]) ) zArg++;
16784	if( zArg[0]==0 ) break;
16785	if( zArg[0]=='-' ){
16786	onOff = 0;
16787	zArg++;
16788	}else if( zArg[0]=='+' ){
16789	onOff = 1;
16790	zArg++;
16791	}
16792	while( !isalpha(zArg[0]) ){
16793	if( zArg[0]==0 ) break;
16794	zArg++;
16795	}
16796	if( zArg[0]=='x' && isalpha(zArg[1]) ) zArg++;
16797	for(n=0; isalpha(zArg[n]); n++){}
16798	for(jj=0; jj<(int)(sizeof(aKw)/sizeof(aKw[0])); jj++){
16799	if( sqlite3_strnicmp(aKw[jj].z,(const char*)zArg,n)==0 ){
16800	if( onOff ){
16801	pInfo->mTrace \|= aKw[jj].m;
16802	}else{
16803	pInfo->mTrace &= ~aKw[jj].m;
16804	}
16805	break;
16806	}
16807	}
16808	zArg += n;
16809	}
16810	}
16811	}
16812	sqlite3_snprintf(sizeof(zBuf), zBuf, "PRAGMA,[%s,%s]",a[1],a[2]);
16813	zOp = zBuf;
16814	break;
16815	}
16816	case SQLITE_FCNTL_BUSYHANDLER: zOp = "BUSYHANDLER"; break;
16817	case SQLITE_FCNTL_TEMPFILENAME: zOp = "TEMPFILENAME"; break;
16818	case SQLITE_FCNTL_MMAP_SIZE: {
16819	sqlite3_int64 iMMap = (sqlite3_int64)pArg;
16820	sqlite3_snprintf(sizeof(zBuf), zBuf, "MMAP_SIZE,%lld",iMMap);
16821	zOp = zBuf;
16822	break;
16823	}
16824	case SQLITE_FCNTL_TRACE: zOp = "TRACE"; break;
16825	case SQLITE_FCNTL_HAS_MOVED: zOp = "HAS_MOVED"; break;
16826	case SQLITE_FCNTL_SYNC: zOp = "SYNC"; break;
16827	case SQLITE_FCNTL_COMMIT_PHASETWO: zOp = "COMMIT_PHASETWO"; break;
16828	case SQLITE_FCNTL_WIN32_SET_HANDLE: zOp = "WIN32_SET_HANDLE"; break;
16829	case SQLITE_FCNTL_WAL_BLOCK: zOp = "WAL_BLOCK"; break;
16830	case SQLITE_FCNTL_ZIPVFS: zOp = "ZIPVFS"; break;
16831	case SQLITE_FCNTL_RBU: zOp = "RBU"; break;
16832	case SQLITE_FCNTL_VFS_POINTER: zOp = "VFS_POINTER"; break;
16833	case SQLITE_FCNTL_JOURNAL_POINTER: zOp = "JOURNAL_POINTER"; break;
16834	case SQLITE_FCNTL_WIN32_GET_HANDLE: zOp = "WIN32_GET_HANDLE"; break;
16835	case SQLITE_FCNTL_PDB: zOp = "PDB"; break;
16836	case SQLITE_FCNTL_BEGIN_ATOMIC_WRITE: zOp = "BEGIN_ATOMIC_WRITE"; break;
16837	case SQLITE_FCNTL_COMMIT_ATOMIC_WRITE: zOp = "COMMIT_ATOMIC_WRITE"; break;
16838	case SQLITE_FCNTL_ROLLBACK_ATOMIC_WRITE: {
16839	zOp = "ROLLBACK_ATOMIC_WRITE";
16840	break;
16841	}
16842	case SQLITE_FCNTL_LOCK_TIMEOUT: {
16843	sqlite3_snprintf(sizeof(zBuf), zBuf, "LOCK_TIMEOUT,%d", (int)pArg);
16844	zOp = zBuf;
16845	break;
16846	}
16847	case SQLITE_FCNTL_DATA_VERSION: zOp = "DATA_VERSION"; break;
16848	case SQLITE_FCNTL_SIZE_LIMIT: zOp = "SIZE_LIMIT"; break;
16849	case SQLITE_FCNTL_CKPT_DONE: zOp = "CKPT_DONE"; break;
16850	case SQLITE_FCNTL_RESERVE_BYTES: zOp = "RESERVED_BYTES"; break;
16851	case SQLITE_FCNTL_CKPT_START: zOp = "CKPT_START"; break;
16852	case SQLITE_FCNTL_EXTERNAL_READER: zOp = "EXTERNAL_READER"; break;
16853	case SQLITE_FCNTL_CKSM_FILE: zOp = "CKSM_FILE"; break;
16854	case SQLITE_FCNTL_RESET_CACHE: zOp = "RESET_CACHE"; break;
16855	case 0xca093fa0: zOp = "DB_UNCHANGED"; break;
16856	default: {
16857	sqlite3_snprintf(sizeof zBuf, zBuf, "%d", op);
16858	zOp = zBuf;
16859	break;
16860	}
16861	}
16862	vfstrace_printf(pInfo, "%s.xFileControl(%s,%s)",
16863	pInfo->zVfsName, p->zFName, zOp);
16864	rc = p->pReal->pMethods->xFileControl(p->pReal, op, pArg);
16865	if( rc==SQLITE_OK ){
16866	switch( op ){
16867	case SQLITE_FCNTL_VFSNAME: {
16868	(char*)pArg = sqlite3_mprintf("vfstrace.%s/%z",
16869	pInfo->zVfsName, (char*)pArg);
16870	zRVal = (char*)pArg;
16871	break;
16872	}
16873	case SQLITE_FCNTL_MMAP_SIZE: {
16874	sqlite3_snprintf(sizeof(zBuf2), zBuf2, "%lld", (sqlite3_int64)pArg);
16875	zRVal = zBuf2;
16876	break;
16877	}
16878	case SQLITE_FCNTL_HAS_MOVED:
16879	case SQLITE_FCNTL_PERSIST_WAL: {
16880	sqlite3_snprintf(sizeof(zBuf2), zBuf2, "%d", (int)pArg);
16881	zRVal = zBuf2;
16882	break;
16883	}
16884	case SQLITE_FCNTL_PRAGMA:
16885	case SQLITE_FCNTL_TEMPFILENAME: {
16886	zRVal = (char*)pArg;
16887	break;
16888	}
16889	}
16890	}
16891	if( zRVal ){
16892	vfstrace_print_errcode(pInfo, " -> %s", rc);
16893	vfstrace_printf(pInfo, ", %s\n", zRVal);
16894	}else{
16895	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16896	}
16897	return rc;
16898	}
16899
16900	/*
16901	** Return the sector-size in bytes for an vfstrace-file.
16902	*/
16903	static int vfstraceSectorSize(sqlite3_file *pFile){
16904	vfstrace_file p = (vfstrace_file )pFile;
16905	vfstrace_info *pInfo = p->pInfo;
16906	int rc;
16907	vfstraceOnOff(pInfo, VTR_SECSZ);
16908	vfstrace_printf(pInfo, "%s.xSectorSize(%s)", pInfo->zVfsName, p->zFName);
16909	rc = p->pReal->pMethods->xSectorSize(p->pReal);
16910	vfstrace_printf(pInfo, " -> %d\n", rc);
16911	return rc;
16912	}
16913
16914	/*
16915	** Return the device characteristic flags supported by an vfstrace-file.
16916	*/
16917	static int vfstraceDeviceCharacteristics(sqlite3_file *pFile){
16918	vfstrace_file p = (vfstrace_file )pFile;
16919	vfstrace_info *pInfo = p->pInfo;
16920	int rc;
16921	vfstraceOnOff(pInfo, VTR_DEVCHAR);
16922	vfstrace_printf(pInfo, "%s.xDeviceCharacteristics(%s)",
16923	pInfo->zVfsName, p->zFName);
16924	rc = p->pReal->pMethods->xDeviceCharacteristics(p->pReal);
16925	vfstrace_printf(pInfo, " -> 0x%08x\n", rc);
16926	return rc;
16927	}
16928
16929	/*
16930	** Shared-memory operations.
16931	*/
16932	static int vfstraceShmLock(sqlite3_file *pFile, int ofst, int n, int flags){
16933	static const char *azLockName[] = {
16934	"WRITE",
16935	"CKPT",
16936	"RECOVER",
16937	"READ0",
16938	"READ1",
16939	"READ2",
16940	"READ3",
16941	"READ4",
16942	};
16943	vfstrace_file p = (vfstrace_file )pFile;
16944	vfstrace_info *pInfo = p->pInfo;
16945	int rc;
16946	char zLck[100];
16947	int i = 0;
16948	vfstraceOnOff(pInfo, VTR_SHMLOCK);
16949	memcpy(zLck, "\|0", 3);
16950	if( flags & SQLITE_SHM_UNLOCK ) strappend(zLck, &i, "\|UNLOCK");
16951	if( flags & SQLITE_SHM_LOCK ) strappend(zLck, &i, "\|LOCK");
16952	if( flags & SQLITE_SHM_SHARED ) strappend(zLck, &i, "\|SHARED");
16953	if( flags & SQLITE_SHM_EXCLUSIVE ) strappend(zLck, &i, "\|EXCLUSIVE");
16954	if( flags & ~(0xf) ){
16955	sqlite3_snprintf(sizeof(zLck)-i, &zLck[i], "\|0x%x", flags);
16956	}
16957	if( ofst>=0 && ofst<(int)(sizeof(azLockName)/sizeof(azLockName[0])) ){
16958	vfstrace_printf(pInfo, "%s.xShmLock(%s,ofst=%d(%s),n=%d,%s)",
16959	pInfo->zVfsName, p->zFName, ofst, azLockName[ofst],
16960	n, &zLck[1]);
16961	}else{
16962	vfstrace_printf(pInfo, "%s.xShmLock(%s,ofst=5d,n=%d,%s)",
16963	pInfo->zVfsName, p->zFName, ofst,
16964	n, &zLck[1]);
16965	}
16966	rc = p->pReal->pMethods->xShmLock(p->pReal, ofst, n, flags);
16967	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16968	return rc;
16969	}
16970	static int vfstraceShmMap(
16971	sqlite3_file *pFile,
16972	int iRegion,
16973	int szRegion,
16974	int isWrite,
16975	void volatile **pp
16976	){
16977	vfstrace_file p = (vfstrace_file )pFile;
16978	vfstrace_info *pInfo = p->pInfo;
16979	int rc;
16980	vfstraceOnOff(pInfo, VTR_SHMMAP);
16981	vfstrace_printf(pInfo, "%s.xShmMap(%s,iRegion=%d,szRegion=%d,isWrite=%d,*)",
16982	pInfo->zVfsName, p->zFName, iRegion, szRegion, isWrite);
16983	rc = p->pReal->pMethods->xShmMap(p->pReal, iRegion, szRegion, isWrite, pp);
16984	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
16985	return rc;
16986	}
16987	static void vfstraceShmBarrier(sqlite3_file *pFile){
16988	vfstrace_file p = (vfstrace_file )pFile;
16989	vfstrace_info *pInfo = p->pInfo;
16990	vfstraceOnOff(pInfo, VTR_SHMBAR);
16991	vfstrace_printf(pInfo, "%s.xShmBarrier(%s)\n", pInfo->zVfsName, p->zFName);
16992	p->pReal->pMethods->xShmBarrier(p->pReal);
16993	}
16994	static int vfstraceShmUnmap(sqlite3_file *pFile, int delFlag){
16995	vfstrace_file p = (vfstrace_file )pFile;
16996	vfstrace_info *pInfo = p->pInfo;
16997	int rc;
16998	vfstraceOnOff(pInfo, VTR_SHMUNMAP);
16999	vfstrace_printf(pInfo, "%s.xShmUnmap(%s,delFlag=%d)",
17000	pInfo->zVfsName, p->zFName, delFlag);
17001	rc = p->pReal->pMethods->xShmUnmap(p->pReal, delFlag);
17002	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
17003	return rc;
17004	}
17005
17006
17007
17008	/*
17009	** Open an vfstrace file handle.
17010	*/
17011	static int vfstraceOpen(
17012	sqlite3_vfs *pVfs,
17013	const char *zName,
17014	sqlite3_file *pFile,
17015	int flags,
17016	int *pOutFlags
17017	){
17018	int rc;
17019	vfstrace_file p = (vfstrace_file )pFile;
17020	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17021	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17022	p->pInfo = pInfo;
17023	p->zFName = zName ? fileTail(zName) : "<temp>";
17024	p->pReal = (sqlite3_file *)&p[1];
17025	rc = pRoot->xOpen(pRoot, zName, p->pReal, flags, pOutFlags);
17026	vfstraceOnOff(pInfo, VTR_OPEN);
17027	vfstrace_printf(pInfo, "%s.xOpen(%s,flags=0x%x)",
17028	pInfo->zVfsName, p->zFName, flags);
17029	if( p->pReal->pMethods ){
17030	sqlite3_io_methods pNew = sqlite3_malloc( sizeof(pNew) );
17031	const sqlite3_io_methods *pSub = p->pReal->pMethods;
17032	memset(pNew, 0, sizeof(*pNew));
17033	pNew->iVersion = pSub->iVersion;
17034	pNew->xClose = vfstraceClose;
17035	pNew->xRead = vfstraceRead;
17036	pNew->xWrite = vfstraceWrite;
17037	pNew->xTruncate = vfstraceTruncate;
17038	pNew->xSync = vfstraceSync;
17039	pNew->xFileSize = vfstraceFileSize;
17040	pNew->xLock = vfstraceLock;
17041	pNew->xUnlock = vfstraceUnlock;
17042	pNew->xCheckReservedLock = vfstraceCheckReservedLock;
17043	pNew->xFileControl = vfstraceFileControl;
17044	pNew->xSectorSize = vfstraceSectorSize;
17045	pNew->xDeviceCharacteristics = vfstraceDeviceCharacteristics;
17046	if( pNew->iVersion>=2 ){
17047	pNew->xShmMap = pSub->xShmMap ? vfstraceShmMap : 0;
17048	pNew->xShmLock = pSub->xShmLock ? vfstraceShmLock : 0;
17049	pNew->xShmBarrier = pSub->xShmBarrier ? vfstraceShmBarrier : 0;
17050	pNew->xShmUnmap = pSub->xShmUnmap ? vfstraceShmUnmap : 0;
17051	}
17052	pFile->pMethods = pNew;
17053	}
17054	vfstrace_print_errcode(pInfo, " -> %s", rc);
17055	if( pOutFlags ){
17056	vfstrace_printf(pInfo, ", outFlags=0x%x\n", *pOutFlags);
17057	}else{
17058	vfstrace_printf(pInfo, "\n");
17059	}
17060	return rc;
17061	}
17062
17063	/*
17064	** Delete the file located at zPath. If the dirSync argument is true,
17065	** ensure the file-system modifications are synced to disk before
17066	** returning.
17067	*/
17068	static int vfstraceDelete(sqlite3_vfs pVfs, const char zPath, int dirSync){
17069	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17070	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17071	int rc;
17072	vfstraceOnOff(pInfo, VTR_DELETE);
17073	vfstrace_printf(pInfo, "%s.xDelete(\"%s\",%d)",
17074	pInfo->zVfsName, zPath, dirSync);
17075	rc = pRoot->xDelete(pRoot, zPath, dirSync);
17076	vfstrace_print_errcode(pInfo, " -> %s\n", rc);
17077	return rc;
17078	}
17079
17080	/*
17081	** Test for access permissions. Return true if the requested permission
17082	** is available, or false otherwise.
17083	*/
17084	static int vfstraceAccess(
17085	sqlite3_vfs *pVfs,
17086	const char *zPath,
17087	int flags,
17088	int *pResOut
17089	){
17090	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17091	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17092	int rc;
17093	vfstraceOnOff(pInfo, VTR_ACCESS);
17094	vfstrace_printf(pInfo, "%s.xAccess(\"%s\",%d)",
17095	pInfo->zVfsName, zPath, flags);
17096	rc = pRoot->xAccess(pRoot, zPath, flags, pResOut);
17097	vfstrace_print_errcode(pInfo, " -> %s", rc);
17098	vfstrace_printf(pInfo, ", out=%d\n", *pResOut);
17099	return rc;
17100	}
17101
17102	/*
17103	** Populate buffer zOut with the full canonical pathname corresponding
17104	** to the pathname in zPath. zOut is guaranteed to point to a buffer
17105	** of at least (DEVSYM_MAX_PATHNAME+1) bytes.
17106	*/
17107	static int vfstraceFullPathname(
17108	sqlite3_vfs *pVfs,
17109	const char *zPath,
17110	int nOut,
17111	char *zOut
17112	){
17113	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17114	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17115	int rc;
17116	vfstraceOnOff(pInfo, VTR_FULLPATH);
17117	vfstrace_printf(pInfo, "%s.xFullPathname(\"%s\")",
17118	pInfo->zVfsName, zPath);
17119	rc = pRoot->xFullPathname(pRoot, zPath, nOut, zOut);
17120	vfstrace_print_errcode(pInfo, " -> %s", rc);
17121	vfstrace_printf(pInfo, ", out=\"%.*s\"\n", nOut, zOut);
17122	return rc;
17123	}
17124
17125	/*
17126	** Open the dynamic library located at zPath and return a handle.
17127	*/
17128	static void vfstraceDlOpen(sqlite3_vfs pVfs, const char *zPath){
17129	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17130	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17131	vfstraceOnOff(pInfo, VTR_DLOPEN);
17132	vfstrace_printf(pInfo, "%s.xDlOpen(\"%s\")\n", pInfo->zVfsName, zPath);
17133	return pRoot->xDlOpen(pRoot, zPath);
17134	}
17135
17136	/*
17137	** Populate the buffer zErrMsg (size nByte bytes) with a human readable
17138	** utf-8 string describing the most recent error encountered associated
17139	** with dynamic libraries.
17140	*/
17141	static void vfstraceDlError(sqlite3_vfs pVfs, int nByte, char zErrMsg){
17142	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17143	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17144	vfstraceOnOff(pInfo, VTR_DLERR);
17145	vfstrace_printf(pInfo, "%s.xDlError(%d)", pInfo->zVfsName, nByte);
17146	pRoot->xDlError(pRoot, nByte, zErrMsg);
17147	vfstrace_printf(pInfo, " -> \"%s\"", zErrMsg);
17148	}
17149
17150	/*
17151	** Return a pointer to the symbol zSymbol in the dynamic library pHandle.
17152	*/
17153	static void (vfstraceDlSym(sqlite3_vfs pVfs,void p,const char zSym))(void){
17154	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17155	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17156	vfstrace_printf(pInfo, "%s.xDlSym(\"%s\")\n", pInfo->zVfsName, zSym);
17157	return pRoot->xDlSym(pRoot, p, zSym);
17158	}
17159
17160	/*
17161	** Close the dynamic library handle pHandle.
17162	*/
17163	static void vfstraceDlClose(sqlite3_vfs pVfs, void pHandle){
17164	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17165	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17166	vfstraceOnOff(pInfo, VTR_DLCLOSE);
17167	vfstrace_printf(pInfo, "%s.xDlOpen()\n", pInfo->zVfsName);
17168	pRoot->xDlClose(pRoot, pHandle);
17169	}
17170
17171	/*
17172	** Populate the buffer pointed to by zBufOut with nByte bytes of
17173	** random data.
17174	*/
17175	static int vfstraceRandomness(sqlite3_vfs pVfs, int nByte, char zBufOut){
17176	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17177	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17178	vfstraceOnOff(pInfo, VTR_RAND);
17179	vfstrace_printf(pInfo, "%s.xRandomness(%d)\n", pInfo->zVfsName, nByte);
17180	return pRoot->xRandomness(pRoot, nByte, zBufOut);
17181	}
17182
17183	/*
17184	** Sleep for nMicro microseconds. Return the number of microseconds
17185	** actually slept.
17186	*/
17187	static int vfstraceSleep(sqlite3_vfs *pVfs, int nMicro){
17188	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17189	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17190	vfstraceOnOff(pInfo, VTR_SLEEP);
17191	vfstrace_printf(pInfo, "%s.xSleep(%d)\n", pInfo->zVfsName, nMicro);
17192	return pRoot->xSleep(pRoot, nMicro);
17193	}
17194
17195	/*
17196	** Return the current time as a Julian Day number in *pTimeOut.
17197	*/
17198	static int vfstraceCurrentTime(sqlite3_vfs pVfs, double pTimeOut){
17199	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17200	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17201	int rc;
17202	vfstraceOnOff(pInfo, VTR_CURTIME);
17203	vfstrace_printf(pInfo, "%s.xCurrentTime()", pInfo->zVfsName);
17204	rc = pRoot->xCurrentTime(pRoot, pTimeOut);
17205	vfstrace_printf(pInfo, " -> %.17g\n", *pTimeOut);
17206	return rc;
17207	}
17208	static int vfstraceCurrentTimeInt64(sqlite3_vfs pVfs, sqlite3_int64 pTimeOut){
17209	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17210	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17211	int rc;
17212	vfstraceOnOff(pInfo, VTR_CURTIME);
17213	vfstrace_printf(pInfo, "%s.xCurrentTimeInt64()", pInfo->zVfsName);
17214	rc = pRoot->xCurrentTimeInt64(pRoot, pTimeOut);
17215	vfstrace_printf(pInfo, " -> %lld\n", *pTimeOut);
17216	return rc;
17217	}
17218
17219	/*
17220	** Return the most recent error code and message
17221	*/
17222	static int vfstraceGetLastError(sqlite3_vfs pVfs, int nErr, char zErr){
17223	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17224	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17225	int rc;
17226	vfstraceOnOff(pInfo, VTR_LASTERR);
17227	vfstrace_printf(pInfo, "%s.xGetLastError(%d,zBuf)", pInfo->zVfsName, nErr);
17228	if( nErr ) zErr[0] = 0;
17229	rc = pRoot->xGetLastError(pRoot, nErr, zErr);
17230	vfstrace_printf(pInfo, " -> zBuf[] = \"%s\", rc = %d\n", nErr?zErr:"", rc);
17231	return rc;
17232	}
17233
17234	/*
17235	** Override system calls.
17236	*/
17237	static int vfstraceSetSystemCall(
17238	sqlite3_vfs *pVfs,
17239	const char *zName,
17240	sqlite3_syscall_ptr pFunc
17241	){
17242	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17243	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17244	return pRoot->xSetSystemCall(pRoot, zName, pFunc);
17245	}
17246	static sqlite3_syscall_ptr vfstraceGetSystemCall(
17247	sqlite3_vfs *pVfs,
17248	const char *zName
17249	){
17250	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17251	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17252	return pRoot->xGetSystemCall(pRoot, zName);
17253	}
17254	static const char vfstraceNextSystemCall(sqlite3_vfs pVfs, const char *zName){
17255	vfstrace_info pInfo = (vfstrace_info)pVfs->pAppData;
17256	sqlite3_vfs *pRoot = pInfo->pRootVfs;
17257	return pRoot->xNextSystemCall(pRoot, zName);
17258	}
17259
17260
17261	/*
17262	** Clients invoke this routine to construct a new trace-vfs shim.
17263	**
17264	** Return SQLITE_OK on success.
17265	**
17266	** SQLITE_NOMEM is returned in the case of a memory allocation error.
17267	** SQLITE_NOTFOUND is returned if zOldVfsName does not exist.
17268	*/
17269	int vfstrace_register(
17270	const char zTraceName, / Name of the newly constructed VFS */
17271	const char zOldVfsName, / Name of the underlying VFS */
17272	int (xOut)(const char,void), / Output routine. ex: fputs */
17273	void pOutArg, / 2nd argument to xOut. ex: stderr */
17274	int makeDefault /* True to make the new VFS the default */
17275	){
17276	sqlite3_vfs *pNew;
17277	sqlite3_vfs *pRoot;
17278	vfstrace_info *pInfo;
17279	size_t nName;
17280	size_t nByte;
17281
17282	pRoot = sqlite3_vfs_find(zOldVfsName);
17283	if( pRoot==0 ) return SQLITE_NOTFOUND;
17284	nName = strlen(zTraceName);
17285	nByte = sizeof(pNew) + sizeof(pInfo) + nName + 1;
17286	pNew = sqlite3_malloc64( nByte );
17287	if( pNew==0 ) return SQLITE_NOMEM;
17288	memset(pNew, 0, nByte);
17289	pInfo = (vfstrace_info*)&pNew[1];
17290	pNew->iVersion = pRoot->iVersion;
17291	pNew->szOsFile = pRoot->szOsFile + sizeof(vfstrace_file);
17292	pNew->mxPathname = pRoot->mxPathname;
17293	pNew->zName = (char*)&pInfo[1];
17294	memcpy((char*)&pInfo[1], zTraceName, nName+1);
17295	pNew->pAppData = pInfo;
17296	pNew->xOpen = vfstraceOpen;
17297	pNew->xDelete = vfstraceDelete;
17298	pNew->xAccess = vfstraceAccess;
17299	pNew->xFullPathname = vfstraceFullPathname;
17300	pNew->xDlOpen = pRoot->xDlOpen==0 ? 0 : vfstraceDlOpen;
17301	pNew->xDlError = pRoot->xDlError==0 ? 0 : vfstraceDlError;
17302	pNew->xDlSym = pRoot->xDlSym==0 ? 0 : vfstraceDlSym;
17303	pNew->xDlClose = pRoot->xDlClose==0 ? 0 : vfstraceDlClose;
17304	pNew->xRandomness = vfstraceRandomness;
17305	pNew->xSleep = vfstraceSleep;
17306	pNew->xCurrentTime = vfstraceCurrentTime;
17307	pNew->xGetLastError = pRoot->xGetLastError==0 ? 0 : vfstraceGetLastError;
17308	if( pNew->iVersion>=2 ){
17309	pNew->xCurrentTimeInt64 = pRoot->xCurrentTimeInt64==0 ? 0 :
17310	vfstraceCurrentTimeInt64;
17311	if( pNew->iVersion>=3 ){
17312	pNew->xSetSystemCall = pRoot->xSetSystemCall==0 ? 0 :
17313	vfstraceSetSystemCall;
17314	pNew->xGetSystemCall = pRoot->xGetSystemCall==0 ? 0 :
17315	vfstraceGetSystemCall;
17316	pNew->xNextSystemCall = pRoot->xNextSystemCall==0 ? 0 :
17317	vfstraceNextSystemCall;
17318	}
17319	}
17320	pInfo->pRootVfs = pRoot;
17321	pInfo->xOut = xOut;
17322	pInfo->pOutArg = pOutArg;
17323	pInfo->zVfsName = pNew->zName;
17324	pInfo->pTraceVfs = pNew;
17325	pInfo->mTrace = 0xffffffff;
17326	pInfo->bOn = 1;
17327	vfstrace_printf(pInfo, "%s.enabled_for(\"%s\")\n",
17328	pInfo->zVfsName, pRoot->zName);
17329	return sqlite3_vfs_register(pNew, makeDefault);
17330	}
17331
17332	/*
17333	** Look for the named VFS. If it is a TRACEVFS, then unregister it
17334	** and delete it.
17335	*/
17336	void vfstrace_unregister(const char *zTraceName){
17337	sqlite3_vfs *pVfs = sqlite3_vfs_find(zTraceName);
17338	if( pVfs==0 ) return;
17339	if( pVfs->xOpen!=vfstraceOpen ) return;
17340	sqlite3_vfs_unregister(pVfs);
17341	sqlite3_free(pVfs);
17342	}
17343
17344	/*********************** End ../ext/misc/vfstrace.c ******************/
17345
17346	#if !defined(SQLITE_OMIT_VIRTUALTABLE) && defined(SQLITE_ENABLE_DBPAGE_VTAB)
17347	#define SQLITE_SHELL_HAVE_RECOVER 1
17348	#else
17349	#define SQLITE_SHELL_HAVE_RECOVER 0
	@@ -18241,11 +19370,11 @@
19370	}while( strstr(z,zBuf)!=0 );
19371	return zBuf;
19372	}
19373
19374	/*
19375	** Implementation of scalar SQL function "escape_crlf". The argument passed to
19376	** this function is the output of built-in function quote(). If the first
19377	** character of the input is "'", indicating that the value passed to quote()
19378	** was a text value, then this function searches the input for "\n" and "\r"
19379	** characters and adds a wrapper similar to the following:
19380	**
	@@ -18252,11 +19381,11 @@
19381	** replace(replace(<input>, '\n', char(10), '\r', char(13));
19382	**
19383	** Or, if the first character of the input is not "'", then a copy of the input
19384	** is returned.
19385	*/
19386	static void recoverEscapeCrlf(
19387	sqlite3_context *context,
19388	int argc,
19389	sqlite3_value **argv
19390	){
19391	const char zText = (const char)sqlite3_value_text(argv[0]);
	@@ -18467,11 +19596,11 @@
19596	void (xFunc)(sqlite3_context,int,sqlite3_value **);
19597	} aFunc[] = {
19598	{ "getpage", 1, recoverGetPage },
19599	{ "page_is_used", 1, recoverPageIsUsed },
19600	{ "read_i32", 2, recoverReadI32 },
19601	{ "escape_crlf", 1, recoverEscapeCrlf },
19602	};
19603
19604	const int flags = SQLITE_OPEN_URI\|SQLITE_OPEN_CREATE\|SQLITE_OPEN_READWRITE;
19605	sqlite3 db = 0; / New database handle */
19606	int ii; /* For iterating through aFunc[] */
	@@ -18820,11 +19949,11 @@
19949	assert( pTab->aCol[ii].iField>=0 && pTab->aCol[ii].iBind>=1 );
19950	zSql = recoverMPrintf(p, "%z%s%Q", zSql, zSep, pTab->aCol[ii].zCol);
19951
19952	if( bSql ){
19953	zBind = recoverMPrintf(p,
19954	"%z%sescape_crlf(quote(?%d))", zBind, zSqlSep, pTab->aCol[ii].iBind
19955	);
19956	zSqlSep = "\|\|', '\|\|";
19957	}else{
19958	zBind = recoverMPrintf(p, "%z%s?%d", zBind, zSep, pTab->aCol[ii].iBind);
19959	}
	@@ -19322,10 +20451,12 @@
20451	apVal[iField] = sqlite3_value_dup( pVal );
20452	if( apVal[iField]==0 ){
20453	recoverError(p, SQLITE_NOMEM, 0);
20454	}
20455	p1->nVal = iField+1;
20456	}else if( pTab->nCol==0 ){
20457	p1->nVal = pTab->nCol;
20458	}
20459	p1->iPrevCell = iCell;
20460	p1->iPrevPage = iPage;
20461	}
20462	}else{
	@@ -20437,10 +21568,11 @@
21568	u8 nEqpLevel; /* Depth of the EQP output graph */
21569	u8 eTraceType; /* SHELL_TRACE_* value for type of trace */
21570	u8 bSafeMode; /* True to prohibit unsafe operations */
21571	u8 bSafeModePersist; /* The long-term value of bSafeMode */
21572	u8 eRestoreState; /* See comments above doAutoDetectRestore() */
21573	u8 crlfMode; /* Do NL-to-CRLF translations when enabled (maybe) */
21574	ColModeOpts cmOpts; /* Option values affecting columnar mode output */
21575	unsigned statsOn; /* True to display memory stats before each finalize */
21576	unsigned mEqpLines; /* Mask of vertical lines in the EQP output graph */
21577	int inputNesting; /* Track nesting level of .read and other redirects */
21578	int outCount; /* Revert to stdout when reaching zero */
	@@ -20582,10 +21714,11 @@
21714	#define MODE_Table 15 /* MySQL-style table formatting */
21715	#define MODE_Box 16 /* Unicode box-drawing characters */
21716	#define MODE_Count 17 /* Output only a count of the rows of output */
21717	#define MODE_Off 18 /* No query output shown */
21718	#define MODE_ScanExp 19 /* Like MODE_Explain, but for ".scanstats vm" */
21719	#define MODE_Www 20 /* Full web-page output */
21720
21721	static const char *modeDescr[] = {
21722	"line",
21723	"column",
21724	"list",
	@@ -20602,11 +21735,13 @@
21735	"json",
21736	"markdown",
21737	"table",
21738	"box",
21739	"count",
21740	"off",
21741	"scanexp",
21742	"www",
21743	};
21744
21745	/*
21746	** These are the column/row/line separators used by the various
21747	** import/export modes.
	@@ -20630,11 +21765,11 @@
21765	** A callback for the sqlite3_log() interface.
21766	*/
21767	static void shellLog(void pArg, int iErrCode, const char zMsg){
21768	ShellState p = (ShellState)pArg;
21769	if( p->pLog==0 ) return;
21770	sqlite3_fprintf(p->pLog, "(%d) %s\n", iErrCode, zMsg);
21771	fflush(p->pLog);
21772	}
21773
21774	/*
21775	** SQL function: shell_putsnl(X)
	@@ -20645,13 +21780,13 @@
21780	static void shellPutsFunc(
21781	sqlite3_context *pCtx,
21782	int nVal,
21783	sqlite3_value **apVal
21784	){
21785	ShellState p = (ShellState)sqlite3_user_data(pCtx);
21786	(void)nVal;
21787	sqlite3_fprintf(p->out, "%s\n", sqlite3_value_text(apVal[0]));
21788	sqlite3_result_value(pCtx, apVal[0]);
21789	}
21790
21791	/*
21792	** If in safe mode, print an error message described by the arguments
	@@ -20666,11 +21801,11 @@
21801	va_list ap;
21802	char *zMsg;
21803	va_start(ap, zErrMsg);
21804	zMsg = sqlite3_vmprintf(zErrMsg, ap);
21805	va_end(ap);
21806	sqlite3_fprintf(stderr, "line %d: %s\n", p->lineno, zMsg);
21807	exit(1);
21808	}
21809	}
21810
21811	/*
	@@ -20699,11 +21834,11 @@
21834	char *zTempFile = 0;
21835	sqlite3 *db;
21836	char *zCmd = 0;
21837	int bBin;
21838	int rc;
21839	int hasCRLF = 0;
21840	FILE *f = 0;
21841	sqlite3_int64 sz;
21842	sqlite3_int64 x;
21843	unsigned char *p = 0;
21844
	@@ -20733,11 +21868,11 @@
21868	}
21869	}
21870	bBin = sqlite3_value_type(argv[0])==SQLITE_BLOB;
21871	/* When writing the file to be edited, do \n to \r\n conversions on systems
21872	** that want \r\n line endings */
21873	f = sqlite3_fopen(zTempFile, bBin ? "wb" : "w");
21874	if( f==0 ){
21875	sqlite3_result_error(context, "edit() cannot open temp file", -1);
21876	goto edit_func_end;
21877	}
21878	sz = sqlite3_value_bytes(argv[0]);
	@@ -20744,11 +21879,11 @@
21879	if( bBin ){
21880	x = fwrite(sqlite3_value_blob(argv[0]), 1, (size_t)sz, f);
21881	}else{
21882	const char z = (const char)sqlite3_value_text(argv[0]);
21883	/* Remember whether or not the value originally contained \r\n */
21884	if( z && strstr(z,"\r\n")!=0 ) hasCRLF = 1;
21885	x = fwrite(sqlite3_value_text(argv[0]), 1, (size_t)sz, f);
21886	}
21887	fclose(f);
21888	f = 0;
21889	if( x!=sz ){
	@@ -20764,11 +21899,11 @@
21899	sqlite3_free(zCmd);
21900	if( rc ){
21901	sqlite3_result_error(context, "EDITOR returned non-zero", -1);
21902	goto edit_func_end;
21903	}
21904	f = sqlite3_fopen(zTempFile, "rb");
21905	if( f==0 ){
21906	sqlite3_result_error(context,
21907	"edit() cannot reopen temp file after edit", -1);
21908	goto edit_func_end;
21909	}
	@@ -20789,11 +21924,11 @@
21924	}
21925	if( bBin ){
21926	sqlite3_result_blob64(context, p, sz, sqlite3_free);
21927	}else{
21928	sqlite3_int64 i, j;
21929	if( hasCRLF ){
21930	/* If the original contains \r\n then do no conversions back to \n */
21931	}else{
21932	/* If the file did not originally contain \r\n then convert any new
21933	** \r\n back into \n */
21934	p[sz] = 0;
	@@ -20830,15 +21965,30 @@
21965	p->mode = p->modePrior;
21966	p->shellFlgs = p->priorShFlgs;
21967	memcpy(p->colSeparator, p->colSepPrior, sizeof(p->colSeparator));
21968	memcpy(p->rowSeparator, p->rowSepPrior, sizeof(p->rowSeparator));
21969	}
21970
21971	/*
21972	** Set output mode to text or binary for Windows.
21973	*/
21974	static void setCrlfMode(ShellState *p){
21975	#ifdef _WIN32
21976	if( p->crlfMode ){
21977	sqlite3_fsetmode(p->out, _O_TEXT);
21978	}else{
21979	sqlite3_fsetmode(p->out, _O_BINARY);
21980	}
21981	#else
21982	UNUSED_PARAMETER(p);
21983	#endif
21984	}
21985
21986	/*
21987	** Output the given string as a hex-encoded blob (eg. X'1234' )
21988	*/
21989	static void output_hex_blob(FILE out, const void pBlob, int nBlob){
21990	int i;
21991	unsigned char aBlob = (unsigned char)pBlob;
21992
21993	char zStr = sqlite3_malloc(nBlob2 + 1);
21994	shell_check_oom(zStr);
	@@ -20851,11 +22001,11 @@
22001	zStr[i*2] = aHex[ (aBlob[i] >> 4) ];
22002	zStr[i*2+1] = aHex[ (aBlob[i] & 0x0F) ];
22003	}
22004	zStr[i*2] = '\0';
22005
22006	sqlite3_fprintf(out, "X'%s'", zStr);
22007	sqlite3_free(zStr);
22008	}
22009
22010	/*
22011	** Find a string that is not found anywhere in z[]. Return a pointer
	@@ -20881,46 +22031,40 @@
22031	/*
22032	** Output the given string as a quoted string using SQL quoting conventions.
22033	**
22034	** See also: output_quoted_escaped_string()
22035	*/
22036	static void output_quoted_string(ShellState p, const char z){
22037	int i;
22038	char c;
22039	FILE *out = p->out;
22040	sqlite3_fsetmode(out, _O_BINARY);


22041	if( z==0 ) return;
22042	for(i=0; (c = z[i])!=0 && c!='\''; i++){}
22043	if( c==0 ){
22044	sqlite3_fprintf(out, "'%s'",z);
22045	}else{
22046	sqlite3_fputs("'", out);
22047	while( *z ){
22048	for(i=0; (c = z[i])!=0 && c!='\''; i++){}
22049	if( c=='\'' ) i++;
22050	if( i ){
22051	sqlite3_fprintf(out, "%.*s", i, z);
22052	z += i;
22053	}
22054	if( c=='\'' ){
22055	sqlite3_fputs("'", out);
22056	continue;
22057	}
22058	if( c==0 ){
22059	break;
22060	}
22061	z++;
22062	}
22063	sqlite3_fputs("'", out);
22064	}
22065	setCrlfMode(p);




22066	}
22067
22068	/*
22069	** Output the given string as a quoted string using SQL quoting conventions.
22070	** Additionallly , escape the "\n" and "\r" characters so that they do not
	@@ -20928,20 +22072,18 @@
22072	** systems.
22073	**
22074	** This is like output_quoted_string() but with the addition of the \r\n
22075	** escape mechanism.
22076	*/
22077	static void output_quoted_escaped_string(ShellState p, const char z){
22078	int i;
22079	char c;
22080	FILE *out = p->out;
22081	sqlite3_fsetmode(out, _O_BINARY);


22082	for(i=0; (c = z[i])!=0 && c!='\'' && c!='\n' && c!='\r'; i++){}
22083	if( c==0 ){
22084	sqlite3_fprintf(out, "'%s'",z);
22085	}else{
22086	const char *zNL = 0;
22087	const char *zCR = 0;
22088	int nNL = 0;
22089	int nCR = 0;
	@@ -20949,52 +22091,48 @@
22091	for(i=0; z[i]; i++){
22092	if( z[i]=='\n' ) nNL++;
22093	if( z[i]=='\r' ) nCR++;
22094	}
22095	if( nNL ){
22096	sqlite3_fputs("replace(", out);
22097	zNL = unused_string(z, "\\n", "\\012", zBuf1);
22098	}
22099	if( nCR ){
22100	sqlite3_fputs("replace(", out);
22101	zCR = unused_string(z, "\\r", "\\015", zBuf2);
22102	}
22103	sqlite3_fputs("'", out);
22104	while( *z ){
22105	for(i=0; (c = z[i])!=0 && c!='\n' && c!='\r' && c!='\''; i++){}
22106	if( c=='\'' ) i++;
22107	if( i ){
22108	sqlite3_fprintf(out, "%.*s", i, z);
22109	z += i;
22110	}
22111	if( c=='\'' ){
22112	sqlite3_fputs("'", out);
22113	continue;
22114	}
22115	if( c==0 ){
22116	break;
22117	}
22118	z++;
22119	if( c=='\n' ){
22120	sqlite3_fputs(zNL, out);
22121	continue;
22122	}
22123	sqlite3_fputs(zCR, out);
22124	}
22125	sqlite3_fputs("'", out);
22126	if( nCR ){
22127	sqlite3_fprintf(out, ",'%s',char(13))", zCR);
22128	}
22129	if( nNL ){
22130	sqlite3_fprintf(out, ",'%s',char(10))", zNL);
22131	}
22132	}
22133	setCrlfMode(p);




22134	}
22135
22136	/*
22137	** Find earliest of chars within s specified in zAny.
22138	** With ns == ~0, is like strpbrk(s,zAny) and s must be 0-terminated.
	@@ -21010,26 +22148,64 @@
22148	}
22149	++zAny;
22150	}
22151	return pcFirst;
22152	}
22153
22154	/* Skip over as much z[] input char sequence as is valid UTF-8,
22155	** limited per nAccept char's or whole characters and containing
22156	** no char cn such that ((1<<cn) & ccm)!=0. On return, the
22157	** sequence z:return (inclusive:exclusive) is validated UTF-8.
22158	** Limit: nAccept>=0 => char count, nAccept<0 => character
22159	*/
22160	const char zSkipValidUtf8(const char z, int nAccept, long ccm){
22161	int ng = (nAccept<0)? -nAccept : 0;
22162	const char *pcLimit = (nAccept>=0)? z+nAccept : 0;
22163	assert(z!=0);
22164	while( (pcLimit)? (z<pcLimit) : (ng-- != 0) ){
22165	unsigned char c = (u8)z;
22166	if( c<0x7f ){
22167	if( ccm != 0L && c < 0x20 && ((1L<<c) & ccm) != 0 ) return z;
22168	++z; /* ASCII */
22169	}else if( (c & 0xC0) != 0xC0 ) return z; /* not a lead byte */
22170	else{
22171	const char zt = z+1; / Got lead byte, look at trail bytes.*/
22172	do{
22173	if( pcLimit && zt >= pcLimit ) return z;
22174	else{
22175	char ct = *zt++;
22176	if( ct==0 \|\| (zt-z)>4 \|\| (ct & 0xC0)!=0x80 ){
22177	/* Trailing bytes are too few, too many, or invalid. */
22178	return z;
22179	}
22180	}
22181	} while( ((c <<= 1) & 0x40) == 0x40 ); /* Eat lead byte's count. */
22182	z = zt;
22183	}
22184	}
22185	return z;
22186	}
22187
22188
22189	/*
22190	** Output the given string as a quoted according to C or TCL quoting rules.
22191	*/
22192	static void output_c_string(FILE out, const char z){
22193	char c;
22194	static const char *zq = "\"";
22195	static long ctrlMask = ~0L;
22196	static const char zDQBSRO = "\"\\\x7f"; / double-quote, backslash, rubout */
22197	char ace[3] = "\\?";
22198	char cbsSay;
22199	sqlite3_fputs(zq, out);
22200	while( *z!=0 ){
22201	const char *pcDQBSRO = anyOfInStr(z, zDQBSRO, ~(size_t)0);
22202	const char *pcPast = zSkipValidUtf8(z, INT_MAX, ctrlMask);
22203	const char *pcEnd = (pcDQBSRO && pcDQBSRO < pcPast)? pcDQBSRO : pcPast;
22204	if( pcEnd > z ){
22205	sqlite3_fprintf(out, "%.*s", (int)(pcEnd-z), z);
22206	}
22207	if( (c = *pcEnd)==0 ) break;
22208	++pcEnd;
22209	switch( c ){
22210	case '\\': case '"':
22211	cbsSay = (char)c;
	@@ -21040,47 +22216,47 @@
22216	case '\f': cbsSay = 'f'; break;
22217	default: cbsSay = 0; break;
22218	}
22219	if( cbsSay ){
22220	ace[1] = cbsSay;
22221	sqlite3_fputs(ace, out);
22222	}else if( !isprint(c&0xff) ){
22223	sqlite3_fprintf(out, "\\%03o", c&0xff);
22224	}else{
22225	ace[1] = (char)c;
22226	sqlite3_fputs(ace+1, out);
22227	}
22228	z = pcEnd;
22229	}
22230	sqlite3_fputs(zq, out);
22231	}
22232
22233	/*
22234	** Output the given string as quoted according to JSON quoting rules.
22235	*/
22236	static void output_json_string(FILE out, const char z, i64 n){
22237	unsigned char c;
22238	static const char *zq = "\"";
22239	static long ctrlMask = ~0L;
22240	static const char *zDQBS = "\"\\";
22241	const char *pcLimit;
22242	char ace[3] = "\\?";
22243	char cbsSay;
22244
22245	if( z==0 ) z = "";
22246	pcLimit = z + ((n<0)? strlen(z) : (size_t)n);
22247	sqlite3_fputs(zq, out);
22248	while( z < pcLimit ){
22249	const char *pcDQBS = anyOfInStr(z, zDQBS, pcLimit-z);
22250	const char *pcPast = zSkipValidUtf8(z, (int)(pcLimit-z), ctrlMask);
22251	const char *pcEnd = (pcDQBS && pcDQBS < pcPast)? pcDQBS : pcPast;
22252	if( pcEnd > z ){
22253	sqlite3_fprintf(out, "%.*s", (int)(pcEnd-z), z);
22254	z = pcEnd;
22255	}
22256	if( z >= pcLimit ) break;
22257	c = (unsigned char)*(z++);
22258	switch( c ){
22259	case '"': case '\\':
22260	cbsSay = (char)c;
22261	break;
22262	case '\b': cbsSay = 'b'; break;
	@@ -21090,26 +22266,26 @@
22266	case '\t': cbsSay = 't'; break;
22267	default: cbsSay = 0; break;
22268	}
22269	if( cbsSay ){
22270	ace[1] = cbsSay;
22271	sqlite3_fputs(ace, out);
22272	}else if( c<=0x1f \|\| c>=0x7f ){
22273	sqlite3_fprintf(out, "\\u%04x", c);
22274	}else{
22275	ace[1] = (char)c;
22276	sqlite3_fputs(ace+1, out);
22277	}
22278	}
22279	sqlite3_fputs(zq, out);
22280	}
22281
22282	/*
22283	** Output the given string with characters that are special to
22284	** HTML escaped.
22285	*/
22286	static void output_html_string(FILE out, const char z){
22287	int i;
22288	if( z==0 ) z = "";
22289	while( *z ){
22290	for(i=0; z[i]
22291	&& z[i]!='<'
	@@ -21117,22 +22293,22 @@
22293	&& z[i]!='>'
22294	&& z[i]!='\"'
22295	&& z[i]!='\'';
22296	i++){}
22297	if( i>0 ){
22298	sqlite3_fprintf(out, "%.*s",i,z);
22299	}
22300	if( z[i]=='<' ){
22301	sqlite3_fputs("<", out);
22302	}else if( z[i]=='&' ){
22303	sqlite3_fputs("&", out);
22304	}else if( z[i]=='>' ){
22305	sqlite3_fputs(">", out);
22306	}else if( z[i]=='\"' ){
22307	sqlite3_fputs(""", out);
22308	}else if( z[i]=='\'' ){
22309	sqlite3_fputs("'", out);
22310	}else{
22311	break;
22312	}
22313	z += i + 1;
22314	}
	@@ -21167,11 +22343,11 @@
22343	** the null value. Strings are quoted if necessary. The separator
22344	** is only issued if bSep is true.
22345	*/
22346	static void output_csv(ShellState p, const char z, int bSep){
22347	if( z==0 ){
22348	sqlite3_fprintf(p->out, "%s",p->nullValue);
22349	}else{
22350	unsigned i;
22351	for(i=0; z[i]; i++){
22352	if( needCsvQuote[((unsigned char*)z)[i]] ){
22353	i = 0;
	@@ -21179,18 +22355,18 @@
22355	}
22356	}
22357	if( i==0 \|\| strstr(z, p->colSeparator)!=0 ){
22358	char *zQuoted = sqlite3_mprintf("\"%w\"", z);
22359	shell_check_oom(zQuoted);
22360	sqlite3_fputs(zQuoted, p->out);
22361	sqlite3_free(zQuoted);
22362	}else{
22363	sqlite3_fputs(z, p->out);
22364	}
22365	}
22366	if( bSep ){
22367	sqlite3_fputs(p->colSeparator, p->out);
22368	}
22369	}
22370
22371	/*
22372	** This routine runs when the user presses Ctrl-C
	@@ -21294,20 +22470,20 @@
22470	const char *az[4];
22471	az[0] = zA1;
22472	az[1] = zA2;
22473	az[2] = zA3;
22474	az[3] = zA4;
22475	sqlite3_fprintf(p->out, "authorizer: %s", azAction[op]);
22476	for(i=0; i<4; i++){
22477	sqlite3_fputs(" ", p->out);
22478	if( az[i] ){
22479	output_c_string(p->out, az[i]);
22480	}else{
22481	sqlite3_fputs("NULL", p->out);
22482	}
22483	}
22484	sqlite3_fputs("\n", p->out);
22485	if( p->bSafeMode ) (void)safeModeAuth(pClientData, op, zA1, zA2, zA3, zA4);
22486	return SQLITE_OK;
22487	}
22488	#endif
22489
	@@ -21319,11 +22495,11 @@
22495	**
22496	** If the schema statement in z[] contains a start-of-comment and if
22497	** sqlite3_complete() returns false, try to terminate the comment before
22498	** printing the result. https://sqlite.org/forum/forumpost/d7be961c5c
22499	*/
22500	static void printSchemaLine(FILE out, const char z, const char *zTail){
22501	char *zToFree = 0;
22502	if( z==0 ) return;
22503	if( zTail==0 ) return;
22504	if( zTail[0]==';' && (strstr(z, "/*")!=0 \|\| strstr(z,"--")!=0) ){
22505	const char *zOrig = z;
	@@ -21341,20 +22517,20 @@
22517	}
22518	sqlite3_free(zNew);
22519	}
22520	}
22521	if( sqlite3_strglob("CREATE TABLE ['\"]*", z)==0 ){
22522	sqlite3_fprintf(out, "CREATE TABLE IF NOT EXISTS %s%s", z+13, zTail);
22523	}else{
22524	sqlite3_fprintf(out, "%s%s", z, zTail);
22525	}
22526	sqlite3_free(zToFree);
22527	}
22528	static void printSchemaLineN(FILE out, char z, int n, const char *zTail){
22529	char c = z[n];
22530	z[n] = 0;
22531	printSchemaLine(out, z, zTail);
22532	z[n] = c;
22533	}
22534
22535	/*
22536	** Return true if string z[] has nothing but whitespace and comments to the
	@@ -21378,11 +22554,11 @@
22554	EQPGraphRow *pNew;
22555	i64 nText;
22556	if( zText==0 ) return;
22557	nText = strlen(zText);
22558	if( p->autoEQPtest ){
22559	sqlite3_fprintf(p->out, "%d,%d,%s\n", iEqpId, p2, zText);
22560	}
22561	pNew = sqlite3_malloc64( sizeof(*pNew) + nText );
22562	shell_check_oom(pNew);
22563	pNew->iEqpId = iEqpId;
22564	pNew->iParentId = p2;
	@@ -21426,11 +22602,12 @@
22602	i64 n = strlen(p->sGraph.zPrefix);
22603	char *z;
22604	for(pRow = eqp_next_row(p, iEqpId, 0); pRow; pRow = pNext){
22605	pNext = eqp_next_row(p, iEqpId, pRow);
22606	z = pRow->zText;
22607	sqlite3_fprintf(p->out, "%s%s%s\n", p->sGraph.zPrefix,
22608	pNext ? "\|--" : "`--", z);
22609	if( n<(i64)sizeof(p->sGraph.zPrefix)-7 ){
22610	memcpy(&p->sGraph.zPrefix[n], pNext ? "\| " : " ", 4);
22611	eqp_render_level(p, pRow->iEqpId);
22612	p->sGraph.zPrefix[n] = 0;
22613	}
	@@ -21446,17 +22623,17 @@
22623	if( pRow->zText[0]=='-' ){
22624	if( pRow->pNext==0 ){
22625	eqp_reset(p);
22626	return;
22627	}
22628	sqlite3_fprintf(p->out, "%s\n", pRow->zText+3);
22629	p->sGraph.pRow = pRow->pNext;
22630	sqlite3_free(pRow);
22631	}else if( nCycle>0 ){
22632	sqlite3_fprintf(p->out, "QUERY PLAN (cycles=%lld [100%%])\n", nCycle);
22633	}else{
22634	sqlite3_fputs("QUERY PLAN\n", p->out);
22635	}
22636	p->sGraph.zPrefix[0] = 0;
22637	eqp_render_level(p, 0);
22638	eqp_reset(p);
22639	}
	@@ -21468,33 +22645,33 @@
22645	*/
22646	static int progress_handler(void *pClientData) {
22647	ShellState p = (ShellState)pClientData;
22648	p->nProgress++;
22649	if( p->nProgress>=p->mxProgress && p->mxProgress>0 ){
22650	sqlite3_fprintf(p->out, "Progress limit reached (%u)\n", p->nProgress);
22651	if( p->flgProgress & SHELL_PROGRESS_RESET ) p->nProgress = 0;
22652	if( p->flgProgress & SHELL_PROGRESS_ONCE ) p->mxProgress = 0;
22653	return 1;
22654	}
22655	if( (p->flgProgress & SHELL_PROGRESS_QUIET)==0 ){
22656	sqlite3_fprintf(p->out, "Progress %u\n", p->nProgress);
22657	}
22658	return 0;
22659	}
22660	#endif /* SQLITE_OMIT_PROGRESS_CALLBACK */
22661
22662	/*
22663	** Print N dashes
22664	*/
22665	static void print_dashes(FILE *out, int N){
22666	const char zDash[] = "--------------------------------------------------";
22667	const int nDash = sizeof(zDash) - 1;
22668	while( N>nDash ){
22669	sqlite3_fputs(zDash, out);
22670	N -= nDash;
22671	}
22672	sqlite3_fprintf(out, "%.*s", N, zDash);
22673	}
22674
22675	/*
22676	** Print a markdown or table-style row separator using ascii-art
22677	*/
	@@ -21503,19 +22680,19 @@
22680	int nArg,
22681	const char *zSep
22682	){
22683	int i;
22684	if( nArg>0 ){
22685	sqlite3_fputs(zSep, p->out);
22686	print_dashes(p->out, p->actualWidth[0]+2);
22687	for(i=1; i<nArg; i++){
22688	sqlite3_fputs(zSep, p->out);
22689	print_dashes(p->out, p->actualWidth[i]+2);
22690	}
22691	sqlite3_fputs(zSep, p->out);
22692	}
22693	sqlite3_fputs("\n", p->out);
22694	}
22695
22696	/*
22697	** This is the callback routine that the shell
22698	** invokes for each row of a query result.
	@@ -21541,13 +22718,13 @@
22718	if( azArg==0 ) break;
22719	for(i=0; i<nArg; i++){
22720	int len = strlen30(azCol[i] ? azCol[i] : "");
22721	if( len>w ) w = len;
22722	}
22723	if( p->cnt++>0 ) sqlite3_fputs(p->rowSeparator, p->out);
22724	for(i=0; i<nArg; i++){
22725	sqlite3_fprintf(p->out, "%*s = %s%s", w, azCol[i],
22726	azArg[i] ? azArg[i] : p->nullValue, p->rowSeparator);
22727	}
22728	break;
22729	}
22730	case MODE_ScanExp:
	@@ -21571,16 +22748,16 @@
22748	if( nArg>nWidth ) nArg = nWidth;
22749
22750	/* If this is the first row seen, print out the headers */
22751	if( p->cnt++==0 ){
22752	for(i=0; i<nArg; i++){
22753	utf8_width_print(p->out, aWidth[i], azCol[ aMap[i] ]);
22754	sqlite3_fputs(i==nArg-1 ? "\n" : " ", p->out);
22755	}
22756	for(i=0; i<nArg; i++){
22757	print_dashes(p->out, aWidth[i]);
22758	sqlite3_fputs(i==nArg-1 ? "\n" : " ", p->out);
22759	}
22760	}
22761
22762	/* If there is no data, exit early. */
22763	if( azArg==0 ) break;
	@@ -21594,21 +22771,21 @@
22771	w = strlenChar(zVal);
22772	zSep = " ";
22773	}
22774	if( i==iIndent && p->aiIndent && p->pStmt ){
22775	if( p->iIndent<p->nIndent ){
22776	sqlite3_fprintf(p->out, "%*.s", p->aiIndent[p->iIndent], "");
22777	}
22778	p->iIndent++;
22779	}
22780	utf8_width_print(p->out, w, zVal ? zVal : p->nullValue);
22781	sqlite3_fputs(i==nArg-1 ? "\n" : zSep, p->out);
22782	}
22783	break;
22784	}
22785	case MODE_Semi: { /* .schema and .fullschema output */
22786	printSchemaLine(p->out, azArg[0], ";\n");
22787	break;
22788	}
22789	case MODE_Pretty: { /* .schema and .fullschema with --indent */
22790	char *z;
22791	int j;
	@@ -21619,11 +22796,11 @@
22796	assert( nArg==1 );
22797	if( azArg[0]==0 ) break;
22798	if( sqlite3_strlike("CREATE VIEW%", azArg[0], 0)==0
22799	\|\| sqlite3_strlike("CREATE TRIG%", azArg[0], 0)==0
22800	){
22801	sqlite3_fprintf(p->out, "%s;\n", azArg[0]);
22802	break;
22803	}
22804	z = sqlite3_mprintf("%s", azArg[0]);
22805	shell_check_oom(z);
22806	j = 0;
	@@ -21652,255 +22829,265 @@
22829	}else if( c=='(' ){
22830	nParen++;
22831	}else if( c==')' ){
22832	nParen--;
22833	if( nLine>0 && nParen==0 && j>0 ){
22834	printSchemaLineN(p->out, z, j, "\n");
22835	j = 0;
22836	}
22837	}
22838	z[j++] = c;
22839	if( nParen==1 && cEnd==0
22840	&& (c=='(' \|\| c=='\n' \|\| (c==',' && !wsToEol(z+i+1)))
22841	){
22842	if( c=='\n' ) j--;
22843	printSchemaLineN(p->out, z, j, "\n ");
22844	j = 0;
22845	nLine++;
22846	while( IsSpace(z[i+1]) ){ i++; }
22847	}
22848	}
22849	z[j] = 0;
22850	}
22851	printSchemaLine(p->out, z, ";\n");
22852	sqlite3_free(z);
22853	break;
22854	}
22855	case MODE_List: {
22856	if( p->cnt++==0 && p->showHeader ){
22857	for(i=0; i<nArg; i++){
22858	sqlite3_fprintf(p->out, "%s%s", azCol[i],
22859	i==nArg-1 ? p->rowSeparator : p->colSeparator);
22860	}
22861	}
22862	if( azArg==0 ) break;
22863	for(i=0; i<nArg; i++){
22864	char *z = azArg[i];
22865	if( z==0 ) z = p->nullValue;
22866	sqlite3_fputs(z, p->out);
22867	sqlite3_fputs((i<nArg-1)? p->colSeparator : p->rowSeparator, p->out);
22868	}
22869	break;
22870	}
22871	case MODE_Www:
22872	case MODE_Html: {
22873	if( p->cnt==0 && p->cMode==MODE_Www ){
22874	sqlite3_fputs(
22875	"</PRE>\n"
22876	"<TABLE border='1' cellspacing='0' cellpadding='2'>\n"
22877	,p->out
22878	);
22879	}
22880	if( p->cnt==0 && (p->showHeader \|\| p->cMode==MODE_Www) ){
22881	sqlite3_fputs("<TR>", p->out);
22882	for(i=0; i<nArg; i++){
22883	sqlite3_fputs("<TH>", p->out);
22884	output_html_string(p->out, azCol[i]);
22885	sqlite3_fputs("</TH>\n", p->out);
22886	}
22887	sqlite3_fputs("</TR>\n", p->out);
22888	}
22889	p->cnt++;
22890	if( azArg==0 ) break;
22891	sqlite3_fputs("<TR>", p->out);
22892	for(i=0; i<nArg; i++){
22893	sqlite3_fputs("<TD>", p->out);
22894	output_html_string(p->out, azArg[i] ? azArg[i] : p->nullValue);
22895	sqlite3_fputs("</TD>\n", p->out);
22896	}
22897	sqlite3_fputs("</TR>\n", p->out);
22898	break;
22899	}
22900	case MODE_Tcl: {
22901	if( p->cnt++==0 && p->showHeader ){
22902	for(i=0; i<nArg; i++){
22903	output_c_string(p->out, azCol[i] ? azCol[i] : "");
22904	if(i<nArg-1) sqlite3_fputs(p->colSeparator, p->out);
22905	}
22906	sqlite3_fputs(p->rowSeparator, p->out);
22907	}
22908	if( azArg==0 ) break;
22909	for(i=0; i<nArg; i++){
22910	output_c_string(p->out, azArg[i] ? azArg[i] : p->nullValue);
22911	if(i<nArg-1) sqlite3_fputs(p->colSeparator, p->out);
22912	}
22913	sqlite3_fputs(p->rowSeparator, p->out);
22914	break;
22915	}
22916	case MODE_Csv: {
22917	sqlite3_fsetmode(p->out, _O_BINARY);
22918	if( p->cnt++==0 && p->showHeader ){
22919	for(i=0; i<nArg; i++){
22920	output_csv(p, azCol[i] ? azCol[i] : "", i<nArg-1);
22921	}
22922	sqlite3_fputs(p->rowSeparator, p->out);
22923	}
22924	if( nArg>0 ){
22925	for(i=0; i<nArg; i++){
22926	output_csv(p, azArg[i], i<nArg-1);
22927	}
22928	sqlite3_fputs(p->rowSeparator, p->out);
22929	}
22930	setCrlfMode(p);
22931	break;
22932	}
22933	case MODE_Insert: {
22934	if( azArg==0 ) break;
22935	sqlite3_fprintf(p->out, "INSERT INTO %s",p->zDestTable);
22936	if( p->showHeader ){
22937	sqlite3_fputs("(", p->out);
22938	for(i=0; i<nArg; i++){
22939	if( i>0 ) sqlite3_fputs(",", p->out);
22940	if( quoteChar(azCol[i]) ){
22941	char *z = sqlite3_mprintf("\"%w\"", azCol[i]);
22942	shell_check_oom(z);
22943	sqlite3_fputs(z, p->out);
22944	sqlite3_free(z);
22945	}else{
22946	sqlite3_fprintf(p->out, "%s", azCol[i]);
22947	}
22948	}
22949	sqlite3_fputs(")", p->out);
22950	}
22951	p->cnt++;
22952	for(i=0; i<nArg; i++){
22953	sqlite3_fputs(i>0 ? "," : " VALUES(", p->out);
22954	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
22955	sqlite3_fputs("NULL", p->out);
22956	}else if( aiType && aiType[i]==SQLITE_TEXT ){
22957	if( ShellHasFlag(p, SHFLG_Newlines) ){
22958	output_quoted_string(p, azArg[i]);
22959	}else{
22960	output_quoted_escaped_string(p, azArg[i]);
22961	}
22962	}else if( aiType && aiType[i]==SQLITE_INTEGER ){
22963	sqlite3_fputs(azArg[i], p->out);
22964	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
22965	char z[50];
22966	double r = sqlite3_column_double(p->pStmt, i);
22967	sqlite3_uint64 ur;
22968	memcpy(&ur,&r,sizeof(r));
22969	if( ur==0x7ff0000000000000LL ){
22970	sqlite3_fputs("9.0e+999", p->out);
22971	}else if( ur==0xfff0000000000000LL ){
22972	sqlite3_fputs("-9.0e+999", p->out);
22973	}else{
22974	sqlite3_int64 ir = (sqlite3_int64)r;
22975	if( r==(double)ir ){
22976	sqlite3_snprintf(50,z,"%lld.0", ir);
22977	}else{
22978	sqlite3_snprintf(50,z,"%!.20g", r);
22979	}
22980	sqlite3_fputs(z, p->out);
22981	}
22982	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
22983	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
22984	int nBlob = sqlite3_column_bytes(p->pStmt, i);
22985	output_hex_blob(p->out, pBlob, nBlob);
22986	}else if( isNumber(azArg[i], 0) ){
22987	sqlite3_fputs(azArg[i], p->out);
22988	}else if( ShellHasFlag(p, SHFLG_Newlines) ){
22989	output_quoted_string(p, azArg[i]);
22990	}else{
22991	output_quoted_escaped_string(p, azArg[i]);
22992	}
22993	}
22994	sqlite3_fputs(");\n", p->out);
22995	break;
22996	}
22997	case MODE_Json: {
22998	if( azArg==0 ) break;
22999	if( p->cnt==0 ){
23000	sqlite3_fputs("[{", p->out);
23001	}else{
23002	sqlite3_fputs(",\n{", p->out);
23003	}
23004	p->cnt++;
23005	for(i=0; i<nArg; i++){
23006	output_json_string(p->out, azCol[i], -1);
23007	sqlite3_fputs(":", p->out);
23008	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
23009	sqlite3_fputs("null", p->out);
23010	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
23011	char z[50];
23012	double r = sqlite3_column_double(p->pStmt, i);
23013	sqlite3_uint64 ur;
23014	memcpy(&ur,&r,sizeof(r));
23015	if( ur==0x7ff0000000000000LL ){
23016	sqlite3_fputs("9.0e+999", p->out);
23017	}else if( ur==0xfff0000000000000LL ){
23018	sqlite3_fputs("-9.0e+999", p->out);
23019	}else{
23020	sqlite3_snprintf(50,z,"%!.20g", r);
23021	sqlite3_fputs(z, p->out);
23022	}
23023	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
23024	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
23025	int nBlob = sqlite3_column_bytes(p->pStmt, i);
23026	output_json_string(p->out, pBlob, nBlob);
23027	}else if( aiType && aiType[i]==SQLITE_TEXT ){
23028	output_json_string(p->out, azArg[i], -1);
23029	}else{
23030	sqlite3_fputs(azArg[i], p->out);
23031	}
23032	if( i<nArg-1 ){
23033	sqlite3_fputs(",", p->out);
23034	}
23035	}
23036	sqlite3_fputs("}", p->out);
23037	break;
23038	}
23039	case MODE_Quote: {
23040	if( azArg==0 ) break;
23041	if( p->cnt==0 && p->showHeader ){
23042	for(i=0; i<nArg; i++){
23043	if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
23044	output_quoted_string(p, azCol[i]);
23045	}
23046	sqlite3_fputs(p->rowSeparator, p->out);
23047	}
23048	p->cnt++;
23049	for(i=0; i<nArg; i++){
23050	if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
23051	if( (azArg[i]==0) \|\| (aiType && aiType[i]==SQLITE_NULL) ){
23052	sqlite3_fputs("NULL", p->out);
23053	}else if( aiType && aiType[i]==SQLITE_TEXT ){
23054	output_quoted_string(p, azArg[i]);
23055	}else if( aiType && aiType[i]==SQLITE_INTEGER ){
23056	sqlite3_fputs(azArg[i], p->out);
23057	}else if( aiType && aiType[i]==SQLITE_FLOAT ){
23058	char z[50];
23059	double r = sqlite3_column_double(p->pStmt, i);
23060	sqlite3_snprintf(50,z,"%!.20g", r);
23061	sqlite3_fputs(z, p->out);
23062	}else if( aiType && aiType[i]==SQLITE_BLOB && p->pStmt ){
23063	const void *pBlob = sqlite3_column_blob(p->pStmt, i);
23064	int nBlob = sqlite3_column_bytes(p->pStmt, i);
23065	output_hex_blob(p->out, pBlob, nBlob);
23066	}else if( isNumber(azArg[i], 0) ){
23067	sqlite3_fputs(azArg[i], p->out);
23068	}else{
23069	output_quoted_string(p, azArg[i]);
23070	}
23071	}
23072	sqlite3_fputs(p->rowSeparator, p->out);
23073	break;
23074	}
23075	case MODE_Ascii: {
23076	if( p->cnt++==0 && p->showHeader ){
23077	for(i=0; i<nArg; i++){
23078	if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
23079	sqlite3_fputs(azCol[i] ? azCol[i] : "", p->out);
23080	}
23081	sqlite3_fputs(p->rowSeparator, p->out);
23082	}
23083	if( azArg==0 ) break;
23084	for(i=0; i<nArg; i++){
23085	if( i>0 ) sqlite3_fputs(p->colSeparator, p->out);
23086	sqlite3_fputs(azArg[i] ? azArg[i] : p->nullValue, p->out);
23087	}
23088	sqlite3_fputs(p->rowSeparator, p->out);
23089	break;
23090	}
23091	case MODE_EQP: {
23092	eqp_append(p, atoi(azArg[0]), atoi(azArg[1]), azArg[3]);
23093	break;
	@@ -21975,11 +23162,11 @@
23162	"INSERT INTO selftest(tno,op,cmd,ans)"
23163	" SELECT rowid*10,op,cmd,ans FROM [_shell$self];\n"
23164	"DROP TABLE [_shell$self];"
23165	,0,0,&zErrMsg);
23166	if( zErrMsg ){
23167	sqlite3_fprintf(stderr, "SELFTEST initialization failure: %s\n", zErrMsg);
23168	sqlite3_free(zErrMsg);
23169	}
23170	sqlite3_exec(p->db, "RELEASE selftest_init",0,0,0);
23171	}
23172
	@@ -22078,36 +23265,37 @@
23265	int i;
23266	const char *z;
23267	rc = sqlite3_prepare_v2(p->db, zSelect, -1, &pSelect, 0);
23268	if( rc!=SQLITE_OK \|\| !pSelect ){
23269	char *zContext = shell_error_context(zSelect, p->db);
23270	sqlite3_fprintf(p->out, "/** ERROR: (%d) %s ***/\n%s",
23271	rc, sqlite3_errmsg(p->db), zContext);
23272	sqlite3_free(zContext);
23273	if( (rc&0xff)!=SQLITE_CORRUPT ) p->nErr++;
23274	return rc;
23275	}
23276	rc = sqlite3_step(pSelect);
23277	nResult = sqlite3_column_count(pSelect);
23278	while( rc==SQLITE_ROW ){
23279	z = (const char*)sqlite3_column_text(pSelect, 0);
23280	sqlite3_fprintf(p->out, "%s", z);
23281	for(i=1; i<nResult; i++){
23282	sqlite3_fprintf(p->out, ",%s", sqlite3_column_text(pSelect, i));
23283	}
23284	if( z==0 ) z = "";
23285	while( z[0] && (z[0]!='-' \|\| z[1]!='-') ) z++;
23286	if( z[0] ){
23287	sqlite3_fputs("\n;\n", p->out);
23288	}else{
23289	sqlite3_fputs(";\n", p->out);
23290	}
23291	rc = sqlite3_step(pSelect);
23292	}
23293	rc = sqlite3_finalize(pSelect);
23294	if( rc!=SQLITE_OK ){
23295	sqlite3_fprintf(p->out, "/** ERROR: (%d) %s ***/\n",
23296	rc, sqlite3_errmsg(p->db));
23297	if( (rc&0xff)!=SQLITE_CORRUPT ) p->nErr++;
23298	}
23299	return rc;
23300	}
23301
	@@ -22139,17 +23327,17 @@
23327
23328	#ifdef __linux__
23329	/*
23330	** Attempt to display I/O stats on Linux using /proc/PID/io
23331	*/
23332	static void displayLinuxIoStats(FILE *out){
23333	FILE *in;
23334	char z[200];
23335	sqlite3_snprintf(sizeof(z), z, "/proc/%d/io", getpid());
23336	in = sqlite3_fopen(z, "rb");
23337	if( in==0 ) return;
23338	while( sqlite3_fgets(z, sizeof(z), in)!=0 ){
23339	static const struct {
23340	const char *zPattern;
23341	const char *zDesc;
23342	} aTrans[] = {
23343	{ "rchar: ", "Bytes received by read():" },
	@@ -22162,11 +23350,11 @@
23350	};
23351	int i;
23352	for(i=0; i<ArraySize(aTrans); i++){
23353	int n = strlen30(aTrans[i].zPattern);
23354	if( cli_strncmp(aTrans[i].zPattern, z, n)==0 ){
23355	sqlite3_fprintf(out, "%-36s %s", aTrans[i].zDesc, &z[n]);
23356	break;
23357	}
23358	}
23359	}
23360	fclose(in);
	@@ -22175,10 +23363,11 @@
23363
23364	/*
23365	** Display a single line of status using 64-bit values.
23366	*/
23367	static void displayStatLine(
23368	FILE out, / Write to this channel */
23369	char zLabel, / Label for this one line */
23370	char zFormat, / Format for the result */
23371	int iStatusCtrl, /* Which status to display */
23372	int bReset /* True to reset the stats */
23373	){
	@@ -22193,11 +23382,11 @@
23382	if( nPercent>1 ){
23383	sqlite3_snprintf(sizeof(zLine), zLine, zFormat, iCur, iHiwtr);
23384	}else{
23385	sqlite3_snprintf(sizeof(zLine), zLine, zFormat, iHiwtr);
23386	}
23387	sqlite3_fprintf(out, "%-36s %s\n", zLabel, zLine);
23388	}
23389
23390	/*
23391	** Display memory stats.
23392	*/
	@@ -22206,130 +23395,152 @@
23395	ShellState pArg, / Pointer to ShellState */
23396	int bReset /* True to reset the stats */
23397	){
23398	int iCur;
23399	int iHiwtr;
23400	FILE *out;
23401	if( pArg==0 \|\| pArg->out==0 ) return 0;
23402	out = pArg->out;
23403
23404	if( pArg->pStmt && pArg->statsOn==2 ){
23405	int nCol, i, x;
23406	sqlite3_stmt *pStmt = pArg->pStmt;
23407	char z[100];
23408	nCol = sqlite3_column_count(pStmt);
23409	sqlite3_fprintf(out, "%-36s %d\n", "Number of output columns:", nCol);
23410	for(i=0; i<nCol; i++){
23411	sqlite3_snprintf(sizeof(z),z,"Column %d %nname:", i, &x);
23412	sqlite3_fprintf(out, "%-36s %s\n", z, sqlite3_column_name(pStmt,i));
23413	#ifndef SQLITE_OMIT_DECLTYPE
23414	sqlite3_snprintf(30, z+x, "declared type:");
23415	sqlite3_fprintf(out, "%-36s %s\n", z, sqlite3_column_decltype(pStmt, i));
23416	#endif
23417	#ifdef SQLITE_ENABLE_COLUMN_METADATA
23418	sqlite3_snprintf(30, z+x, "database name:");
23419	sqlite3_fprintf(out, "%-36s %s\n", z,
23420	sqlite3_column_database_name(pStmt,i));
23421	sqlite3_snprintf(30, z+x, "table name:");
23422	sqlite3_fprintf(out, "%-36s %s\n", z, sqlite3_column_table_name(pStmt,i));
23423	sqlite3_snprintf(30, z+x, "origin name:");
23424	sqlite3_fprintf(out, "%-36s %s\n", z,sqlite3_column_origin_name(pStmt,i));
23425	#endif
23426	}
23427	}
23428
23429	if( pArg->statsOn==3 ){
23430	if( pArg->pStmt ){
23431	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_VM_STEP,bReset);
23432	sqlite3_fprintf(out, "VM-steps: %d\n", iCur);
23433	}
23434	return 0;
23435	}
23436
23437	displayStatLine(out, "Memory Used:",
23438	"%lld (max %lld) bytes", SQLITE_STATUS_MEMORY_USED, bReset);
23439	displayStatLine(out, "Number of Outstanding Allocations:",
23440	"%lld (max %lld)", SQLITE_STATUS_MALLOC_COUNT, bReset);
23441	if( pArg->shellFlgs & SHFLG_Pagecache ){
23442	displayStatLine(out, "Number of Pcache Pages Used:",
23443	"%lld (max %lld) pages", SQLITE_STATUS_PAGECACHE_USED, bReset);
23444	}
23445	displayStatLine(out, "Number of Pcache Overflow Bytes:",
23446	"%lld (max %lld) bytes", SQLITE_STATUS_PAGECACHE_OVERFLOW, bReset);
23447	displayStatLine(out, "Largest Allocation:",
23448	"%lld bytes", SQLITE_STATUS_MALLOC_SIZE, bReset);
23449	displayStatLine(out, "Largest Pcache Allocation:",
23450	"%lld bytes", SQLITE_STATUS_PAGECACHE_SIZE, bReset);
23451	#ifdef YYTRACKMAXSTACKDEPTH
23452	displayStatLine(out, "Deepest Parser Stack:",
23453	"%lld (max %lld)", SQLITE_STATUS_PARSER_STACK, bReset);
23454	#endif
23455
23456	if( db ){
23457	if( pArg->shellFlgs & SHFLG_Lookaside ){
23458	iHiwtr = iCur = -1;
23459	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_USED,
23460	&iCur, &iHiwtr, bReset);
23461	sqlite3_fprintf(out,
23462	"Lookaside Slots Used: %d (max %d)\n", iCur, iHiwtr);
23463	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_HIT,
23464	&iCur, &iHiwtr, bReset);
23465	sqlite3_fprintf(out,
23466	"Successful lookaside attempts: %d\n", iHiwtr);
23467	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_MISS_SIZE,
23468	&iCur, &iHiwtr, bReset);
23469	sqlite3_fprintf(out,
23470	"Lookaside failures due to size: %d\n", iHiwtr);
23471	sqlite3_db_status(db, SQLITE_DBSTATUS_LOOKASIDE_MISS_FULL,
23472	&iCur, &iHiwtr, bReset);
23473	sqlite3_fprintf(out,
23474	"Lookaside failures due to OOM: %d\n", iHiwtr);
23475	}
23476	iHiwtr = iCur = -1;
23477	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_USED, &iCur, &iHiwtr, bReset);
23478	sqlite3_fprintf(out,
23479	"Pager Heap Usage: %d bytes\n", iCur);
23480	iHiwtr = iCur = -1;
23481	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_HIT, &iCur, &iHiwtr, 1);
23482	sqlite3_fprintf(out,
23483	"Page cache hits: %d\n", iCur);
23484	iHiwtr = iCur = -1;
23485	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_MISS, &iCur, &iHiwtr, 1);
23486	sqlite3_fprintf(out,
23487	"Page cache misses: %d\n", iCur);
23488	iHiwtr = iCur = -1;
23489	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_WRITE, &iCur, &iHiwtr, 1);
23490	sqlite3_fprintf(out,
23491	"Page cache writes: %d\n", iCur);
23492	iHiwtr = iCur = -1;
23493	sqlite3_db_status(db, SQLITE_DBSTATUS_CACHE_SPILL, &iCur, &iHiwtr, 1);
23494	sqlite3_fprintf(out,
23495	"Page cache spills: %d\n", iCur);
23496	iHiwtr = iCur = -1;
23497	sqlite3_db_status(db, SQLITE_DBSTATUS_SCHEMA_USED, &iCur, &iHiwtr, bReset);
23498	sqlite3_fprintf(out,
23499	"Schema Heap Usage: %d bytes\n", iCur);
23500	iHiwtr = iCur = -1;
23501	sqlite3_db_status(db, SQLITE_DBSTATUS_STMT_USED, &iCur, &iHiwtr, bReset);
23502	sqlite3_fprintf(out,
23503	"Statement Heap/Lookaside Usage: %d bytes\n", iCur);
23504	}
23505
23506	if( pArg->pStmt ){
23507	int iHit, iMiss;
23508	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FULLSCAN_STEP,
23509	bReset);
23510	sqlite3_fprintf(out,
23511	"Fullscan Steps: %d\n", iCur);
23512	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_SORT, bReset);
23513	sqlite3_fprintf(out,
23514	"Sort Operations: %d\n", iCur);
23515	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_AUTOINDEX,bReset);
23516	sqlite3_fprintf(out,
23517	"Autoindex Inserts: %d\n", iCur);
23518	iHit = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FILTER_HIT,
23519	bReset);
23520	iMiss = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_FILTER_MISS,
23521	bReset);
23522	if( iHit \|\| iMiss ){
23523	sqlite3_fprintf(out,
23524	"Bloom filter bypass taken: %d/%d\n", iHit, iHit+iMiss);
23525	}
23526	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_VM_STEP, bReset);
23527	sqlite3_fprintf(out,
23528	"Virtual Machine Steps: %d\n", iCur);
23529	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_REPREPARE,bReset);
23530	sqlite3_fprintf(out,
23531	"Reprepare operations: %d\n", iCur);
23532	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_RUN, bReset);
23533	sqlite3_fprintf(out,
23534	"Number of times run: %d\n", iCur);
23535	iCur = sqlite3_stmt_status(pArg->pStmt, SQLITE_STMTSTATUS_MEMUSED, bReset);
23536	sqlite3_fprintf(out,
23537	"Memory used by prepared stmt: %d\n", iCur);
23538	}
23539
23540	#ifdef __linux__
23541	displayLinuxIoStats(pArg->out);
23542	#endif
23543
23544	/* Do not remove this machine readable comment: extra-stats-output-here */
23545
23546	return 0;
	@@ -22721,21 +23932,21 @@
23932	#define BOX_1234 "\342\224\274" /* U+253c -\|- */
23933
23934	/* Draw horizontal line N characters long using unicode box
23935	** characters
23936	*/
23937	static void print_box_line(FILE *out, int N){
23938	const char zDash[] =
23939	BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24
23940	BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24 BOX_24;
23941	const int nDash = sizeof(zDash) - 1;
23942	N *= 3;
23943	while( N>nDash ){
23944	sqlite3_fputs(zDash, out);
23945	N -= nDash;
23946	}
23947	sqlite3_fprintf(out, "%.*s", N, zDash);
23948	}
23949
23950	/*
23951	** Draw a horizontal separator for a MODE_Box table.
23952	*/
	@@ -22746,19 +23957,19 @@
23957	const char *zSep2,
23958	const char *zSep3
23959	){
23960	int i;
23961	if( nArg>0 ){
23962	sqlite3_fputs(zSep1, p->out);
23963	print_box_line(p->out, p->actualWidth[0]+2);
23964	for(i=1; i<nArg; i++){
23965	sqlite3_fputs(zSep2, p->out);
23966	print_box_line(p->out, p->actualWidth[i]+2);
23967	}
23968	sqlite3_fputs(zSep3, p->out);
23969	}
23970	sqlite3_fputs("\n", p->out);
23971	}
23972
23973	/*
23974	** z[] is a line of text that is to be displayed the .mode box or table or
23975	** similar tabular formats. z[] might contain control characters such
	@@ -22788,16 +23999,26 @@
23999	}
24000	if( mxWidth<0 ) mxWidth = -mxWidth;
24001	if( mxWidth==0 ) mxWidth = 1000000;
24002	i = j = n = 0;
24003	while( n<mxWidth ){
24004	unsigned char c = z[i];
24005	if( c>=0xc0 ){
24006	int u;
24007	int len = decodeUtf8(&z[i], &u);
24008	i += len;
24009	j += len;
24010	n += cli_wcwidth(u);
24011	continue;
24012	}
24013	if( c>=' ' ){
24014	n++;
24015	i++;
24016	j++;
24017	continue;
24018	}
24019	if( c=='\t' ){
24020	do{
24021	n++;
24022	j++;
24023	}while( (n&7)!=0 && n<mxWidth );
24024	i++;
	@@ -22835,13 +24056,21 @@
24056	}
24057	zOut = malloc( j+1 );
24058	shell_check_oom(zOut);
24059	i = j = n = 0;
24060	while( i<k ){
24061	unsigned char c = z[i];
24062	if( c>=0xc0 ){
24063	int u;
24064	int len = decodeUtf8(&z[i], &u);
24065	do{ zOut[j++] = z[i++]; }while( (--len)>0 );
24066	n += cli_wcwidth(u);
24067	continue;
24068	}
24069	if( c>=' ' ){
24070	n++;
24071	zOut[j++] = z[i++];
24072	continue;
24073	}
24074	if( z[i]=='\t' ){
24075	do{
24076	n++;
	@@ -23017,97 +24246,99 @@
24246	rowSep = "\n";
24247	if( p->showHeader ){
24248	for(i=0; i<nColumn; i++){
24249	w = p->actualWidth[i];
24250	if( p->colWidth[i]<0 ) w = -w;
24251	utf8_width_print(p->out, w, azData[i]);
24252	sqlite3_fputs(i==nColumn-1?"\n":" ", p->out);
24253	}
24254	for(i=0; i<nColumn; i++){
24255	print_dashes(p->out, p->actualWidth[i]);
24256	sqlite3_fputs(i==nColumn-1?"\n":" ", p->out);
24257	}
24258	}
24259	break;
24260	}
24261	case MODE_Table: {
24262	colSep = " \| ";
24263	rowSep = " \|\n";
24264	print_row_separator(p, nColumn, "+");
24265	sqlite3_fputs("\| ", p->out);
24266	for(i=0; i<nColumn; i++){
24267	w = p->actualWidth[i];
24268	n = strlenChar(azData[i]);
24269	sqlite3_fprintf(p->out, "%s%s%s", (w-n)/2, "",
24270	azData[i], (w-n+1)/2, "");
24271	sqlite3_fputs(i==nColumn-1?" \|\n":" \| ", p->out);
24272	}
24273	print_row_separator(p, nColumn, "+");
24274	break;
24275	}
24276	case MODE_Markdown: {
24277	colSep = " \| ";
24278	rowSep = " \|\n";
24279	sqlite3_fputs("\| ", p->out);
24280	for(i=0; i<nColumn; i++){
24281	w = p->actualWidth[i];
24282	n = strlenChar(azData[i]);
24283	sqlite3_fprintf(p->out, "%s%s%s", (w-n)/2, "",
24284	azData[i], (w-n+1)/2, "");
24285	sqlite3_fputs(i==nColumn-1?" \|\n":" \| ", p->out);
24286	}
24287	print_row_separator(p, nColumn, "\|");
24288	break;
24289	}
24290	case MODE_Box: {
24291	colSep = " " BOX_13 " ";
24292	rowSep = " " BOX_13 "\n";
24293	print_box_row_separator(p, nColumn, BOX_23, BOX_234, BOX_34);
24294	sqlite3_fputs(BOX_13 " ", p->out);
24295	for(i=0; i<nColumn; i++){
24296	w = p->actualWidth[i];
24297	n = strlenChar(azData[i]);
24298	sqlite3_fprintf(p->out, "%s%s%s%s",
24299	(w-n)/2, "", azData[i], (w-n+1)/2, "",
24300	i==nColumn-1?" "BOX_13"\n":" "BOX_13" ");
24301	}
24302	print_box_row_separator(p, nColumn, BOX_123, BOX_1234, BOX_134);
24303	break;
24304	}
24305	}
24306	for(i=nColumn, j=0; i<nTotal; i++, j++){
24307	if( j==0 && p->cMode!=MODE_Column ){
24308	sqlite3_fputs(p->cMode==MODE_Box?BOX_13" ":"\| ", p->out);
24309	}
24310	z = azData[i];
24311	if( z==0 ) z = p->nullValue;
24312	w = p->actualWidth[j];
24313	if( p->colWidth[j]<0 ) w = -w;
24314	utf8_width_print(p->out, w, z);
24315	if( j==nColumn-1 ){
24316	sqlite3_fputs(rowSep, p->out);
24317	if( bMultiLineRowExists && abRowDiv[i/nColumn-1] && i+1<nTotal ){
24318	if( p->cMode==MODE_Table ){
24319	print_row_separator(p, nColumn, "+");
24320	}else if( p->cMode==MODE_Box ){
24321	print_box_row_separator(p, nColumn, BOX_123, BOX_1234, BOX_134);
24322	}else if( p->cMode==MODE_Column ){
24323	sqlite3_fputs("\n", p->out);
24324	}
24325	}
24326	j = -1;
24327	if( seenInterrupt ) goto columnar_end;
24328	}else{
24329	sqlite3_fputs(colSep, p->out);
24330	}
24331	}
24332	if( p->cMode==MODE_Table ){
24333	print_row_separator(p, nColumn, "+");
24334	}else if( p->cMode==MODE_Box ){
24335	print_box_row_separator(p, nColumn, BOX_12, BOX_124, BOX_14);
24336	}
24337	columnar_end:
24338	if( seenInterrupt ){
24339	sqlite3_fputs("Interrupt\n", p->out);
24340	}
24341	nData = (nRow+1)*nColumn;
24342	for(i=0; i<nData; i++){
24343	z = azData[i];
24344	if( z!=zEmpty && z!=zShowNull ) free(azData[i]);
	@@ -23190,11 +24421,13 @@
24421	}
24422	}
24423	} while( SQLITE_ROW == rc );
24424	sqlite3_free(pData);
24425	if( pArg->cMode==MODE_Json ){
24426	sqlite3_fputs("]\n", pArg->out);
24427	}else if( pArg->cMode==MODE_Www ){
24428	sqlite3_fputs("</TABLE>\n<PRE>\n", pArg->out);
24429	}else if( pArg->cMode==MODE_Count ){
24430	char zBuf[200];
24431	sqlite3_snprintf(sizeof(zBuf), zBuf, "%llu row%s\n",
24432	nRow, nRow!=1 ? "s" : "");
24433	printf("%s", zBuf);
	@@ -23239,10 +24472,11 @@
24472	int bCancel,
24473	char **pzErr
24474	){
24475	int rc = SQLITE_OK;
24476	sqlite3expert *p = pState->expert.pExpert;
24477	FILE *out = pState->out;
24478	assert( p );
24479	assert( bCancel \|\| pzErr==0 \|\| *pzErr==0 );
24480	if( bCancel==0 ){
24481	int bVerbose = pState->expert.bVerbose;
24482
	@@ -23251,24 +24485,25 @@
24485	int nQuery = sqlite3_expert_count(p);
24486	int i;
24487
24488	if( bVerbose ){
24489	const char *zCand = sqlite3_expert_report(p,0,EXPERT_REPORT_CANDIDATES);
24490	sqlite3_fputs("-- Candidates -----------------------------\n", out);
24491	sqlite3_fprintf(out, "%s\n", zCand);
24492	}
24493	for(i=0; i<nQuery; i++){
24494	const char *zSql = sqlite3_expert_report(p, i, EXPERT_REPORT_SQL);
24495	const char *zIdx = sqlite3_expert_report(p, i, EXPERT_REPORT_INDEXES);
24496	const char *zEQP = sqlite3_expert_report(p, i, EXPERT_REPORT_PLAN);
24497	if( zIdx==0 ) zIdx = "(no new indexes)\n";
24498	if( bVerbose ){
24499	sqlite3_fprintf(out,
24500	"-- Query %d --------------------------------\n"
24501	"%s\n\n"
24502	,i+1, zSql);
24503	}
24504	sqlite3_fprintf(out, "%s\n%s\n", zIdx, zEQP);

24505	}
24506	}
24507	}
24508	sqlite3_expert_destroy(p);
24509	pState->expert.pExpert = 0;
	@@ -23299,30 +24534,31 @@
24534	if( n>=2 && 0==cli_strncmp(z, "-verbose", n) ){
24535	pState->expert.bVerbose = 1;
24536	}
24537	else if( n>=2 && 0==cli_strncmp(z, "-sample", n) ){
24538	if( i==(nArg-1) ){
24539	sqlite3_fprintf(stderr, "option requires an argument: %s\n", z);
24540	rc = SQLITE_ERROR;
24541	}else{
24542	iSample = (int)integerValue(azArg[++i]);
24543	if( iSample<0 \|\| iSample>100 ){
24544	sqlite3_fprintf(stderr,"value out of range: %s\n", azArg[i]);
24545	rc = SQLITE_ERROR;
24546	}
24547	}
24548	}
24549	else{
24550	sqlite3_fprintf(stderr,"unknown option: %s\n", z);
24551	rc = SQLITE_ERROR;
24552	}
24553	}
24554
24555	if( rc==SQLITE_OK ){
24556	pState->expert.pExpert = sqlite3_expert_new(pState->db, &zErr);
24557	if( pState->expert.pExpert==0 ){
24558	sqlite3_fprintf(stderr,
24559	"sqlite3_expert_new: %s\n", zErr ? zErr : "out of memory");
24560	rc = SQLITE_ERROR;
24561	}else{
24562	sqlite3_expert_config(
24563	pState->expert.pExpert, EXPERT_CONFIG_SAMPLE, iSample
24564	);
	@@ -23647,33 +24883,33 @@
24883	if( zType==0 ) return 0;
24884	dataOnly = (p->shellFlgs & SHFLG_DumpDataOnly)!=0;
24885	noSys = (p->shellFlgs & SHFLG_DumpNoSys)!=0;
24886
24887	if( cli_strcmp(zTable, "sqlite_sequence")==0 && !noSys ){
24888	/* no-op */
24889	}else if( sqlite3_strglob("sqlite_stat?", zTable)==0 && !noSys ){
24890	if( !dataOnly ) sqlite3_fputs("ANALYZE sqlite_schema;\n", p->out);
24891	}else if( cli_strncmp(zTable, "sqlite_", 7)==0 ){
24892	return 0;
24893	}else if( dataOnly ){
24894	/* no-op */
24895	}else if( cli_strncmp(zSql, "CREATE VIRTUAL TABLE", 20)==0 ){
24896	char *zIns;
24897	if( !p->writableSchema ){
24898	sqlite3_fputs("PRAGMA writable_schema=ON;\n", p->out);
24899	p->writableSchema = 1;
24900	}
24901	zIns = sqlite3_mprintf(
24902	"INSERT INTO sqlite_schema(type,name,tbl_name,rootpage,sql)"
24903	"VALUES('table','%q','%q',0,'%q');",
24904	zTable, zTable, zSql);
24905	shell_check_oom(zIns);
24906	sqlite3_fprintf(p->out, "%s\n", zIns);
24907	sqlite3_free(zIns);
24908	return 0;
24909	}else{
24910	printSchemaLine(p->out, zSql, ";\n");
24911	}
24912
24913	if( cli_strcmp(zType, "table")==0 ){
24914	ShellText sSelect;
24915	ShellText sTable;
	@@ -23727,11 +24963,11 @@
24963	savedMode = p->mode;
24964	p->zDestTable = sTable.z;
24965	p->mode = p->cMode = MODE_Insert;
24966	rc = shell_exec(p, sSelect.z, 0);
24967	if( (rc&0xff)==SQLITE_CORRUPT ){
24968	sqlite3_fputs("/**** CORRUPTION ERROR *****/\n", p->out);
24969	toggleSelectOrder(p->db);
24970	shell_exec(p, sSelect.z, 0);
24971	toggleSelectOrder(p->db);
24972	}
24973	p->zDestTable = savedDestTable;
	@@ -23758,28 +24994,28 @@
24994	char *zErr = 0;
24995	rc = sqlite3_exec(p->db, zQuery, dump_callback, p, &zErr);
24996	if( rc==SQLITE_CORRUPT ){
24997	char *zQ2;
24998	int len = strlen30(zQuery);
24999	sqlite3_fputs("/**** CORRUPTION ERROR *****/\n", p->out);
25000	if( zErr ){
25001	sqlite3_fprintf(p->out, "/**** %s ****/\n", zErr);
25002	sqlite3_free(zErr);
25003	zErr = 0;
25004	}
25005	zQ2 = malloc( len+100 );
25006	if( zQ2==0 ) return rc;
25007	sqlite3_snprintf(len+100, zQ2, "%s ORDER BY rowid DESC", zQuery);
25008	rc = sqlite3_exec(p->db, zQ2, dump_callback, p, &zErr);
25009	if( rc ){
25010	sqlite3_fprintf(p->out, "/**** ERROR: %s ****/\n", zErr);
25011	}else{
25012	rc = SQLITE_CORRUPT;
25013	}

25014	free(zQ2);
25015	}
25016	sqlite3_free(zErr);
25017	return rc;
25018	}
25019
25020	/*
25021	** Text of help messages.
	@@ -23832,18 +25068,17 @@
25068	#ifndef SQLITE_SHELL_FIDDLE
25069	".check GLOB Fail if output since .testcase does not match",
25070	".clone NEWDB Clone data into NEWDB from the existing database",
25071	#endif
25072	".connection [close] [#] Open or close an auxiliary database connection",
25073	".crlf ?on\|off? Whether or not to use \\r\\n line endings",


25074	".databases List names and files of attached databases",
25075	".dbconfig ?op? ?val? List or change sqlite3_db_config() options",
25076	#if SQLITE_SHELL_HAVE_RECOVER
25077	".dbinfo ?DB? Show status information about the database",
25078	#endif
25079	".dbtotxt Hex dump of the database file",
25080	".dump ?OBJECTS? Render database content as SQL",
25081	" Options:",
25082	" --data-only Output only INSERT statements",
25083	" --newlines Allow unescaped newline characters in output",
25084	" --nosys Omit system tables (ex: \"sqlite_stat1\")",
	@@ -23940,13 +25175,15 @@
25175	#endif
25176	".nullvalue STRING Use STRING in place of NULL values",
25177	#ifndef SQLITE_SHELL_FIDDLE
25178	".once ?OPTIONS? ?FILE? Output for the next SQL command only to FILE",
25179	" If FILE begins with '\|' then open as a pipe",
25180	" --bom Put a UTF8 byte-order mark at the beginning",
25181	" -e Send output to the system text editor",
25182	" --plain Use text/plain output instead of HTML for -w option",
25183	" -w Send output as HTML to a web browser (same as \".www\")",
25184	" -x Send output as CSV to a spreadsheet (same as \".excel\")",
25185	/* Note that .open is (partially) available in WASM builds but is
25186	** currently only intended to be used by the fiddle tool, not
25187	** end users, so is "undocumented." */
25188	".open ?OPTIONS? ?FILE? Close existing database and reopen FILE",
25189	" Options:",
	@@ -23965,10 +25202,12 @@
25202	".output ?FILE? Send output to FILE or stdout if FILE is omitted",
25203	" If FILE begins with '\|' then open it as a pipe.",
25204	" Options:",
25205	" --bom Prefix output with a UTF8 byte-order mark",
25206	" -e Send output to the system text editor",
25207	" --plain Use text/plain for -w option",
25208	" -w Send output to a web browser",
25209	" -x Send output as CSV to a spreadsheet",
25210	#endif
25211	".parameter CMD ... Manage SQL parameter bindings",
25212	" clear Erase all bindings",
25213	" init Initialize the TEMP table that holds bindings",
	@@ -24078,10 +25317,14 @@
25317	".vfsinfo ?AUX? Information about the top-level VFS",
25318	".vfslist List all available VFSes",
25319	".vfsname ?AUX? Print the name of the VFS stack",
25320	".width NUM1 NUM2 ... Set minimum column widths for columnar output",
25321	" Negative values right-justify",
25322	#ifndef SQLITE_SHELL_FIDDLE
25323	".www Display output of the next command in web browser",
25324	" --plain Show results as text/plain, not as HTML",
25325	#endif
25326	};
25327
25328	/*
25329	** Output help text.
25330	**
	@@ -24126,24 +25369,24 @@
25369	hh &= ~HH_Summary;
25370	break;
25371	}
25372	if( ((hw^hh)&HH_Undoc)==0 ){
25373	if( (hh&HH_Summary)!=0 ){
25374	sqlite3_fprintf(out, ".%s\n", azHelp[i]+1);
25375	++n;
25376	}else if( (hw&HW_SummaryOnly)==0 ){
25377	sqlite3_fprintf(out, "%s\n", azHelp[i]);
25378	}
25379	}
25380	}
25381	}else{
25382	/* Seek documented commands for which zPattern is an exact prefix */
25383	zPat = sqlite3_mprintf(".%s*", zPattern);
25384	shell_check_oom(zPat);
25385	for(i=0; i<ArraySize(azHelp); i++){
25386	if( sqlite3_strglob(zPat, azHelp[i])==0 ){
25387	sqlite3_fprintf(out, "%s\n", azHelp[i]);
25388	j = i+1;
25389	n++;
25390	}
25391	}
25392	sqlite3_free(zPat);
	@@ -24150,11 +25393,11 @@
25393	if( n ){
25394	if( n==1 ){
25395	/* when zPattern is a prefix of exactly one command, then include
25396	** the details of that command, which should begin at offset j */
25397	while( j<ArraySize(azHelp)-1 && azHelp[j][0]==' ' ){
25398	sqlite3_fprintf(out, "%s\n", azHelp[j]);
25399	j++;
25400	}
25401	}
25402	return n;
25403	}
	@@ -24167,14 +25410,14 @@
25410	while( i<ArraySize(azHelp)-1 && azHelp[i+1][0]==' ' ) ++i;
25411	continue;
25412	}
25413	if( azHelp[i][0]=='.' ) j = i;
25414	if( sqlite3_strlike(zPat, azHelp[i], 0)==0 ){
25415	sqlite3_fprintf(out, "%s\n", azHelp[j]);
25416	while( j<ArraySize(azHelp)-1 && azHelp[j+1][0]==' ' ){
25417	j++;
25418	sqlite3_fprintf(out, "%s\n", azHelp[j]);
25419	}
25420	i = j;
25421	n++;
25422	}
25423	}
	@@ -24200,35 +25443,35 @@
25443	**
25444	** NULL is returned if any error is encountered. The final value of *pnByte
25445	** is undefined in this case.
25446	*/
25447	static char readFile(const char zName, int *pnByte){
25448	FILE *in = sqlite3_fopen(zName, "rb");
25449	long nIn;
25450	size_t nRead;
25451	char *pBuf;
25452	int rc;
25453	if( in==0 ) return 0;
25454	rc = fseek(in, 0, SEEK_END);
25455	if( rc!=0 ){
25456	sqlite3_fprintf(stderr,"Error: '%s' not seekable\n", zName);
25457	fclose(in);
25458	return 0;
25459	}
25460	nIn = ftell(in);
25461	rewind(in);
25462	pBuf = sqlite3_malloc64( nIn+1 );
25463	if( pBuf==0 ){
25464	sqlite3_fputs("Error: out of memory\n", stderr);
25465	fclose(in);
25466	return 0;
25467	}
25468	nRead = fread(pBuf, nIn, 1, in);
25469	fclose(in);
25470	if( nRead!=1 ){
25471	sqlite3_free(pBuf);
25472	sqlite3_fprintf(stderr,"Error: cannot read '%s'\n", zName);
25473	return 0;
25474	}
25475	pBuf[nIn] = 0;
25476	if( pnByte ) *pnByte = nIn;
25477	return pBuf;
	@@ -24290,11 +25533,11 @@
25533	** archive and the dfltZip flag is true, then assume it is a ZIP archive.
25534	** Otherwise, assume an ordinary database regardless of the filename if
25535	** the type cannot be determined from content.
25536	*/
25537	int deduceDatabaseType(const char *zName, int dfltZip){
25538	FILE *f = sqlite3_fopen(zName, "rb");
25539	size_t n;
25540	int rc = SHELL_OPEN_UNSPEC;
25541	char zBuf[100];
25542	if( f==0 ){
25543	if( dfltZip && sqlite3_strlike("%.zip",zName,0)==0 ){
	@@ -24343,13 +25586,13 @@
25586	FILE *in;
25587	const char *zDbFilename = p->pAuxDb->zDbFilename;
25588	unsigned int x[16];
25589	char zLine[1000];
25590	if( zDbFilename ){
25591	in = sqlite3_fopen(zDbFilename, "r");
25592	if( in==0 ){
25593	sqlite3_fprintf(stderr,"cannot open \"%s\" for reading\n", zDbFilename);
25594	return 0;
25595	}
25596	nLine = 0;
25597	}else{
25598	in = p->in;
	@@ -24356,24 +25599,24 @@
25599	nLine = p->lineno;
25600	if( in==0 ) in = stdin;
25601	}
25602	*pnData = 0;
25603	nLine++;
25604	if( sqlite3_fgets(zLine, sizeof(zLine), in)==0 ) goto readHexDb_error;
25605	rc = sscanf(zLine, "\| size %d pagesize %d", &n, &pgsz);
25606	if( rc!=2 ) goto readHexDb_error;
25607	if( n<0 ) goto readHexDb_error;
25608	if( pgsz<512 \|\| pgsz>65536 \|\| (pgsz&(pgsz-1))!=0 ) goto readHexDb_error;
25609	n = (n+pgsz-1)&~(pgsz-1); /* Round n up to the next multiple of pgsz */
25610	a = sqlite3_malloc( n ? n : 1 );
25611	shell_check_oom(a);
25612	memset(a, 0, n);
25613	if( pgsz<512 \|\| pgsz>65536 \|\| (pgsz & (pgsz-1))!=0 ){
25614	sqlite3_fputs("invalid pagesize\n", stderr);
25615	goto readHexDb_error;
25616	}
25617	for(nLine++; sqlite3_fgets(zLine, sizeof(zLine), in)!=0; nLine++){
25618	rc = sscanf(zLine, "\| page %d offset %d", &j, &k);
25619	if( rc==2 ){
25620	iOffset = k;
25621	continue;
25622	}
	@@ -24401,18 +25644,18 @@
25644
25645	readHexDb_error:
25646	if( in!=p->in ){
25647	fclose(in);
25648	}else{
25649	while( sqlite3_fgets(zLine, sizeof(zLine), p->in)!=0 ){
25650	nLine++;
25651	if(cli_strncmp(zLine, "\| end ", 6)==0 ) break;
25652	}
25653	p->lineno = nLine;
25654	}
25655	sqlite3_free(a);
25656	sqlite3_fprintf(stderr,"Error on line %d of --hexdb input\n", nLine);
25657	return 0;
25658	}
25659	#endif /* SQLITE_OMIT_DESERIALIZE */
25660
25661	/*
	@@ -24483,22 +25726,24 @@
25726	SQLITE_OPEN_READWRITE\|SQLITE_OPEN_CREATE\|p->openFlags, 0);
25727	break;
25728	}
25729	}
25730	if( p->db==0 \|\| SQLITE_OK!=sqlite3_errcode(p->db) ){
25731	sqlite3_fprintf(stderr,"Error: unable to open database \"%s\": %s\n",
25732	zDbFilename, sqlite3_errmsg(p->db));
25733	if( (openFlags & OPEN_DB_KEEPALIVE)==0 ){
25734	exit(1);
25735	}
25736	sqlite3_close(p->db);
25737	sqlite3_open(":memory:", &p->db);
25738	if( p->db==0 \|\| SQLITE_OK!=sqlite3_errcode(p->db) ){
25739	sqlite3_fputs("Also: unable to open substitute in-memory database.\n",
25740	stderr);
25741	exit(1);
25742	}else{
25743	sqlite3_fprintf(stderr,
25744	"Notice: using substitute in-memory database instead of \"%s\"\n",
25745	zDbFilename);
25746	}
25747	}
25748	globalDb = p->db;
25749	sqlite3_db_config(p->db, SQLITE_DBCONFIG_STMT_SCANSTATUS, (int)0, (int*)0);
	@@ -24511,10 +25756,11 @@
25756	}
25757
25758	#ifndef SQLITE_OMIT_LOAD_EXTENSION
25759	sqlite3_enable_load_extension(p->db, 1);
25760	#endif
25761	sqlite3_sha_init(p->db, 0, 0);
25762	sqlite3_shathree_init(p->db, 0, 0);
25763	sqlite3_uint_init(p->db, 0, 0);
25764	sqlite3_stmtrand_init(p->db, 0, 0);
25765	sqlite3_decimal_init(p->db, 0, 0);
25766	sqlite3_percentile_init(p->db, 0, 0);
	@@ -24605,11 +25851,11 @@
25851	}
25852	rc = sqlite3_deserialize(p->db, "main", aData, nData, nData,
25853	SQLITE_DESERIALIZE_RESIZEABLE \|
25854	SQLITE_DESERIALIZE_FREEONCLOSE);
25855	if( rc ){
25856	sqlite3_fprintf(stderr,"Error: sqlite3_deserialize() returns %d\n", rc);
25857	}
25858	if( p->szMax>0 ){
25859	sqlite3_file_control(p->db, "main", SQLITE_FCNTL_SIZE_LIMIT, &p->szMax);
25860	}
25861	}
	@@ -24629,15 +25875,17 @@
25875	** Attempt to close the database connection. Report errors.
25876	*/
25877	void close_db(sqlite3 *db){
25878	int rc = sqlite3_close(db);
25879	if( rc ){
25880	sqlite3_fprintf(stderr,
25881	"Error: sqlite3_close() returns %d: %s\n", rc, sqlite3_errmsg(db));
25882	}
25883	}
25884
25885	#if (HAVE_READLINE \|\| HAVE_EDITLINE) \
25886	&& !defined(SQLITE_OMIT_READLINE_COMPLETION)
25887	/*
25888	** Readline completion callbacks
25889	*/
25890	static char readline_completion_generator(const char text, int state){
25891	static sqlite3_stmt *pStmt = 0;
	@@ -24671,19 +25919,26 @@
25919	#elif HAVE_LINENOISE
25920	/*
25921	** Linenoise completion callback. Note that the 3rd argument is from
25922	** the "msteveb" version of linenoise, not the "antirez" version.
25923	*/
25924	static void linenoise_completion(
25925	const char *zLine,
25926	linenoiseCompletions *lc
25927	#if HAVE_LINENOISE==2
25928	,void *pUserData
25929	#endif
25930	){
25931	i64 nLine = strlen(zLine);
25932	i64 i, iStart;
25933	sqlite3_stmt *pStmt = 0;
25934	char *zSql;
25935	char zBuf[1000];
25936
25937	#if HAVE_LINENOISE==2
25938	UNUSED_PARAMETER(pUserData);
25939	#endif
25940	if( nLine>(i64)sizeof(zBuf)-30 ) return;
25941	if( zLine[0]=='.' \|\| zLine[0]=='#') return;
25942	for(i=nLine-1; i>=0 && (isalnum(zLine[i]) \|\| zLine[i]=='_'); i--){}
25943	if( i==nLine-1 ) return;
25944	iStart = i+1;
	@@ -24793,11 +26048,12 @@
26048	return 1;
26049	}
26050	if( sqlite3_stricmp(zArg, "off")==0 \|\| sqlite3_stricmp(zArg,"no")==0 ){
26051	return 0;
26052	}
26053	sqlite3_fprintf(stderr,
26054	"ERROR: Not a boolean value: \"%s\". Assuming \"no\".\n", zArg);
26055	return 0;
26056	}
26057
26058	/*
26059	** Set or clear a shell flag according to a boolean value.
	@@ -24820,22 +26076,22 @@
26076	/*
26077	** Try to open an output file. The names "stdout" and "stderr" are
26078	** recognized and do the right thing. NULL is returned if the output
26079	** filename is "off".
26080	*/
26081	static FILE output_file_open(const char zFile){
26082	FILE *f;
26083	if( cli_strcmp(zFile,"stdout")==0 ){
26084	f = stdout;
26085	}else if( cli_strcmp(zFile, "stderr")==0 ){
26086	f = stderr;
26087	}else if( cli_strcmp(zFile, "off")==0 ){
26088	f = 0;
26089	}else{
26090	f = sqlite3_fopen(zFile, "w");
26091	if( f==0 ){
26092	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zFile);
26093	}
26094	}
26095	return f;
26096	}
26097
	@@ -24884,16 +26140,17 @@
26140	if( nSql>1000000000 ) nSql = 1000000000;
26141	while( nSql>0 && zSql[nSql-1]==';' ){ nSql--; }
26142	switch( mType ){
26143	case SQLITE_TRACE_ROW:
26144	case SQLITE_TRACE_STMT: {
26145	sqlite3_fprintf(p->traceOut, "%.*s;\n", (int)nSql, zSql);
26146	break;
26147	}
26148	case SQLITE_TRACE_PROFILE: {
26149	sqlite3_int64 nNanosec = pX ? (sqlite3_int64)pX : 0;
26150	sqlite3_fprintf(p->traceOut,
26151	"%.*s; -- %lld ns\n", (int)nSql, zSql, nNanosec);
26152	break;
26153	}
26154	}
26155	return 0;
26156	}
	@@ -24996,14 +26253,15 @@
26253	do{ p->n--; }while( p->z[p->n]!=cQuote );
26254	p->cTerm = c;
26255	break;
26256	}
26257	if( pc==cQuote && c!='\r' ){
26258	sqlite3_fprintf(stderr,"%s:%d: unescaped %c character\n",
26259	p->zFile, p->nLine, cQuote);
26260	}
26261	if( c==EOF ){
26262	sqlite3_fprintf(stderr,"%s:%d: unterminated %c-quoted field\n",
26263	p->zFile, startLine, cQuote);
26264	p->cTerm = c;
26265	break;
26266	}
26267	import_append_char(p, c);
	@@ -25098,11 +26356,11 @@
26356
26357	zQuery = sqlite3_mprintf("SELECT * FROM \"%w\"", zTable);
26358	shell_check_oom(zQuery);
26359	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
26360	if( rc ){
26361	sqlite3_fprintf(stderr,"Error %d: %s on [%s]\n",
26362	sqlite3_extended_errcode(p->db), sqlite3_errmsg(p->db), zQuery);
26363	goto end_data_xfer;
26364	}
26365	n = sqlite3_column_count(pQuery);
26366	zInsert = sqlite3_malloc64(200 + nTable + n*3);
	@@ -25115,11 +26373,11 @@
26373	i += 2;
26374	}
26375	memcpy(zInsert+i, ");", 3);
26376	rc = sqlite3_prepare_v2(newDb, zInsert, -1, &pInsert, 0);
26377	if( rc ){
26378	sqlite3_fprintf(stderr,"Error %d: %s on [%s]\n",
26379	sqlite3_extended_errcode(newDb), sqlite3_errmsg(newDb), zInsert);
26380	goto end_data_xfer;
26381	}
26382	for(k=0; k<2; k++){
26383	while( (rc = sqlite3_step(pQuery))==SQLITE_ROW ){
	@@ -25151,11 +26409,11 @@
26409	}
26410	}
26411	} /* End for */
26412	rc = sqlite3_step(pInsert);
26413	if( rc!=SQLITE_OK && rc!=SQLITE_ROW && rc!=SQLITE_DONE ){
26414	sqlite3_fprintf(stderr,"Error %d: %s\n",
26415	sqlite3_extended_errcode(newDb), sqlite3_errmsg(newDb));
26416	}
26417	sqlite3_reset(pInsert);
26418	cnt++;
26419	if( (cnt%spinRate)==0 ){
	@@ -25169,11 +26427,11 @@
26427	zQuery = sqlite3_mprintf("SELECT * FROM \"%w\" ORDER BY rowid DESC;",
26428	zTable);
26429	shell_check_oom(zQuery);
26430	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
26431	if( rc ){
26432	sqlite3_fprintf(stderr,"Warning: cannot step \"%s\" backwards", zTable);
26433	break;
26434	}
26435	} /* End for(k=0...) */
26436
26437	end_data_xfer:
	@@ -25206,23 +26464,24 @@
26464	zQuery = sqlite3_mprintf("SELECT name, sql FROM sqlite_schema"
26465	" WHERE %s ORDER BY rowid ASC", zWhere);
26466	shell_check_oom(zQuery);
26467	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
26468	if( rc ){
26469	sqlite3_fprintf(stderr,
26470	"Error: (%d) %s on [%s]\n", sqlite3_extended_errcode(p->db),
26471	sqlite3_errmsg(p->db), zQuery);
26472	goto end_schema_xfer;
26473	}
26474	while( (rc = sqlite3_step(pQuery))==SQLITE_ROW ){
26475	zName = sqlite3_column_text(pQuery, 0);
26476	zSql = sqlite3_column_text(pQuery, 1);
26477	if( zName==0 \|\| zSql==0 ) continue;
26478	if( sqlite3_stricmp((char*)zName, "sqlite_sequence")!=0 ){
26479	sqlite3_fprintf(stdout, "%s... ", zName); fflush(stdout);
26480	sqlite3_exec(newDb, (const char*)zSql, 0, 0, &zErrMsg);
26481	if( zErrMsg ){
26482	sqlite3_fprintf(stderr,"Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
26483	sqlite3_free(zErrMsg);
26484	zErrMsg = 0;
26485	}
26486	}
26487	if( xForEach ){
	@@ -25236,23 +26495,23 @@
26495	zQuery = sqlite3_mprintf("SELECT name, sql FROM sqlite_schema"
26496	" WHERE %s ORDER BY rowid DESC", zWhere);
26497	shell_check_oom(zQuery);
26498	rc = sqlite3_prepare_v2(p->db, zQuery, -1, &pQuery, 0);
26499	if( rc ){
26500	sqlite3_fprintf(stderr,"Error: (%d) %s on [%s]\n",
26501	sqlite3_extended_errcode(p->db), sqlite3_errmsg(p->db), zQuery);
26502	goto end_schema_xfer;
26503	}
26504	while( sqlite3_step(pQuery)==SQLITE_ROW ){
26505	zName = sqlite3_column_text(pQuery, 0);
26506	zSql = sqlite3_column_text(pQuery, 1);
26507	if( zName==0 \|\| zSql==0 ) continue;
26508	if( sqlite3_stricmp((char*)zName, "sqlite_sequence")==0 ) continue;
26509	sqlite3_fprintf(stdout, "%s... ", zName); fflush(stdout);
26510	sqlite3_exec(newDb, (const char*)zSql, 0, 0, &zErrMsg);
26511	if( zErrMsg ){
26512	sqlite3_fprintf(stderr,"Error: %s\nSQL: [%s]\n", zErrMsg, zSql);
26513	sqlite3_free(zErrMsg);
26514	zErrMsg = 0;
26515	}
26516	if( xForEach ){
26517	xForEach(p, newDb, (const char*)zName);
	@@ -25272,16 +26531,17 @@
26531	*/
26532	static void tryToClone(ShellState p, const char zNewDb){
26533	int rc;
26534	sqlite3 *newDb = 0;
26535	if( access(zNewDb,0)==0 ){
26536	sqlite3_fprintf(stderr,"File \"%s\" already exists.\n", zNewDb);
26537	return;
26538	}
26539	rc = sqlite3_open(zNewDb, &newDb);
26540	if( rc ){
26541	sqlite3_fprintf(stderr,
26542	"Cannot create output database: %s\n", sqlite3_errmsg(newDb));
26543	}else{
26544	sqlite3_exec(p->db, "PRAGMA writable_schema=ON;", 0, 0, 0);
26545	sqlite3_exec(newDb, "BEGIN EXCLUSIVE;", 0, 0, 0);
26546	tryToCloneSchema(p, newDb, "type='table'", tryToCloneData);
26547	tryToCloneSchema(p, newDb, "type!='table'", 0);
	@@ -25294,14 +26554,22 @@
26554	#ifndef SQLITE_SHELL_FIDDLE
26555	/*
26556	** Change the output stream (file or pipe or console) to something else.
26557	*/
26558	static void output_redir(ShellState p, FILE pfNew){
26559	if( p->out != stdout ){
26560	sqlite3_fputs("Output already redirected.\n", stderr);
26561	}else{
26562	p->out = pfNew;
26563	setCrlfMode(p);
26564	if( p->mode==MODE_Www ){
26565	sqlite3_fputs(
26566	"<!DOCTYPE html>\n"
26567	"<HTML><BODY><PRE>\n",
26568	p->out
26569	);
26570	}
26571	}
26572	}
26573
26574	/*
26575	** Change the output file back to stdout.
	@@ -25314,10 +26582,13 @@
26582	if( p->outfile[0]=='\|' ){
26583	#ifndef SQLITE_OMIT_POPEN
26584	pclose(p->out);
26585	#endif
26586	}else{
26587	if( p->mode==MODE_Www ){
26588	sqlite3_fputs("</PRE></BODY></HTML>\n", p->out);
26589	}
26590	output_file_close(p->out);
26591	#ifndef SQLITE_NOHAVE_SYSTEM
26592	if( p->doXdgOpen ){
26593	const char *zXdgOpenCmd =
26594	#if defined(_WIN32)
	@@ -25328,11 +26599,11 @@
26599	"xdg-open";
26600	#endif
26601	char *zCmd;
26602	zCmd = sqlite3_mprintf("%s %s", zXdgOpenCmd, p->zTempFile);
26603	if( system(zCmd) ){
26604	sqlite3_fprintf(stderr,"Failed: [%s]\n", zCmd);
26605	}else{
26606	/* Give the start/open/xdg-open command some time to get
26607	** going before we continue, and potential delete the
26608	** p->zTempFile data file out from under it */
26609	sqlite3_sleep(2000);
	@@ -25343,28 +26614,34 @@
26614	}
26615	#endif /* !defined(SQLITE_NOHAVE_SYSTEM) */
26616	}
26617	p->outfile[0] = 0;
26618	p->out = stdout;
26619	setCrlfMode(p);
26620	}
26621	#else
26622	# define output_redir(SS,pfO)
26623	# define output_reset(SS)
26624	#endif
26625
26626	/*
26627	** Run an SQL command and return the single integer result.
26628	*/
26629	static int db_int(sqlite3 db, const char zSql, ...){
26630	sqlite3_stmt *pStmt;
26631	int res = 0;
26632	char *z;
26633	va_list ap;
26634	va_start(ap, zSql);
26635	z = sqlite3_vmprintf(zSql, ap);
26636	va_end(ap);
26637	sqlite3_prepare_v2(db, z, -1, &pStmt, 0);
26638	if( pStmt && sqlite3_step(pStmt)==SQLITE_ROW ){
26639	res = sqlite3_column_int(pStmt,0);
26640	}
26641	sqlite3_finalize(pStmt);
26642	sqlite3_free(z);
26643	return res;
26644	}
26645
26646	#if SQLITE_SHELL_HAVE_RECOVER
26647	/*
	@@ -25419,11 +26696,11 @@
26696	if( p->db==0 ) return 1;
26697	rc = sqlite3_prepare_v2(p->db,
26698	"SELECT data FROM sqlite_dbpage(?1) WHERE pgno=1",
26699	-1, &pStmt, 0);
26700	if( rc ){
26701	sqlite3_fprintf(stderr,"error: %s\n", sqlite3_errmsg(p->db));
26702	sqlite3_finalize(pStmt);
26703	return 1;
26704	}
26705	sqlite3_bind_text(pStmt, 1, zDb, -1, SQLITE_STATIC);
26706	if( sqlite3_step(pStmt)==SQLITE_ROW
	@@ -25432,58 +26709,149 @@
26709	const u8 *pb = sqlite3_column_blob(pStmt,0);
26710	shell_check_oom(pb);
26711	memcpy(aHdr, pb, 100);
26712	sqlite3_finalize(pStmt);
26713	}else{
26714	sqlite3_fputs("unable to read database header\n", stderr);
26715	sqlite3_finalize(pStmt);
26716	return 1;
26717	}
26718	i = get2byteInt(aHdr+16);
26719	if( i==1 ) i = 65536;
26720	sqlite3_fprintf(p->out, "%-20s %d\n", "database page size:", i);
26721	sqlite3_fprintf(p->out, "%-20s %d\n", "write format:", aHdr[18]);
26722	sqlite3_fprintf(p->out, "%-20s %d\n", "read format:", aHdr[19]);
26723	sqlite3_fprintf(p->out, "%-20s %d\n", "reserved bytes:", aHdr[20]);
26724	for(i=0; i<ArraySize(aField); i++){
26725	int ofst = aField[i].ofst;
26726	unsigned int val = get4byteInt(aHdr + ofst);
26727	sqlite3_fprintf(p->out, "%-20s %u", aField[i].zName, val);
26728	switch( ofst ){
26729	case 56: {
26730	if( val==1 ) sqlite3_fputs(" (utf8)", p->out);
26731	if( val==2 ) sqlite3_fputs(" (utf16le)", p->out);
26732	if( val==3 ) sqlite3_fputs(" (utf16be)", p->out);
26733	}
26734	}
26735	sqlite3_fputs("\n", p->out);
26736	}
26737	if( zDb==0 ){
26738	zSchemaTab = sqlite3_mprintf("main.sqlite_schema");
26739	}else if( cli_strcmp(zDb,"temp")==0 ){
26740	zSchemaTab = sqlite3_mprintf("%s", "sqlite_temp_schema");
26741	}else{
26742	zSchemaTab = sqlite3_mprintf("\"%w\".sqlite_schema", zDb);
26743	}
26744	for(i=0; i<ArraySize(aQuery); i++){
26745	int val = db_int(p->db, aQuery[i].zSql, zSchemaTab);
26746	sqlite3_fprintf(p->out, "%-20s %d\n", aQuery[i].zName, val);


26747	}
26748	sqlite3_free(zSchemaTab);
26749	sqlite3_file_control(p->db, zDb, SQLITE_FCNTL_DATA_VERSION, &iDataVersion);
26750	sqlite3_fprintf(p->out, "%-20s %u\n", "data version", iDataVersion);
26751	return 0;
26752	}
26753	#endif /* SQLITE_SHELL_HAVE_RECOVER */
26754
26755	/*
26756	** Implementation of the ".dbtotxt" command.
26757	**
26758	** Return 1 on error, 2 to exit, and 0 otherwise.
26759	*/
26760	static int shell_dbtotxt_command(ShellState p, int nArg, char *azArg){
26761	sqlite3_stmt *pStmt = 0;
26762	sqlite3_int64 nPage = 0;
26763	int pgSz = 0;
26764	const char *zTail;
26765	char *zName = 0;
26766	int rc, i, j;
26767	unsigned char bShow[256]; /* Characters ok to display */
26768
26769	UNUSED_PARAMETER(nArg);
26770	UNUSED_PARAMETER(azArg);
26771	memset(bShow, '.', sizeof(bShow));
26772	for(i=' '; i<='~'; i++){
26773	if( i!='{' && i!='}' && i!='"' && i!='\\' ) bShow[i] = (unsigned char)i;
26774	}
26775	rc = sqlite3_prepare_v2(p->db, "PRAGMA page_size", -1, &pStmt, 0);
26776	if( rc ) goto dbtotxt_error;
26777	rc = 0;
26778	if( sqlite3_step(pStmt)!=SQLITE_ROW ) goto dbtotxt_error;
26779	pgSz = sqlite3_column_int(pStmt, 0);
26780	sqlite3_finalize(pStmt);
26781	pStmt = 0;
26782	if( pgSz<512 \|\| pgSz>65536 \|\| (pgSz&(pgSz-1))!=0 ) goto dbtotxt_error;
26783	rc = sqlite3_prepare_v2(p->db, "PRAGMA page_count", -1, &pStmt, 0);
26784	if( rc ) goto dbtotxt_error;
26785	rc = 0;
26786	if( sqlite3_step(pStmt)!=SQLITE_ROW ) goto dbtotxt_error;
26787	nPage = sqlite3_column_int64(pStmt, 0);
26788	sqlite3_finalize(pStmt);
26789	pStmt = 0;
26790	if( nPage<1 ) goto dbtotxt_error;
26791	rc = sqlite3_prepare_v2(p->db, "PRAGMA databases", -1, &pStmt, 0);
26792	if( rc ) goto dbtotxt_error;
26793	if( sqlite3_step(pStmt)!=SQLITE_ROW ){
26794	zTail = "unk.db";
26795	}else{
26796	const char zFilename = (const char)sqlite3_column_text(pStmt, 2);
26797	if( zFilename==0 \|\| zFilename[0]==0 ) zFilename = "unk.db";
26798	zTail = strrchr(zFilename, '/');
26799	#if defined(_WIN32)
26800	if( zTail==0 ) zTail = strrchr(zFilename, '\\');
26801	#endif
26802	}
26803	zName = strdup(zTail);
26804	shell_check_oom(zName);
26805	sqlite3_fprintf(p->out, "\| size %lld pagesize %d filename %s\n",
26806	nPage*pgSz, pgSz, zName);
26807	sqlite3_finalize(pStmt);
26808	pStmt = 0;
26809	rc = sqlite3_prepare_v2(p->db,
26810	"SELECT pgno, data FROM sqlite_dbpage ORDER BY pgno", -1, &pStmt, 0);
26811	if( rc ) goto dbtotxt_error;
26812	while( sqlite3_step(pStmt)==SQLITE_ROW ){
26813	sqlite3_int64 pgno = sqlite3_column_int64(pStmt, 0);
26814	const u8 *aData = sqlite3_column_blob(pStmt, 1);
26815	int seenPageLabel = 0;
26816	for(i=0; i<pgSz; i+=16){
26817	const u8 *aLine = aData+i;
26818	for(j=0; j<16 && aLine[j]==0; j++){}
26819	if( j==16 ) continue;
26820	if( !seenPageLabel ){
26821	sqlite3_fprintf(p->out, "\| page %lld offset %lld\n", pgno, pgno*pgSz);
26822	seenPageLabel = 1;
26823	}
26824	sqlite3_fprintf(p->out, "\| %5d:", i);
26825	for(j=0; j<16; j++) sqlite3_fprintf(p->out, " %02x", aLine[j]);
26826	sqlite3_fprintf(p->out, " ");
26827	for(j=0; j<16; j++){
26828	unsigned char c = (unsigned char)aLine[j];
26829	sqlite3_fprintf(p->out, "%c", bShow[c]);
26830	}
26831	sqlite3_fprintf(p->out, "\n");
26832	}
26833	}
26834	sqlite3_finalize(pStmt);
26835	sqlite3_fprintf(p->out, "\| end %s\n", zName);
26836	free(zName);
26837	return 0;
26838
26839	dbtotxt_error:
26840	if( rc ){
26841	sqlite3_fprintf(stderr, "ERROR: %s\n", sqlite3_errmsg(p->db));
26842	}
26843	sqlite3_finalize(pStmt);
26844	free(zName);
26845	return 1;
26846	}
26847
26848	/*
26849	** Print the given string as an error message.
26850	*/
26851	static void shellEmitError(const char *zErr){
26852	sqlite3_fprintf(stderr,"Error: %s\n", zErr);
26853	}
26854	/*
26855	** Print the current sqlite3_errmsg() value to stderr and return 1.
26856	*/
26857	static int shellDatabaseError(sqlite3 *db){
	@@ -25726,10 +27094,11 @@
27094	int bGroupByParent = 0; /* If -groupbyparent is present */
27095	int i; /* To iterate through azArg[] */
27096	const char zIndent = ""; / How much to indent CREATE INDEX by */
27097	int rc; /* Return code */
27098	sqlite3_stmt pSql = 0; / Compiled version of SQL statement below */
27099	FILE out = pState->out; / Send output here */
27100
27101	/*
27102	** This SELECT statement returns one row for each foreign key constraint
27103	** in the schema of the main database. The column values are:
27104	**
	@@ -25801,11 +27170,12 @@
27170	else if( n>1 && sqlite3_strnicmp("-groupbyparent", azArg[i], n)==0 ){
27171	bGroupByParent = 1;
27172	zIndent = " ";
27173	}
27174	else{
27175	sqlite3_fprintf(stderr,
27176	"Usage: %s %s ?-verbose? ?-groupbyparent?\n", azArg[0], azArg[1]);
27177	return SQLITE_ERROR;
27178	}
27179	}
27180
27181	/* Register the fkey_collate_clause() SQL function */
	@@ -25845,44 +27215,45 @@
27215	}
27216	rc = sqlite3_finalize(pExplain);
27217	if( rc!=SQLITE_OK ) break;
27218
27219	if( res<0 ){
27220	sqlite3_fputs("Error: internal error", stderr);
27221	break;
27222	}else{
27223	if( bGroupByParent
27224	&& (bVerbose \|\| res==0)
27225	&& (zPrev==0 \|\| sqlite3_stricmp(zParent, zPrev))
27226	){
27227	sqlite3_fprintf(out, "-- Parent table %s\n", zParent);
27228	sqlite3_free(zPrev);
27229	zPrev = sqlite3_mprintf("%s", zParent);
27230	}
27231
27232	if( res==0 ){
27233	sqlite3_fprintf(out, "%s%s --> %s\n", zIndent, zCI, zTarget);
27234	}else if( bVerbose ){
27235	sqlite3_fprintf(out,
27236	"%s/* no extra indexes required for %s -> %s */\n",
27237	zIndent, zFrom, zTarget
27238	);
27239	}
27240	}
27241	}
27242	sqlite3_free(zPrev);
27243
27244	if( rc!=SQLITE_OK ){
27245	sqlite3_fprintf(stderr,"%s\n", sqlite3_errmsg(db));
27246	}
27247
27248	rc2 = sqlite3_finalize(pSql);
27249	if( rc==SQLITE_OK && rc2!=SQLITE_OK ){
27250	rc = rc2;
27251	sqlite3_fprintf(stderr,"%s\n", sqlite3_errmsg(db));
27252	}
27253	}else{
27254	sqlite3_fprintf(stderr,"%s\n", sqlite3_errmsg(db));
27255	}
27256
27257	return rc;
27258	}
27259
	@@ -25898,13 +27269,13 @@
27269	n = (nArg>=2 ? strlen30(azArg[1]) : 0);
27270	if( n<1 \|\| sqlite3_strnicmp(azArg[1], "fkey-indexes", n) ) goto usage;
27271	return lintFkeyIndexes(pState, azArg, nArg);
27272
27273	usage:
27274	sqlite3_fprintf(stderr,"Usage %s sub-command ?switches...?\n", azArg[0]);
27275	sqlite3_fprintf(stderr, "Where sub-commands are:\n");
27276	sqlite3_fprintf(stderr, " fkey-indexes\n");
27277	return SQLITE_ERROR;
27278	}
27279
27280	static void shellPrepare(
27281	sqlite3 *db,
	@@ -25914,11 +27285,12 @@
27285	){
27286	*ppStmt = 0;
27287	if( *pRc==SQLITE_OK ){
27288	int rc = sqlite3_prepare_v2(db, zSql, -1, ppStmt, 0);
27289	if( rc!=SQLITE_OK ){
27290	sqlite3_fprintf(stderr,
27291	"sql error: %s (%d)\n", sqlite3_errmsg(db), sqlite3_errcode(db));
27292	*pRc = rc;
27293	}
27294	}
27295	}
27296
	@@ -25958,11 +27330,11 @@
27330	if( pStmt ){
27331	sqlite3 *db = sqlite3_db_handle(pStmt);
27332	int rc = sqlite3_finalize(pStmt);
27333	if( *pRc==SQLITE_OK ){
27334	if( rc!=SQLITE_OK ){
27335	sqlite3_fprintf(stderr,"SQL error: %s\n", sqlite3_errmsg(db));
27336	}
27337	*pRc = rc;
27338	}
27339	}
27340	}
	@@ -25980,11 +27352,11 @@
27352	){
27353	int rc = sqlite3_reset(pStmt);
27354	if( *pRc==SQLITE_OK ){
27355	if( rc!=SQLITE_OK ){
27356	sqlite3 *db = sqlite3_db_handle(pStmt);
27357	sqlite3_fprintf(stderr,"SQL error: %s\n", sqlite3_errmsg(db));
27358	}
27359	*pRc = rc;
27360	}
27361	}
27362	#endif /* !defined SQLITE_OMIT_VIRTUALTABLE */
	@@ -26009,10 +27381,11 @@
27381	char zSrcTable; / "sqlar", "zipfile($file)" or "zip" */
27382	const char zFile; / --file argument, or NULL */
27383	const char zDir; / --directory argument, or NULL */
27384	char *azArg; / Array of command arguments */
27385	ShellState p; / Shell state */
27386	FILE out; / Output to this stream */
27387	sqlite3 db; / Database containing the archive */
27388	};
27389
27390	/*
27391	** Print a usage message for the .ar command to stderr and return SQLITE_ERROR.
	@@ -26032,13 +27405,13 @@
27405	va_start(ap, zFmt);
27406	z = sqlite3_vmprintf(zFmt, ap);
27407	va_end(ap);
27408	shellEmitError(z);
27409	if( pAr->fromCmdLine ){
27410	sqlite3_fputs("Use \"-A\" for more help\n", stderr);
27411	}else{
27412	sqlite3_fputs("Use \".archive --help\" for more help\n", stderr);
27413	}
27414	sqlite3_free(z);
27415	return SQLITE_ERROR;
27416	}
27417
	@@ -26134,11 +27507,11 @@
27507	};
27508	int nSwitch = sizeof(aSwitch) / sizeof(struct ArSwitch);
27509	struct ArSwitch *pEnd = &aSwitch[nSwitch];
27510
27511	if( nArg<=1 ){
27512	sqlite3_fprintf(stderr, "Wrong number of arguments. Usage:\n");
27513	return arUsage(stderr);
27514	}else{
27515	char *z = azArg[1];
27516	if( z[0]!='-' ){
27517	/* Traditional style [tar] invocation */
	@@ -26240,11 +27613,11 @@
27613	}
27614	}
27615	}
27616	}
27617	if( pAr->eCmd==0 ){
27618	sqlite3_fprintf(stderr, "Required argument missing. Usage:\n");
27619	return arUsage(stderr);
27620	}
27621	return SQLITE_OK;
27622	}
27623
	@@ -26283,11 +27656,11 @@
27656	if( SQLITE_ROW==sqlite3_step(pTest) ){
27657	bOk = 1;
27658	}
27659	shellReset(&rc, pTest);
27660	if( rc==SQLITE_OK && bOk==0 ){
27661	sqlite3_fprintf(stderr,"not found in archive: %s\n", z);
27662	rc = SQLITE_ERROR;
27663	}
27664	}
27665	shellFinalize(&rc, pTest);
27666	}
	@@ -26350,19 +27723,19 @@
27723	arWhereClause(&rc, pAr, &zWhere);
27724
27725	shellPreparePrintf(pAr->db, &rc, &pSql, zSql, azCols[pAr->bVerbose],
27726	pAr->zSrcTable, zWhere);
27727	if( pAr->bDryRun ){
27728	sqlite3_fprintf(pAr->out, "%s\n", sqlite3_sql(pSql));
27729	}else{
27730	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
27731	if( pAr->bVerbose ){
27732	sqlite3_fprintf(pAr->out, "%s % 10d %s %s\n",
27733	sqlite3_column_text(pSql, 0), sqlite3_column_int(pSql, 1),
27734	sqlite3_column_text(pSql, 2),sqlite3_column_text(pSql, 3));
27735	}else{
27736	sqlite3_fprintf(pAr->out, "%s\n", sqlite3_column_text(pSql, 0));
27737	}
27738	}
27739	}
27740	shellFinalize(&rc, pSql);
27741	sqlite3_free(zWhere);
	@@ -26385,11 +27758,11 @@
27758	}
27759	if( rc==SQLITE_OK ){
27760	zSql = sqlite3_mprintf("DELETE FROM %s WHERE %s;",
27761	pAr->zSrcTable, zWhere);
27762	if( pAr->bDryRun ){
27763	sqlite3_fprintf(pAr->out, "%s\n", zSql);
27764	}else{
27765	char *zErr = 0;
27766	rc = sqlite3_exec(pAr->db, "SAVEPOINT ar;", 0, 0, 0);
27767	if( rc==SQLITE_OK ){
27768	rc = sqlite3_exec(pAr->db, zSql, 0, 0, &zErr);
	@@ -26398,11 +27771,11 @@
27771	}else{
27772	rc = sqlite3_exec(pAr->db, "RELEASE ar;", 0, 0, 0);
27773	}
27774	}
27775	if( zErr ){
27776	sqlite3_fprintf(stdout, "ERROR: %s\n", zErr); /* stdout? */
27777	sqlite3_free(zErr);
27778	}
27779	}
27780	}
27781	sqlite3_free(zWhere);
	@@ -26462,15 +27835,15 @@
27835	** populating them changes the timestamp). */
27836	for(i=0; i<2; i++){
27837	j = sqlite3_bind_parameter_index(pSql, "$dirOnly");
27838	sqlite3_bind_int(pSql, j, i);
27839	if( pAr->bDryRun ){
27840	sqlite3_fprintf(pAr->out, "%s\n", sqlite3_sql(pSql));
27841	}else{
27842	while( rc==SQLITE_OK && SQLITE_ROW==sqlite3_step(pSql) ){
27843	if( i==0 && pAr->bVerbose ){
27844	sqlite3_fprintf(pAr->out, "%s\n", sqlite3_column_text(pSql, 0));
27845	}
27846	}
27847	}
27848	shellReset(&rc, pSql);
27849	}
	@@ -26486,17 +27859,17 @@
27859	** Run the SQL statement in zSql. Or if doing a --dryrun, merely print it out.
27860	*/
27861	static int arExecSql(ArCommand pAr, const char zSql){
27862	int rc;
27863	if( pAr->bDryRun ){
27864	sqlite3_fprintf(pAr->out, "%s\n", zSql);
27865	rc = SQLITE_OK;
27866	}else{
27867	char *zErr = 0;
27868	rc = sqlite3_exec(pAr->db, zSql, 0, 0, &zErr);
27869	if( zErr ){
27870	sqlite3_fprintf(stdout, "ERROR: %s\n", zErr);
27871	sqlite3_free(zErr);
27872	}
27873	}
27874	return rc;
27875	}
	@@ -26641,10 +28014,11 @@
28014	cmd.fromCmdLine = fromCmdLine;
28015	rc = arParseCommand(azArg, nArg, &cmd);
28016	if( rc==SQLITE_OK ){
28017	int eDbType = SHELL_OPEN_UNSPEC;
28018	cmd.p = pState;
28019	cmd.out = pState->out;
28020	cmd.db = pState->db;
28021	if( cmd.zFile ){
28022	eDbType = deduceDatabaseType(cmd.zFile, 1);
28023	}else{
28024	eDbType = pState->openMode;
	@@ -26667,17 +28041,18 @@
28041	}else{
28042	flags = SQLITE_OPEN_READONLY;
28043	}
28044	cmd.db = 0;
28045	if( cmd.bDryRun ){
28046	sqlite3_fprintf(cmd.out, "-- open database '%s'%s\n", cmd.zFile,
28047	eDbType==SHELL_OPEN_APPENDVFS ? " using 'apndvfs'" : "");
28048	}
28049	rc = sqlite3_open_v2(cmd.zFile, &cmd.db, flags,
28050	eDbType==SHELL_OPEN_APPENDVFS ? "apndvfs" : 0);
28051	if( rc!=SQLITE_OK ){
28052	sqlite3_fprintf(stderr, "cannot open file: %s (%s)\n",
28053	cmd.zFile, sqlite3_errmsg(cmd.db));
28054	goto end_ar_command;
28055	}
28056	sqlite3_fileio_init(cmd.db, 0, 0);
28057	sqlite3_sqlar_init(cmd.db, 0, 0);
28058	sqlite3_create_function(cmd.db, "shell_putsnl", 1, SQLITE_UTF8, cmd.p,
	@@ -26686,11 +28061,11 @@
28061	}
28062	if( cmd.zSrcTable==0 && cmd.bZip==0 && cmd.eCmd!=AR_CMD_HELP ){
28063	if( cmd.eCmd!=AR_CMD_CREATE
28064	&& sqlite3_table_column_metadata(cmd.db,0,"sqlar","name",0,0,0,0,0)
28065	){
28066	sqlite3_fprintf(stderr, "database does not contain an 'sqlar' table\n");
28067	rc = SQLITE_ERROR;
28068	goto end_ar_command;
28069	}
28070	cmd.zSrcTable = sqlite3_mprintf("sqlar");
28071	}
	@@ -26744,11 +28119,11 @@
28119	** This function is used as a callback by the recover extension. Simply
28120	** print the supplied SQL statement to stdout.
28121	*/
28122	static int recoverSqlCb(void pCtx, const char zSql){
28123	ShellState pState = (ShellState)pCtx;
28124	sqlite3_fprintf(pState->out, "%s;\n", zSql);
28125	return SQLITE_OK;
28126	}
28127
28128	/*
28129	** This function is called to recover data from the database. A script
	@@ -26787,11 +28162,11 @@
28162	}else
28163	if( n<=10 && memcmp("-no-rowids", z, n)==0 ){
28164	bRowids = 0;
28165	}
28166	else{
28167	sqlite3_fprintf(stderr,"unexpected option: %s\n", azArg[i]);
28168	showHelp(pState->out, azArg[0]);
28169	return 1;
28170	}
28171	}
28172
	@@ -26806,11 +28181,11 @@
28181
28182	sqlite3_recover_run(p);
28183	if( sqlite3_recover_errcode(p)!=SQLITE_OK ){
28184	const char *zErr = sqlite3_recover_errmsg(p);
28185	int errCode = sqlite3_recover_errcode(p);
28186	sqlite3_fprintf(stderr,"sql error: %s (%d)\n", zErr, errCode);
28187	}
28188	rc = sqlite3_recover_finish(p);
28189	return rc;
28190	}
28191	#endif /* SQLITE_SHELL_HAVE_RECOVER */
	@@ -26828,25 +28203,25 @@
28203	i64 nError = 0;
28204	const char *zErr = 0;
28205	while( SQLITE_OK==sqlite3_intck_step(p) ){
28206	const char *zMsg = sqlite3_intck_message(p);
28207	if( zMsg ){
28208	sqlite3_fprintf(pState->out, "%s\n", zMsg);
28209	nError++;
28210	}
28211	nStep++;
28212	if( nStepPerUnlock && (nStep % nStepPerUnlock)==0 ){
28213	sqlite3_intck_unlock(p);
28214	}
28215	}
28216	rc = sqlite3_intck_error(p, &zErr);
28217	if( zErr ){
28218	sqlite3_fprintf(stderr,"%s\n", zErr);
28219	}
28220	sqlite3_intck_close(p);
28221
28222	sqlite3_fprintf(pState->out, "%lld steps, %lld errors\n", nStep, nError);
28223	}
28224
28225	return rc;
28226	}
28227
	@@ -26865,11 +28240,11 @@
28240	*/
28241	#ifdef SHELL_DEBUG
28242	#define rc_err_oom_die(rc) \
28243	if( rc==SQLITE_NOMEM ) shell_check_oom(0); \
28244	else if(!(rc==SQLITE_OK\|\|rc==SQLITE_DONE)) \
28245	sqlite3_fprintf(stderr,"E:%d\n",rc), assert(0)
28246	#else
28247	static void rc_err_oom_die(int rc){
28248	if( rc==SQLITE_NOMEM ) shell_check_oom(0);
28249	assert(rc==SQLITE_OK\|\|rc==SQLITE_DONE);
28250	}
	@@ -27023,12 +28398,12 @@
28398	}else if( *pDb==0 ){
28399	return 0;
28400	}else{
28401	/* Formulate the columns spec, close the DB, zero pDb. /
28402	char *zColsSpec = 0;
28403	int hasDupes = db_int(*pDb, "%s", zHasDupes);
28404	int nDigits = (hasDupes)? db_int(*pDb, "%s", zColDigits) : 0;
28405	if( hasDupes ){
28406	#ifdef SHELL_COLUMN_RENAME_CLEAN
28407	rc = sqlite3_exec(*pDb, zDedoctor, 0, 0, 0);
28408	rc_err_oom_die(rc);
28409	#endif
	@@ -27039,11 +28414,11 @@
28414	sqlite3_bind_int(pStmt, 1, nDigits);
28415	rc = sqlite3_step(pStmt);
28416	sqlite3_finalize(pStmt);
28417	if( rc!=SQLITE_DONE ) rc_err_oom_die(SQLITE_NOMEM);
28418	}
28419	assert(db_int(pDb, "%s", zHasDupes)==0); / Consider: remove this */
28420	rc = sqlite3_prepare_v2(*pDb, zCollectVar, -1, &pStmt, 0);
28421	rc_err_oom_die(rc);
28422	rc = sqlite3_step(pStmt);
28423	if( rc==SQLITE_ROW ){
28424	zColsSpec = sqlite3_mprintf("%s", sqlite3_column_text(pStmt, 0));
	@@ -27082,12 +28457,13 @@
28457	shellPreparePrintf(p->db, &rc, &pStmt,
28458	"SELECT 1 FROM sqlite_schema o WHERE "
28459	"sql LIKE 'CREATE VIRTUAL TABLE%%' AND %s", zLike ? zLike : "true"
28460	);
28461	if( rc==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW ){
28462	sqlite3_fputs("/* WARNING: "
28463	"Script requires that SQLITE_DBCONFIG_DEFENSIVE be disabled */\n",
28464	p->out
28465	);
28466	}
28467	shellFinalize(&rc, pStmt);
28468	return rc;
28469	}
	@@ -27114,16 +28490,18 @@
28490	return SQLITE_OK;
28491	}
28492	if( faultsim_state.iCnt ){
28493	if( faultsim_state.iCnt>0 ) faultsim_state.iCnt--;
28494	if( faultsim_state.eVerbose>=2 ){
28495	sqlite3_fprintf(stdout,
28496	"FAULT-SIM id=%d no-fault (cnt=%d)\n", iArg, faultsim_state.iCnt);
28497	}
28498	return SQLITE_OK;
28499	}
28500	if( faultsim_state.eVerbose>=1 ){
28501	sqlite3_fprintf(stdout,
28502	"FAULT-SIM id=%d returns %d\n", iArg, faultsim_state.iErr);
28503	}
28504	faultsim_state.iCnt = faultsim_state.iInterval;
28505	faultsim_state.nHit++;
28506	if( faultsim_state.nRepeat>0 && faultsim_state.nRepeat<=faultsim_state.nHit ){
28507	faultsim_state.iCnt = -1;
	@@ -27182,11 +28560,11 @@
28560	clearTempFile(p);
28561
28562	#ifndef SQLITE_OMIT_AUTHORIZATION
28563	if( c=='a' && cli_strncmp(azArg[0], "auth", n)==0 ){
28564	if( nArg!=2 ){
28565	sqlite3_fprintf(stderr, "Usage: .auth ON\|OFF\n");
28566	rc = 1;
28567	goto meta_command_exit;
28568	}
28569	open_db(p, 0);
28570	if( booleanValue(azArg[1]) ){
	@@ -27229,32 +28607,32 @@
28607	}else
28608	if( cli_strcmp(z, "-async")==0 ){
28609	bAsync = 1;
28610	}else
28611	{
28612	sqlite3_fprintf(stderr,"unknown option: %s\n", azArg[j]);
28613	return 1;
28614	}
28615	}else if( zDestFile==0 ){
28616	zDestFile = azArg[j];
28617	}else if( zDb==0 ){
28618	zDb = zDestFile;
28619	zDestFile = azArg[j];
28620	}else{
28621	sqlite3_fprintf(stderr, "Usage: .backup ?DB? ?OPTIONS? FILENAME\n");
28622	return 1;
28623	}
28624	}
28625	if( zDestFile==0 ){
28626	sqlite3_fprintf(stderr, "missing FILENAME argument on .backup\n");
28627	return 1;
28628	}
28629	if( zDb==0 ) zDb = "main";
28630	rc = sqlite3_open_v2(zDestFile, &pDest,
28631	SQLITE_OPEN_READWRITE\|SQLITE_OPEN_CREATE, zVfs);
28632	if( rc!=SQLITE_OK ){
28633	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zDestFile);
28634	close_db(pDest);
28635	return 1;
28636	}
28637	if( bAsync ){
28638	sqlite3_exec(pDest, "PRAGMA synchronous=OFF; PRAGMA journal_mode=OFF;",
	@@ -27286,23 +28664,14 @@
28664	eputz("Usage: .bail on\|off\n");
28665	rc = 1;
28666	}
28667	}else
28668
28669	/* Undocumented. Legacy only. See "crlf" below */
28670	if( c=='b' && n>=3 && cli_strncmp(azArg[0], "binary", n)==0 ){
28671	eputz("The \".binary\" command is deprecated.\n");
28672	rc = 1;









28673	}else
28674
28675	/* The undocumented ".breakpoint" command causes a call to the no-op
28676	** routine named test_breakpoint().
28677	*/
	@@ -27320,11 +28689,11 @@
28689	sqlite3_free(z);
28690	#else
28691	rc = chdir(azArg[1]);
28692	#endif
28693	if( rc ){
28694	sqlite3_fprintf(stderr,"Cannot change to directory \"%s\"\n", azArg[1]);
28695	rc = 1;
28696	}
28697	}else{
28698	eputz("Usage: .cd DIRECTORY\n");
28699	rc = 1;
	@@ -27353,15 +28722,16 @@
28722	eputz("Usage: .check GLOB-PATTERN\n");
28723	rc = 2;
28724	}else if( (zRes = readFile("testcase-out.txt", 0))==0 ){
28725	rc = 2;
28726	}else if( testcase_glob(azArg[1],zRes)==0 ){
28727	sqlite3_fprintf(stderr,
28728	"testcase-%s FAILED\n Expected: [%s]\n Got: [%s]\n",
28729	p->zTestcase, azArg[1], zRes);
28730	rc = 1;
28731	}else{
28732	sqlite3_fprintf(p->out, "testcase-%s ok\n", p->zTestcase);
28733	p->nCheck++;
28734	}
28735	sqlite3_free(zRes);
28736	}else
28737	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
	@@ -27390,13 +28760,13 @@
28760	zFile = "(memory)";
28761	}else if( zFile[0]==0 ){
28762	zFile = "(temporary-file)";
28763	}
28764	if( p->pAuxDb == &p->aAuxDb[i] ){
28765	sqlite3_fprintf(stdout, "ACTIVE %d: %s\n", i, zFile);
28766	}else if( p->aAuxDb[i].db!=0 ){
28767	sqlite3_fprintf(stdout, " %d: %s\n", i, zFile);
28768	}
28769	}
28770	}else if( nArg==2 && IsDigit(azArg[1][0]) && azArg[1][1]==0 ){
28771	int i = azArg[1][0] - '0';
28772	if( p->pAuxDb != &p->aAuxDb[i] && i>=0 && i<ArraySize(p->aAuxDb) ){
	@@ -27422,24 +28792,22 @@
28792	eputz("Usage: .connection [close] [CONNECTION-NUMBER]\n");
28793	rc = 1;
28794	}
28795	}else
28796
28797	if( c=='c' && n==4
28798	&& (cli_strncmp(azArg[0], "crlf", n)==0
28799	\|\| cli_strncmp(azArg[0], "crnl",n)==0)
28800	){
28801	if( nArg==2 ){
28802	#ifdef _WIN32
28803	p->crlfMode = booleanValue(azArg[1]);
28804	#else
28805	p->crlfMode = 0;




28806	#endif


28807	}
28808	sqlite3_fprintf(stderr, "crlf is %s\n", p->crlfMode ? "ON" : "OFF");
28809	}else
28810
28811	if( c=='d' && n>1 && cli_strncmp(azArg[0], "databases", n)==0 ){
28812	char **azName = 0;
28813	int nName = 0;
	@@ -27465,11 +28833,11 @@
28833	sqlite3_finalize(pStmt);
28834	for(i=0; i<nName; i++){
28835	int eTxn = sqlite3_txn_state(p->db, azName[i*2]);
28836	int bRdonly = sqlite3_db_readonly(p->db, azName[i*2]);
28837	const char z = azName[i2+1];
28838	sqlite3_fprintf(p->out, "%s: %s %s%s\n",
28839	azName[i*2], z && z[0] ? z : "\"\"", bRdonly ? "r/o" : "r/w",
28840	eTxn==SQLITE_TXN_NONE ? "" :
28841	eTxn==SQLITE_TXN_READ ? " read-txn" : " write-txn");
28842	free(azName[i*2]);
28843	free(azName[i*2+1]);
	@@ -27507,15 +28875,16 @@
28875	if( nArg>1 && cli_strcmp(azArg[1], aDbConfig[ii].zName)!=0 ) continue;
28876	if( nArg>=3 ){
28877	sqlite3_db_config(p->db, aDbConfig[ii].op, booleanValue(azArg[2]), 0);
28878	}
28879	sqlite3_db_config(p->db, aDbConfig[ii].op, -1, &v);
28880	sqlite3_fprintf(p->out, "%19s %s\n",
28881	aDbConfig[ii].zName, v ? "on" : "off");
28882	if( nArg>1 ) break;
28883	}
28884	if( nArg>1 && ii==ArraySize(aDbConfig) ){
28885	sqlite3_fprintf(stderr,"Error: unknown dbconfig \"%s\"\n", azArg[1]);
28886	eputz("Enter \".dbconfig\" with no arguments for a list\n");
28887	}
28888	}else
28889
28890	#if SQLITE_SHELL_HAVE_RECOVER
	@@ -27561,11 +28930,12 @@
28930	}else
28931	if( cli_strcmp(z,"nosys")==0 ){
28932	ShellSetFlag(p, SHFLG_DumpNoSys);
28933	}else
28934	{
28935	sqlite3_fprintf(stderr,
28936	"Unknown option \"%s\" on \".dump\"\n", azArg[i]);
28937	rc = 1;
28938	sqlite3_free(zLike);
28939	goto meta_command_exit;
28940	}
28941	}else{
	@@ -27596,12 +28966,12 @@
28966	outputDumpWarning(p, zLike);
28967	if( (p->shellFlgs & SHFLG_DumpDataOnly)==0 ){
28968	/* When playing back a "dump", the content might appear in an order
28969	** which causes immediate foreign key constraints to be violated.
28970	** So disable foreign-key constraint enforcement to prevent problems. */
28971	sqlite3_fputs("PRAGMA foreign_keys=OFF;\n", p->out);
28972	sqlite3_fputs("BEGIN TRANSACTION;\n", p->out);
28973	}
28974	p->writableSchema = 0;
28975	p->showHeader = 0;
28976	/* Set writable_schema=ON since doing so forces SQLite to initialize
28977	** as much of the schema as it can even if the sqlite_schema table is
	@@ -27629,17 +28999,17 @@
28999	run_table_dump_query(p, zSql);
29000	sqlite3_free(zSql);
29001	}
29002	sqlite3_free(zLike);
29003	if( p->writableSchema ){
29004	sqlite3_fputs("PRAGMA writable_schema=OFF;\n", p->out);
29005	p->writableSchema = 0;
29006	}
29007	sqlite3_exec(p->db, "PRAGMA writable_schema=OFF;", 0, 0, 0);
29008	sqlite3_exec(p->db, "RELEASE dump;", 0, 0, 0);
29009	if( (p->shellFlgs & SHFLG_DumpDataOnly)==0 ){
29010	sqlite3_fputs(p->nErr?"ROLLBACK; -- due to errors\n":"COMMIT;\n", p->out);
29011	}
29012	p->showHeader = savedShowHeader;
29013	p->shellFlgs = savedShellFlags;
29014	}else
29015
	@@ -27649,10 +29019,14 @@
29019	}else{
29020	eputz("Usage: .echo on\|off\n");
29021	rc = 1;
29022	}
29023	}else
29024
29025	if( c=='d' && n>=3 && cli_strncmp(azArg[0], "dbtotxt", n)==0 ){
29026	rc = shell_dbtotxt_command(p, nArg, azArg);
29027	}else
29028
29029	if( c=='e' && cli_strncmp(azArg[0], "eqp", n)==0 ){
29030	if( nArg==2 ){
29031	p->autoEQPtest = 0;
29032	if( p->autoEQPtrace ){
	@@ -27715,11 +29089,12 @@
29089	}else
29090
29091	#ifndef SQLITE_OMIT_VIRTUALTABLE
29092	if( c=='e' && cli_strncmp(azArg[0], "expert", n)==0 ){
29093	if( p->bSafeMode ){
29094	sqlite3_fprintf(stderr,
29095	"Cannot run experimental commands such as \"%s\" in safe mode\n",
29096	azArg[0]);
29097	rc = 1;
29098	}else{
29099	open_db(p, 0);
29100	expertDotCommand(p, azArg, nArg);
	@@ -27772,13 +29147,14 @@
29147	if( zCmd[0]=='-' && zCmd[1] ) zCmd++;
29148	}
29149
29150	/* --help lists all file-controls */
29151	if( cli_strcmp(zCmd,"help")==0 ){
29152	sqlite3_fputs("Available file-controls:\n", p->out);
29153	for(i=0; i<ArraySize(aCtrl); i++){
29154	sqlite3_fprintf(p->out,
29155	" .filectrl %s %s\n", aCtrl[i].zCtrlName, aCtrl[i].zUsage);
29156	}
29157	rc = 1;
29158	goto meta_command_exit;
29159	}
29160
	@@ -27789,19 +29165,19 @@
29165	if( cli_strncmp(zCmd, aCtrl[i].zCtrlName, n2)==0 ){
29166	if( filectrl<0 ){
29167	filectrl = aCtrl[i].ctrlCode;
29168	iCtrl = i;
29169	}else{
29170	sqlite3_fprintf(stderr,"Error: ambiguous file-control: \"%s\"\n"
29171	"Use \".filectrl --help\" for help\n", zCmd);
29172	rc = 1;
29173	goto meta_command_exit;
29174	}
29175	}
29176	}
29177	if( filectrl<0 ){
29178	sqlite3_fprintf(stderr,"Error: unknown file-control: %s\n"
29179	"Use \".filectrl --help\" for help\n", zCmd);
29180	}else{
29181	switch(filectrl){
29182	case SQLITE_FCNTL_SIZE_LIMIT: {
29183	if( nArg!=2 && nArg!=3 ) break;
	@@ -27841,11 +29217,11 @@
29217	case SQLITE_FCNTL_TEMPFILENAME: {
29218	char *z = 0;
29219	if( nArg!=2 ) break;
29220	sqlite3_file_control(p->db, zSchema, filectrl, &z);
29221	if( z ){
29222	sqlite3_fprintf(p->out, "%s\n", z);
29223	sqlite3_free(z);
29224	}
29225	isOk = 2;
29226	break;
29227	}
	@@ -27855,23 +29231,24 @@
29231	x = atoi(azArg[2]);
29232	sqlite3_file_control(p->db, zSchema, filectrl, &x);
29233	}
29234	x = -1;
29235	sqlite3_file_control(p->db, zSchema, filectrl, &x);
29236	sqlite3_fprintf(p->out, "%d\n", x);
29237	isOk = 2;
29238	break;
29239	}
29240	}
29241	}
29242	if( isOk==0 && iCtrl>=0 ){
29243	sqlite3_fprintf(p->out, "Usage: .filectrl %s %s\n",
29244	zCmd, aCtrl[iCtrl].zUsage);
29245	rc = 1;
29246	}else if( isOk==1 ){
29247	char zBuf[100];
29248	sqlite3_snprintf(sizeof(zBuf), zBuf, "%lld", iRes);
29249	sqlite3_fprintf(p->out, "%s\n", zBuf);
29250	}
29251	}else
29252
29253	if( c=='f' && cli_strncmp(azArg[0], "fullschema", n)==0 ){
29254	ShellState data;
	@@ -27908,19 +29285,19 @@
29285	doStats = sqlite3_step(pStmt)==SQLITE_ROW;
29286	sqlite3_finalize(pStmt);
29287	}
29288	}
29289	if( doStats==0 ){
29290	sqlite3_fputs("/* No STAT tables available */\n", p->out);
29291	}else{
29292	sqlite3_fputs("ANALYZE sqlite_schema;\n", p->out);
29293	data.cMode = data.mode = MODE_Insert;
29294	data.zDestTable = "sqlite_stat1";
29295	shell_exec(&data, "SELECT * FROM sqlite_stat1", 0);
29296	data.zDestTable = "sqlite_stat4";
29297	shell_exec(&data, "SELECT * FROM sqlite_stat4", 0);
29298	sqlite3_fputs("ANALYZE sqlite_schema;\n", p->out);
29299	}
29300	}else
29301
29302	if( c=='h' && cli_strncmp(azArg[0], "headers", n)==0 ){
29303	if( nArg==2 ){
	@@ -27934,11 +29311,11 @@
29311
29312	if( c=='h' && cli_strncmp(azArg[0], "help", n)==0 ){
29313	if( nArg>=2 ){
29314	n = showHelp(p->out, azArg[1]);
29315	if( n==0 ){
29316	sqlite3_fprintf(p->out, "Nothing matches '%s'\n", azArg[1]);
29317	}
29318	}else{
29319	showHelp(p->out, 0);
29320	}
29321	}else
	@@ -27977,11 +29354,11 @@
29354	if( zFile==0 ){
29355	zFile = z;
29356	}else if( zTable==0 ){
29357	zTable = z;
29358	}else{
29359	sqlite3_fprintf(p->out, "ERROR: extra argument: \"%s\". Usage:\n",z);
29360	showHelp(p->out, "import");
29361	goto meta_command_exit;
29362	}
29363	}else if( cli_strcmp(z,"-v")==0 ){
29364	eVerbose++;
	@@ -27998,17 +29375,17 @@
29375	sCtx.cColSep = ',';
29376	sCtx.cRowSep = '\n';
29377	xRead = csv_read_one_field;
29378	useOutputMode = 0;
29379	}else{
29380	sqlite3_fprintf(p->out, "ERROR: unknown option: \"%s\". Usage:\n", z);
29381	showHelp(p->out, "import");
29382	goto meta_command_exit;
29383	}
29384	}
29385	if( zTable==0 ){
29386	sqlite3_fprintf(p->out, "ERROR: missing %s argument. Usage:\n",
29387	zFile==0 ? "FILE" : "TABLE");
29388	showHelp(p->out, "import");
29389	goto meta_command_exit;
29390	}
29391	seenInterrupt = 0;
	@@ -28054,32 +29431,32 @@
29431	if( sCtx.zFile[0]=='\|' ){
29432	#ifdef SQLITE_OMIT_POPEN
29433	eputz("Error: pipes are not supported in this OS\n");
29434	goto meta_command_exit;
29435	#else
29436	sCtx.in = sqlite3_popen(sCtx.zFile+1, "r");
29437	sCtx.zFile = "<pipe>";
29438	sCtx.xCloser = pclose;
29439	#endif
29440	}else{
29441	sCtx.in = sqlite3_fopen(sCtx.zFile, "rb");
29442	sCtx.xCloser = fclose;
29443	}
29444	if( sCtx.in==0 ){
29445	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zFile);
29446	goto meta_command_exit;
29447	}
29448	if( eVerbose>=2 \|\| (eVerbose>=1 && useOutputMode) ){
29449	char zSep[2];
29450	zSep[1] = 0;
29451	zSep[0] = sCtx.cColSep;
29452	sqlite3_fputs("Column separator ", p->out);
29453	output_c_string(p->out, zSep);
29454	sqlite3_fputs(", row separator ", p->out);
29455	zSep[0] = sCtx.cRowSep;
29456	output_c_string(p->out, zSep);
29457	sqlite3_fputs("\n", p->out);
29458	}
29459	sCtx.z = sqlite3_malloc64(120);
29460	if( sCtx.z==0 ){
29461	import_cleanup(&sCtx);
29462	shell_out_of_memory();
	@@ -28087,11 +29464,15 @@
29464	/* Below, resources must be freed before exit. */
29465	while( (nSkip--)>0 ){
29466	while( xRead(&sCtx) && sCtx.cTerm==sCtx.cColSep ){}
29467	}
29468	import_append_char(&sCtx, 0); /* To ensure sCtx.z is allocated */
29469	if( sqlite3_table_column_metadata(p->db, zSchema, zTable,0,0,0,0,0,0)
29470	&& 0==db_int(p->db, "SELECT count(*) FROM \"%w\".sqlite_schema"
29471	" WHERE name=%Q AND type='view'",
29472	zSchema ? zSchema : "main", zTable)
29473	){
29474	/* Table does not exist. Create it. */
29475	sqlite3 *dbCols = 0;
29476	char *zRenames = 0;
29477	char *zColDefs;
29478	zCreate = sqlite3_mprintf("CREATE TABLE \"%w\".\"%w\"",
	@@ -28100,18 +29481,18 @@
29481	zAutoColumn(sCtx.z, &dbCols, 0);
29482	if( sCtx.cTerm!=sCtx.cColSep ) break;
29483	}
29484	zColDefs = zAutoColumn(0, &dbCols, &zRenames);
29485	if( zRenames!=0 ){
29486	sqlite3_fprintf((stdin_is_interactive && p->in==stdin)? p->out : stderr,
29487	"Columns renamed during .import %s due to duplicates:\n"
29488	"%s\n", sCtx.zFile, zRenames);
29489	sqlite3_free(zRenames);
29490	}
29491	assert(dbCols==0);
29492	if( zColDefs==0 ){
29493	sqlite3_fprintf(stderr,"%s: empty file\n", sCtx.zFile);
29494	import_cleanup(&sCtx);
29495	rc = 1;
29496	sqlite3_free(zCreate);
29497	goto meta_command_exit;
29498	}
	@@ -28119,17 +29500,20 @@
29500	if( zCreate==0 ){
29501	import_cleanup(&sCtx);
29502	shell_out_of_memory();
29503	}
29504	if( eVerbose>=1 ){
29505	sqlite3_fprintf(p->out, "%s\n", zCreate);
29506	}
29507	rc = sqlite3_exec(p->db, zCreate, 0, 0, 0);
29508	if( rc ){
29509	sqlite3_fprintf(stderr,
29510	"%s failed:\n%s\n", zCreate, sqlite3_errmsg(p->db));
29511	}
29512	sqlite3_free(zCreate);
29513	zCreate = 0;
29514	if( rc ){

29515	import_cleanup(&sCtx);
29516	rc = 1;
29517	goto meta_command_exit;
29518	}
29519	}
	@@ -28180,11 +29564,11 @@
29564	}
29565	zSql[j++] = ')';
29566	zSql[j] = 0;
29567	assert( j<nByte );
29568	if( eVerbose>=2 ){
29569	sqlite3_fprintf(p->out, "Insert using: %s\n", zSql);
29570	}
29571	rc = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0);
29572	sqlite3_free(zSql);
29573	zSql = 0;
29574	if( rc ){
	@@ -28219,11 +29603,11 @@
29603	if( z==0 && (xRead==csv_read_one_field) && i==nCol-1 && i>0 ){
29604	z = "";
29605	}
29606	sqlite3_bind_text(pStmt, i+1, z, -1, SQLITE_TRANSIENT);
29607	if( i<nCol-1 && sCtx.cTerm!=sCtx.cColSep ){
29608	sqlite3_fprintf(stderr,"%s:%d: expected %d columns but found %d"
29609	" - filling the rest with NULL\n",
29610	sCtx.zFile, startLine, nCol, i+1);
29611	i += 2;
29612	while( i<=nCol ){ sqlite3_bind_null(pStmt, i); i++; }
29613	}
	@@ -28231,18 +29615,19 @@
29615	if( sCtx.cTerm==sCtx.cColSep ){
29616	do{
29617	xRead(&sCtx);
29618	i++;
29619	}while( sCtx.cTerm==sCtx.cColSep );
29620	sqlite3_fprintf(stderr,
29621	"%s:%d: expected %d columns but found %d - extras ignored\n",
29622	sCtx.zFile, startLine, nCol, i);
29623	}
29624	if( i>=nCol ){
29625	sqlite3_step(pStmt);
29626	rc = sqlite3_reset(pStmt);
29627	if( rc!=SQLITE_OK ){
29628	sqlite3_fprintf(stderr,"%s:%d: INSERT failed: %s\n",
29629	sCtx.zFile, startLine, sqlite3_errmsg(p->db));
29630	sCtx.nErr++;
29631	}else{
29632	sCtx.nRow++;
29633	}
	@@ -28251,11 +29636,12 @@
29636
29637	import_cleanup(&sCtx);
29638	sqlite3_finalize(pStmt);
29639	if( needCommit ) sqlite3_exec(p->db, "COMMIT", 0, 0, 0);
29640	if( eVerbose>0 ){
29641	sqlite3_fprintf(p->out,
29642	"Added %d rows with %d errors using %d lines of input\n",
29643	sCtx.nRow, sCtx.nErr, sCtx.nLine-1);
29644	}
29645	}else
29646	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
29647
	@@ -28267,11 +29653,11 @@
29653	int tnum = 0;
29654	int isWO = 0; /* True if making an imposter of a WITHOUT ROWID table */
29655	int lenPK = 0; /* Length of the PRIMARY KEY string for isWO tables */
29656	int i;
29657	if( !ShellHasFlag(p,SHFLG_TestingMode) ){
29658	sqlite3_fprintf(stderr,".%s unavailable without --unsafe-testing\n",
29659	"imposter");
29660	rc = 1;
29661	goto meta_command_exit;
29662	}
29663	if( !(nArg==3 \|\| (nArg==2 && sqlite3_stricmp(azArg[1],"off")==0)) ){
	@@ -28333,11 +29719,11 @@
29719	zCollist = sqlite3_mprintf("%z,\"%w\"", zCollist, zCol);
29720	}
29721	}
29722	sqlite3_finalize(pStmt);
29723	if( i==0 \|\| tnum==0 ){
29724	sqlite3_fprintf(stderr,"no such index: \"%s\"\n", azArg[1]);
29725	rc = 1;
29726	sqlite3_free(zCollist);
29727	goto meta_command_exit;
29728	}
29729	if( lenPK==0 ) lenPK = 100000;
	@@ -28348,18 +29734,20 @@
29734	rc = sqlite3_test_control(SQLITE_TESTCTRL_IMPOSTER, p->db, "main", 1, tnum);
29735	if( rc==SQLITE_OK ){
29736	rc = sqlite3_exec(p->db, zSql, 0, 0, 0);
29737	sqlite3_test_control(SQLITE_TESTCTRL_IMPOSTER, p->db, "main", 0, 0);
29738	if( rc ){
29739	sqlite3_fprintf(stderr,
29740	"Error in [%s]: %s\n", zSql, sqlite3_errmsg(p->db));
29741	}else{
29742	sqlite3_fprintf(stdout, "%s;\n", zSql);
29743	sqlite3_fprintf(stdout,
29744	"WARNING: writing to an imposter table will corrupt"
29745	" the \"%s\" %s!\n", azArg[1], isWO ? "table" : "index");
29746	}
29747	}else{
29748	sqlite3_fprintf(stderr,"SQLITE_TESTCTRL_IMPOSTER returns %d\n", rc);
29749	rc = 1;
29750	}
29751	sqlite3_free(zSql);
29752	}else
29753	#endif /* !defined(SQLITE_OMIT_TEST_CONTROL) */
	@@ -28369,11 +29757,11 @@
29757	if( nArg==2 ){
29758	iArg = integerValue(azArg[1]);
29759	if( iArg==0 ) iArg = -1;
29760	}
29761	if( (nArg!=1 && nArg!=2) \|\| iArg<0 ){
29762	sqlite3_fprintf(stderr,"%s","Usage: .intck STEPS_PER_UNLOCK\n");
29763	rc = 1;
29764	goto meta_command_exit;
29765	}
29766	open_db(p, 0);
29767	rc = intckDatabaseCmd(p, iArg);
	@@ -28388,13 +29776,13 @@
29776	sqlite3IoTrace = 0;
29777	}else if( cli_strcmp(azArg[1], "-")==0 ){
29778	sqlite3IoTrace = iotracePrintf;
29779	iotrace = stdout;
29780	}else{
29781	iotrace = sqlite3_fopen(azArg[1], "w");
29782	if( iotrace==0 ){
29783	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", azArg[1]);
29784	sqlite3IoTrace = 0;
29785	rc = 1;
29786	}else{
29787	sqlite3IoTrace = iotracePrintf;
29788	}
	@@ -28422,11 +29810,11 @@
29810	};
29811	int i, n2;
29812	open_db(p, 0);
29813	if( nArg==1 ){
29814	for(i=0; i<ArraySize(aLimit); i++){
29815	sqlite3_fprintf(stdout, "%20s %d\n", aLimit[i].zLimitName,
29816	sqlite3_limit(p->db, aLimit[i].limitCode, -1));
29817	}
29818	}else if( nArg>3 ){
29819	eputz("Usage: .limit NAME ?NEW-VALUE?\n");
29820	rc = 1;
	@@ -28437,28 +29825,28 @@
29825	for(i=0; i<ArraySize(aLimit); i++){
29826	if( sqlite3_strnicmp(aLimit[i].zLimitName, azArg[1], n2)==0 ){
29827	if( iLimit<0 ){
29828	iLimit = i;
29829	}else{
29830	sqlite3_fprintf(stderr,"ambiguous limit: \"%s\"\n", azArg[1]);
29831	rc = 1;
29832	goto meta_command_exit;
29833	}
29834	}
29835	}
29836	if( iLimit<0 ){
29837	sqlite3_fprintf(stderr,"unknown limit: \"%s\"\n"
29838	"enter \".limits\" with no arguments for a list.\n",
29839	azArg[1]);
29840	rc = 1;
29841	goto meta_command_exit;
29842	}
29843	if( nArg==3 ){
29844	sqlite3_limit(p->db, aLimit[iLimit].limitCode,
29845	(int)integerValue(azArg[2]));
29846	}
29847	sqlite3_fprintf(stdout, "%20s %d\n", aLimit[iLimit].zLimitName,
29848	sqlite3_limit(p->db, aLimit[iLimit].limitCode, -1));
29849	}
29850	}else
29851
29852	if( c=='l' && n>2 && cli_strncmp(azArg[0], "lint", n)==0 ){
	@@ -28503,11 +29891,11 @@
29891	" than \"on\" or \"off\"\n");
29892	zFile = "off";
29893	}
29894	output_file_close(p->pLog);
29895	if( cli_strcmp(zFile,"on")==0 ) zFile = "stdout";
29896	p->pLog = output_file_open(zFile);
29897	}
29898	}else
29899
29900	if( c=='m' && cli_strncmp(azArg[0], "mode", n)==0 ){
29901	const char *zMode = 0;
	@@ -28537,35 +29925,37 @@
29925	cmOpts = cmo;
29926	}
29927	}else if( zTabname==0 ){
29928	zTabname = z;
29929	}else if( z[0]=='-' ){
29930	sqlite3_fprintf(stderr,"unknown option: %s\n", z);
29931	eputz("options:\n"
29932	" --noquote\n"
29933	" --quote\n"
29934	" --wordwrap on/off\n"
29935	" --wrap N\n"
29936	" --ww\n");
29937	rc = 1;
29938	goto meta_command_exit;
29939	}else{
29940	sqlite3_fprintf(stderr,"extra argument: \"%s\"\n", z);
29941	rc = 1;
29942	goto meta_command_exit;
29943	}
29944	}
29945	if( zMode==0 ){
29946	if( p->mode==MODE_Column
29947	\|\| (p->mode>=MODE_Markdown && p->mode<=MODE_Box)
29948	){
29949	sqlite3_fprintf(p->out,
29950	"current output mode: %s --wrap %d --wordwrap %s --%squote\n",
29951	modeDescr[p->mode], p->cmOpts.iWrap,
29952	p->cmOpts.bWordWrap ? "on" : "off",
29953	p->cmOpts.bQuote ? "" : "no");
29954	}else{
29955	sqlite3_fprintf(p->out,
29956	"current output mode: %s\n", modeDescr[p->mode]);
29957	}
29958	zMode = modeDescr[p->mode];
29959	}
29960	n2 = strlen30(zMode);
29961	if( cli_strncmp(zMode,"lines",n2)==0 ){
	@@ -28634,11 +30024,11 @@
30024	if( c=='n' && cli_strcmp(azArg[0], "nonce")==0 ){
30025	if( nArg!=2 ){
30026	eputz("Usage: .nonce NONCE\n");
30027	rc = 1;
30028	}else if( p->zNonce==0 \|\| cli_strcmp(azArg[1],p->zNonce)!=0 ){
30029	sqlite3_fprintf(stderr,"line %d: incorrect nonce: \"%s\"\n",
30030	p->lineno, azArg[1]);
30031	exit(1);
30032	}else{
30033	p->bSafeMode = 0;
30034	return 0; /* Return immediately to bypass the safe mode reset
	@@ -28689,15 +30079,15 @@
30079	p->szMax = integerValue(azArg[++iName]);
30080	#endif /* SQLITE_OMIT_DESERIALIZE */
30081	}else
30082	#endif /* !SQLITE_SHELL_FIDDLE */
30083	if( z[0]=='-' ){
30084	sqlite3_fprintf(stderr,"unknown option: %s\n", z);
30085	rc = 1;
30086	goto meta_command_exit;
30087	}else if( zFN ){
30088	sqlite3_fprintf(stderr,"extra argument: \"%s\"\n", z);
30089	rc = 1;
30090	goto meta_command_exit;
30091	}else{
30092	zFN = z;
30093	}
	@@ -28735,11 +30125,11 @@
30125	zNewFilename = 0;
30126	}
30127	p->pAuxDb->zDbFilename = zNewFilename;
30128	open_db(p, OPEN_DB_KEEPALIVE);
30129	if( p->db==0 ){
30130	sqlite3_fprintf(stderr,"Error: cannot open '%s'\n", zNewFilename);
30131	sqlite3_free(zNewFilename);
30132	}else{
30133	p->pAuxDb->zFreeOnClose = zNewFilename;
30134	}
30135	}
	@@ -28753,22 +30143,26 @@
30143	#ifndef SQLITE_SHELL_FIDDLE
30144	if( (c=='o'
30145	&& (cli_strncmp(azArg[0], "output", n)==0
30146	\|\| cli_strncmp(azArg[0], "once", n)==0))
30147	\|\| (c=='e' && n==5 && cli_strcmp(azArg[0],"excel")==0)
30148	\|\| (c=='w' && n==3 && cli_strcmp(azArg[0],"www")==0)
30149	){
30150	char *zFile = 0;

30151	int i;
30152	int eMode = 0;
30153	int bOnce = 0; /* 0: .output, 1: .once, 2: .excel/.www */
30154	int bPlain = 0; /* --plain option */
30155	static const char *zBomUtf8 = "\357\273\277";
30156	const char *zBom = 0;
30157
30158	failIfSafeMode(p, "cannot run .%s in safe mode", azArg[0]);
30159	if( c=='e' ){
30160	eMode = 'x';
30161	bOnce = 2;
30162	}else if( c=='w' ){
30163	eMode = 'w';
30164	bOnce = 2;
30165	}else if( cli_strncmp(azArg[0],"once",n)==0 ){
30166	bOnce = 1;
30167	}
30168	for(i=1; i<nArg; i++){
	@@ -28775,28 +30169,34 @@
30169	char *z = azArg[i];
30170	if( z[0]=='-' ){
30171	if( z[1]=='-' ) z++;
30172	if( cli_strcmp(z,"-bom")==0 ){
30173	zBom = zBomUtf8;
30174	}else if( cli_strcmp(z,"-plain")==0 ){
30175	bPlain = 1;
30176	}else if( c=='o' && cli_strcmp(z,"-x")==0 ){
30177	eMode = 'x'; /* spreadsheet */
30178	}else if( c=='o' && cli_strcmp(z,"-e")==0 ){
30179	eMode = 'e'; /* text editor */
30180	}else if( c=='o' && cli_strcmp(z,"-w")==0 ){
30181	eMode = 'w'; /* Web browser */
30182	}else{
30183	sqlite3_fprintf(p->out,
30184	"ERROR: unknown option: \"%s\". Usage:\n", azArg[i]);
30185	showHelp(p->out, azArg[0]);
30186	rc = 1;
30187	goto meta_command_exit;
30188	}
30189	}else if( zFile==0 && eMode==0 ){
30190	zFile = sqlite3_mprintf("%s", z);
30191	if( zFile && zFile[0]=='\|' ){
30192	while( i+1<nArg ) zFile = sqlite3_mprintf("%z %s", zFile, azArg[++i]);
30193	break;
30194	}
30195	}else{
30196	sqlite3_fprintf(p->out,
30197	"ERROR: extra parameter: \"%s\". Usage:\n", azArg[i]);
30198	showHelp(p->out, azArg[0]);
30199	rc = 1;
30200	sqlite3_free(zFile);
30201	goto meta_command_exit;
30202	}
	@@ -28809,24 +30209,31 @@
30209	}else{
30210	p->outCount = 0;
30211	}
30212	output_reset(p);
30213	#ifndef SQLITE_NOHAVE_SYSTEM
30214	if( eMode=='e' \|\| eMode=='x' \|\| eMode=='w' ){
30215	p->doXdgOpen = 1;
30216	outputModePush(p);
30217	if( eMode=='x' ){
30218	/* spreadsheet mode. Output as CSV. */
30219	newTempFile(p, "csv");
30220	ShellClearFlag(p, SHFLG_Echo);
30221	p->mode = MODE_Csv;
30222	sqlite3_snprintf(sizeof(p->colSeparator), p->colSeparator, SEP_Comma);
30223	sqlite3_snprintf(sizeof(p->rowSeparator), p->rowSeparator, SEP_CrLf);
30224	#ifdef _WIN32
30225	zBom = zBomUtf8; /* Always include the BOM on Windows, as Excel does
30226	** not work without it. */
30227	#endif
30228	}else if( eMode=='w' ){
30229	/* web-browser mode. */
30230	newTempFile(p, "html");
30231	if( !bPlain ) p->mode = MODE_Www;
30232	}else{
30233	/* text editor mode */
30234	newTempFile(p, "txt");

30235	}
30236	sqlite3_free(zFile);
30237	zFile = sqlite3_mprintf("%s", p->zTempFile);
30238	}
30239	#endif /* SQLITE_NOHAVE_SYSTEM */
	@@ -28835,30 +30242,36 @@
30242	#ifdef SQLITE_OMIT_POPEN
30243	eputz("Error: pipes are not supported in this OS\n");
30244	rc = 1;
30245	output_redir(p, stdout);
30246	#else
30247	FILE *pfPipe = sqlite3_popen(zFile + 1, "w");
30248	if( pfPipe==0 ){
30249	sqlite3_fprintf(stderr,"Error: cannot open pipe \"%s\"\n", zFile + 1);
30250	rc = 1;
30251	}else{
30252	output_redir(p, pfPipe);
30253	if( zBom ) sqlite3_fputs(zBom, pfPipe);
30254	sqlite3_snprintf(sizeof(p->outfile), p->outfile, "%s", zFile);
30255	}
30256	#endif
30257	}else{
30258	FILE *pfFile = output_file_open(zFile);
30259	if( pfFile==0 ){
30260	if( cli_strcmp(zFile,"off")!=0 ){
30261	sqlite3_fprintf(stderr,"Error: cannot write to \"%s\"\n", zFile);
30262	}
30263	rc = 1;
30264	} else {
30265	output_redir(p, pfFile);
30266	if( zBom ) sqlite3_fputs(zBom, pfFile);
30267	if( bPlain && eMode=='w' ){
30268	sqlite3_fputs(
30269	"<!DOCTYPE html>\n<BODY>\n<PLAINTEXT>\n",
30270	pfFile
30271	);
30272	}
30273	sqlite3_snprintf(sizeof(p->outfile), p->outfile, "%s", zFile);
30274	}
30275	}
30276	sqlite3_free(zFile);
30277	}else
	@@ -28895,11 +30308,12 @@
30308	if( len ){
30309	rx = sqlite3_prepare_v2(p->db,
30310	"SELECT key, quote(value) "
30311	"FROM temp.sqlite_parameters;", -1, &pStmt, 0);
30312	while( rx==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW ){
30313	sqlite3_fprintf(p->out,
30314	"%-*s %s\n", len, sqlite3_column_text(pStmt,0),
30315	sqlite3_column_text(pStmt,1));
30316	}
30317	sqlite3_finalize(pStmt);
30318	}
30319	}else
	@@ -28940,11 +30354,11 @@
30354	"VALUES(%Q,%Q);", zKey, zValue);
30355	shell_check_oom(zSql);
30356	rx = sqlite3_prepare_v2(p->db, zSql, -1, &pStmt, 0);
30357	sqlite3_free(zSql);
30358	if( rx!=SQLITE_OK ){
30359	sqlite3_fprintf(p->out, "Error: %s\n", sqlite3_errmsg(p->db));
30360	sqlite3_finalize(pStmt);
30361	pStmt = 0;
30362	rc = 1;
30363	}
30364	}
	@@ -28969,14 +30383,14 @@
30383	}else
30384
30385	if( c=='p' && n>=3 && cli_strncmp(azArg[0], "print", n)==0 ){
30386	int i;
30387	for(i=1; i<nArg; i++){
30388	if( i>1 ) sqlite3_fputs(" ", p->out);
30389	sqlite3_fputs(azArg[i], p->out);
30390	}
30391	sqlite3_fputs("\n", p->out);
30392	}else
30393
30394	#ifndef SQLITE_OMIT_PROGRESS_CALLBACK
30395	if( c=='p' && n>=3 && cli_strncmp(azArg[0], "progress", n)==0 ){
30396	int i;
	@@ -29009,11 +30423,11 @@
30423	}else{
30424	p->mxProgress = (int)integerValue(azArg[++i]);
30425	}
30426	continue;
30427	}
30428	sqlite3_fprintf(stderr,"Error: unknown option: \"%s\"\n", azArg[i]);
30429	rc = 1;
30430	goto meta_command_exit;
30431	}else{
30432	nn = (int)integerValue(z);
30433	}
	@@ -29050,23 +30464,22 @@
30464	}
30465	if( azArg[1][0]=='\|' ){
30466	#ifdef SQLITE_OMIT_POPEN
30467	eputz("Error: pipes are not supported in this OS\n");
30468	rc = 1;

30469	#else
30470	p->in = sqlite3_popen(azArg[1]+1, "r");
30471	if( p->in==0 ){
30472	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", azArg[1]);
30473	rc = 1;
30474	}else{
30475	rc = process_input(p);
30476	pclose(p->in);
30477	}
30478	#endif
30479	}else if( (p->in = openChrSource(azArg[1]))==0 ){
30480	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", azArg[1]);
30481	rc = 1;
30482	}else{
30483	rc = process_input(p);
30484	fclose(p->in);
30485	}
	@@ -29095,11 +30508,11 @@
30508	rc = 1;
30509	goto meta_command_exit;
30510	}
30511	rc = sqlite3_open(zSrcFile, &pSrc);
30512	if( rc!=SQLITE_OK ){
30513	sqlite3_fprintf(stderr,"Error: cannot open \"%s\"\n", zSrcFile);
30514	close_db(pSrc);
30515	return 1;
30516	}
30517	open_db(p, 0);
30518	pBackup = sqlite3_backup_init(p->db, zDb, pSrc, "main");
	@@ -29127,11 +30540,14 @@
30540	}
30541	close_db(pSrc);
30542	}else
30543	#endif /* !defined(SQLITE_SHELL_FIDDLE) */
30544
30545	if( c=='s' &&
30546	(cli_strncmp(azArg[0], "scanstats", n)==0 \|\|
30547	cli_strncmp(azArg[0], "scanstatus", n)==0)
30548	){
30549	if( nArg==2 ){
30550	if( cli_strcmp(azArg[1], "vm")==0 ){
30551	p->scanstatsOn = 3;
30552	}else
30553	if( cli_strcmp(azArg[1], "est")==0 ){
	@@ -29178,11 +30594,11 @@
30594	}else if( optionMatch(azArg[ii],"debug") ){
30595	bDebug = 1;
30596	}else if( optionMatch(azArg[ii],"nosys") ){
30597	bNoSystemTabs = 1;
30598	}else if( azArg[ii][0]=='-' ){
30599	sqlite3_fprintf(stderr,"Unknown option: \"%s\"\n", azArg[ii]);
30600	rc = 1;
30601	goto meta_command_exit;
30602	}else if( zName==0 ){
30603	zName = azArg[ii];
30604	}else{
	@@ -29279,11 +30695,11 @@
30695	appendText(&sSelect, "name NOT LIKE 'sqlite_%%' AND ", 0);
30696	}
30697	appendText(&sSelect, "sql IS NOT NULL"
30698	" ORDER BY snum, rowid", 0);
30699	if( bDebug ){
30700	sqlite3_fprintf(p->out, "SQL: %s;\n", sSelect.z);
30701	}else{
30702	rc = sqlite3_exec(p->db, sSelect.z, callback, &data, &zErrMsg);
30703	}
30704	freeText(&sSelect);
30705	}
	@@ -29340,11 +30756,12 @@
30756	session_not_open:
30757	eputz("ERROR: No sessions are open\n");
30758	}else{
30759	rc = sqlite3session_attach(pSession->p, azCmd[1]);
30760	if( rc ){
30761	sqlite3_fprintf(stderr,
30762	"ERROR: sqlite3session_attach() returns %d\n",rc);
30763	rc = 0;
30764	}
30765	}
30766	}else
30767
	@@ -29357,13 +30774,13 @@
30774	){
30775	FILE *out = 0;
30776	failIfSafeMode(p, "cannot run \".session %s\" in safe mode", azCmd[0]);
30777	if( nCmd!=2 ) goto session_syntax_error;
30778	if( pSession->p==0 ) goto session_not_open;
30779	out = sqlite3_fopen(azCmd[1], "wb");
30780	if( out==0 ){
30781	sqlite3_fprintf(stderr,"ERROR: cannot open \"%s\" for writing\n",
30782	azCmd[1]);
30783	}else{
30784	int szChng;
30785	void *pChng;
30786	if( azCmd[0][0]=='c' ){
	@@ -29370,16 +30787,17 @@
30787	rc = sqlite3session_changeset(pSession->p, &szChng, &pChng);
30788	}else{
30789	rc = sqlite3session_patchset(pSession->p, &szChng, &pChng);
30790	}
30791	if( rc ){
30792	sqlite3_fprintf(stdout, "Error: error code %d\n", rc);
30793	rc = 0;
30794	}
30795	if( pChng
30796	&& fwrite(pChng, szChng, 1, out)!=1 ){
30797	sqlite3_fprintf(stderr,
30798	"ERROR: Failed to write entire %d-byte output\n", szChng);
30799	}
30800	sqlite3_free(pChng);
30801	fclose(out);
30802	}
30803	}else
	@@ -29402,11 +30820,12 @@
30820	int ii;
30821	if( nCmd>2 ) goto session_syntax_error;
30822	ii = nCmd==1 ? -1 : booleanValue(azCmd[1]);
30823	if( pAuxDb->nSession ){
30824	ii = sqlite3session_enable(pSession->p, ii);
30825	sqlite3_fprintf(p->out,
30826	"session %s enable flag = %d\n", pSession->zName, ii);
30827	}
30828	}else
30829
30830	/* .session filter GLOB ....
30831	** Set a list of GLOB patterns of table names to be excluded.
	@@ -29437,11 +30856,12 @@
30856	int ii;
30857	if( nCmd>2 ) goto session_syntax_error;
30858	ii = nCmd==1 ? -1 : booleanValue(azCmd[1]);
30859	if( pAuxDb->nSession ){
30860	ii = sqlite3session_indirect(pSession->p, ii);
30861	sqlite3_fprintf(p->out,
30862	"session %s indirect flag = %d\n", pSession->zName, ii);
30863	}
30864	}else
30865
30866	/* .session isempty
30867	** Determine if the session is empty
	@@ -29449,20 +30869,21 @@
30869	if( cli_strcmp(azCmd[0], "isempty")==0 ){
30870	int ii;
30871	if( nCmd!=1 ) goto session_syntax_error;
30872	if( pAuxDb->nSession ){
30873	ii = sqlite3session_isempty(pSession->p);
30874	sqlite3_fprintf(p->out,
30875	"session %s isempty flag = %d\n", pSession->zName, ii);
30876	}
30877	}else
30878
30879	/* .session list
30880	** List all currently open sessions
30881	*/
30882	if( cli_strcmp(azCmd[0],"list")==0 ){
30883	for(i=0; i<pAuxDb->nSession; i++){
30884	sqlite3_fprintf(p->out, "%d %s\n", i, pAuxDb->aSession[i].zName);
30885	}
30886	}else
30887
30888	/* .session open DB NAME
30889	** Open a new session called NAME on the attached database DB.
	@@ -29473,22 +30894,23 @@
30894	if( nCmd!=3 ) goto session_syntax_error;
30895	zName = azCmd[2];
30896	if( zName[0]==0 ) goto session_syntax_error;
30897	for(i=0; i<pAuxDb->nSession; i++){
30898	if( cli_strcmp(pAuxDb->aSession[i].zName,zName)==0 ){
30899	sqlite3_fprintf(stderr,"Session \"%s\" already exists\n", zName);
30900	goto meta_command_exit;
30901	}
30902	}
30903	if( pAuxDb->nSession>=ArraySize(pAuxDb->aSession) ){
30904	sqlite3_fprintf(stderr,
30905	"Maximum of %d sessions\n", ArraySize(pAuxDb->aSession));
30906	goto meta_command_exit;
30907	}
30908	pSession = &pAuxDb->aSession[pAuxDb->nSession];
30909	rc = sqlite3session_create(p->db, azCmd[1], &pSession->p);
30910	if( rc ){
30911	sqlite3_fprintf(stderr,"Cannot open session: error code=%d\n", rc);
30912	rc = 0;
30913	goto meta_command_exit;
30914	}
30915	pSession->nFilter = 0;
30916	sqlite3session_table_filter(pSession->p, session_filter, pSession);
	@@ -29508,20 +30930,20 @@
30930	if( c=='s' && n>=10 && cli_strncmp(azArg[0], "selftest-", 9)==0 ){
30931	if( cli_strncmp(azArg[0]+9, "boolean", n-9)==0 ){
30932	int i, v;
30933	for(i=1; i<nArg; i++){
30934	v = booleanValue(azArg[i]);
30935	sqlite3_fprintf(p->out, "%s: %d 0x%x\n", azArg[i], v, v);
30936	}
30937	}
30938	if( cli_strncmp(azArg[0]+9, "integer", n-9)==0 ){
30939	int i; sqlite3_int64 v;
30940	for(i=1; i<nArg; i++){
30941	char zBuf[200];
30942	v = integerValue(azArg[i]);
30943	sqlite3_snprintf(sizeof(zBuf),zBuf,"%s: %lld 0x%llx\n", azArg[i],v,v);
30944	sqlite3_fputs(zBuf, p->out);
30945	}
30946	}
30947	}else
30948	#endif
30949
	@@ -29544,12 +30966,13 @@
30966	}else
30967	if( cli_strcmp(z,"-v")==0 ){
30968	bVerbose++;
30969	}else
30970	{
30971	sqlite3_fprintf(stderr,
30972	"Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
30973	sqlite3_fputs("Should be one of: --init -v\n", stderr);
30974	rc = 1;
30975	goto meta_command_exit;
30976	}
30977	}
30978	if( sqlite3_table_column_metadata(p->db,"main","selftest",0,0,0,0,0,0)
	@@ -29590,46 +31013,47 @@
31013	if( zOp==0 ) continue;
31014	if( zSql==0 ) continue;
31015	if( zAns==0 ) continue;
31016	k = 0;
31017	if( bVerbose>0 ){
31018	sqlite3_fprintf(stdout, "%d: %s %s\n", tno, zOp, zSql);
31019	}
31020	if( cli_strcmp(zOp,"memo")==0 ){
31021	sqlite3_fprintf(p->out, "%s\n", zSql);
31022	}else
31023	if( cli_strcmp(zOp,"run")==0 ){
31024	char *zErrMsg = 0;
31025	str.n = 0;
31026	str.z[0] = 0;
31027	rc = sqlite3_exec(p->db, zSql, captureOutputCallback, &str, &zErrMsg);
31028	nTest++;
31029	if( bVerbose ){
31030	sqlite3_fprintf(p->out, "Result: %s\n", str.z);
31031	}
31032	if( rc \|\| zErrMsg ){
31033	nErr++;
31034	rc = 1;
31035	sqlite3_fprintf(p->out, "%d: error-code-%d: %s\n", tno, rc,zErrMsg);
31036	sqlite3_free(zErrMsg);
31037	}else if( cli_strcmp(zAns,str.z)!=0 ){
31038	nErr++;
31039	rc = 1;
31040	sqlite3_fprintf(p->out, "%d: Expected: [%s]\n", tno, zAns);
31041	sqlite3_fprintf(p->out, "%d: Got: [%s]\n", tno, str.z);
31042	}
31043	}
31044	else{
31045	sqlite3_fprintf(stderr,
31046	"Unknown operation \"%s\" on selftest line %d\n", zOp, tno);
31047	rc = 1;
31048	break;
31049	}
31050	} /* End loop over rows of content from SELFTEST */
31051	sqlite3_finalize(pStmt);
31052	} /* End loop over k */
31053	freeText(&str);
31054	sqlite3_fprintf(p->out, "%d errors out of %d tests\n", nErr, nTest);
31055	}else
31056
31057	if( c=='s' && cli_strncmp(azArg[0], "separator", n)==0 ){
31058	if( nArg<2 \|\| nArg>3 ){
31059	eputz("Usage: .separator COL ?ROW?\n");
	@@ -29673,11 +31097,12 @@
31097	}else
31098	if( cli_strcmp(z,"debug")==0 ){
31099	bDebug = 1;
31100	}else
31101	{
31102	sqlite3_fprintf(stderr,
31103	"Unknown option \"%s\" on \"%s\"\n", azArg[i], azArg[0]);
31104	showHelp(p->out, azArg[0]);
31105	rc = 1;
31106	goto meta_command_exit;
31107	}
31108	}else if( zLike ){
	@@ -29751,11 +31176,11 @@
31176	}
31177	shell_check_oom(zSql);
31178	freeText(&sQuery);
31179	freeText(&sSql);
31180	if( bDebug ){
31181	sqlite3_fprintf(p->out, "%s\n", zSql);
31182	}else{
31183	shell_exec(p, zSql, 0);
31184	}
31185	#if !defined(SQLITE_OMIT_SCHEMA_PRAGMAS) && !defined(SQLITE_OMIT_VIRTUALTABLE)
31186	{
	@@ -29781,11 +31206,11 @@
31206	"\|\|group_concat('CAST(CAST('\|\|cname\|\|' AS BLOB) AS TEXT)<>'\|\|cname\n"
31207	"\|\| ' AND typeof('\|\|cname\|\|')=''text'' ',\n"
31208	"' OR ') as query, tname from tabcols group by tname)"
31209	, zRevText);
31210	shell_check_oom(zRevText);
31211	if( bDebug ) sqlite3_fprintf(p->out, "%s\n", zRevText);
31212	lrc = sqlite3_prepare_v2(p->db, zRevText, -1, &pStmt, 0);
31213	if( lrc!=SQLITE_OK ){
31214	/* assert(lrc==SQLITE_NOMEM); // might also be SQLITE_ERROR if the
31215	** user does cruel and unnatural things like ".limit expr_depth 0". */
31216	rc = 1;
	@@ -29794,19 +31219,20 @@
31219	lrc = SQLITE_ROW==sqlite3_step(pStmt);
31220	if( lrc ){
31221	const char zGenQuery = (char)sqlite3_column_text(pStmt,0);
31222	sqlite3_stmt *pCheckStmt;
31223	lrc = sqlite3_prepare_v2(p->db, zGenQuery, -1, &pCheckStmt, 0);
31224	if( bDebug ) sqlite3_fprintf(p->out, "%s\n", zGenQuery);
31225	if( lrc!=SQLITE_OK ){
31226	rc = 1;
31227	}else{
31228	if( SQLITE_ROW==sqlite3_step(pCheckStmt) ){
31229	double countIrreversible = sqlite3_column_double(pCheckStmt, 0);
31230	if( countIrreversible>0 ){
31231	int sz = (int)(countIrreversible + 0.5);
31232	sqlite3_fprintf(stderr,
31233	"Digest includes %d invalidly encoded text field%s.\n",
31234	sz, (sz>1)? "s": "");
31235	}
31236	}
31237	sqlite3_finalize(pCheckStmt);
31238	}
	@@ -29836,15 +31262,15 @@
31262	zCmd = sqlite3_mprintf(strchr(azArg[1],' ')==0?"%s":"\"%s\"", azArg[1]);
31263	for(i=2; i<nArg && zCmd!=0; i++){
31264	zCmd = sqlite3_mprintf(strchr(azArg[i],' ')==0?"%z %s":"%z \"%s\"",
31265	zCmd, azArg[i]);
31266	}
31267	/consoleRestore();/
31268	x = zCmd!=0 ? system(zCmd) : 1;
31269	/consoleRenewSetup();/
31270	sqlite3_free(zCmd);
31271	if( x ) sqlite3_fprintf(stderr,"System command returns %d\n", x);
31272	}else
31273	#endif /* !defined(SQLITE_NOHAVE_SYSTEM) && !defined(SQLITE_SHELL_FIDDLE) */
31274
31275	if( c=='s' && cli_strncmp(azArg[0], "show", n)==0 ){
31276	static const char *azBool[] = { "off", "on", "trigger", "full"};
	@@ -29853,50 +31279,52 @@
31279	if( nArg!=1 ){
31280	eputz("Usage: .show\n");
31281	rc = 1;
31282	goto meta_command_exit;
31283	}
31284	sqlite3_fprintf(p->out, "%12.12s: %s\n","echo",
31285	azBool[ShellHasFlag(p, SHFLG_Echo)]);
31286	sqlite3_fprintf(p->out, "%12.12s: %s\n","eqp", azBool[p->autoEQP&3]);
31287	sqlite3_fprintf(p->out, "%12.12s: %s\n","explain",
31288	p->mode==MODE_Explain ? "on" : p->autoExplain ? "auto" : "off");
31289	sqlite3_fprintf(p->out, "%12.12s: %s\n","headers",
31290	azBool[p->showHeader!=0]);
31291	if( p->mode==MODE_Column
31292	\|\| (p->mode>=MODE_Markdown && p->mode<=MODE_Box)
31293	){
31294	sqlite3_fprintf(p->out,
31295	"%12.12s: %s --wrap %d --wordwrap %s --%squote\n", "mode",
31296	modeDescr[p->mode], p->cmOpts.iWrap,
31297	p->cmOpts.bWordWrap ? "on" : "off",
31298	p->cmOpts.bQuote ? "" : "no");
31299	}else{
31300	sqlite3_fprintf(p->out, "%12.12s: %s\n","mode", modeDescr[p->mode]);
31301	}
31302	sqlite3_fprintf(p->out, "%12.12s: ", "nullvalue");
31303	output_c_string(p->out, p->nullValue);
31304	sqlite3_fputs("\n", p->out);
31305	sqlite3_fprintf(p->out, "%12.12s: %s\n","output",
31306	strlen30(p->outfile) ? p->outfile : "stdout");
31307	sqlite3_fprintf(p->out, "%12.12s: ", "colseparator");
31308	output_c_string(p->out, p->colSeparator);
31309	sqlite3_fputs("\n", p->out);
31310	sqlite3_fprintf(p->out, "%12.12s: ", "rowseparator");
31311	output_c_string(p->out, p->rowSeparator);
31312	sqlite3_fputs("\n", p->out);
31313	switch( p->statsOn ){
31314	case 0: zOut = "off"; break;
31315	default: zOut = "on"; break;
31316	case 2: zOut = "stmt"; break;
31317	case 3: zOut = "vmstep"; break;
31318	}
31319	sqlite3_fprintf(p->out, "%12.12s: %s\n","stats", zOut);
31320	sqlite3_fprintf(p->out, "%12.12s: ", "width");
31321	for (i=0;i<p->nWidth;i++) {
31322	sqlite3_fprintf(p->out, "%d ", p->colWidth[i]);
31323	}
31324	sqlite3_fputs("\n", p->out);
31325	sqlite3_fprintf(p->out, "%12.12s: %s\n", "filename",
31326	p->pAuxDb->zDbFilename ? p->pAuxDb->zDbFilename : "");
31327	}else
31328
31329	if( c=='s' && cli_strncmp(azArg[0], "stats", n)==0 ){
31330	if( nArg==2 ){
	@@ -30010,13 +31438,14 @@
31438	if( nPrintCol<1 ) nPrintCol = 1;
31439	nPrintRow = (nRow + nPrintCol - 1)/nPrintCol;
31440	for(i=0; i<nPrintRow; i++){
31441	for(j=i; j<nRow; j+=nPrintRow){
31442	char *zSp = j<nPrintRow ? "" : " ";
31443	sqlite3_fprintf(p->out,
31444	"%s%-*s", zSp, maxlen, azResult[j] ? azResult[j]:"");
31445	}
31446	sqlite3_fputs("\n", p->out);
31447	}
31448	}
31449
31450	for(ii=0; ii<nRow; ii++) sqlite3_free(azResult[ii]);
31451	sqlite3_free(azResult);
	@@ -30024,11 +31453,11 @@
31453
31454	#ifndef SQLITE_SHELL_FIDDLE
31455	/* Begin redirecting output to the file "testcase-out.txt" */
31456	if( c=='t' && cli_strcmp(azArg[0],"testcase")==0 ){
31457	output_reset(p);
31458	p->out = output_file_open("testcase-out.txt");
31459	if( p->out==0 ){
31460	eputz("Error: cannot open 'testcase-out.txt'\n");
31461	}
31462	if( nArg>=2 ){
31463	sqlite3_snprintf(sizeof(p->zTestcase), p->zTestcase, "%s", azArg[1]);
	@@ -30068,11 +31497,10 @@
31497	{"prng_save", SQLITE_TESTCTRL_PRNG_SAVE, 0, "" },
31498	{"prng_seed", SQLITE_TESTCTRL_PRNG_SEED, 0, "SEED ?db?" },
31499	{"seek_count", SQLITE_TESTCTRL_SEEK_COUNT, 0, "" },
31500	{"sorter_mmap", SQLITE_TESTCTRL_SORTER_MMAP, 0, "NMAX" },
31501	{"tune", SQLITE_TESTCTRL_TUNE, 1, "ID VALUE" },

31502	};
31503	int testctrl = -1;
31504	int iCtrl = -1;
31505	int rc2 = 0; /* 0: usage. 1: %d 2: %x 3: no-output */
31506	int isOk = 0;
	@@ -30088,14 +31516,14 @@
31516	if( zCmd[0]=='-' && zCmd[1] ) zCmd++;
31517	}
31518
31519	/* --help lists all test-controls */
31520	if( cli_strcmp(zCmd,"help")==0 ){
31521	sqlite3_fputs("Available test-controls:\n", p->out);
31522	for(i=0; i<ArraySize(aCtrl); i++){
31523	if( aCtrl[i].unSafe && !ShellHasFlag(p,SHFLG_TestingMode) ) continue;
31524	sqlite3_fprintf(p->out, " .testctrl %s %s\n",
31525	aCtrl[i].zCtrlName, aCtrl[i].zUsage);
31526	}
31527	rc = 1;
31528	goto meta_command_exit;
31529	}
	@@ -30108,19 +31536,19 @@
31536	if( cli_strncmp(zCmd, aCtrl[i].zCtrlName, n2)==0 ){
31537	if( testctrl<0 ){
31538	testctrl = aCtrl[i].ctrlCode;
31539	iCtrl = i;
31540	}else{
31541	sqlite3_fprintf(stderr,"Error: ambiguous test-control: \"%s\"\n"
31542	"Use \".testctrl --help\" for help\n", zCmd);
31543	rc = 1;
31544	goto meta_command_exit;
31545	}
31546	}
31547	}
31548	if( testctrl<0 ){
31549	sqlite3_fprintf(stderr,"Error: unknown test-control: %s\n"
31550	"Use \".testctrl --help\" for help\n", zCmd);
31551	}else{
31552	switch(testctrl){
31553
31554	/* Special processing for .testctrl opt MASK ...
	@@ -30170,11 +31598,13 @@
31598	{ 0x10000000, 1, "OrderBySubq" },
31599	{ 0xffffffff, 0, "All" },
31600	};
31601	unsigned int curOpt;
31602	unsigned int newOpt;
31603	unsigned int m;
31604	int ii;
31605	int nOff;
31606	sqlite3_test_control(SQLITE_TESTCTRL_GETOPT, p->db, &curOpt);
31607	newOpt = curOpt;
31608	for(ii=2; ii<nArg; ii++){
31609	const char *z = azArg[ii];
31610	int useLabel = 0;
	@@ -30192,16 +31622,17 @@
31622	int jj;
31623	for(jj=0; jj<ArraySize(aLabel); jj++){
31624	if( sqlite3_stricmp(zLabel, aLabel[jj].zLabel)==0 ) break;
31625	}
31626	if( jj>=ArraySize(aLabel) ){
31627	sqlite3_fprintf(stderr,
31628	"Error: no such optimization: \"%s\"\n", zLabel);
31629	sqlite3_fputs("Should be one of:", stderr);
31630	for(jj=0; jj<ArraySize(aLabel); jj++){
31631	sqlite3_fprintf(stderr," %s", aLabel[jj].zLabel);
31632	}
31633	sqlite3_fputs("\n", stderr);
31634	rc = 1;
31635	goto meta_command_exit;
31636	}
31637	if( useLabel=='+' ){
31638	newOpt &= ~aLabel[jj].mask;
	@@ -30210,28 +31641,32 @@
31641	}
31642	}
31643	}
31644	if( curOpt!=newOpt ){
31645	sqlite3_test_control(SQLITE_TESTCTRL_OPTIMIZATIONS,p->db,newOpt);
31646	}
31647	for(ii=nOff=0, m=1; ii<32; ii++, m <<= 1){
31648	if( m & newOpt ) nOff++;
31649	}
31650	if( nOff<12 ){
31651	sqlite3_fputs("+All", p->out);
31652	for(ii=0; ii<ArraySize(aLabel); ii++){
31653	if( !aLabel[ii].bDsply ) continue;
31654	if( (newOpt & aLabel[ii].mask)!=0 ){
31655	sqlite3_fprintf(p->out, " -%s", aLabel[ii].zLabel);
31656	}
31657	}
31658	}else{
31659	sqlite3_fputs("-All", p->out);
31660	for(ii=0; ii<ArraySize(aLabel); ii++){
31661	if( !aLabel[ii].bDsply ) continue;
31662	if( (newOpt & aLabel[ii].mask)==0 ){
31663	sqlite3_fprintf(p->out, " +%s", aLabel[ii].zLabel);


31664	}
31665	}
31666	}
31667	sqlite3_fputs("\n", p->out);
31668	rc2 = isOk = 3;
31669	break;
31670	}
31671
31672	/* sqlite3_test_control(int, db, int) */
	@@ -30267,11 +31702,11 @@
31702	if( nArg==3 \|\| nArg==4 ){
31703	int ii = (int)integerValue(azArg[2]);
31704	sqlite3 *db;
31705	if( ii==0 && cli_strcmp(azArg[2],"random")==0 ){
31706	sqlite3_randomness(sizeof(ii),&ii);
31707	sqlite3_fprintf(stdout, "-- random seed: %d\n", ii);
31708	}
31709	if( nArg==3 ){
31710	db = 0;
31711	}else{
31712	db = p->db;
	@@ -30301,25 +31736,10 @@
31736	rc2 = sqlite3_test_control(testctrl, opt);
31737	isOk = 3;
31738	}
31739	break;
31740















31741	/* sqlite3_test_control(sqlite3) /
31742	case SQLITE_TESTCTRL_INTERNAL_FUNCTIONS:
31743	rc2 = sqlite3_test_control(testctrl, p->db);
31744	isOk = 3;
31745	break;
	@@ -30335,11 +31755,11 @@
31755	break;
31756
31757	case SQLITE_TESTCTRL_SEEK_COUNT: {
31758	u64 x = 0;
31759	rc2 = sqlite3_test_control(testctrl, p->db, &x);
31760	sqlite3_fprintf(p->out, "%llu\n", x);
31761	isOk = 3;
31762	break;
31763	}
31764
31765	#ifdef YYCOVERAGE
	@@ -30366,15 +31786,15 @@
31786	int id = 1;
31787	while(1){
31788	int val = 0;
31789	rc2 = sqlite3_test_control(testctrl, -id, &val);
31790	if( rc2!=SQLITE_OK ) break;
31791	if( id>1 ) sqlite3_fputs(" ", p->out);
31792	sqlite3_fprintf(p->out, "%d: %d", id, val);
31793	id++;
31794	}
31795	if( id>1 ) sqlite3_fputs("\n", p->out);
31796	isOk = 3;
31797	}
31798	break;
31799	}
31800	#endif
	@@ -30412,18 +31832,26 @@
31832	}else if( cli_strcmp(z,"reset")==0 ){
31833	faultsim_state.iCnt = faultsim_state.nSkip;
31834	faultsim_state.nHit = 0;
31835	sqlite3_test_control(testctrl, faultsim_callback);
31836	}else if( cli_strcmp(z,"status")==0 ){
31837	sqlite3_fprintf(p->out, "faultsim.iId: %d\n",
31838	faultsim_state.iId);
31839	sqlite3_fprintf(p->out, "faultsim.iErr: %d\n",
31840	faultsim_state.iErr);
31841	sqlite3_fprintf(p->out, "faultsim.iCnt: %d\n",
31842	faultsim_state.iCnt);
31843	sqlite3_fprintf(p->out, "faultsim.nHit: %d\n",
31844	faultsim_state.nHit);
31845	sqlite3_fprintf(p->out, "faultsim.iInterval: %d\n",
31846	faultsim_state.iInterval);
31847	sqlite3_fprintf(p->out, "faultsim.eVerbose: %d\n",
31848	faultsim_state.eVerbose);
31849	sqlite3_fprintf(p->out, "faultsim.nRepeat: %d\n",
31850	faultsim_state.nRepeat);
31851	sqlite3_fprintf(p->out, "faultsim.nSkip: %d\n",
31852	faultsim_state.nSkip);
31853	}else if( cli_strcmp(z,"-v")==0 ){
31854	if( faultsim_state.eVerbose<2 ) faultsim_state.eVerbose++;
31855	}else if( cli_strcmp(z,"-q")==0 ){
31856	if( faultsim_state.eVerbose>0 ) faultsim_state.eVerbose--;
31857	}else if( cli_strcmp(z,"-id")==0 && kk+1<nArg ){
	@@ -30437,19 +31865,20 @@
31865	}else if( cli_strcmp(z,"-skip")==0 && kk+1<nArg ){
31866	faultsim_state.nSkip = atoi(azArg[++kk]);
31867	}else if( cli_strcmp(z,"-?")==0 \|\| sqlite3_strglob("help",z)==0){
31868	bShowHelp = 1;
31869	}else{
31870	sqlite3_fprintf(stderr,
31871	"Unrecognized fault_install argument: \"%s\"\n",
31872	azArg[kk]);
31873	rc = 1;
31874	bShowHelp = 1;
31875	break;
31876	}
31877	}
31878	if( bShowHelp ){
31879	sqlite3_fputs(
31880	"Usage: .testctrl fault_install ARGS\n"
31881	"Possible arguments:\n"
31882	" off Disable faultsim\n"
31883	" on Activate faultsim\n"
31884	" reset Reset the trigger counter\n"
	@@ -30459,23 +31888,25 @@
31888	" --errcode N When triggered, return N as error code\n"
31889	" --id ID Trigger only for the ID specified\n"
31890	" --interval N Trigger only after every N-th call\n"
31891	" --repeat N Turn off after N hits. 0 means never\n"
31892	" --skip N Skip the first N encounters\n"
31893	,p->out
31894	);
31895	}
31896	break;
31897	}
31898	}
31899	}
31900	if( isOk==0 && iCtrl>=0 ){
31901	sqlite3_fprintf(p->out,
31902	"Usage: .testctrl %s %s\n", zCmd,aCtrl[iCtrl].zUsage);
31903	rc = 1;
31904	}else if( isOk==1 ){
31905	sqlite3_fprintf(p->out, "%d\n", rc2);
31906	}else if( isOk==2 ){
31907	sqlite3_fprintf(p->out, "0x%08x\n", rc2);
31908	}
31909	}else
31910	#endif /* !defined(SQLITE_UNTESTABLE) */
31911
31912	if( c=='t' && n>4 && cli_strncmp(azArg[0], "timeout", n)==0 ){
	@@ -30526,17 +31957,17 @@
31957	}
31958	else if( optionMatch(z, "close") ){
31959	mType \|= SQLITE_TRACE_CLOSE;
31960	}
31961	else {
31962	sqlite3_fprintf(stderr,"Unknown option \"%s\" on \".trace\"\n", z);
31963	rc = 1;
31964	goto meta_command_exit;
31965	}
31966	}else{
31967	output_file_close(p->traceOut);
31968	p->traceOut = output_file_open(z);
31969	}
31970	}
31971	if( p->traceOut==0 ){
31972	sqlite3_trace_v2(p->db, 0, 0, 0);
31973	}else{
	@@ -30569,103 +32000,40 @@
32000	}
32001	}
32002	}else
32003	#endif
32004































































32005	if( c=='v' && cli_strncmp(azArg[0], "version", n)==0 ){
32006	char zPtrSz = sizeof(void)==8 ? "64-bit" : "32-bit";
32007	sqlite3_fprintf(p->out, "SQLite %s %s\n" /extra-version-info/,
32008	sqlite3_libversion(), sqlite3_sourceid());
32009	#if SQLITE_HAVE_ZLIB
32010	sqlite3_fprintf(p->out, "zlib version %s\n", zlibVersion());
32011	#endif
32012	#define CTIMEOPT_VAL_(opt) #opt
32013	#define CTIMEOPT_VAL(opt) CTIMEOPT_VAL_(opt)
32014	#if defined(__clang__) && defined(__clang_major__)
32015	sqlite3_fprintf(p->out, "clang-" CTIMEOPT_VAL(__clang_major__) "."
32016	CTIMEOPT_VAL(__clang_minor__) "."
32017	CTIMEOPT_VAL(__clang_patchlevel__) " (%s)\n", zPtrSz);
32018	#elif defined(_MSC_VER)
32019	sqlite3_fprintf(p->out, "msvc-" CTIMEOPT_VAL(_MSC_VER) " (%s)\n", zPtrSz);
32020	#elif defined(__GNUC__) && defined(__VERSION__)
32021	sqlite3_fprintf(p->out, "gcc-" __VERSION__ " (%s)\n", zPtrSz);
32022	#endif
32023	}else
32024
32025	if( c=='v' && cli_strncmp(azArg[0], "vfsinfo", n)==0 ){
32026	const char *zDbName = nArg==2 ? azArg[1] : "main";
32027	sqlite3_vfs *pVfs = 0;
32028	if( p->db ){
32029	sqlite3_file_control(p->db, zDbName, SQLITE_FCNTL_VFS_POINTER, &pVfs);
32030	if( pVfs ){
32031	sqlite3_fprintf(p->out, "vfs.zName = \"%s\"\n", pVfs->zName);
32032	sqlite3_fprintf(p->out, "vfs.iVersion = %d\n", pVfs->iVersion);
32033	sqlite3_fprintf(p->out, "vfs.szOsFile = %d\n", pVfs->szOsFile);
32034	sqlite3_fprintf(p->out, "vfs.mxPathname = %d\n", pVfs->mxPathname);
32035	}
32036	}
32037	}else
32038
32039	if( c=='v' && cli_strncmp(azArg[0], "vfslist", n)==0 ){
	@@ -30673,17 +32041,17 @@
32041	sqlite3_vfs *pCurrent = 0;
32042	if( p->db ){
32043	sqlite3_file_control(p->db, "main", SQLITE_FCNTL_VFS_POINTER, &pCurrent);
32044	}
32045	for(pVfs=sqlite3_vfs_find(0); pVfs; pVfs=pVfs->pNext){
32046	sqlite3_fprintf(p->out, "vfs.zName = \"%s\"%s\n", pVfs->zName,
32047	pVfs==pCurrent ? " <--- CURRENT" : "");
32048	sqlite3_fprintf(p->out, "vfs.iVersion = %d\n", pVfs->iVersion);
32049	sqlite3_fprintf(p->out, "vfs.szOsFile = %d\n", pVfs->szOsFile);
32050	sqlite3_fprintf(p->out, "vfs.mxPathname = %d\n", pVfs->mxPathname);
32051	if( pVfs->pNext ){
32052	sqlite3_fputs("-----------------------------------\n", p->out);
32053	}
32054	}
32055	}else
32056
32057	if( c=='v' && cli_strncmp(azArg[0], "vfsname", n)==0 ){
	@@ -30690,11 +32058,11 @@
32058	const char *zDbName = nArg==2 ? azArg[1] : "main";
32059	char *zVfsName = 0;
32060	if( p->db ){
32061	sqlite3_file_control(p->db, zDbName, SQLITE_FCNTL_VFSNAME, &zVfsName);
32062	if( zVfsName ){
32063	sqlite3_fprintf(p->out, "%s\n", zVfsName);
32064	sqlite3_free(zVfsName);
32065	}
32066	}
32067	}else
32068
	@@ -30714,11 +32082,11 @@
32082	p->colWidth[j-1] = (int)integerValue(azArg[j]);
32083	}
32084	}else
32085
32086	{
32087	sqlite3_fprintf(stderr,"Error: unknown command or invalid arguments: "
32088	" \"%s\". Enter \".help\" for help\n", azArg[0]);
32089	rc = 1;
32090	}
32091
32092	meta_command_exit:
	@@ -30755,11 +32123,10 @@
32123	SCAN_TRACKER_REFTYPE pst){
32124	char cin;
32125	char cWait = (char)qss; /* intentional narrowing loss */
32126	if( cWait==0 ){
32127	PlainScan:

32128	while( (cin = *zLine++)!=0 ){
32129	if( IsSpace(cin) )
32130	continue;
32131	switch (cin){
32132	case '-':
	@@ -30807,11 +32174,10 @@
32174	switch( cWait ){
32175	case '*':
32176	if( *zLine != '/' )
32177	continue;
32178	++zLine;

32179	CONTINUE_PROMPT_AWAITC(pst, 0);
32180	qss = QSS_SETV(qss, 0);
32181	goto PlainScan;
32182	case '`': case '\'': case '"':
32183	if(*zLine==cWait){
	@@ -30819,11 +32185,10 @@
32185	++zLine;
32186	continue;
32187	}
32188	deliberate_fall_through;
32189	case ']':

32190	CONTINUE_PROMPT_AWAITC(pst, 0);
32191	qss = QSS_SETV(qss, 0);
32192	goto PlainScan;
32193	default: assert(0);
32194	}
	@@ -30966,11 +32331,11 @@
32331	open_db(p, 0);
32332	if( ShellHasFlag(p,SHFLG_Backslash) ) resolve_backslashes(zSql);
32333	if( p->flgProgress & SHELL_PROGRESS_RESET ) p->nProgress = 0;
32334	BEGIN_TIMER;
32335	rc = shell_exec(p, zSql, &zErrMsg);
32336	END_TIMER(p->out);
32337	if( rc \|\| zErrMsg ){
32338	char zPrefix[100];
32339	const char *zErrorTail;
32340	const char *zErrorType;
32341	if( zErrMsg==0 ){
	@@ -30990,28 +32355,31 @@
32355	sqlite3_snprintf(sizeof(zPrefix), zPrefix,
32356	"%s near line %d:", zErrorType, startline);
32357	}else{
32358	sqlite3_snprintf(sizeof(zPrefix), zPrefix, "%s:", zErrorType);
32359	}
32360	sqlite3_fprintf(stderr,"%s %s\n", zPrefix, zErrorTail);
32361	sqlite3_free(zErrMsg);
32362	zErrMsg = 0;
32363	return 1;
32364	}else if( ShellHasFlag(p, SHFLG_CountChanges) ){
32365	char zLineBuf[2000];
32366	sqlite3_snprintf(sizeof(zLineBuf), zLineBuf,
32367	"changes: %lld total_changes: %lld",
32368	sqlite3_changes64(p->db), sqlite3_total_changes64(p->db));
32369	sqlite3_fprintf(p->out, "%s\n", zLineBuf);
32370	}
32371
32372	if( doAutoDetectRestore(p, zSql) ) return 1;
32373	return 0;
32374	}
32375
32376	static void echo_group_input(ShellState p, const char zDo){
32377	if( ShellHasFlag(p, SHFLG_Echo) ){
32378	sqlite3_fprintf(p->out, "%s\n", zDo);
32379	fflush(p->out);
32380	}
32381	}
32382
32383	#ifdef SQLITE_SHELL_FIDDLE
32384	/*
32385	** Alternate one_input_line() impl for wasm mode. This is not in the primary
	@@ -31065,11 +32433,11 @@
32433	i64 startline = 0; /* Line number for start of current input */
32434	QuickScanState qss = QSS_Start; /* Accumulated line status (so far) */
32435
32436	if( p->inputNesting==MAX_INPUT_NESTING ){
32437	/* This will be more informative in a later version. */
32438	sqlite3_fprintf(stderr,"Input nesting limit (%d) reached at line %d."
32439	" Check recursion.\n", MAX_INPUT_NESTING, p->lineno);
32440	return 1;
32441	}
32442	++p->inputNesting;
32443	p->lineno = 0;
	@@ -31077,11 +32445,11 @@
32445	while( errCnt==0 \|\| !bail_on_error \|\| (p->in==0 && stdin_is_interactive) ){
32446	fflush(p->out);
32447	zLine = one_input_line(p->in, zLine, nSql>0);
32448	if( zLine==0 ){
32449	/* End of input */
32450	if( p->in==0 && stdin_is_interactive ) sqlite3_fputs("\n", p->out);
32451	break;
32452	}
32453	if( seenInterrupt ){
32454	if( p->in!=0 ) break;
32455	seenInterrupt = 0;
	@@ -31297,19 +32665,19 @@
32665	}
32666	zBuf = sqlite3_mprintf("%s/.sqliterc",home_dir);
32667	shell_check_oom(zBuf);
32668	sqliterc = zBuf;
32669	}
32670	p->in = sqlite3_fopen(sqliterc,"rb");
32671	if( p->in ){
32672	if( stdin_is_interactive ){
32673	sqlite3_fprintf(stderr,"-- Loading resources from %s\n", sqliterc);
32674	}
32675	if( process_input(p) && bail_on_error ) exit(1);
32676	fclose(p->in);
32677	}else if( sqliterc_override!=0 ){
32678	sqlite3_fprintf(stderr,"cannot open: \"%s\"\n", sqliterc);
32679	if( bail_on_error ) exit(1);
32680	}
32681	p->in = inSaved;
32682	p->lineno = savedLineno;
32683	sqlite3_free(zBuf);
	@@ -31374,23 +32742,21 @@
32742	" -table set output mode to 'table'\n"
32743	" -tabs set output mode to 'tabs'\n"
32744	" -unsafe-testing allow unsafe commands and modes for testing\n"
32745	" -version show SQLite version\n"
32746	" -vfs NAME use NAME as the default VFS\n"

32747	" -vfstrace enable tracing of all VFS calls\n"

32748	#ifdef SQLITE_HAVE_ZLIB
32749	" -zip open the file as a ZIP Archive\n"
32750	#endif
32751	;
32752	static void usage(int showDetail){
32753	sqlite3_fprintf(stderr,"Usage: %s [OPTIONS] [FILENAME [SQL]]\n"
32754	"FILENAME is the name of an SQLite database. A new database is created\n"
32755	"if the file does not previously exist. Defaults to :memory:.\n", Argv0);
32756	if( showDetail ){
32757	sqlite3_fprintf(stderr,"OPTIONS include:\n%s", zOptions);
32758	}else{
32759	eputz("Use the -help option for additional information\n");
32760	}
32761	exit(0);
32762	}
	@@ -31411,10 +32777,13 @@
32777	*/
32778	static void main_init(ShellState *data) {
32779	memset(data, 0, sizeof(*data));
32780	data->normalMode = data->cMode = data->mode = MODE_List;
32781	data->autoExplain = 1;
32782	#ifdef _WIN32
32783	data->crlfMode = 1;
32784	#endif
32785	data->pAuxDb = &data->aAuxDb[0];
32786	memcpy(data->colSeparator,SEP_Column, 2);
32787	memcpy(data->rowSeparator,SEP_Row, 2);
32788	data->showHeader = 0;
32789	data->shellFlgs = SHFLG_Lookaside;
	@@ -31446,29 +32815,39 @@
32815	SetConsoleTextAttribute(out, defaultScreenInfo.wAttributes);
32816	#endif
32817	}
32818	#else
32819	static void printBold(const char *zText){
32820	sqlite3_fprintf(stdout, "\033[1m%s\033[0m", zText);
32821	}
32822	#endif
32823
32824	/*
32825	** Get the argument to an --option. Throw an error and die if no argument
32826	** is available.
32827	*/
32828	static char cmdline_option_value(int argc, char *argv, int i){
32829	if( i==argc ){
32830	sqlite3_fprintf(stderr,
32831	"%s: Error: missing argument to %s\n", argv[0], argv[argc-1]);
32832	exit(1);
32833	}
32834	return argv[i];
32835	}
32836
32837	static void sayAbnormalExit(void){
32838	if( seenInterrupt ) eputz("Program interrupted.\n");
32839	}
32840
32841	/* Routine to output from vfstrace
32842	*/
32843	static int vfstraceOut(const char z, void pArg){
32844	ShellState p = (ShellState)pArg;
32845	sqlite3_fputs(z, p->out);
32846	fflush(p->out);
32847	return 1;
32848	}
32849
32850	#ifndef SQLITE_SHELL_IS_UTF8
32851	# if (defined(_WIN32) \|\| defined(WIN32)) \
32852	&& (defined(_MSC_VER) \|\| (defined(UNICODE) && defined(__GNUC__)))
32853	# define SQLITE_SHELL_IS_UTF8 (0)
	@@ -31493,19 +32872,19 @@
32872	char *zErrMsg = 0;
32873	#ifdef SQLITE_SHELL_FIDDLE
32874	# define data shellState
32875	#else
32876	ShellState data;

32877	#endif
32878	const char *zInitFile = 0;
32879	int i;
32880	int rc = 0;
32881	int warnInmemoryDb = 0;
32882	int readStdin = 1;
32883	int nCmd = 0;
32884	int nOptsEnd = argc;
32885	int bEnableVfstrace = 0;
32886	char **azCmd = 0;
32887	const char zVfs = 0; / Value of -vfs command-line option */
32888	#if !SQLITE_SHELL_IS_UTF8
32889	char **argvToFree = 0;
32890	int argcToFree = 0;
	@@ -31515,25 +32894,29 @@
32894	#ifdef SQLITE_SHELL_FIDDLE
32895	stdin_is_interactive = 0;
32896	stdout_is_console = 1;
32897	data.wasm.zDefaultDbName = "/fiddle.sqlite3";
32898	#else
32899	stdin_is_interactive = isatty(0);
32900	stdout_is_console = isatty(1);


32901	#endif
32902	atexit(sayAbnormalExit);
32903	#ifdef SQLITE_DEBUG
32904	mem_main_enter = sqlite3_memory_used();
32905	#endif
32906	#if !defined(_WIN32_WCE)
32907	if( getenv("SQLITE_DEBUG_BREAK") ){
32908	if( isatty(0) && isatty(2) ){
32909	char zLine[100];
32910	sqlite3_fprintf(stderr,
32911	"attach debugger to process %d and press ENTER to continue...",
32912	GETPID());
32913	if( sqlite3_fgets(zLine, sizeof(zLine), stdin)!=0
32914	&& cli_strcmp(zLine,"stop")==0
32915	){
32916	exit(1);
32917	}
32918	}else{
32919	#if defined(_WIN32) \|\| defined(WIN32)
32920	#if SQLITE_OS_WINRT
32921	__debugbreak();
32922	#else
	@@ -31554,11 +32937,12 @@
32937	}
32938	#endif
32939
32940	#if USE_SYSTEM_SQLITE+0!=1
32941	if( cli_strncmp(sqlite3_sourceid(),SQLITE_SOURCE_ID,60)!=0 ){
32942	sqlite3_fprintf(stderr,
32943	"SQLite header and source version mismatch\n%s\n%s\n",
32944	sqlite3_sourceid(), SQLITE_SOURCE_ID);
32945	exit(1);
32946	}
32947	#endif
32948	main_init(&data);
	@@ -31696,21 +33080,12 @@
33080	switch( n ){
33081	case 0: sqlite3_config(SQLITE_CONFIG_SINGLETHREAD); break;
33082	case 2: sqlite3_config(SQLITE_CONFIG_MULTITHREAD); break;
33083	default: sqlite3_config(SQLITE_CONFIG_SERIALIZED); break;
33084	}

33085	}else if( cli_strcmp(z,"-vfstrace")==0 ){
33086	bEnableVfstrace = 1;








33087	#ifdef SQLITE_ENABLE_MULTIPLEX
33088	}else if( cli_strcmp(z,"-multiplex")==0 ){
33089	extern int sqlite3_multiplex_initialize(const char*,int);
33090	sqlite3_multiplex_initialize(0, 1);
33091	#endif
	@@ -31762,11 +33137,11 @@
33137	}else if( cli_strcmp(z,"-safe")==0 ){
33138	/* no-op - catch this on the second pass */
33139	}
33140	}
33141	#ifndef SQLITE_SHELL_FIDDLE
33142	if( !bEnableVfstrace ) verify_uninitialized();
33143	#endif
33144
33145
33146	#ifdef SQLITE_SHELL_INIT_PROC
33147	{
	@@ -31786,25 +33161,29 @@
33161	if( zVfs ){
33162	sqlite3_vfs *pVfs = sqlite3_vfs_find(zVfs);
33163	if( pVfs ){
33164	sqlite3_vfs_register(pVfs, 1);
33165	}else{
33166	sqlite3_fprintf(stderr,"no such VFS: \"%s\"\n", zVfs);
33167	exit(1);
33168	}
33169	}
33170
33171	if( data.pAuxDb->zDbFilename==0 ){
33172	#ifndef SQLITE_OMIT_MEMORYDB
33173	data.pAuxDb->zDbFilename = ":memory:";
33174	warnInmemoryDb = argc==1;
33175	#else
33176	sqlite3_fprintf(stderr,
33177	"%s: Error: no database filename specified\n", Argv0);
33178	return 1;
33179	#endif
33180	}
33181	data.out = stdout;
33182	if( bEnableVfstrace ){
33183	vfstrace_register("trace",0,vfstraceOut, &data, 1);
33184	}
33185	#ifndef SQLITE_SHELL_FIDDLE
33186	sqlite3_appendvfs_init(0,0,0);
33187	#endif
33188
33189	/* Go ahead and open the database file if it already exists. If the
	@@ -31913,11 +33292,11 @@
33292	*/
33293	ShellSetFlag(&data, SHFLG_Backslash);
33294	}else if( cli_strcmp(z,"-bail")==0 ){
33295	/* No-op. The bail_on_error flag should already be set. */
33296	}else if( cli_strcmp(z,"-version")==0 ){
33297	sqlite3_fprintf(stdout, "%s %s (%d-bit)\n",
33298	sqlite3_libversion(), sqlite3_sourceid(), 8(int)sizeof(char));
33299	return 0;
33300	}else if( cli_strcmp(z,"-interactive")==0 ){
33301	/* Need to check for interactive override here to so that it can
33302	** affect console setup (for Windows only) and testing thereof.
	@@ -31951,14 +33330,12 @@
33330	}else if( cli_strcmp(z,"-sorterref")==0 ){
33331	i++;
33332	#endif
33333	}else if( cli_strcmp(z,"-vfs")==0 ){
33334	i++;

33335	}else if( cli_strcmp(z,"-vfstrace")==0 ){
33336	i++;

33337	#ifdef SQLITE_ENABLE_MULTIPLEX
33338	}else if( cli_strcmp(z,"-multiplex")==0 ){
33339	i++;
33340	#endif
33341	}else if( cli_strcmp(z,"-help")==0 ){
	@@ -31978,18 +33355,18 @@
33355	rc = shell_exec(&data, z, &zErrMsg);
33356	if( zErrMsg!=0 ){
33357	shellEmitError(zErrMsg);
33358	if( bail_on_error ) return rc!=0 ? rc : 1;
33359	}else if( rc!=0 ){
33360	sqlite3_fprintf(stderr,"Error: unable to process SQL \"%s\"\n", z);
33361	if( bail_on_error ) return rc;
33362	}
33363	}
33364	#if !defined(SQLITE_OMIT_VIRTUALTABLE) && defined(SQLITE_HAVE_ZLIB)
33365	}else if( cli_strncmp(z, "-A", 2)==0 ){
33366	if( nCmd>0 ){
33367	sqlite3_fprintf(stderr,"Error: cannot mix regular SQL or dot-commands"
33368	" with \"%s\"\n", z);
33369	return 1;
33370	}
33371	open_db(&data, OPEN_DB_ZIPFILE);
33372	if( z[2] ){
	@@ -32004,11 +33381,11 @@
33381	}else if( cli_strcmp(z,"-safe")==0 ){
33382	data.bSafeMode = data.bSafeModePersist = 1;
33383	}else if( cli_strcmp(z,"-unsafe-testing")==0 ){
33384	/* Acted upon in first pass. */
33385	}else{
33386	sqlite3_fprintf(stderr,"%s: Error: unknown option: %s\n", Argv0, z);
33387	eputz("Use -help for a list of options.\n");
33388	return 1;
33389	}
33390	data.cMode = data.mode;
33391	}
	@@ -32031,11 +33408,12 @@
33408	rc = shell_exec(&data, azCmd[i], &zErrMsg);
33409	if( zErrMsg \|\| rc ){
33410	if( zErrMsg!=0 ){
33411	shellEmitError(zErrMsg);
33412	}else{
33413	sqlite3_fprintf(stderr,
33414	"Error: unable to process SQL: %s\n", azCmd[i]);
33415	}
33416	sqlite3_free(zErrMsg);
33417	if( rc==0 ) rc = 1;
33418	goto shell_main_exit;
33419	}
	@@ -32046,18 +33424,14 @@
33424	*/
33425	if( stdin_is_interactive ){
33426	char *zHome;
33427	char *zHistory;
33428	int nHistory;
33429	sqlite3_fprintf(stdout,
33430	"SQLite version %s %.19s\n" /extra-version-info/




33431	"Enter \".help\" for usage hints.\n",
33432	sqlite3_libversion(), sqlite3_sourceid());
33433	if( warnInmemoryDb ){
33434	sputz(stdout, "Connected to a ");
33435	printBold("transient in-memory database");
33436	sputz(stdout, ".\nUse \".open FILENAME\" to reopen on a"
33437	" persistent database.\n");
	@@ -32070,13 +33444,15 @@
33444	if( (zHistory = malloc(nHistory))!=0 ){
33445	sqlite3_snprintf(nHistory, zHistory,"%s/.sqlite_history", zHome);
33446	}
33447	}
33448	if( zHistory ){ shell_read_history(zHistory); }
33449	#if (HAVE_READLINE \|\| HAVE_EDITLINE) && !defined(SQLITE_OMIT_READLINE_COMPLETION)
33450	rl_attempted_completion_function = readline_completion;
33451	#elif HAVE_LINENOISE==1
33452	linenoiseSetCompletionCallback(linenoise_completion);
33453	#elif HAVE_LINENOISE==2
33454	linenoiseSetCompletionCallback(linenoise_completion, NULL);
33455	#endif
33456	data.in = 0;
33457	rc = process_input(&data);
33458	if( zHistory ){
	@@ -32122,13 +33498,16 @@
33498	free(data.colWidth);
33499	free(data.zNonce);
33500	/* Clear the global data structure so that valgrind will detect memory
33501	** leaks */
33502	memset(&data, 0, sizeof(data));
33503	if( bEnableVfstrace ){
33504	vfstrace_unregister("trace");
33505	}
33506	#ifdef SQLITE_DEBUG
33507	if( sqlite3_memory_used()>mem_main_enter ){
33508	sqlite3_fprintf(stderr,"Memory leaked: %u bytes\n",
33509	(unsigned int)(sqlite3_memory_used()-mem_main_enter));
33510	}
33511	#endif
33512	#else /* SQLITE_SHELL_FIDDLE... */
33513	shell_main_exit:
	@@ -32164,11 +33543,11 @@
33543	return pVfs;
33544	}
33545
33546	/* Only for emcc experimentation purposes. */
33547	sqlite3 * fiddle_db_arg(sqlite3 *arg){
33548	sqlite3_fprintf(stdout, "fiddle_db_arg(%p)\n", (const void*)arg);
33549	return arg;
33550	}
33551
33552	/*
33553	** Intended to be called via a SharedWorker() while a separate
	@@ -32201,11 +33580,11 @@
33580	while( sqlite3_txn_state(globalDb,0)>0 ){
33581	/*
33582	** Resolve problem reported in
33583	** https://sqlite.org/forum/forumpost/0b41a25d65
33584	*/
33585	sqlite3_fputs("Rolling back in-progress transaction.\n", stdout);
33586	sqlite3_exec(globalDb,"ROLLBACK", 0, 0, 0);
33587	}
33588	rc = sqlite3_db_config(globalDb, SQLITE_DBCONFIG_RESET_DATABASE, 1, 0);
33589	if( 0==rc ) sqlite3_exec(globalDb, "VACUUM", 0, 0, 0);
33590	sqlite3_db_config(globalDb, SQLITE_DBCONFIG_RESET_DATABASE, 0, 0);
33591

M extsrc/sqlite3.c

+2361 -1405

		--- extsrc/sqlite3.c
		+++ extsrc/sqlite3.c
		@@ -1,8 +1,8 @@
1	1	/******************************************************************************
2	2	** This file is an amalgamation of many separate C source files from SQLite
3		-** version 3.47.0. By combining all the individual C code files into this
	3	+** version 3.48.0. By combining all the individual C code files into this
4	4	** single large file, the entire code can be compiled as a single translation
5	5	** unit. This allows many compilers to do optimizations that would not be
6	6	** possible if the files were compiled separately. Performance improvements
7	7	** of 5% or more are commonly seen when SQLite is compiled as a single
8	8	** translation unit.
		@@ -16,12 +16,15 @@
16	16	** if you want a wrapper to interface SQLite with your choice of programming
17	17	** language. The code for the "sqlite3" command-line shell is also in a
18	18	** separate file. This file contains only code for the core SQLite library.
19	19	**
20	20	** The content in this amalgamation comes from Fossil check-in
21		-** 7891a266c4425722ae8b9231397ef9e42e24.
	21	+** e2bae4143afd07de1ae55a6d2606a3b541a5 with changes in files:
	22	+**
	23	+**
22	24	*/
	25	+#ifndef SQLITE_AMALGAMATION
23	26	#define SQLITE_CORE 1
24	27	#define SQLITE_AMALGAMATION 1
25	28	#ifndef SQLITE_PRIVATE
26	29	# define SQLITE_PRIVATE static
27	30	#endif
		@@ -460,13 +463,13 @@
460	463	**
461	464	** See also: [sqlite3_libversion()],
462	465	** [sqlite3_libversion_number()], [sqlite3_sourceid()],
463	466	** [sqlite_version()] and [sqlite_source_id()].
464	467	*/
465		-#define SQLITE_VERSION "3.47.0"
466		-#define SQLITE_VERSION_NUMBER 3047000
467		-#define SQLITE_SOURCE_ID "2024-09-02 21:59:31 7891a266c4425722ae8b9231397ef9e42e2432be9e6b70632dfaf9ff15300d2c"
	468	+#define SQLITE_VERSION "3.48.0"
	469	+#define SQLITE_VERSION_NUMBER 3048000
	470	+#define SQLITE_SOURCE_ID "2024-12-09 20:46:36 e2bae4143afd07de1ae55a6d2606a3b541a5b94568aa41f6a96e5d1245471653"
468	471
469	472	/*
470	473	** CAPI3REF: Run-Time Library Version Numbers
471	474	** KEYWORDS: sqlite3_version sqlite3_sourceid
472	475	**
		@@ -966,10 +969,17 @@
966	969	**
967	970	** The SQLITE_IOCAP_BATCH_ATOMIC property means that the underlying
968	971	** filesystem supports doing multiple write operations atomically when those
969	972	** write operations are bracketed by [SQLITE_FCNTL_BEGIN_ATOMIC_WRITE] and
970	973	** [SQLITE_FCNTL_COMMIT_ATOMIC_WRITE].
	974	+**
	975	+** The SQLITE_IOCAP_SUBPAGE_READ property means that it is ok to read
	976	+** from the database file in amounts that are not a multiple of the
	977	+** page size and that do not begin at a page boundary. Without this
	978	+** property, SQLite is careful to only do full-page reads and write
	979	+** on aligned pages, with the one exception that it will do a sub-page
	980	+** read of the first page to access the database header.
971	981	*/
972	982	#define SQLITE_IOCAP_ATOMIC 0x00000001
973	983	#define SQLITE_IOCAP_ATOMIC512 0x00000002
974	984	#define SQLITE_IOCAP_ATOMIC1K 0x00000004
975	985	#define SQLITE_IOCAP_ATOMIC2K 0x00000008
		@@ -982,10 +992,11 @@
982	992	#define SQLITE_IOCAP_SEQUENTIAL 0x00000400
983	993	#define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN 0x00000800
984	994	#define SQLITE_IOCAP_POWERSAFE_OVERWRITE 0x00001000
985	995	#define SQLITE_IOCAP_IMMUTABLE 0x00002000
986	996	#define SQLITE_IOCAP_BATCH_ATOMIC 0x00004000
	997	+#define SQLITE_IOCAP_SUBPAGE_READ 0x00008000
987	998
988	999	/*
989	1000	** CAPI3REF: File Locking Levels
990	1001	**
991	1002	** SQLite uses one of these integer values as the second
		@@ -1128,10 +1139,11 @@
1128	1139	** <li> [SQLITE_IOCAP_SEQUENTIAL]
1129	1140	** <li> [SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN]
1130	1141	** <li> [SQLITE_IOCAP_POWERSAFE_OVERWRITE]
1131	1142	** <li> [SQLITE_IOCAP_IMMUTABLE]
1132	1143	** <li> [SQLITE_IOCAP_BATCH_ATOMIC]
	1144	+** <li> [SQLITE_IOCAP_SUBPAGE_READ]
1133	1145	** </ul>
1134	1146	**
1135	1147	** The SQLITE_IOCAP_ATOMIC property means that all writes of
1136	1148	** any size are atomic. The SQLITE_IOCAP_ATOMICnnn values
1137	1149	** mean that writes of blocks that are nnn bytes in size and
		@@ -1405,10 +1417,15 @@
1405	1417	** The [SQLITE_FCNTL_WIN32_SET_HANDLE] opcode is used for debugging. This
1406	1418	** opcode causes the xFileControl method to swap the file handle with the one
1407	1419	** pointed to by the pArg argument. This capability is used during testing
1408	1420	** and only needs to be supported when SQLITE_TEST is defined.
1409	1421	**
	1422	+** <li>[[SQLITE_FCNTL_NULL_IO]]
	1423	+** The [SQLITE_FCNTL_NULL_IO] opcode sets the low-level file descriptor
	1424	+** or file handle for the [sqlite3_file] object such that it will no longer
	1425	+** read or write to the database file.
	1426	+**
1410	1427	** <li>[[SQLITE_FCNTL_WAL_BLOCK]]
1411	1428	** The [SQLITE_FCNTL_WAL_BLOCK] is a signal to the VFS layer that it might
1412	1429	** be advantageous to block on the next WAL lock if the lock is not immediately
1413	1430	** available. The WAL subsystem issues this signal during rare
1414	1431	** circumstances in order to fix a problem with priority inversion.
		@@ -1558,10 +1575,11 @@
1558	1575	#define SQLITE_FCNTL_RESERVE_BYTES 38
1559	1576	#define SQLITE_FCNTL_CKPT_START 39
1560	1577	#define SQLITE_FCNTL_EXTERNAL_READER 40
1561	1578	#define SQLITE_FCNTL_CKSM_FILE 41
1562	1579	#define SQLITE_FCNTL_RESET_CACHE 42
	1580	+#define SQLITE_FCNTL_NULL_IO 43
1563	1581
1564	1582	/* deprecated names */
1565	1583	#define SQLITE_GET_LOCKPROXYFILE SQLITE_FCNTL_GET_LOCKPROXYFILE
1566	1584	#define SQLITE_SET_LOCKPROXYFILE SQLITE_FCNTL_SET_LOCKPROXYFILE
1567	1585	#define SQLITE_LAST_ERRNO SQLITE_FCNTL_LAST_ERRNO
		@@ -2936,14 +2954,18 @@
2936	2954	**
2937	2955	** ^These functions return the number of rows modified, inserted or
2938	2956	** deleted by the most recently completed INSERT, UPDATE or DELETE
2939	2957	** statement on the database connection specified by the only parameter.
2940	2958	** The two functions are identical except for the type of the return value
2941		-** and that if the number of rows modified by the most recent INSERT, UPDATE
	2959	+** and that if the number of rows modified by the most recent INSERT, UPDATE,
2942	2960	** or DELETE is greater than the maximum value supported by type "int", then
2943	2961	** the return value of sqlite3_changes() is undefined. ^Executing any other
2944	2962	** type of SQL statement does not modify the value returned by these functions.
	2963	+** For the purposes of this interface, a CREATE TABLE AS SELECT statement
	2964	+** does not count as an INSERT, UPDATE or DELETE statement and hence the rows
	2965	+** added to the new table by the CREATE TABLE AS SELECT statement are not
	2966	+** counted.
2945	2967	**
2946	2968	** ^Only changes made directly by the INSERT, UPDATE or DELETE statement are
2947	2969	** considered - auxiliary changes caused by [CREATE TRIGGER \| triggers],
2948	2970	** [foreign key actions] or [REPLACE] constraint resolution are not counted.
2949	2971	**
		@@ -4499,15 +4521,26 @@
4499	4521	**
4500	4522	** [[SQLITE_PREPARE_NO_VTAB]] <dt>SQLITE_PREPARE_NO_VTAB</dt>
4501	4523	** <dd>The SQLITE_PREPARE_NO_VTAB flag causes the SQL compiler
4502	4524	** to return an error (error code SQLITE_ERROR) if the statement uses
4503	4525	** any virtual tables.
	4526	+**
	4527	+** [[SQLITE_PREPARE_DONT_LOG]] <dt>SQLITE_PREPARE_DONT_LOG</dt>
	4528	+** <dd>The SQLITE_PREPARE_DONT_LOG flag prevents SQL compiler
	4529	+** errors from being sent to the error log defined by
	4530	+** [SQLITE_CONFIG_LOG]. This can be used, for example, to do test
	4531	+** compiles to see if some SQL syntax is well-formed, without generating
	4532	+** messages on the global error log when it is not. If the test compile
	4533	+** fails, the sqlite3_prepare_v3() call returns the same error indications
	4534	+** with or without this flag; it just omits the call to [sqlite3_log()] that
	4535	+** logs the error.
4504	4536	** </dl>
4505	4537	*/
4506	4538	#define SQLITE_PREPARE_PERSISTENT 0x01
4507	4539	#define SQLITE_PREPARE_NORMALIZE 0x02
4508	4540	#define SQLITE_PREPARE_NO_VTAB 0x04
	4541	+#define SQLITE_PREPARE_DONT_LOG 0x10
4509	4542
4510	4543	/*
4511	4544	** CAPI3REF: Compiling An SQL Statement
4512	4545	** KEYWORDS: {SQL statement compiler}
4513	4546	** METHOD: sqlite3
		@@ -4536,17 +4569,21 @@
4536	4569	** and sqlite3_prepare_v3()
4537	4570	** interfaces use UTF-8, and sqlite3_prepare16(), sqlite3_prepare16_v2(),
4538	4571	** and sqlite3_prepare16_v3() use UTF-16.
4539	4572	**
4540	4573	** ^If the nByte argument is negative, then zSql is read up to the
4541		-** first zero terminator. ^If nByte is positive, then it is the
4542		-** number of bytes read from zSql. ^If nByte is zero, then no prepared
	4574	+** first zero terminator. ^If nByte is positive, then it is the maximum
	4575	+** number of bytes read from zSql. When nByte is positive, zSql is read
	4576	+** up to the first zero terminator or until the nByte bytes have been read,
	4577	+** whichever comes first. ^If nByte is zero, then no prepared
4543	4578	** statement is generated.
4544	4579	** If the caller knows that the supplied string is nul-terminated, then
4545	4580	** there is a small performance advantage to passing an nByte parameter that
4546	4581	** is the number of bytes in the input string <i>including</i>
4547	4582	** the nul-terminator.
	4583	+** Note that nByte measure the length of the input in bytes, not
	4584	+** characters, even for the UTF-16 interfaces.
4548	4585	**
4549	4586	** ^If pzTail is not NULL then *pzTail is made to point to the first byte
4550	4587	** past the end of the first SQL statement in zSql. These routines only
4551	4588	** compile the first statement in zSql, so *pzTail is left pointing to
4552	4589	** what remains uncompiled.
		@@ -5913,11 +5950,11 @@
5913	5950	** The SQLITE_SUBTYPE flag indicates to SQLite that a function might call
5914	5951	** [sqlite3_value_subtype()] to inspect the sub-types of its arguments.
5915	5952	** This flag instructs SQLite to omit some corner-case optimizations that
5916	5953	** might disrupt the operation of the [sqlite3_value_subtype()] function,
5917	5954	** causing it to return zero rather than the correct subtype().
5918		-** SQL functions that invokes [sqlite3_value_subtype()] should have this
	5955	+** All SQL functions that invoke [sqlite3_value_subtype()] should have this
5919	5956	** property. If the SQLITE_SUBTYPE property is omitted, then the return
5920	5957	** value from [sqlite3_value_subtype()] might sometimes be zero even though
5921	5958	** a non-zero subtype was specified by the function argument expression.
5922	5959	**
5923	5960	** [[SQLITE_RESULT_SUBTYPE]] <dt>SQLITE_RESULT_SUBTYPE</dt><dd>
		@@ -8678,11 +8715,11 @@
8678	8715	#define SQLITE_TESTCTRL_EXTRA_SCHEMA_CHECKS 29
8679	8716	#define SQLITE_TESTCTRL_SEEK_COUNT 30
8680	8717	#define SQLITE_TESTCTRL_TRACEFLAGS 31
8681	8718	#define SQLITE_TESTCTRL_TUNE 32
8682	8719	#define SQLITE_TESTCTRL_LOGEST 33
8683		-#define SQLITE_TESTCTRL_USELONGDOUBLE 34
	8720	+#define SQLITE_TESTCTRL_USELONGDOUBLE 34 /* NOT USED */
8684	8721	#define SQLITE_TESTCTRL_LAST 34 /* Largest TESTCTRL */
8685	8722
8686	8723	/*
8687	8724	** CAPI3REF: SQL Keyword Checking
8688	8725	**
		@@ -9654,10 +9691,20 @@
9654	9691	** threads may safely make multiple concurrent calls to sqlite3_backup_step().
9655	9692	** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
9656	9693	** APIs are not strictly speaking threadsafe. If they are invoked at the
9657	9694	** same time as another thread is invoking sqlite3_backup_step() it is
9658	9695	** possible that they return invalid values.
	9696	+**
	9697	+** <b>Alternatives To Using The Backup API</b>
	9698	+**
	9699	+** Other techniques for safely creating a consistent backup of an SQLite
	9700	+** database include:
	9701	+**
	9702	+** <ul>
	9703	+** <li> The [VACUUM INTO] command.
	9704	+** <li> The [sqlite3_rsync] utility program.
	9705	+** </ul>
9659	9706	*/
9660	9707	SQLITE_API sqlite3_backup *sqlite3_backup_init(
9661	9708	sqlite3 pDest, / Destination database handle */
9662	9709	const char zDestName, / Destination database name */
9663	9710	sqlite3 pSource, / Source database handle */
		@@ -10852,10 +10899,18 @@
10852	10899	** schema S in database connection D. ^On success, the
10853	10900	** [sqlite3_snapshot_get(D,S,P)] interface writes a pointer to the newly
10854	10901	** created [sqlite3_snapshot] object into *P and returns SQLITE_OK.
10855	10902	** If there is not already a read-transaction open on schema S when
10856	10903	** this function is called, one is opened automatically.
	10904	+**
	10905	+** If a read-transaction is opened by this function, then it is guaranteed
	10906	+** that the returned snapshot object may not be invalidated by a database
	10907	+** writer or checkpointer until after the read-transaction is closed. This
	10908	+** is not guaranteed if a read-transaction is already open when this
	10909	+** function is called. In that case, any subsequent write or checkpoint
	10910	+** operation on the database may invalidate the returned snapshot handle,
	10911	+** even while the read-transaction remains open.
10857	10912	**
10858	10913	** The following must be true for this function to succeed. If any of
10859	10914	** the following statements are false when sqlite3_snapshot_get() is
10860	10915	** called, SQLITE_ERROR is returned. The final value of *P is undefined
10861	10916	** in this case.
		@@ -11172,11 +11227,11 @@
11172	11227	#endif
11173	11228
11174	11229	#if 0
11175	11230	} /* End of the 'extern "C"' block */
11176	11231	#endif
11177		-#endif /* SQLITE3_H */
	11232	+/* #endif for SQLITE3_H will be added by mksqlite3.tcl */
11178	11233
11179	11234	/****** Begin file sqlite3rtree.h *******/
11180	11235	/*
11181	11236	** 2010 August 30
11182	11237	**
		@@ -13423,17 +13478,32 @@
13423	13478	** This is used to access token iToken of phrase hit iIdx within the
13424	13479	** current row. If iIdx is less than zero or greater than or equal to the
13425	13480	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
13426	13481	** output variable (*ppToken) is set to point to a buffer containing the
13427	13482	** matching document token, and (*pnToken) to the size of that buffer in
13428		-** bytes. This API is not available if the specified token matches a
13429		-** prefix query term. In that case both output variables are always set
13430		-** to 0.
	13483	+** bytes.
13431	13484	**
13432	13485	** The output text is not a copy of the document text that was tokenized.
13433	13486	** It is the output of the tokenizer module. For tokendata=1 tables, this
13434	13487	** includes any embedded 0x00 and trailing data.
	13488	+**
	13489	+** This API may be slow in some cases if the token identified by parameters
	13490	+** iIdx and iToken matched a prefix token in the query. In most cases, the
	13491	+** first call to this API for each prefix token in the query is forced
	13492	+** to scan the portion of the full-text index that matches the prefix
	13493	+** token to collect the extra data required by this API. If the prefix
	13494	+** token matches a large number of token instances in the document set,
	13495	+** this may be a performance problem.
	13496	+**
	13497	+** If the user knows in advance that a query may use this API for a
	13498	+** prefix token, FTS5 may be configured to collect all required data as part
	13499	+** of the initial querying of the full-text index, avoiding the second scan
	13500	+** entirely. This also causes prefix queries that do not use this API to
	13501	+** run more slowly and use more memory. FTS5 may be configured in this way
	13502	+** either on a per-table basis using the [FTS5 insttoken \| 'insttoken']
	13503	+** option, or on a per-query basis using the
	13504	+** [fts5_insttoken \| fts5_insttoken()] user function.
13435	13505	**
13436	13506	** This API can be quite slow if used with an FTS5 table created with the
13437	13507	** "detail=none" or "detail=column" option.
13438	13508	**
13439	13509	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
		@@ -13521,11 +13591,10 @@
13521	13591	** CUSTOM TOKENIZERS
13522	13592	**
13523	13593	** Applications may also register custom tokenizer types. A tokenizer
13524	13594	** is registered by providing fts5 with a populated instance of the
13525	13595	** following structure. All structure methods must be defined, setting
13526		-**
13527	13596	** any member of the fts5_tokenizer struct to NULL leads to undefined
13528	13597	** behaviour. The structure methods are expected to function as follows:
13529	13598	**
13530	13599	** xCreate:
13531	13600	** This function is used to allocate and initialize a tokenizer instance.
		@@ -13865,10 +13934,11 @@
13865	13934	#endif
13866	13935
13867	13936	#endif /* _FTS5_H */
13868	13937
13869	13938	/****** End of fts5.h *******/
	13939	+#endif /* SQLITE3_H */
13870	13940
13871	13941	/************ End of sqlite3.h *******************************************/
13872	13942	/************ Continuing where we left off in sqliteInt.h ****************/
13873	13943
13874	13944	/*
		@@ -13910,10 +13980,11 @@
13910	13980	** to count the size: 2^31-1 or 2147483647.
13911	13981	*/
13912	13982	#ifndef SQLITE_MAX_LENGTH
13913	13983	# define SQLITE_MAX_LENGTH 1000000000
13914	13984	#endif
	13985	+#define SQLITE_MIN_LENGTH 30 /* Minimum value for the length limit */
13915	13986
13916	13987	/*
13917	13988	** This is the maximum number of
13918	13989	**
13919	13990	** * Columns in a table
		@@ -14809,10 +14880,11 @@
14809	14880	#include <stdio.h>
14810	14881	#include <stdlib.h>
14811	14882	#include <string.h>
14812	14883	#include <assert.h>
14813	14884	#include <stddef.h>
	14885	+#include <ctype.h>
14814	14886
14815	14887	/*
14816	14888	** Use a macro to replace memcpy() if compiled with SQLITE_INLINE_MEMCPY.
14817	14889	** This allows better measurements of where memcpy() is used when running
14818	14890	** cachegrind. But this macro version of memcpy() is very slow so it
		@@ -14831,11 +14903,10 @@
14831	14903	#ifdef SQLITE_OMIT_FLOATING_POINT
14832	14904	# define double sqlite_int64
14833	14905	# define float sqlite_int64
14834	14906	# define fabs(X) ((X)<0?-(X):(X))
14835	14907	# define sqlite3IsOverflow(X) 0
14836		-# define LONGDOUBLE_TYPE sqlite_int64
14837	14908	# ifndef SQLITE_BIG_DBL
14838	14909	# define SQLITE_BIG_DBL (((sqlite3_int64)1)<<50)
14839	14910	# endif
14840	14911	# define SQLITE_OMIT_DATETIME_FUNCS 1
14841	14912	# define SQLITE_OMIT_TRACE 1
		@@ -15006,13 +15077,10 @@
15006	15077	# define INT8_TYPE int8_t
15007	15078	# else
15008	15079	# define INT8_TYPE signed char
15009	15080	# endif
15010	15081	#endif
15011		-#ifndef LONGDOUBLE_TYPE
15012		-# define LONGDOUBLE_TYPE long double
15013		-#endif
15014	15082	typedef sqlite_int64 i64; /* 8-byte signed integer */
15015	15083	typedef sqlite_uint64 u64; /* 8-byte unsigned integer */
15016	15084	typedef UINT32_TYPE u32; /* 4-byte unsigned integer */
15017	15085	typedef UINT16_TYPE u16; /* 2-byte unsigned integer */
15018	15086	typedef INT16_TYPE i16; /* 2-byte signed integer */
		@@ -16391,10 +16459,13 @@
16391	16459	struct KeyInfo, / First argument to compare function */
16392	16460	BtCursor pCursor / Space to write cursor structure */
16393	16461	);
16394	16462	SQLITE_PRIVATE BtCursor *sqlite3BtreeFakeValidCursor(void);
16395	16463	SQLITE_PRIVATE int sqlite3BtreeCursorSize(void);
	16464	+#ifdef SQLITE_DEBUG
	16465	+SQLITE_PRIVATE int sqlite3BtreeClosesWithCursor(Btree,BtCursor);
	16466	+#endif
16396	16467	SQLITE_PRIVATE void sqlite3BtreeCursorZero(BtCursor*);
16397	16468	SQLITE_PRIVATE void sqlite3BtreeCursorHintFlags(BtCursor*, unsigned);
16398	16469	#ifdef SQLITE_ENABLE_CURSOR_HINTS
16399	16470	SQLITE_PRIVATE void sqlite3BtreeCursorHint(BtCursor*, int, ...);
16400	16471	#endif
		@@ -17004,11 +17075,11 @@
17004	17075
17005	17076	/*
17006	17077	** Additional non-public SQLITE_PREPARE_* flags
17007	17078	*/
17008	17079	#define SQLITE_PREPARE_SAVESQL 0x80 /* Preserve SQL text */
17009		-#define SQLITE_PREPARE_MASK 0x0f /* Mask of public flags */
	17080	+#define SQLITE_PREPARE_MASK 0x1f /* Mask of public flags */
17010	17081
17011	17082	/*
17012	17083	** Prototypes for the VDBE interface. See comments on the implementation
17013	17084	** for a description of what each of these routines does.
17014	17085	*/
		@@ -17719,51 +17790,15 @@
17719	17790	struct FuncDefHash {
17720	17791	FuncDef a[SQLITE_FUNC_HASH_SZ]; / Hash table for functions */
17721	17792	};
17722	17793	#define SQLITE_FUNC_HASH(C,L) (((C)+(L))%SQLITE_FUNC_HASH_SZ)
17723	17794
17724		-#if defined(SQLITE_USER_AUTHENTICATION)
17725		-# warning "The SQLITE_USER_AUTHENTICATION extension is deprecated. \
17726		- See ext/userauth/user-auth.txt for details."
17727		-#endif
17728		-#ifdef SQLITE_USER_AUTHENTICATION
17729		-/*
17730		-** Information held in the "sqlite3" database connection object and used
17731		-** to manage user authentication.
17732		-*/
17733		-typedef struct sqlite3_userauth sqlite3_userauth;
17734		-struct sqlite3_userauth {
17735		- u8 authLevel; /* Current authentication level */
17736		- int nAuthPW; /* Size of the zAuthPW in bytes */
17737		- char zAuthPW; / Password used to authenticate */
17738		- char zAuthUser; / User name used to authenticate */
17739		-};
17740		-
17741		-/* Allowed values for sqlite3_userauth.authLevel */
17742		-#define UAUTH_Unknown 0 /* Authentication not yet checked */
17743		-#define UAUTH_Fail 1 /* User authentication failed */
17744		-#define UAUTH_User 2 /* Authenticated as a normal user */
17745		-#define UAUTH_Admin 3 /* Authenticated as an administrator */
17746		-
17747		-/* Functions used only by user authorization logic */
17748		-SQLITE_PRIVATE int sqlite3UserAuthTable(const char*);
17749		-SQLITE_PRIVATE int sqlite3UserAuthCheckLogin(sqlite3,const char,u8*);
17750		-SQLITE_PRIVATE void sqlite3UserAuthInit(sqlite3*);
17751		-SQLITE_PRIVATE void sqlite3CryptFunc(sqlite3_context,int,sqlite3_value*);
17752		-
17753		-#endif /* SQLITE_USER_AUTHENTICATION */
17754		-
17755	17795	/*
17756	17796	** typedef for the authorization callback function.
17757	17797	*/
17758		-#ifdef SQLITE_USER_AUTHENTICATION
17759		- typedef int (sqlite3_xauth)(void,int,const char,const char,const char*,
17760		- const char, const char);
17761		-#else
17762		- typedef int (sqlite3_xauth)(void,int,const char,const char,const char*,
17763		- const char*);
17764		-#endif
	17798	+typedef int (sqlite3_xauth)(void,int,const char,const char,const char*,
	17799	+ const char*);
17765	17800
17766	17801	#ifndef SQLITE_OMIT_DEPRECATED
17767	17802	/* This is an extra SQLITE_TRACE macro that indicates "legacy" tracing
17768	17803	** in the style of sqlite3_trace()
17769	17804	*/
		@@ -17920,13 +17955,10 @@
17920	17955	sqlite3 pUnlockConnection; / Connection to watch for unlock */
17921	17956	void pUnlockArg; / Argument to xUnlockNotify */
17922	17957	void (xUnlockNotify)(void , int); / Unlock notify callback */
17923	17958	sqlite3 pNextBlocked; / Next in list of all blocked connections */
17924	17959	#endif
17925		-#ifdef SQLITE_USER_AUTHENTICATION
17926		- sqlite3_userauth auth; /* User authentication information */
17927		-#endif
17928	17960	};
17929	17961
17930	17962	/*
17931	17963	** A macro to discover the encoding of a database.
17932	17964	*/
		@@ -19221,11 +19253,11 @@
19221	19253	#define EP_Quoted 0x4000000 /* TK_ID was originally quoted */
19222	19254	#define EP_Static 0x8000000 /* Held in memory not obtained from malloc() */
19223	19255	#define EP_IsTrue 0x10000000 /* Always has boolean value of TRUE */
19224	19256	#define EP_IsFalse 0x20000000 /* Always has boolean value of FALSE */
19225	19257	#define EP_FromDDL 0x40000000 /* Originates from sqlite_schema */
19226		- /* 0x80000000 // Available */
	19258	+#define EP_SubtArg 0x80000000 /* Is argument to SQLITE_SUBTYPE function */
19227	19259
19228	19260	/* The EP_Propagate mask is a set of properties that automatically propagate
19229	19261	** upwards into parent nodes.
19230	19262	*/
19231	19263	#define EP_Propagate (EP_Collate\|EP_Subquery\|EP_HasFunc)
		@@ -19777,11 +19809,11 @@
19777	19809	**
19778	19810	** SRT_Set The result must be a single column. Store each
19779	19811	** row of result as the key in table pDest->iSDParm.
19780	19812	** Apply the affinity pDest->affSdst before storing
19781	19813	** results. if pDest->iSDParm2 is positive, then it is
19782		-** a regsiter holding a Bloom filter for the IN operator
	19814	+** a register holding a Bloom filter for the IN operator
19783	19815	** that should be populated in addition to the
19784	19816	** pDest->iSDParm table. This SRT is used to
19785	19817	** implement "IN (SELECT ...)".
19786	19818	**
19787	19819	** SRT_EphemTab Create an temporary table pDest->iSDParm and store
		@@ -20376,11 +20408,10 @@
20376	20408	u8 bFullMutex; /* True to enable full mutexing */
20377	20409	u8 bOpenUri; /* True to interpret filenames as URIs */
20378	20410	u8 bUseCis; /* Use covering indices for full-scans */
20379	20411	u8 bSmallMalloc; /* Avoid large memory allocations if true */
20380	20412	u8 bExtraSchemaChecks; /* Verify type,name,tbl_name in schema */
20381		- u8 bUseLongDouble; /* Make use of long double */
20382	20413	#ifdef SQLITE_DEBUG
20383	20414	u8 bJsonSelfcheck; /* Double-check JSON parsing */
20384	20415	#endif
20385	20416	int mxStrlen; /* Maximum string length */
20386	20417	int neverCorrupt; /* Database is always well-formed */
		@@ -20751,19 +20782,10 @@
20751	20782	*/
20752	20783	#if defined(SQLITE_ENABLE_FTS4) && !defined(SQLITE_ENABLE_FTS3)
20753	20784	# define SQLITE_ENABLE_FTS3 1
20754	20785	#endif
20755	20786
20756		-/*
20757		-** The ctype.h header is needed for non-ASCII systems. It is also
20758		-** needed by FTS3 when FTS3 is included in the amalgamation.
20759		-*/
20760		-#if !defined(SQLITE_ASCII) \|\| \
20761		- (defined(SQLITE_ENABLE_FTS3) && defined(SQLITE_AMALGAMATION))
20762		-# include <ctype.h>
20763		-#endif
20764		-
20765	20787	/*
20766	20788	** The following macros mimic the standard library functions toupper(),
20767	20789	** isspace(), isalnum(), isdigit() and isxdigit(), respectively. The
20768	20790	** sqlite versions only work for ASCII characters, regardless of locale.
20769	20791	*/
		@@ -21381,11 +21403,11 @@
21381	21403	SQLITE_PRIVATE int sqlite3AtoF(const char z, double, int, u8);
21382	21404	SQLITE_PRIVATE int sqlite3GetInt32(const char , int);
21383	21405	SQLITE_PRIVATE int sqlite3GetUInt32(const char, u32);
21384	21406	SQLITE_PRIVATE int sqlite3Atoi(const char*);
21385	21407	#ifndef SQLITE_OMIT_UTF16
21386		-SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *pData, int nChar);
	21408	+SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *pData, int nByte, int nChar);
21387	21409	#endif
21388	21410	SQLITE_PRIVATE int sqlite3Utf8CharLen(const char *pData, int nByte);
21389	21411	SQLITE_PRIVATE u32 sqlite3Utf8Read(const u8**);
21390	21412	SQLITE_PRIVATE int sqlite3Utf8ReadLimited(const u8, int, u32);
21391	21413	SQLITE_PRIVATE LogEst sqlite3LogEst(u64);
		@@ -22839,13 +22861,10 @@
22839	22861	"UNLINK_AFTER_CLOSE",
22840	22862	#endif
22841	22863	#ifdef SQLITE_UNTESTABLE
22842	22864	"UNTESTABLE",
22843	22865	#endif
22844		-#ifdef SQLITE_USER_AUTHENTICATION
22845		- "USER_AUTHENTICATION",
22846		-#endif
22847	22866	#ifdef SQLITE_USE_ALLOCA
22848	22867	"USE_ALLOCA",
22849	22868	#endif
22850	22869	#ifdef SQLITE_USE_FCNTL_TRACE
22851	22870	"USE_FCNTL_TRACE",
		@@ -23117,11 +23136,10 @@
23117	23136	SQLITE_THREADSAFE==1, /* bFullMutex */
23118	23137	SQLITE_USE_URI, /* bOpenUri */
23119	23138	SQLITE_ALLOW_COVERING_INDEX_SCAN, /* bUseCis */
23120	23139	0, /* bSmallMalloc */
23121	23140	1, /* bExtraSchemaChecks */
23122		- sizeof(LONGDOUBLE_TYPE)>8, /* bUseLongDouble */
23123	23141	#ifdef SQLITE_DEBUG
23124	23142	0, /* bJsonSelfcheck */
23125	23143	#endif
23126	23144	0x7ffffffe, /* mxStrlen */
23127	23145	0, /* neverCorrupt */
		@@ -23837,13 +23855,15 @@
23837	23855	UnpackedRecord pNewUnpacked; / Unpacked version of new.* record */
23838	23856	int iNewReg; /* Register for new.* values */
23839	23857	int iBlobWrite; /* Value returned by preupdate_blobwrite() */
23840	23858	i64 iKey1; /* First key value passed to hook */
23841	23859	i64 iKey2; /* Second key value passed to hook */
	23860	+ Mem oldipk; /* Memory cell holding "old" IPK value */
23842	23861	Mem aNew; / Array of new.* values */
23843	23862	Table pTab; / Schema object being updated */
23844	23863	Index pPk; / PK index if pTab is WITHOUT ROWID */
	23864	+ sqlite3_value *apDflt; / Array of default values, if required */
23845	23865	};
23846	23866
23847	23867	/*
23848	23868	** An instance of this object is used to pass an vector of values into
23849	23869	** OP_VFilter, the xFilter method of a virtual table. The vector is the
		@@ -29285,20 +29305,33 @@
29285	29305
29286	29306	#ifndef NDEBUG
29287	29307	/*
29288	29308	** The sqlite3_mutex_held() and sqlite3_mutex_notheld() routine are
29289	29309	** intended for use inside assert() statements.
	29310	+**
	29311	+** Because these routines raise false-positive alerts in TSAN, disable
	29312	+** them (make them always return 1) when compiling with TSAN.
29290	29313	*/
29291	29314	SQLITE_API int sqlite3_mutex_held(sqlite3_mutex *p){
	29315	+# if defined(__has_feature)
	29316	+# if __has_feature(thread_sanitizer)
	29317	+ p = 0;
	29318	+# endif
	29319	+# endif
29292	29320	assert( p==0 \|\| sqlite3GlobalConfig.mutex.xMutexHeld );
29293	29321	return p==0 \|\| sqlite3GlobalConfig.mutex.xMutexHeld(p);
29294	29322	}
29295	29323	SQLITE_API int sqlite3_mutex_notheld(sqlite3_mutex *p){
	29324	+# if defined(__has_feature)
	29325	+# if __has_feature(thread_sanitizer)
	29326	+ p = 0;
	29327	+# endif
	29328	+# endif
29296	29329	assert( p==0 \|\| sqlite3GlobalConfig.mutex.xMutexNotheld );
29297	29330	return p==0 \|\| sqlite3GlobalConfig.mutex.xMutexNotheld(p);
29298	29331	}
29299		-#endif
	29332	+#endif /* NDEBUG */
29300	29333
29301	29334	#endif /* !defined(SQLITE_MUTEX_OMIT) */
29302	29335
29303	29336	/************ End of mutex.c *********************************************/
29304	29337	/************ Begin file mutex_noop.c ************************************/
		@@ -32272,10 +32305,11 @@
32272	32305	&& (ExprHasProperty(pExpr,EP_OuterON\|EP_InnerON) \|\| pExpr->w.iOfst<=0)
32273	32306	){
32274	32307	pExpr = pExpr->pLeft;
32275	32308	}
32276	32309	if( pExpr==0 ) return;
	32310	+ if( ExprHasProperty(pExpr, EP_FromDDL) ) return;
32277	32311	db->errByteOffset = pExpr->w.iOfst;
32278	32312	}
32279	32313
32280	32314	/*
32281	32315	** Enlarge the memory allocation on a StrAccum object so that it is
		@@ -33001,11 +33035,11 @@
33001	33035	}
33002	33036	if( pItem->fg.isCte ){
33003	33037	sqlite3_str_appendf(&x, " CteUse=0x%p", pItem->u2.pCteUse);
33004	33038	}
33005	33039	if( pItem->fg.isOn \|\| (pItem->fg.isUsing==0 && pItem->u3.pOn!=0) ){
33006		- sqlite3_str_appendf(&x, " ON");
	33040	+ sqlite3_str_appendf(&x, " isOn");
33007	33041	}
33008	33042	if( pItem->fg.isTabFunc ) sqlite3_str_appendf(&x, " isTabFunc");
33009	33043	if( pItem->fg.isCorrelated ) sqlite3_str_appendf(&x, " isCorrelated");
33010	33044	if( pItem->fg.isMaterialized ) sqlite3_str_appendf(&x, " isMaterialized");
33011	33045	if( pItem->fg.viaCoroutine ) sqlite3_str_appendf(&x, " viaCoroutine");
		@@ -34085,10 +34119,14 @@
34085	34119	**
34086	34120	** This routines are given external linkage so that they will always be
34087	34121	** accessible to the debugging, and to avoid warnings about unused
34088	34122	** functions. But these routines only exist in debugging builds, so they
34089	34123	** do not contaminate the interface.
	34124	+**
	34125	+** See Also:
	34126	+**
	34127	+** sqlite3ShowWhereTerm() in where.c
34090	34128	*/
34091	34129	SQLITE_PRIVATE void sqlite3ShowExpr(const Expr *p){ sqlite3TreeViewExpr(0,p,0); }
34092	34130	SQLITE_PRIVATE void sqlite3ShowExprList(const ExprList *p){ sqlite3TreeViewExprList(0,p,0,0);}
34093	34131	SQLITE_PRIVATE void sqlite3ShowIdList(const IdList *p){ sqlite3TreeViewIdList(0,p,0,0); }
34094	34132	SQLITE_PRIVATE void sqlite3ShowSrcList(const SrcList *p){ sqlite3TreeViewSrcList(0,p); }
		@@ -34687,11 +34725,11 @@
34687	34725	*/
34688	34726	#define READ_UTF8(zIn, zTerm, c) \
34689	34727	c = *(zIn++); \
34690	34728	if( c>=0xc0 ){ \
34691	34729	c = sqlite3Utf8Trans1[c-0xc0]; \
34692		- while( zIn!=zTerm && (*zIn & 0xc0)==0x80 ){ \
	34730	+ while( zIn<zTerm && (*zIn & 0xc0)==0x80 ){ \
34693	34731	c = (c<<6) + (0x3f & *(zIn++)); \
34694	34732	} \
34695	34733	if( c<0x80 \
34696	34734	\|\| (c&0xFFFFF800)==0xD800 \
34697	34735	\|\| (c&0xFFFFFFFE)==0xFFFE ){ c = 0xFFFD; } \
		@@ -35065,24 +35103,26 @@
35065	35103	assert( m.z \|\| db->mallocFailed );
35066	35104	return m.z;
35067	35105	}
35068	35106
35069	35107	/*
35070		-** zIn is a UTF-16 encoded unicode string at least nChar characters long.
	35108	+** zIn is a UTF-16 encoded unicode string at least nByte bytes long.
35071	35109	** Return the number of bytes in the first nChar unicode characters
35072		-** in pZ. nChar must be non-negative.
	35110	+** in pZ. nChar must be non-negative. Surrogate pairs count as a single
	35111	+** character.
35073	35112	*/
35074		-SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *zIn, int nChar){
	35113	+SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *zIn, int nByte, int nChar){
35075	35114	int c;
35076	35115	unsigned char const *z = zIn;
	35116	+ unsigned char const *zEnd = &z[nByte-1];
35077	35117	int n = 0;
35078	35118
35079	35119	if( SQLITE_UTF16NATIVE==SQLITE_UTF16LE ) z++;
35080		- while( n<nChar ){
	35120	+ while( n<nChar && ALWAYS(z<=zEnd) ){
35081	35121	c = z[0];
35082	35122	z += 2;
35083		- if( c>=0xd8 && c<0xdc && z[0]>=0xdc && z[0]<0xe0 ) z += 2;
	35123	+ if( c>=0xd8 && c<0xdc && z<=zEnd && z[0]>=0xdc && z[0]<0xe0 ) z += 2;
35084	35124	n++;
35085	35125	}
35086	35126	return (int)(z-(unsigned char const *)zIn)
35087	35127	- (SQLITE_UTF16NATIVE==SQLITE_UTF16LE);
35088	35128	}
		@@ -35659,10 +35699,12 @@
35659	35699	int esign = 1; /* sign of exponent */
35660	35700	int e = 0; /* exponent */
35661	35701	int eValid = 1; /* True exponent is either not used or is well-formed */
35662	35702	int nDigit = 0; /* Number of digits processed */
35663	35703	int eType = 1; /* 1: pure integer, 2+: fractional -1 or less: bad UTF16 */
	35704	+ u64 s2; /* round-tripped significand */
	35705	+ double rr[2];
35664	35706
35665	35707	assert( enc==SQLITE_UTF8 \|\| enc==SQLITE_UTF16LE \|\| enc==SQLITE_UTF16BE );
35666	35708	pResult = 0.0; / Default return value, in case of an error */
35667	35709	if( length==0 ) return 0;
35668	35710
		@@ -35761,81 +35803,65 @@
35761	35803
35762	35804	/* adjust exponent by d, and update sign */
35763	35805	e = (e*esign) + d;
35764	35806
35765	35807	/* Try to adjust the exponent to make it smaller */
35766		- while( e>0 && s<(LARGEST_UINT64/10) ){
	35808	+ while( e>0 && s<((LARGEST_UINT64-0x7ff)/10) ){
35767	35809	s *= 10;
35768	35810	e--;
35769	35811	}
35770	35812	while( e<0 && (s%10)==0 ){
35771	35813	s /= 10;
35772	35814	e++;
35773	35815	}
35774	35816
35775		- if( e==0 ){
35776		- *pResult = s;
35777		- }else if( sqlite3Config.bUseLongDouble ){
35778		- LONGDOUBLE_TYPE r = (LONGDOUBLE_TYPE)s;
35779		- if( e>0 ){
35780		- while( e>=100 ){ e-=100; r *= 1.0e+100L; }
35781		- while( e>=10 ){ e-=10; r *= 1.0e+10L; }
35782		- while( e>=1 ){ e-=1; r *= 1.0e+01L; }
35783		- }else{
35784		- while( e<=-100 ){ e+=100; r *= 1.0e-100L; }
35785		- while( e<=-10 ){ e+=10; r *= 1.0e-10L; }
35786		- while( e<=-1 ){ e+=1; r *= 1.0e-01L; }
35787		- }
35788		- assert( r>=0.0 );
35789		- if( r>+1.7976931348623157081452742373e+308L ){
35790		-#ifdef INFINITY
35791		- *pResult = +INFINITY;
35792		-#else
35793		- pResult = 1.0e30810.0;
35794		-#endif
35795		- }else{
35796		- *pResult = (double)r;
35797		- }
35798		- }else{
35799		- double rr[2];
35800		- u64 s2;
35801		- rr[0] = (double)s;
35802		- s2 = (u64)rr[0];
35803		-#if defined(_MSC_VER) && _MSC_VER<1700
35804		- if( s2==0x8000000000000000LL ){ s2 = 2(u64)(0.5rr[0]); }
35805		-#endif
	35817	+ rr[0] = (double)s;
	35818	+ assert( sizeof(s2)==sizeof(rr[0]) );
	35819	+#ifdef SQLITE_DEBUG
	35820	+ rr[1] = 18446744073709549568.0;
	35821	+ memcpy(&s2, &rr[1], sizeof(s2));
	35822	+ assert( s2==0x43efffffffffffffLL );
	35823	+#endif
	35824	+ /* Largest double that can be safely converted to u64
	35825	+ ** vvvvvvvvvvvvvvvvvvvvvv */
	35826	+ if( rr[0]<=18446744073709549568.0 ){
	35827	+ s2 = (u64)rr[0];
35806	35828	rr[1] = s>=s2 ? (double)(s - s2) : -(double)(s2 - s);
35807		- if( e>0 ){
35808		- while( e>=100 ){
35809		- e -= 100;
35810		- dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
35811		- }
35812		- while( e>=10 ){
35813		- e -= 10;
35814		- dekkerMul2(rr, 1.0e+10, 0.0);
35815		- }
35816		- while( e>=1 ){
35817		- e -= 1;
35818		- dekkerMul2(rr, 1.0e+01, 0.0);
35819		- }
35820		- }else{
35821		- while( e<=-100 ){
35822		- e += 100;
35823		- dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
35824		- }
35825		- while( e<=-10 ){
35826		- e += 10;
35827		- dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
35828		- }
35829		- while( e<=-1 ){
35830		- e += 1;
35831		- dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
35832		- }
35833		- }
35834		- *pResult = rr[0]+rr[1];
35835		- if( sqlite3IsNaN(pResult) ) pResult = 1e300*1e300;
35836		- }
	35829	+ }else{
	35830	+ rr[1] = 0.0;
	35831	+ }
	35832	+ assert( rr[1]<=1.0e-10rr[0] ); / Equal only when rr[0]==0.0 */
	35833	+
	35834	+ if( e>0 ){
	35835	+ while( e>=100 ){
	35836	+ e -= 100;
	35837	+ dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
	35838	+ }
	35839	+ while( e>=10 ){
	35840	+ e -= 10;
	35841	+ dekkerMul2(rr, 1.0e+10, 0.0);
	35842	+ }
	35843	+ while( e>=1 ){
	35844	+ e -= 1;
	35845	+ dekkerMul2(rr, 1.0e+01, 0.0);
	35846	+ }
	35847	+ }else{
	35848	+ while( e<=-100 ){
	35849	+ e += 100;
	35850	+ dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
	35851	+ }
	35852	+ while( e<=-10 ){
	35853	+ e += 10;
	35854	+ dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
	35855	+ }
	35856	+ while( e<=-1 ){
	35857	+ e += 1;
	35858	+ dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
	35859	+ }
	35860	+ }
	35861	+ *pResult = rr[0]+rr[1];
	35862	+ if( sqlite3IsNaN(pResult) ) pResult = 1e300*1e300;
35837	35863	if( sign<0 ) pResult = -pResult;
35838	35864	assert( !sqlite3IsNaN(*pResult) );
35839	35865
35840	35866	atof_return:
35841	35867	/* return true if number and no extra non-whitespace characters after */
		@@ -36152,13 +36178,14 @@
36152	36178	*/
36153	36179	SQLITE_PRIVATE void sqlite3FpDecode(FpDecode *p, double r, int iRound, int mxRound){
36154	36180	int i;
36155	36181	u64 v;
36156	36182	int e, exp = 0;
	36183	+ double rr[2];
	36184	+
36157	36185	p->isSpecial = 0;
36158	36186	p->z = p->zBuf;
36159		-
36160	36187	assert( mxRound>0 );
36161	36188
36162	36189	/* Convert negative numbers to positive. Deal with Infinity, 0.0, and
36163	36190	** NaN. */
36164	36191	if( r<0.0 ){
		@@ -36182,66 +36209,49 @@
36182	36209	return;
36183	36210	}
36184	36211
36185	36212	/* Multiply r by powers of ten until it lands somewhere in between
36186	36213	** 1.0e+19 and 1.0e+17.
36187		- */
36188		- if( sqlite3Config.bUseLongDouble ){
36189		- LONGDOUBLE_TYPE rr = r;
36190		- if( rr>=1.0e+19 ){
36191		- while( rr>=1.0e+119L ){ exp+=100; rr *= 1.0e-100L; }
36192		- while( rr>=1.0e+29L ){ exp+=10; rr *= 1.0e-10L; }
36193		- while( rr>=1.0e+19L ){ exp++; rr *= 1.0e-1L; }
36194		- }else{
36195		- while( rr<1.0e-97L ){ exp-=100; rr *= 1.0e+100L; }
36196		- while( rr<1.0e+07L ){ exp-=10; rr *= 1.0e+10L; }
36197		- while( rr<1.0e+17L ){ exp--; rr *= 1.0e+1L; }
36198		- }
36199		- v = (u64)rr;
36200		- }else{
36201		- /* If high-precision floating point is not available using "long double",
36202		- ** then use Dekker-style double-double computation to increase the
36203		- ** precision.
36204		- **
36205		- ** The error terms on constants like 1.0e+100 computed using the
36206		- ** decimal extension, for example as follows:
36207		- **
36208		- ** SELECT decimal_exp(decimal_sub('1.0e+100',decimal(1.0e+100)));
36209		- */
36210		- double rr[2];
36211		- rr[0] = r;
36212		- rr[1] = 0.0;
36213		- if( rr[0]>9.223372036854774784e+18 ){
36214		- while( rr[0]>9.223372036854774784e+118 ){
36215		- exp += 100;
36216		- dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
36217		- }
36218		- while( rr[0]>9.223372036854774784e+28 ){
36219		- exp += 10;
36220		- dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
36221		- }
36222		- while( rr[0]>9.223372036854774784e+18 ){
36223		- exp += 1;
36224		- dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
36225		- }
36226		- }else{
36227		- while( rr[0]<9.223372036854774784e-83 ){
36228		- exp -= 100;
36229		- dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
36230		- }
36231		- while( rr[0]<9.223372036854774784e+07 ){
36232		- exp -= 10;
36233		- dekkerMul2(rr, 1.0e+10, 0.0);
36234		- }
36235		- while( rr[0]<9.22337203685477478e+17 ){
36236		- exp -= 1;
36237		- dekkerMul2(rr, 1.0e+01, 0.0);
36238		- }
36239		- }
36240		- v = rr[1]<0.0 ? (u64)rr[0]-(u64)(-rr[1]) : (u64)rr[0]+(u64)rr[1];
36241		- }
36242		-
	36214	+ **
	36215	+ ** Use Dekker-style double-double computation to increase the
	36216	+ ** precision.
	36217	+ **
	36218	+ ** The error terms on constants like 1.0e+100 computed using the
	36219	+ ** decimal extension, for example as follows:
	36220	+ **
	36221	+ ** SELECT decimal_exp(decimal_sub('1.0e+100',decimal(1.0e+100)));
	36222	+ */
	36223	+ rr[0] = r;
	36224	+ rr[1] = 0.0;
	36225	+ if( rr[0]>9.223372036854774784e+18 ){
	36226	+ while( rr[0]>9.223372036854774784e+118 ){
	36227	+ exp += 100;
	36228	+ dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
	36229	+ }
	36230	+ while( rr[0]>9.223372036854774784e+28 ){
	36231	+ exp += 10;
	36232	+ dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
	36233	+ }
	36234	+ while( rr[0]>9.223372036854774784e+18 ){
	36235	+ exp += 1;
	36236	+ dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
	36237	+ }
	36238	+ }else{
	36239	+ while( rr[0]<9.223372036854774784e-83 ){
	36240	+ exp -= 100;
	36241	+ dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
	36242	+ }
	36243	+ while( rr[0]<9.223372036854774784e+07 ){
	36244	+ exp -= 10;
	36245	+ dekkerMul2(rr, 1.0e+10, 0.0);
	36246	+ }
	36247	+ while( rr[0]<9.22337203685477478e+17 ){
	36248	+ exp -= 1;
	36249	+ dekkerMul2(rr, 1.0e+01, 0.0);
	36250	+ }
	36251	+ }
	36252	+ v = rr[1]<0.0 ? (u64)rr[0]-(u64)(-rr[1]) : (u64)rr[0]+(u64)rr[1];
36243	36253
36244	36254	/* Extract significant digits. */
36245	36255	i = sizeof(p->zBuf)-1;
36246	36256	assert( v>0 );
36247	36257	while( v ){ p->zBuf[i--] = (v%10) + '0'; v /= 10; }
		@@ -37008,108 +37018,10 @@
37008	37018	i += pIn[i+1];
37009	37019	}while( i<mx );
37010	37020	return 0;
37011	37021	}
37012	37022
37013		-/*
37014		-** High-resolution hardware timer used for debugging and testing only.
37015		-*/
37016		-#if defined(VDBE_PROFILE) \
37017		- \|\| defined(SQLITE_PERFORMANCE_TRACE) \
37018		- \|\| defined(SQLITE_ENABLE_STMT_SCANSTATUS)
37019		-/************ Include hwtime.h in the middle of util.c *******************/
37020		-/************ Begin file hwtime.h ****************************************/
37021		-/*
37022		-** 2008 May 27
37023		-**
37024		-** The author disclaims copyright to this source code. In place of
37025		-** a legal notice, here is a blessing:
37026		-**
37027		-** May you do good and not evil.
37028		-** May you find forgiveness for yourself and forgive others.
37029		-** May you share freely, never taking more than you give.
37030		-**
37031		-******************************************************************************
37032		-**
37033		-** This file contains inline asm code for retrieving "high-performance"
37034		-** counters for x86 and x86_64 class CPUs.
37035		-*/
37036		-#ifndef SQLITE_HWTIME_H
37037		-#define SQLITE_HWTIME_H
37038		-
37039		-/*
37040		-** The following routine only works on Pentium-class (or newer) processors.
37041		-** It uses the RDTSC opcode to read the cycle count value out of the
37042		-** processor and returns that value. This can be used for high-res
37043		-** profiling.
37044		-*/
37045		-#if !defined(__STRICT_ANSI__) && \
37046		- (defined(__GNUC__) \|\| defined(_MSC_VER)) && \
37047		- (defined(i386) \|\| defined(__i386__) \|\| defined(_M_IX86))
37048		-
37049		- #if defined(__GNUC__)
37050		-
37051		- __inline__ sqlite_uint64 sqlite3Hwtime(void){
37052		- unsigned int lo, hi;
37053		- __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
37054		- return (sqlite_uint64)hi << 32 \| lo;
37055		- }
37056		-
37057		- #elif defined(_MSC_VER)
37058		-
37059		- __declspec(naked) __inline sqlite_uint64 __cdecl sqlite3Hwtime(void){
37060		- __asm {
37061		- rdtsc
37062		- ret ; return value at EDX:EAX
37063		- }
37064		- }
37065		-
37066		- #endif
37067		-
37068		-#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__x86_64__))
37069		-
37070		- __inline__ sqlite_uint64 sqlite3Hwtime(void){
37071		- unsigned int lo, hi;
37072		- __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
37073		- return (sqlite_uint64)hi << 32 \| lo;
37074		- }
37075		-
37076		-#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__ppc__))
37077		-
37078		- __inline__ sqlite_uint64 sqlite3Hwtime(void){
37079		- unsigned long long retval;
37080		- unsigned long junk;
37081		- __asm__ __volatile__ ("\n\
37082		- 1: mftbu %1\n\
37083		- mftb %L0\n\
37084		- mftbu %0\n\
37085		- cmpw %0,%1\n\
37086		- bne 1b"
37087		- : "=r" (retval), "=r" (junk));
37088		- return retval;
37089		- }
37090		-
37091		-#else
37092		-
37093		- /*
37094		- ** asm() is needed for hardware timing support. Without asm(),
37095		- ** disable the sqlite3Hwtime() routine.
37096		- **
37097		- ** sqlite3Hwtime() is only used for some obscure debugging
37098		- ** and analysis configurations, not in any deliverable, so this
37099		- ** should not be a great loss.
37100		- */
37101		-SQLITE_PRIVATE sqlite_uint64 sqlite3Hwtime(void){ return ((sqlite_uint64)0); }
37102		-
37103		-#endif
37104		-
37105		-#endif /* !defined(SQLITE_HWTIME_H) */
37106		-
37107		-/************ End of hwtime.h ********************************************/
37108		-/************ Continuing where we left off in util.c *********************/
37109		-#endif
37110		-
37111	37023	/************ End of util.c **********************************************/
37112	37024	/************ Begin file hash.c ******************************************/
37113	37025	/*
37114	37026	** 2001 September 22
37115	37027	**
		@@ -38787,11 +38699,11 @@
38787	38699	# define F_SETLKW 7
38788	38700	# endif
38789	38701	# endif
38790	38702	#else /* !SQLITE_WASI */
38791	38703	# ifndef HAVE_FCHMOD
38792		-# define HAVE_FCHMOD
	38704	+# define HAVE_FCHMOD 1
38793	38705	# endif
38794	38706	#endif /* SQLITE_WASI */
38795	38707
38796	38708	#ifdef SQLITE_WASI
38797	38709	# define osGetpid(X) (pid_t)1
		@@ -41038,58 +40950,37 @@
41038	40950	** file by this or any other process. If such a lock is held, set *pResOut
41039	40951	** to a non-zero value otherwise *pResOut is set to zero. The return value
41040	40952	** is set to SQLITE_OK unless an I/O error occurs during lock checking.
41041	40953	*/
41042	40954	static int flockCheckReservedLock(sqlite3_file id, int pResOut){
41043		- int rc = SQLITE_OK;
41044		- int reserved = 0;
	40955	+#ifdef SQLITE_DEBUG
41045	40956	unixFile pFile = (unixFile)id;
	40957	+#else
	40958	+ UNUSED_PARAMETER(id);
	40959	+#endif
41046	40960
41047	40961	SimulateIOError( return SQLITE_IOERR_CHECKRESERVEDLOCK; );
41048	40962
41049	40963	assert( pFile );
41050		-
41051		- /* Check if a thread in this process holds such a lock */
41052		- if( pFile->eFileLock>SHARED_LOCK ){
41053		- reserved = 1;
41054		- }
41055		-
41056		- /* Otherwise see if some other process holds it. */
41057		- if( !reserved ){
41058		- /* attempt to get the lock */
41059		- int lrc = robust_flock(pFile->h, LOCK_EX \| LOCK_NB);
41060		- if( !lrc ){
41061		- /* got the lock, unlock it */
41062		- lrc = robust_flock(pFile->h, LOCK_UN);
41063		- if ( lrc ) {
41064		- int tErrno = errno;
41065		- /* unlock failed with an error */
41066		- lrc = SQLITE_IOERR_UNLOCK;
41067		- storeLastErrno(pFile, tErrno);
41068		- rc = lrc;
41069		- }
41070		- } else {
41071		- int tErrno = errno;
41072		- reserved = 1;
41073		- /* someone else might have it reserved */
41074		- lrc = sqliteErrorFromPosixError(tErrno, SQLITE_IOERR_LOCK);
41075		- if( IS_LOCK_ERROR(lrc) ){
41076		- storeLastErrno(pFile, tErrno);
41077		- rc = lrc;
41078		- }
41079		- }
41080		- }
41081		- OSTRACE(("TEST WR-LOCK %d %d %d (flock)\n", pFile->h, rc, reserved));
41082		-
41083		-#ifdef SQLITE_IGNORE_FLOCK_LOCK_ERRORS
41084		- if( (rc & 0xff) == SQLITE_IOERR ){
41085		- rc = SQLITE_OK;
41086		- reserved=1;
41087		- }
41088		-#endif /* SQLITE_IGNORE_FLOCK_LOCK_ERRORS */
41089		- *pResOut = reserved;
41090		- return rc;
	40964	+ assert( pFile->eFileLock<=SHARED_LOCK );
	40965	+
	40966	+ /* The flock VFS only ever takes exclusive locks (see function flockLock).
	40967	+ ** Therefore, if this connection is holding any lock at all, no other
	40968	+ ** connection may be holding a RESERVED lock. So set *pResOut to 0
	40969	+ ** in this case.
	40970	+ **
	40971	+ ** Or, this connection may be holding no lock. In that case, set *pResOut to
	40972	+ ** 0 as well. The caller will then attempt to take an EXCLUSIVE lock on the
	40973	+ ** db in order to roll the hot journal back. If there is another connection
	40974	+ ** holding a lock, that attempt will fail and an SQLITE_BUSY returned to
	40975	+ ** the user. With other VFS, we try to avoid this, in order to allow a reader
	40976	+ ** to proceed while a writer is preparing its transaction. But that won't
	40977	+ ** work with the flock VFS - as it always takes EXCLUSIVE locks - so it is
	40978	+ ** not a problem in this case. */
	40979	+ *pResOut = 0;
	40980	+
	40981	+ return SQLITE_OK;
41091	40982	}
41092	40983
41093	40984	/*
41094	40985	** Lock the file with the lock specified by parameter eFileLock - one
41095	40986	** of the following:
		@@ -42582,10 +42473,15 @@
42582	42473	int rc = osIoctl(pFile->h, F2FS_IOC_ABORT_VOLATILE_WRITE);
42583	42474	return rc ? SQLITE_IOERR_ROLLBACK_ATOMIC : SQLITE_OK;
42584	42475	}
42585	42476	#endif /* __linux__ && SQLITE_ENABLE_BATCH_ATOMIC_WRITE */
42586	42477
	42478	+ case SQLITE_FCNTL_NULL_IO: {
	42479	+ osClose(pFile->h);
	42480	+ pFile->h = -1;
	42481	+ return SQLITE_OK;
	42482	+ }
42587	42483	case SQLITE_FCNTL_LOCKSTATE: {
42588	42484	(int)pArg = pFile->eFileLock;
42589	42485	return SQLITE_OK;
42590	42486	}
42591	42487	case SQLITE_FCNTL_LAST_ERRNO: {
		@@ -42723,10 +42619,11 @@
42723	42619
42724	42620	/* Set the POWERSAFE_OVERWRITE flag if requested. */
42725	42621	if( pFd->ctrlFlags & UNIXFILE_PSOW ){
42726	42622	pFd->deviceCharacteristics \|= SQLITE_IOCAP_POWERSAFE_OVERWRITE;
42727	42623	}
	42624	+ pFd->deviceCharacteristics \|= SQLITE_IOCAP_SUBPAGE_READ;
42728	42625
42729	42626	pFd->sectorSize = SQLITE_DEFAULT_SECTOR_SIZE;
42730	42627	}
42731	42628	}
42732	42629	#else
		@@ -42773,19 +42670,19 @@
42773	42670	0;
42774	42671	}else if( !strcmp(fsInfo.f_basetype, "qnx4") ){
42775	42672	pFile->sectorSize = fsInfo.f_bsize;
42776	42673	pFile->deviceCharacteristics =
42777	42674	/* full bitset of atomics from max sector size and smaller */
42778		- ((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2 \|
	42675	+ (((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2) \|
42779	42676	SQLITE_IOCAP_SEQUENTIAL \| /* The ram filesystem has no write behind
42780	42677	** so it is ordered */
42781	42678	0;
42782	42679	}else if( strstr(fsInfo.f_basetype, "dos") ){
42783	42680	pFile->sectorSize = fsInfo.f_bsize;
42784	42681	pFile->deviceCharacteristics =
42785	42682	/* full bitset of atomics from max sector size and smaller */
42786		- ((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2 \|
	42683	+ (((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2) \|
42787	42684	SQLITE_IOCAP_SEQUENTIAL \| /* The ram filesystem has no write behind
42788	42685	** so it is ordered */
42789	42686	0;
42790	42687	}else{
42791	42688	pFile->deviceCharacteristics =
		@@ -50462,10 +50359,15 @@
50462	50359	OSTRACE(("FCNTL oldFile=%p, newFile=%p, rc=SQLITE_OK\n",
50463	50360	hOldFile, pFile->h));
50464	50361	return SQLITE_OK;
50465	50362	}
50466	50363	#endif
	50364	+ case SQLITE_FCNTL_NULL_IO: {
	50365	+ (void)osCloseHandle(pFile->h);
	50366	+ pFile->h = NULL;
	50367	+ return SQLITE_OK;
	50368	+ }
50467	50369	case SQLITE_FCNTL_TEMPFILENAME: {
50468	50370	char *zTFile = 0;
50469	50371	int rc = winGetTempname(pFile->pVfs, &zTFile);
50470	50372	if( rc==SQLITE_OK ){
50471	50373	(char*)pArg = zTFile;
		@@ -50523,11 +50425,11 @@
50523	50425	/*
50524	50426	** Return a vector of device characteristics.
50525	50427	*/
50526	50428	static int winDeviceCharacteristics(sqlite3_file *id){
50527	50429	winFile p = (winFile)id;
50528		- return SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN \|
	50430	+ return SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN \| SQLITE_IOCAP_SUBPAGE_READ \|
50529	50431	((p->ctrlFlags & WINFILE_PSOW)?SQLITE_IOCAP_POWERSAFE_OVERWRITE:0);
50530	50432	}
50531	50433
50532	50434	/*
50533	50435	** Windows will only let you create file view mappings
		@@ -51911,11 +51813,11 @@
51911	51813	*/
51912	51814	char zTmpname = 0; / For temporary filename, if necessary. */
51913	51815
51914	51816	int rc = SQLITE_OK; /* Function Return Code */
51915	51817	#if !defined(NDEBUG) \|\| SQLITE_OS_WINCE
51916		- int eType = flags&0xFFFFFF00; /* Type of file to open */
	51818	+ int eType = flags&0x0FFF00; /* Type of file to open */
51917	51819	#endif
51918	51820
51919	51821	int isExclusive = (flags & SQLITE_OPEN_EXCLUSIVE);
51920	51822	int isDelete = (flags & SQLITE_OPEN_DELETEONCLOSE);
51921	51823	int isCreate = (flags & SQLITE_OPEN_CREATE);
		@@ -58131,24 +58033,32 @@
58131	58033	#ifdef SQLITE_DIRECT_OVERFLOW_READ
58132	58034	/*
58133	58035	** Return true if page pgno can be read directly from the database file
58134	58036	** by the b-tree layer. This is the case if:
58135	58037	**
58136		-** * the database file is open,
58137		-** * there are no dirty pages in the cache, and
58138		-** * the desired page is not currently in the wal file.
	58038	+** (1) the database file is open
	58039	+** (2) the VFS for the database is able to do unaligned sub-page reads
	58040	+** (3) there are no dirty pages in the cache, and
	58041	+** (4) the desired page is not currently in the wal file.
58139	58042	*/
58140	58043	SQLITE_PRIVATE int sqlite3PagerDirectReadOk(Pager *pPager, Pgno pgno){
58141		- if( pPager->fd->pMethods==0 ) return 0;
58142		- if( sqlite3PCacheIsDirty(pPager->pPCache) ) return 0;
	58044	+ assert( pPager!=0 );
	58045	+ assert( pPager->fd!=0 );
	58046	+ if( pPager->fd->pMethods==0 ) return 0; /* Case (1) */
	58047	+ if( sqlite3PCacheIsDirty(pPager->pPCache) ) return 0; /* Failed (3) */
58143	58048	#ifndef SQLITE_OMIT_WAL
58144	58049	if( pPager->pWal ){
58145	58050	u32 iRead = 0;
58146	58051	(void)sqlite3WalFindFrame(pPager->pWal, pgno, &iRead);
58147		- return iRead==0;
	58052	+ return iRead==0; /* Condition (4) */
58148	58053	}
58149	58054	#endif
	58055	+ assert( pPager->fd->pMethods->xDeviceCharacteristics!=0 );
	58056	+ if( (pPager->fd->pMethods->xDeviceCharacteristics(pPager->fd)
	58057	+ & SQLITE_IOCAP_SUBPAGE_READ)==0 ){
	58058	+ return 0; /* Case (2) */
	58059	+ }
58150	58060	return 1;
58151	58061	}
58152	58062	#endif
58153	58063
58154	58064	#ifndef SQLITE_OMIT_WAL
		@@ -65171,11 +65081,11 @@
65171	65081	** 20: Salt-2, a different random integer changing with each ckpt
65172	65082	** 24: Checksum-1 (first part of checksum for first 24 bytes of header).
65173	65083	** 28: Checksum-2 (second part of checksum for first 24 bytes of header).
65174	65084	**
65175	65085	** Immediately following the wal-header are zero or more frames. Each
65176		-** frame consists of a 24-byte frame-header followed by a <page-size> bytes
	65086	+** frame consists of a 24-byte frame-header followed by <page-size> bytes
65177	65087	** of page data. The frame-header is six big-endian 32-bit unsigned
65178	65088	** integer values, as follows:
65179	65089	**
65180	65090	** 0: Page number.
65181	65091	** 4: For commit records, the size of the database image in pages
		@@ -65668,10 +65578,11 @@
65668	65578	int nSehTry; /* Number of nested SEH_TRY{} blocks */
65669	65579	u8 lockError; /* True if a locking error has occurred */
65670	65580	#endif
65671	65581	#ifdef SQLITE_ENABLE_SNAPSHOT
65672	65582	WalIndexHdr pSnapshot; / Start transaction here if not NULL */
	65583	+ int bGetSnapshot; /* Transaction opened for sqlite3_get_snapshot() */
65673	65584	#endif
65674	65585	#ifdef SQLITE_ENABLE_SETLK_TIMEOUT
65675	65586	sqlite3 *db;
65676	65587	#endif
65677	65588	};
		@@ -67560,11 +67471,11 @@
67560	67471	return SQLITE_IOERR_IN_PAGE;
67561	67472	}
67562	67473
67563	67474	/*
67564	67475	** Assert that the Wal.lockMask mask, which indicates the locks held
67565		-** by the connenction, is consistent with the Wal.readLock, Wal.writeLock
	67476	+** by the connection, is consistent with the Wal.readLock, Wal.writeLock
67566	67477	** and Wal.ckptLock variables. To be used as:
67567	67478	**
67568	67479	** assert( walAssertLockmask(pWal) );
67569	67480	*/
67570	67481	static int walAssertLockmask(Wal *pWal){
		@@ -68224,11 +68135,11 @@
68224	68135	assert( pWal->apWiData[0]!=0 );
68225	68136	pInfo = walCkptInfo(pWal);
68226	68137	SEH_INJECT_FAULT;
68227	68138	if( !useWal && AtomicLoad(&pInfo->nBackfill)==pWal->hdr.mxFrame
68228	68139	#ifdef SQLITE_ENABLE_SNAPSHOT
68229		- && (pWal->pSnapshot==0 \|\| pWal->hdr.mxFrame==0)
	68140	+ && ((pWal->bGetSnapshot==0 && pWal->pSnapshot==0) \|\| pWal->hdr.mxFrame==0)
68230	68141	#endif
68231	68142	){
68232	68143	/* The WAL has been completely backfilled (or it is empty).
68233	68144	** and can be safely ignored.
68234	68145	*/
		@@ -69624,11 +69535,24 @@
69624	69535	*/
69625	69536	SQLITE_PRIVATE void sqlite3WalSnapshotOpen(
69626	69537	Wal *pWal,
69627	69538	sqlite3_snapshot *pSnapshot
69628	69539	){
69629		- pWal->pSnapshot = (WalIndexHdr*)pSnapshot;
	69540	+ if( pSnapshot && ((WalIndexHdr*)pSnapshot)->iVersion==0 ){
	69541	+ /* iVersion==0 means that this is a call to sqlite3_snapshot_get(). In
	69542	+ ** this case set the bGetSnapshot flag so that if the call to
	69543	+ ** sqlite3_snapshot_get() is about to read transaction on this wal
	69544	+ ** file, it does not take read-lock 0 if the wal file has been completely
	69545	+ ** checkpointed. Taking read-lock 0 would work, but then it would be
	69546	+ ** possible for a subsequent writer to destroy the snapshot even while
	69547	+ ** this connection is holding its read-transaction open. This is contrary
	69548	+ ** to user expectations, so we avoid it by not taking read-lock 0. */
	69549	+ pWal->bGetSnapshot = 1;
	69550	+ }else{
	69551	+ pWal->pSnapshot = (WalIndexHdr*)pSnapshot;
	69552	+ pWal->bGetSnapshot = 0;
	69553	+ }
69630	69554	}
69631	69555
69632	69556	/*
69633	69557	** Return a +ve value if snapshot p1 is newer than p2. A -ve value if
69634	69558	** p1 is older than p2 and zero if p1 and p2 are the same snapshot.
		@@ -75504,10 +75428,29 @@
75504	75428	** this routine.
75505	75429	*/
75506	75430	SQLITE_PRIVATE int sqlite3BtreeCursorSize(void){
75507	75431	return ROUND8(sizeof(BtCursor));
75508	75432	}
	75433	+
	75434	+#ifdef SQLITE_DEBUG
	75435	+/*
	75436	+** Return true if and only if the Btree object will be automatically
	75437	+** closed with the BtCursor closes. This is used within assert() statements
	75438	+** only.
	75439	+*/
	75440	+SQLITE_PRIVATE int sqlite3BtreeClosesWithCursor(
	75441	+ Btree pBtree, / the btree object */
	75442	+ BtCursor pCur / Corresponding cursor */
	75443	+){
	75444	+ BtShared *pBt = pBtree->pBt;
	75445	+ if( (pBt->openFlags & BTREE_SINGLE)==0 ) return 0;
	75446	+ if( pBt->pCursor!=pCur ) return 0;
	75447	+ if( pCur->pNext!=0 ) return 0;
	75448	+ if( pCur->pBtree!=pBtree ) return 0;
	75449	+ return 1;
	75450	+}
	75451	+#endif
75509	75452
75510	75453	/*
75511	75454	** Initialize memory that will be converted into a BtCursor object.
75512	75455	**
75513	75456	** The simple approach here would be to memset() the entire object
		@@ -84538,11 +84481,12 @@
84538	84481	if( apVal==0 ){
84539	84482	rc = SQLITE_NOMEM_BKPT;
84540	84483	goto value_from_function_out;
84541	84484	}
84542	84485	for(i=0; i<nVal; i++){
84543		- rc = sqlite3ValueFromExpr(db, pList->a[i].pExpr, enc, aff, &apVal[i]);
	84486	+ rc = sqlite3Stat4ValueFromExpr(pCtx->pParse, pList->a[i].pExpr, aff,
	84487	+ &apVal[i]);
84544	84488	if( apVal[i]==0 \|\| rc!=SQLITE_OK ) goto value_from_function_out;
84545	84489	}
84546	84490	}
84547	84491
84548	84492	pVal = valueNew(db, pCtx);
		@@ -89571,11 +89515,11 @@
89571	89515	/* The following two functions are used only within testcase() to prove
89572	89516	** test coverage. These functions do no exist for production builds.
89573	89517	** We must use separate SQLITE_NOINLINE functions here, since otherwise
89574	89518	** optimizer code movement causes gcov to become very confused.
89575	89519	*/
89576		-#if defined(SQLITE_COVERAGE_TEST) \|\| defined(SQLITE_DEBUG)
	89520	+#if defined(SQLITE_COVERAGE_TEST) \|\| defined(SQLITE_DEBUG)
89577	89521	static int SQLITE_NOINLINE doubleLt(double a, double b){ return a<b; }
89578	89522	static int SQLITE_NOINLINE doubleEq(double a, double b){ return a==b; }
89579	89523	#endif
89580	89524
89581	89525	/*
		@@ -89586,17 +89530,10 @@
89586	89530	SQLITE_PRIVATE int sqlite3IntFloatCompare(i64 i, double r){
89587	89531	if( sqlite3IsNaN(r) ){
89588	89532	/* SQLite considers NaN to be a NULL. And all integer values are greater
89589	89533	** than NULL */
89590	89534	return 1;
89591		- }
89592		- if( sqlite3Config.bUseLongDouble ){
89593		- LONGDOUBLE_TYPE x = (LONGDOUBLE_TYPE)i;
89594		- testcase( x<r );
89595		- testcase( x>r );
89596		- testcase( x==r );
89597		- return (x<r) ? -1 : (x>r);
89598	89535	}else{
89599	89536	i64 y;
89600	89537	if( r<-9223372036854775808.0 ) return +1;
89601	89538	if( r>=9223372036854775808.0 ) return -1;
89602	89539	y = (i64)r;
		@@ -90595,17 +90532,25 @@
90595	90532	db->xPreUpdateCallback(db->pPreUpdateArg, db, op, zDb, zTbl, iKey1, iKey2);
90596	90533	db->pPreUpdate = 0;
90597	90534	sqlite3DbFree(db, preupdate.aRecord);
90598	90535	vdbeFreeUnpacked(db, preupdate.keyinfo.nKeyField+1, preupdate.pUnpacked);
90599	90536	vdbeFreeUnpacked(db, preupdate.keyinfo.nKeyField+1, preupdate.pNewUnpacked);
	90537	+ sqlite3VdbeMemRelease(&preupdate.oldipk);
90600	90538	if( preupdate.aNew ){
90601	90539	int i;
90602	90540	for(i=0; i<pCsr->nField; i++){
90603	90541	sqlite3VdbeMemRelease(&preupdate.aNew[i]);
90604	90542	}
90605	90543	sqlite3DbNNFreeNN(db, preupdate.aNew);
90606	90544	}
	90545	+ if( preupdate.apDflt ){
	90546	+ int i;
	90547	+ for(i=0; i<pTab->nCol; i++){
	90548	+ sqlite3ValueFree(preupdate.apDflt[i]);
	90549	+ }
	90550	+ sqlite3DbFree(db, preupdate.apDflt);
	90551	+ }
90607	90552	}
90608	90553	#endif /* SQLITE_ENABLE_PREUPDATE_HOOK */
90609	90554
90610	90555	/************ End of vdbeaux.c *******************************************/
90611	90556	/************ Begin file vdbeapi.c ***************************************/
		@@ -92230,10 +92175,21 @@
92230	92175	** A successful evaluation of this routine acquires the mutex on p.
92231	92176	** the mutex is released if any kind of error occurs.
92232	92177	**
92233	92178	** The error code stored in database p->db is overwritten with the return
92234	92179	** value in any case.
	92180	+**
	92181	+** (tag-20240917-01) If vdbeUnbind(p,(u32)(i-1)) returns SQLITE_OK,
	92182	+** that means all of the the following will be true:
	92183	+**
	92184	+** p!=0
	92185	+** p->pVar!=0
	92186	+** i>0
	92187	+** i<=p->nVar
	92188	+**
	92189	+** An assert() is normally added after vdbeUnbind() to help static analyzers
	92190	+** realize this.
92235	92191	*/
92236	92192	static int vdbeUnbind(Vdbe *p, unsigned int i){
92237	92193	Mem *pVar;
92238	92194	if( vdbeSafetyNotNull(p) ){
92239	92195	return SQLITE_MISUSE_BKPT;
		@@ -92287,10 +92243,11 @@
92287	92243	Mem *pVar;
92288	92244	int rc;
92289	92245
92290	92246	rc = vdbeUnbind(p, (u32)(i-1));
92291	92247	if( rc==SQLITE_OK ){
	92248	+ assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92292	92249	if( zData!=0 ){
92293	92250	pVar = &p->aVar[i-1];
92294	92251	rc = sqlite3VdbeMemSetStr(pVar, zData, nData, encoding, xDel);
92295	92252	if( rc==SQLITE_OK && encoding!=0 ){
92296	92253	rc = sqlite3VdbeChangeEncoding(pVar, ENC(p->db));
		@@ -92336,10 +92293,11 @@
92336	92293	SQLITE_API int sqlite3_bind_double(sqlite3_stmt *pStmt, int i, double rValue){
92337	92294	int rc;
92338	92295	Vdbe p = (Vdbe )pStmt;
92339	92296	rc = vdbeUnbind(p, (u32)(i-1));
92340	92297	if( rc==SQLITE_OK ){
	92298	+ assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92341	92299	sqlite3VdbeMemSetDouble(&p->aVar[i-1], rValue);
92342	92300	sqlite3_mutex_leave(p->db->mutex);
92343	92301	}
92344	92302	return rc;
92345	92303	}
		@@ -92349,10 +92307,11 @@
92349	92307	SQLITE_API int sqlite3_bind_int64(sqlite3_stmt *pStmt, int i, sqlite_int64 iValue){
92350	92308	int rc;
92351	92309	Vdbe p = (Vdbe )pStmt;
92352	92310	rc = vdbeUnbind(p, (u32)(i-1));
92353	92311	if( rc==SQLITE_OK ){
	92312	+ assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92354	92313	sqlite3VdbeMemSetInt64(&p->aVar[i-1], iValue);
92355	92314	sqlite3_mutex_leave(p->db->mutex);
92356	92315	}
92357	92316	return rc;
92358	92317	}
		@@ -92359,10 +92318,11 @@
92359	92318	SQLITE_API int sqlite3_bind_null(sqlite3_stmt *pStmt, int i){
92360	92319	int rc;
92361	92320	Vdbe p = (Vdbe)pStmt;
92362	92321	rc = vdbeUnbind(p, (u32)(i-1));
92363	92322	if( rc==SQLITE_OK ){
	92323	+ assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92364	92324	sqlite3_mutex_leave(p->db->mutex);
92365	92325	}
92366	92326	return rc;
92367	92327	}
92368	92328	SQLITE_API int sqlite3_bind_pointer(
		@@ -92374,10 +92334,11 @@
92374	92334	){
92375	92335	int rc;
92376	92336	Vdbe p = (Vdbe)pStmt;
92377	92337	rc = vdbeUnbind(p, (u32)(i-1));
92378	92338	if( rc==SQLITE_OK ){
	92339	+ assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92379	92340	sqlite3VdbeMemSetPointer(&p->aVar[i-1], pPtr, zPTtype, xDestructor);
92380	92341	sqlite3_mutex_leave(p->db->mutex);
92381	92342	}else if( xDestructor ){
92382	92343	xDestructor(pPtr);
92383	92344	}
		@@ -92455,10 +92416,11 @@
92455	92416	SQLITE_API int sqlite3_bind_zeroblob(sqlite3_stmt *pStmt, int i, int n){
92456	92417	int rc;
92457	92418	Vdbe p = (Vdbe )pStmt;
92458	92419	rc = vdbeUnbind(p, (u32)(i-1));
92459	92420	if( rc==SQLITE_OK ){
	92421	+ assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92460	92422	#ifndef SQLITE_OMIT_INCRBLOB
92461	92423	sqlite3VdbeMemSetZeroBlob(&p->aVar[i-1], n);
92462	92424	#else
92463	92425	rc = sqlite3VdbeMemSetZeroBlob(&p->aVar[i-1], n);
92464	92426	#endif
		@@ -92789,41 +92751,68 @@
92789	92751	if( iIdx>=p->pCsr->nField \|\| iIdx<0 ){
92790	92752	rc = SQLITE_RANGE;
92791	92753	goto preupdate_old_out;
92792	92754	}
92793	92755
92794		- /* If the old.* record has not yet been loaded into memory, do so now. */
92795		- if( p->pUnpacked==0 ){
92796		- u32 nRec;
92797		- u8 *aRec;
92798		-
92799		- assert( p->pCsr->eCurType==CURTYPE_BTREE );
92800		- nRec = sqlite3BtreePayloadSize(p->pCsr->uc.pCursor);
92801		- aRec = sqlite3DbMallocRaw(db, nRec);
92802		- if( !aRec ) goto preupdate_old_out;
92803		- rc = sqlite3BtreePayload(p->pCsr->uc.pCursor, 0, nRec, aRec);
92804		- if( rc==SQLITE_OK ){
92805		- p->pUnpacked = vdbeUnpackRecord(&p->keyinfo, nRec, aRec);
92806		- if( !p->pUnpacked ) rc = SQLITE_NOMEM;
92807		- }
92808		- if( rc!=SQLITE_OK ){
92809		- sqlite3DbFree(db, aRec);
92810		- goto preupdate_old_out;
92811		- }
92812		- p->aRecord = aRec;
92813		- }
92814		-
92815		- pMem = *ppValue = &p->pUnpacked->aMem[iIdx];
92816	92756	if( iIdx==p->pTab->iPKey ){
	92757	+ *ppValue = pMem = &p->oldipk;
92817	92758	sqlite3VdbeMemSetInt64(pMem, p->iKey1);
92818		- }else if( iIdx>=p->pUnpacked->nField ){
92819		- ppValue = (sqlite3_value )columnNullValue();
92820		- }else if( p->pTab->aCol[iIdx].affinity==SQLITE_AFF_REAL ){
92821		- if( pMem->flags & (MEM_Int\|MEM_IntReal) ){
92822		- testcase( pMem->flags & MEM_Int );
92823		- testcase( pMem->flags & MEM_IntReal );
92824		- sqlite3VdbeMemRealify(pMem);
	92759	+ }else{
	92760	+
	92761	+ /* If the old.* record has not yet been loaded into memory, do so now. */
	92762	+ if( p->pUnpacked==0 ){
	92763	+ u32 nRec;
	92764	+ u8 *aRec;
	92765	+
	92766	+ assert( p->pCsr->eCurType==CURTYPE_BTREE );
	92767	+ nRec = sqlite3BtreePayloadSize(p->pCsr->uc.pCursor);
	92768	+ aRec = sqlite3DbMallocRaw(db, nRec);
	92769	+ if( !aRec ) goto preupdate_old_out;
	92770	+ rc = sqlite3BtreePayload(p->pCsr->uc.pCursor, 0, nRec, aRec);
	92771	+ if( rc==SQLITE_OK ){
	92772	+ p->pUnpacked = vdbeUnpackRecord(&p->keyinfo, nRec, aRec);
	92773	+ if( !p->pUnpacked ) rc = SQLITE_NOMEM;
	92774	+ }
	92775	+ if( rc!=SQLITE_OK ){
	92776	+ sqlite3DbFree(db, aRec);
	92777	+ goto preupdate_old_out;
	92778	+ }
	92779	+ p->aRecord = aRec;
	92780	+ }
	92781	+
	92782	+ pMem = *ppValue = &p->pUnpacked->aMem[iIdx];
	92783	+ if( iIdx>=p->pUnpacked->nField ){
	92784	+ /* This occurs when the table has been extended using ALTER TABLE
	92785	+ ** ADD COLUMN. The value to return is the default value of the column. */
	92786	+ Column *pCol = &p->pTab->aCol[iIdx];
	92787	+ if( pCol->iDflt>0 ){
	92788	+ if( p->apDflt==0 ){
	92789	+ int nByte = sizeof(sqlite3_value)p->pTab->nCol;
	92790	+ p->apDflt = (sqlite3_value**)sqlite3DbMallocZero(db, nByte);
	92791	+ if( p->apDflt==0 ) goto preupdate_old_out;
	92792	+ }
	92793	+ if( p->apDflt[iIdx]==0 ){
	92794	+ sqlite3_value *pVal = 0;
	92795	+ Expr *pDflt;
	92796	+ assert( p->pTab!=0 && IsOrdinaryTable(p->pTab) );
	92797	+ pDflt = p->pTab->u.tab.pDfltList->a[pCol->iDflt-1].pExpr;
	92798	+ rc = sqlite3ValueFromExpr(db, pDflt, ENC(db), pCol->affinity, &pVal);
	92799	+ if( rc==SQLITE_OK && pVal==0 ){
	92800	+ rc = SQLITE_CORRUPT_BKPT;
	92801	+ }
	92802	+ p->apDflt[iIdx] = pVal;
	92803	+ }
	92804	+ *ppValue = p->apDflt[iIdx];
	92805	+ }else{
	92806	+ ppValue = (sqlite3_value )columnNullValue();
	92807	+ }
	92808	+ }else if( p->pTab->aCol[iIdx].affinity==SQLITE_AFF_REAL ){
	92809	+ if( pMem->flags & (MEM_Int\|MEM_IntReal) ){
	92810	+ testcase( pMem->flags & MEM_Int );
	92811	+ testcase( pMem->flags & MEM_IntReal );
	92812	+ sqlite3VdbeMemRealify(pMem);
	92813	+ }
92825	92814	}
92826	92815	}
92827	92816
92828	92817	preupdate_old_out:
92829	92818	sqlite3Error(db, rc);
		@@ -93367,10 +93356,108 @@
93367	93356	** commenting and indentation practices when changing or adding code.
93368	93357	*/
93369	93358	/* #include "sqliteInt.h" */
93370	93359	/* #include "vdbeInt.h" */
93371	93360
	93361	+/*
	93362	+** High-resolution hardware timer used for debugging and testing only.
	93363	+*/
	93364	+#if defined(VDBE_PROFILE) \
	93365	+ \|\| defined(SQLITE_PERFORMANCE_TRACE) \
	93366	+ \|\| defined(SQLITE_ENABLE_STMT_SCANSTATUS)
	93367	+/************ Include hwtime.h in the middle of vdbe.c *******************/
	93368	+/************ Begin file hwtime.h ****************************************/
	93369	+/*
	93370	+** 2008 May 27
	93371	+**
	93372	+** The author disclaims copyright to this source code. In place of
	93373	+** a legal notice, here is a blessing:
	93374	+**
	93375	+** May you do good and not evil.
	93376	+** May you find forgiveness for yourself and forgive others.
	93377	+** May you share freely, never taking more than you give.
	93378	+**
	93379	+******************************************************************************
	93380	+**
	93381	+** This file contains inline asm code for retrieving "high-performance"
	93382	+** counters for x86 and x86_64 class CPUs.
	93383	+*/
	93384	+#ifndef SQLITE_HWTIME_H
	93385	+#define SQLITE_HWTIME_H
	93386	+
	93387	+/*
	93388	+** The following routine only works on Pentium-class (or newer) processors.
	93389	+** It uses the RDTSC opcode to read the cycle count value out of the
	93390	+** processor and returns that value. This can be used for high-res
	93391	+** profiling.
	93392	+*/
	93393	+#if !defined(__STRICT_ANSI__) && \
	93394	+ (defined(__GNUC__) \|\| defined(_MSC_VER)) && \
	93395	+ (defined(i386) \|\| defined(__i386__) \|\| defined(_M_IX86))
	93396	+
	93397	+ #if defined(__GNUC__)
	93398	+
	93399	+ __inline__ sqlite_uint64 sqlite3Hwtime(void){
	93400	+ unsigned int lo, hi;
	93401	+ __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
	93402	+ return (sqlite_uint64)hi << 32 \| lo;
	93403	+ }
	93404	+
	93405	+ #elif defined(_MSC_VER)
	93406	+
	93407	+ __declspec(naked) __inline sqlite_uint64 __cdecl sqlite3Hwtime(void){
	93408	+ __asm {
	93409	+ rdtsc
	93410	+ ret ; return value at EDX:EAX
	93411	+ }
	93412	+ }
	93413	+
	93414	+ #endif
	93415	+
	93416	+#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__x86_64__))
	93417	+
	93418	+ __inline__ sqlite_uint64 sqlite3Hwtime(void){
	93419	+ unsigned int lo, hi;
	93420	+ __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
	93421	+ return (sqlite_uint64)hi << 32 \| lo;
	93422	+ }
	93423	+
	93424	+#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__ppc__))
	93425	+
	93426	+ __inline__ sqlite_uint64 sqlite3Hwtime(void){
	93427	+ unsigned long long retval;
	93428	+ unsigned long junk;
	93429	+ __asm__ __volatile__ ("\n\
	93430	+ 1: mftbu %1\n\
	93431	+ mftb %L0\n\
	93432	+ mftbu %0\n\
	93433	+ cmpw %0,%1\n\
	93434	+ bne 1b"
	93435	+ : "=r" (retval), "=r" (junk));
	93436	+ return retval;
	93437	+ }
	93438	+
	93439	+#else
	93440	+
	93441	+ /*
	93442	+ ** asm() is needed for hardware timing support. Without asm(),
	93443	+ ** disable the sqlite3Hwtime() routine.
	93444	+ **
	93445	+ ** sqlite3Hwtime() is only used for some obscure debugging
	93446	+ ** and analysis configurations, not in any deliverable, so this
	93447	+ ** should not be a great loss.
	93448	+ */
	93449	+SQLITE_PRIVATE sqlite_uint64 sqlite3Hwtime(void){ return ((sqlite_uint64)0); }
	93450	+
	93451	+#endif
	93452	+
	93453	+#endif /* !defined(SQLITE_HWTIME_H) */
	93454	+
	93455	+/************ End of hwtime.h ********************************************/
	93456	+/************ Continuing where we left off in vdbe.c *********************/
	93457	+#endif
	93458	+
93372	93459	/*
93373	93460	** Invoke this macro on memory cells just prior to changing the
93374	93461	** value of the cell. This macro verifies that shallow copies are
93375	93462	** not misused. A shallow copy of a string or blob just copies a
93376	93463	** pointer to the string or blob, not the content. If the original
		@@ -97875,12 +97962,17 @@
97875	97962	0, pCx->uc.pCursor);
97876	97963	pCx->isTable = 1;
97877	97964	}
97878	97965	}
97879	97966	pCx->isOrdered = (pOp->p5!=BTREE_UNORDERED);
	97967	+ assert( p->apCsr[pOp->p1]==pCx );
97880	97968	if( rc ){
	97969	+ assert( !sqlite3BtreeClosesWithCursor(pCx->ub.pBtx, pCx->uc.pCursor) );
97881	97970	sqlite3BtreeClose(pCx->ub.pBtx);
	97971	+ p->apCsr[pOp->p1] = 0; /* Not required; helps with static analysis */
	97972	+ }else{
	97973	+ assert( sqlite3BtreeClosesWithCursor(pCx->ub.pBtx, pCx->uc.pCursor) );
97882	97974	}
97883	97975	}
97884	97976	}
97885	97977	if( rc ) goto abort_due_to_error;
97886	97978	pCx->nullRow = 1;
		@@ -102394,11 +102486,11 @@
102394	102486	** element.
102395	102487	**
102396	102488	** As with all opcodes, the meanings of the parameters for OP_Explain
102397	102489	** are subject to change from one release to the next. Applications
102398	102490	** should not attempt to interpret or use any of the information
102399		-** contined in the OP_Explain opcode. The information provided by this
	102491	+** contained in the OP_Explain opcode. The information provided by this
102400	102492	** opcode is intended for testing and debugging use only.
102401	102493	*/
102402	102494	default: { /* This is really OP_Noop, OP_Explain */
102403	102495	assert( pOp->opcode==OP_Noop \|\| pOp->opcode==OP_Explain );
102404	102496
		@@ -107490,11 +107582,11 @@
107490	107582	** non-VIEW candidate plus multiple VIEW candidates. In other
107491	107583	** words non-VIEW candidate terms take precedence over VIEWs.
107492	107584	*/
107493	107585	if( cntTab==0
107494	107586	\|\| (cntTab==1
107495		- && ALWAYS(pMatch!=0)
	107587	+ && pMatch!=0
107496	107588	&& ALWAYS(pMatch->pSTab!=0)
107497	107589	&& (pMatch->pSTab->tabFlags & TF_Ephemeral)!=0
107498	107590	&& (pTab->tabFlags & TF_Ephemeral)==0)
107499	107591	){
107500	107592	cntTab = 1;
		@@ -108123,12 +108215,12 @@
108123	108215	}
108124	108216
108125	108217	/* Resolve function names
108126	108218	*/
108127	108219	case TK_FUNCTION: {
108128		- ExprList pList = pExpr->x.pList; / The argument list */
108129		- int n = pList ? pList->nExpr : 0; /* Number of arguments */
	108220	+ ExprList pList; / The argument list */
	108221	+ int n; /* Number of arguments */
108130	108222	int no_such_func = 0; /* True if no such function exists */
108131	108223	int wrong_num_args = 0; /* True if wrong number of arguments */
108132	108224	int is_agg = 0; /* True if is an aggregate function */
108133	108225	const char zId; / The function name. */
108134	108226	FuncDef pDef; / Information about the function */
		@@ -108137,10 +108229,12 @@
108137	108229	#ifndef SQLITE_OMIT_WINDOWFUNC
108138	108230	Window *pWin = (IsWindowFunc(pExpr) ? pExpr->y.pWin : 0);
108139	108231	#endif
108140	108232	assert( !ExprHasProperty(pExpr, EP_xIsSelect\|EP_IntValue) );
108141	108233	assert( pExpr->pLeft==0 \|\| pExpr->pLeft->op==TK_ORDER );
	108234	+ pList = pExpr->x.pList;
	108235	+ n = pList ? pList->nExpr : 0;
108142	108236	zId = pExpr->u.zToken;
108143	108237	pDef = sqlite3FindFunction(pParse->db, zId, n, enc, 0);
108144	108238	if( pDef==0 ){
108145	108239	pDef = sqlite3FindFunction(pParse->db, zId, -2, enc, 0);
108146	108240	if( pDef==0 ){
		@@ -108185,10 +108279,28 @@
108185	108279	pExpr->op = TK_NULL;
108186	108280	return WRC_Prune;
108187	108281	}
108188	108282	}
108189	108283	#endif
	108284	+
	108285	+ /* If the function may call sqlite3_value_subtype(), then set the
	108286	+ ** EP_SubtArg flag on all of its argument expressions. This prevents
	108287	+ ** where.c from replacing the expression with a value read from an
	108288	+ ** index on the same expression, which will not have the correct
	108289	+ ** subtype. Also set the flag if the function expression itself is
	108290	+ ** an EP_SubtArg expression. In this case subtypes are required as
	108291	+ ** the function may return a value with a subtype back to its
	108292	+ ** caller using sqlite3_result_value(). */
	108293	+ if( (pDef->funcFlags & SQLITE_SUBTYPE)
	108294	+ \|\| ExprHasProperty(pExpr, EP_SubtArg)
	108295	+ ){
	108296	+ int ii;
	108297	+ for(ii=0; ii<n; ii++){
	108298	+ ExprSetProperty(pList->a[ii].pExpr, EP_SubtArg);
	108299	+ }
	108300	+ }
	108301	+
108190	108302	if( pDef->funcFlags & (SQLITE_FUNC_CONSTANT\|SQLITE_FUNC_SLOCHNG) ){
108191	108303	/* For the purposes of the EP_ConstFunc flag, date and time
108192	108304	** functions and other functions that change slowly are considered
108193	108305	** constant because they are constant for the duration of one query.
108194	108306	** This allows them to be factored out of inner loops. */
		@@ -111951,11 +112063,11 @@
111951	112063	**
111952	112064	** (4) If pSrc is the right operand of a LEFT JOIN, then...
111953	112065	** (4a) pExpr must come from an ON clause..
111954	112066	** (4b) and specifically the ON clause associated with the LEFT JOIN.
111955	112067	**
111956		-** (5) If pSrc is not the right operand of a LEFT JOIN or the left
	112068	+** (5) If pSrc is the right operand of a LEFT JOIN or the left
111957	112069	** operand of a RIGHT JOIN, then pExpr must be from the WHERE
111958	112070	** clause, not an ON clause.
111959	112071	**
111960	112072	** (6) Either:
111961	112073	**
		@@ -113858,10 +113970,63 @@
113858	113970	}
113859	113971	#endif /* !defined(SQLITE_UNTESTABLE) */
113860	113972	}
113861	113973	return target;
113862	113974	}
	113975	+
	113976	+/*
	113977	+** Expression Node callback for sqlite3ExprCanReturnSubtype().
	113978	+**
	113979	+** Only a function call is able to return a subtype. So if the node
	113980	+** is not a function call, return WRC_Prune immediately.
	113981	+**
	113982	+** A function call is able to return a subtype if it has the
	113983	+** SQLITE_RESULT_SUBTYPE property.
	113984	+**
	113985	+** Assume that every function is able to pass-through a subtype from
	113986	+** one of its argument (using sqlite3_result_value()). Most functions
	113987	+** are not this way, but we don't have a mechanism to distinguish those
	113988	+** that are from those that are not, so assume they all work this way.
	113989	+** That means that if one of its arguments is another function and that
	113990	+** other function is able to return a subtype, then this function is
	113991	+** able to return a subtype.
	113992	+*/
	113993	+static int exprNodeCanReturnSubtype(Walker pWalker, Expr pExpr){
	113994	+ int n;
	113995	+ FuncDef *pDef;
	113996	+ sqlite3 *db;
	113997	+ if( pExpr->op!=TK_FUNCTION ){
	113998	+ return WRC_Prune;
	113999	+ }
	114000	+ assert( ExprUseXList(pExpr) );
	114001	+ db = pWalker->pParse->db;
	114002	+ n = ALWAYS(pExpr->x.pList) ? pExpr->x.pList->nExpr : 0;
	114003	+ pDef = sqlite3FindFunction(db, pExpr->u.zToken, n, ENC(db), 0);
	114004	+ if( NEVER(pDef==0) \|\| (pDef->funcFlags & SQLITE_RESULT_SUBTYPE)!=0 ){
	114005	+ pWalker->eCode = 1;
	114006	+ return WRC_Prune;
	114007	+ }
	114008	+ return WRC_Continue;
	114009	+}
	114010	+
	114011	+/*
	114012	+** Return TRUE if expression pExpr is able to return a subtype.
	114013	+**
	114014	+** A TRUE return does not guarantee that a subtype will be returned.
	114015	+** It only indicates that a subtype return is possible. False positives
	114016	+** are acceptable as they only disable an optimization. False negatives,
	114017	+** on the other hand, can lead to incorrect answers.
	114018	+*/
	114019	+static int sqlite3ExprCanReturnSubtype(Parse pParse, Expr pExpr){
	114020	+ Walker w;
	114021	+ memset(&w, 0, sizeof(w));
	114022	+ w.pParse = pParse;
	114023	+ w.xExprCallback = exprNodeCanReturnSubtype;
	114024	+ sqlite3WalkExpr(&w, pExpr);
	114025	+ return w.eCode;
	114026	+}
	114027	+
113863	114028
113864	114029	/*
113865	114030	** Check to see if pExpr is one of the indexed expressions on pParse->pIdxEpr.
113866	114031	** If it is, then resolve the expression by reading from the index and
113867	114032	** return the register into which the value has been read. If pExpr is
		@@ -113891,10 +114056,21 @@
113891	114056	){
113892	114057	/* Affinity mismatch on a generated column */
113893	114058	continue;
113894	114059	}
113895	114060
	114061	+
	114062	+ /* Functions that might set a subtype should not be replaced by the
	114063	+ ** value taken from an expression index if they are themselves an
	114064	+ ** argument to another scalar function or aggregate.
	114065	+ ** https://sqlite.org/forum/forumpost/68d284c86b082c3e */
	114066	+ if( ExprHasProperty(pExpr, EP_SubtArg)
	114067	+ && sqlite3ExprCanReturnSubtype(pParse, pExpr)
	114068	+ ){
	114069	+ continue;
	114070	+ }
	114071	+
113896	114072	v = pParse->pVdbe;
113897	114073	assert( v!=0 );
113898	114074	if( p->bMaybeNullRow ){
113899	114075	/* If the index is on a NULL row due to an outer join, then we
113900	114076	** cannot extract the value from the index. The value must be
		@@ -115421,35 +115597,41 @@
115421	115597	**
115422	115598	** Additionally, if pExpr is a simple SQL value and the value is the
115423	115599	** same as that currently bound to variable pVar, non-zero is returned.
115424	115600	** Otherwise, if the values are not the same or if pExpr is not a simple
115425	115601	** SQL value, zero is returned.
	115602	+**
	115603	+** If the SQLITE_EnableQPSG flag is set on the database connection, then
	115604	+** this routine always returns false.
115426	115605	*/
115427		-static int exprCompareVariable(
	115606	+static SQLITE_NOINLINE int exprCompareVariable(
115428	115607	const Parse *pParse,
115429	115608	const Expr *pVar,
115430	115609	const Expr *pExpr
115431	115610	){
115432		- int res = 0;
	115611	+ int res = 2;
115433	115612	int iVar;
115434	115613	sqlite3_value pL, pR = 0;
115435	115614
	115615	+ if( pExpr->op==TK_VARIABLE && pVar->iColumn==pExpr->iColumn ){
	115616	+ return 0;
	115617	+ }
	115618	+ if( (pParse->db->flags & SQLITE_EnableQPSG)!=0 ) return 2;
115436	115619	sqlite3ValueFromExpr(pParse->db, pExpr, SQLITE_UTF8, SQLITE_AFF_BLOB, &pR);
115437	115620	if( pR ){
115438	115621	iVar = pVar->iColumn;
115439	115622	sqlite3VdbeSetVarmask(pParse->pVdbe, iVar);
115440	115623	pL = sqlite3VdbeGetBoundValue(pParse->pReprepare, iVar, SQLITE_AFF_BLOB);
115441	115624	if( pL ){
115442	115625	if( sqlite3_value_type(pL)==SQLITE_TEXT ){
115443	115626	sqlite3_value_text(pL); /* Make sure the encoding is UTF-8 */
115444	115627	}
115445		- res = 0==sqlite3MemCompare(pL, pR, 0);
	115628	+ res = sqlite3MemCompare(pL, pR, 0) ? 2 : 0;
115446	115629	}
115447	115630	sqlite3ValueFree(pR);
115448	115631	sqlite3ValueFree(pL);
115449	115632	}
115450		-
115451	115633	return res;
115452	115634	}
115453	115635
115454	115636	/*
115455	115637	** Do a deep comparison of two expression trees. Return 0 if the two
		@@ -115471,16 +115653,14 @@
115471	115653	** can be sure the expressions are the same. In the places where
115472	115654	** this routine is used, it does not hurt to get an extra 2 - that
115473	115655	** just might result in some slightly slower code. But returning
115474	115656	** an incorrect 0 or 1 could lead to a malfunction.
115475	115657	**
115476		-** If pParse is not NULL then TK_VARIABLE terms in pA with bindings in
115477		-** pParse->pReprepare can be matched against literals in pB. The
115478		-** pParse->pVdbe->expmask bitmask is updated for each variable referenced.
115479		-** If pParse is NULL (the normal case) then any TK_VARIABLE term in
115480		-** Argument pParse should normally be NULL. If it is not NULL and pA or
115481		-** pB causes a return value of 2.
	115658	+** If pParse is not NULL and SQLITE_EnableQPSG is off then TK_VARIABLE
	115659	+** terms in pA with bindings in pParse->pReprepare can be matched against
	115660	+** literals in pB. The pParse->pVdbe->expmask bitmask is updated for
	115661	+** each variable referenced.
115482	115662	*/
115483	115663	SQLITE_PRIVATE int sqlite3ExprCompare(
115484	115664	const Parse *pParse,
115485	115665	const Expr *pA,
115486	115666	const Expr *pB,
		@@ -115488,12 +115668,12 @@
115488	115668	){
115489	115669	u32 combinedFlags;
115490	115670	if( pA==0 \|\| pB==0 ){
115491	115671	return pB==pA ? 0 : 2;
115492	115672	}
115493		- if( pParse && pA->op==TK_VARIABLE && exprCompareVariable(pParse, pA, pB) ){
115494		- return 0;
	115673	+ if( pParse && pA->op==TK_VARIABLE ){
	115674	+ return exprCompareVariable(pParse, pA, pB);
115495	115675	}
115496	115676	combinedFlags = pA->flags \| pB->flags;
115497	115677	if( combinedFlags & EP_IntValue ){
115498	115678	if( (pA->flags&pB->flags&EP_IntValue)!=0 && pA->u.iValue==pB->u.iValue ){
115499	115679	return 0;
		@@ -115683,23 +115863,75 @@
115683	115863	return exprImpliesNotNull(pParse, p->pLeft, pNN, iTab, 1);
115684	115864	}
115685	115865	}
115686	115866	return 0;
115687	115867	}
	115868	+
	115869	+/*
	115870	+** Return true if the boolean value of the expression is always either
	115871	+** FALSE or NULL.
	115872	+*/
	115873	+static int sqlite3ExprIsNotTrue(Expr *pExpr){
	115874	+ int v;
	115875	+ if( pExpr->op==TK_NULL ) return 1;
	115876	+ if( pExpr->op==TK_TRUEFALSE && sqlite3ExprTruthValue(pExpr)==0 ) return 1;
	115877	+ v = 1;
	115878	+ if( sqlite3ExprIsInteger(pExpr, &v, 0) && v==0 ) return 1;
	115879	+ return 0;
	115880	+}
	115881	+
	115882	+/*
	115883	+** Return true if the expression is one of the following:
	115884	+**
	115885	+** CASE WHEN x THEN y END
	115886	+** CASE WHEN x THEN y ELSE NULL END
	115887	+** CASE WHEN x THEN y ELSE false END
	115888	+** iif(x,y)
	115889	+** iif(x,y,NULL)
	115890	+** iif(x,y,false)
	115891	+*/
	115892	+static int sqlite3ExprIsIIF(sqlite3 db, const Expr pExpr){
	115893	+ ExprList *pList;
	115894	+ if( pExpr->op==TK_FUNCTION ){
	115895	+ const char *z = pExpr->u.zToken;
	115896	+ FuncDef *pDef;
	115897	+ if( (z[0]!='i' && z[0]!='I') ) return 0;
	115898	+ if( pExpr->x.pList==0 ) return 0;
	115899	+ pDef = sqlite3FindFunction(db, z, pExpr->x.pList->nExpr, ENC(db), 0);
	115900	+#ifdef SQLITE_ENABLE_UNKNOWN_SQL_FUNCTION
	115901	+ if( pDef==0 ) return 0;
	115902	+#else
	115903	+ if( NEVER(pDef==0) ) return 0;
	115904	+#endif
	115905	+ if( (pDef->funcFlags & SQLITE_FUNC_INLINE)==0 ) return 0;
	115906	+ if( SQLITE_PTR_TO_INT(pDef->pUserData)!=INLINEFUNC_iif ) return 0;
	115907	+ }else if( pExpr->op==TK_CASE ){
	115908	+ if( pExpr->pLeft!=0 ) return 0;
	115909	+ }else{
	115910	+ return 0;
	115911	+ }
	115912	+ pList = pExpr->x.pList;
	115913	+ assert( pList!=0 );
	115914	+ if( pList->nExpr==2 ) return 1;
	115915	+ if( pList->nExpr==3 && sqlite3ExprIsNotTrue(pList->a[2].pExpr) ) return 1;
	115916	+ return 0;
	115917	+}
115688	115918
115689	115919	/*
115690	115920	** Return true if we can prove the pE2 will always be true if pE1 is
115691	115921	** true. Return false if we cannot complete the proof or if pE2 might
115692	115922	** be false. Examples:
115693	115923	**
115694		-** pE1: x==5 pE2: x==5 Result: true
115695		-** pE1: x>0 pE2: x==5 Result: false
115696		-** pE1: x=21 pE2: x=21 OR y=43 Result: true
115697		-** pE1: x!=123 pE2: x IS NOT NULL Result: true
115698		-** pE1: x!=?1 pE2: x IS NOT NULL Result: true
115699		-** pE1: x IS NULL pE2: x IS NOT NULL Result: false
115700		-** pE1: x IS ?2 pE2: x IS NOT NULL Result: false
	115924	+** pE1: x==5 pE2: x==5 Result: true
	115925	+** pE1: x>0 pE2: x==5 Result: false
	115926	+** pE1: x=21 pE2: x=21 OR y=43 Result: true
	115927	+** pE1: x!=123 pE2: x IS NOT NULL Result: true
	115928	+** pE1: x!=?1 pE2: x IS NOT NULL Result: true
	115929	+** pE1: x IS NULL pE2: x IS NOT NULL Result: false
	115930	+** pE1: x IS ?2 pE2: x IS NOT NULL Result: false
	115931	+** pE1: iif(x,y) pE2: x Result: true
	115932	+** PE1: iif(x,y,0) pE2: x Result: true
115701	115933	**
115702	115934	** When comparing TK_COLUMN nodes between pE1 and pE2, if pE2 has
115703	115935	** Expr.iTable<0 then assume a table number given by iTab.
115704	115936	**
115705	115937	** If pParse is not NULL, then the values of bound variables in pE1 are
		@@ -115729,10 +115961,13 @@
115729	115961	if( pE2->op==TK_NOTNULL
115730	115962	&& exprImpliesNotNull(pParse, pE1, pE2->pLeft, iTab, 0)
115731	115963	){
115732	115964	return 1;
115733	115965	}
	115966	+ if( sqlite3ExprIsIIF(pParse->db, pE1) ){
	115967	+ return sqlite3ExprImpliesExpr(pParse,pE1->x.pList->a[0].pExpr,pE2,iTab);
	115968	+ }
115734	115969	return 0;
115735	115970	}
115736	115971
115737	115972	/* This is a helper function to impliesNotNullRow(). In this routine,
115738	115973	** set pWalker->eCode to one only if both of the input expressions
		@@ -120689,12 +120924,12 @@
120689	120924	int nIdxCol = 1; /* Number of columns in stat4 records */
120690	120925
120691	120926	char zIndex; / Index name */
120692	120927	Index pIdx; / Pointer to the index object */
120693	120928	int nSample; /* Number of samples */
120694		- int nByte; /* Bytes of space required */
120695		- int i; /* Bytes of space required */
	120929	+ i64 nByte; /* Bytes of space required */
	120930	+ i64 i; /* Bytes of space required */
120696	120931	tRowcnt pSpace; / Available allocated memory space */
120697	120932	u8 pPtr; / Available memory as a u8 for easier manipulation */
120698	120933
120699	120934	zIndex = (char *)sqlite3_column_text(pStmt, 0);
120700	120935	if( zIndex==0 ) continue;
		@@ -121140,19 +121375,10 @@
121140	121375	rc = sqlite3Init(db, &zErrDyn);
121141	121376	}
121142	121377	sqlite3BtreeLeaveAll(db);
121143	121378	assert( zErrDyn==0 \|\| rc!=SQLITE_OK );
121144	121379	}
121145		-#ifdef SQLITE_USER_AUTHENTICATION
121146		- if( rc==SQLITE_OK && !REOPEN_AS_MEMDB(db) ){
121147		- u8 newAuth = 0;
121148		- rc = sqlite3UserAuthCheckLogin(db, zName, &newAuth);
121149		- if( newAuth<db->auth.authLevel ){
121150		- rc = SQLITE_AUTH_USER;
121151		- }
121152		- }
121153		-#endif
121154	121380	if( rc ){
121155	121381	if( ALWAYS(!REOPEN_AS_MEMDB(db)) ){
121156	121382	int iDb = db->nDb - 1;
121157	121383	assert( iDb>=2 );
121158	121384	if( db->aDb[iDb].pBt ){
		@@ -121646,15 +121872,11 @@
121646	121872	sqlite3 db = pParse->db; / Database handle */
121647	121873	char zDb = db->aDb[iDb].zDbSName; / Schema name of attached database */
121648	121874	int rc; /* Auth callback return code */
121649	121875
121650	121876	if( db->init.busy ) return SQLITE_OK;
121651		- rc = db->xAuth(db->pAuthArg, SQLITE_READ, zTab,zCol,zDb,pParse->zAuthContext
121652		-#ifdef SQLITE_USER_AUTHENTICATION
121653		- ,db->auth.zAuthUser
121654		-#endif
121655		- );
	121877	+ rc = db->xAuth(db->pAuthArg, SQLITE_READ, zTab,zCol,zDb,pParse->zAuthContext);
121656	121878	if( rc==SQLITE_DENY ){
121657	121879	char *z = sqlite3_mprintf("%s.%s", zTab, zCol);
121658	121880	if( db->nDb>2 \|\| iDb!=0 ) z = sqlite3_mprintf("%s.%z", zDb, z);
121659	121881	sqlite3ErrorMsg(pParse, "access to %z is prohibited", z);
121660	121882	pParse->rc = SQLITE_AUTH;
		@@ -121757,15 +121979,11 @@
121757	121979	testcase( zArg1==0 );
121758	121980	testcase( zArg2==0 );
121759	121981	testcase( zArg3==0 );
121760	121982	testcase( pParse->zAuthContext==0 );
121761	121983
121762		- rc = db->xAuth(db->pAuthArg, code, zArg1, zArg2, zArg3, pParse->zAuthContext
121763		-#ifdef SQLITE_USER_AUTHENTICATION
121764		- ,db->auth.zAuthUser
121765		-#endif
121766		- );
	121984	+ rc = db->xAuth(db->pAuthArg,code,zArg1,zArg2,zArg3,pParse->zAuthContext);
121767	121985	if( rc==SQLITE_DENY ){
121768	121986	sqlite3ErrorMsg(pParse, "not authorized");
121769	121987	pParse->rc = SQLITE_AUTH;
121770	121988	}else if( rc!=SQLITE_OK && rc!=SQLITE_IGNORE ){
121771	121989	rc = SQLITE_DENY;
		@@ -121994,21 +122212,10 @@
121994	122212	sqlite3VdbeJumpHere(v, addrRewind);
121995	122213	}
121996	122214	}
121997	122215	sqlite3VdbeAddOp0(v, OP_Halt);
121998	122216
121999		-#if SQLITE_USER_AUTHENTICATION && !defined(SQLITE_OMIT_SHARED_CACHE)
122000		- if( pParse->nTableLock>0 && db->init.busy==0 ){
122001		- sqlite3UserAuthInit(db);
122002		- if( db->auth.authLevel<UAUTH_User ){
122003		- sqlite3ErrorMsg(pParse, "user not authenticated");
122004		- pParse->rc = SQLITE_AUTH_USER;
122005		- return;
122006		- }
122007		- }
122008		-#endif
122009		-
122010	122217	/* The cookie mask contains one bit for each database file open.
122011	122218	** (Bit 0 is for main, bit 1 is for temp, and so forth.) Bits are
122012	122219	** set for each database that is used. Generate code to start a
122013	122220	** transaction on each used database and to verify the schema cookie
122014	122221	** on each used database.
		@@ -122133,20 +122340,10 @@
122133	122340	sqlite3DbFree(db, zSql);
122134	122341	memcpy(PARSE_TAIL(pParse), saveBuf, PARSE_TAIL_SZ);
122135	122342	pParse->nested--;
122136	122343	}
122137	122344
122138		-#if SQLITE_USER_AUTHENTICATION
122139		-/*
122140		-** Return TRUE if zTable is the name of the system table that stores the
122141		-** list of users and their access credentials.
122142		-*/
122143		-SQLITE_PRIVATE int sqlite3UserAuthTable(const char *zTable){
122144		- return sqlite3_stricmp(zTable, "sqlite_user")==0;
122145		-}
122146		-#endif
122147		-
122148	122345	/*
122149	122346	** Locate the in-memory structure that describes a particular database
122150	122347	** table given the name of that table and (optionally) the name of the
122151	122348	** database containing the table. Return NULL if not found.
122152	122349	**
		@@ -122161,17 +122358,10 @@
122161	122358	Table *p = 0;
122162	122359	int i;
122163	122360
122164	122361	/* All mutexes are required for schema access. Make sure we hold them. */
122165	122362	assert( zDatabase!=0 \|\| sqlite3BtreeHoldsAllMutexes(db) );
122166		-#if SQLITE_USER_AUTHENTICATION
122167		- /* Only the admin user is allowed to know that the sqlite_user table
122168		- ** exists */
122169		- if( db->auth.authLevel<UAUTH_Admin && sqlite3UserAuthTable(zName)!=0 ){
122170		- return 0;
122171		- }
122172		-#endif
122173	122363	if( zDatabase ){
122174	122364	for(i=0; i<db->nDb; i++){
122175	122365	if( sqlite3StrICmp(zDatabase, db->aDb[i].zDbSName)==0 ) break;
122176	122366	}
122177	122367	if( i>=db->nDb ){
		@@ -125826,13 +126016,10 @@
125826	126016
125827	126017	assert( pTab!=0 );
125828	126018	if( sqlite3StrNICmp(pTab->zName, "sqlite_", 7)==0
125829	126019	&& db->init.busy==0
125830	126020	&& pTblName!=0
125831		-#if SQLITE_USER_AUTHENTICATION
125832		- && sqlite3UserAuthTable(pTab->zName)==0
125833		-#endif
125834	126021	){
125835	126022	sqlite3ErrorMsg(pParse, "table %s may not be indexed", pTab->zName);
125836	126023	goto exit_create_index;
125837	126024	}
125838	126025	#ifndef SQLITE_OMIT_VIEW
		@@ -128224,10 +128411,11 @@
128224	128411	** 4) The table is a shadow table, the database connection is in
128225	128412	** defensive mode, and the current sqlite3_prepare()
128226	128413	** is for a top-level SQL statement.
128227	128414	*/
128228	128415	static int vtabIsReadOnly(Parse pParse, Table pTab){
	128416	+ assert( IsVirtual(pTab) );
128229	128417	if( sqlite3GetVTable(pParse->db, pTab)->pMod->pModule->xUpdate==0 ){
128230	128418	return 1;
128231	128419	}
128232	128420
128233	128421	/* Within triggers:
		@@ -131716,11 +131904,17 @@
131716	131904	#ifdef SQLITE_DEBUG
131717	131905	/*
131718	131906	** Implementation of fpdecode(x,y,z) function.
131719	131907	**
131720	131908	** x is a real number that is to be decoded. y is the precision.
131721		-** z is the maximum real precision.
	131909	+** z is the maximum real precision. Return a string that shows the
	131910	+** results of the sqlite3FpDecode() function.
	131911	+**
	131912	+** Used for testing and debugging only, specifically testing and debugging
	131913	+** of the sqlite3FpDecode() function. This SQL function does not appear
	131914	+** in production builds. This function is not an API and is subject to
	131915	+** modification or removal in future versions of SQLite.
131722	131916	*/
131723	131917	static void fpdecodeFunc(
131724	131918	sqlite3_context *context,
131725	131919	int argc,
131726	131920	sqlite3_value **argv
		@@ -131740,10 +131934,86 @@
131740	131934	sqlite3_snprintf(sizeof(zBuf), zBuf, "NaN");
131741	131935	}else{
131742	131936	sqlite3_snprintf(sizeof(zBuf), zBuf, "%c%.*s/%d", s.sign, s.n, s.z, s.iDP);
131743	131937	}
131744	131938	sqlite3_result_text(context, zBuf, -1, SQLITE_TRANSIENT);
	131939	+}
	131940	+#endif /* SQLITE_DEBUG */
	131941	+
	131942	+#ifdef SQLITE_DEBUG
	131943	+/*
	131944	+** Implementation of parseuri(uri,flags) function.
	131945	+**
	131946	+** Required Arguments:
	131947	+** "uri" The URI to parse.
	131948	+** "flags" Bitmask of flags, as if to sqlite3_open_v2().
	131949	+**
	131950	+** Additional arguments beyond the first two make calls to
	131951	+** sqlite3_uri_key() for integers and sqlite3_uri_parameter for
	131952	+** anything else.
	131953	+**
	131954	+** The result is a string showing the results of calling sqlite3ParseUri().
	131955	+**
	131956	+** Used for testing and debugging only, specifically testing and debugging
	131957	+** of the sqlite3ParseUri() function. This SQL function does not appear
	131958	+** in production builds. This function is not an API and is subject to
	131959	+** modification or removal in future versions of SQLite.
	131960	+*/
	131961	+static void parseuriFunc(
	131962	+ sqlite3_context *ctx,
	131963	+ int argc,
	131964	+ sqlite3_value **argv
	131965	+){
	131966	+ sqlite3_str *pResult;
	131967	+ const char *zVfs;
	131968	+ const char *zUri;
	131969	+ unsigned int flgs;
	131970	+ int rc;
	131971	+ sqlite3_vfs *pVfs = 0;
	131972	+ char *zFile = 0;
	131973	+ char *zErr = 0;
	131974	+
	131975	+ if( argc<2 ) return;
	131976	+ pVfs = sqlite3_vfs_find(0);
	131977	+ assert( pVfs );
	131978	+ zVfs = pVfs->zName;
	131979	+ zUri = (const char*)sqlite3_value_text(argv[0]);
	131980	+ if( zUri==0 ) return;
	131981	+ flgs = (unsigned int)sqlite3_value_int(argv[1]);
	131982	+ rc = sqlite3ParseUri(zVfs, zUri, &flgs, &pVfs, &zFile, &zErr);
	131983	+ pResult = sqlite3_str_new(0);
	131984	+ if( pResult ){
	131985	+ int i;
	131986	+ sqlite3_str_appendf(pResult, "rc=%d", rc);
	131987	+ sqlite3_str_appendf(pResult, ", flags=0x%x", flgs);
	131988	+ sqlite3_str_appendf(pResult, ", vfs=%Q", pVfs ? pVfs->zName: 0);
	131989	+ sqlite3_str_appendf(pResult, ", err=%Q", zErr);
	131990	+ sqlite3_str_appendf(pResult, ", file=%Q", zFile);
	131991	+ if( zFile ){
	131992	+ const char *z = zFile;
	131993	+ z += sqlite3Strlen30(z)+1;
	131994	+ while( z[0] ){
	131995	+ sqlite3_str_appendf(pResult, ", %Q", z);
	131996	+ z += sqlite3Strlen30(z)+1;
	131997	+ }
	131998	+ for(i=2; i<argc; i++){
	131999	+ const char *zArg;
	132000	+ if( sqlite3_value_type(argv[i])==SQLITE_INTEGER ){
	132001	+ int k = sqlite3_value_int(argv[i]);
	132002	+ sqlite3_str_appendf(pResult, ", '%d:%q'",k,sqlite3_uri_key(zFile, k));
	132003	+ }else if( (zArg = (const char*)sqlite3_value_text(argv[i]))!=0 ){
	132004	+ sqlite3_str_appendf(pResult, ", '%q:%q'",
	132005	+ zArg, sqlite3_uri_parameter(zFile,zArg));
	132006	+ }else{
	132007	+ sqlite3_str_appendf(pResult, ", NULL");
	132008	+ }
	132009	+ }
	132010	+ }
	132011	+ sqlite3_result_text(ctx, sqlite3_str_finish(pResult), -1, sqlite3_free);
	132012	+ }
	132013	+ sqlite3_free_filename(zFile);
	132014	+ sqlite3_free(zErr);
131745	132015	}
131746	132016	#endif /* SQLITE_DEBUG */
131747	132017
131748	132018	/*
131749	132019	** All of the FuncDef structures in the aBuiltinFunc[] array above
		@@ -131777,13 +132047,10 @@
131777	132047	#endif
131778	132048	#ifndef SQLITE_OMIT_LOAD_EXTENSION
131779	132049	SFUNCTION(load_extension, 1, 0, 0, loadExt ),
131780	132050	SFUNCTION(load_extension, 2, 0, 0, loadExt ),
131781	132051	#endif
131782		-#if SQLITE_USER_AUTHENTICATION
131783		- FUNCTION(sqlite_crypt, 2, 0, 0, sqlite3CryptFunc ),
131784		-#endif
131785	132052	#ifndef SQLITE_OMIT_COMPILEOPTION_DIAGS
131786	132053	DFUNCTION(sqlite_compileoption_used,1, 0, 0, compileoptionusedFunc ),
131787	132054	DFUNCTION(sqlite_compileoption_get, 1, 0, 0, compileoptiongetFunc ),
131788	132055	#endif /* SQLITE_OMIT_COMPILEOPTION_DIAGS */
131789	132056	INLINE_FUNC(unlikely, 1, INLINEFUNC_unlikely, SQLITE_FUNC_UNLIKELY),
		@@ -131805,11 +132072,12 @@
131805	132072	FUNCTION(max, -1, 1, 1, minmaxFunc ),
131806	132073	FUNCTION(max, 0, 1, 1, 0 ),
131807	132074	WAGGREGATE(max, 1, 1, 1, minmaxStep, minMaxFinalize, minMaxValue, 0,
131808	132075	SQLITE_FUNC_MINMAX\|SQLITE_FUNC_ANYORDER ),
131809	132076	FUNCTION2(typeof, 1, 0, 0, typeofFunc, SQLITE_FUNC_TYPEOF),
131810		- FUNCTION2(subtype, 1, 0, 0, subtypeFunc, SQLITE_FUNC_TYPEOF),
	132077	+ FUNCTION2(subtype, 1, 0, 0, subtypeFunc,
	132078	+ SQLITE_FUNC_TYPEOF\|SQLITE_SUBTYPE),
131811	132079	FUNCTION2(length, 1, 0, 0, lengthFunc, SQLITE_FUNC_LENGTH),
131812	132080	FUNCTION2(octet_length, 1, 0, 0, bytelengthFunc,SQLITE_FUNC_BYTELEN),
131813	132081	FUNCTION(instr, 2, 0, 0, instrFunc ),
131814	132082	FUNCTION(printf, -1, 0, 0, printfFunc ),
131815	132083	FUNCTION(format, -1, 0, 0, printfFunc ),
		@@ -131816,10 +132084,11 @@
131816	132084	FUNCTION(unicode, 1, 0, 0, unicodeFunc ),
131817	132085	FUNCTION(char, -1, 0, 0, charFunc ),
131818	132086	FUNCTION(abs, 1, 0, 0, absFunc ),
131819	132087	#ifdef SQLITE_DEBUG
131820	132088	FUNCTION(fpdecode, 3, 0, 0, fpdecodeFunc ),
	132089	+ FUNCTION(parseuri, -1, 0, 0, parseuriFunc ),
131821	132090	#endif
131822	132091	#ifndef SQLITE_OMIT_FLOATING_POINT
131823	132092	FUNCTION(round, 1, 0, 0, roundFunc ),
131824	132093	FUNCTION(round, 2, 0, 0, roundFunc ),
131825	132094	#endif
		@@ -131910,15 +132179,18 @@
131910	132179	MFUNCTION(atanh, 1, atanh, math1Func ),
131911	132180	#endif
131912	132181	MFUNCTION(sqrt, 1, sqrt, math1Func ),
131913	132182	MFUNCTION(radians, 1, degToRad, math1Func ),
131914	132183	MFUNCTION(degrees, 1, radToDeg, math1Func ),
131915		- FUNCTION(pi, 0, 0, 0, piFunc ),
	132184	+ MFUNCTION(pi, 0, 0, piFunc ),
131916	132185	#endif /* SQLITE_ENABLE_MATH_FUNCTIONS */
131917	132186	FUNCTION(sign, 1, 0, 0, signFunc ),
131918	132187	INLINE_FUNC(coalesce, -1, INLINEFUNC_coalesce, 0 ),
	132188	+ INLINE_FUNC(iif, 2, INLINEFUNC_iif, 0 ),
131919	132189	INLINE_FUNC(iif, 3, INLINEFUNC_iif, 0 ),
	132190	+ INLINE_FUNC(if, 2, INLINEFUNC_iif, 0 ),
	132191	+ INLINE_FUNC(if, 3, INLINEFUNC_iif, 0 ),
131920	132192	};
131921	132193	#ifndef SQLITE_OMIT_ALTERTABLE
131922	132194	sqlite3AlterFunctions();
131923	132195	#endif
131924	132196	sqlite3WindowFunctions();
		@@ -140427,16 +140699,10 @@
140427	140699	if( db->autoCommit==0 ){
140428	140700	/* Foreign key support may not be enabled or disabled while not
140429	140701	** in auto-commit mode. */
140430	140702	mask &= ~(SQLITE_ForeignKeys);
140431	140703	}
140432		-#if SQLITE_USER_AUTHENTICATION
140433		- if( db->auth.authLevel==UAUTH_User ){
140434		- /* Do not allow non-admin users to modify the schema arbitrarily */
140435		- mask &= ~(SQLITE_WriteSchema);
140436		- }
140437		-#endif
140438	140704
140439	140705	if( sqlite3GetBoolean(zRight, 0) ){
140440	140706	if( (mask & SQLITE_WriteSchema)==0
140441	140707	\|\| (db->flags & SQLITE_Defensive)==0
140442	140708	){
		@@ -140568,11 +140834,12 @@
140568	140834	pTab = sqliteHashData(k);
140569	140835	if( pTab->nCol==0 ){
140570	140836	char zSql = sqlite3MPrintf(db, "SELECTFROM\"%w\"", pTab->zName);
140571	140837	if( zSql ){
140572	140838	sqlite3_stmt *pDummy = 0;
140573		- (void)sqlite3_prepare(db, zSql, -1, &pDummy, 0);
	140839	+ (void)sqlite3_prepare_v3(db, zSql, -1, SQLITE_PREPARE_DONT_LOG,
	140840	+ &pDummy, 0);
140574	140841	(void)sqlite3_finalize(pDummy);
140575	140842	sqlite3DbFree(db, zSql);
140576	140843	}
140577	140844	if( db->mallocFailed ){
140578	140845	sqlite3ErrorMsg(db->pParse, "out of memory");
		@@ -141044,10 +141311,11 @@
141044	141311	}
141045	141312	aRoot[0] = cnt;
141046	141313
141047	141314	/* Make sure sufficient number of registers have been allocated */
141048	141315	sqlite3TouchRegister(pParse, 8+cnt);
	141316	+ sqlite3VdbeAddOp3(v, OP_Null, 0, 8, 8+cnt);
141049	141317	sqlite3ClearTempRegCache(pParse);
141050	141318
141051	141319	/* Do the b-tree integrity checks */
141052	141320	sqlite3VdbeAddOp4(v, OP_IntegrityCk, 1, cnt, 8, (char*)aRoot,P4_INTARRAY);
141053	141321	sqlite3VdbeChangeP5(v, (u8)i);
		@@ -142668,18 +142936,11 @@
142668	142936	encoding = (u8)meta[BTREE_TEXT_ENCODING-1] & 3;
142669	142937	if( encoding==0 ) encoding = SQLITE_UTF8;
142670	142938	#else
142671	142939	encoding = SQLITE_UTF8;
142672	142940	#endif
142673		- if( db->nVdbeActive>0 && encoding!=ENC(db)
142674		- && (db->mDbFlags & DBFLAG_Vacuum)==0
142675		- ){
142676		- rc = SQLITE_LOCKED;
142677		- goto initone_error_out;
142678		- }else{
142679		- sqlite3SetTextEncoding(db, encoding);
142680		- }
	142941	+ sqlite3SetTextEncoding(db, encoding);
142681	142942	}else{
142682	142943	/* If opening an attached database, the encoding much match ENC(db) */
142683	142944	if( (meta[BTREE_TEXT_ENCODING-1] & 3)!=ENC(db) ){
142684	142945	sqlite3SetString(pzErrMsg, db, "attached databases must use the same"
142685	142946	" text encoding as main database");
		@@ -143369,16 +143630,28 @@
143369	143630	#endif
143370	143631	*ppStmt = 0;
143371	143632	if( !sqlite3SafetyCheckOk(db)\|\|zSql==0 ){
143372	143633	return SQLITE_MISUSE_BKPT;
143373	143634	}
	143635	+
	143636	+ /* Make sure nBytes is non-negative and correct. It should be the
	143637	+ ** number of bytes until the end of the input buffer or until the first
	143638	+ ** U+0000 character. If the input nBytes is odd, convert it into
	143639	+ ** an even number. If the input nBytes is negative, then the input
	143640	+ ** must be terminated by at least one U+0000 character */
143374	143641	if( nBytes>=0 ){
143375	143642	int sz;
143376	143643	const char z = (const char)zSql;
143377	143644	for(sz=0; sz<nBytes && (z[sz]!=0 \|\| z[sz+1]!=0); sz += 2){}
143378	143645	nBytes = sz;
	143646	+ }else{
	143647	+ int sz;
	143648	+ const char z = (const char)zSql;
	143649	+ for(sz=0; z[sz]!=0 \|\| z[sz+1]!=0; sz += 2){}
	143650	+ nBytes = sz;
143379	143651	}
	143652	+
143380	143653	sqlite3_mutex_enter(db->mutex);
143381	143654	zSql8 = sqlite3Utf16to8(db, zSql, nBytes, SQLITE_UTF16NATIVE);
143382	143655	if( zSql8 ){
143383	143656	rc = sqlite3LockAndPrepare(db, zSql8, -1, prepFlags, 0, ppStmt, &zTail8);
143384	143657	}
		@@ -143388,11 +143661,11 @@
143388	143661	** equivalent pointer into the UTF-16 string by counting the unicode
143389	143662	** characters between zSql8 and zTail8, and then returning a pointer
143390	143663	** the same number of characters into the UTF-16 string.
143391	143664	*/
143392	143665	int chars_parsed = sqlite3Utf8CharLen(zSql8, (int)(zTail8-zSql8));
143393		- pzTail = (u8 )zSql + sqlite3Utf16ByteLen(zSql, chars_parsed);
	143666	+ pzTail = (u8 )zSql + sqlite3Utf16ByteLen(zSql, nBytes, chars_parsed);
143394	143667	}
143395	143668	sqlite3DbFree(db, zSql8);
143396	143669	rc = sqlite3ApiExit(db, rc);
143397	143670	sqlite3_mutex_leave(db->mutex);
143398	143671	return rc;
		@@ -147361,36 +147634,36 @@
147361	147634	return pExpr;
147362	147635	}
147363	147636	if( pSubst->isOuterJoin ){
147364	147637	ExprSetProperty(pNew, EP_CanBeNull);
147365	147638	}
	147639	+ if( pNew->op==TK_TRUEFALSE ){
	147640	+ pNew->u.iValue = sqlite3ExprTruthValue(pNew);
	147641	+ pNew->op = TK_INTEGER;
	147642	+ ExprSetProperty(pNew, EP_IntValue);
	147643	+ }
	147644	+
	147645	+ /* Ensure that the expression now has an implicit collation sequence,
	147646	+ ** just as it did when it was a column of a view or sub-query. */
	147647	+ {
	147648	+ CollSeq *pNat = sqlite3ExprCollSeq(pSubst->pParse, pNew);
	147649	+ CollSeq *pColl = sqlite3ExprCollSeq(pSubst->pParse,
	147650	+ pSubst->pCList->a[iColumn].pExpr
	147651	+ );
	147652	+ if( pNat!=pColl \|\| (pNew->op!=TK_COLUMN && pNew->op!=TK_COLLATE) ){
	147653	+ pNew = sqlite3ExprAddCollateString(pSubst->pParse, pNew,
	147654	+ (pColl ? pColl->zName : "BINARY")
	147655	+ );
	147656	+ }
	147657	+ }
	147658	+ ExprClearProperty(pNew, EP_Collate);
147366	147659	if( ExprHasProperty(pExpr,EP_OuterON\|EP_InnerON) ){
147367	147660	sqlite3SetJoinExpr(pNew, pExpr->w.iJoin,
147368	147661	pExpr->flags & (EP_OuterON\|EP_InnerON));
147369	147662	}
147370	147663	sqlite3ExprDelete(db, pExpr);
147371	147664	pExpr = pNew;
147372		- if( pExpr->op==TK_TRUEFALSE ){
147373		- pExpr->u.iValue = sqlite3ExprTruthValue(pExpr);
147374		- pExpr->op = TK_INTEGER;
147375		- ExprSetProperty(pExpr, EP_IntValue);
147376		- }
147377		-
147378		- /* Ensure that the expression now has an implicit collation sequence,
147379		- ** just as it did when it was a column of a view or sub-query. */
147380		- {
147381		- CollSeq *pNat = sqlite3ExprCollSeq(pSubst->pParse, pExpr);
147382		- CollSeq *pColl = sqlite3ExprCollSeq(pSubst->pParse,
147383		- pSubst->pCList->a[iColumn].pExpr
147384		- );
147385		- if( pNat!=pColl \|\| (pExpr->op!=TK_COLUMN && pExpr->op!=TK_COLLATE) ){
147386		- pExpr = sqlite3ExprAddCollateString(pSubst->pParse, pExpr,
147387		- (pColl ? pColl->zName : "BINARY")
147388		- );
147389		- }
147390		- }
147391		- ExprClearProperty(pExpr, EP_Collate);
147392	147665	}
147393	147666	}
147394	147667	}else{
147395	147668	if( pExpr->op==TK_IF_NULL_ROW && pExpr->iTable==pSubst->iTable ){
147396	147669	pExpr->iTable = pSubst->iNewTable;
		@@ -148123,20 +148396,20 @@
148123	148396	}
148124	148397
148125	148398	/* Transfer the FROM clause terms from the subquery into the
148126	148399	** outer query.
148127	148400	*/
	148401	+ iNewParent = pSubSrc->a[0].iCursor;
148128	148402	for(i=0; i<nSubSrc; i++){
148129	148403	SrcItem *pItem = &pSrc->a[i+iFrom];
148130	148404	assert( pItem->fg.isTabFunc==0 );
148131	148405	assert( pItem->fg.isSubquery
148132	148406	\|\| pItem->fg.fixedSchema
148133	148407	\|\| pItem->u4.zDatabase==0 );
148134	148408	if( pItem->fg.isUsing ) sqlite3IdListDelete(db, pItem->u3.pUsing);
148135	148409	*pItem = pSubSrc->a[i];
148136	148410	pItem->fg.jointype \|= ltorj;
148137		- iNewParent = pSubSrc->a[i].iCursor;
148138	148411	memset(&pSubSrc->a[i], 0, sizeof(pSubSrc->a[i]));
148139	148412	}
148140	148413	pSrc->a[iFrom].fg.jointype &= JT_LTORJ;
148141	148414	pSrc->a[iFrom].fg.jointype \|= jointype \| ltorj;
148142	148415
		@@ -148172,10 +148445,11 @@
148172	148445	pSub->pOrderBy = 0;
148173	148446	}
148174	148447	pWhere = pSub->pWhere;
148175	148448	pSub->pWhere = 0;
148176	148449	if( isOuterJoin>0 ){
	148450	+ assert( pSubSrc->nSrc==1 );
148177	148451	sqlite3SetJoinExpr(pWhere, iNewParent, EP_OuterON);
148178	148452	}
148179	148453	if( pWhere ){
148180	148454	if( pParent->pWhere ){
148181	148455	pParent->pWhere = sqlite3PExpr(pParse, TK_AND, pWhere, pParent->pWhere);
		@@ -151271,11 +151545,11 @@
151271	151545	sqlite3TreeViewSelect(0, p, 0);
151272	151546	}
151273	151547	#endif
151274	151548	assert( pSubq->pSelect && (pSub->selFlags & SF_PushDown)!=0 );
151275	151549	}else{
151276		- TREETRACE(0x4000,pParse,p,("WHERE-lcause push-down not possible\n"));
	151550	+ TREETRACE(0x4000,pParse,p,("WHERE-clause push-down not possible\n"));
151277	151551	}
151278	151552
151279	151553	/* Convert unused result columns of the subquery into simple NULL
151280	151554	** expressions, to avoid unneeded searching and computation.
151281	151555	** tag-select-0440
		@@ -156980,10 +157254,11 @@
156980	157254	assert( sParse.zErrMsg==0 );
156981	157255	if( !pTab->aCol ){
156982	157256	Table *pNew = sParse.pNewTable;
156983	157257	Index *pIdx;
156984	157258	pTab->aCol = pNew->aCol;
	157259	+ assert( IsOrdinaryTable(pNew) );
156985	157260	sqlite3ExprListDelete(db, pNew->u.tab.pDfltList);
156986	157261	pTab->nNVCol = pTab->nCol = pNew->nCol;
156987	157262	pTab->tabFlags \|= pNew->tabFlags & (TF_WithoutRowid\|TF_NoVisibleRowid);
156988	157263	pNew->nCol = 0;
156989	157264	pNew->aCol = 0;
		@@ -158044,13 +158319,21 @@
158044	158319	SQLITE_PRIVATE int sqlite3WhereExplainBloomFilter(
158045	158320	const Parse pParse, / Parse context */
158046	158321	const WhereInfo pWInfo, / WHERE clause */
158047	158322	const WhereLevel pLevel / Bloom filter on this level */
158048	158323	);
	158324	+SQLITE_PRIVATE void sqlite3WhereAddExplainText(
	158325	+ Parse pParse, / Parse context */
	158326	+ int addr,
	158327	+ SrcList pTabList, / Table list this loop refers to */
	158328	+ WhereLevel pLevel, / Scan to write OP_Explain opcode for */
	158329	+ u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
	158330	+);
158049	158331	#else
158050	158332	# define sqlite3WhereExplainOneScan(u,v,w,x) 0
158051	158333	# define sqlite3WhereExplainBloomFilter(u,v,w) 0
	158334	+# define sqlite3WhereAddExplainText(u,v,w,x,y)
158052	158335	#endif /* SQLITE_OMIT_EXPLAIN */
158053	158336	#ifdef SQLITE_ENABLE_STMT_SCANSTATUS
158054	158337	SQLITE_PRIVATE void sqlite3WhereAddScanStatus(
158055	158338	Vdbe v, / Vdbe to add scanstatus entry to */
158056	158339	SrcList pSrclist, / FROM clause pLvl reads data from */
		@@ -158248,42 +158531,42 @@
158248	158531	}
158249	158532	sqlite3_str_append(pStr, ")", 1);
158250	158533	}
158251	158534
158252	158535	/*
158253		-** This function is a no-op unless currently processing an EXPLAIN QUERY PLAN
158254		-** command, or if stmt_scanstatus_v2() stats are enabled, or if SQLITE_DEBUG
158255		-** was defined at compile-time. If it is not a no-op, a single OP_Explain
158256		-** opcode is added to the output to describe the table scan strategy in pLevel.
158257		-**
158258		-** If an OP_Explain opcode is added to the VM, its address is returned.
158259		-** Otherwise, if no OP_Explain is coded, zero is returned.
	158536	+** This function sets the P4 value of an existing OP_Explain opcode to
	158537	+** text describing the loop in pLevel. If the OP_Explain opcode already has
	158538	+** a P4 value, it is freed before it is overwritten.
158260	158539	*/
158261		-SQLITE_PRIVATE int sqlite3WhereExplainOneScan(
	158540	+SQLITE_PRIVATE void sqlite3WhereAddExplainText(
158262	158541	Parse pParse, / Parse context */
	158542	+ int addr, /* Address of OP_Explain opcode */
158263	158543	SrcList pTabList, / Table list this loop refers to */
158264	158544	WhereLevel pLevel, / Scan to write OP_Explain opcode for */
158265	158545	u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
158266	158546	){
158267		- int ret = 0;
158268	158547	#if !defined(SQLITE_DEBUG)
158269	158548	if( sqlite3ParseToplevel(pParse)->explain==2 \|\| IS_STMT_SCANSTATUS(pParse->db) )
158270	158549	#endif
158271	158550	{
	158551	+ VdbeOp *pOp = sqlite3VdbeGetOp(pParse->pVdbe, addr);
	158552	+
158272	158553	SrcItem *pItem = &pTabList->a[pLevel->iFrom];
158273		- Vdbe v = pParse->pVdbe; / VM being constructed */
158274	158554	sqlite3 db = pParse->db; / Database handle */
158275	158555	int isSearch; /* True for a SEARCH. False for SCAN. */
158276	158556	WhereLoop pLoop; / The controlling WhereLoop object */
158277	158557	u32 flags; /* Flags that describe this loop */
	158558	+#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_EXPLAIN)
158278	158559	char zMsg; / Text to add to EQP output */
	158560	+#endif
158279	158561	StrAccum str; /* EQP output string */
158280	158562	char zBuf[100]; /* Initial space for EQP output string */
158281	158563
	158564	+ if( db->mallocFailed ) return;
	158565	+
158282	158566	pLoop = pLevel->pWLoop;
158283	158567	flags = pLoop->wsFlags;
158284		- if( (flags&WHERE_MULTI_OR) \|\| (wctrlFlags&WHERE_OR_SUBCLAUSE) ) return 0;
158285	158568
158286	158569	isSearch = (flags&(WHERE_BTM_LIMIT\|WHERE_TOP_LIMIT))!=0
158287	158570	\|\| ((flags&WHERE_VIRTUALTABLE)==0 && (pLoop->u.btree.nEq>0))
158288	158571	\|\| (wctrlFlags&(WHERE_ORDERBY_MIN\|WHERE_ORDERBY_MAX));
158289	158572
		@@ -158303,11 +158586,11 @@
158303	158586	}
158304	158587	}else if( flags & WHERE_PARTIALIDX ){
158305	158588	zFmt = "AUTOMATIC PARTIAL COVERING INDEX";
158306	158589	}else if( flags & WHERE_AUTO_INDEX ){
158307	158590	zFmt = "AUTOMATIC COVERING INDEX";
158308		- }else if( flags & WHERE_IDX_ONLY ){
	158591	+ }else if( flags & (WHERE_IDX_ONLY\|WHERE_EXPRIDX) ){
158309	158592	zFmt = "COVERING INDEX %s";
158310	158593	}else{
158311	158594	zFmt = "INDEX %s";
158312	158595	}
158313	158596	if( zFmt ){
		@@ -158355,15 +158638,54 @@
158355	158638	sqlite3LogEstToInt(pLoop->nOut));
158356	158639	}else{
158357	158640	sqlite3_str_append(&str, " (~1 row)", 9);
158358	158641	}
158359	158642	#endif
	158643	+#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_EXPLAIN)
158360	158644	zMsg = sqlite3StrAccumFinish(&str);
158361	158645	sqlite3ExplainBreakpoint("",zMsg);
158362		- ret = sqlite3VdbeAddOp4(v, OP_Explain, sqlite3VdbeCurrentAddr(v),
158363		- pParse->addrExplain, pLoop->rRun,
158364		- zMsg, P4_DYNAMIC);
	158646	+#endif
	158647	+
	158648	+ assert( pOp->opcode==OP_Explain );
	158649	+ assert( pOp->p4type==P4_DYNAMIC \|\| pOp->p4.z==0 );
	158650	+ sqlite3DbFree(db, pOp->p4.z);
	158651	+ pOp->p4type = P4_DYNAMIC;
	158652	+ pOp->p4.z = sqlite3StrAccumFinish(&str);
	158653	+ }
	158654	+}
	158655	+
	158656	+
	158657	+/*
	158658	+** This function is a no-op unless currently processing an EXPLAIN QUERY PLAN
	158659	+** command, or if stmt_scanstatus_v2() stats are enabled, or if SQLITE_DEBUG
	158660	+** was defined at compile-time. If it is not a no-op, a single OP_Explain
	158661	+** opcode is added to the output to describe the table scan strategy in pLevel.
	158662	+**
	158663	+** If an OP_Explain opcode is added to the VM, its address is returned.
	158664	+** Otherwise, if no OP_Explain is coded, zero is returned.
	158665	+*/
	158666	+SQLITE_PRIVATE int sqlite3WhereExplainOneScan(
	158667	+ Parse pParse, / Parse context */
	158668	+ SrcList pTabList, / Table list this loop refers to */
	158669	+ WhereLevel pLevel, / Scan to write OP_Explain opcode for */
	158670	+ u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
	158671	+){
	158672	+ int ret = 0;
	158673	+#if !defined(SQLITE_DEBUG)
	158674	+ if( sqlite3ParseToplevel(pParse)->explain==2 \|\| IS_STMT_SCANSTATUS(pParse->db) )
	158675	+#endif
	158676	+ {
	158677	+ if( (pLevel->pWLoop->wsFlags & WHERE_MULTI_OR)==0
	158678	+ && (wctrlFlags & WHERE_OR_SUBCLAUSE)==0
	158679	+ ){
	158680	+ Vdbe *v = pParse->pVdbe;
	158681	+ int addr = sqlite3VdbeCurrentAddr(v);
	158682	+ ret = sqlite3VdbeAddOp3(
	158683	+ v, OP_Explain, addr, pParse->addrExplain, pLevel->pWLoop->rRun
	158684	+ );
	158685	+ sqlite3WhereAddExplainText(pParse, addr, pTabList, pLevel, wctrlFlags);
	158686	+ }
158365	158687	}
158366	158688	return ret;
158367	158689	}
158368	158690
158369	158691	/*
		@@ -158458,13 +158780,14 @@
158458	158780	if( wsFlags & WHERE_INDEXED ){
158459	158781	sqlite3VdbeScanStatusRange(v, addrExplain, -1, pLvl->iIdxCur);
158460	158782	}
158461	158783	}else{
158462	158784	int addr;
	158785	+ VdbeOp *pOp;
158463	158786	assert( pSrclist->a[pLvl->iFrom].fg.isSubquery );
158464	158787	addr = pSrclist->a[pLvl->iFrom].u4.pSubq->addrFillSub;
158465		- VdbeOp *pOp = sqlite3VdbeGetOp(v, addr-1);
	158788	+ pOp = sqlite3VdbeGetOp(v, addr-1);
158466	158789	assert( sqlite3VdbeDb(v)->mallocFailed \|\| pOp->opcode==OP_InitCoroutine );
158467	158790	assert( sqlite3VdbeDb(v)->mallocFailed \|\| pOp->p2>addr );
158468	158791	sqlite3VdbeScanStatusRange(v, addrExplain, addr, pOp->p2-1);
158469	158792	}
158470	158793	}
		@@ -158713,10 +159036,11 @@
158713	159036	if( pOrigLhs ){
158714	159037	sqlite3ExprListDelete(db, pOrigLhs);
158715	159038	pNew->pLeft->x.pList = pLhs;
158716	159039	}
158717	159040	pSelect->pEList = pRhs;
	159041	+ pSelect->selId = ++pParse->nSelect; /* Req'd for SubrtnSig validity */
158718	159042	if( pLhs && pLhs->nExpr==1 ){
158719	159043	/* Take care here not to generate a TK_VECTOR containing only a
158720	159044	** single value. Since the parser never creates such a vector, some
158721	159045	** of the subroutines do not handle this case. */
158722	159046	Expr *p = pLhs->a[0].pExpr;
		@@ -161260,24 +161584,29 @@
161260	161584	}else if( op==TK_STRING ){
161261	161585	assert( !ExprHasProperty(pRight, EP_IntValue) );
161262	161586	z = (u8*)pRight->u.zToken;
161263	161587	}
161264	161588	if( z ){
161265		-
161266		- /* Count the number of prefix characters prior to the first wildcard.
161267		- ** If the underlying database has a UTF16LE encoding, then only consider
161268		- ** ASCII characters. Note that the encoding of z[] is UTF8 - we are
161269		- ** dealing with only UTF8 here in this code, but the database engine
161270		- ** itself might be processing content using a different encoding. */
	161589	+ /* Count the number of prefix bytes prior to the first wildcard.
	161590	+ ** or U+fffd character. If the underlying database has a UTF16LE
	161591	+ ** encoding, then only consider ASCII characters. Note that the
	161592	+ ** encoding of z[] is UTF8 - we are dealing with only UTF8 here in
	161593	+ ** this code, but the database engine itself might be processing
	161594	+ ** content using a different encoding. */
161271	161595	cnt = 0;
161272	161596	while( (c=z[cnt])!=0 && c!=wc[0] && c!=wc[1] && c!=wc[2] ){
161273	161597	cnt++;
161274		- if( c==wc[3] && z[cnt]!=0 ){
	161598	+ if( c==wc[3] && z[cnt]>0 && z[cnt]<0x80 ){
161275	161599	cnt++;
161276		- }else if( c>=0x80 && ENC(db)==SQLITE_UTF16LE ){
161277		- cnt--;
161278		- break;
	161600	+ }else if( c>=0x80 ){
	161601	+ const u8 *z2 = z+cnt-1;
	161602	+ if( sqlite3Utf8Read(&z2)==0xfffd \|\| ENC(db)==SQLITE_UTF16LE ){
	161603	+ cnt--;
	161604	+ break;
	161605	+ }else{
	161606	+ cnt = (int)(z2-z);
	161607	+ }
161279	161608	}
161280	161609	}
161281	161610
161282	161611	/* The optimization is possible only if (1) the pattern does not begin
161283	161612	** with a wildcard and if (2) the non-wildcard prefix does not end with
		@@ -161285,11 +161614,11 @@
161285	161614	** a single escape character. The second condition is necessary so
161286	161615	** that we can increment the prefix key to find an upper bound for the
161287	161616	** range search. The third is because the caller assumes that the pattern
161288	161617	** consists of at least one character after all escapes have been
161289	161618	** removed. */
161290		- if( (cnt>1 \|\| (cnt>0 && z[0]!=wc[3])) && 255!=(u8)z[cnt-1] ){
	161619	+ if( (cnt>1 \|\| (cnt>0 && z[0]!=wc[3])) && ALWAYS(255!=(u8)z[cnt-1]) ){
161291	161620	Expr *pPrefix;
161292	161621
161293	161622	/* A "complete" match if the pattern ends with "" or "%" /
161294	161623	*pisComplete = c==wc[0] && z[cnt+1]==0 && ENC(db)!=SQLITE_UTF16LE;
161295	161624
		@@ -163780,11 +164109,11 @@
163780	164109	\|\| pTerm->pExpr->w.iJoin != pSrc->iCursor
163781	164110	){
163782	164111	return 0;
163783	164112	}
163784	164113	if( (pSrc->fg.jointype & (JT_LEFT\|JT_RIGHT))!=0
163785		- && ExprHasProperty(pTerm->pExpr, EP_InnerON)
	164114	+ && NEVER(ExprHasProperty(pTerm->pExpr, EP_InnerON))
163786	164115	){
163787	164116	return 0;
163788	164117	}
163789	164118	return 1;
163790	164119	}
		@@ -164577,13 +164906,15 @@
164577	164906	** Whether or not an error is returned, it is the responsibility of the
164578	164907	** caller to eventually free p->idxStr if p->needToFreeIdxStr indicates
164579	164908	** that this is required.
164580	164909	*/
164581	164910	static int vtabBestIndex(Parse pParse, Table pTab, sqlite3_index_info *p){
164582		- sqlite3_vtab *pVtab = sqlite3GetVTable(pParse->db, pTab)->pVtab;
164583	164911	int rc;
	164912	+ sqlite3_vtab *pVtab;
164584	164913
	164914	+ assert( IsVirtual(pTab) );
	164915	+ pVtab = sqlite3GetVTable(pParse->db, pTab)->pVtab;
164585	164916	whereTraceIndexInfoInputs(p, pTab);
164586	164917	pParse->db->nSchemaLock++;
164587	164918	rc = pVtab->pModule->xBestIndex(pVtab, p);
164588	164919	pParse->db->nSchemaLock--;
164589	164920	whereTraceIndexInfoOutputs(p, pTab);
		@@ -165271,11 +165602,11 @@
165271	165602	return rc;
165272	165603	}
165273	165604	#endif /* SQLITE_ENABLE_STAT4 */
165274	165605
165275	165606
165276		-#ifdef WHERETRACE_ENABLED
	165607	+#if defined(WHERETRACE_ENABLED) \|\| defined(SQLITE_DEBUG)
165277	165608	/*
165278	165609	** Print the content of a WhereTerm object
165279	165610	*/
165280	165611	SQLITE_PRIVATE void sqlite3WhereTermPrint(WhereTerm *pTerm, int iTerm){
165281	165612	if( pTerm==0 ){
		@@ -165314,10 +165645,13 @@
165314	165645	sqlite3DebugPrintf(" iParent=%d", pTerm->iParent);
165315	165646	}
165316	165647	sqlite3DebugPrintf("\n");
165317	165648	sqlite3TreeViewExpr(0, pTerm->pExpr, 0);
165318	165649	}
	165650	+}
	165651	+SQLITE_PRIVATE void sqlite3ShowWhereTerm(WhereTerm *pTerm){
	165652	+ sqlite3WhereTermPrint(pTerm, 0);
165319	165653	}
165320	165654	#endif
165321	165655
165322	165656	#ifdef WHERETRACE_ENABLED
165323	165657	/*
		@@ -165522,11 +165856,11 @@
165522	165856	** Return TRUE if X is a proper subset of Y but is of equal or less cost.
165523	165857	** In other words, return true if all constraints of X are also part of Y
165524	165858	** and Y has additional constraints that might speed the search that X lacks
165525	165859	** but the cost of running X is not more than the cost of running Y.
165526	165860	**
165527		-** In other words, return true if the cost relationwship between X and Y
	165861	+** In other words, return true if the cost relationship between X and Y
165528	165862	** is inverted and needs to be adjusted.
165529	165863	**
165530	165864	** Case 1:
165531	165865	**
165532	165866	** (1a) X and Y use the same index.
		@@ -166500,11 +166834,10 @@
166500	166834	pParse = pWC->pWInfo->pParse;
166501	166835	while( pWhere->op==TK_AND ){
166502	166836	if( !whereUsablePartialIndex(iTab,jointype,pWC,pWhere->pLeft) ) return 0;
166503	166837	pWhere = pWhere->pRight;
166504	166838	}
166505		- if( pParse->db->flags & SQLITE_EnableQPSG ) pParse = 0;
166506	166839	for(i=0, pTerm=pWC->a; i<pWC->nTerm; i++, pTerm++){
166507	166840	Expr *pExpr;
166508	166841	pExpr = pTerm->pExpr;
166509	166842	if( (!ExprHasProperty(pExpr, EP_OuterON) \|\| pExpr->w.iJoin==iTab)
166510	166843	&& ((jointype & JT_OUTER)==0 \|\| ExprHasProperty(pExpr, EP_OuterON))
		@@ -169140,10 +169473,11 @@
169140	169473	hasRightJoin = (pWInfo->pTabList->a[0].fg.jointype & JT_LTORJ)!=0;
169141	169474	for(i=pWInfo->nLevel-1; i>=1; i--){
169142	169475	WhereTerm pTerm, pEnd;
169143	169476	SrcItem *pItem;
169144	169477	WhereLoop *pLoop;
	169478	+ Bitmask m1;
169145	169479	pLoop = pWInfo->a[i].pWLoop;
169146	169480	pItem = &pWInfo->pTabList->a[pLoop->iTab];
169147	169481	if( (pItem->fg.jointype & (JT_LEFT\|JT_RIGHT))!=JT_LEFT ) continue;
169148	169482	if( (pWInfo->wctrlFlags & WHERE_WANT_DISTINCT)==0
169149	169483	&& (pLoop->wsFlags & WHERE_ONEROW)==0
		@@ -169160,17 +169494,20 @@
169160	169494	break;
169161	169495	}
169162	169496	}
169163	169497	if( hasRightJoin
169164	169498	&& ExprHasProperty(pTerm->pExpr, EP_InnerON)
169165		- && pTerm->pExpr->w.iJoin==pItem->iCursor
	169499	+ && NEVER(pTerm->pExpr->w.iJoin==pItem->iCursor)
169166	169500	){
169167	169501	break; /* restriction (5) */
169168	169502	}
169169	169503	}
169170	169504	if( pTerm<pEnd ) continue;
169171		- WHERETRACE(0xffffffff, ("-> drop loop %c not used\n", pLoop->cId));
	169505	+ WHERETRACE(0xffffffff,("-> omit unused FROM-clause term %c\n",pLoop->cId));
	169506	+ m1 = MASKBIT(i)-1;
	169507	+ testcase( ((pWInfo->revMask>>1) & ~m1)!=0 );
	169508	+ pWInfo->revMask = (m1 & pWInfo->revMask) \| ((pWInfo->revMask>>1) & ~m1);
169172	169509	notReady &= ~pLoop->maskSelf;
169173	169510	for(pTerm=pWInfo->sWC.a; pTerm<pEnd; pTerm++){
169174	169511	if( (pTerm->prereqAll & pLoop->maskSelf)!=0 ){
169175	169512	pTerm->wtFlags \|= TERM_CODED;
169176	169513	}
		@@ -169237,62 +169574,10 @@
169237	169574	nSearch += pLoop->nOut;
169238	169575	if( pWInfo->nOutStarDelta ) nSearch += pLoop->rStarDelta;
169239	169576	}
169240	169577	}
169241	169578
169242		-/*
169243		-** Expression Node callback for sqlite3ExprCanReturnSubtype().
169244		-**
169245		-** Only a function call is able to return a subtype. So if the node
169246		-** is not a function call, return WRC_Prune immediately.
169247		-**
169248		-** A function call is able to return a subtype if it has the
169249		-** SQLITE_RESULT_SUBTYPE property.
169250		-**
169251		-** Assume that every function is able to pass-through a subtype from
169252		-** one of its argument (using sqlite3_result_value()). Most functions
169253		-** are not this way, but we don't have a mechanism to distinguish those
169254		-** that are from those that are not, so assume they all work this way.
169255		-** That means that if one of its arguments is another function and that
169256		-** other function is able to return a subtype, then this function is
169257		-** able to return a subtype.
169258		-*/
169259		-static int exprNodeCanReturnSubtype(Walker pWalker, Expr pExpr){
169260		- int n;
169261		- FuncDef *pDef;
169262		- sqlite3 *db;
169263		- if( pExpr->op!=TK_FUNCTION ){
169264		- return WRC_Prune;
169265		- }
169266		- assert( ExprUseXList(pExpr) );
169267		- db = pWalker->pParse->db;
169268		- n = pExpr->x.pList ? pExpr->x.pList->nExpr : 0;
169269		- pDef = sqlite3FindFunction(db, pExpr->u.zToken, n, ENC(db), 0);
169270		- if( pDef==0 \|\| (pDef->funcFlags & SQLITE_RESULT_SUBTYPE)!=0 ){
169271		- pWalker->eCode = 1;
169272		- return WRC_Prune;
169273		- }
169274		- return WRC_Continue;
169275		-}
169276		-
169277		-/*
169278		-** Return TRUE if expression pExpr is able to return a subtype.
169279		-**
169280		-** A TRUE return does not guarantee that a subtype will be returned.
169281		-** It only indicates that a subtype return is possible. False positives
169282		-** are acceptable as they only disable an optimization. False negatives,
169283		-** on the other hand, can lead to incorrect answers.
169284		-*/
169285		-static int sqlite3ExprCanReturnSubtype(Parse pParse, Expr pExpr){
169286		- Walker w;
169287		- memset(&w, 0, sizeof(w));
169288		- w.pParse = pParse;
169289		- w.xExprCallback = exprNodeCanReturnSubtype;
169290		- sqlite3WalkExpr(&w, pExpr);
169291		- return w.eCode;
169292		-}
169293		-
169294	169579	/*
169295	169580	** The index pIdx is used by a query and contains one or more expressions.
169296	169581	** In other words pIdx is an index on an expression. iIdxCur is the cursor
169297	169582	** number for the index and iDataCur is the cursor number for the corresponding
169298	169583	** table.
		@@ -169322,16 +169607,10 @@
169322	169607	pExpr = sqlite3ColumnExpr(pTab, &pTab->aCol[j]);
169323	169608	}else{
169324	169609	continue;
169325	169610	}
169326	169611	if( sqlite3ExprIsConstant(0,pExpr) ) continue;
169327		- if( pExpr->op==TK_FUNCTION && sqlite3ExprCanReturnSubtype(pParse,pExpr) ){
169328		- /* Functions that might set a subtype should not be replaced by the
169329		- ** value taken from an expression index since the index omits the
169330		- ** subtype. https://sqlite.org/forum/forumpost/68d284c86b082c3e */
169331		- continue;
169332		- }
169333	169612	p = sqlite3DbMallocRaw(pParse->db, sizeof(IndexedExpr));
169334	169613	if( p==0 ) break;
169335	169614	p->pIENext = pParse->pIdxEpr;
169336	169615	#ifdef WHERETRACE_ENABLED
169337	169616	if( sqlite3WhereTrace & 0x200 ){
		@@ -170434,18 +170713,32 @@
170434	170713	x = sqlite3TableColumnToIndex(pIdx, x);
170435	170714	if( x>=0 ){
170436	170715	pOp->p2 = x;
170437	170716	pOp->p1 = pLevel->iIdxCur;
170438	170717	OpcodeRewriteTrace(db, k, pOp);
170439		- }else{
170440		- /* Unable to translate the table reference into an index
170441		- ** reference. Verify that this is harmless - that the
170442		- ** table being referenced really is open.
170443		- */
	170718	+ }else if( pLoop->wsFlags & (WHERE_IDX_ONLY\|WHERE_EXPRIDX) ){
170444	170719	if( pLoop->wsFlags & WHERE_IDX_ONLY ){
	170720	+ /* An error. pLoop is supposed to be a covering index loop,
	170721	+ ** and yet the VM code refers to a column of the table that
	170722	+ ** is not part of the index. */
170445	170723	sqlite3ErrorMsg(pParse, "internal query planner error");
170446	170724	pParse->rc = SQLITE_INTERNAL;
	170725	+ }else{
	170726	+ /* The WHERE_EXPRIDX flag is set by the planner when it is likely
	170727	+ ** that pLoop is a covering index loop, but it is not possible
	170728	+ ** to be 100% sure. In this case, any OP_Explain opcode
	170729	+ ** corresponding to this loop describes the index as a "COVERING
	170730	+ ** INDEX". But, pOp proves that pLoop is not actually a covering
	170731	+ ** index loop. So clear the WHERE_EXPRIDX flag and rewrite the
	170732	+ ** text that accompanies the OP_Explain opcode, if any. */
	170733	+ pLoop->wsFlags &= ~WHERE_EXPRIDX;
	170734	+ sqlite3WhereAddExplainText(pParse,
	170735	+ pLevel->addrBody-1,
	170736	+ pTabList,
	170737	+ pLevel,
	170738	+ pWInfo->wctrlFlags
	170739	+ );
170447	170740	}
170448	170741	}
170449	170742	}else if( pOp->opcode==OP_Rowid ){
170450	170743	pOp->p1 = pLevel->iIdxCur;
170451	170744	pOp->opcode = OP_IdxRowid;
		@@ -172149,10 +172442,11 @@
172149	172442	for(pWin=pMWin; pWin; pWin=pWin->pNextWin){
172150	172443	FuncDef *pFunc = pWin->pWFunc;
172151	172444	int regArg;
172152	172445	int nArg = pWin->bExprArgs ? 0 : windowArgCount(pWin);
172153	172446	int i;
	172447	+ int addrIf = 0;
172154	172448
172155	172449	assert( bInverse==0 \|\| pWin->eStart!=TK_UNBOUNDED );
172156	172450
172157	172451	/* All OVER clauses in the same window function aggregate step must
172158	172452	** be the same. */
		@@ -172164,10 +172458,22 @@
172164	172458	}else{
172165	172459	sqlite3VdbeAddOp3(v, OP_Column, pMWin->iEphCsr, pWin->iArgCol+i, reg+i);
172166	172460	}
172167	172461	}
172168	172462	regArg = reg;
	172463	+
	172464	+ if( pWin->pFilter ){
	172465	+ int regTmp;
	172466	+ assert( ExprUseXList(pWin->pOwner) );
	172467	+ assert( pWin->bExprArgs \|\| !nArg \|\|nArg==pWin->pOwner->x.pList->nExpr );
	172468	+ assert( pWin->bExprArgs \|\| nArg \|\|pWin->pOwner->x.pList==0 );
	172469	+ regTmp = sqlite3GetTempReg(pParse);
	172470	+ sqlite3VdbeAddOp3(v, OP_Column, csr, pWin->iArgCol+nArg,regTmp);
	172471	+ addrIf = sqlite3VdbeAddOp3(v, OP_IfNot, regTmp, 0, 1);
	172472	+ VdbeCoverage(v);
	172473	+ sqlite3ReleaseTempReg(pParse, regTmp);
	172474	+ }
172169	172475
172170	172476	if( pMWin->regStartRowid==0
172171	172477	&& (pFunc->funcFlags & SQLITE_FUNC_MINMAX)
172172	172478	&& (pWin->eStart!=TK_UNBOUNDED)
172173	172479	){
		@@ -172184,29 +172490,17 @@
172184	172490	sqlite3VdbeAddOp1(v, OP_Delete, pWin->csrApp);
172185	172491	sqlite3VdbeJumpHere(v, sqlite3VdbeCurrentAddr(v)-2);
172186	172492	}
172187	172493	sqlite3VdbeJumpHere(v, addrIsNull);
172188	172494	}else if( pWin->regApp ){
	172495	+ assert( pWin->pFilter==0 );
172189	172496	assert( pFunc->zName==nth_valueName
172190	172497	\|\| pFunc->zName==first_valueName
172191	172498	);
172192	172499	assert( bInverse==0 \|\| bInverse==1 );
172193	172500	sqlite3VdbeAddOp2(v, OP_AddImm, pWin->regApp+1-bInverse, 1);
172194	172501	}else if( pFunc->xSFunc!=noopStepFunc ){
172195		- int addrIf = 0;
172196		- if( pWin->pFilter ){
172197		- int regTmp;
172198		- assert( ExprUseXList(pWin->pOwner) );
172199		- assert( pWin->bExprArgs \|\| !nArg \|\|nArg==pWin->pOwner->x.pList->nExpr );
172200		- assert( pWin->bExprArgs \|\| nArg \|\|pWin->pOwner->x.pList==0 );
172201		- regTmp = sqlite3GetTempReg(pParse);
172202		- sqlite3VdbeAddOp3(v, OP_Column, csr, pWin->iArgCol+nArg,regTmp);
172203		- addrIf = sqlite3VdbeAddOp3(v, OP_IfNot, regTmp, 0, 1);
172204		- VdbeCoverage(v);
172205		- sqlite3ReleaseTempReg(pParse, regTmp);
172206		- }
172207		-
172208	172502	if( pWin->bExprArgs ){
172209	172503	int iOp = sqlite3VdbeCurrentAddr(v);
172210	172504	int iEnd;
172211	172505
172212	172506	assert( ExprUseXList(pWin->pOwner) );
		@@ -172233,12 +172527,13 @@
172233	172527	sqlite3VdbeAppendP4(v, pFunc, P4_FUNCDEF);
172234	172528	sqlite3VdbeChangeP5(v, (u8)nArg);
172235	172529	if( pWin->bExprArgs ){
172236	172530	sqlite3ReleaseTempRange(pParse, regArg, nArg);
172237	172531	}
172238		- if( addrIf ) sqlite3VdbeJumpHere(v, addrIf);
172239	172532	}
	172533	+
	172534	+ if( addrIf ) sqlite3VdbeJumpHere(v, addrIf);
172240	172535	}
172241	172536	}
172242	172537
172243	172538	/*
172244	172539	** Values that may be passed as the second argument to windowCodeOp().
		@@ -173660,10 +173955,17 @@
173660	173955	** Then the "b" IdList records the list "a,b,c".
173661	173956	*/
173662	173957	struct TrigEvent { int a; IdList * b; };
173663	173958
173664	173959	struct FrameBound { int eType; Expr *pExpr; };
	173960	+
	173961	+/*
	173962	+** Generate a syntax error
	173963	+*/
	173964	+static void parserSyntaxError(Parse pParse, Token p){
	173965	+ sqlite3ErrorMsg(pParse, "near \"%T\": syntax error", p);
	173966	+}
173665	173967
173666	173968	/*
173667	173969	** Disable lookaside memory allocation for objects that might be
173668	173970	** shared across database connections.
173669	173971	*/
		@@ -177553,11 +177855,15 @@
177553	177855	}
177554	177856	break;
177555	177857	case 84: /* cmd ::= select */
177556	177858	{
177557	177859	SelectDest dest = {SRT_Output, 0, 0, 0, 0, 0, 0};
177558		- sqlite3Select(pParse, yymsp[0].minor.yy555, &dest);
	177860	+ if( (pParse->db->mDbFlags & DBFLAG_EncodingFixed)!=0
	177861	+ \|\| sqlite3ReadSchema(pParse)==SQLITE_OK
	177862	+ ){
	177863	+ sqlite3Select(pParse, yymsp[0].minor.yy555, &dest);
	177864	+ }
177559	177865	sqlite3SelectDelete(pParse->db, yymsp[0].minor.yy555);
177560	177866	}
177561	177867	break;
177562	177868	case 85: /* select ::= WITH wqlist selectnowith */
177563	177869	{yymsp[-2].minor.yy555 = attachWithToSelect(pParse,yymsp[0].minor.yy555,yymsp[-1].minor.yy59);}
		@@ -178024,11 +178330,11 @@
178024	178330	** that look like this: #1 #2 ... These terms refer to registers
178025	178331	** in the virtual machine. #N is the N-th register. */
178026	178332	Token t = yymsp[0].minor.yy0; /A-overwrites-X/
178027	178333	assert( t.n>=2 );
178028	178334	if( pParse->nested==0 ){
178029		- sqlite3ErrorMsg(pParse, "near \"%T\": syntax error", &t);
	178335	+ parserSyntaxError(pParse, &t);
178030	178336	yymsp[0].minor.yy454 = 0;
178031	178337	}else{
178032	178338	yymsp[0].minor.yy454 = sqlite3PExpr(pParse, TK_REGISTER, 0, 0);
178033	178339	if( yymsp[0].minor.yy454 ) sqlite3GetInt32(&t.z[1], &yymsp[0].minor.yy454->iTable);
178034	178340	}
		@@ -178872,11 +179178,11 @@
178872	179178	#define TOKEN yyminor
178873	179179	/********** Begin %syntax_error code **************************************/
178874	179180
178875	179181	UNUSED_PARAMETER(yymajor); /* Silence some compiler warnings */
178876	179182	if( TOKEN.z[0] ){
178877		- sqlite3ErrorMsg(pParse, "near \"%T\": syntax error", &TOKEN);
	179183	+ parserSyntaxError(pParse, &TOKEN);
178878	179184	}else{
178879	179185	sqlite3ErrorMsg(pParse, "incomplete input");
178880	179186	}
178881	179187	/********** End %syntax_error code ****************************************/
178882	179188	sqlite3ParserARG_STORE /* Suppress warning about unused %extra_argument variable */
		@@ -180363,11 +180669,13 @@
180363	180669	}
180364	180670	if( pParse->zErrMsg \|\| (pParse->rc!=SQLITE_OK && pParse->rc!=SQLITE_DONE) ){
180365	180671	if( pParse->zErrMsg==0 ){
180366	180672	pParse->zErrMsg = sqlite3MPrintf(db, "%s", sqlite3ErrStr(pParse->rc));
180367	180673	}
180368		- sqlite3_log(pParse->rc, "%s in \"%s\"", pParse->zErrMsg, pParse->zTail);
	180674	+ if( (pParse->prepFlags & SQLITE_PREPARE_DONT_LOG)==0 ){
	180675	+ sqlite3_log(pParse->rc, "%s in \"%s\"", pParse->zErrMsg, pParse->zTail);
	180676	+ }
180369	180677	nErr++;
180370	180678	}
180371	180679	pParse->zTail = zSql;
180372	180680	#ifndef SQLITE_OMIT_VIRTUALTABLE
180373	180681	sqlite3_free(pParse->apVtabLock);
		@@ -181071,36 +181379,10 @@
181071	181379	** all database files specified with a relative pathname.
181072	181380	**
181073	181381	** See also the "PRAGMA data_store_directory" SQL command.
181074	181382	*/
181075	181383	SQLITE_API char *sqlite3_data_directory = 0;
181076		-
181077		-/*
181078		-** Determine whether or not high-precision (long double) floating point
181079		-** math works correctly on CPU currently running.
181080		-*/
181081		-static SQLITE_NOINLINE int hasHighPrecisionDouble(int rc){
181082		- if( sizeof(LONGDOUBLE_TYPE)<=8 ){
181083		- /* If the size of "long double" is not more than 8, then
181084		- ** high-precision math is not possible. */
181085		- return 0;
181086		- }else{
181087		- /* Just because sizeof(long double)>8 does not mean that the underlying
181088		- ** hardware actually supports high-precision floating point. For example,
181089		- ** clearing the 0x100 bit in the floating-point control word on Intel
181090		- ** processors will make long double work like double, even though long
181091		- ** double takes up more space. The only way to determine if long double
181092		- ** actually works is to run an experiment. */
181093		- LONGDOUBLE_TYPE a, b, c;
181094		- rc++;
181095		- a = 1.0+rc*0.1;
181096		- b = 1.0e+18+rc*25.0;
181097		- c = a+b;
181098		- return b!=c;
181099		- }
181100		-}
181101		-
181102	181384
181103	181385	/*
181104	181386	** Initialize SQLite.
181105	181387	**
181106	181388	** This routine must be called to initialize the memory allocation,
		@@ -181292,17 +181574,10 @@
181292	181574	if( bRunExtraInit ){
181293	181575	int SQLITE_EXTRA_INIT(const char*);
181294	181576	rc = SQLITE_EXTRA_INIT(0);
181295	181577	}
181296	181578	#endif
181297		-
181298		- /* Experimentally determine if high-precision floating point is
181299		- ** available. */
181300		-#ifndef SQLITE_OMIT_WSD
181301		- sqlite3Config.bUseLongDouble = hasHighPrecisionDouble(rc);
181302		-#endif
181303		-
181304	181579	return rc;
181305	181580	}
181306	181581
181307	181582	/*
181308	181583	** Undo the effects of sqlite3_initialize(). Must not be called while
		@@ -182369,14 +182644,10 @@
182369	182644	#endif
182370	182645
182371	182646	sqlite3Error(db, SQLITE_OK); /* Deallocates any cached error strings. */
182372	182647	sqlite3ValueFree(db->pErr);
182373	182648	sqlite3CloseExtensions(db);
182374		-#if SQLITE_USER_AUTHENTICATION
182375		- sqlite3_free(db->auth.zAuthUser);
182376		- sqlite3_free(db->auth.zAuthPW);
182377		-#endif
182378	182649
182379	182650	db->eOpenState = SQLITE_STATE_ERROR;
182380	182651
182381	182652	/* The temp-database schema is allocated differently from the other schema
182382	182653	** objects (using sqliteMalloc() directly, instead of sqlite3BtreeSchema()).
		@@ -183875,12 +184146,12 @@
183875	184146	}
183876	184147	oldLimit = db->aLimit[limitId];
183877	184148	if( newLimit>=0 ){ /* IMP: R-52476-28732 */
183878	184149	if( newLimit>aHardLimit[limitId] ){
183879	184150	newLimit = aHardLimit[limitId]; /* IMP: R-51463-25634 */
183880		- }else if( newLimit<1 && limitId==SQLITE_LIMIT_LENGTH ){
183881		- newLimit = 1;
	184151	+ }else if( newLimit<SQLITE_MIN_LENGTH && limitId==SQLITE_LIMIT_LENGTH ){
	184152	+ newLimit = SQLITE_MIN_LENGTH;
183882	184153	}
183883	184154	db->aLimit[limitId] = newLimit;
183884	184155	}
183885	184156	return oldLimit; /* IMP: R-53341-35419 */
183886	184157	}
		@@ -184395,10 +184666,11 @@
184395	184666	testcase( (1<<(flags&7))==0x04 ); /* READWRITE */
184396	184667	testcase( (1<<(flags&7))==0x40 ); /* READWRITE \| CREATE */
184397	184668	if( ((1<<(flags&7)) & 0x46)==0 ){
184398	184669	rc = SQLITE_MISUSE_BKPT; /* IMP: R-18321-05872 */
184399	184670	}else{
	184671	+ if( zFilename==0 ) zFilename = ":memory:";
184400	184672	rc = sqlite3ParseUri(zVfs, zFilename, &flags, &db->pVfs, &zOpen, &zErrMsg);
184401	184673	}
184402	184674	if( rc!=SQLITE_OK ){
184403	184675	if( rc==SQLITE_NOMEM ) sqlite3OomFault(db);
184404	184676	sqlite3ErrorWithMsg(db, rc, zErrMsg ? "%s" : 0, zErrMsg);
		@@ -185237,10 +185509,11 @@
185237	185509	#ifndef SQLITE_OMIT_WINDOWFUNC
185238	185510	sqlite3ShowWindow(0);
185239	185511	sqlite3ShowWinFunc(0);
185240	185512	#endif
185241	185513	sqlite3ShowSelect(0);
	185514	+ sqlite3ShowWhereTerm(0);
185242	185515	}
185243	185516	#endif
185244	185517	break;
185245	185518	}
185246	185519
		@@ -185549,28 +185822,10 @@
185549	185822	*pI1 = rLogEst;
185550	185823	*pU64 = sqlite3LogEstToInt(rLogEst);
185551	185824	pI2 = sqlite3LogEst(pU64);
185552	185825	break;
185553	185826	}
185554		-
185555		-#if !defined(SQLITE_OMIT_WSD)
185556		- /* sqlite3_test_control(SQLITE_TESTCTRL_USELONGDOUBLE, int X);
185557		- **
185558		- ** X<0 Make no changes to the bUseLongDouble. Just report value.
185559		- ** X==0 Disable bUseLongDouble
185560		- ** X==1 Enable bUseLongDouble
185561		- ** X>=2 Set bUseLongDouble to its default value for this platform
185562		- */
185563		- case SQLITE_TESTCTRL_USELONGDOUBLE: {
185564		- int b = va_arg(ap, int);
185565		- if( b>=2 ) b = hasHighPrecisionDouble(b);
185566		- if( b>=0 ) sqlite3Config.bUseLongDouble = b>0;
185567		- rc = sqlite3Config.bUseLongDouble!=0;
185568		- break;
185569		- }
185570		-#endif
185571		-
185572	185827
185573	185828	#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_WSD)
185574	185829	/* sqlite3_test_control(SQLITE_TESTCTRL_TUNE, id, *piValue)
185575	185830	**
185576	185831	** If "id" is an integer between 1 and SQLITE_NTUNE then set the value
		@@ -185875,11 +186130,15 @@
185875	186130	if( db->autoCommit==0 ){
185876	186131	int iDb = sqlite3FindDbName(db, zDb);
185877	186132	if( iDb==0 \|\| iDb>1 ){
185878	186133	Btree *pBt = db->aDb[iDb].pBt;
185879	186134	if( SQLITE_TXN_WRITE!=sqlite3BtreeTxnState(pBt) ){
	186135	+ Pager *pPager = sqlite3BtreePager(pBt);
	186136	+ i64 dummy = 0;
	186137	+ sqlite3PagerSnapshotOpen(pPager, (sqlite3_snapshot*)&dummy);
185880	186138	rc = sqlite3BtreeBeginTrans(pBt, 0, 0);
	186139	+ sqlite3PagerSnapshotOpen(pPager, 0);
185881	186140	if( rc==SQLITE_OK ){
185882	186141	rc = sqlite3PagerSnapshotGet(sqlite3BtreePager(pBt), ppSnapshot);
185883	186142	}
185884	186143	}
185885	186144	}
		@@ -189665,14 +189924,19 @@
189665	189924
189666	189925	assert_fts3_nc( p!=0 && p1!=0 && p2!=0 );
189667	189926	if( *p1==POS_COLUMN ){
189668	189927	p1++;
189669	189928	p1 += fts3GetVarint32(p1, &iCol1);
	189929	+ /* iCol1==0 indicates corruption. Column 0 does not have a POS_COLUMN
	189930	+ ** entry, so this is actually end-of-doclist. */
	189931	+ if( iCol1==0 ) return 0;
189670	189932	}
189671	189933	if( *p2==POS_COLUMN ){
189672	189934	p2++;
189673	189935	p2 += fts3GetVarint32(p2, &iCol2);
	189936	+ /* As above, iCol2==0 indicates corruption. */
	189937	+ if( iCol2==0 ) return 0;
189674	189938	}
189675	189939
189676	189940	while( 1 ){
189677	189941	if( iCol1==iCol2 ){
189678	189942	char *pSave = p;
		@@ -192839,11 +193103,11 @@
192839	193103	for(p=pExpr; p->pLeft; p=p->pLeft){
192840	193104	assert( p->pRight->pPhrase->doclist.nList>0 );
192841	193105	nTmp += p->pRight->pPhrase->doclist.nList;
192842	193106	}
192843	193107	nTmp += p->pPhrase->doclist.nList;
192844		- aTmp = sqlite3_malloc64(nTmp*2);
	193108	+ aTmp = sqlite3_malloc64(nTmp*2 + FTS3_VARINT_MAX);
192845	193109	if( !aTmp ){
192846	193110	*pRc = SQLITE_NOMEM;
192847	193111	res = 0;
192848	193112	}else{
192849	193113	char *aPoslist = p->pPhrase->doclist.pList;
		@@ -193490,11 +193754,11 @@
193490	193754	SQLITE_PRIVATE int sqlite3Fts3Corrupt(){
193491	193755	return SQLITE_CORRUPT_VTAB;
193492	193756	}
193493	193757	#endif
193494	193758
193495		-#if !SQLITE_CORE
	193759	+#if !defined(SQLITE_CORE)
193496	193760	/*
193497	193761	** Initialize API pointer table, if required.
193498	193762	*/
193499	193763	#ifdef _WIN32
193500	193764	__declspec(dllexport)
		@@ -194392,14 +194656,15 @@
194392	194656	rc = pModule->xNext(pCursor, &zByte, &nByte, &iBegin, &iEnd, &iPos);
194393	194657	if( rc==SQLITE_OK ){
194394	194658	Fts3PhraseToken *pToken;
194395	194659
194396	194660	p = fts3ReallocOrFree(p, nSpace + ii*sizeof(Fts3PhraseToken));
194397		- if( !p ) goto no_mem;
194398		-
194399	194661	zTemp = fts3ReallocOrFree(zTemp, nTemp + nByte);
194400		- if( !zTemp ) goto no_mem;
	194662	+ if( !zTemp \|\| !p ){
	194663	+ rc = SQLITE_NOMEM;
	194664	+ goto getnextstring_out;
	194665	+ }
194401	194666
194402	194667	assert( nToken==ii );
194403	194668	pToken = &((Fts3Phrase *)(&p[1]))->aToken[ii];
194404	194669	memset(pToken, 0, sizeof(Fts3PhraseToken));
194405	194670
		@@ -194410,53 +194675,51 @@
194410	194675	pToken->isPrefix = (iEnd<nInput && zInput[iEnd]=='*');
194411	194676	pToken->bFirst = (iBegin>0 && zInput[iBegin-1]=='^');
194412	194677	nToken = ii+1;
194413	194678	}
194414	194679	}
194415		-
194416		- pModule->xClose(pCursor);
194417		- pCursor = 0;
194418	194680	}
194419	194681
194420	194682	if( rc==SQLITE_DONE ){
194421	194683	int jj;
194422	194684	char *zBuf = 0;
194423	194685
194424	194686	p = fts3ReallocOrFree(p, nSpace + nToken*sizeof(Fts3PhraseToken) + nTemp);
194425		- if( !p ) goto no_mem;
	194687	+ if( !p ){
	194688	+ rc = SQLITE_NOMEM;
	194689	+ goto getnextstring_out;
	194690	+ }
194426	194691	memset(p, 0, (char )&(((Fts3Phrase )&p[1])->aToken[0])-(char *)p);
194427	194692	p->eType = FTSQUERY_PHRASE;
194428	194693	p->pPhrase = (Fts3Phrase *)&p[1];
194429	194694	p->pPhrase->iColumn = pParse->iDefaultCol;
194430	194695	p->pPhrase->nToken = nToken;
194431	194696
194432	194697	zBuf = (char *)&p->pPhrase->aToken[nToken];
	194698	+ assert( nTemp==0 \|\| zTemp );
194433	194699	if( zTemp ){
194434	194700	memcpy(zBuf, zTemp, nTemp);
194435		- sqlite3_free(zTemp);
194436		- }else{
194437		- assert( nTemp==0 );
194438	194701	}
194439	194702
194440	194703	for(jj=0; jj<p->pPhrase->nToken; jj++){
194441	194704	p->pPhrase->aToken[jj].z = zBuf;
194442	194705	zBuf += p->pPhrase->aToken[jj].n;
194443	194706	}
194444	194707	rc = SQLITE_OK;
194445	194708	}
194446	194709
194447		- *ppExpr = p;
194448		- return rc;
194449		-no_mem:
194450		-
	194710	+ getnextstring_out:
194451	194711	if( pCursor ){
194452	194712	pModule->xClose(pCursor);
194453	194713	}
194454	194714	sqlite3_free(zTemp);
194455		- sqlite3_free(p);
194456		- *ppExpr = 0;
194457		- return SQLITE_NOMEM;
	194715	+ if( rc!=SQLITE_OK ){
	194716	+ sqlite3_free(p);
	194717	+ p = 0;
	194718	+ }
	194719	+ *ppExpr = p;
	194720	+ return rc;
194458	194721	}
194459	194722
194460	194723	/*
194461	194724	** The output variable *ppExpr is populated with an allocated Fts3Expr
194462	194725	** structure, or set to 0 if the end of the input buffer is reached.
		@@ -208867,11 +209130,13 @@
208867	209130	int rawKey = 1;
208868	209131	x = pParse->aBlob[iRoot];
208869	209132	zPath++;
208870	209133	if( zPath[0]=='"' ){
208871	209134	zKey = zPath + 1;
208872		- for(i=1; zPath[i] && zPath[i]!='"'; i++){}
	209135	+ for(i=1; zPath[i] && zPath[i]!='"'; i++){
	209136	+ if( zPath[i]=='\\' && zPath[i+1]!=0 ) i++;
	209137	+ }
208873	209138	nKey = i-1;
208874	209139	if( zPath[i] ){
208875	209140	i++;
208876	209141	}else{
208877	209142	return JSON_LOOKUP_PATHERROR;
		@@ -217779,11 +218044,11 @@
217779	218044	return sqlite3_create_function_v2(db, zQueryFunc, -1, SQLITE_ANY,
217780	218045	(void *)pGeomCtx, geomCallback, 0, 0, rtreeFreeCallback
217781	218046	);
217782	218047	}
217783	218048
217784		-#if !SQLITE_CORE
	218049	+#ifndef SQLITE_CORE
217785	218050	#ifdef _WIN32
217786	218051	__declspec(dllexport)
217787	218052	#endif
217788	218053	SQLITE_API int sqlite3_rtree_init(
217789	218054	sqlite3 *db,
		@@ -218370,11 +218635,11 @@
218370	218635	}
218371	218636
218372	218637	return rc;
218373	218638	}
218374	218639
218375		-#if !SQLITE_CORE
	218640	+#ifndef SQLITE_CORE
218376	218641	#ifdef _WIN32
218377	218642	__declspec(dllexport)
218378	218643	#endif
218379	218644	SQLITE_API int sqlite3_icu_init(
218380	218645	sqlite3 *db,
		@@ -219628,10 +219893,31 @@
219628	219893	struct RbuFrame {
219629	219894	u32 iDbPage;
219630	219895	u32 iWalFrame;
219631	219896	};
219632	219897
	219898	+#ifndef UNUSED_PARAMETER
	219899	+/*
	219900	+** The following macros are used to suppress compiler warnings and to
	219901	+** make it clear to human readers when a function parameter is deliberately
	219902	+** left unused within the body of a function. This usually happens when
	219903	+** a function is called via a function pointer. For example the
	219904	+** implementation of an SQL aggregate step callback may not use the
	219905	+** parameter indicating the number of arguments passed to the aggregate,
	219906	+** if it knows that this is enforced elsewhere.
	219907	+**
	219908	+** When a function parameter is not used at all within the body of a function,
	219909	+** it is generally named "NotUsed" or "NotUsed2" to make things even clearer.
	219910	+** However, these macros may also be used to suppress warnings related to
	219911	+** parameters that may or may not be used depending on compilation options.
	219912	+** For example those parameters only used in assert() statements. In these
	219913	+** cases the parameters are named as per the usual conventions.
	219914	+*/
	219915	+#define UNUSED_PARAMETER(x) (void)(x)
	219916	+#define UNUSED_PARAMETER2(x,y) UNUSED_PARAMETER(x),UNUSED_PARAMETER(y)
	219917	+#endif
	219918	+
219633	219919	/*
219634	219920	** RBU handle.
219635	219921	**
219636	219922	** nPhaseOneStep:
219637	219923	** If the RBU database contains an rbu_count table, this value is set to
		@@ -219679,11 +219965,11 @@
219679	219965	char zState; / Path to state db (or NULL if zRbu) */
219680	219966	char zStateDb[5]; /* Db name for state ("stat" or "main") */
219681	219967	int rc; /* Value returned by last rbu_step() call */
219682	219968	char zErrmsg; / Error message if rc!=SQLITE_OK */
219683	219969	int nStep; /* Rows processed for current object */
219684		- int nProgress; /* Rows processed for all objects */
	219970	+ sqlite3_int64 nProgress; /* Rows processed for all objects */
219685	219971	RbuObjIter objiter; /* Iterator for skipping through tbl/idx */
219686	219972	const char zVfsName; / Name of automatically created rbu vfs */
219687	219973	rbu_file pTargetFd; / File handle open on target db */
219688	219974	int nPagePerSector; /* Pages per sector for pTargetFd */
219689	219975	i64 iOalSz;
		@@ -219796,11 +220082,11 @@
219796	220082	unsigned char *zStart = z;
219797	220083	while( (c = zValue[0x7f&*(z++)])>=0 ){
219798	220084	v = (v<<6) + c;
219799	220085	}
219800	220086	z--;
219801		- *pLen -= z - zStart;
	220087	+ *pLen -= (int)(z - zStart);
219802	220088	pz = (char)z;
219803	220089	return v;
219804	220090	}
219805	220091
219806	220092	#if RBU_ENABLE_DELTA_CKSUM
		@@ -219981,10 +220267,11 @@
219981	220267	int nOut;
219982	220268	int nOut2;
219983	220269	char *aOut;
219984	220270
219985	220271	assert( argc==2 );
	220272	+ UNUSED_PARAMETER(argc);
219986	220273
219987	220274	nOrig = sqlite3_value_bytes(argv[0]);
219988	220275	aOrig = (const char*)sqlite3_value_blob(argv[0]);
219989	220276	nDelta = sqlite3_value_bytes(argv[1]);
219990	220277	aDelta = (const char*)sqlite3_value_blob(argv[1]);
		@@ -221560,17 +221847,17 @@
221560	221847	nParen++;
221561	221848	}
221562	221849	else if( c==')' ){
221563	221850	nParen--;
221564	221851	if( nParen==0 ){
221565		- int nSpan = &zSql[i] - pIter->aIdxCol[iIdxCol].zSpan;
	221852	+ int nSpan = (int)(&zSql[i] - pIter->aIdxCol[iIdxCol].zSpan);
221566	221853	pIter->aIdxCol[iIdxCol++].nSpan = nSpan;
221567	221854	i++;
221568	221855	break;
221569	221856	}
221570	221857	}else if( c==',' && nParen==1 ){
221571		- int nSpan = &zSql[i] - pIter->aIdxCol[iIdxCol].zSpan;
	221858	+ int nSpan = (int)(&zSql[i] - pIter->aIdxCol[iIdxCol].zSpan);
221572	221859	pIter->aIdxCol[iIdxCol++].nSpan = nSpan;
221573	221860	pIter->aIdxCol[iIdxCol].zSpan = &zSql[i+1];
221574	221861	}else if( c=='"' \|\| c=='\'' \|\| c=='`' ){
221575	221862	for(i++; 1; i++){
221576	221863	if( zSql[i]==c ){
		@@ -222256,10 +222543,12 @@
222256	222543	int i, sz;
222257	222544	sz = (int)strlen(z)&0xffffff;
222258	222545	for(i=sz-1; i>0 && z[i]!='/' && z[i]!='.'; i--){}
222259	222546	if( z[i]=='.' && sz>i+4 ) memmove(&z[i+1], &z[sz-3], 4);
222260	222547	}
	222548	+#else
	222549	+ UNUSED_PARAMETER2(zBase,z);
222261	222550	#endif
222262	222551	}
222263	222552
222264	222553	/*
222265	222554	** Return the current wal-index header checksum for the target database
		@@ -222840,11 +223129,11 @@
222840	223129	"INSERT OR REPLACE INTO %s.rbu_state(k, v) VALUES "
222841	223130	"(%d, %d), "
222842	223131	"(%d, %Q), "
222843	223132	"(%d, %Q), "
222844	223133	"(%d, %d), "
222845		- "(%d, %d), "
	223134	+ "(%d, %lld), "
222846	223135	"(%d, %lld), "
222847	223136	"(%d, %lld), "
222848	223137	"(%d, %lld), "
222849	223138	"(%d, %lld), "
222850	223139	"(%d, %Q) ",
		@@ -223198,10 +223487,11 @@
223198	223487	char *zErrmsg = 0;
223199	223488	int rc;
223200	223489	sqlite3 *db = (rbuIsVacuum(p) ? p->dbRbu : p->dbMain);
223201	223490
223202	223491	assert( nVal==1 );
	223492	+ UNUSED_PARAMETER(nVal);
223203	223493
223204	223494	rc = prepareFreeAndCollectError(db, &pStmt, &zErrmsg,
223205	223495	sqlite3_mprintf("SELECT count(*) FROM sqlite_schema "
223206	223496	"WHERE type='index' AND tbl_name = %Q", sqlite3_value_text(apVal[0]))
223207	223497	);
		@@ -223473,11 +223763,11 @@
223473	223763	const char *zTarget,
223474	223764	const char *zState
223475	223765	){
223476	223766	if( zTarget==0 ){ return rbuMisuseError(); }
223477	223767	if( zState ){
223478		- int n = strlen(zState);
	223768	+ size_t n = strlen(zState);
223479	223769	if( n>=7 && 0==memcmp("-vactmp", &zState[n-7], 7) ){
223480	223770	return rbuMisuseError();
223481	223771	}
223482	223772	}
223483	223773	/* TODO: Check that both arguments are non-NULL */
		@@ -223690,10 +223980,11 @@
223690	223980	/*
223691	223981	** Default xRename callback for RBU.
223692	223982	*/
223693	223983	static int xDefaultRename(void pArg, const char zOld, const char *zNew){
223694	223984	int rc = SQLITE_OK;
	223985	+ UNUSED_PARAMETER(pArg);
223695	223986	#if defined(_WIN32_WCE)
223696	223987	{
223697	223988	LPWSTR zWideOld;
223698	223989	LPWSTR zWideNew;
223699	223990
		@@ -224594,10 +224885,13 @@
224594	224885
224595	224886	/*
224596	224887	** No-op.
224597	224888	*/
224598	224889	static int rbuVfsGetLastError(sqlite3_vfs pVfs, int a, char b){
	224890	+ UNUSED_PARAMETER(pVfs);
	224891	+ UNUSED_PARAMETER(a);
	224892	+ UNUSED_PARAMETER(b);
224599	224893	return 0;
224600	224894	}
224601	224895
224602	224896	/*
224603	224897	** Deregister and destroy an RBU vfs created by an earlier call to
		@@ -225650,11 +225944,17 @@
225650	225944	** schema for the database file that is to be read. The default schema is
225651	225945	** "main".
225652	225946	**
225653	225947	** The data field of sqlite_dbpage table can be updated. The new
225654	225948	** value must be a BLOB which is the correct page size, otherwise the
225655		-** update fails. Rows may not be deleted or inserted.
	225949	+** update fails. INSERT operations also work, and operate as if they
	225950	+** where REPLACE. The size of the database can be extended by INSERT-ing
	225951	+** new pages on the end.
	225952	+**
	225953	+** Rows may not be deleted. However, doing an INSERT to page number N
	225954	+** with NULL page data causes the N-th page and all subsequent pages to be
	225955	+** deleted and the database to be truncated.
225656	225956	*/
225657	225957
225658	225958	/* #include "sqliteInt.h" Requires access to internal data structures */
225659	225959	#if (defined(SQLITE_ENABLE_DBPAGE_VTAB) \|\| defined(SQLITE_TEST)) \
225660	225960	&& !defined(SQLITE_OMIT_VIRTUALTABLE)
		@@ -225673,17 +225973,18 @@
225673	225973	};
225674	225974
225675	225975	struct DbpageTable {
225676	225976	sqlite3_vtab base; /* Base class. Must be first */
225677	225977	sqlite3 db; / The database */
	225978	+ int iDbTrunc; /* Database to truncate */
	225979	+ Pgno pgnoTrunc; /* Size to truncate to */
225678	225980	};
225679	225981
225680	225982	/* Columns */
225681	225983	#define DBPAGE_COLUMN_PGNO 0
225682	225984	#define DBPAGE_COLUMN_DATA 1
225683	225985	#define DBPAGE_COLUMN_SCHEMA 2
225684		-
225685	225986
225686	225987
225687	225988	/*
225688	225989	** Connect to or create a dbpagevfs virtual table.
225689	225990	*/
		@@ -225943,15 +226244,15 @@
225943	226244	DbpageTable pTab = (DbpageTable )pVtab;
225944	226245	Pgno pgno;
225945	226246	DbPage *pDbPage = 0;
225946	226247	int rc = SQLITE_OK;
225947	226248	char *zErr = 0;
225948		- const char *zSchema;
225949	226249	int iDb;
225950	226250	Btree *pBt;
225951	226251	Pager *pPager;
225952	226252	int szPage;
	226253	+ int isInsert;
225953	226254
225954	226255	(void)pRowid;
225955	226256	if( pTab->db->flags & SQLITE_Defensive ){
225956	226257	zErr = "read-only";
225957	226258	goto update_fail;
		@@ -225958,45 +226259,62 @@
225958	226259	}
225959	226260	if( argc==1 ){
225960	226261	zErr = "cannot delete";
225961	226262	goto update_fail;
225962	226263	}
225963		- pgno = sqlite3_value_int(argv[0]);
225964		- if( sqlite3_value_type(argv[0])==SQLITE_NULL
225965		- \|\| (Pgno)sqlite3_value_int(argv[1])!=pgno
225966		- ){
225967		- zErr = "cannot insert";
225968		- goto update_fail;
225969		- }
225970		- zSchema = (const char*)sqlite3_value_text(argv[4]);
225971		- iDb = ALWAYS(zSchema) ? sqlite3FindDbName(pTab->db, zSchema) : -1;
225972		- if( NEVER(iDb<0) ){
225973		- zErr = "no such schema";
225974		- goto update_fail;
	226264	+ if( sqlite3_value_type(argv[0])==SQLITE_NULL ){
	226265	+ pgno = (Pgno)sqlite3_value_int(argv[2]);
	226266	+ isInsert = 1;
	226267	+ }else{
	226268	+ pgno = sqlite3_value_int(argv[0]);
	226269	+ if( (Pgno)sqlite3_value_int(argv[1])!=pgno ){
	226270	+ zErr = "cannot insert";
	226271	+ goto update_fail;
	226272	+ }
	226273	+ isInsert = 0;
	226274	+ }
	226275	+ if( sqlite3_value_type(argv[4])==SQLITE_NULL ){
	226276	+ iDb = 0;
	226277	+ }else{
	226278	+ const char zSchema = (const char)sqlite3_value_text(argv[4]);
	226279	+ iDb = sqlite3FindDbName(pTab->db, zSchema);
	226280	+ if( iDb<0 ){
	226281	+ zErr = "no such schema";
	226282	+ goto update_fail;
	226283	+ }
225975	226284	}
225976	226285	pBt = pTab->db->aDb[iDb].pBt;
225977		- if( NEVER(pgno<1) \|\| NEVER(pBt==0) \|\| NEVER(pgno>sqlite3BtreeLastPage(pBt)) ){
	226286	+ if( pgno<1 \|\| NEVER(pBt==0) ){
225978	226287	zErr = "bad page number";
225979	226288	goto update_fail;
225980	226289	}
225981	226290	szPage = sqlite3BtreeGetPageSize(pBt);
225982	226291	if( sqlite3_value_type(argv[3])!=SQLITE_BLOB
225983	226292	\|\| sqlite3_value_bytes(argv[3])!=szPage
225984	226293	){
225985		- zErr = "bad page value";
225986		- goto update_fail;
	226294	+ if( sqlite3_value_type(argv[3])==SQLITE_NULL && isInsert && pgno>1 ){
	226295	+ /* "INSERT INTO dbpage($PGNO,NULL)" causes page number $PGNO and
	226296	+ ** all subsequent pages to be deleted. */
	226297	+ pTab->iDbTrunc = iDb;
	226298	+ pgno--;
	226299	+ pTab->pgnoTrunc = pgno;
	226300	+ }else{
	226301	+ zErr = "bad page value";
	226302	+ goto update_fail;
	226303	+ }
225987	226304	}
225988	226305	pPager = sqlite3BtreePager(pBt);
225989	226306	rc = sqlite3PagerGet(pPager, pgno, (DbPage**)&pDbPage, 0);
225990	226307	if( rc==SQLITE_OK ){
225991	226308	const void *pData = sqlite3_value_blob(argv[3]);
225992		- assert( pData!=0 \|\| pTab->db->mallocFailed );
225993		- if( pData
225994		- && (rc = sqlite3PagerWrite(pDbPage))==SQLITE_OK
225995		- ){
225996		- memcpy(sqlite3PagerGetData(pDbPage), pData, szPage);
	226309	+ if( (rc = sqlite3PagerWrite(pDbPage))==SQLITE_OK && pData ){
	226310	+ unsigned char *aPage = sqlite3PagerGetData(pDbPage);
	226311	+ memcpy(aPage, pData, szPage);
	226312	+ pTab->pgnoTrunc = 0;
225997	226313	}
	226314	+ }else{
	226315	+ pTab->pgnoTrunc = 0;
225998	226316	}
225999	226317	sqlite3PagerUnref(pDbPage);
226000	226318	return rc;
226001	226319
226002	226320	update_fail:
		@@ -226015,13 +226333,35 @@
226015	226333	int i;
226016	226334	for(i=0; i<db->nDb; i++){
226017	226335	Btree *pBt = db->aDb[i].pBt;
226018	226336	if( pBt ) (void)sqlite3BtreeBeginTrans(pBt, 1, 0);
226019	226337	}
	226338	+ pTab->pgnoTrunc = 0;
	226339	+ return SQLITE_OK;
	226340	+}
	226341	+
	226342	+/* Invoke sqlite3PagerTruncate() as necessary, just prior to COMMIT
	226343	+*/
	226344	+static int dbpageSync(sqlite3_vtab *pVtab){
	226345	+ DbpageTable pTab = (DbpageTable )pVtab;
	226346	+ if( pTab->pgnoTrunc>0 ){
	226347	+ Btree *pBt = pTab->db->aDb[pTab->iDbTrunc].pBt;
	226348	+ Pager *pPager = sqlite3BtreePager(pBt);
	226349	+ sqlite3PagerTruncateImage(pPager, pTab->pgnoTrunc);
	226350	+ }
	226351	+ pTab->pgnoTrunc = 0;
	226352	+ return SQLITE_OK;
	226353	+}
	226354	+
	226355	+/* Cancel any pending truncate.
	226356	+*/
	226357	+static int dbpageRollbackTo(sqlite3_vtab *pVtab, int notUsed1){
	226358	+ DbpageTable pTab = (DbpageTable )pVtab;
	226359	+ pTab->pgnoTrunc = 0;
	226360	+ (void)notUsed1;
226020	226361	return SQLITE_OK;
226021	226362	}
226022		-
226023	226363
226024	226364	/*
226025	226365	** Invoke this routine to register the "dbpage" virtual table module
226026	226366	*/
226027	226367	SQLITE_PRIVATE int sqlite3DbpageRegister(sqlite3 *db){
		@@ -226039,18 +226379,18 @@
226039	226379	dbpageEof, /* xEof - check for end of scan */
226040	226380	dbpageColumn, /* xColumn - read data */
226041	226381	dbpageRowid, /* xRowid - read data */
226042	226382	dbpageUpdate, /* xUpdate */
226043	226383	dbpageBegin, /* xBegin */
226044		- 0, /* xSync */
	226384	+ dbpageSync, /* xSync */
226045	226385	0, /* xCommit */
226046	226386	0, /* xRollback */
226047	226387	0, /* xFindMethod */
226048	226388	0, /* xRename */
226049	226389	0, /* xSavepoint */
226050	226390	0, /* xRelease */
226051		- 0, /* xRollbackTo */
	226391	+ dbpageRollbackTo, /* xRollbackTo */
226052	226392	0, /* xShadowName */
226053	226393	0 /* xIntegrity */
226054	226394	};
226055	226395	return sqlite3_create_module(db, "sqlite_dbpage", &dbpage_module, 0);
226056	226396	}
		@@ -226134,10 +226474,14 @@
226134	226474	/*
226135	226475	** An object of this type is used internally as an abstraction for
226136	226476	** input data. Input data may be supplied either as a single large buffer
226137	226477	** (e.g. sqlite3changeset_start()) or using a stream function (e.g.
226138	226478	** sqlite3changeset_start_strm()).
	226479	+**
	226480	+** bNoDiscard:
	226481	+** If true, then the only time data is discarded is as a result of explicit
	226482	+** sessionDiscardData() calls. Not within every sessionInputBuffer() call.
226139	226483	*/
226140	226484	struct SessionInput {
226141	226485	int bNoDiscard; /* If true, do not discard in InputBuffer() */
226142	226486	int iCurrent; /* Offset in aData[] of current change */
226143	226487	int iNext; /* Offset in aData[] of next change */
		@@ -227817,20 +228161,23 @@
227817	228161	/* Figure out how large an allocation is required */
227818	228162	nByte = sizeof(SessionChange);
227819	228163	for(i=0; i<(pTab->nCol-pTab->bRowid); i++){
227820	228164	sqlite3_value *p = 0;
227821	228165	if( op!=SQLITE_INSERT ){
227822		- TESTONLY(int trc = ) pSession->hook.xOld(pSession->hook.pCtx, i, &p);
227823		- assert( trc==SQLITE_OK );
	228166	+ /* This may fail if the column has a non-NULL default and was added
	228167	+ ** using ALTER TABLE ADD COLUMN after this record was created. */
	228168	+ rc = pSession->hook.xOld(pSession->hook.pCtx, i, &p);
227824	228169	}else if( pTab->abPK[i] ){
227825	228170	TESTONLY(int trc = ) pSession->hook.xNew(pSession->hook.pCtx, i, &p);
227826	228171	assert( trc==SQLITE_OK );
227827	228172	}
227828	228173
227829		- /* This may fail if SQLite value p contains a utf-16 string that must
227830		- ** be converted to utf-8 and an OOM error occurs while doing so. */
227831		- rc = sessionSerializeValue(0, p, &nByte);
	228174	+ if( rc==SQLITE_OK ){
	228175	+ /* This may fail if SQLite value p contains a utf-16 string that must
	228176	+ ** be converted to utf-8 and an OOM error occurs while doing so. */
	228177	+ rc = sessionSerializeValue(0, p, &nByte);
	228178	+ }
227832	228179	if( rc!=SQLITE_OK ) goto error_out;
227833	228180	}
227834	228181	if( pTab->bRowid ){
227835	228182	nByte += 9; /* Size of rowid field - an integer */
227836	228183	}
		@@ -231184,19 +231531,25 @@
231184	231531	int rc = SQLITE_OK; /* Return code */
231185	231532	const char zTab = 0; / Name of current table */
231186	231533	int nTab = 0; /* Result of sqlite3Strlen30(zTab) */
231187	231534	SessionApplyCtx sApply; /* changeset_apply() context object */
231188	231535	int bPatchset;
	231536	+ u64 savedFlag = db->flags & SQLITE_FkNoAction;
231189	231537
231190	231538	assert( xConflict!=0 );
	231539	+
	231540	+ sqlite3_mutex_enter(sqlite3_db_mutex(db));
	231541	+ if( flags & SQLITE_CHANGESETAPPLY_FKNOACTION ){
	231542	+ db->flags \|= ((u64)SQLITE_FkNoAction);
	231543	+ db->aDb[0].pSchema->schema_cookie -= 32;
	231544	+ }
231191	231545
231192	231546	pIter->in.bNoDiscard = 1;
231193	231547	memset(&sApply, 0, sizeof(sApply));
231194	231548	sApply.bRebase = (ppRebase && pnRebase);
231195	231549	sApply.bInvertConstraints = !!(flags & SQLITE_CHANGESETAPPLY_INVERT);
231196	231550	sApply.bIgnoreNoop = !!(flags & SQLITE_CHANGESETAPPLY_IGNORENOOP);
231197		- sqlite3_mutex_enter(sqlite3_db_mutex(db));
231198	231551	if( (flags & SQLITE_CHANGESETAPPLY_NOSAVEPOINT)==0 ){
231199	231552	rc = sqlite3_exec(db, "SAVEPOINT changeset_apply", 0, 0, 0);
231200	231553	}
231201	231554	if( rc==SQLITE_OK ){
231202	231555	rc = sqlite3_exec(db, "PRAGMA defer_foreign_keys = 1", 0, 0, 0);
		@@ -231354,10 +231707,16 @@
231354	231707	sqlite3_finalize(sApply.pDelete);
231355	231708	sqlite3_finalize(sApply.pSelect);
231356	231709	sqlite3_free((char)sApply.azCol); / cast works around VC++ bug */
231357	231710	sqlite3_free((char*)sApply.constraints.aBuf);
231358	231711	sqlite3_free((char*)sApply.rebase.aBuf);
	231712	+
	231713	+ if( (flags & SQLITE_CHANGESETAPPLY_FKNOACTION) && savedFlag==0 ){
	231714	+ assert( db->flags & SQLITE_FkNoAction );
	231715	+ db->flags &= ~((u64)SQLITE_FkNoAction);
	231716	+ db->aDb[0].pSchema->schema_cookie -= 32;
	231717	+ }
231359	231718	sqlite3_mutex_leave(sqlite3_db_mutex(db));
231360	231719	return rc;
231361	231720	}
231362	231721
231363	231722	/*
		@@ -231382,28 +231741,17 @@
231382	231741	int flags
231383	231742	){
231384	231743	sqlite3_changeset_iter pIter; / Iterator to skip through changeset */
231385	231744	int bInv = !!(flags & SQLITE_CHANGESETAPPLY_INVERT);
231386	231745	int rc = sessionChangesetStart(&pIter, 0, 0, nChangeset, pChangeset, bInv, 1);
231387		- u64 savedFlag = db->flags & SQLITE_FkNoAction;
231388		-
231389		- if( flags & SQLITE_CHANGESETAPPLY_FKNOACTION ){
231390		- db->flags \|= ((u64)SQLITE_FkNoAction);
231391		- db->aDb[0].pSchema->schema_cookie -= 32;
231392		- }
231393	231746
231394	231747	if( rc==SQLITE_OK ){
231395	231748	rc = sessionChangesetApply(
231396	231749	db, pIter, xFilter, xConflict, pCtx, ppRebase, pnRebase, flags
231397	231750	);
231398	231751	}
231399	231752
231400		- if( (flags & SQLITE_CHANGESETAPPLY_FKNOACTION) && savedFlag==0 ){
231401		- assert( db->flags & SQLITE_FkNoAction );
231402		- db->flags &= ~((u64)SQLITE_FkNoAction);
231403		- db->aDb[0].pSchema->schema_cookie -= 32;
231404		- }
231405	231753	return rc;
231406	231754	}
231407	231755
231408	231756	/*
231409	231757	** Apply the changeset passed via pChangeset/nChangeset to the main database
		@@ -231720,10 +232068,13 @@
231720	232068	if( op==SQLITE_INSERT \|\| (op==SQLITE_DELETE && pGrp->bPatch==0) ){
231721	232069	/* Append the missing default column values to the record. */
231722	232070	sessionAppendBlob(pOut, aRec, nRec, &rc);
231723	232071	if( rc==SQLITE_OK && pTab->pDfltStmt==0 ){
231724	232072	rc = sessionPrepareDfltStmt(pGrp->db, pTab, &pTab->pDfltStmt);
	232073	+ if( rc==SQLITE_OK && SQLITE_ROW!=sqlite3_step(pTab->pDfltStmt) ){
	232074	+ rc = sqlite3_errcode(pGrp->db);
	232075	+ }
231725	232076	}
231726	232077	for(ii=nCol; rc==SQLITE_OK && ii<pTab->nCol; ii++){
231727	232078	int eType = sqlite3_column_type(pTab->pDfltStmt, ii);
231728	232079	sessionAppendByte(pOut, eType, &rc);
231729	232080	switch( eType ){
		@@ -231736,10 +232087,11 @@
231736	232087	double rVal = sqlite3_column_int64(pTab->pDfltStmt, ii);
231737	232088	memcpy(&iVal, &rVal, sizeof(i64));
231738	232089	}
231739	232090	if( SQLITE_OK==sessionBufferGrow(pOut, 8, &rc) ){
231740	232091	sessionPutI64(&pOut->aBuf[pOut->nBuf], iVal);
	232092	+ pOut->nBuf += 8;
231741	232093	}
231742	232094	break;
231743	232095	}
231744	232096
231745	232097	case SQLITE_BLOB:
		@@ -231874,10 +232226,12 @@
231874	232226	SessionChange *pExist = 0;
231875	232227	SessionChange **pp = 0;
231876	232228	SessionTable *pTab = 0;
231877	232229	u8 *aRec = &pIter->in.aData[pIter->in.iCurrent + 2];
231878	232230	int nRec = (pIter->in.iNext - pIter->in.iCurrent) - 2;
	232231	+
	232232	+ assert( nRec>0 );
231879	232233
231880	232234	/* Ensure that only changesets, or only patchsets, but not a mixture
231881	232235	** of both, are being combined. It is an error to try to combine a
231882	232236	** changeset and a patchset. */
231883	232237	if( pGrp->pList==0 ){
		@@ -231952,10 +232306,11 @@
231952	232306	){
231953	232307	u8 *aRec;
231954	232308	int nRec;
231955	232309	int rc = SQLITE_OK;
231956	232310
	232311	+ pIter->in.bNoDiscard = 1;
231957	232312	while( SQLITE_ROW==(sessionChangesetNext(pIter, &aRec, &nRec, 0)) ){
231958	232313	rc = sessionOneChangeToHash(pGrp, pIter, bRebase);
231959	232314	if( rc!=SQLITE_OK ) break;
231960	232315	}
231961	232316
		@@ -232583,20 +232938,46 @@
232583	232938	#endif /* SQLITE_ENABLE_SESSION && SQLITE_ENABLE_PREUPDATE_HOOK */
232584	232939
232585	232940	/************ End of sqlite3session.c ************************************/
232586	232941	/************ Begin file fts5.c ******************************************/
232587	232942
232588		-
	232943	+/*
	232944	+** This, the "fts5.c" source file, is a composite file that is itself
	232945	+** assembled from the following files:
	232946	+**
	232947	+** fts5.h
	232948	+** fts5Int.h
	232949	+** fts5parse.h <--- Generated from fts5parse.y by Lemon
	232950	+** fts5parse.c <--- Generated from fts5parse.y by Lemon
	232951	+** fts5_aux.c
	232952	+** fts5_buffer.c
	232953	+** fts5_config.c
	232954	+** fts5_expr.c
	232955	+** fts5_hash.c
	232956	+** fts5_index.c
	232957	+** fts5_main.c
	232958	+** fts5_storage.c
	232959	+** fts5_tokenize.c
	232960	+** fts5_unicode2.c
	232961	+** fts5_varint.c
	232962	+** fts5_vocab.c
	232963	+*/
232589	232964	#if !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_FTS5)
232590	232965
232591	232966	#if !defined(NDEBUG) && !defined(SQLITE_DEBUG)
232592	232967	# define NDEBUG 1
232593	232968	#endif
232594	232969	#if defined(NDEBUG) && defined(SQLITE_DEBUG)
232595	232970	# undef NDEBUG
232596	232971	#endif
232597	232972
	232973	+#ifdef HAVE_STDINT_H
	232974	+/* #include <stdint.h> */
	232975	+#endif
	232976	+#ifdef HAVE_INTTYPES_H
	232977	+/* #include <inttypes.h> */
	232978	+#endif
232598	232979	/*
232599	232980	** 2014 May 31
232600	232981	**
232601	232982	** The author disclaims copyright to this source code. In place of
232602	232983	** a legal notice, here is a blessing:
		@@ -232893,17 +233274,32 @@
232893	233274	** This is used to access token iToken of phrase hit iIdx within the
232894	233275	** current row. If iIdx is less than zero or greater than or equal to the
232895	233276	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
232896	233277	** output variable (*ppToken) is set to point to a buffer containing the
232897	233278	** matching document token, and (*pnToken) to the size of that buffer in
232898		-** bytes. This API is not available if the specified token matches a
232899		-** prefix query term. In that case both output variables are always set
232900		-** to 0.
	233279	+** bytes.
232901	233280	**
232902	233281	** The output text is not a copy of the document text that was tokenized.
232903	233282	** It is the output of the tokenizer module. For tokendata=1 tables, this
232904	233283	** includes any embedded 0x00 and trailing data.
	233284	+**
	233285	+** This API may be slow in some cases if the token identified by parameters
	233286	+** iIdx and iToken matched a prefix token in the query. In most cases, the
	233287	+** first call to this API for each prefix token in the query is forced
	233288	+** to scan the portion of the full-text index that matches the prefix
	233289	+** token to collect the extra data required by this API. If the prefix
	233290	+** token matches a large number of token instances in the document set,
	233291	+** this may be a performance problem.
	233292	+**
	233293	+** If the user knows in advance that a query may use this API for a
	233294	+** prefix token, FTS5 may be configured to collect all required data as part
	233295	+** of the initial querying of the full-text index, avoiding the second scan
	233296	+** entirely. This also causes prefix queries that do not use this API to
	233297	+** run more slowly and use more memory. FTS5 may be configured in this way
	233298	+** either on a per-table basis using the [FTS5 insttoken \| 'insttoken']
	233299	+** option, or on a per-query basis using the
	233300	+** [fts5_insttoken \| fts5_insttoken()] user function.
232905	233301	**
232906	233302	** This API can be quite slow if used with an FTS5 table created with the
232907	233303	** "detail=none" or "detail=column" option.
232908	233304	**
232909	233305	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
		@@ -232991,11 +233387,10 @@
232991	233387	** CUSTOM TOKENIZERS
232992	233388	**
232993	233389	** Applications may also register custom tokenizer types. A tokenizer
232994	233390	** is registered by providing fts5 with a populated instance of the
232995	233391	** following structure. All structure methods must be defined, setting
232996		-**
232997	233392	** any member of the fts5_tokenizer struct to NULL leads to undefined
232998	233393	** behaviour. The structure methods are expected to function as follows:
232999	233394	**
233000	233395	** xCreate:
233001	233396	** This function is used to allocate and initialize a tokenizer instance.
		@@ -233560,10 +233955,11 @@
233560	233955	u8 abUnindexed; / True for unindexed columns */
233561	233956	int nPrefix; /* Number of prefix indexes */
233562	233957	int aPrefix; / Sizes in bytes of nPrefix prefix indexes */
233563	233958	int eContent; /* An FTS5_CONTENT value */
233564	233959	int bContentlessDelete; /* "contentless_delete=" option (dflt==0) */
	233960	+ int bContentlessUnindexed; /* "contentless_unindexed=" option (dflt=0) */
233565	233961	char zContent; / content table */
233566	233962	char zContentRowid; / "content_rowid=" option value */
233567	233963	int bColumnsize; /* "columnsize=" option value (dflt==1) */
233568	233964	int bTokendata; /* "tokendata=" option value (dflt==0) */
233569	233965	int bLocale; /* "locale=" option value (dflt==0) */
		@@ -233582,11 +233978,12 @@
233582	233978	int nUsermerge; /* 'usermerge' setting */
233583	233979	int nHashSize; /* Bytes of memory for in-memory hash */
233584	233980	char zRank; / Name of rank function */
233585	233981	char zRankArgs; / Arguments to rank function */
233586	233982	int bSecureDelete; /* 'secure-delete' */
233587		- int nDeleteMerge; /* 'deletemerge' */
	233983	+ int nDeleteMerge; /* 'deletemerge' */
	233984	+ int bPrefixInsttoken; /* 'prefix-insttoken' */
233588	233985
233589	233986	/* If non-NULL, points to sqlite3_vtab.base.zErrmsg. Often NULL. */
233590	233987	char **pzErrmsg;
233591	233988
233592	233989	#ifdef SQLITE_DEBUG
		@@ -233598,13 +233995,14 @@
233598	233995	** the expected version if the 'secure-delete' option has ever been
233599	233996	** set on the table. */
233600	233997	#define FTS5_CURRENT_VERSION 4
233601	233998	#define FTS5_CURRENT_VERSION_SECUREDELETE 5
233602	233999
233603		-#define FTS5_CONTENT_NORMAL 0
233604		-#define FTS5_CONTENT_NONE 1
233605		-#define FTS5_CONTENT_EXTERNAL 2
	234000	+#define FTS5_CONTENT_NORMAL 0
	234001	+#define FTS5_CONTENT_NONE 1
	234002	+#define FTS5_CONTENT_EXTERNAL 2
	234003	+#define FTS5_CONTENT_UNINDEXED 3
233606	234004
233607	234005	#define FTS5_DETAIL_FULL 0
233608	234006	#define FTS5_DETAIL_NONE 1
233609	234007	#define FTS5_DETAIL_COLUMNS 2
233610	234008
		@@ -233838,11 +234236,18 @@
233838	234236	static int sqlite3Fts5StructureTest(Fts5Index, void);
233839	234237
233840	234238	/*
233841	234239	** Used by xInstToken():
233842	234240	*/
233843		-static int sqlite3Fts5IterToken(Fts5IndexIter, i64, int, int, const char, int);
	234241	+static int sqlite3Fts5IterToken(
	234242	+ Fts5IndexIter *pIndexIter,
	234243	+ const char *pToken, int nToken,
	234244	+ i64 iRowid,
	234245	+ int iCol,
	234246	+ int iOff,
	234247	+ const char *ppOut, int pnOut
	234248	+);
233844	234249
233845	234250	/*
233846	234251	** Insert or remove data to or from the index. Each time a document is
233847	234252	** added to or removed from the index, this function is called one or more
233848	234253	** times.
		@@ -233972,20 +234377,17 @@
233972	234377
233973	234378	static Fts5Table sqlite3Fts5TableFromCsrid(Fts5Global, i64);
233974	234379
233975	234380	static int sqlite3Fts5FlushToDisk(Fts5Table*);
233976	234381
233977		-static int sqlite3Fts5ExtractText(
233978		- Fts5Config *pConfig,
233979		- sqlite3_value pVal, / Value to extract text from */
233980		- int bContent, /* Loaded from content table */
233981		- int pbResetTokenizer, / OUT: True if ClearLocale() required */
233982		- const char *ppText, / OUT: Pointer to text buffer */
233983		- int pnText / OUT: Size of (ppText) in bytes /
233984		-);
233985		-
233986	234382	static void sqlite3Fts5ClearLocale(Fts5Config *pConfig);
	234383	+static void sqlite3Fts5SetLocale(Fts5Config pConfig, const char pLoc, int nLoc);
	234384	+
	234385	+static int sqlite3Fts5IsLocaleValue(Fts5Config pConfig, sqlite3_value pVal);
	234386	+static int sqlite3Fts5DecodeLocaleValue(sqlite3_value *pVal,
	234387	+ const char *ppText, int pnText, const char *ppLoc, int pnLoc
	234388	+);
233987	234389
233988	234390	/*
233989	234391	** End of interface to code in fts5.c.
233990	234392	**************************************************************************/
233991	234393
		@@ -234063,11 +234465,11 @@
234063	234465
234064	234466	static int sqlite3Fts5DropAll(Fts5Config*);
234065	234467	static int sqlite3Fts5CreateTable(Fts5Config, const char, const char, int, char *);
234066	234468
234067	234469	static int sqlite3Fts5StorageDelete(Fts5Storage p, i64, sqlite3_value*, int);
234068		-static int sqlite3Fts5StorageContentInsert(Fts5Storage p, sqlite3_value, i64);
	234470	+static int sqlite3Fts5StorageContentInsert(Fts5Storage p, int, sqlite3_value, i64);
234069	234471	static int sqlite3Fts5StorageIndexInsert(Fts5Storage p, sqlite3_value*, i64);
234070	234472
234071	234473	static int sqlite3Fts5StorageIntegrity(Fts5Storage *p, int iArg);
234072	234474
234073	234475	static int sqlite3Fts5StorageStmt(Fts5Storage p, int eStmt, sqlite3_stmt, char*);
		@@ -237270,10 +237672,11 @@
237270	237672	const char zArg, / Argument to parse */
237271	237673	char *pzErr / OUT: Error message */
237272	237674	){
237273	237675	int rc = SQLITE_OK;
237274	237676	int nCmd = (int)strlen(zCmd);
	237677	+
237275	237678	if( sqlite3_strnicmp("prefix", zCmd, nCmd)==0 ){
237276	237679	const int nByte = sizeof(int) * FTS5_MAX_PREFIX_INDEXES;
237277	237680	const char *p;
237278	237681	int bFirst = 1;
237279	237682	if( pConfig->aPrefix==0 ){
		@@ -237388,10 +237791,20 @@
237388	237791	}else{
237389	237792	pConfig->bContentlessDelete = (zArg[0]=='1');
237390	237793	}
237391	237794	return rc;
237392	237795	}
	237796	+
	237797	+ if( sqlite3_strnicmp("contentless_unindexed", zCmd, nCmd)==0 ){
	237798	+ if( (zArg[0]!='0' && zArg[0]!='1') \|\| zArg[1]!='\0' ){
	237799	+ *pzErr = sqlite3_mprintf("malformed contentless_delete=... directive");
	237800	+ rc = SQLITE_ERROR;
	237801	+ }else{
	237802	+ pConfig->bContentlessUnindexed = (zArg[0]=='1');
	237803	+ }
	237804	+ return rc;
	237805	+ }
237393	237806
237394	237807	if( sqlite3_strnicmp("content_rowid", zCmd, nCmd)==0 ){
237395	237808	if( pConfig->zContentRowid ){
237396	237809	*pzErr = sqlite3_mprintf("multiple content_rowid=... directives");
237397	237810	rc = SQLITE_ERROR;
		@@ -237506,11 +237919,12 @@
237506	237919
237507	237920	static int fts5ConfigParseColumn(
237508	237921	Fts5Config *p,
237509	237922	char *zCol,
237510	237923	char *zArg,
237511		- char **pzErr
	237924	+ char **pzErr,
	237925	+ int *pbUnindexed
237512	237926	){
237513	237927	int rc = SQLITE_OK;
237514	237928	if( 0==sqlite3_stricmp(zCol, FTS5_RANK_NAME)
237515	237929	\|\| 0==sqlite3_stricmp(zCol, FTS5_ROWID_NAME)
237516	237930	){
		@@ -237517,10 +237931,11 @@
237517	237931	*pzErr = sqlite3_mprintf("reserved fts5 column name: %s", zCol);
237518	237932	rc = SQLITE_ERROR;
237519	237933	}else if( zArg ){
237520	237934	if( 0==sqlite3_stricmp(zArg, "unindexed") ){
237521	237935	p->abUnindexed[p->nCol] = 1;
	237936	+ *pbUnindexed = 1;
237522	237937	}else{
237523	237938	*pzErr = sqlite3_mprintf("unrecognized column option: %s", zArg);
237524	237939	rc = SQLITE_ERROR;
237525	237940	}
237526	237941	}
		@@ -237537,15 +237952,30 @@
237537	237952	int rc = SQLITE_OK;
237538	237953	Fts5Buffer buf = {0, 0, 0};
237539	237954
237540	237955	sqlite3Fts5BufferAppendPrintf(&rc, &buf, "T.%Q", p->zContentRowid);
237541	237956	if( p->eContent!=FTS5_CONTENT_NONE ){
	237957	+ assert( p->eContent==FTS5_CONTENT_EXTERNAL
	237958	+ \|\| p->eContent==FTS5_CONTENT_NORMAL
	237959	+ \|\| p->eContent==FTS5_CONTENT_UNINDEXED
	237960	+ );
237542	237961	for(i=0; i<p->nCol; i++){
237543	237962	if( p->eContent==FTS5_CONTENT_EXTERNAL ){
237544	237963	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.%Q", p->azCol[i]);
237545		- }else{
	237964	+ }else if( p->eContent==FTS5_CONTENT_NORMAL \|\| p->abUnindexed[i] ){
237546	237965	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.c%d", i);
	237966	+ }else{
	237967	+ sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", NULL");
	237968	+ }
	237969	+ }
	237970	+ }
	237971	+ if( p->eContent==FTS5_CONTENT_NORMAL && p->bLocale ){
	237972	+ for(i=0; i<p->nCol; i++){
	237973	+ if( p->abUnindexed[i]==0 ){
	237974	+ sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.l%d", i);
	237975	+ }else{
	237976	+ sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", NULL");
237547	237977	}
237548	237978	}
237549	237979	}
237550	237980
237551	237981	assert( p->zContentExprlist==0 );
		@@ -237575,10 +238005,11 @@
237575	238005	){
237576	238006	int rc = SQLITE_OK; /* Return code */
237577	238007	Fts5Config pRet; / New object to return */
237578	238008	int i;
237579	238009	sqlite3_int64 nByte;
	238010	+ int bUnindexed = 0; /* True if there are one or more UNINDEXED */
237580	238011
237581	238012	ppOut = pRet = (Fts5Config)sqlite3_malloc(sizeof(Fts5Config));
237582	238013	if( pRet==0 ) return SQLITE_NOMEM;
237583	238014	memset(pRet, 0, sizeof(Fts5Config));
237584	238015	pRet->pGlobal = pGlobal;
		@@ -237634,11 +238065,11 @@
237634	238065	ALWAYS(zOne)?zOne:"",
237635	238066	zTwo?zTwo:"",
237636	238067	pzErr
237637	238068	);
237638	238069	}else{
237639		- rc = fts5ConfigParseColumn(pRet, zOne, zTwo, pzErr);
	238070	+ rc = fts5ConfigParseColumn(pRet, zOne, zTwo, pzErr, &bUnindexed);
237640	238071	zOne = 0;
237641	238072	}
237642	238073	}
237643	238074	}
237644	238075
		@@ -237665,18 +238096,34 @@
237665	238096	*pzErr = sqlite3_mprintf(
237666	238097	"contentless_delete=1 is incompatible with columnsize=0"
237667	238098	);
237668	238099	rc = SQLITE_ERROR;
237669	238100	}
	238101	+
	238102	+ /* We only allow contentless_unindexed=1 if the table is actually a
	238103	+ ** contentless one.
	238104	+ */
	238105	+ if( rc==SQLITE_OK
	238106	+ && pRet->bContentlessUnindexed
	238107	+ && pRet->eContent!=FTS5_CONTENT_NONE
	238108	+ ){
	238109	+ *pzErr = sqlite3_mprintf(
	238110	+ "contentless_unindexed=1 requires a contentless table"
	238111	+ );
	238112	+ rc = SQLITE_ERROR;
	238113	+ }
237670	238114
237671	238115	/* If no zContent option was specified, fill in the default values. */
237672	238116	if( rc==SQLITE_OK && pRet->zContent==0 ){
237673	238117	const char *zTail = 0;
237674	238118	assert( pRet->eContent==FTS5_CONTENT_NORMAL
237675	238119	\|\| pRet->eContent==FTS5_CONTENT_NONE
237676	238120	);
237677	238121	if( pRet->eContent==FTS5_CONTENT_NORMAL ){
	238122	+ zTail = "content";
	238123	+ }else if( bUnindexed && pRet->bContentlessUnindexed ){
	238124	+ pRet->eContent = FTS5_CONTENT_UNINDEXED;
237678	238125	zTail = "content";
237679	238126	}else if( pRet->bColumnsize ){
237680	238127	zTail = "docsize";
237681	238128	}
237682	238129
		@@ -238010,10 +238457,23 @@
238010	238457	if( bVal<0 ){
238011	238458	*pbBadkey = 1;
238012	238459	}else{
238013	238460	pConfig->bSecureDelete = (bVal ? 1 : 0);
238014	238461	}
	238462	+ }
	238463	+
	238464	+ else if( 0==sqlite3_stricmp(zKey, "insttoken") ){
	238465	+ int bVal = -1;
	238466	+ if( SQLITE_INTEGER==sqlite3_value_numeric_type(pVal) ){
	238467	+ bVal = sqlite3_value_int(pVal);
	238468	+ }
	238469	+ if( bVal<0 ){
	238470	+ *pbBadkey = 1;
	238471	+ }else{
	238472	+ pConfig->bPrefixInsttoken = (bVal ? 1 : 0);
	238473	+ }
	238474	+
238015	238475	}else{
238016	238476	*pbBadkey = 1;
238017	238477	}
238018	238478	return rc;
238019	238479	}
		@@ -241145,11 +241605,11 @@
241145	241605	&& memcmp(pT->pTerm, pToken, pT->nQueryTerm)==0
241146	241606	){
241147	241607	int rc = sqlite3Fts5PoslistWriterAppend(
241148	241608	&pExpr->apExprPhrase[i]->poslist, &p->aPopulator[i].writer, p->iOff
241149	241609	);
241150		- if( rc==SQLITE_OK && pExpr->pConfig->bTokendata && !pT->bPrefix ){
	241610	+ if( rc==SQLITE_OK && (pExpr->pConfig->bTokendata \|\| pT->bPrefix) ){
241151	241611	int iCol = p->iOff>>32;
241152	241612	int iTokOff = p->iOff & 0x7FFFFFFF;
241153	241613	rc = sqlite3Fts5IndexIterWriteTokendata(
241154	241614	pT->pIter, pToken, nToken, iRowid, iCol, iTokOff
241155	241615	);
		@@ -241338,19 +241798,18 @@
241338	241798	pPhrase = pExpr->apExprPhrase[iPhrase];
241339	241799	if( iToken<0 \|\| iToken>=pPhrase->nTerm ){
241340	241800	return SQLITE_RANGE;
241341	241801	}
241342	241802	pTerm = &pPhrase->aTerm[iToken];
241343		- if( pTerm->bPrefix==0 ){
241344		- if( pExpr->pConfig->bTokendata ){
241345		- rc = sqlite3Fts5IterToken(
241346		- pTerm->pIter, iRowid, iCol, iOff+iToken, ppOut, pnOut
241347		- );
241348		- }else{
241349		- *ppOut = pTerm->pTerm;
241350		- *pnOut = pTerm->nFullTerm;
241351		- }
	241803	+ if( pExpr->pConfig->bTokendata \|\| pTerm->bPrefix ){
	241804	+ rc = sqlite3Fts5IterToken(
	241805	+ pTerm->pIter, pTerm->pTerm, pTerm->nQueryTerm,
	241806	+ iRowid, iCol, iOff+iToken, ppOut, pnOut
	241807	+ );
	241808	+ }else{
	241809	+ *ppOut = pTerm->pTerm;
	241810	+ *pnOut = pTerm->nFullTerm;
241352	241811	}
241353	241812	return rc;
241354	241813	}
241355	241814
241356	241815	/*
		@@ -246847,10 +247306,15 @@
246847	247306	if( nPercent>=pConfig->nDeleteMerge && nPercent>nBest ){
246848	247307	iRet = ii;
246849	247308	nBest = nPercent;
246850	247309	}
246851	247310	}
	247311	+
	247312	+ /* If pLvl is already the input level to an ongoing merge, look no
	247313	+ ** further for a merge candidate. The caller should be allowed to
	247314	+ ** continue merging from pLvl first. */
	247315	+ if( pLvl->nMerge ) break;
246852	247316	}
246853	247317	}
246854	247318	return iRet;
246855	247319	}
246856	247320
		@@ -248156,10 +248620,387 @@
248156	248620	fts5BufferFree(&tmp);
248157	248621	memset(&out.p[out.n], 0, FTS5_DATA_ZERO_PADDING);
248158	248622	*p1 = out;
248159	248623	}
248160	248624
	248625	+
	248626	+/*
	248627	+** Iterate through a range of entries in the FTS index, invoking the xVisit
	248628	+** callback for each of them.
	248629	+**
	248630	+** Parameter pToken points to an nToken buffer containing an FTS index term
	248631	+** (i.e. a document term with the preceding 1 byte index identifier -
	248632	+** FTS5_MAIN_PREFIX or similar). If bPrefix is true, then the call visits
	248633	+** all entries for terms that have pToken/nToken as a prefix. If bPrefix
	248634	+** is false, then only entries with pToken/nToken as the entire key are
	248635	+** visited.
	248636	+**
	248637	+** If the current table is a tokendata=1 table, then if bPrefix is true then
	248638	+** each index term is treated separately. However, if bPrefix is false, then
	248639	+** all index terms corresponding to pToken/nToken are collapsed into a single
	248640	+** term before the callback is invoked.
	248641	+**
	248642	+** The callback invoked for each entry visited is specified by paramter xVisit.
	248643	+** Each time it is invoked, it is passed a pointer to the Fts5Index object,
	248644	+** a copy of the 7th paramter to this function (pCtx) and a pointer to the
	248645	+** iterator that indicates the current entry. If the current entry is the
	248646	+** first with a new term (i.e. different from that of the previous entry,
	248647	+** including the very first term), then the final two parameters are passed
	248648	+** a pointer to the term and its size in bytes, respectively. If the current
	248649	+** entry is not the first associated with its term, these two parameters
	248650	+** are passed 0.
	248651	+**
	248652	+** If parameter pColset is not NULL, then it is used to filter entries before
	248653	+** the callback is invoked.
	248654	+*/
	248655	+static int fts5VisitEntries(
	248656	+ Fts5Index p, / Fts5 index object */
	248657	+ Fts5Colset pColset, / Columns filter to apply, or NULL */
	248658	+ u8 pToken, / Buffer containing token */
	248659	+ int nToken, /* Size of buffer pToken in bytes */
	248660	+ int bPrefix, /* True for a prefix scan */
	248661	+ void (xVisit)(Fts5Index, void pCtx, Fts5Iter pIter, const u8*, int),
	248662	+ void pCtx / Passed as second argument to xVisit() */
	248663	+){
	248664	+ const int flags = (bPrefix ? FTS5INDEX_QUERY_SCAN : 0)
	248665	+ \| FTS5INDEX_QUERY_SKIPEMPTY
	248666	+ \| FTS5INDEX_QUERY_NOOUTPUT;
	248667	+ Fts5Iter p1 = 0; / Iterator used to gather data from index */
	248668	+ int bNewTerm = 1;
	248669	+ Fts5Structure *pStruct = fts5StructureRead(p);
	248670	+
	248671	+ fts5MultiIterNew(p, pStruct, flags, pColset, pToken, nToken, -1, 0, &p1);
	248672	+ fts5IterSetOutputCb(&p->rc, p1);
	248673	+ for( /* no-op */ ;
	248674	+ fts5MultiIterEof(p, p1)==0;
	248675	+ fts5MultiIterNext2(p, p1, &bNewTerm)
	248676	+ ){
	248677	+ Fts5SegIter *pSeg = &p1->aSeg[ p1->aFirst[1].iFirst ];
	248678	+ int nNew = 0;
	248679	+ const u8 *pNew = 0;
	248680	+
	248681	+ p1->xSetOutputs(p1, pSeg);
	248682	+ if( p->rc ) break;
	248683	+
	248684	+ if( bNewTerm ){
	248685	+ nNew = pSeg->term.n;
	248686	+ pNew = pSeg->term.p;
	248687	+ if( nNew<nToken \|\| memcmp(pToken, pNew, nToken) ) break;
	248688	+ }
	248689	+
	248690	+ xVisit(p, pCtx, p1, pNew, nNew);
	248691	+ }
	248692	+ fts5MultiIterFree(p1);
	248693	+
	248694	+ fts5StructureRelease(pStruct);
	248695	+ return p->rc;
	248696	+}
	248697	+
	248698	+
	248699	+/*
	248700	+** Usually, a tokendata=1 iterator (struct Fts5TokenDataIter) accumulates an
	248701	+** array of these for each row it visits (so all iRowid fields are the same).
	248702	+** Or, for an iterator used by an "ORDER BY rank" query, it accumulates an
	248703	+** array of these for the entire query (in which case iRowid fields may take
	248704	+** a variety of values).
	248705	+**
	248706	+** Each instance in the array indicates the iterator (and therefore term)
	248707	+** associated with position iPos of rowid iRowid. This is used by the
	248708	+** xInstToken() API.
	248709	+**
	248710	+** iRowid:
	248711	+** Rowid for the current entry.
	248712	+**
	248713	+** iPos:
	248714	+** Position of current entry within row. In the usual ((iCol<<32)+iOff)
	248715	+** format (e.g. see macros FTS5_POS2COLUMN() and FTS5_POS2OFFSET()).
	248716	+**
	248717	+** iIter:
	248718	+** If the Fts5TokenDataIter iterator that the entry is part of is
	248719	+** actually an iterator (i.e. with nIter>0, not just a container for
	248720	+** Fts5TokenDataMap structures), then this variable is an index into
	248721	+** the apIter[] array. The corresponding term is that which the iterator
	248722	+** at apIter[iIter] currently points to.
	248723	+**
	248724	+** Or, if the Fts5TokenDataIter iterator is just a container object
	248725	+** (nIter==0), then iIter is an index into the term.p[] buffer where
	248726	+** the term is stored.
	248727	+**
	248728	+** nByte:
	248729	+** In the case where iIter is an index into term.p[], this variable
	248730	+** is the size of the term in bytes. If iIter is an index into apIter[],
	248731	+** this variable is unused.
	248732	+*/
	248733	+struct Fts5TokenDataMap {
	248734	+ i64 iRowid; /* Row this token is located in */
	248735	+ i64 iPos; /* Position of token */
	248736	+ int iIter; /* Iterator token was read from */
	248737	+ int nByte; /* Length of token in bytes (or 0) */
	248738	+};
	248739	+
	248740	+/*
	248741	+** An object used to supplement Fts5Iter for tokendata=1 iterators.
	248742	+**
	248743	+** This object serves two purposes. The first is as a container for an array
	248744	+** of Fts5TokenDataMap structures, which are used to find the token required
	248745	+** when the xInstToken() API is used. This is done by the nMapAlloc, nMap and
	248746	+** aMap[] variables.
	248747	+*/
	248748	+struct Fts5TokenDataIter {
	248749	+ int nMapAlloc; /* Allocated size of aMap[] in entries */
	248750	+ int nMap; /* Number of valid entries in aMap[] */
	248751	+ Fts5TokenDataMap aMap; / Array of (rowid+pos -> token) mappings */
	248752	+
	248753	+ /* The following are used for prefix-queries only. */
	248754	+ Fts5Buffer terms;
	248755	+
	248756	+ /* The following are used for other full-token tokendata queries only. */
	248757	+ int nIter;
	248758	+ int nIterAlloc;
	248759	+ Fts5PoslistReader *aPoslistReader;
	248760	+ int *aPoslistToIter;
	248761	+ Fts5Iter *apIter[1];
	248762	+};
	248763	+
	248764	+/*
	248765	+** The two input arrays - a1[] and a2[] - are in sorted order. This function
	248766	+** merges the two arrays together and writes the result to output array
	248767	+** aOut[]. aOut[] is guaranteed to be large enough to hold the result.
	248768	+**
	248769	+** Duplicate entries are copied into the output. So the size of the output
	248770	+** array is always (n1+n2) entries.
	248771	+*/
	248772	+static void fts5TokendataMerge(
	248773	+ Fts5TokenDataMap a1, int n1, / Input array 1 */
	248774	+ Fts5TokenDataMap a2, int n2, / Input array 2 */
	248775	+ Fts5TokenDataMap aOut / Output array */
	248776	+){
	248777	+ int i1 = 0;
	248778	+ int i2 = 0;
	248779	+
	248780	+ assert( n1>=0 && n2>=0 );
	248781	+ while( i1<n1 \|\| i2<n2 ){
	248782	+ Fts5TokenDataMap *pOut = &aOut[i1+i2];
	248783	+ if( i2>=n2 \|\| (i1<n1 && (
	248784	+ a1[i1].iRowid<a2[i2].iRowid
	248785	+ \|\| (a1[i1].iRowid==a2[i2].iRowid && a1[i1].iPos<=a2[i2].iPos)
	248786	+ ))){
	248787	+ memcpy(pOut, &a1[i1], sizeof(Fts5TokenDataMap));
	248788	+ i1++;
	248789	+ }else{
	248790	+ memcpy(pOut, &a2[i2], sizeof(Fts5TokenDataMap));
	248791	+ i2++;
	248792	+ }
	248793	+ }
	248794	+}
	248795	+
	248796	+
	248797	+/*
	248798	+** Append a mapping to the token-map belonging to object pT.
	248799	+*/
	248800	+static void fts5TokendataIterAppendMap(
	248801	+ Fts5Index *p,
	248802	+ Fts5TokenDataIter *pT,
	248803	+ int iIter,
	248804	+ int nByte,
	248805	+ i64 iRowid,
	248806	+ i64 iPos
	248807	+){
	248808	+ if( p->rc==SQLITE_OK ){
	248809	+ if( pT->nMap==pT->nMapAlloc ){
	248810	+ int nNew = pT->nMapAlloc ? pT->nMapAlloc*2 : 64;
	248811	+ int nAlloc = nNew * sizeof(Fts5TokenDataMap);
	248812	+ Fts5TokenDataMap *aNew;
	248813	+
	248814	+ aNew = (Fts5TokenDataMap*)sqlite3_realloc(pT->aMap, nAlloc);
	248815	+ if( aNew==0 ){
	248816	+ p->rc = SQLITE_NOMEM;
	248817	+ return;
	248818	+ }
	248819	+
	248820	+ pT->aMap = aNew;
	248821	+ pT->nMapAlloc = nNew;
	248822	+ }
	248823	+
	248824	+ pT->aMap[pT->nMap].iRowid = iRowid;
	248825	+ pT->aMap[pT->nMap].iPos = iPos;
	248826	+ pT->aMap[pT->nMap].iIter = iIter;
	248827	+ pT->aMap[pT->nMap].nByte = nByte;
	248828	+ pT->nMap++;
	248829	+ }
	248830	+}
	248831	+
	248832	+/*
	248833	+** Sort the contents of the pT->aMap[] array.
	248834	+**
	248835	+** The sorting algorithm requries a malloc(). If this fails, an error code
	248836	+** is left in Fts5Index.rc before returning.
	248837	+*/
	248838	+static void fts5TokendataIterSortMap(Fts5Index p, Fts5TokenDataIter pT){
	248839	+ Fts5TokenDataMap *aTmp = 0;
	248840	+ int nByte = pT->nMap * sizeof(Fts5TokenDataMap);
	248841	+
	248842	+ aTmp = (Fts5TokenDataMap*)sqlite3Fts5MallocZero(&p->rc, nByte);
	248843	+ if( aTmp ){
	248844	+ Fts5TokenDataMap *a1 = pT->aMap;
	248845	+ Fts5TokenDataMap *a2 = aTmp;
	248846	+ i64 nHalf;
	248847	+
	248848	+ for(nHalf=1; nHalf<pT->nMap; nHalf=nHalf*2){
	248849	+ int i1;
	248850	+ for(i1=0; i1<pT->nMap; i1+=(nHalf*2)){
	248851	+ int n1 = MIN(nHalf, pT->nMap-i1);
	248852	+ int n2 = MIN(nHalf, pT->nMap-i1-n1);
	248853	+ fts5TokendataMerge(&a1[i1], n1, &a1[i1+n1], n2, &a2[i1]);
	248854	+ }
	248855	+ SWAPVAL(Fts5TokenDataMap*, a1, a2);
	248856	+ }
	248857	+
	248858	+ if( a1!=pT->aMap ){
	248859	+ memcpy(pT->aMap, a1, pT->nMap*sizeof(Fts5TokenDataMap));
	248860	+ }
	248861	+ sqlite3_free(aTmp);
	248862	+
	248863	+#ifdef SQLITE_DEBUG
	248864	+ {
	248865	+ int ii;
	248866	+ for(ii=1; ii<pT->nMap; ii++){
	248867	+ Fts5TokenDataMap *p1 = &pT->aMap[ii-1];
	248868	+ Fts5TokenDataMap *p2 = &pT->aMap[ii];
	248869	+ assert( p1->iRowid<p2->iRowid
	248870	+ \|\| (p1->iRowid==p2->iRowid && p1->iPos<=p2->iPos)
	248871	+ );
	248872	+ }
	248873	+ }
	248874	+#endif
	248875	+ }
	248876	+}
	248877	+
	248878	+/*
	248879	+** Delete an Fts5TokenDataIter structure and its contents.
	248880	+*/
	248881	+static void fts5TokendataIterDelete(Fts5TokenDataIter *pSet){
	248882	+ if( pSet ){
	248883	+ int ii;
	248884	+ for(ii=0; ii<pSet->nIter; ii++){
	248885	+ fts5MultiIterFree(pSet->apIter[ii]);
	248886	+ }
	248887	+ fts5BufferFree(&pSet->terms);
	248888	+ sqlite3_free(pSet->aPoslistReader);
	248889	+ sqlite3_free(pSet->aMap);
	248890	+ sqlite3_free(pSet);
	248891	+ }
	248892	+}
	248893	+
	248894	+
	248895	+/*
	248896	+** fts5VisitEntries() context object used by fts5SetupPrefixIterTokendata()
	248897	+** to pass data to prefixIterSetupTokendataCb().
	248898	+*/
	248899	+typedef struct TokendataSetupCtx TokendataSetupCtx;
	248900	+struct TokendataSetupCtx {
	248901	+ Fts5TokenDataIter pT; / Object being populated with mappings */
	248902	+ int iTermOff; /* Offset of current term in terms.p[] */
	248903	+ int nTermByte; /* Size of current term in bytes */
	248904	+};
	248905	+
	248906	+/*
	248907	+** fts5VisitEntries() callback used by fts5SetupPrefixIterTokendata(). This
	248908	+** callback adds an entry to the Fts5TokenDataIter.aMap[] array for each
	248909	+** position in the current position-list. It doesn't matter that some of
	248910	+** these may be out of order - they will be sorted later.
	248911	+*/
	248912	+static void prefixIterSetupTokendataCb(
	248913	+ Fts5Index *p,
	248914	+ void *pCtx,
	248915	+ Fts5Iter *p1,
	248916	+ const u8 *pNew,
	248917	+ int nNew
	248918	+){
	248919	+ TokendataSetupCtx pSetup = (TokendataSetupCtx)pCtx;
	248920	+ int iPosOff = 0;
	248921	+ i64 iPos = 0;
	248922	+
	248923	+ if( pNew ){
	248924	+ pSetup->nTermByte = nNew-1;
	248925	+ pSetup->iTermOff = pSetup->pT->terms.n;
	248926	+ fts5BufferAppendBlob(&p->rc, &pSetup->pT->terms, nNew-1, pNew+1);
	248927	+ }
	248928	+
	248929	+ while( 0==sqlite3Fts5PoslistNext64(
	248930	+ p1->base.pData, p1->base.nData, &iPosOff, &iPos
	248931	+ ) ){
	248932	+ fts5TokendataIterAppendMap(p,
	248933	+ pSetup->pT, pSetup->iTermOff, pSetup->nTermByte, p1->base.iRowid, iPos
	248934	+ );
	248935	+ }
	248936	+}
	248937	+
	248938	+
	248939	+/*
	248940	+** Context object passed by fts5SetupPrefixIter() to fts5VisitEntries().
	248941	+*/
	248942	+typedef struct PrefixSetupCtx PrefixSetupCtx;
	248943	+struct PrefixSetupCtx {
	248944	+ void (xMerge)(Fts5Index, Fts5Buffer, int, Fts5Buffer);
	248945	+ void (xAppend)(Fts5Index, u64, Fts5Iter, Fts5Buffer);
	248946	+ i64 iLastRowid;
	248947	+ int nMerge;
	248948	+ Fts5Buffer *aBuf;
	248949	+ int nBuf;
	248950	+ Fts5Buffer doclist;
	248951	+ TokendataSetupCtx *pTokendata;
	248952	+};
	248953	+
	248954	+/*
	248955	+** fts5VisitEntries() callback used by fts5SetupPrefixIter()
	248956	+*/
	248957	+static void prefixIterSetupCb(
	248958	+ Fts5Index *p,
	248959	+ void *pCtx,
	248960	+ Fts5Iter *p1,
	248961	+ const u8 *pNew,
	248962	+ int nNew
	248963	+){
	248964	+ PrefixSetupCtx pSetup = (PrefixSetupCtx)pCtx;
	248965	+ const int nMerge = pSetup->nMerge;
	248966	+
	248967	+ if( p1->base.nData>0 ){
	248968	+ if( p1->base.iRowid<=pSetup->iLastRowid && pSetup->doclist.n>0 ){
	248969	+ int i;
	248970	+ for(i=0; p->rc==SQLITE_OK && pSetup->doclist.n; i++){
	248971	+ int i1 = i*nMerge;
	248972	+ int iStore;
	248973	+ assert( i1+nMerge<=pSetup->nBuf );
	248974	+ for(iStore=i1; iStore<i1+nMerge; iStore++){
	248975	+ if( pSetup->aBuf[iStore].n==0 ){
	248976	+ fts5BufferSwap(&pSetup->doclist, &pSetup->aBuf[iStore]);
	248977	+ fts5BufferZero(&pSetup->doclist);
	248978	+ break;
	248979	+ }
	248980	+ }
	248981	+ if( iStore==i1+nMerge ){
	248982	+ pSetup->xMerge(p, &pSetup->doclist, nMerge, &pSetup->aBuf[i1]);
	248983	+ for(iStore=i1; iStore<i1+nMerge; iStore++){
	248984	+ fts5BufferZero(&pSetup->aBuf[iStore]);
	248985	+ }
	248986	+ }
	248987	+ }
	248988	+ pSetup->iLastRowid = 0;
	248989	+ }
	248990	+
	248991	+ pSetup->xAppend(
	248992	+ p, (u64)p1->base.iRowid-(u64)pSetup->iLastRowid, p1, &pSetup->doclist
	248993	+ );
	248994	+ pSetup->iLastRowid = p1->base.iRowid;
	248995	+ }
	248996	+
	248997	+ if( pSetup->pTokendata ){
	248998	+ prefixIterSetupTokendataCb(p, (void*)pSetup->pTokendata, p1, pNew, nNew);
	248999	+ }
	249000	+}
	249001	+
248161	249002	static void fts5SetupPrefixIter(
248162	249003	Fts5Index p, / Index to read from */
248163	249004	int bDesc, /* True for "ORDER BY rowid DESC" */
248164	249005	int iIdx, /* Index to scan for data */
248165	249006	u8 pToken, / Buffer containing prefix to match */
		@@ -248166,137 +249007,89 @@
248166	249007	int nToken, /* Size of buffer pToken in bytes */
248167	249008	Fts5Colset pColset, / Restrict matches to these columns */
248168	249009	Fts5Iter *ppIter / OUT: New iterator */
248169	249010	){
248170	249011	Fts5Structure *pStruct;
248171		- Fts5Buffer *aBuf;
248172		- int nBuf = 32;
248173		- int nMerge = 1;
	249012	+ PrefixSetupCtx s;
	249013	+ TokendataSetupCtx s2;
248174	249014
248175		- void (xMerge)(Fts5Index, Fts5Buffer, int, Fts5Buffer);
248176		- void (xAppend)(Fts5Index, u64, Fts5Iter, Fts5Buffer);
	249015	+ memset(&s, 0, sizeof(s));
	249016	+ memset(&s2, 0, sizeof(s2));
	249017	+
	249018	+ s.nMerge = 1;
	249019	+ s.iLastRowid = 0;
	249020	+ s.nBuf = 32;
	249021	+ if( iIdx==0
	249022	+ && p->pConfig->eDetail==FTS5_DETAIL_FULL
	249023	+ && p->pConfig->bPrefixInsttoken
	249024	+ ){
	249025	+ s.pTokendata = &s2;
	249026	+ s2.pT = (Fts5TokenDataIter)fts5IdxMalloc(p, sizeof(s2.pT));
	249027	+ }
	249028	+
248177	249029	if( p->pConfig->eDetail==FTS5_DETAIL_NONE ){
248178		- xMerge = fts5MergeRowidLists;
248179		- xAppend = fts5AppendRowid;
	249030	+ s.xMerge = fts5MergeRowidLists;
	249031	+ s.xAppend = fts5AppendRowid;
248180	249032	}else{
248181		- nMerge = FTS5_MERGE_NLIST-1;
248182		- nBuf = nMerge8; / Sufficient to merge (16^8)==(2^32) lists */
248183		- xMerge = fts5MergePrefixLists;
248184		- xAppend = fts5AppendPoslist;
	249033	+ s.nMerge = FTS5_MERGE_NLIST-1;
	249034	+ s.nBuf = s.nMerge8; / Sufficient to merge (16^8)==(2^32) lists */
	249035	+ s.xMerge = fts5MergePrefixLists;
	249036	+ s.xAppend = fts5AppendPoslist;
248185	249037	}
248186	249038
248187		- aBuf = (Fts5Buffer)fts5IdxMalloc(p, sizeof(Fts5Buffer)nBuf);
	249039	+ s.aBuf = (Fts5Buffer)fts5IdxMalloc(p, sizeof(Fts5Buffer)s.nBuf);
248188	249040	pStruct = fts5StructureRead(p);
248189		- assert( p->rc!=SQLITE_OK \|\| (aBuf && pStruct) );
	249041	+ assert( p->rc!=SQLITE_OK \|\| (s.aBuf && pStruct) );
248190	249042
248191	249043	if( p->rc==SQLITE_OK ){
248192		- const int flags = FTS5INDEX_QUERY_SCAN
248193		- \| FTS5INDEX_QUERY_SKIPEMPTY
248194		- \| FTS5INDEX_QUERY_NOOUTPUT;
	249044	+ void pCtx = (void)&s;
248195	249045	int i;
248196		- i64 iLastRowid = 0;
248197		- Fts5Iter p1 = 0; / Iterator used to gather data from index */
248198	249046	Fts5Data *pData;
248199		- Fts5Buffer doclist;
248200		- int bNewTerm = 1;
248201		-
248202		- memset(&doclist, 0, sizeof(doclist));
248203	249047
248204	249048	/* If iIdx is non-zero, then it is the number of a prefix-index for
248205	249049	** prefixes 1 character longer than the prefix being queried for. That
248206	249050	** index contains all the doclists required, except for the one
248207	249051	** corresponding to the prefix itself. That one is extracted from the
248208	249052	** main term index here. */
248209	249053	if( iIdx!=0 ){
248210		- int dummy = 0;
248211		- const int f2 = FTS5INDEX_QUERY_SKIPEMPTY\|FTS5INDEX_QUERY_NOOUTPUT;
248212	249054	pToken[0] = FTS5_MAIN_PREFIX;
248213		- fts5MultiIterNew(p, pStruct, f2, pColset, pToken, nToken, -1, 0, &p1);
248214		- fts5IterSetOutputCb(&p->rc, p1);
248215		- for(;
248216		- fts5MultiIterEof(p, p1)==0;
248217		- fts5MultiIterNext2(p, p1, &dummy)
248218		- ){
248219		- Fts5SegIter *pSeg = &p1->aSeg[ p1->aFirst[1].iFirst ];
248220		- p1->xSetOutputs(p1, pSeg);
248221		- if( p1->base.nData ){
248222		- xAppend(p, (u64)p1->base.iRowid-(u64)iLastRowid, p1, &doclist);
248223		- iLastRowid = p1->base.iRowid;
248224		- }
248225		- }
248226		- fts5MultiIterFree(p1);
	249055	+ fts5VisitEntries(p, pColset, pToken, nToken, 0, prefixIterSetupCb, pCtx);
248227	249056	}
248228	249057
248229	249058	pToken[0] = FTS5_MAIN_PREFIX + iIdx;
248230		- fts5MultiIterNew(p, pStruct, flags, pColset, pToken, nToken, -1, 0, &p1);
248231		- fts5IterSetOutputCb(&p->rc, p1);
248232		-
248233		- for( /* no-op */ ;
248234		- fts5MultiIterEof(p, p1)==0;
248235		- fts5MultiIterNext2(p, p1, &bNewTerm)
248236		- ){
248237		- Fts5SegIter *pSeg = &p1->aSeg[ p1->aFirst[1].iFirst ];
248238		- int nTerm = pSeg->term.n;
248239		- const u8 *pTerm = pSeg->term.p;
248240		- p1->xSetOutputs(p1, pSeg);
248241		-
248242		- assert_nc( memcmp(pToken, pTerm, MIN(nToken, nTerm))<=0 );
248243		- if( bNewTerm ){
248244		- if( nTerm<nToken \|\| memcmp(pToken, pTerm, nToken) ) break;
248245		- }
248246		-
248247		- if( p1->base.nData==0 ) continue;
248248		- if( p1->base.iRowid<=iLastRowid && doclist.n>0 ){
248249		- for(i=0; p->rc==SQLITE_OK && doclist.n; i++){
248250		- int i1 = i*nMerge;
248251		- int iStore;
248252		- assert( i1+nMerge<=nBuf );
248253		- for(iStore=i1; iStore<i1+nMerge; iStore++){
248254		- if( aBuf[iStore].n==0 ){
248255		- fts5BufferSwap(&doclist, &aBuf[iStore]);
248256		- fts5BufferZero(&doclist);
248257		- break;
248258		- }
248259		- }
248260		- if( iStore==i1+nMerge ){
248261		- xMerge(p, &doclist, nMerge, &aBuf[i1]);
248262		- for(iStore=i1; iStore<i1+nMerge; iStore++){
248263		- fts5BufferZero(&aBuf[iStore]);
248264		- }
248265		- }
248266		- }
248267		- iLastRowid = 0;
248268		- }
248269		-
248270		- xAppend(p, (u64)p1->base.iRowid-(u64)iLastRowid, p1, &doclist);
248271		- iLastRowid = p1->base.iRowid;
248272		- }
248273		-
248274		- assert( (nBuf%nMerge)==0 );
248275		- for(i=0; i<nBuf; i+=nMerge){
	249059	+ fts5VisitEntries(p, pColset, pToken, nToken, 1, prefixIterSetupCb, pCtx);
	249060	+
	249061	+ assert( (s.nBuf%s.nMerge)==0 );
	249062	+ for(i=0; i<s.nBuf; i+=s.nMerge){
248276	249063	int iFree;
248277	249064	if( p->rc==SQLITE_OK ){
248278		- xMerge(p, &doclist, nMerge, &aBuf[i]);
	249065	+ s.xMerge(p, &s.doclist, s.nMerge, &s.aBuf[i]);
248279	249066	}
248280		- for(iFree=i; iFree<i+nMerge; iFree++){
248281		- fts5BufferFree(&aBuf[iFree]);
	249067	+ for(iFree=i; iFree<i+s.nMerge; iFree++){
	249068	+ fts5BufferFree(&s.aBuf[iFree]);
248282	249069	}
248283	249070	}
248284		- fts5MultiIterFree(p1);
248285	249071
248286		- pData = fts5IdxMalloc(p, sizeof(*pData)+doclist.n+FTS5_DATA_ZERO_PADDING);
	249072	+ pData = fts5IdxMalloc(p, sizeof(*pData)+s.doclist.n+FTS5_DATA_ZERO_PADDING);
248287	249073	if( pData ){
248288	249074	pData->p = (u8*)&pData[1];
248289		- pData->nn = pData->szLeaf = doclist.n;
248290		- if( doclist.n ) memcpy(pData->p, doclist.p, doclist.n);
	249075	+ pData->nn = pData->szLeaf = s.doclist.n;
	249076	+ if( s.doclist.n ) memcpy(pData->p, s.doclist.p, s.doclist.n);
248291	249077	fts5MultiIterNew2(p, pData, bDesc, ppIter);
248292	249078	}
248293		- fts5BufferFree(&doclist);
	249079	+
	249080	+ if( p->rc==SQLITE_OK && s.pTokendata ){
	249081	+ fts5TokendataIterSortMap(p, s2.pT);
	249082	+ (*ppIter)->pTokenDataIter = s2.pT;
	249083	+ s2.pT = 0;
	249084	+ }
248294	249085	}
248295	249086
	249087	+ fts5TokendataIterDelete(s2.pT);
	249088	+ fts5BufferFree(&s.doclist);
248296	249089	fts5StructureRelease(pStruct);
248297		- sqlite3_free(aBuf);
	249090	+ sqlite3_free(s.aBuf);
248298	249091	}
248299	249092
248300	249093
248301	249094	/*
248302	249095	** Indicate that all subsequent calls to sqlite3Fts5IndexWrite() pertain
		@@ -248546,42 +249339,10 @@
248546	249339	static void fts5SegIterSetEOF(Fts5SegIter *pSeg){
248547	249340	fts5DataRelease(pSeg->pLeaf);
248548	249341	pSeg->pLeaf = 0;
248549	249342	}
248550	249343
248551		-/*
248552		-** Usually, a tokendata=1 iterator (struct Fts5TokenDataIter) accumulates an
248553		-** array of these for each row it visits. Or, for an iterator used by an
248554		-** "ORDER BY rank" query, it accumulates an array of these for the entire
248555		-** query.
248556		-**
248557		-** Each instance in the array indicates the iterator (and therefore term)
248558		-** associated with position iPos of rowid iRowid. This is used by the
248559		-** xInstToken() API.
248560		-*/
248561		-struct Fts5TokenDataMap {
248562		- i64 iRowid; /* Row this token is located in */
248563		- i64 iPos; /* Position of token */
248564		- int iIter; /* Iterator token was read from */
248565		-};
248566		-
248567		-/*
248568		-** An object used to supplement Fts5Iter for tokendata=1 iterators.
248569		-*/
248570		-struct Fts5TokenDataIter {
248571		- int nIter;
248572		- int nIterAlloc;
248573		-
248574		- int nMap;
248575		- int nMapAlloc;
248576		- Fts5TokenDataMap *aMap;
248577		-
248578		- Fts5PoslistReader *aPoslistReader;
248579		- int *aPoslistToIter;
248580		- Fts5Iter *apIter[1];
248581		-};
248582		-
248583	249344	/*
248584	249345	** This function appends iterator pAppend to Fts5TokenDataIter pIn and
248585	249346	** returns the result.
248586	249347	*/
248587	249348	static Fts5TokenDataIter *fts5AppendTokendataIter(
		@@ -248614,58 +249375,10 @@
248614	249375	assert( pRet==0 \|\| pRet->nIter<=pRet->nIterAlloc );
248615	249376
248616	249377	return pRet;
248617	249378	}
248618	249379
248619		-/*
248620		-** Delete an Fts5TokenDataIter structure and its contents.
248621		-*/
248622		-static void fts5TokendataIterDelete(Fts5TokenDataIter *pSet){
248623		- if( pSet ){
248624		- int ii;
248625		- for(ii=0; ii<pSet->nIter; ii++){
248626		- fts5MultiIterFree(pSet->apIter[ii]);
248627		- }
248628		- sqlite3_free(pSet->aPoslistReader);
248629		- sqlite3_free(pSet->aMap);
248630		- sqlite3_free(pSet);
248631		- }
248632		-}
248633		-
248634		-/*
248635		-** Append a mapping to the token-map belonging to object pT.
248636		-*/
248637		-static void fts5TokendataIterAppendMap(
248638		- Fts5Index *p,
248639		- Fts5TokenDataIter *pT,
248640		- int iIter,
248641		- i64 iRowid,
248642		- i64 iPos
248643		-){
248644		- if( p->rc==SQLITE_OK ){
248645		- if( pT->nMap==pT->nMapAlloc ){
248646		- int nNew = pT->nMapAlloc ? pT->nMapAlloc*2 : 64;
248647		- int nByte = nNew * sizeof(Fts5TokenDataMap);
248648		- Fts5TokenDataMap *aNew;
248649		-
248650		- aNew = (Fts5TokenDataMap*)sqlite3_realloc(pT->aMap, nByte);
248651		- if( aNew==0 ){
248652		- p->rc = SQLITE_NOMEM;
248653		- return;
248654		- }
248655		-
248656		- pT->aMap = aNew;
248657		- pT->nMapAlloc = nNew;
248658		- }
248659		-
248660		- pT->aMap[pT->nMap].iRowid = iRowid;
248661		- pT->aMap[pT->nMap].iPos = iPos;
248662		- pT->aMap[pT->nMap].iIter = iIter;
248663		- pT->nMap++;
248664		- }
248665		-}
248666		-
248667	249380	/*
248668	249381	** The iterator passed as the only argument must be a tokendata=1 iterator
248669	249382	** (pIter->pTokenDataIter!=0). This function sets the iterator output
248670	249383	** variables (pIter->base.*) according to the contents of the current
248671	249384	** row.
		@@ -248702,11 +249415,11 @@
248702	249415	int eDetail = pIter->pIndex->pConfig->eDetail;
248703	249416	pIter->base.bEof = 0;
248704	249417	pIter->base.iRowid = iRowid;
248705	249418
248706	249419	if( nHit==1 && eDetail==FTS5_DETAIL_FULL ){
248707		- fts5TokendataIterAppendMap(pIter->pIndex, pT, iMin, iRowid, -1);
	249420	+ fts5TokendataIterAppendMap(pIter->pIndex, pT, iMin, 0, iRowid, -1);
248708	249421	}else
248709	249422	if( nHit>1 && eDetail!=FTS5_DETAIL_NONE ){
248710	249423	int nReader = 0;
248711	249424	int nByte = 0;
248712	249425	i64 iPrev = 0;
		@@ -248955,10 +249668,11 @@
248955	249668
248956	249669	if( p->rc==SQLITE_OK ){
248957	249670	pRet = fts5MultiIterAlloc(p, 0);
248958	249671	}
248959	249672	if( pRet ){
	249673	+ pRet->nSeg = 0;
248960	249674	pRet->pTokenDataIter = pSet;
248961	249675	if( pSet ){
248962	249676	fts5IterSetOutputsTokendata(pRet);
248963	249677	}else{
248964	249678	pRet->base.bEof = 1;
		@@ -248969,11 +249683,10 @@
248969	249683
248970	249684	fts5StructureRelease(pStruct);
248971	249685	fts5BufferFree(&bSeek);
248972	249686	return pRet;
248973	249687	}
248974		-
248975	249688
248976	249689	/*
248977	249690	** Open a new iterator to iterate though all rowid that match the
248978	249691	** specified token or token prefix.
248979	249692	*/
		@@ -248995,10 +249708,15 @@
248995	249708	int iIdx = 0; /* Index to search */
248996	249709	int iPrefixIdx = 0; /* +1 prefix index */
248997	249710	int bTokendata = pConfig->bTokendata;
248998	249711	if( nToken>0 ) memcpy(&buf.p[1], pToken, nToken);
248999	249712
	249713	+ /* The NOTOKENDATA flag is set when each token in a tokendata=1 table
	249714	+ ** should be treated individually, instead of merging all those with
	249715	+ ** a common prefix into a single entry. This is used, for example, by
	249716	+ ** queries performed as part of an integrity-check, or by the fts5vocab
	249717	+ ** module. */
249000	249718	if( flags & (FTS5INDEX_QUERY_NOTOKENDATA\|FTS5INDEX_QUERY_SCAN) ){
249001	249719	bTokendata = 0;
249002	249720	}
249003	249721
249004	249722	/* Figure out which index to search and set iIdx accordingly. If this
		@@ -249025,11 +249743,11 @@
249025	249743	if( nIdxChar==nChar+1 ) iPrefixIdx = iIdx;
249026	249744	}
249027	249745	}
249028	249746
249029	249747	if( bTokendata && iIdx==0 ){
249030		- buf.p[0] = '0';
	249748	+ buf.p[0] = FTS5_MAIN_PREFIX;
249031	249749	pRet = fts5SetupTokendataIter(p, buf.p, nToken+1, pColset);
249032	249750	}else if( iIdx<=pConfig->nPrefix ){
249033	249751	/* Straight index lookup */
249034	249752	Fts5Structure *pStruct = fts5StructureRead(p);
249035	249753	buf.p[0] = (u8)(FTS5_MAIN_PREFIX + iIdx);
		@@ -249038,11 +249756,11 @@
249038	249756	pColset, buf.p, nToken+1, -1, 0, &pRet
249039	249757	);
249040	249758	fts5StructureRelease(pStruct);
249041	249759	}
249042	249760	}else{
249043		- /* Scan multiple terms in the main index */
	249761	+ /* Scan multiple terms in the main index for a prefix query. */
249044	249762	int bDesc = (flags & FTS5INDEX_QUERY_DESC)!=0;
249045	249763	fts5SetupPrefixIter(p, bDesc, iPrefixIdx, buf.p, nToken+1, pColset,&pRet);
249046	249764	if( pRet==0 ){
249047	249765	assert( p->rc!=SQLITE_OK );
249048	249766	}else{
		@@ -249074,11 +249792,12 @@
249074	249792	** Move to the next matching rowid.
249075	249793	*/
249076	249794	static int sqlite3Fts5IterNext(Fts5IndexIter *pIndexIter){
249077	249795	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249078	249796	assert( pIter->pIndex->rc==SQLITE_OK );
249079		- if( pIter->pTokenDataIter ){
	249797	+ if( pIter->nSeg==0 ){
	249798	+ assert( pIter->pTokenDataIter );
249080	249799	fts5TokendataIterNext(pIter, 0, 0);
249081	249800	}else{
249082	249801	fts5MultiIterNext(pIter->pIndex, pIter, 0, 0);
249083	249802	}
249084	249803	return fts5IndexReturn(pIter->pIndex);
		@@ -249111,11 +249830,12 @@
249111	249830	** definition of "at or after" depends on whether this iterator iterates
249112	249831	** in ascending or descending rowid order.
249113	249832	*/
249114	249833	static int sqlite3Fts5IterNextFrom(Fts5IndexIter *pIndexIter, i64 iMatch){
249115	249834	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249116		- if( pIter->pTokenDataIter ){
	249835	+ if( pIter->nSeg==0 ){
	249836	+ assert( pIter->pTokenDataIter );
249117	249837	fts5TokendataIterNext(pIter, 1, iMatch);
249118	249838	}else{
249119	249839	fts5MultiIterNextFrom(pIter->pIndex, pIter, iMatch);
249120	249840	}
249121	249841	return fts5IndexReturn(pIter->pIndex);
		@@ -249129,32 +249849,87 @@
249129	249849	const char z = (const char)fts5MultiIterTerm((Fts5Iter*)pIndexIter, &n);
249130	249850	assert_nc( z \|\| n<=1 );
249131	249851	*pn = n-1;
249132	249852	return (z ? &z[1] : 0);
249133	249853	}
	249854	+
	249855	+/*
	249856	+** pIter is a prefix query. This function populates pIter->pTokenDataIter
	249857	+** with an Fts5TokenDataIter object containing mappings for all rows
	249858	+** matched by the query.
	249859	+*/
	249860	+static int fts5SetupPrefixIterTokendata(
	249861	+ Fts5Iter *pIter,
	249862	+ const char pToken, / Token prefix to search for */
	249863	+ int nToken /* Size of pToken in bytes */
	249864	+){
	249865	+ Fts5Index *p = pIter->pIndex;
	249866	+ Fts5Buffer token = {0, 0, 0};
	249867	+ TokendataSetupCtx ctx;
	249868	+
	249869	+ memset(&ctx, 0, sizeof(ctx));
	249870	+
	249871	+ fts5BufferGrow(&p->rc, &token, nToken+1);
	249872	+ ctx.pT = (Fts5TokenDataIter)sqlite3Fts5MallocZero(&p->rc, sizeof(ctx.pT));
	249873	+
	249874	+ if( p->rc==SQLITE_OK ){
	249875	+
	249876	+ /* Fill in the token prefix to search for */
	249877	+ token.p[0] = FTS5_MAIN_PREFIX;
	249878	+ memcpy(&token.p[1], pToken, nToken);
	249879	+ token.n = nToken+1;
	249880	+
	249881	+ fts5VisitEntries(
	249882	+ p, 0, token.p, token.n, 1, prefixIterSetupTokendataCb, (void*)&ctx
	249883	+ );
	249884	+
	249885	+ fts5TokendataIterSortMap(p, ctx.pT);
	249886	+ }
	249887	+
	249888	+ if( p->rc==SQLITE_OK ){
	249889	+ pIter->pTokenDataIter = ctx.pT;
	249890	+ }else{
	249891	+ fts5TokendataIterDelete(ctx.pT);
	249892	+ }
	249893	+ fts5BufferFree(&token);
	249894	+
	249895	+ return fts5IndexReturn(p);
	249896	+}
249134	249897
249135	249898	/*
249136	249899	** This is used by xInstToken() to access the token at offset iOff, column
249137	249900	** iCol of row iRowid. The token is returned via output variables *ppOut
249138	249901	** and *pnOut. The iterator passed as the first argument must be a tokendata=1
249139	249902	** iterator (pIter->pTokenDataIter!=0).
	249903	+**
	249904	+** pToken/nToken:
249140	249905	*/
249141	249906	static int sqlite3Fts5IterToken(
249142	249907	Fts5IndexIter *pIndexIter,
	249908	+ const char *pToken, int nToken,
249143	249909	i64 iRowid,
249144	249910	int iCol,
249145	249911	int iOff,
249146	249912	const char *ppOut, int pnOut
249147	249913	){
249148	249914	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249149	249915	Fts5TokenDataIter *pT = pIter->pTokenDataIter;
249150		- Fts5TokenDataMap *aMap = pT->aMap;
249151	249916	i64 iPos = (((i64)iCol)<<32) + iOff;
249152		-
	249917	+ Fts5TokenDataMap *aMap = 0;
249153	249918	int i1 = 0;
249154		- int i2 = pT->nMap;
	249919	+ int i2 = 0;
249155	249920	int iTest = 0;
	249921	+
	249922	+ assert( pT \|\| (pToken && pIter->nSeg>0) );
	249923	+ if( pT==0 ){
	249924	+ int rc = fts5SetupPrefixIterTokendata(pIter, pToken, nToken);
	249925	+ if( rc!=SQLITE_OK ) return rc;
	249926	+ pT = pIter->pTokenDataIter;
	249927	+ }
	249928	+
	249929	+ i2 = pT->nMap;
	249930	+ aMap = pT->aMap;
249156	249931
249157	249932	while( i2>i1 ){
249158	249933	iTest = (i1 + i2) / 2;
249159	249934
249160	249935	if( aMap[iTest].iRowid<iRowid ){
		@@ -249174,13 +249949,19 @@
249174	249949	}
249175	249950	}
249176	249951	}
249177	249952
249178	249953	if( i2>i1 ){
249179		- Fts5Iter *pMap = pT->apIter[aMap[iTest].iIter];
249180		- ppOut = (const char)pMap->aSeg[0].term.p+1;
249181		- *pnOut = pMap->aSeg[0].term.n-1;
	249954	+ if( pIter->nSeg==0 ){
	249955	+ Fts5Iter *pMap = pT->apIter[aMap[iTest].iIter];
	249956	+ ppOut = (const char)pMap->aSeg[0].term.p+1;
	249957	+ *pnOut = pMap->aSeg[0].term.n-1;
	249958	+ }else{
	249959	+ Fts5TokenDataMap *p = &aMap[iTest];
	249960	+ ppOut = (const char)&pT->terms.p[p->iIter];
	249961	+ *pnOut = aMap[iTest].nByte;
	249962	+ }
249182	249963	}
249183	249964
249184	249965	return SQLITE_OK;
249185	249966	}
249186	249967
		@@ -249188,11 +249969,13 @@
249188	249969	** Clear any existing entries from the token-map associated with the
249189	249970	** iterator passed as the only argument.
249190	249971	*/
249191	249972	static void sqlite3Fts5IndexIterClearTokendata(Fts5IndexIter *pIndexIter){
249192	249973	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249193		- if( pIter && pIter->pTokenDataIter ){
	249974	+ if( pIter && pIter->pTokenDataIter
	249975	+ && (pIter->nSeg==0 \|\| pIter->pIndex->pConfig->eDetail!=FTS5_DETAIL_FULL)
	249976	+ ){
249194	249977	pIter->pTokenDataIter->nMap = 0;
249195	249978	}
249196	249979	}
249197	249980
249198	249981	/*
		@@ -249208,21 +249991,33 @@
249208	249991	i64 iRowid, int iCol, int iOff
249209	249992	){
249210	249993	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249211	249994	Fts5TokenDataIter *pT = pIter->pTokenDataIter;
249212	249995	Fts5Index *p = pIter->pIndex;
249213		- int ii;
	249996	+ i64 iPos = (((i64)iCol)<<32) + iOff;
249214	249997
249215	249998	assert( p->pConfig->eDetail!=FTS5_DETAIL_FULL );
249216		- assert( pIter->pTokenDataIter );
249217		-
249218		- for(ii=0; ii<pT->nIter; ii++){
249219		- Fts5Buffer *pTerm = &pT->apIter[ii]->aSeg[0].term;
249220		- if( nToken==pTerm->n-1 && memcmp(pToken, pTerm->p+1, nToken)==0 ) break;
249221		- }
249222		- if( ii<pT->nIter ){
249223		- fts5TokendataIterAppendMap(p, pT, ii, iRowid, (((i64)iCol)<<32) + iOff);
	249999	+ assert( pIter->pTokenDataIter \|\| pIter->nSeg>0 );
	250000	+ if( pIter->nSeg>0 ){
	250001	+ /* This is a prefix term iterator. */
	250002	+ if( pT==0 ){
	250003	+ pT = (Fts5TokenDataIter)sqlite3Fts5MallocZero(&p->rc, sizeof(pT));
	250004	+ pIter->pTokenDataIter = pT;
	250005	+ }
	250006	+ if( pT ){
	250007	+ fts5TokendataIterAppendMap(p, pT, pT->terms.n, nToken, iRowid, iPos);
	250008	+ fts5BufferAppendBlob(&p->rc, &pT->terms, nToken, (const u8*)pToken);
	250009	+ }
	250010	+ }else{
	250011	+ int ii;
	250012	+ for(ii=0; ii<pT->nIter; ii++){
	250013	+ Fts5Buffer *pTerm = &pT->apIter[ii]->aSeg[0].term;
	250014	+ if( nToken==pTerm->n-1 && memcmp(pToken, pTerm->p+1, nToken)==0 ) break;
	250015	+ }
	250016	+ if( ii<pT->nIter ){
	250017	+ fts5TokendataIterAppendMap(p, pT, ii, 0, iRowid, iPos);
	250018	+ }
249224	250019	}
249225	250020	return fts5IndexReturn(p);
249226	250021	}
249227	250022
249228	250023	/*
		@@ -250771,11 +251566,11 @@
250771	251566	return rc;
250772	251567	}
250773	251568
250774	251569	/*
250775	251570	** We must have a single struct=? constraint that will be passed through
250776		-** into the xFilter method. If there is no valid stmt=? constraint,
	251571	+** into the xFilter method. If there is no valid struct=? constraint,
250777	251572	** then return an SQLITE_CONSTRAINT error.
250778	251573	*/
250779	251574	static int fts5structBestIndexMethod(
250780	251575	sqlite3_vtab *tab,
250781	251576	sqlite3_index_info *pIdxInfo
		@@ -251113,12 +251908,22 @@
251113	251908	i64 iNextId; /* Used to allocate unique cursor ids */
251114	251909	Fts5Auxiliary pAux; / First in list of all aux. functions */
251115	251910	Fts5TokenizerModule pTok; / First in list of all tokenizer modules */
251116	251911	Fts5TokenizerModule pDfltTok; / Default tokenizer module */
251117	251912	Fts5Cursor pCsr; / First in list of all open cursors */
	251913	+ u32 aLocaleHdr[4];
251118	251914	};
251119	251915
	251916	+/*
	251917	+** Size of header on fts5_locale() values. And macro to access a buffer
	251918	+** containing a copy of the header from an Fts5Config pointer.
	251919	+*/
	251920	+#define FTS5_LOCALE_HDR_SIZE ((int)sizeof( ((Fts5Global*)0)->aLocaleHdr ))
	251921	+#define FTS5_LOCALE_HDR(pConfig) ((const u8*)(pConfig->pGlobal->aLocaleHdr))
	251922	+
	251923	+#define FTS5_INSTTOKEN_SUBTYPE 73
	251924	+
251120	251925	/*
251121	251926	** Each auxiliary function registered with the FTS5 module is represented
251122	251927	** by an object of the following type. All such objects are stored as part
251123	251928	** of the Fts5Global.pAux list.
251124	251929	*/
		@@ -251277,16 +252082,10 @@
251277	252082	#define FTS5CSR_REQUIRE_POSLIST 0x40
251278	252083
251279	252084	#define BitFlagAllTest(x,y) (((x) & (y))==(y))
251280	252085	#define BitFlagTest(x,y) (((x) & (y))!=0)
251281	252086
251282		-/*
251283		-** The subtype value and header bytes used by fts5_locale().
251284		-*/
251285		-#define FTS5_LOCALE_SUBTYPE ((unsigned int)'L')
251286		-#define FTS5_LOCALE_HEADER "\x00\xE0\xB2\xEB"
251287		-
251288	252087
251289	252088	/*
251290	252089	** Macros to Set(), Clear() and Test() cursor flags.
251291	252090	*/
251292	252091	#define CsrFlagSet(pCsr, flag) ((pCsr)->csrflags \|= (flag))
		@@ -251359,14 +252158,20 @@
251359	252158	#else
251360	252159	# define fts5CheckTransactionState(x,y,z)
251361	252160	#endif
251362	252161
251363	252162	/*
251364		-** Return true if pTab is a contentless table.
	252163	+** Return true if pTab is a contentless table. If parameter bIncludeUnindexed
	252164	+** is true, this includes contentless tables that store UNINDEXED columns
	252165	+** only.
251365	252166	*/
251366		-static int fts5IsContentless(Fts5FullTable *pTab){
251367		- return pTab->p.pConfig->eContent==FTS5_CONTENT_NONE;
	252167	+static int fts5IsContentless(Fts5FullTable *pTab, int bIncludeUnindexed){
	252168	+ int eContent = pTab->p.pConfig->eContent;
	252169	+ return (
	252170	+ eContent==FTS5_CONTENT_NONE
	252171	+ \|\| (bIncludeUnindexed && eContent==FTS5_CONTENT_UNINDEXED)
	252172	+ );
251368	252173	}
251369	252174
251370	252175	/*
251371	252176	** Delete a virtual table handle allocated by fts5InitVtab().
251372	252177	*/
		@@ -251653,10 +252458,11 @@
251653	252458	){
251654	252459	/* A MATCH operator or equivalent */
251655	252460	if( p->usable==0 \|\| iCol<0 ){
251656	252461	/* As there exists an unusable MATCH constraint this is an
251657	252462	** unusable plan. Return SQLITE_CONSTRAINT. */
	252463	+ idxStr[iIdxStr] = 0;
251658	252464	return SQLITE_CONSTRAINT;
251659	252465	}else{
251660	252466	if( iCol==nCol+1 ){
251661	252467	if( bSeenRank ) continue;
251662	252468	idxStr[iIdxStr++] = 'r';
		@@ -252286,11 +253092,11 @@
252286	253092	** Arrange for subsequent calls to sqlite3Fts5Tokenize() to use the locale
252287	253093	** specified by pLocale/nLocale. The buffer indicated by pLocale must remain
252288	253094	** valid until after the final call to sqlite3Fts5Tokenize() that will use
252289	253095	** the locale.
252290	253096	*/
252291		-static void fts5SetLocale(
	253097	+static void sqlite3Fts5SetLocale(
252292	253098	Fts5Config *pConfig,
252293	253099	const char *zLocale,
252294	253100	int nLocale
252295	253101	){
252296	253102	Fts5TokenizerConfig *pT = &pConfig->t;
		@@ -252297,141 +253103,88 @@
252297	253103	pT->pLocale = zLocale;
252298	253104	pT->nLocale = nLocale;
252299	253105	}
252300	253106
252301	253107	/*
252302		-** Clear any locale configured by an earlier call to fts5SetLocale() or
252303		-** sqlite3Fts5ExtractText().
	253108	+** Clear any locale configured by an earlier call to sqlite3Fts5SetLocale().
252304	253109	*/
252305	253110	static void sqlite3Fts5ClearLocale(Fts5Config *pConfig){
252306		- fts5SetLocale(pConfig, 0, 0);
	253111	+ sqlite3Fts5SetLocale(pConfig, 0, 0);
252307	253112	}
252308	253113
252309	253114	/*
252310		-** This function is used to extract utf-8 text from an sqlite3_value. This
252311		-** is usually done in order to tokenize it. For example, when:
252312		-**
252313		-** * a value is written to an fts5 table,
252314		-** * a value is deleted from an FTS5_CONTENT_NORMAL table,
252315		-** * a value containing a query expression is passed to xFilter()
252316		-**
252317		-** and so on.
252318		-**
252319		-** This function handles 2 cases:
252320		-**
252321		-** 1) Ordinary values. The text can be extracted from these using
252322		-** sqlite3_value_text().
252323		-**
252324		-** 2) Combination text/locale blobs created by fts5_locale(). There
252325		-** are several cases for these:
252326		-**
252327		-** * Blobs tagged with FTS5_LOCALE_SUBTYPE.
252328		-** * Blobs read from the content table of a locale=1 external-content
252329		-** table, and
252330		-** * Blobs read from the content table of a locale=1 regular
252331		-** content table.
252332		-**
252333		-** The first two cases above should have the 4 byte FTS5_LOCALE_HEADER
252334		-** header. It is an error if a blob with the subtype or a blob read
252335		-** from the content table of an external content table does not have
252336		-** the required header. A blob read from the content table of a regular
252337		-** locale=1 table does not have the header. This is to save space.
252338		-**
252339		-** If successful, SQLITE_OK is returned and output parameters (*ppText)
252340		-** and (*pnText) are set to point to a buffer containing the extracted utf-8
252341		-** text and its length in bytes, respectively. The buffer is not
252342		-** nul-terminated. It has the same lifetime as the sqlite3_value object
252343		-** from which it is extracted.
252344		-**
252345		-** Parameter bContent must be true if the value was read from an indexed
252346		-** column (i.e. not UNINDEXED) of the on disk content.
252347		-**
252348		-** If pbResetTokenizer is not NULL and if case (2) is used, then
252349		-** fts5SetLocale() is called to ensure subsequent sqlite3Fts5Tokenize() calls
252350		-** use the locale. In this case (*pbResetTokenizer) is set to true before
252351		-** returning, to indicate that the caller must call sqlite3Fts5ClearLocale()
252352		-** to clear the locale after tokenizing the text.
	253115	+** Return true if the value passed as the only argument is an
	253116	+** fts5_locale() value.
252353	253117	*/
252354		-static int sqlite3Fts5ExtractText(
252355		- Fts5Config *pConfig,
252356		- sqlite3_value pVal, / Value to extract text from */
252357		- int bContent, /* True if indexed table content */
252358		- int pbResetTokenizer, / OUT: True if xSetLocale(NULL) required */
252359		- const char *ppText, / OUT: Pointer to text buffer */
252360		- int pnText / OUT: Size of (ppText) in bytes /
252361		-){
252362		- const char *pText = 0;
252363		- int nText = 0;
252364		- int rc = SQLITE_OK;
252365		- int bDecodeBlob = 0;
252366		-
252367		- assert( pbResetTokenizer==0 \|\| *pbResetTokenizer==0 );
252368		- assert( bContent==0 \|\| pConfig->eContent!=FTS5_CONTENT_NONE );
252369		- assert( bContent==0 \|\| sqlite3_value_subtype(pVal)==0 );
252370		-
	253118	+static int sqlite3Fts5IsLocaleValue(Fts5Config pConfig, sqlite3_value pVal){
	253119	+ int ret = 0;
252371	253120	if( sqlite3_value_type(pVal)==SQLITE_BLOB ){
252372		- if( sqlite3_value_subtype(pVal)==FTS5_LOCALE_SUBTYPE
252373		- \|\| (bContent && pConfig->bLocale)
252374		- ){
252375		- bDecodeBlob = 1;
252376		- }
252377		- }
252378		-
252379		- if( bDecodeBlob ){
252380		- const int SZHDR = sizeof(FTS5_LOCALE_HEADER)-1;
	253121	+ /* Call sqlite3_value_bytes() after sqlite3_value_blob() in this case.
	253122	+ ** If the blob was created using zeroblob(), then sqlite3_value_blob()
	253123	+ ** may call malloc(). If this malloc() fails, then the values returned
	253124	+ ** by both value_blob() and value_bytes() will be 0. If value_bytes() were
	253125	+ ** called first, then the NULL pointer returned by value_blob() might
	253126	+ ** be dereferenced. */
252381	253127	const u8 *pBlob = sqlite3_value_blob(pVal);
252382	253128	int nBlob = sqlite3_value_bytes(pVal);
252383		-
252384		- /* Unless this blob was read from the %_content table of an
252385		- ** FTS5_CONTENT_NORMAL table, it should have the 4 byte fts5_locale()
252386		- ** header. Check for this. If it is not found, return an error. */
252387		- if( (!bContent \|\| pConfig->eContent!=FTS5_CONTENT_NORMAL) ){
252388		- if( nBlob<SZHDR \|\| memcmp(FTS5_LOCALE_HEADER, pBlob, SZHDR) ){
252389		- rc = SQLITE_ERROR;
252390		- }else{
252391		- pBlob += 4;
252392		- nBlob -= 4;
252393		- }
252394		- }
252395		-
252396		- if( rc==SQLITE_OK ){
252397		- int nLocale = 0;
252398		-
252399		- for(nLocale=0; nLocale<nBlob; nLocale++){
252400		- if( pBlob[nLocale]==0x00 ) break;
252401		- }
252402		- if( nLocale==nBlob \|\| nLocale==0 ){
252403		- rc = SQLITE_ERROR;
252404		- }else{
252405		- pText = (const char*)&pBlob[nLocale+1];
252406		- nText = nBlob-nLocale-1;
252407		-
252408		- if( pbResetTokenizer ){
252409		- fts5SetLocale(pConfig, (const char*)pBlob, nLocale);
252410		- *pbResetTokenizer = 1;
252411		- }
252412		- }
252413		- }
252414		-
252415		- }else{
252416		- pText = (const char*)sqlite3_value_text(pVal);
252417		- nText = sqlite3_value_bytes(pVal);
252418		- }
252419		-
252420		- *ppText = pText;
252421		- *pnText = nText;
252422		- return rc;
	253129	+ if( nBlob>FTS5_LOCALE_HDR_SIZE
	253130	+ && 0==memcmp(pBlob, FTS5_LOCALE_HDR(pConfig), FTS5_LOCALE_HDR_SIZE)
	253131	+ ){
	253132	+ ret = 1;
	253133	+ }
	253134	+ }
	253135	+ return ret;
	253136	+}
	253137	+
	253138	+/*
	253139	+** Value pVal is guaranteed to be an fts5_locale() value, according to
	253140	+** sqlite3Fts5IsLocaleValue(). This function extracts the text and locale
	253141	+** from the value and returns them separately.
	253142	+**
	253143	+** If successful, SQLITE_OK is returned and (ppText) and (ppLoc) set
	253144	+** to point to buffers containing the text and locale, as utf-8,
	253145	+** respectively. In this case output parameters (pnText) and (pnLoc) are
	253146	+** set to the sizes in bytes of these two buffers.
	253147	+**
	253148	+** Or, if an error occurs, then an SQLite error code is returned. The final
	253149	+** value of the four output parameters is undefined in this case.
	253150	+*/
	253151	+static int sqlite3Fts5DecodeLocaleValue(
	253152	+ sqlite3_value *pVal,
	253153	+ const char **ppText,
	253154	+ int *pnText,
	253155	+ const char **ppLoc,
	253156	+ int *pnLoc
	253157	+){
	253158	+ const char *p = sqlite3_value_blob(pVal);
	253159	+ int n = sqlite3_value_bytes(pVal);
	253160	+ int nLoc = 0;
	253161	+
	253162	+ assert( sqlite3_value_type(pVal)==SQLITE_BLOB );
	253163	+ assert( n>FTS5_LOCALE_HDR_SIZE );
	253164	+
	253165	+ for(nLoc=FTS5_LOCALE_HDR_SIZE; p[nLoc]; nLoc++){
	253166	+ if( nLoc==(n-1) ){
	253167	+ return SQLITE_MISMATCH;
	253168	+ }
	253169	+ }
	253170	+ *ppLoc = &p[FTS5_LOCALE_HDR_SIZE];
	253171	+ *pnLoc = nLoc - FTS5_LOCALE_HDR_SIZE;
	253172	+
	253173	+ *ppText = &p[nLoc+1];
	253174	+ *pnText = n - nLoc - 1;
	253175	+ return SQLITE_OK;
252423	253176	}
252424	253177
252425	253178	/*
252426	253179	** Argument pVal is the text of a full-text search expression. It may or
252427	253180	** may not have been wrapped by fts5_locale(). This function extracts
252428	253181	** the text of the expression, and sets output variable (*pzText) to
252429	253182	** point to a nul-terminated buffer containing the expression.
252430	253183	**
252431		-** If pVal was an fts5_locale() value, then fts5SetLocale() is called to
252432		-** set the tokenizer to use the specified locale.
	253184	+** If pVal was an fts5_locale() value, then sqlite3Fts5SetLocale() is called
	253185	+** to set the tokenizer to use the specified locale.
252433	253186	**
252434	253187	** If output variable (*pbFreeAndReset) is set to true, then the caller
252435	253188	** is required to (a) call sqlite3Fts5ClearLocale() to reset the tokenizer
252436	253189	** locale, and (b) call sqlite3_free() to free (*pzText).
252437	253190	*/
		@@ -252439,28 +253192,26 @@
252439	253192	Fts5Config pConfig, / Fts5 configuration */
252440	253193	sqlite3_value pVal, / Value to extract expression text from */
252441	253194	char *pzText, / OUT: nul-terminated buffer of text */
252442	253195	int pbFreeAndReset / OUT: Free (pzText) and clear locale /
252443	253196	){
252444		- const char *zText = 0;
252445		- int nText = 0;
252446	253197	int rc = SQLITE_OK;
252447		- int bReset = 0;
252448		-
252449		- *pbFreeAndReset = 0;
252450		- rc = sqlite3Fts5ExtractText(pConfig, pVal, 0, &bReset, &zText, &nText);
252451		- if( rc==SQLITE_OK ){
252452		- if( bReset ){
252453		- pzText = sqlite3Fts5Mprintf(&rc, "%.s", nText, zText);
252454		- if( rc!=SQLITE_OK ){
252455		- sqlite3Fts5ClearLocale(pConfig);
252456		- }else{
252457		- *pbFreeAndReset = 1;
252458		- }
252459		- }else{
252460		- pzText = (char)zText;
252461		- }
	253198	+
	253199	+ if( sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
	253200	+ const char *pText = 0;
	253201	+ int nText = 0;
	253202	+ const char *pLoc = 0;
	253203	+ int nLoc = 0;
	253204	+ rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
	253205	+ pzText = sqlite3Fts5Mprintf(&rc, "%.s", nText, pText);
	253206	+ if( rc==SQLITE_OK ){
	253207	+ sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
	253208	+ }
	253209	+ *pbFreeAndReset = 1;
	253210	+ }else{
	253211	+ pzText = (char)sqlite3_value_text(pVal);
	253212	+ *pbFreeAndReset = 0;
252462	253213	}
252463	253214
252464	253215	return rc;
252465	253216	}
252466	253217
		@@ -252493,10 +253244,11 @@
252493	253244	sqlite3_value pRowidEq = 0; / rowid = ? expression (or NULL) */
252494	253245	sqlite3_value pRowidLe = 0; / rowid <= ? expression (or NULL) */
252495	253246	sqlite3_value pRowidGe = 0; / rowid >= ? expression (or NULL) */
252496	253247	int iCol; /* Column on LHS of MATCH operator */
252497	253248	char **pzErrmsg = pConfig->pzErrmsg;
	253249	+ int bPrefixInsttoken = pConfig->bPrefixInsttoken;
252498	253250	int i;
252499	253251	int iIdxStr = 0;
252500	253252	Fts5Expr *pExpr = 0;
252501	253253
252502	253254	assert( pConfig->bLock==0 );
		@@ -252528,10 +253280,13 @@
252528	253280	int bInternal = 0;
252529	253281
252530	253282	rc = fts5ExtractExprText(pConfig, apVal[i], &zText, &bFreeAndReset);
252531	253283	if( rc!=SQLITE_OK ) goto filter_out;
252532	253284	if( zText==0 ) zText = "";
	253285	+ if( sqlite3_value_subtype(apVal[i])==FTS5_INSTTOKEN_SUBTYPE ){
	253286	+ pConfig->bPrefixInsttoken = 1;
	253287	+ }
252533	253288
252534	253289	iCol = 0;
252535	253290	do{
252536	253291	iCol = iCol*10 + (idxStr[iIdxStr]-'0');
252537	253292	iIdxStr++;
		@@ -252668,10 +253423,11 @@
252668	253423	}
252669	253424
252670	253425	filter_out:
252671	253426	sqlite3Fts5ExprFree(pExpr);
252672	253427	pConfig->pzErrmsg = pzErrmsg;
	253428	+ pConfig->bPrefixInsttoken = bPrefixInsttoken;
252673	253429	return rc;
252674	253430	}
252675	253431
252676	253432	/*
252677	253433	** This is the xEof method of the virtual table. SQLite calls this
		@@ -252808,11 +253564,11 @@
252808	253564	}else{
252809	253565	rc = sqlite3Fts5StorageDeleteAll(pTab->pStorage);
252810	253566	}
252811	253567	bLoadConfig = 1;
252812	253568	}else if( 0==sqlite3_stricmp("rebuild", zCmd) ){
252813		- if( pConfig->eContent==FTS5_CONTENT_NONE ){
	253569	+ if( fts5IsContentless(pTab, 1) ){
252814	253570	fts5SetVtabError(pTab,
252815	253571	"'rebuild' may not be used with a contentless fts5 table"
252816	253572	);
252817	253573	rc = SQLITE_ERROR;
252818	253574	}else{
		@@ -252877,17 +253633,78 @@
252877	253633	sqlite3_value **apVal,
252878	253634	i64 *piRowid
252879	253635	){
252880	253636	int rc = *pRc;
252881	253637	if( rc==SQLITE_OK ){
252882		- rc = sqlite3Fts5StorageContentInsert(pTab->pStorage, apVal, piRowid);
	253638	+ rc = sqlite3Fts5StorageContentInsert(pTab->pStorage, 0, apVal, piRowid);
252883	253639	}
252884	253640	if( rc==SQLITE_OK ){
252885	253641	rc = sqlite3Fts5StorageIndexInsert(pTab->pStorage, apVal, *piRowid);
252886	253642	}
252887	253643	*pRc = rc;
252888	253644	}
	253645	+
	253646	+/*
	253647	+**
	253648	+** This function is called when the user attempts an UPDATE on a contentless
	253649	+** table. Parameter bRowidModified is true if the UPDATE statement modifies
	253650	+** the rowid value. Parameter apVal[] contains the new values for each user
	253651	+** defined column of the fts5 table. pConfig is the configuration object of the
	253652	+** table being updated (guaranteed to be contentless). The contentless_delete=1
	253653	+** and contentless_unindexed=1 options may or may not be set.
	253654	+**
	253655	+** This function returns SQLITE_OK if the UPDATE can go ahead, or an SQLite
	253656	+** error code if it cannot. In this case an error message is also loaded into
	253657	+** pConfig. Output parameter (*pbContent) is set to true if the caller should
	253658	+** update the %_content table only - not the FTS index or any other shadow
	253659	+** table. This occurs when an UPDATE modifies only UNINDEXED columns of the
	253660	+** table.
	253661	+**
	253662	+** An UPDATE may proceed if:
	253663	+**
	253664	+** * The only columns modified are UNINDEXED columns, or
	253665	+**
	253666	+** * The contentless_delete=1 option was specified and all of the indexed
	253667	+** columns (not a subset) have been modified.
	253668	+*/
	253669	+static int fts5ContentlessUpdate(
	253670	+ Fts5Config *pConfig,
	253671	+ sqlite3_value **apVal,
	253672	+ int bRowidModified,
	253673	+ int *pbContent
	253674	+){
	253675	+ int ii;
	253676	+ int bSeenIndex = 0; /* Have seen modified indexed column */
	253677	+ int bSeenIndexNC = 0; /* Have seen unmodified indexed column */
	253678	+ int rc = SQLITE_OK;
	253679	+
	253680	+ for(ii=0; ii<pConfig->nCol; ii++){
	253681	+ if( pConfig->abUnindexed[ii]==0 ){
	253682	+ if( sqlite3_value_nochange(apVal[ii]) ){
	253683	+ bSeenIndexNC++;
	253684	+ }else{
	253685	+ bSeenIndex++;
	253686	+ }
	253687	+ }
	253688	+ }
	253689	+
	253690	+ if( bSeenIndex==0 && bRowidModified==0 ){
	253691	+ *pbContent = 1;
	253692	+ }else{
	253693	+ if( bSeenIndexNC \|\| pConfig->bContentlessDelete==0 ){
	253694	+ rc = SQLITE_ERROR;
	253695	+ sqlite3Fts5ConfigErrmsg(pConfig,
	253696	+ (pConfig->bContentlessDelete ?
	253697	+ "%s a subset of columns on fts5 contentless-delete table: %s" :
	253698	+ "%s contentless fts5 table: %s")
	253699	+ , "cannot UPDATE", pConfig->zName
	253700	+ );
	253701	+ }
	253702	+ }
	253703	+
	253704	+ return rc;
	253705	+}
252889	253706
252890	253707	/*
252891	253708	** This function is the implementation of the xUpdate callback used by
252892	253709	** FTS3 virtual tables. It is invoked by SQLite each time a row is to be
252893	253710	** inserted, updated or deleted.
		@@ -252971,48 +253788,38 @@
252971	253788	}
252972	253789
252973	253790	assert( eType0==SQLITE_INTEGER \|\| eType0==SQLITE_NULL );
252974	253791	assert( nArg!=1 \|\| eType0==SQLITE_INTEGER );
252975	253792
252976		- /* Filter out attempts to run UPDATE or DELETE on contentless tables.
252977		- ** This is not suported. Except - they are both supported if the CREATE
252978		- ** VIRTUAL TABLE statement contained "contentless_delete=1". */
252979		- if( eType0==SQLITE_INTEGER
252980		- && pConfig->eContent==FTS5_CONTENT_NONE
252981		- && pConfig->bContentlessDelete==0
252982		- ){
252983		- pTab->p.base.zErrMsg = sqlite3_mprintf(
252984		- "cannot %s contentless fts5 table: %s",
252985		- (nArg>1 ? "UPDATE" : "DELETE from"), pConfig->zName
252986		- );
252987		- rc = SQLITE_ERROR;
252988		- }
252989		-
252990	253793	/* DELETE */
252991		- else if( nArg==1 ){
252992		- i64 iDel = sqlite3_value_int64(apVal[0]); /* Rowid to delete */
252993		- rc = sqlite3Fts5StorageDelete(pTab->pStorage, iDel, 0, 0);
252994		- bUpdateOrDelete = 1;
	253794	+ if( nArg==1 ){
	253795	+ /* It is only possible to DELETE from a contentless table if the
	253796	+ ** contentless_delete=1 flag is set. */
	253797	+ if( fts5IsContentless(pTab, 1) && pConfig->bContentlessDelete==0 ){
	253798	+ fts5SetVtabError(pTab,
	253799	+ "cannot DELETE from contentless fts5 table: %s", pConfig->zName
	253800	+ );
	253801	+ rc = SQLITE_ERROR;
	253802	+ }else{
	253803	+ i64 iDel = sqlite3_value_int64(apVal[0]); /* Rowid to delete */
	253804	+ rc = sqlite3Fts5StorageDelete(pTab->pStorage, iDel, 0, 0);
	253805	+ bUpdateOrDelete = 1;
	253806	+ }
252995	253807	}
252996	253808
252997	253809	/* INSERT or UPDATE */
252998	253810	else{
252999	253811	int eType1 = sqlite3_value_numeric_type(apVal[1]);
253000	253812
253001		- /* Ensure that no fts5_locale() values are written to locale=0 tables.
253002		- ** And that no blobs except fts5_locale() blobs are written to indexed
253003		- ** (i.e. not UNINDEXED) columns of locale=1 tables. */
253004		- int ii;
253005		- for(ii=0; ii<pConfig->nCol; ii++){
253006		- if( sqlite3_value_type(apVal[ii+2])==SQLITE_BLOB ){
253007		- int bSub = (sqlite3_value_subtype(apVal[ii+2])==FTS5_LOCALE_SUBTYPE);
253008		- if( (pConfig->bLocale && !bSub && pConfig->abUnindexed[ii]==0)
253009		- \|\| (pConfig->bLocale==0 && bSub)
253010		- ){
253011		- if( pConfig->bLocale==0 ){
253012		- fts5SetVtabError(pTab, "fts5_locale() requires locale=1");
253013		- }
	253813	+ /* It is an error to write an fts5_locale() value to a table without
	253814	+ ** the locale=1 option. */
	253815	+ if( pConfig->bLocale==0 ){
	253816	+ int ii;
	253817	+ for(ii=0; ii<pConfig->nCol; ii++){
	253818	+ sqlite3_value *pVal = apVal[ii+2];
	253819	+ if( sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
	253820	+ fts5SetVtabError(pTab, "fts5_locale() requires locale=1");
253014	253821	rc = SQLITE_MISMATCH;
253015	253822	goto update_out;
253016	253823	}
253017	253824	}
253018	253825	}
		@@ -253028,39 +253835,59 @@
253028	253835	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253029	253836	}
253030	253837
253031	253838	/* UPDATE */
253032	253839	else{
253033		- i64 iOld = sqlite3_value_int64(apVal[0]); /* Old rowid */
253034		- i64 iNew = sqlite3_value_int64(apVal[1]); /* New rowid */
253035		- if( eType1!=SQLITE_INTEGER ){
253036		- rc = SQLITE_MISMATCH;
253037		- }else if( iOld!=iNew ){
253038		- if( eConflict==SQLITE_REPLACE ){
253039		- rc = sqlite3Fts5StorageDelete(pTab->pStorage, iOld, 0, 1);
253040		- if( rc==SQLITE_OK ){
253041		- rc = sqlite3Fts5StorageDelete(pTab->pStorage, iNew, 0, 0);
253042		- }
253043		- fts5StorageInsert(&rc, pTab, apVal, pRowid);
253044		- }else{
253045		- rc = sqlite3Fts5StorageFindDeleteRow(pTab->pStorage, iOld);
253046		- if( rc==SQLITE_OK ){
253047		- rc = sqlite3Fts5StorageContentInsert(pTab->pStorage,apVal,pRowid);
253048		- }
253049		- if( rc==SQLITE_OK ){
253050		- rc = sqlite3Fts5StorageDelete(pTab->pStorage, iOld, 0, 1);
253051		- }
253052		- if( rc==SQLITE_OK ){
253053		- rc = sqlite3Fts5StorageIndexInsert(pTab->pStorage, apVal,*pRowid);
253054		- }
253055		- }
253056		- }else{
253057		- rc = sqlite3Fts5StorageDelete(pTab->pStorage, iOld, 0, 1);
253058		- fts5StorageInsert(&rc, pTab, apVal, pRowid);
253059		- }
253060		- bUpdateOrDelete = 1;
253061		- sqlite3Fts5StorageReleaseDeleteRow(pTab->pStorage);
	253840	+ Fts5Storage *pStorage = pTab->pStorage;
	253841	+ i64 iOld = sqlite3_value_int64(apVal[0]); /* Old rowid */
	253842	+ i64 iNew = sqlite3_value_int64(apVal[1]); /* New rowid */
	253843	+ int bContent = 0; /* Content only update */
	253844	+
	253845	+ /* If this is a contentless table (including contentless_unindexed=1
	253846	+ ** tables), check if the UPDATE may proceed. */
	253847	+ if( fts5IsContentless(pTab, 1) ){
	253848	+ rc = fts5ContentlessUpdate(pConfig, &apVal[2], iOld!=iNew, &bContent);
	253849	+ if( rc!=SQLITE_OK ) goto update_out;
	253850	+ }
	253851	+
	253852	+ if( eType1!=SQLITE_INTEGER ){
	253853	+ rc = SQLITE_MISMATCH;
	253854	+ }else if( iOld!=iNew ){
	253855	+ assert( bContent==0 );
	253856	+ if( eConflict==SQLITE_REPLACE ){
	253857	+ rc = sqlite3Fts5StorageDelete(pStorage, iOld, 0, 1);
	253858	+ if( rc==SQLITE_OK ){
	253859	+ rc = sqlite3Fts5StorageDelete(pStorage, iNew, 0, 0);
	253860	+ }
	253861	+ fts5StorageInsert(&rc, pTab, apVal, pRowid);
	253862	+ }else{
	253863	+ rc = sqlite3Fts5StorageFindDeleteRow(pStorage, iOld);
	253864	+ if( rc==SQLITE_OK ){
	253865	+ rc = sqlite3Fts5StorageContentInsert(pStorage, 0, apVal, pRowid);
	253866	+ }
	253867	+ if( rc==SQLITE_OK ){
	253868	+ rc = sqlite3Fts5StorageDelete(pStorage, iOld, 0, 0);
	253869	+ }
	253870	+ if( rc==SQLITE_OK ){
	253871	+ rc = sqlite3Fts5StorageIndexInsert(pStorage, apVal, *pRowid);
	253872	+ }
	253873	+ }
	253874	+ }else if( bContent ){
	253875	+ /* This occurs when an UPDATE on a contentless table affects only
	253876	+ ** UNINDEXED columns. This is a no-op for contentless_unindexed=0
	253877	+ ** tables, or a write to the %_content table only for =1 tables. */
	253878	+ assert( fts5IsContentless(pTab, 1) );
	253879	+ rc = sqlite3Fts5StorageFindDeleteRow(pStorage, iOld);
	253880	+ if( rc==SQLITE_OK ){
	253881	+ rc = sqlite3Fts5StorageContentInsert(pStorage, 1, apVal, pRowid);
	253882	+ }
	253883	+ }else{
	253884	+ rc = sqlite3Fts5StorageDelete(pStorage, iOld, 0, 1);
	253885	+ fts5StorageInsert(&rc, pTab, apVal, pRowid);
	253886	+ }
	253887	+ bUpdateOrDelete = 1;
	253888	+ sqlite3Fts5StorageReleaseDeleteRow(pStorage);
253062	253889	}
253063	253890
253064	253891	}
253065	253892	}
253066	253893
		@@ -253169,15 +253996,15 @@
253169	253996	){
253170	253997	Fts5Cursor pCsr = (Fts5Cursor)pCtx;
253171	253998	Fts5Table pTab = (Fts5Table)(pCsr->base.pVtab);
253172	253999	int rc = SQLITE_OK;
253173	254000
253174		- fts5SetLocale(pTab->pConfig, pLoc, nLoc);
	254001	+ sqlite3Fts5SetLocale(pTab->pConfig, pLoc, nLoc);
253175	254002	rc = sqlite3Fts5Tokenize(pTab->pConfig,
253176	254003	FTS5_TOKENIZE_AUX, pText, nText, pUserData, xToken
253177	254004	);
253178		- fts5SetLocale(pTab->pConfig, 0, 0);
	254005	+ sqlite3Fts5SetLocale(pTab->pConfig, 0, 0);
253179	254006
253180	254007	return rc;
253181	254008	}
253182	254009
253183	254010	/*
		@@ -253200,10 +254027,53 @@
253200	254027
253201	254028	static int fts5ApiPhraseSize(Fts5Context *pCtx, int iPhrase){
253202	254029	Fts5Cursor pCsr = (Fts5Cursor)pCtx;
253203	254030	return sqlite3Fts5ExprPhraseSize(pCsr->pExpr, iPhrase);
253204	254031	}
	254032	+
	254033	+/*
	254034	+** Argument pStmt is an SQL statement of the type used by Fts5Cursor. This
	254035	+** function extracts the text value of column iCol of the current row.
	254036	+** Additionally, if there is an associated locale, it invokes
	254037	+** sqlite3Fts5SetLocale() to configure the tokenizer. In all cases the caller
	254038	+** should invoke sqlite3Fts5ClearLocale() to clear the locale at some point
	254039	+** after this function returns.
	254040	+**
	254041	+** If successful, (*ppText) is set to point to a buffer containing the text
	254042	+** value as utf-8 and SQLITE_OK returned. (*pnText) is set to the size of that
	254043	+** buffer in bytes. It is not guaranteed to be nul-terminated. If an error
	254044	+** occurs, an SQLite error code is returned. The final values of the two
	254045	+** output parameters are undefined in this case.
	254046	+*/
	254047	+static int fts5TextFromStmt(
	254048	+ Fts5Config *pConfig,
	254049	+ sqlite3_stmt *pStmt,
	254050	+ int iCol,
	254051	+ const char **ppText,
	254052	+ int *pnText
	254053	+){
	254054	+ sqlite3_value *pVal = sqlite3_column_value(pStmt, iCol+1);
	254055	+ const char *pLoc = 0;
	254056	+ int nLoc = 0;
	254057	+ int rc = SQLITE_OK;
	254058	+
	254059	+ if( pConfig->bLocale
	254060	+ && pConfig->eContent==FTS5_CONTENT_EXTERNAL
	254061	+ && sqlite3Fts5IsLocaleValue(pConfig, pVal)
	254062	+ ){
	254063	+ rc = sqlite3Fts5DecodeLocaleValue(pVal, ppText, pnText, &pLoc, &nLoc);
	254064	+ }else{
	254065	+ ppText = (const char)sqlite3_value_text(pVal);
	254066	+ *pnText = sqlite3_value_bytes(pVal);
	254067	+ if( pConfig->bLocale && pConfig->eContent==FTS5_CONTENT_NORMAL ){
	254068	+ pLoc = (const char*)sqlite3_column_text(pStmt, iCol+1+pConfig->nCol);
	254069	+ nLoc = sqlite3_column_bytes(pStmt, iCol+1+pConfig->nCol);
	254070	+ }
	254071	+ }
	254072	+ sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
	254073	+ return rc;
	254074	+}
253205	254075
253206	254076	static int fts5ApiColumnText(
253207	254077	Fts5Context *pCtx,
253208	254078	int iCol,
253209	254079	const char **pz,
		@@ -253214,20 +254084,18 @@
253214	254084	Fts5Table pTab = (Fts5Table)(pCsr->base.pVtab);
253215	254085
253216	254086	assert( pCsr->ePlan!=FTS5_PLAN_SPECIAL );
253217	254087	if( iCol<0 \|\| iCol>=pTab->pConfig->nCol ){
253218	254088	rc = SQLITE_RANGE;
253219		- }else if( fts5IsContentless((Fts5FullTable*)(pCsr->base.pVtab)) ){
	254089	+ }else if( fts5IsContentless((Fts5FullTable*)(pCsr->base.pVtab), 0) ){
253220	254090	*pz = 0;
253221	254091	*pn = 0;
253222	254092	}else{
253223	254093	rc = fts5SeekCursor(pCsr, 0);
253224	254094	if( rc==SQLITE_OK ){
253225		- Fts5Config *pConfig = pTab->pConfig;
253226		- int bContent = (pConfig->abUnindexed[iCol]==0);
253227		- sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
253228		- sqlite3Fts5ExtractText(pConfig, pVal, bContent, 0, pz, pn);
	254095	+ rc = fts5TextFromStmt(pTab->pConfig, pCsr->pStmt, iCol, pz, pn);
	254096	+ sqlite3Fts5ClearLocale(pTab->pConfig);
253229	254097	}
253230	254098	}
253231	254099	return rc;
253232	254100	}
253233	254101
		@@ -253249,11 +254117,11 @@
253249	254117	int bLive = (pCsr->pSorter==0);
253250	254118
253251	254119	if( iPhrase<0 \|\| iPhrase>=sqlite3Fts5ExprPhraseCount(pCsr->pExpr) ){
253252	254120	rc = SQLITE_RANGE;
253253	254121	}else if( pConfig->eDetail!=FTS5_DETAIL_FULL
253254		- && pConfig->eContent==FTS5_CONTENT_NONE
	254122	+ && fts5IsContentless((Fts5FullTable*)pCsr->base.pVtab, 1)
253255	254123	){
253256	254124	*pa = 0;
253257	254125	*pn = 0;
253258	254126	return SQLITE_OK;
253259	254127	}else if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_POSLIST) ){
		@@ -253265,21 +254133,19 @@
253265	254133	if( aPopulator==0 ) rc = SQLITE_NOMEM;
253266	254134	if( rc==SQLITE_OK ){
253267	254135	rc = fts5SeekCursor(pCsr, 0);
253268	254136	}
253269	254137	for(i=0; i<pConfig->nCol && rc==SQLITE_OK; i++){
253270		- sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, i+1);
253271	254138	const char *z = 0;
253272	254139	int n = 0;
253273		- int bReset = 0;
253274		- rc = sqlite3Fts5ExtractText(pConfig, pVal, 1, &bReset, &z, &n);
	254140	+ rc = fts5TextFromStmt(pConfig, pCsr->pStmt, i, &z, &n);
253275	254141	if( rc==SQLITE_OK ){
253276	254142	rc = sqlite3Fts5ExprPopulatePoslists(
253277	254143	pConfig, pCsr->pExpr, aPopulator, i, z, n
253278	254144	);
253279	254145	}
253280		- if( bReset ) sqlite3Fts5ClearLocale(pConfig);
	254146	+ sqlite3Fts5ClearLocale(pConfig);
253281	254147	}
253282	254148	sqlite3_free(aPopulator);
253283	254149
253284	254150	if( pCsr->pSorter ){
253285	254151	sqlite3Fts5ExprCheckPoslists(pCsr->pExpr, pCsr->pSorter->iRowid);
		@@ -253447,11 +254313,11 @@
253447	254313
253448	254314	if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_DOCSIZE) ){
253449	254315	if( pConfig->bColumnsize ){
253450	254316	i64 iRowid = fts5CursorRowid(pCsr);
253451	254317	rc = sqlite3Fts5StorageDocsize(pTab->pStorage, iRowid, pCsr->aColumnSize);
253452		- }else if( pConfig->zContent==0 ){
	254318	+ }else if( !pConfig->zContent \|\| pConfig->eContent==FTS5_CONTENT_UNINDEXED ){
253453	254319	int i;
253454	254320	for(i=0; i<pConfig->nCol; i++){
253455	254321	if( pConfig->abUnindexed[i]==0 ){
253456	254322	pCsr->aColumnSize[i] = -1;
253457	254323	}
		@@ -253461,21 +254327,18 @@
253461	254327	rc = fts5SeekCursor(pCsr, 0);
253462	254328	for(i=0; rc==SQLITE_OK && i<pConfig->nCol; i++){
253463	254329	if( pConfig->abUnindexed[i]==0 ){
253464	254330	const char *z = 0;
253465	254331	int n = 0;
253466		- int bReset = 0;
253467		- sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, i+1);
253468		-
253469	254332	pCsr->aColumnSize[i] = 0;
253470		- rc = sqlite3Fts5ExtractText(pConfig, pVal, 1, &bReset, &z, &n);
	254333	+ rc = fts5TextFromStmt(pConfig, pCsr->pStmt, i, &z, &n);
253471	254334	if( rc==SQLITE_OK ){
253472	254335	rc = sqlite3Fts5Tokenize(pConfig, FTS5_TOKENIZE_AUX,
253473	254336	z, n, (void*)&pCsr->aColumnSize[i], fts5ColumnSizeCb
253474	254337	);
253475		- if( bReset ) sqlite3Fts5ClearLocale(pConfig);
253476	254338	}
	254339	+ sqlite3Fts5ClearLocale(pConfig);
253477	254340	}
253478	254341	}
253479	254342	}
253480	254343	CsrFlagClear(pCsr, FTS5CSR_REQUIRE_DOCSIZE);
253481	254344	}
		@@ -253738,46 +254601,23 @@
253738	254601	assert( pCsr->ePlan!=FTS5_PLAN_SPECIAL );
253739	254602	if( iCol<0 \|\| iCol>=pConfig->nCol ){
253740	254603	rc = SQLITE_RANGE;
253741	254604	}else if(
253742	254605	pConfig->abUnindexed[iCol]==0
253743		- && pConfig->eContent!=FTS5_CONTENT_NONE
	254606	+ && 0==fts5IsContentless((Fts5FullTable*)pCsr->base.pVtab, 1)
253744	254607	&& pConfig->bLocale
253745	254608	){
253746	254609	rc = fts5SeekCursor(pCsr, 0);
253747	254610	if( rc==SQLITE_OK ){
253748		- /* Load the value into pVal. pVal is a locale/text pair iff:
253749		- **
253750		- ** 1) It is an SQLITE_BLOB, and
253751		- ** 2) Either the subtype is FTS5_LOCALE_SUBTYPE, or else the
253752		- ** value was loaded from an FTS5_CONTENT_NORMAL table, and
253753		- ** 3) It does not begin with an 0x00 byte.
253754		- */
253755		- sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
253756		- if( sqlite3_value_type(pVal)==SQLITE_BLOB ){
253757		- const u8 pBlob = (const u8)sqlite3_value_blob(pVal);
253758		- int nBlob = sqlite3_value_bytes(pVal);
253759		- if( pConfig->eContent==FTS5_CONTENT_EXTERNAL ){
253760		- const int SZHDR = sizeof(FTS5_LOCALE_HEADER)-1;
253761		- if( nBlob<SZHDR \|\| memcmp(FTS5_LOCALE_HEADER, pBlob, SZHDR) ){
253762		- rc = SQLITE_ERROR;
253763		- }
253764		- pBlob += 4;
253765		- nBlob -= 4;
253766		- }
253767		- if( rc==SQLITE_OK ){
253768		- int nLocale = 0;
253769		- for(nLocale=0; nLocale<nBlob && pBlob[nLocale]!=0x00; nLocale++);
253770		- if( nLocale==nBlob \|\| nLocale==0 ){
253771		- rc = SQLITE_ERROR;
253772		- }else{
253773		- /* A locale/text pair */
253774		- pzLocale = (const char)pBlob;
253775		- *pnLocale = nLocale;
253776		- }
253777		- }
253778		- }
	254611	+ const char *zDummy = 0;
	254612	+ int nDummy = 0;
	254613	+ rc = fts5TextFromStmt(pConfig, pCsr->pStmt, iCol, &zDummy, &nDummy);
	254614	+ if( rc==SQLITE_OK ){
	254615	+ *pzLocale = pConfig->t.pLocale;
	254616	+ *pnLocale = pConfig->t.nLocale;
	254617	+ }
	254618	+ sqlite3Fts5ClearLocale(pConfig);
253779	254619	}
253780	254620	}
253781	254621
253782	254622	return rc;
253783	254623	}
		@@ -253994,61 +254834,10 @@
253994	254834
253995	254835	sqlite3_result_blob(pCtx, val.p, val.n, sqlite3_free);
253996	254836	return rc;
253997	254837	}
253998	254838
253999		-/*
254000		-** Value pVal was read from column iCol of the FTS5 table. This function
254001		-** returns it to the owner of pCtx via a call to an sqlite3_result_xxx()
254002		-** function. This function deals with the same cases as
254003		-** sqlite3Fts5ExtractText():
254004		-**
254005		-** 1) Ordinary values. These can be returned using sqlite3_result_value().
254006		-**
254007		-** 2) Blobs from fts5_locale(). The text is extracted from these and
254008		-** returned via sqlite3_result_text(). The locale is discarded.
254009		-*/
254010		-static void fts5ExtractValueFromColumn(
254011		- sqlite3_context *pCtx,
254012		- Fts5Config *pConfig,
254013		- int iCol,
254014		- sqlite3_value *pVal
254015		-){
254016		- assert( pConfig->eContent!=FTS5_CONTENT_NONE );
254017		-
254018		- if( pConfig->bLocale
254019		- && sqlite3_value_type(pVal)==SQLITE_BLOB
254020		- && pConfig->abUnindexed[iCol]==0
254021		- ){
254022		- const int SZHDR = sizeof(FTS5_LOCALE_HEADER)-1;
254023		- const u8 *pBlob = sqlite3_value_blob(pVal);
254024		- int nBlob = sqlite3_value_bytes(pVal);
254025		- int ii;
254026		-
254027		- if( pConfig->eContent==FTS5_CONTENT_EXTERNAL ){
254028		- if( nBlob<SZHDR \|\| memcmp(pBlob, FTS5_LOCALE_HEADER, SZHDR) ){
254029		- sqlite3_result_error_code(pCtx, SQLITE_ERROR);
254030		- return;
254031		- }else{
254032		- pBlob += 4;
254033		- nBlob -= 4;
254034		- }
254035		- }
254036		-
254037		- for(ii=0; ii<nBlob && pBlob[ii]; ii++);
254038		- if( ii==0 \|\| ii==nBlob ){
254039		- sqlite3_result_error_code(pCtx, SQLITE_ERROR);
254040		- }else{
254041		- const char pText = (const char)&pBlob[ii+1];
254042		- sqlite3_result_text(pCtx, pText, nBlob-ii-1, SQLITE_TRANSIENT);
254043		- }
254044		- return;
254045		- }
254046		-
254047		- sqlite3_result_value(pCtx, pVal);
254048		-}
254049		-
254050	254839	/*
254051	254840	** This is the xColumn method, called by SQLite to request a value from
254052	254841	** the row that the supplied cursor currently points to.
254053	254842	*/
254054	254843	static int fts5ColumnMethod(
		@@ -254087,26 +254876,31 @@
254087	254876	if( pCsr->pRank \|\| SQLITE_OK==(rc = fts5FindRankFunction(pCsr)) ){
254088	254877	fts5ApiInvoke(pCsr->pRank, pCsr, pCtx, pCsr->nRankArg, pCsr->apRankArg);
254089	254878	}
254090	254879	}
254091	254880	}else{
254092		- /* A column created by the user containing values. */
254093		- int bNochange = sqlite3_vtab_nochange(pCtx);
254094		-
254095		- if( fts5IsContentless(pTab) ){
254096		- if( bNochange && pConfig->bContentlessDelete ){
254097		- fts5ResultError(pCtx, "cannot UPDATE a subset of "
254098		- "columns on fts5 contentless-delete table: %s", pConfig->zName
254099		- );
254100		- }
254101		- }else if( bNochange==0 \|\| pConfig->eContent!=FTS5_CONTENT_NORMAL ){
	254881	+ if( !sqlite3_vtab_nochange(pCtx) && pConfig->eContent!=FTS5_CONTENT_NONE ){
254102	254882	pConfig->pzErrmsg = &pTab->p.base.zErrMsg;
254103	254883	rc = fts5SeekCursor(pCsr, 1);
254104	254884	if( rc==SQLITE_OK ){
254105	254885	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
254106		- fts5ExtractValueFromColumn(pCtx, pConfig, iCol, pVal);
	254886	+ if( pConfig->bLocale
	254887	+ && pConfig->eContent==FTS5_CONTENT_EXTERNAL
	254888	+ && sqlite3Fts5IsLocaleValue(pConfig, pVal)
	254889	+ ){
	254890	+ const char *z = 0;
	254891	+ int n = 0;
	254892	+ rc = fts5TextFromStmt(pConfig, pCsr->pStmt, iCol, &z, &n);
	254893	+ if( rc==SQLITE_OK ){
	254894	+ sqlite3_result_text(pCtx, z, n, SQLITE_TRANSIENT);
	254895	+ }
	254896	+ sqlite3Fts5ClearLocale(pConfig);
	254897	+ }else{
	254898	+ sqlite3_result_value(pCtx, pVal);
	254899	+ }
254107	254900	}
	254901	+
254108	254902	pConfig->pzErrmsg = 0;
254109	254903	}
254110	254904	}
254111	254905
254112	254906	return rc;
		@@ -254625,11 +255419,11 @@
254625	255419	int nArg, /* Number of args */
254626	255420	sqlite3_value *apUnused / Function arguments */
254627	255421	){
254628	255422	assert( nArg==0 );
254629	255423	UNUSED_PARAM2(nArg, apUnused);
254630		- sqlite3_result_text(pCtx, "fts5: 2024-09-02 18:41:59 e6bec37ea1ca51e1d048941ce4c5211d8fc5c5e3556a1441f9c79b036843f9e3", -1, SQLITE_TRANSIENT);
	255424	+ sqlite3_result_text(pCtx, "fts5: 2024-12-09 20:46:36 e2bae4143afd07de1ae55a6d2606a3b541a5b94568aa41f6a96e5d1245471653", -1, SQLITE_TRANSIENT);
254631	255425	}
254632	255426
254633	255427	/*
254634	255428	** Implementation of fts5_locale(LOCALE, TEXT) function.
254635	255429	**
		@@ -254664,36 +255458,48 @@
254664	255458	nText = sqlite3_value_bytes(apArg[1]);
254665	255459
254666	255460	if( zLocale==0 \|\| zLocale[0]=='\0' ){
254667	255461	sqlite3_result_text(pCtx, zText, nText, SQLITE_TRANSIENT);
254668	255462	}else{
	255463	+ Fts5Global p = (Fts5Global)sqlite3_user_data(pCtx);
254669	255464	u8 *pBlob = 0;
254670	255465	u8 *pCsr = 0;
254671	255466	int nBlob = 0;
254672		- const int nHdr = 4;
254673		- assert( sizeof(FTS5_LOCALE_HEADER)==nHdr+1 );
254674	255467
254675		- nBlob = nHdr + nLocale + 1 + nText;
	255468	+ nBlob = FTS5_LOCALE_HDR_SIZE + nLocale + 1 + nText;
254676	255469	pBlob = (u8*)sqlite3_malloc(nBlob);
254677	255470	if( pBlob==0 ){
254678	255471	sqlite3_result_error_nomem(pCtx);
254679	255472	return;
254680	255473	}
254681	255474
254682	255475	pCsr = pBlob;
254683		- memcpy(pCsr, FTS5_LOCALE_HEADER, nHdr);
254684		- pCsr += nHdr;
	255476	+ memcpy(pCsr, (const u8*)p->aLocaleHdr, FTS5_LOCALE_HDR_SIZE);
	255477	+ pCsr += FTS5_LOCALE_HDR_SIZE;
254685	255478	memcpy(pCsr, zLocale, nLocale);
254686	255479	pCsr += nLocale;
254687	255480	(*pCsr++) = 0x00;
254688	255481	if( zText ) memcpy(pCsr, zText, nText);
254689	255482	assert( &pCsr[nText]==&pBlob[nBlob] );
254690	255483
254691	255484	sqlite3_result_blob(pCtx, pBlob, nBlob, sqlite3_free);
254692		- sqlite3_result_subtype(pCtx, FTS5_LOCALE_SUBTYPE);
254693	255485	}
254694	255486	}
	255487	+
	255488	+/*
	255489	+** Implementation of fts5_insttoken() function.
	255490	+*/
	255491	+static void fts5InsttokenFunc(
	255492	+ sqlite3_context pCtx, / Function call context */
	255493	+ int nArg, /* Number of args */
	255494	+ sqlite3_value *apArg / Function arguments */
	255495	+){
	255496	+ assert( nArg==1 );
	255497	+ (void)nArg;
	255498	+ sqlite3_result_value(pCtx, apArg[0]);
	255499	+ sqlite3_result_subtype(pCtx, FTS5_INSTTOKEN_SUBTYPE);
	255500	+}
254695	255501
254696	255502	/*
254697	255503	** Return true if zName is the extension on one of the shadow tables used
254698	255504	** by this module.
254699	255505	*/
		@@ -254789,10 +255595,20 @@
254789	255595	pGlobal->api.xCreateFunction = fts5CreateAux;
254790	255596	pGlobal->api.xCreateTokenizer = fts5CreateTokenizer;
254791	255597	pGlobal->api.xFindTokenizer = fts5FindTokenizer;
254792	255598	pGlobal->api.xCreateTokenizer_v2 = fts5CreateTokenizer_v2;
254793	255599	pGlobal->api.xFindTokenizer_v2 = fts5FindTokenizer_v2;
	255600	+
	255601	+ /* Initialize pGlobal->aLocaleHdr[] to a 128-bit pseudo-random vector.
	255602	+ ** The constants below were generated randomly. */
	255603	+ sqlite3_randomness(sizeof(pGlobal->aLocaleHdr), pGlobal->aLocaleHdr);
	255604	+ pGlobal->aLocaleHdr[0] ^= 0xF924976D;
	255605	+ pGlobal->aLocaleHdr[1] ^= 0x16596E13;
	255606	+ pGlobal->aLocaleHdr[2] ^= 0x7C80BEAA;
	255607	+ pGlobal->aLocaleHdr[3] ^= 0x9B03A67F;
	255608	+ assert( sizeof(pGlobal->aLocaleHdr)==16 );
	255609	+
254794	255610	rc = sqlite3_create_module_v2(db, "fts5", &fts5Mod, p, fts5ModuleDestroy);
254795	255611	if( rc==SQLITE_OK ) rc = sqlite3Fts5IndexInit(db);
254796	255612	if( rc==SQLITE_OK ) rc = sqlite3Fts5ExprInit(pGlobal, db);
254797	255613	if( rc==SQLITE_OK ) rc = sqlite3Fts5AuxInit(&pGlobal->api);
254798	255614	if( rc==SQLITE_OK ) rc = sqlite3Fts5TokenizerInit(&pGlobal->api);
		@@ -254810,13 +255626,20 @@
254810	255626	);
254811	255627	}
254812	255628	if( rc==SQLITE_OK ){
254813	255629	rc = sqlite3_create_function(
254814	255630	db, "fts5_locale", 2,
254815		- SQLITE_UTF8\|SQLITE_INNOCUOUS\|SQLITE_RESULT_SUBTYPE,
	255631	+ SQLITE_UTF8\|SQLITE_INNOCUOUS\|SQLITE_RESULT_SUBTYPE\|SQLITE_SUBTYPE,
254816	255632	p, fts5LocaleFunc, 0, 0
254817	255633	);
	255634	+ }
	255635	+ if( rc==SQLITE_OK ){
	255636	+ rc = sqlite3_create_function(
	255637	+ db, "fts5_insttoken", 1,
	255638	+ SQLITE_UTF8\|SQLITE_INNOCUOUS\|SQLITE_RESULT_SUBTYPE,
	255639	+ p, fts5InsttokenFunc, 0, 0
	255640	+ );
254818	255641	}
254819	255642	}
254820	255643
254821	255644	/* If SQLITE_FTS5_ENABLE_TEST_MI is defined, assume that the file
254822	255645	** fts5_test_mi.c is compiled and linked into the executable. And call
		@@ -255015,24 +255838,39 @@
255015	255838	);
255016	255839	break;
255017	255840
255018	255841	case FTS5_STMT_INSERT_CONTENT:
255019	255842	case FTS5_STMT_REPLACE_CONTENT: {
255020		- int nCol = pC->nCol + 1;
255021		- char *zBind;
	255843	+ char *zBind = 0;
255022	255844	int i;
255023	255845
255024		- zBind = sqlite3_malloc64(1 + nCol*2);
255025		- if( zBind ){
255026		- for(i=0; i<nCol; i++){
255027		- zBind[i*2] = '?';
255028		- zBind[i*2 + 1] = ',';
255029		- }
255030		- zBind[i*2-1] = '\0';
255031		- zSql = sqlite3_mprintf(azStmt[eStmt], pC->zDb, pC->zName, zBind);
255032		- sqlite3_free(zBind);
255033		- }
	255846	+ assert( pC->eContent==FTS5_CONTENT_NORMAL
	255847	+ \|\| pC->eContent==FTS5_CONTENT_UNINDEXED
	255848	+ );
	255849	+
	255850	+ /* Add bindings for the "c*" columns - those that store the actual
	255851	+ ** table content. If eContent==NORMAL, then there is one binding
	255852	+ ** for each column. Or, if eContent==UNINDEXED, then there are only
	255853	+ ** bindings for the UNINDEXED columns. */
	255854	+ for(i=0; rc==SQLITE_OK && i<(pC->nCol+1); i++){
	255855	+ if( !i \|\| pC->eContent==FTS5_CONTENT_NORMAL \|\| pC->abUnindexed[i-1] ){
	255856	+ zBind = sqlite3Fts5Mprintf(&rc, "%z%s?%d", zBind, zBind?",":"",i+1);
	255857	+ }
	255858	+ }
	255859	+
	255860	+ /* Add bindings for any "l*" columns. Only non-UNINDEXED columns
	255861	+ ** require these. */
	255862	+ if( pC->bLocale && pC->eContent==FTS5_CONTENT_NORMAL ){
	255863	+ for(i=0; rc==SQLITE_OK && i<pC->nCol; i++){
	255864	+ if( pC->abUnindexed[i]==0 ){
	255865	+ zBind = sqlite3Fts5Mprintf(&rc, "%z,?%d", zBind, pC->nCol+i+2);
	255866	+ }
	255867	+ }
	255868	+ }
	255869	+
	255870	+ zSql = sqlite3Fts5Mprintf(&rc, azStmt[eStmt], pC->zDb, pC->zName,zBind);
	255871	+ sqlite3_free(zBind);
255034	255872	break;
255035	255873	}
255036	255874
255037	255875	case FTS5_STMT_REPLACE_DOCSIZE:
255038	255876	zSql = sqlite3_mprintf(azStmt[eStmt], pC->zDb, pC->zName,
		@@ -255214,23 +256052,37 @@
255214	256052	p->aTotalSize = (i64*)&p[1];
255215	256053	p->pConfig = pConfig;
255216	256054	p->pIndex = pIndex;
255217	256055
255218	256056	if( bCreate ){
255219		- if( pConfig->eContent==FTS5_CONTENT_NORMAL ){
	256057	+ if( pConfig->eContent==FTS5_CONTENT_NORMAL
	256058	+ \|\| pConfig->eContent==FTS5_CONTENT_UNINDEXED
	256059	+ ){
255220	256060	int nDefn = 32 + pConfig->nCol*10;
255221		- char zDefn = sqlite3_malloc64(32 + (sqlite3_int64)pConfig->nCol 10);
	256061	+ char zDefn = sqlite3_malloc64(32 + (sqlite3_int64)pConfig->nCol 20);
255222	256062	if( zDefn==0 ){
255223	256063	rc = SQLITE_NOMEM;
255224	256064	}else{
255225	256065	int i;
255226	256066	int iOff;
255227	256067	sqlite3_snprintf(nDefn, zDefn, "id INTEGER PRIMARY KEY");
255228	256068	iOff = (int)strlen(zDefn);
255229	256069	for(i=0; i<pConfig->nCol; i++){
255230		- sqlite3_snprintf(nDefn-iOff, &zDefn[iOff], ", c%d", i);
255231		- iOff += (int)strlen(&zDefn[iOff]);
	256070	+ if( pConfig->eContent==FTS5_CONTENT_NORMAL
	256071	+ \|\| pConfig->abUnindexed[i]
	256072	+ ){
	256073	+ sqlite3_snprintf(nDefn-iOff, &zDefn[iOff], ", c%d", i);
	256074	+ iOff += (int)strlen(&zDefn[iOff]);
	256075	+ }
	256076	+ }
	256077	+ if( pConfig->bLocale ){
	256078	+ for(i=0; i<pConfig->nCol; i++){
	256079	+ if( pConfig->abUnindexed[i]==0 ){
	256080	+ sqlite3_snprintf(nDefn-iOff, &zDefn[iOff], ", l%d", i);
	256081	+ iOff += (int)strlen(&zDefn[iOff]);
	256082	+ }
	256083	+ }
255232	256084	}
255233	256085	rc = sqlite3Fts5CreateTable(pConfig, "content", zDefn, 0, pzErr);
255234	256086	}
255235	256087	sqlite3_free(zDefn);
255236	256088	}
		@@ -255379,33 +256231,43 @@
255379	256231	for(iCol=1; rc==SQLITE_OK && iCol<=pConfig->nCol; iCol++){
255380	256232	if( pConfig->abUnindexed[iCol-1]==0 ){
255381	256233	sqlite3_value *pVal = 0;
255382	256234	const char *pText = 0;
255383	256235	int nText = 0;
255384		- int bReset = 0;
	256236	+ const char *pLoc = 0;
	256237	+ int nLoc = 0;
255385	256238
255386	256239	assert( pSeek==0 \|\| apVal==0 );
255387	256240	assert( pSeek!=0 \|\| apVal!=0 );
255388	256241	if( pSeek ){
255389	256242	pVal = sqlite3_column_value(pSeek, iCol);
255390	256243	}else{
255391	256244	pVal = apVal[iCol-1];
255392	256245	}
255393	256246
255394		- rc = sqlite3Fts5ExtractText(
255395		- pConfig, pVal, pSeek!=0, &bReset, &pText, &nText
255396		- );
	256247	+ if( pConfig->bLocale && sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
	256248	+ rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
	256249	+ }else{
	256250	+ pText = (const char*)sqlite3_value_text(pVal);
	256251	+ nText = sqlite3_value_bytes(pVal);
	256252	+ if( pConfig->bLocale && pSeek ){
	256253	+ pLoc = (const char*)sqlite3_column_text(pSeek, iCol + pConfig->nCol);
	256254	+ nLoc = sqlite3_column_bytes(pSeek, iCol + pConfig->nCol);
	256255	+ }
	256256	+ }
	256257	+
255397	256258	if( rc==SQLITE_OK ){
	256259	+ sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
255398	256260	ctx.szCol = 0;
255399	256261	rc = sqlite3Fts5Tokenize(pConfig, FTS5_TOKENIZE_DOCUMENT,
255400	256262	pText, nText, (void*)&ctx, fts5StorageInsertCallback
255401	256263	);
255402	256264	p->aTotalSize[iCol-1] -= (i64)ctx.szCol;
255403	256265	if( rc==SQLITE_OK && p->aTotalSize[iCol-1]<0 ){
255404	256266	rc = FTS5_CORRUPT;
255405	256267	}
255406		- if( bReset ) sqlite3Fts5ClearLocale(pConfig);
	256268	+ sqlite3Fts5ClearLocale(pConfig);
255407	256269	}
255408	256270	}
255409	256271	}
255410	256272	if( rc==SQLITE_OK && p->nTotalRow<1 ){
255411	256273	rc = FTS5_CORRUPT;
		@@ -255446,11 +256308,13 @@
255446	256308	i64 iOrigin = 0;
255447	256309	sqlite3_stmt *pLookup = 0;
255448	256310	int rc = SQLITE_OK;
255449	256311
255450	256312	assert( p->pConfig->bContentlessDelete );
255451		- assert( p->pConfig->eContent==FTS5_CONTENT_NONE );
	256313	+ assert( p->pConfig->eContent==FTS5_CONTENT_NONE
	256314	+ \|\| p->pConfig->eContent==FTS5_CONTENT_UNINDEXED
	256315	+ );
255452	256316
255453	256317	/* Look up the origin of the document in the %_docsize table. Store
255454	256318	** this in stack variable iOrigin. */
255455	256319	rc = fts5StorageGetStmt(p, FTS5_STMT_LOOKUP_DOCSIZE, &pLookup, 0);
255456	256320	if( rc==SQLITE_OK ){
		@@ -255570,10 +256434,16 @@
255570	256434	}
255571	256435
255572	256436	if( rc==SQLITE_OK ){
255573	256437	if( p->pConfig->bContentlessDelete ){
255574	256438	rc = fts5StorageContentlessDelete(p, iDel);
	256439	+ if( rc==SQLITE_OK
	256440	+ && bSaveRow
	256441	+ && p->pConfig->eContent==FTS5_CONTENT_UNINDEXED
	256442	+ ){
	256443	+ rc = sqlite3Fts5StorageFindDeleteRow(p, iDel);
	256444	+ }
255575	256445	}else{
255576	256446	rc = fts5StorageDeleteFromIndex(p, iDel, apVal, bSaveRow);
255577	256447	}
255578	256448	}
255579	256449
		@@ -255586,11 +256456,13 @@
255586	256456	rc = sqlite3_reset(pDel);
255587	256457	}
255588	256458	}
255589	256459
255590	256460	/* Delete the %_content record */
255591		- if( pConfig->eContent==FTS5_CONTENT_NORMAL ){
	256461	+ if( pConfig->eContent==FTS5_CONTENT_NORMAL
	256462	+ \|\| pConfig->eContent==FTS5_CONTENT_UNINDEXED
	256463	+ ){
255592	256464	if( rc==SQLITE_OK ){
255593	256465	rc = fts5StorageGetStmt(p, FTS5_STMT_DELETE_CONTENT, &pDel, 0);
255594	256466	}
255595	256467	if( rc==SQLITE_OK ){
255596	256468	sqlite3_bind_int64(pDel, 1, iDel);
		@@ -255618,12 +256490,17 @@
255618	256490	pConfig->zDb, pConfig->zName,
255619	256491	pConfig->zDb, pConfig->zName
255620	256492	);
255621	256493	if( rc==SQLITE_OK && pConfig->bColumnsize ){
255622	256494	rc = fts5ExecPrintf(pConfig->db, 0,
255623		- "DELETE FROM %Q.'%q_docsize';",
255624		- pConfig->zDb, pConfig->zName
	256495	+ "DELETE FROM %Q.'%q_docsize';", pConfig->zDb, pConfig->zName
	256496	+ );
	256497	+ }
	256498	+
	256499	+ if( rc==SQLITE_OK && pConfig->eContent==FTS5_CONTENT_UNINDEXED ){
	256500	+ rc = fts5ExecPrintf(pConfig->db, 0,
	256501	+ "DELETE FROM %Q.'%q_content';", pConfig->zDb, pConfig->zName
255625	256502	);
255626	256503	}
255627	256504
255628	256505	/* Reinitialize the %_data table. This call creates the initial structure
255629	256506	** and averages records. */
		@@ -255660,24 +256537,39 @@
255660	256537	sqlite3Fts5BufferZero(&buf);
255661	256538	rc = sqlite3Fts5IndexBeginWrite(p->pIndex, 0, iRowid);
255662	256539	for(ctx.iCol=0; rc==SQLITE_OK && ctx.iCol<pConfig->nCol; ctx.iCol++){
255663	256540	ctx.szCol = 0;
255664	256541	if( pConfig->abUnindexed[ctx.iCol]==0 ){
255665		- int bReset = 0; /* True if tokenizer locale must be reset */
255666	256542	int nText = 0; /* Size of pText in bytes */
255667	256543	const char pText = 0; / Pointer to buffer containing text value */
	256544	+ int nLoc = 0; /* Size of pLoc in bytes */
	256545	+ const char pLoc = 0; / Pointer to buffer containing text value */
	256546	+
255668	256547	sqlite3_value *pVal = sqlite3_column_value(pScan, ctx.iCol+1);
	256548	+ if( pConfig->eContent==FTS5_CONTENT_EXTERNAL
	256549	+ && sqlite3Fts5IsLocaleValue(pConfig, pVal)
	256550	+ ){
	256551	+ rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
	256552	+ }else{
	256553	+ pText = (const char*)sqlite3_value_text(pVal);
	256554	+ nText = sqlite3_value_bytes(pVal);
	256555	+ if( pConfig->bLocale ){
	256556	+ int iCol = ctx.iCol + 1 + pConfig->nCol;
	256557	+ pLoc = (const char*)sqlite3_column_text(pScan, iCol);
	256558	+ nLoc = sqlite3_column_bytes(pScan, iCol);
	256559	+ }
	256560	+ }
255669	256561
255670		- rc = sqlite3Fts5ExtractText(pConfig, pVal, 1, &bReset, &pText, &nText);
255671	256562	if( rc==SQLITE_OK ){
	256563	+ sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
255672	256564	rc = sqlite3Fts5Tokenize(pConfig,
255673	256565	FTS5_TOKENIZE_DOCUMENT,
255674	256566	pText, nText,
255675	256567	(void*)&ctx,
255676	256568	fts5StorageInsertCallback
255677	256569	);
255678		- if( bReset ) sqlite3Fts5ClearLocale(pConfig);
	256570	+ sqlite3Fts5ClearLocale(pConfig);
255679	256571	}
255680	256572	}
255681	256573	sqlite3Fts5BufferAppendVarint(&rc, &buf, ctx.szCol);
255682	256574	p->aTotalSize[ctx.iCol] += (i64)ctx.szCol;
255683	256575	}
		@@ -255740,53 +256632,75 @@
255740	256632	/*
255741	256633	** Insert a new row into the FTS content table.
255742	256634	*/
255743	256635	static int sqlite3Fts5StorageContentInsert(
255744	256636	Fts5Storage *p,
	256637	+ int bReplace, /* True to use REPLACE instead of INSERT */
255745	256638	sqlite3_value **apVal,
255746	256639	i64 *piRowid
255747	256640	){
255748	256641	Fts5Config *pConfig = p->pConfig;
255749	256642	int rc = SQLITE_OK;
255750	256643
255751	256644	/* Insert the new row into the %_content table. */
255752		- if( pConfig->eContent!=FTS5_CONTENT_NORMAL ){
	256645	+ if( pConfig->eContent!=FTS5_CONTENT_NORMAL
	256646	+ && pConfig->eContent!=FTS5_CONTENT_UNINDEXED
	256647	+ ){
255753	256648	if( sqlite3_value_type(apVal[1])==SQLITE_INTEGER ){
255754	256649	*piRowid = sqlite3_value_int64(apVal[1]);
255755	256650	}else{
255756	256651	rc = fts5StorageNewRowid(p, piRowid);
255757	256652	}
255758	256653	}else{
255759	256654	sqlite3_stmt pInsert = 0; / Statement to write %_content table */
255760	256655	int i; /* Counter variable */
255761		- rc = fts5StorageGetStmt(p, FTS5_STMT_INSERT_CONTENT, &pInsert, 0);
255762		- for(i=1; rc==SQLITE_OK && i<=pConfig->nCol+1; i++){
255763		- sqlite3_value *pVal = apVal[i];
255764		- if( sqlite3_value_nochange(pVal) && p->pSavedRow ){
255765		- /* This is an UPDATE statement, and column (i-2) was not modified.
255766		- ** Retrieve the value from Fts5Storage.pSavedRow instead. */
255767		- pVal = sqlite3_column_value(p->pSavedRow, i-1);
255768		- }else if( sqlite3_value_subtype(pVal)==FTS5_LOCALE_SUBTYPE ){
255769		- assert( pConfig->bLocale );
255770		- assert( i>1 );
255771		- if( pConfig->abUnindexed[i-2] ){
255772		- /* At attempt to insert an fts5_locale() value into an UNINDEXED
255773		- ** column. Strip the locale away and just bind the text. */
	256656	+
	256657	+ assert( FTS5_STMT_INSERT_CONTENT+1==FTS5_STMT_REPLACE_CONTENT );
	256658	+ assert( bReplace==0 \|\| bReplace==1 );
	256659	+ rc = fts5StorageGetStmt(p, FTS5_STMT_INSERT_CONTENT+bReplace, &pInsert, 0);
	256660	+ if( pInsert ) sqlite3_clear_bindings(pInsert);
	256661	+
	256662	+ /* Bind the rowid value */
	256663	+ sqlite3_bind_value(pInsert, 1, apVal[1]);
	256664	+
	256665	+ /* Loop through values for user-defined columns. i=2 is the leftmost
	256666	+ ** user-defined column. As is column 1 of pSavedRow. */
	256667	+ for(i=2; rc==SQLITE_OK && i<=pConfig->nCol+1; i++){
	256668	+ int bUnindexed = pConfig->abUnindexed[i-2];
	256669	+ if( pConfig->eContent==FTS5_CONTENT_NORMAL \|\| bUnindexed ){
	256670	+ sqlite3_value *pVal = apVal[i];
	256671	+
	256672	+ if( sqlite3_value_nochange(pVal) && p->pSavedRow ){
	256673	+ /* This is an UPDATE statement, and user-defined column (i-2) was not
	256674	+ ** modified. Retrieve the value from Fts5Storage.pSavedRow. */
	256675	+ pVal = sqlite3_column_value(p->pSavedRow, i-1);
	256676	+ if( pConfig->bLocale && bUnindexed==0 ){
	256677	+ sqlite3_bind_value(pInsert, pConfig->nCol + i,
	256678	+ sqlite3_column_value(p->pSavedRow, pConfig->nCol + i - 1)
	256679	+ );
	256680	+ }
	256681	+ }else if( sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
255774	256682	const char *pText = 0;
	256683	+ const char *pLoc = 0;
255775	256684	int nText = 0;
255776		- rc = sqlite3Fts5ExtractText(pConfig, pVal, 0, 0, &pText, &nText);
255777		- sqlite3_bind_text(pInsert, i, pText, nText, SQLITE_TRANSIENT);
255778		- }else{
255779		- const u8 pBlob = (const u8)sqlite3_value_blob(pVal);
255780		- int nBlob = sqlite3_value_bytes(pVal);
255781		- assert( nBlob>4 );
255782		- sqlite3_bind_blob(pInsert, i, pBlob+4, nBlob-4, SQLITE_TRANSIENT);
255783		- }
255784		- continue;
255785		- }
255786		-
255787		- rc = sqlite3_bind_value(pInsert, i, pVal);
	256685	+ int nLoc = 0;
	256686	+ assert( pConfig->bLocale );
	256687	+
	256688	+ rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
	256689	+ if( rc==SQLITE_OK ){
	256690	+ sqlite3_bind_text(pInsert, i, pText, nText, SQLITE_TRANSIENT);
	256691	+ if( bUnindexed==0 ){
	256692	+ int iLoc = pConfig->nCol + i;
	256693	+ sqlite3_bind_text(pInsert, iLoc, pLoc, nLoc, SQLITE_TRANSIENT);
	256694	+ }
	256695	+ }
	256696	+
	256697	+ continue;
	256698	+ }
	256699	+
	256700	+ rc = sqlite3_bind_value(pInsert, i, pVal);
	256701	+ }
255788	256702	}
255789	256703	if( rc==SQLITE_OK ){
255790	256704	sqlite3_step(pInsert);
255791	256705	rc = sqlite3_reset(pInsert);
255792	256706	}
		@@ -255817,27 +256731,41 @@
255817	256731	rc = sqlite3Fts5IndexBeginWrite(p->pIndex, 0, iRowid);
255818	256732	}
255819	256733	for(ctx.iCol=0; rc==SQLITE_OK && ctx.iCol<pConfig->nCol; ctx.iCol++){
255820	256734	ctx.szCol = 0;
255821	256735	if( pConfig->abUnindexed[ctx.iCol]==0 ){
255822		- int bReset = 0; /* True if tokenizer locale must be reset */
255823	256736	int nText = 0; /* Size of pText in bytes */
255824	256737	const char pText = 0; / Pointer to buffer containing text value */
	256738	+ int nLoc = 0; /* Size of pText in bytes */
	256739	+ const char pLoc = 0; / Pointer to buffer containing text value */
	256740	+
255825	256741	sqlite3_value *pVal = apVal[ctx.iCol+2];
255826		- int bDisk = 0;
255827	256742	if( p->pSavedRow && sqlite3_value_nochange(pVal) ){
255828	256743	pVal = sqlite3_column_value(p->pSavedRow, ctx.iCol+1);
255829		- bDisk = 1;
	256744	+ if( pConfig->eContent==FTS5_CONTENT_NORMAL && pConfig->bLocale ){
	256745	+ int iCol = ctx.iCol + 1 + pConfig->nCol;
	256746	+ pLoc = (const char*)sqlite3_column_text(p->pSavedRow, iCol);
	256747	+ nLoc = sqlite3_column_bytes(p->pSavedRow, iCol);
	256748	+ }
	256749	+ }else{
	256750	+ pVal = apVal[ctx.iCol+2];
	256751	+ }
	256752	+
	256753	+ if( pConfig->bLocale && sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
	256754	+ rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
	256755	+ }else{
	256756	+ pText = (const char*)sqlite3_value_text(pVal);
	256757	+ nText = sqlite3_value_bytes(pVal);
255830	256758	}
255831		- rc = sqlite3Fts5ExtractText(pConfig, pVal, bDisk, &bReset, &pText,&nText);
	256759	+
255832	256760	if( rc==SQLITE_OK ){
255833		- assert( bReset==0 \|\| pConfig->bLocale );
	256761	+ sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
255834	256762	rc = sqlite3Fts5Tokenize(pConfig,
255835	256763	FTS5_TOKENIZE_DOCUMENT, pText, nText, (void*)&ctx,
255836	256764	fts5StorageInsertCallback
255837	256765	);
255838		- if( bReset ) sqlite3Fts5ClearLocale(pConfig);
	256766	+ sqlite3Fts5ClearLocale(pConfig);
255839	256767	}
255840	256768	}
255841	256769	sqlite3Fts5BufferAppendVarint(&rc, &buf, ctx.szCol);
255842	256770	p->aTotalSize[ctx.iCol] += (i64)ctx.szCol;
255843	256771	}
		@@ -255998,41 +256926,65 @@
255998	256926	}
255999	256927	if( rc==SQLITE_OK && pConfig->eDetail==FTS5_DETAIL_NONE ){
256000	256928	rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
256001	256929	}
256002	256930	for(i=0; rc==SQLITE_OK && i<pConfig->nCol; i++){
256003		- if( pConfig->abUnindexed[i] ) continue;
256004		- ctx.iCol = i;
256005		- ctx.szCol = 0;
256006		- if( pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
256007		- rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
256008		- }
256009		- if( rc==SQLITE_OK ){
256010		- int bReset = 0; /* True if tokenizer locale must be reset */
256011		- int nText = 0; /* Size of pText in bytes */
256012		- const char pText = 0; / Pointer to buffer containing text value */
256013		-
256014		- rc = sqlite3Fts5ExtractText(pConfig,
256015		- sqlite3_column_value(pScan, i+1), 1, &bReset, &pText, &nText
256016		- );
256017		- if( rc==SQLITE_OK ){
	256931	+ if( pConfig->abUnindexed[i]==0 ){
	256932	+ const char *pText = 0;
	256933	+ int nText = 0;
	256934	+ const char *pLoc = 0;
	256935	+ int nLoc = 0;
	256936	+ sqlite3_value *pVal = sqlite3_column_value(pScan, i+1);
	256937	+
	256938	+ if( pConfig->eContent==FTS5_CONTENT_EXTERNAL
	256939	+ && sqlite3Fts5IsLocaleValue(pConfig, pVal)
	256940	+ ){
	256941	+ rc = sqlite3Fts5DecodeLocaleValue(
	256942	+ pVal, &pText, &nText, &pLoc, &nLoc
	256943	+ );
	256944	+ }else{
	256945	+ if( pConfig->eContent==FTS5_CONTENT_NORMAL && pConfig->bLocale ){
	256946	+ int iCol = i + 1 + pConfig->nCol;
	256947	+ pLoc = (const char*)sqlite3_column_text(pScan, iCol);
	256948	+ nLoc = sqlite3_column_bytes(pScan, iCol);
	256949	+ }
	256950	+ pText = (const char*)sqlite3_value_text(pVal);
	256951	+ nText = sqlite3_value_bytes(pVal);
	256952	+ }
	256953	+
	256954	+ ctx.iCol = i;
	256955	+ ctx.szCol = 0;
	256956	+
	256957	+ if( rc==SQLITE_OK && pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
	256958	+ rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
	256959	+ }
	256960	+
	256961	+ if( rc==SQLITE_OK ){
	256962	+ sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
256018	256963	rc = sqlite3Fts5Tokenize(pConfig,
256019	256964	FTS5_TOKENIZE_DOCUMENT,
256020	256965	pText, nText,
256021	256966	(void*)&ctx,
256022	256967	fts5StorageIntegrityCallback
256023	256968	);
256024		- if( bReset ) sqlite3Fts5ClearLocale(pConfig);
256025		- }
256026		- }
256027		- if( rc==SQLITE_OK && pConfig->bColumnsize && ctx.szCol!=aColSize[i] ){
256028		- rc = FTS5_CORRUPT;
256029		- }
256030		- aTotalSize[i] += ctx.szCol;
256031		- if( pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
256032		- sqlite3Fts5TermsetFree(ctx.pTermset);
256033		- ctx.pTermset = 0;
	256969	+ sqlite3Fts5ClearLocale(pConfig);
	256970	+ }
	256971	+
	256972	+ /* If this is not a columnsize=0 database, check that the number
	256973	+ ** of tokens in the value matches the aColSize[] value read from
	256974	+ ** the %_docsize table. */
	256975	+ if( rc==SQLITE_OK
	256976	+ && pConfig->bColumnsize
	256977	+ && ctx.szCol!=aColSize[i]
	256978	+ ){
	256979	+ rc = FTS5_CORRUPT;
	256980	+ }
	256981	+ aTotalSize[i] += ctx.szCol;
	256982	+ if( pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
	256983	+ sqlite3Fts5TermsetFree(ctx.pTermset);
	256984	+ ctx.pTermset = 0;
	256985	+ }
256034	256986	}
256035	256987	}
256036	256988	sqlite3Fts5TermsetFree(ctx.pTermset);
256037	256989	ctx.pTermset = 0;
256038	256990
		@@ -256454,11 +257406,11 @@
256454	257406
256455	257407	#define READ_UTF8(zIn, zTerm, c) \
256456	257408	c = *(zIn++); \
256457	257409	if( c>=0xc0 ){ \
256458	257410	c = sqlite3Utf8Trans1[c-0xc0]; \
256459		- while( zIn!=zTerm && (*zIn & 0xc0)==0x80 ){ \
	257411	+ while( zIn<zTerm && (*zIn & 0xc0)==0x80 ){ \
256460	257412	c = (c<<6) + (0x3f & *(zIn++)); \
256461	257413	} \
256462	257414	if( c<0x80 \
256463	257415	\|\| (c&0xFFFFF800)==0xD800 \
256464	257416	\|\| (c&0xFFFFFFFE)==0xFFFE ){ c = 0xFFFD; } \
		@@ -257609,22 +258561,22 @@
257609	258561	int rc = SQLITE_OK;
257610	258562	char aBuf[32];
257611	258563	char *zOut = aBuf;
257612	258564	int ii;
257613	258565	const unsigned char zIn = (const unsigned char)pText;
257614		- const unsigned char *zEof = &zIn[nText];
257615		- u32 iCode;
	258566	+ const unsigned char *zEof = (zIn ? &zIn[nText] : 0);
	258567	+ u32 iCode = 0;
257616	258568	int aStart[3]; /* Input offset of each character in aBuf[] */
257617	258569
257618	258570	UNUSED_PARAM(unusedFlags);
257619	258571
257620	258572	/* Populate aBuf[] with the characters for the first trigram. */
257621	258573	for(ii=0; ii<3; ii++){
257622	258574	do {
257623	258575	aStart[ii] = zIn - (const unsigned char*)pText;
	258576	+ if( zIn>=zEof ) return SQLITE_OK;
257624	258577	READ_UTF8(zIn, zEof, iCode);
257625		- if( iCode==0 ) return SQLITE_OK;
257626	258578	if( p->bFold ) iCode = sqlite3Fts5UnicodeFold(iCode, p->iFoldParam);
257627	258579	}while( iCode==0 );
257628	258580	WRITE_UTF8(zOut, iCode);
257629	258581	}
257630	258582
		@@ -257641,12 +258593,15 @@
257641	258593	const char *z1;
257642	258594
257643	258595	/* Read characters from the input up until the first non-diacritic */
257644	258596	do {
257645	258597	iNext = zIn - (const unsigned char*)pText;
	258598	+ if( zIn>=zEof ){
	258599	+ iCode = 0;
	258600	+ break;
	258601	+ }
257646	258602	READ_UTF8(zIn, zEof, iCode);
257647		- if( iCode==0 ) break;
257648	258603	if( p->bFold ) iCode = sqlite3Fts5UnicodeFold(iCode, p->iFoldParam);
257649	258604	}while( iCode==0 );
257650	258605
257651	258606	/* Pass the current trigram back to fts5 */
257652	258607	rc = xToken(pCtx, 0, aBuf, zOut-aBuf, aStart[0], iNext);
		@@ -259679,11 +260634,11 @@
259679	260634
259680	260635	return sqlite3_create_module_v2(db, "fts5vocab", &fts5Vocab, p, 0);
259681	260636	}
259682	260637
259683	260638
259684		-
	260639	+/* Here ends the fts5.c composite file. */
259685	260640	#endif /* !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_FTS5) */
259686	260641
259687	260642	/************ End of fts5.c **********************************************/
259688	260643	/************ Begin file stmt.c ******************************************/
259689	260644	/*
		@@ -260035,6 +260990,7 @@
260035	260990	#endif /* !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_STMTVTAB) */
260036	260991
260037	260992	/************ End of stmt.c **********************************************/
260038	260993	/* Return the source-id for this library */
260039	260994	SQLITE_API const char *sqlite3_sourceid(void){ return SQLITE_SOURCE_ID; }
	260995	+#endif /* SQLITE_AMALGAMATION */
260040	260996	/************************ End of sqlite3.c ****************************/
260041	260997

	--- extsrc/sqlite3.c
	+++ extsrc/sqlite3.c
	@@ -1,8 +1,8 @@
1	/******************************************************************************
2	** This file is an amalgamation of many separate C source files from SQLite
3	** version 3.47.0. By combining all the individual C code files into this
4	** single large file, the entire code can be compiled as a single translation
5	** unit. This allows many compilers to do optimizations that would not be
6	** possible if the files were compiled separately. Performance improvements
7	** of 5% or more are commonly seen when SQLite is compiled as a single
8	** translation unit.
	@@ -16,12 +16,15 @@
16	** if you want a wrapper to interface SQLite with your choice of programming
17	** language. The code for the "sqlite3" command-line shell is also in a
18	** separate file. This file contains only code for the core SQLite library.
19	**
20	** The content in this amalgamation comes from Fossil check-in
21	** 7891a266c4425722ae8b9231397ef9e42e24.


22	*/

23	#define SQLITE_CORE 1
24	#define SQLITE_AMALGAMATION 1
25	#ifndef SQLITE_PRIVATE
26	# define SQLITE_PRIVATE static
27	#endif
	@@ -460,13 +463,13 @@
460	**
461	** See also: [sqlite3_libversion()],
462	** [sqlite3_libversion_number()], [sqlite3_sourceid()],
463	** [sqlite_version()] and [sqlite_source_id()].
464	*/
465	#define SQLITE_VERSION "3.47.0"
466	#define SQLITE_VERSION_NUMBER 3047000
467	#define SQLITE_SOURCE_ID "2024-09-02 21:59:31 7891a266c4425722ae8b9231397ef9e42e2432be9e6b70632dfaf9ff15300d2c"
468
469	/*
470	** CAPI3REF: Run-Time Library Version Numbers
471	** KEYWORDS: sqlite3_version sqlite3_sourceid
472	**
	@@ -966,10 +969,17 @@
966	**
967	** The SQLITE_IOCAP_BATCH_ATOMIC property means that the underlying
968	** filesystem supports doing multiple write operations atomically when those
969	** write operations are bracketed by [SQLITE_FCNTL_BEGIN_ATOMIC_WRITE] and
970	** [SQLITE_FCNTL_COMMIT_ATOMIC_WRITE].







971	*/
972	#define SQLITE_IOCAP_ATOMIC 0x00000001
973	#define SQLITE_IOCAP_ATOMIC512 0x00000002
974	#define SQLITE_IOCAP_ATOMIC1K 0x00000004
975	#define SQLITE_IOCAP_ATOMIC2K 0x00000008
	@@ -982,10 +992,11 @@
982	#define SQLITE_IOCAP_SEQUENTIAL 0x00000400
983	#define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN 0x00000800
984	#define SQLITE_IOCAP_POWERSAFE_OVERWRITE 0x00001000
985	#define SQLITE_IOCAP_IMMUTABLE 0x00002000
986	#define SQLITE_IOCAP_BATCH_ATOMIC 0x00004000

987
988	/*
989	** CAPI3REF: File Locking Levels
990	**
991	** SQLite uses one of these integer values as the second
	@@ -1128,10 +1139,11 @@
1128	** <li> [SQLITE_IOCAP_SEQUENTIAL]
1129	** <li> [SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN]
1130	** <li> [SQLITE_IOCAP_POWERSAFE_OVERWRITE]
1131	** <li> [SQLITE_IOCAP_IMMUTABLE]
1132	** <li> [SQLITE_IOCAP_BATCH_ATOMIC]

1133	** </ul>
1134	**
1135	** The SQLITE_IOCAP_ATOMIC property means that all writes of
1136	** any size are atomic. The SQLITE_IOCAP_ATOMICnnn values
1137	** mean that writes of blocks that are nnn bytes in size and
	@@ -1405,10 +1417,15 @@
1405	** The [SQLITE_FCNTL_WIN32_SET_HANDLE] opcode is used for debugging. This
1406	** opcode causes the xFileControl method to swap the file handle with the one
1407	** pointed to by the pArg argument. This capability is used during testing
1408	** and only needs to be supported when SQLITE_TEST is defined.
1409	**





1410	** <li>[[SQLITE_FCNTL_WAL_BLOCK]]
1411	** The [SQLITE_FCNTL_WAL_BLOCK] is a signal to the VFS layer that it might
1412	** be advantageous to block on the next WAL lock if the lock is not immediately
1413	** available. The WAL subsystem issues this signal during rare
1414	** circumstances in order to fix a problem with priority inversion.
	@@ -1558,10 +1575,11 @@
1558	#define SQLITE_FCNTL_RESERVE_BYTES 38
1559	#define SQLITE_FCNTL_CKPT_START 39
1560	#define SQLITE_FCNTL_EXTERNAL_READER 40
1561	#define SQLITE_FCNTL_CKSM_FILE 41
1562	#define SQLITE_FCNTL_RESET_CACHE 42

1563
1564	/* deprecated names */
1565	#define SQLITE_GET_LOCKPROXYFILE SQLITE_FCNTL_GET_LOCKPROXYFILE
1566	#define SQLITE_SET_LOCKPROXYFILE SQLITE_FCNTL_SET_LOCKPROXYFILE
1567	#define SQLITE_LAST_ERRNO SQLITE_FCNTL_LAST_ERRNO
	@@ -2936,14 +2954,18 @@
2936	**
2937	** ^These functions return the number of rows modified, inserted or
2938	** deleted by the most recently completed INSERT, UPDATE or DELETE
2939	** statement on the database connection specified by the only parameter.
2940	** The two functions are identical except for the type of the return value
2941	** and that if the number of rows modified by the most recent INSERT, UPDATE
2942	** or DELETE is greater than the maximum value supported by type "int", then
2943	** the return value of sqlite3_changes() is undefined. ^Executing any other
2944	** type of SQL statement does not modify the value returned by these functions.




2945	**
2946	** ^Only changes made directly by the INSERT, UPDATE or DELETE statement are
2947	** considered - auxiliary changes caused by [CREATE TRIGGER \| triggers],
2948	** [foreign key actions] or [REPLACE] constraint resolution are not counted.
2949	**
	@@ -4499,15 +4521,26 @@
4499	**
4500	** [[SQLITE_PREPARE_NO_VTAB]] <dt>SQLITE_PREPARE_NO_VTAB</dt>
4501	** <dd>The SQLITE_PREPARE_NO_VTAB flag causes the SQL compiler
4502	** to return an error (error code SQLITE_ERROR) if the statement uses
4503	** any virtual tables.










4504	** </dl>
4505	*/
4506	#define SQLITE_PREPARE_PERSISTENT 0x01
4507	#define SQLITE_PREPARE_NORMALIZE 0x02
4508	#define SQLITE_PREPARE_NO_VTAB 0x04

4509
4510	/*
4511	** CAPI3REF: Compiling An SQL Statement
4512	** KEYWORDS: {SQL statement compiler}
4513	** METHOD: sqlite3
	@@ -4536,17 +4569,21 @@
4536	** and sqlite3_prepare_v3()
4537	** interfaces use UTF-8, and sqlite3_prepare16(), sqlite3_prepare16_v2(),
4538	** and sqlite3_prepare16_v3() use UTF-16.
4539	**
4540	** ^If the nByte argument is negative, then zSql is read up to the
4541	** first zero terminator. ^If nByte is positive, then it is the
4542	** number of bytes read from zSql. ^If nByte is zero, then no prepared


4543	** statement is generated.
4544	** If the caller knows that the supplied string is nul-terminated, then
4545	** there is a small performance advantage to passing an nByte parameter that
4546	** is the number of bytes in the input string <i>including</i>
4547	** the nul-terminator.


4548	**
4549	** ^If pzTail is not NULL then *pzTail is made to point to the first byte
4550	** past the end of the first SQL statement in zSql. These routines only
4551	** compile the first statement in zSql, so *pzTail is left pointing to
4552	** what remains uncompiled.
	@@ -5913,11 +5950,11 @@
5913	** The SQLITE_SUBTYPE flag indicates to SQLite that a function might call
5914	** [sqlite3_value_subtype()] to inspect the sub-types of its arguments.
5915	** This flag instructs SQLite to omit some corner-case optimizations that
5916	** might disrupt the operation of the [sqlite3_value_subtype()] function,
5917	** causing it to return zero rather than the correct subtype().
5918	** SQL functions that invokes [sqlite3_value_subtype()] should have this
5919	** property. If the SQLITE_SUBTYPE property is omitted, then the return
5920	** value from [sqlite3_value_subtype()] might sometimes be zero even though
5921	** a non-zero subtype was specified by the function argument expression.
5922	**
5923	** [[SQLITE_RESULT_SUBTYPE]] <dt>SQLITE_RESULT_SUBTYPE</dt><dd>
	@@ -8678,11 +8715,11 @@
8678	#define SQLITE_TESTCTRL_EXTRA_SCHEMA_CHECKS 29
8679	#define SQLITE_TESTCTRL_SEEK_COUNT 30
8680	#define SQLITE_TESTCTRL_TRACEFLAGS 31
8681	#define SQLITE_TESTCTRL_TUNE 32
8682	#define SQLITE_TESTCTRL_LOGEST 33
8683	#define SQLITE_TESTCTRL_USELONGDOUBLE 34
8684	#define SQLITE_TESTCTRL_LAST 34 /* Largest TESTCTRL */
8685
8686	/*
8687	** CAPI3REF: SQL Keyword Checking
8688	**
	@@ -9654,10 +9691,20 @@
9654	** threads may safely make multiple concurrent calls to sqlite3_backup_step().
9655	** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
9656	** APIs are not strictly speaking threadsafe. If they are invoked at the
9657	** same time as another thread is invoking sqlite3_backup_step() it is
9658	** possible that they return invalid values.










9659	*/
9660	SQLITE_API sqlite3_backup *sqlite3_backup_init(
9661	sqlite3 pDest, / Destination database handle */
9662	const char zDestName, / Destination database name */
9663	sqlite3 pSource, / Source database handle */
	@@ -10852,10 +10899,18 @@
10852	** schema S in database connection D. ^On success, the
10853	** [sqlite3_snapshot_get(D,S,P)] interface writes a pointer to the newly
10854	** created [sqlite3_snapshot] object into *P and returns SQLITE_OK.
10855	** If there is not already a read-transaction open on schema S when
10856	** this function is called, one is opened automatically.








10857	**
10858	** The following must be true for this function to succeed. If any of
10859	** the following statements are false when sqlite3_snapshot_get() is
10860	** called, SQLITE_ERROR is returned. The final value of *P is undefined
10861	** in this case.
	@@ -11172,11 +11227,11 @@
11172	#endif
11173
11174	#if 0
11175	} /* End of the 'extern "C"' block */
11176	#endif
11177	#endif /* SQLITE3_H */
11178
11179	/****** Begin file sqlite3rtree.h *******/
11180	/*
11181	** 2010 August 30
11182	**
	@@ -13423,17 +13478,32 @@
13423	** This is used to access token iToken of phrase hit iIdx within the
13424	** current row. If iIdx is less than zero or greater than or equal to the
13425	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
13426	** output variable (*ppToken) is set to point to a buffer containing the
13427	** matching document token, and (*pnToken) to the size of that buffer in
13428	** bytes. This API is not available if the specified token matches a
13429	** prefix query term. In that case both output variables are always set
13430	** to 0.
13431	**
13432	** The output text is not a copy of the document text that was tokenized.
13433	** It is the output of the tokenizer module. For tokendata=1 tables, this
13434	** includes any embedded 0x00 and trailing data.

















13435	**
13436	** This API can be quite slow if used with an FTS5 table created with the
13437	** "detail=none" or "detail=column" option.
13438	**
13439	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
	@@ -13521,11 +13591,10 @@
13521	** CUSTOM TOKENIZERS
13522	**
13523	** Applications may also register custom tokenizer types. A tokenizer
13524	** is registered by providing fts5 with a populated instance of the
13525	** following structure. All structure methods must be defined, setting
13526	**
13527	** any member of the fts5_tokenizer struct to NULL leads to undefined
13528	** behaviour. The structure methods are expected to function as follows:
13529	**
13530	** xCreate:
13531	** This function is used to allocate and initialize a tokenizer instance.
	@@ -13865,10 +13934,11 @@
13865	#endif
13866
13867	#endif /* _FTS5_H */
13868
13869	/****** End of fts5.h *******/

13870
13871	/************ End of sqlite3.h *******************************************/
13872	/************ Continuing where we left off in sqliteInt.h ****************/
13873
13874	/*
	@@ -13910,10 +13980,11 @@
13910	** to count the size: 2^31-1 or 2147483647.
13911	*/
13912	#ifndef SQLITE_MAX_LENGTH
13913	# define SQLITE_MAX_LENGTH 1000000000
13914	#endif

13915
13916	/*
13917	** This is the maximum number of
13918	**
13919	** * Columns in a table
	@@ -14809,10 +14880,11 @@
14809	#include <stdio.h>
14810	#include <stdlib.h>
14811	#include <string.h>
14812	#include <assert.h>
14813	#include <stddef.h>

14814
14815	/*
14816	** Use a macro to replace memcpy() if compiled with SQLITE_INLINE_MEMCPY.
14817	** This allows better measurements of where memcpy() is used when running
14818	** cachegrind. But this macro version of memcpy() is very slow so it
	@@ -14831,11 +14903,10 @@
14831	#ifdef SQLITE_OMIT_FLOATING_POINT
14832	# define double sqlite_int64
14833	# define float sqlite_int64
14834	# define fabs(X) ((X)<0?-(X):(X))
14835	# define sqlite3IsOverflow(X) 0
14836	# define LONGDOUBLE_TYPE sqlite_int64
14837	# ifndef SQLITE_BIG_DBL
14838	# define SQLITE_BIG_DBL (((sqlite3_int64)1)<<50)
14839	# endif
14840	# define SQLITE_OMIT_DATETIME_FUNCS 1
14841	# define SQLITE_OMIT_TRACE 1
	@@ -15006,13 +15077,10 @@
15006	# define INT8_TYPE int8_t
15007	# else
15008	# define INT8_TYPE signed char
15009	# endif
15010	#endif
15011	#ifndef LONGDOUBLE_TYPE
15012	# define LONGDOUBLE_TYPE long double
15013	#endif
15014	typedef sqlite_int64 i64; /* 8-byte signed integer */
15015	typedef sqlite_uint64 u64; /* 8-byte unsigned integer */
15016	typedef UINT32_TYPE u32; /* 4-byte unsigned integer */
15017	typedef UINT16_TYPE u16; /* 2-byte unsigned integer */
15018	typedef INT16_TYPE i16; /* 2-byte signed integer */
	@@ -16391,10 +16459,13 @@
16391	struct KeyInfo, / First argument to compare function */
16392	BtCursor pCursor / Space to write cursor structure */
16393	);
16394	SQLITE_PRIVATE BtCursor *sqlite3BtreeFakeValidCursor(void);
16395	SQLITE_PRIVATE int sqlite3BtreeCursorSize(void);



16396	SQLITE_PRIVATE void sqlite3BtreeCursorZero(BtCursor*);
16397	SQLITE_PRIVATE void sqlite3BtreeCursorHintFlags(BtCursor*, unsigned);
16398	#ifdef SQLITE_ENABLE_CURSOR_HINTS
16399	SQLITE_PRIVATE void sqlite3BtreeCursorHint(BtCursor*, int, ...);
16400	#endif
	@@ -17004,11 +17075,11 @@
17004
17005	/*
17006	** Additional non-public SQLITE_PREPARE_* flags
17007	*/
17008	#define SQLITE_PREPARE_SAVESQL 0x80 /* Preserve SQL text */
17009	#define SQLITE_PREPARE_MASK 0x0f /* Mask of public flags */
17010
17011	/*
17012	** Prototypes for the VDBE interface. See comments on the implementation
17013	** for a description of what each of these routines does.
17014	*/
	@@ -17719,51 +17790,15 @@
17719	struct FuncDefHash {
17720	FuncDef a[SQLITE_FUNC_HASH_SZ]; / Hash table for functions */
17721	};
17722	#define SQLITE_FUNC_HASH(C,L) (((C)+(L))%SQLITE_FUNC_HASH_SZ)
17723
17724	#if defined(SQLITE_USER_AUTHENTICATION)
17725	# warning "The SQLITE_USER_AUTHENTICATION extension is deprecated. \
17726	See ext/userauth/user-auth.txt for details."
17727	#endif
17728	#ifdef SQLITE_USER_AUTHENTICATION
17729	/*
17730	** Information held in the "sqlite3" database connection object and used
17731	** to manage user authentication.
17732	*/
17733	typedef struct sqlite3_userauth sqlite3_userauth;
17734	struct sqlite3_userauth {
17735	u8 authLevel; /* Current authentication level */
17736	int nAuthPW; /* Size of the zAuthPW in bytes */
17737	char zAuthPW; / Password used to authenticate */
17738	char zAuthUser; / User name used to authenticate */
17739	};
17740
17741	/* Allowed values for sqlite3_userauth.authLevel */
17742	#define UAUTH_Unknown 0 /* Authentication not yet checked */
17743	#define UAUTH_Fail 1 /* User authentication failed */
17744	#define UAUTH_User 2 /* Authenticated as a normal user */
17745	#define UAUTH_Admin 3 /* Authenticated as an administrator */
17746
17747	/* Functions used only by user authorization logic */
17748	SQLITE_PRIVATE int sqlite3UserAuthTable(const char*);
17749	SQLITE_PRIVATE int sqlite3UserAuthCheckLogin(sqlite3,const char,u8*);
17750	SQLITE_PRIVATE void sqlite3UserAuthInit(sqlite3*);
17751	SQLITE_PRIVATE void sqlite3CryptFunc(sqlite3_context,int,sqlite3_value*);
17752
17753	#endif /* SQLITE_USER_AUTHENTICATION */
17754
17755	/*
17756	** typedef for the authorization callback function.
17757	*/
17758	#ifdef SQLITE_USER_AUTHENTICATION
17759	typedef int (sqlite3_xauth)(void,int,const char,const char,const char*,
17760	const char, const char);
17761	#else
17762	typedef int (sqlite3_xauth)(void,int,const char,const char,const char*,
17763	const char*);
17764	#endif
17765
17766	#ifndef SQLITE_OMIT_DEPRECATED
17767	/* This is an extra SQLITE_TRACE macro that indicates "legacy" tracing
17768	** in the style of sqlite3_trace()
17769	*/
	@@ -17920,13 +17955,10 @@
17920	sqlite3 pUnlockConnection; / Connection to watch for unlock */
17921	void pUnlockArg; / Argument to xUnlockNotify */
17922	void (xUnlockNotify)(void , int); / Unlock notify callback */
17923	sqlite3 pNextBlocked; / Next in list of all blocked connections */
17924	#endif
17925	#ifdef SQLITE_USER_AUTHENTICATION
17926	sqlite3_userauth auth; /* User authentication information */
17927	#endif
17928	};
17929
17930	/*
17931	** A macro to discover the encoding of a database.
17932	*/
	@@ -19221,11 +19253,11 @@
19221	#define EP_Quoted 0x4000000 /* TK_ID was originally quoted */
19222	#define EP_Static 0x8000000 /* Held in memory not obtained from malloc() */
19223	#define EP_IsTrue 0x10000000 /* Always has boolean value of TRUE */
19224	#define EP_IsFalse 0x20000000 /* Always has boolean value of FALSE */
19225	#define EP_FromDDL 0x40000000 /* Originates from sqlite_schema */
19226	/* 0x80000000 // Available */
19227
19228	/* The EP_Propagate mask is a set of properties that automatically propagate
19229	** upwards into parent nodes.
19230	*/
19231	#define EP_Propagate (EP_Collate\|EP_Subquery\|EP_HasFunc)
	@@ -19777,11 +19809,11 @@
19777	**
19778	** SRT_Set The result must be a single column. Store each
19779	** row of result as the key in table pDest->iSDParm.
19780	** Apply the affinity pDest->affSdst before storing
19781	** results. if pDest->iSDParm2 is positive, then it is
19782	** a regsiter holding a Bloom filter for the IN operator
19783	** that should be populated in addition to the
19784	** pDest->iSDParm table. This SRT is used to
19785	** implement "IN (SELECT ...)".
19786	**
19787	** SRT_EphemTab Create an temporary table pDest->iSDParm and store
	@@ -20376,11 +20408,10 @@
20376	u8 bFullMutex; /* True to enable full mutexing */
20377	u8 bOpenUri; /* True to interpret filenames as URIs */
20378	u8 bUseCis; /* Use covering indices for full-scans */
20379	u8 bSmallMalloc; /* Avoid large memory allocations if true */
20380	u8 bExtraSchemaChecks; /* Verify type,name,tbl_name in schema */
20381	u8 bUseLongDouble; /* Make use of long double */
20382	#ifdef SQLITE_DEBUG
20383	u8 bJsonSelfcheck; /* Double-check JSON parsing */
20384	#endif
20385	int mxStrlen; /* Maximum string length */
20386	int neverCorrupt; /* Database is always well-formed */
	@@ -20751,19 +20782,10 @@
20751	*/
20752	#if defined(SQLITE_ENABLE_FTS4) && !defined(SQLITE_ENABLE_FTS3)
20753	# define SQLITE_ENABLE_FTS3 1
20754	#endif
20755
20756	/*
20757	** The ctype.h header is needed for non-ASCII systems. It is also
20758	** needed by FTS3 when FTS3 is included in the amalgamation.
20759	*/
20760	#if !defined(SQLITE_ASCII) \|\| \
20761	(defined(SQLITE_ENABLE_FTS3) && defined(SQLITE_AMALGAMATION))
20762	# include <ctype.h>
20763	#endif
20764
20765	/*
20766	** The following macros mimic the standard library functions toupper(),
20767	** isspace(), isalnum(), isdigit() and isxdigit(), respectively. The
20768	** sqlite versions only work for ASCII characters, regardless of locale.
20769	*/
	@@ -21381,11 +21403,11 @@
21381	SQLITE_PRIVATE int sqlite3AtoF(const char z, double, int, u8);
21382	SQLITE_PRIVATE int sqlite3GetInt32(const char , int);
21383	SQLITE_PRIVATE int sqlite3GetUInt32(const char, u32);
21384	SQLITE_PRIVATE int sqlite3Atoi(const char*);
21385	#ifndef SQLITE_OMIT_UTF16
21386	SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *pData, int nChar);
21387	#endif
21388	SQLITE_PRIVATE int sqlite3Utf8CharLen(const char *pData, int nByte);
21389	SQLITE_PRIVATE u32 sqlite3Utf8Read(const u8**);
21390	SQLITE_PRIVATE int sqlite3Utf8ReadLimited(const u8, int, u32);
21391	SQLITE_PRIVATE LogEst sqlite3LogEst(u64);
	@@ -22839,13 +22861,10 @@
22839	"UNLINK_AFTER_CLOSE",
22840	#endif
22841	#ifdef SQLITE_UNTESTABLE
22842	"UNTESTABLE",
22843	#endif
22844	#ifdef SQLITE_USER_AUTHENTICATION
22845	"USER_AUTHENTICATION",
22846	#endif
22847	#ifdef SQLITE_USE_ALLOCA
22848	"USE_ALLOCA",
22849	#endif
22850	#ifdef SQLITE_USE_FCNTL_TRACE
22851	"USE_FCNTL_TRACE",
	@@ -23117,11 +23136,10 @@
23117	SQLITE_THREADSAFE==1, /* bFullMutex */
23118	SQLITE_USE_URI, /* bOpenUri */
23119	SQLITE_ALLOW_COVERING_INDEX_SCAN, /* bUseCis */
23120	0, /* bSmallMalloc */
23121	1, /* bExtraSchemaChecks */
23122	sizeof(LONGDOUBLE_TYPE)>8, /* bUseLongDouble */
23123	#ifdef SQLITE_DEBUG
23124	0, /* bJsonSelfcheck */
23125	#endif
23126	0x7ffffffe, /* mxStrlen */
23127	0, /* neverCorrupt */
	@@ -23837,13 +23855,15 @@
23837	UnpackedRecord pNewUnpacked; / Unpacked version of new.* record */
23838	int iNewReg; /* Register for new.* values */
23839	int iBlobWrite; /* Value returned by preupdate_blobwrite() */
23840	i64 iKey1; /* First key value passed to hook */
23841	i64 iKey2; /* Second key value passed to hook */

23842	Mem aNew; / Array of new.* values */
23843	Table pTab; / Schema object being updated */
23844	Index pPk; / PK index if pTab is WITHOUT ROWID */

23845	};
23846
23847	/*
23848	** An instance of this object is used to pass an vector of values into
23849	** OP_VFilter, the xFilter method of a virtual table. The vector is the
	@@ -29285,20 +29305,33 @@
29285
29286	#ifndef NDEBUG
29287	/*
29288	** The sqlite3_mutex_held() and sqlite3_mutex_notheld() routine are
29289	** intended for use inside assert() statements.



29290	*/
29291	SQLITE_API int sqlite3_mutex_held(sqlite3_mutex *p){





29292	assert( p==0 \|\| sqlite3GlobalConfig.mutex.xMutexHeld );
29293	return p==0 \|\| sqlite3GlobalConfig.mutex.xMutexHeld(p);
29294	}
29295	SQLITE_API int sqlite3_mutex_notheld(sqlite3_mutex *p){





29296	assert( p==0 \|\| sqlite3GlobalConfig.mutex.xMutexNotheld );
29297	return p==0 \|\| sqlite3GlobalConfig.mutex.xMutexNotheld(p);
29298	}
29299	#endif
29300
29301	#endif /* !defined(SQLITE_MUTEX_OMIT) */
29302
29303	/************ End of mutex.c *********************************************/
29304	/************ Begin file mutex_noop.c ************************************/
	@@ -32272,10 +32305,11 @@
32272	&& (ExprHasProperty(pExpr,EP_OuterON\|EP_InnerON) \|\| pExpr->w.iOfst<=0)
32273	){
32274	pExpr = pExpr->pLeft;
32275	}
32276	if( pExpr==0 ) return;

32277	db->errByteOffset = pExpr->w.iOfst;
32278	}
32279
32280	/*
32281	** Enlarge the memory allocation on a StrAccum object so that it is
	@@ -33001,11 +33035,11 @@
33001	}
33002	if( pItem->fg.isCte ){
33003	sqlite3_str_appendf(&x, " CteUse=0x%p", pItem->u2.pCteUse);
33004	}
33005	if( pItem->fg.isOn \|\| (pItem->fg.isUsing==0 && pItem->u3.pOn!=0) ){
33006	sqlite3_str_appendf(&x, " ON");
33007	}
33008	if( pItem->fg.isTabFunc ) sqlite3_str_appendf(&x, " isTabFunc");
33009	if( pItem->fg.isCorrelated ) sqlite3_str_appendf(&x, " isCorrelated");
33010	if( pItem->fg.isMaterialized ) sqlite3_str_appendf(&x, " isMaterialized");
33011	if( pItem->fg.viaCoroutine ) sqlite3_str_appendf(&x, " viaCoroutine");
	@@ -34085,10 +34119,14 @@
34085	**
34086	** This routines are given external linkage so that they will always be
34087	** accessible to the debugging, and to avoid warnings about unused
34088	** functions. But these routines only exist in debugging builds, so they
34089	** do not contaminate the interface.




34090	*/
34091	SQLITE_PRIVATE void sqlite3ShowExpr(const Expr *p){ sqlite3TreeViewExpr(0,p,0); }
34092	SQLITE_PRIVATE void sqlite3ShowExprList(const ExprList *p){ sqlite3TreeViewExprList(0,p,0,0);}
34093	SQLITE_PRIVATE void sqlite3ShowIdList(const IdList *p){ sqlite3TreeViewIdList(0,p,0,0); }
34094	SQLITE_PRIVATE void sqlite3ShowSrcList(const SrcList *p){ sqlite3TreeViewSrcList(0,p); }
	@@ -34687,11 +34725,11 @@
34687	*/
34688	#define READ_UTF8(zIn, zTerm, c) \
34689	c = *(zIn++); \
34690	if( c>=0xc0 ){ \
34691	c = sqlite3Utf8Trans1[c-0xc0]; \
34692	while( zIn!=zTerm && (*zIn & 0xc0)==0x80 ){ \
34693	c = (c<<6) + (0x3f & *(zIn++)); \
34694	} \
34695	if( c<0x80 \
34696	\|\| (c&0xFFFFF800)==0xD800 \
34697	\|\| (c&0xFFFFFFFE)==0xFFFE ){ c = 0xFFFD; } \
	@@ -35065,24 +35103,26 @@
35065	assert( m.z \|\| db->mallocFailed );
35066	return m.z;
35067	}
35068
35069	/*
35070	** zIn is a UTF-16 encoded unicode string at least nChar characters long.
35071	** Return the number of bytes in the first nChar unicode characters
35072	** in pZ. nChar must be non-negative.

35073	*/
35074	SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *zIn, int nChar){
35075	int c;
35076	unsigned char const *z = zIn;

35077	int n = 0;
35078
35079	if( SQLITE_UTF16NATIVE==SQLITE_UTF16LE ) z++;
35080	while( n<nChar ){
35081	c = z[0];
35082	z += 2;
35083	if( c>=0xd8 && c<0xdc && z[0]>=0xdc && z[0]<0xe0 ) z += 2;
35084	n++;
35085	}
35086	return (int)(z-(unsigned char const *)zIn)
35087	- (SQLITE_UTF16NATIVE==SQLITE_UTF16LE);
35088	}
	@@ -35659,10 +35699,12 @@
35659	int esign = 1; /* sign of exponent */
35660	int e = 0; /* exponent */
35661	int eValid = 1; /* True exponent is either not used or is well-formed */
35662	int nDigit = 0; /* Number of digits processed */
35663	int eType = 1; /* 1: pure integer, 2+: fractional -1 or less: bad UTF16 */


35664
35665	assert( enc==SQLITE_UTF8 \|\| enc==SQLITE_UTF16LE \|\| enc==SQLITE_UTF16BE );
35666	pResult = 0.0; / Default return value, in case of an error */
35667	if( length==0 ) return 0;
35668
	@@ -35761,81 +35803,65 @@
35761
35762	/* adjust exponent by d, and update sign */
35763	e = (e*esign) + d;
35764
35765	/* Try to adjust the exponent to make it smaller */
35766	while( e>0 && s<(LARGEST_UINT64/10) ){
35767	s *= 10;
35768	e--;
35769	}
35770	while( e<0 && (s%10)==0 ){
35771	s /= 10;
35772	e++;
35773	}
35774
35775	if( e==0 ){
35776	*pResult = s;
35777	}else if( sqlite3Config.bUseLongDouble ){
35778	LONGDOUBLE_TYPE r = (LONGDOUBLE_TYPE)s;
35779	if( e>0 ){
35780	while( e>=100 ){ e-=100; r *= 1.0e+100L; }
35781	while( e>=10 ){ e-=10; r *= 1.0e+10L; }
35782	while( e>=1 ){ e-=1; r *= 1.0e+01L; }
35783	}else{
35784	while( e<=-100 ){ e+=100; r *= 1.0e-100L; }
35785	while( e<=-10 ){ e+=10; r *= 1.0e-10L; }
35786	while( e<=-1 ){ e+=1; r *= 1.0e-01L; }
35787	}
35788	assert( r>=0.0 );
35789	if( r>+1.7976931348623157081452742373e+308L ){
35790	#ifdef INFINITY
35791	*pResult = +INFINITY;
35792	#else
35793	pResult = 1.0e30810.0;
35794	#endif
35795	}else{
35796	*pResult = (double)r;
35797	}
35798	}else{
35799	double rr[2];
35800	u64 s2;
35801	rr[0] = (double)s;
35802	s2 = (u64)rr[0];
35803	#if defined(_MSC_VER) && _MSC_VER<1700
35804	if( s2==0x8000000000000000LL ){ s2 = 2(u64)(0.5rr[0]); }
35805	#endif
35806	rr[1] = s>=s2 ? (double)(s - s2) : -(double)(s2 - s);
35807	if( e>0 ){
35808	while( e>=100 ){
35809	e -= 100;
35810	dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
35811	}
35812	while( e>=10 ){
35813	e -= 10;
35814	dekkerMul2(rr, 1.0e+10, 0.0);
35815	}
35816	while( e>=1 ){
35817	e -= 1;
35818	dekkerMul2(rr, 1.0e+01, 0.0);
35819	}
35820	}else{
35821	while( e<=-100 ){
35822	e += 100;
35823	dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
35824	}
35825	while( e<=-10 ){
35826	e += 10;
35827	dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
35828	}
35829	while( e<=-1 ){
35830	e += 1;
35831	dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
35832	}
35833	}
35834	*pResult = rr[0]+rr[1];
35835	if( sqlite3IsNaN(pResult) ) pResult = 1e300*1e300;
35836	}




35837	if( sign<0 ) pResult = -pResult;
35838	assert( !sqlite3IsNaN(*pResult) );
35839
35840	atof_return:
35841	/* return true if number and no extra non-whitespace characters after */
	@@ -36152,13 +36178,14 @@
36152	*/
36153	SQLITE_PRIVATE void sqlite3FpDecode(FpDecode *p, double r, int iRound, int mxRound){
36154	int i;
36155	u64 v;
36156	int e, exp = 0;


36157	p->isSpecial = 0;
36158	p->z = p->zBuf;
36159
36160	assert( mxRound>0 );
36161
36162	/* Convert negative numbers to positive. Deal with Infinity, 0.0, and
36163	** NaN. */
36164	if( r<0.0 ){
	@@ -36182,66 +36209,49 @@
36182	return;
36183	}
36184
36185	/* Multiply r by powers of ten until it lands somewhere in between
36186	** 1.0e+19 and 1.0e+17.
36187	*/
36188	if( sqlite3Config.bUseLongDouble ){
36189	LONGDOUBLE_TYPE rr = r;
36190	if( rr>=1.0e+19 ){
36191	while( rr>=1.0e+119L ){ exp+=100; rr *= 1.0e-100L; }
36192	while( rr>=1.0e+29L ){ exp+=10; rr *= 1.0e-10L; }
36193	while( rr>=1.0e+19L ){ exp++; rr *= 1.0e-1L; }
36194	}else{
36195	while( rr<1.0e-97L ){ exp-=100; rr *= 1.0e+100L; }
36196	while( rr<1.0e+07L ){ exp-=10; rr *= 1.0e+10L; }
36197	while( rr<1.0e+17L ){ exp--; rr *= 1.0e+1L; }
36198	}
36199	v = (u64)rr;
36200	}else{
36201	/* If high-precision floating point is not available using "long double",
36202	** then use Dekker-style double-double computation to increase the
36203	** precision.
36204	**
36205	** The error terms on constants like 1.0e+100 computed using the
36206	** decimal extension, for example as follows:
36207	**
36208	** SELECT decimal_exp(decimal_sub('1.0e+100',decimal(1.0e+100)));
36209	*/
36210	double rr[2];
36211	rr[0] = r;
36212	rr[1] = 0.0;
36213	if( rr[0]>9.223372036854774784e+18 ){
36214	while( rr[0]>9.223372036854774784e+118 ){
36215	exp += 100;
36216	dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
36217	}
36218	while( rr[0]>9.223372036854774784e+28 ){
36219	exp += 10;
36220	dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
36221	}
36222	while( rr[0]>9.223372036854774784e+18 ){
36223	exp += 1;
36224	dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
36225	}
36226	}else{
36227	while( rr[0]<9.223372036854774784e-83 ){
36228	exp -= 100;
36229	dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
36230	}
36231	while( rr[0]<9.223372036854774784e+07 ){
36232	exp -= 10;
36233	dekkerMul2(rr, 1.0e+10, 0.0);
36234	}
36235	while( rr[0]<9.22337203685477478e+17 ){
36236	exp -= 1;
36237	dekkerMul2(rr, 1.0e+01, 0.0);
36238	}
36239	}
36240	v = rr[1]<0.0 ? (u64)rr[0]-(u64)(-rr[1]) : (u64)rr[0]+(u64)rr[1];
36241	}
36242
36243
36244	/* Extract significant digits. */
36245	i = sizeof(p->zBuf)-1;
36246	assert( v>0 );
36247	while( v ){ p->zBuf[i--] = (v%10) + '0'; v /= 10; }
	@@ -37008,108 +37018,10 @@
37008	i += pIn[i+1];
37009	}while( i<mx );
37010	return 0;
37011	}
37012
37013	/*
37014	** High-resolution hardware timer used for debugging and testing only.
37015	*/
37016	#if defined(VDBE_PROFILE) \
37017	\|\| defined(SQLITE_PERFORMANCE_TRACE) \
37018	\|\| defined(SQLITE_ENABLE_STMT_SCANSTATUS)
37019	/************ Include hwtime.h in the middle of util.c *******************/
37020	/************ Begin file hwtime.h ****************************************/
37021	/*
37022	** 2008 May 27
37023	**
37024	** The author disclaims copyright to this source code. In place of
37025	** a legal notice, here is a blessing:
37026	**
37027	** May you do good and not evil.
37028	** May you find forgiveness for yourself and forgive others.
37029	** May you share freely, never taking more than you give.
37030	**
37031	******************************************************************************
37032	**
37033	** This file contains inline asm code for retrieving "high-performance"
37034	** counters for x86 and x86_64 class CPUs.
37035	*/
37036	#ifndef SQLITE_HWTIME_H
37037	#define SQLITE_HWTIME_H
37038
37039	/*
37040	** The following routine only works on Pentium-class (or newer) processors.
37041	** It uses the RDTSC opcode to read the cycle count value out of the
37042	** processor and returns that value. This can be used for high-res
37043	** profiling.
37044	*/
37045	#if !defined(__STRICT_ANSI__) && \
37046	(defined(__GNUC__) \|\| defined(_MSC_VER)) && \
37047	(defined(i386) \|\| defined(__i386__) \|\| defined(_M_IX86))
37048
37049	#if defined(__GNUC__)
37050
37051	__inline__ sqlite_uint64 sqlite3Hwtime(void){
37052	unsigned int lo, hi;
37053	__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
37054	return (sqlite_uint64)hi << 32 \| lo;
37055	}
37056
37057	#elif defined(_MSC_VER)
37058
37059	__declspec(naked) __inline sqlite_uint64 __cdecl sqlite3Hwtime(void){
37060	__asm {
37061	rdtsc
37062	ret ; return value at EDX:EAX
37063	}
37064	}
37065
37066	#endif
37067
37068	#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__x86_64__))
37069
37070	__inline__ sqlite_uint64 sqlite3Hwtime(void){
37071	unsigned int lo, hi;
37072	__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
37073	return (sqlite_uint64)hi << 32 \| lo;
37074	}
37075
37076	#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__ppc__))
37077
37078	__inline__ sqlite_uint64 sqlite3Hwtime(void){
37079	unsigned long long retval;
37080	unsigned long junk;
37081	__asm__ __volatile__ ("\n\
37082	1: mftbu %1\n\
37083	mftb %L0\n\
37084	mftbu %0\n\
37085	cmpw %0,%1\n\
37086	bne 1b"
37087	: "=r" (retval), "=r" (junk));
37088	return retval;
37089	}
37090
37091	#else
37092
37093	/*
37094	** asm() is needed for hardware timing support. Without asm(),
37095	** disable the sqlite3Hwtime() routine.
37096	**
37097	** sqlite3Hwtime() is only used for some obscure debugging
37098	** and analysis configurations, not in any deliverable, so this
37099	** should not be a great loss.
37100	*/
37101	SQLITE_PRIVATE sqlite_uint64 sqlite3Hwtime(void){ return ((sqlite_uint64)0); }
37102
37103	#endif
37104
37105	#endif /* !defined(SQLITE_HWTIME_H) */
37106
37107	/************ End of hwtime.h ********************************************/
37108	/************ Continuing where we left off in util.c *********************/
37109	#endif
37110
37111	/************ End of util.c **********************************************/
37112	/************ Begin file hash.c ******************************************/
37113	/*
37114	** 2001 September 22
37115	**
	@@ -38787,11 +38699,11 @@
38787	# define F_SETLKW 7
38788	# endif
38789	# endif
38790	#else /* !SQLITE_WASI */
38791	# ifndef HAVE_FCHMOD
38792	# define HAVE_FCHMOD
38793	# endif
38794	#endif /* SQLITE_WASI */
38795
38796	#ifdef SQLITE_WASI
38797	# define osGetpid(X) (pid_t)1
	@@ -41038,58 +40950,37 @@
41038	** file by this or any other process. If such a lock is held, set *pResOut
41039	** to a non-zero value otherwise *pResOut is set to zero. The return value
41040	** is set to SQLITE_OK unless an I/O error occurs during lock checking.
41041	*/
41042	static int flockCheckReservedLock(sqlite3_file id, int pResOut){
41043	int rc = SQLITE_OK;
41044	int reserved = 0;
41045	unixFile pFile = (unixFile)id;



41046
41047	SimulateIOError( return SQLITE_IOERR_CHECKRESERVEDLOCK; );
41048
41049	assert( pFile );
41050
41051	/* Check if a thread in this process holds such a lock */
41052	if( pFile->eFileLock>SHARED_LOCK ){
41053	reserved = 1;
41054	}
41055
41056	/* Otherwise see if some other process holds it. */
41057	if( !reserved ){
41058	/* attempt to get the lock */
41059	int lrc = robust_flock(pFile->h, LOCK_EX \| LOCK_NB);
41060	if( !lrc ){
41061	/* got the lock, unlock it */
41062	lrc = robust_flock(pFile->h, LOCK_UN);
41063	if ( lrc ) {
41064	int tErrno = errno;
41065	/* unlock failed with an error */
41066	lrc = SQLITE_IOERR_UNLOCK;
41067	storeLastErrno(pFile, tErrno);
41068	rc = lrc;
41069	}
41070	} else {
41071	int tErrno = errno;
41072	reserved = 1;
41073	/* someone else might have it reserved */
41074	lrc = sqliteErrorFromPosixError(tErrno, SQLITE_IOERR_LOCK);
41075	if( IS_LOCK_ERROR(lrc) ){
41076	storeLastErrno(pFile, tErrno);
41077	rc = lrc;
41078	}
41079	}
41080	}
41081	OSTRACE(("TEST WR-LOCK %d %d %d (flock)\n", pFile->h, rc, reserved));
41082
41083	#ifdef SQLITE_IGNORE_FLOCK_LOCK_ERRORS
41084	if( (rc & 0xff) == SQLITE_IOERR ){
41085	rc = SQLITE_OK;
41086	reserved=1;
41087	}
41088	#endif /* SQLITE_IGNORE_FLOCK_LOCK_ERRORS */
41089	*pResOut = reserved;
41090	return rc;
41091	}
41092
41093	/*
41094	** Lock the file with the lock specified by parameter eFileLock - one
41095	** of the following:
	@@ -42582,10 +42473,15 @@
42582	int rc = osIoctl(pFile->h, F2FS_IOC_ABORT_VOLATILE_WRITE);
42583	return rc ? SQLITE_IOERR_ROLLBACK_ATOMIC : SQLITE_OK;
42584	}
42585	#endif /* __linux__ && SQLITE_ENABLE_BATCH_ATOMIC_WRITE */
42586





42587	case SQLITE_FCNTL_LOCKSTATE: {
42588	(int)pArg = pFile->eFileLock;
42589	return SQLITE_OK;
42590	}
42591	case SQLITE_FCNTL_LAST_ERRNO: {
	@@ -42723,10 +42619,11 @@
42723
42724	/* Set the POWERSAFE_OVERWRITE flag if requested. */
42725	if( pFd->ctrlFlags & UNIXFILE_PSOW ){
42726	pFd->deviceCharacteristics \|= SQLITE_IOCAP_POWERSAFE_OVERWRITE;
42727	}

42728
42729	pFd->sectorSize = SQLITE_DEFAULT_SECTOR_SIZE;
42730	}
42731	}
42732	#else
	@@ -42773,19 +42670,19 @@
42773	0;
42774	}else if( !strcmp(fsInfo.f_basetype, "qnx4") ){
42775	pFile->sectorSize = fsInfo.f_bsize;
42776	pFile->deviceCharacteristics =
42777	/* full bitset of atomics from max sector size and smaller */
42778	((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2 \|
42779	SQLITE_IOCAP_SEQUENTIAL \| /* The ram filesystem has no write behind
42780	** so it is ordered */
42781	0;
42782	}else if( strstr(fsInfo.f_basetype, "dos") ){
42783	pFile->sectorSize = fsInfo.f_bsize;
42784	pFile->deviceCharacteristics =
42785	/* full bitset of atomics from max sector size and smaller */
42786	((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2 \|
42787	SQLITE_IOCAP_SEQUENTIAL \| /* The ram filesystem has no write behind
42788	** so it is ordered */
42789	0;
42790	}else{
42791	pFile->deviceCharacteristics =
	@@ -50462,10 +50359,15 @@
50462	OSTRACE(("FCNTL oldFile=%p, newFile=%p, rc=SQLITE_OK\n",
50463	hOldFile, pFile->h));
50464	return SQLITE_OK;
50465	}
50466	#endif





50467	case SQLITE_FCNTL_TEMPFILENAME: {
50468	char *zTFile = 0;
50469	int rc = winGetTempname(pFile->pVfs, &zTFile);
50470	if( rc==SQLITE_OK ){
50471	(char*)pArg = zTFile;
	@@ -50523,11 +50425,11 @@
50523	/*
50524	** Return a vector of device characteristics.
50525	*/
50526	static int winDeviceCharacteristics(sqlite3_file *id){
50527	winFile p = (winFile)id;
50528	return SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN \|
50529	((p->ctrlFlags & WINFILE_PSOW)?SQLITE_IOCAP_POWERSAFE_OVERWRITE:0);
50530	}
50531
50532	/*
50533	** Windows will only let you create file view mappings
	@@ -51911,11 +51813,11 @@
51911	*/
51912	char zTmpname = 0; / For temporary filename, if necessary. */
51913
51914	int rc = SQLITE_OK; /* Function Return Code */
51915	#if !defined(NDEBUG) \|\| SQLITE_OS_WINCE
51916	int eType = flags&0xFFFFFF00; /* Type of file to open */
51917	#endif
51918
51919	int isExclusive = (flags & SQLITE_OPEN_EXCLUSIVE);
51920	int isDelete = (flags & SQLITE_OPEN_DELETEONCLOSE);
51921	int isCreate = (flags & SQLITE_OPEN_CREATE);
	@@ -58131,24 +58033,32 @@
58131	#ifdef SQLITE_DIRECT_OVERFLOW_READ
58132	/*
58133	** Return true if page pgno can be read directly from the database file
58134	** by the b-tree layer. This is the case if:
58135	**
58136	** * the database file is open,
58137	** * there are no dirty pages in the cache, and
58138	** * the desired page is not currently in the wal file.

58139	*/
58140	SQLITE_PRIVATE int sqlite3PagerDirectReadOk(Pager *pPager, Pgno pgno){
58141	if( pPager->fd->pMethods==0 ) return 0;
58142	if( sqlite3PCacheIsDirty(pPager->pPCache) ) return 0;


58143	#ifndef SQLITE_OMIT_WAL
58144	if( pPager->pWal ){
58145	u32 iRead = 0;
58146	(void)sqlite3WalFindFrame(pPager->pWal, pgno, &iRead);
58147	return iRead==0;
58148	}
58149	#endif





58150	return 1;
58151	}
58152	#endif
58153
58154	#ifndef SQLITE_OMIT_WAL
	@@ -65171,11 +65081,11 @@
65171	** 20: Salt-2, a different random integer changing with each ckpt
65172	** 24: Checksum-1 (first part of checksum for first 24 bytes of header).
65173	** 28: Checksum-2 (second part of checksum for first 24 bytes of header).
65174	**
65175	** Immediately following the wal-header are zero or more frames. Each
65176	** frame consists of a 24-byte frame-header followed by a <page-size> bytes
65177	** of page data. The frame-header is six big-endian 32-bit unsigned
65178	** integer values, as follows:
65179	**
65180	** 0: Page number.
65181	** 4: For commit records, the size of the database image in pages
	@@ -65668,10 +65578,11 @@
65668	int nSehTry; /* Number of nested SEH_TRY{} blocks */
65669	u8 lockError; /* True if a locking error has occurred */
65670	#endif
65671	#ifdef SQLITE_ENABLE_SNAPSHOT
65672	WalIndexHdr pSnapshot; / Start transaction here if not NULL */

65673	#endif
65674	#ifdef SQLITE_ENABLE_SETLK_TIMEOUT
65675	sqlite3 *db;
65676	#endif
65677	};
	@@ -67560,11 +67471,11 @@
67560	return SQLITE_IOERR_IN_PAGE;
67561	}
67562
67563	/*
67564	** Assert that the Wal.lockMask mask, which indicates the locks held
67565	** by the connenction, is consistent with the Wal.readLock, Wal.writeLock
67566	** and Wal.ckptLock variables. To be used as:
67567	**
67568	** assert( walAssertLockmask(pWal) );
67569	*/
67570	static int walAssertLockmask(Wal *pWal){
	@@ -68224,11 +68135,11 @@
68224	assert( pWal->apWiData[0]!=0 );
68225	pInfo = walCkptInfo(pWal);
68226	SEH_INJECT_FAULT;
68227	if( !useWal && AtomicLoad(&pInfo->nBackfill)==pWal->hdr.mxFrame
68228	#ifdef SQLITE_ENABLE_SNAPSHOT
68229	&& (pWal->pSnapshot==0 \|\| pWal->hdr.mxFrame==0)
68230	#endif
68231	){
68232	/* The WAL has been completely backfilled (or it is empty).
68233	** and can be safely ignored.
68234	*/
	@@ -69624,11 +69535,24 @@
69624	*/
69625	SQLITE_PRIVATE void sqlite3WalSnapshotOpen(
69626	Wal *pWal,
69627	sqlite3_snapshot *pSnapshot
69628	){
69629	pWal->pSnapshot = (WalIndexHdr*)pSnapshot;













69630	}
69631
69632	/*
69633	** Return a +ve value if snapshot p1 is newer than p2. A -ve value if
69634	** p1 is older than p2 and zero if p1 and p2 are the same snapshot.
	@@ -75504,10 +75428,29 @@
75504	** this routine.
75505	*/
75506	SQLITE_PRIVATE int sqlite3BtreeCursorSize(void){
75507	return ROUND8(sizeof(BtCursor));
75508	}



















75509
75510	/*
75511	** Initialize memory that will be converted into a BtCursor object.
75512	**
75513	** The simple approach here would be to memset() the entire object
	@@ -84538,11 +84481,12 @@
84538	if( apVal==0 ){
84539	rc = SQLITE_NOMEM_BKPT;
84540	goto value_from_function_out;
84541	}
84542	for(i=0; i<nVal; i++){
84543	rc = sqlite3ValueFromExpr(db, pList->a[i].pExpr, enc, aff, &apVal[i]);

84544	if( apVal[i]==0 \|\| rc!=SQLITE_OK ) goto value_from_function_out;
84545	}
84546	}
84547
84548	pVal = valueNew(db, pCtx);
	@@ -89571,11 +89515,11 @@
89571	/* The following two functions are used only within testcase() to prove
89572	** test coverage. These functions do no exist for production builds.
89573	** We must use separate SQLITE_NOINLINE functions here, since otherwise
89574	** optimizer code movement causes gcov to become very confused.
89575	*/
89576	#if defined(SQLITE_COVERAGE_TEST) \|\| defined(SQLITE_DEBUG)
89577	static int SQLITE_NOINLINE doubleLt(double a, double b){ return a<b; }
89578	static int SQLITE_NOINLINE doubleEq(double a, double b){ return a==b; }
89579	#endif
89580
89581	/*
	@@ -89586,17 +89530,10 @@
89586	SQLITE_PRIVATE int sqlite3IntFloatCompare(i64 i, double r){
89587	if( sqlite3IsNaN(r) ){
89588	/* SQLite considers NaN to be a NULL. And all integer values are greater
89589	** than NULL */
89590	return 1;
89591	}
89592	if( sqlite3Config.bUseLongDouble ){
89593	LONGDOUBLE_TYPE x = (LONGDOUBLE_TYPE)i;
89594	testcase( x<r );
89595	testcase( x>r );
89596	testcase( x==r );
89597	return (x<r) ? -1 : (x>r);
89598	}else{
89599	i64 y;
89600	if( r<-9223372036854775808.0 ) return +1;
89601	if( r>=9223372036854775808.0 ) return -1;
89602	y = (i64)r;
	@@ -90595,17 +90532,25 @@
90595	db->xPreUpdateCallback(db->pPreUpdateArg, db, op, zDb, zTbl, iKey1, iKey2);
90596	db->pPreUpdate = 0;
90597	sqlite3DbFree(db, preupdate.aRecord);
90598	vdbeFreeUnpacked(db, preupdate.keyinfo.nKeyField+1, preupdate.pUnpacked);
90599	vdbeFreeUnpacked(db, preupdate.keyinfo.nKeyField+1, preupdate.pNewUnpacked);

90600	if( preupdate.aNew ){
90601	int i;
90602	for(i=0; i<pCsr->nField; i++){
90603	sqlite3VdbeMemRelease(&preupdate.aNew[i]);
90604	}
90605	sqlite3DbNNFreeNN(db, preupdate.aNew);
90606	}







90607	}
90608	#endif /* SQLITE_ENABLE_PREUPDATE_HOOK */
90609
90610	/************ End of vdbeaux.c *******************************************/
90611	/************ Begin file vdbeapi.c ***************************************/
	@@ -92230,10 +92175,21 @@
92230	** A successful evaluation of this routine acquires the mutex on p.
92231	** the mutex is released if any kind of error occurs.
92232	**
92233	** The error code stored in database p->db is overwritten with the return
92234	** value in any case.











92235	*/
92236	static int vdbeUnbind(Vdbe *p, unsigned int i){
92237	Mem *pVar;
92238	if( vdbeSafetyNotNull(p) ){
92239	return SQLITE_MISUSE_BKPT;
	@@ -92287,10 +92243,11 @@
92287	Mem *pVar;
92288	int rc;
92289
92290	rc = vdbeUnbind(p, (u32)(i-1));
92291	if( rc==SQLITE_OK ){

92292	if( zData!=0 ){
92293	pVar = &p->aVar[i-1];
92294	rc = sqlite3VdbeMemSetStr(pVar, zData, nData, encoding, xDel);
92295	if( rc==SQLITE_OK && encoding!=0 ){
92296	rc = sqlite3VdbeChangeEncoding(pVar, ENC(p->db));
	@@ -92336,10 +92293,11 @@
92336	SQLITE_API int sqlite3_bind_double(sqlite3_stmt *pStmt, int i, double rValue){
92337	int rc;
92338	Vdbe p = (Vdbe )pStmt;
92339	rc = vdbeUnbind(p, (u32)(i-1));
92340	if( rc==SQLITE_OK ){

92341	sqlite3VdbeMemSetDouble(&p->aVar[i-1], rValue);
92342	sqlite3_mutex_leave(p->db->mutex);
92343	}
92344	return rc;
92345	}
	@@ -92349,10 +92307,11 @@
92349	SQLITE_API int sqlite3_bind_int64(sqlite3_stmt *pStmt, int i, sqlite_int64 iValue){
92350	int rc;
92351	Vdbe p = (Vdbe )pStmt;
92352	rc = vdbeUnbind(p, (u32)(i-1));
92353	if( rc==SQLITE_OK ){

92354	sqlite3VdbeMemSetInt64(&p->aVar[i-1], iValue);
92355	sqlite3_mutex_leave(p->db->mutex);
92356	}
92357	return rc;
92358	}
	@@ -92359,10 +92318,11 @@
92359	SQLITE_API int sqlite3_bind_null(sqlite3_stmt *pStmt, int i){
92360	int rc;
92361	Vdbe p = (Vdbe)pStmt;
92362	rc = vdbeUnbind(p, (u32)(i-1));
92363	if( rc==SQLITE_OK ){

92364	sqlite3_mutex_leave(p->db->mutex);
92365	}
92366	return rc;
92367	}
92368	SQLITE_API int sqlite3_bind_pointer(
	@@ -92374,10 +92334,11 @@
92374	){
92375	int rc;
92376	Vdbe p = (Vdbe)pStmt;
92377	rc = vdbeUnbind(p, (u32)(i-1));
92378	if( rc==SQLITE_OK ){

92379	sqlite3VdbeMemSetPointer(&p->aVar[i-1], pPtr, zPTtype, xDestructor);
92380	sqlite3_mutex_leave(p->db->mutex);
92381	}else if( xDestructor ){
92382	xDestructor(pPtr);
92383	}
	@@ -92455,10 +92416,11 @@
92455	SQLITE_API int sqlite3_bind_zeroblob(sqlite3_stmt *pStmt, int i, int n){
92456	int rc;
92457	Vdbe p = (Vdbe )pStmt;
92458	rc = vdbeUnbind(p, (u32)(i-1));
92459	if( rc==SQLITE_OK ){

92460	#ifndef SQLITE_OMIT_INCRBLOB
92461	sqlite3VdbeMemSetZeroBlob(&p->aVar[i-1], n);
92462	#else
92463	rc = sqlite3VdbeMemSetZeroBlob(&p->aVar[i-1], n);
92464	#endif
	@@ -92789,41 +92751,68 @@
92789	if( iIdx>=p->pCsr->nField \|\| iIdx<0 ){
92790	rc = SQLITE_RANGE;
92791	goto preupdate_old_out;
92792	}
92793
92794	/* If the old.* record has not yet been loaded into memory, do so now. */
92795	if( p->pUnpacked==0 ){
92796	u32 nRec;
92797	u8 *aRec;
92798
92799	assert( p->pCsr->eCurType==CURTYPE_BTREE );
92800	nRec = sqlite3BtreePayloadSize(p->pCsr->uc.pCursor);
92801	aRec = sqlite3DbMallocRaw(db, nRec);
92802	if( !aRec ) goto preupdate_old_out;
92803	rc = sqlite3BtreePayload(p->pCsr->uc.pCursor, 0, nRec, aRec);
92804	if( rc==SQLITE_OK ){
92805	p->pUnpacked = vdbeUnpackRecord(&p->keyinfo, nRec, aRec);
92806	if( !p->pUnpacked ) rc = SQLITE_NOMEM;
92807	}
92808	if( rc!=SQLITE_OK ){
92809	sqlite3DbFree(db, aRec);
92810	goto preupdate_old_out;
92811	}
92812	p->aRecord = aRec;
92813	}
92814
92815	pMem = *ppValue = &p->pUnpacked->aMem[iIdx];
92816	if( iIdx==p->pTab->iPKey ){

92817	sqlite3VdbeMemSetInt64(pMem, p->iKey1);
92818	}else if( iIdx>=p->pUnpacked->nField ){
92819	ppValue = (sqlite3_value )columnNullValue();
92820	}else if( p->pTab->aCol[iIdx].affinity==SQLITE_AFF_REAL ){
92821	if( pMem->flags & (MEM_Int\|MEM_IntReal) ){
92822	testcase( pMem->flags & MEM_Int );
92823	testcase( pMem->flags & MEM_IntReal );
92824	sqlite3VdbeMemRealify(pMem);
















































92825	}
92826	}
92827
92828	preupdate_old_out:
92829	sqlite3Error(db, rc);
	@@ -93367,10 +93356,108 @@
93367	** commenting and indentation practices when changing or adding code.
93368	*/
93369	/* #include "sqliteInt.h" */
93370	/* #include "vdbeInt.h" */
93371


































































































93372	/*
93373	** Invoke this macro on memory cells just prior to changing the
93374	** value of the cell. This macro verifies that shallow copies are
93375	** not misused. A shallow copy of a string or blob just copies a
93376	** pointer to the string or blob, not the content. If the original
	@@ -97875,12 +97962,17 @@
97875	0, pCx->uc.pCursor);
97876	pCx->isTable = 1;
97877	}
97878	}
97879	pCx->isOrdered = (pOp->p5!=BTREE_UNORDERED);

97880	if( rc ){

97881	sqlite3BtreeClose(pCx->ub.pBtx);



97882	}
97883	}
97884	}
97885	if( rc ) goto abort_due_to_error;
97886	pCx->nullRow = 1;
	@@ -102394,11 +102486,11 @@
102394	** element.
102395	**
102396	** As with all opcodes, the meanings of the parameters for OP_Explain
102397	** are subject to change from one release to the next. Applications
102398	** should not attempt to interpret or use any of the information
102399	** contined in the OP_Explain opcode. The information provided by this
102400	** opcode is intended for testing and debugging use only.
102401	*/
102402	default: { /* This is really OP_Noop, OP_Explain */
102403	assert( pOp->opcode==OP_Noop \|\| pOp->opcode==OP_Explain );
102404
	@@ -107490,11 +107582,11 @@
107490	** non-VIEW candidate plus multiple VIEW candidates. In other
107491	** words non-VIEW candidate terms take precedence over VIEWs.
107492	*/
107493	if( cntTab==0
107494	\|\| (cntTab==1
107495	&& ALWAYS(pMatch!=0)
107496	&& ALWAYS(pMatch->pSTab!=0)
107497	&& (pMatch->pSTab->tabFlags & TF_Ephemeral)!=0
107498	&& (pTab->tabFlags & TF_Ephemeral)==0)
107499	){
107500	cntTab = 1;
	@@ -108123,12 +108215,12 @@
108123	}
108124
108125	/* Resolve function names
108126	*/
108127	case TK_FUNCTION: {
108128	ExprList pList = pExpr->x.pList; / The argument list */
108129	int n = pList ? pList->nExpr : 0; /* Number of arguments */
108130	int no_such_func = 0; /* True if no such function exists */
108131	int wrong_num_args = 0; /* True if wrong number of arguments */
108132	int is_agg = 0; /* True if is an aggregate function */
108133	const char zId; / The function name. */
108134	FuncDef pDef; / Information about the function */
	@@ -108137,10 +108229,12 @@
108137	#ifndef SQLITE_OMIT_WINDOWFUNC
108138	Window *pWin = (IsWindowFunc(pExpr) ? pExpr->y.pWin : 0);
108139	#endif
108140	assert( !ExprHasProperty(pExpr, EP_xIsSelect\|EP_IntValue) );
108141	assert( pExpr->pLeft==0 \|\| pExpr->pLeft->op==TK_ORDER );


108142	zId = pExpr->u.zToken;
108143	pDef = sqlite3FindFunction(pParse->db, zId, n, enc, 0);
108144	if( pDef==0 ){
108145	pDef = sqlite3FindFunction(pParse->db, zId, -2, enc, 0);
108146	if( pDef==0 ){
	@@ -108185,10 +108279,28 @@
108185	pExpr->op = TK_NULL;
108186	return WRC_Prune;
108187	}
108188	}
108189	#endif


















108190	if( pDef->funcFlags & (SQLITE_FUNC_CONSTANT\|SQLITE_FUNC_SLOCHNG) ){
108191	/* For the purposes of the EP_ConstFunc flag, date and time
108192	** functions and other functions that change slowly are considered
108193	** constant because they are constant for the duration of one query.
108194	** This allows them to be factored out of inner loops. */
	@@ -111951,11 +112063,11 @@
111951	**
111952	** (4) If pSrc is the right operand of a LEFT JOIN, then...
111953	** (4a) pExpr must come from an ON clause..
111954	** (4b) and specifically the ON clause associated with the LEFT JOIN.
111955	**
111956	** (5) If pSrc is not the right operand of a LEFT JOIN or the left
111957	** operand of a RIGHT JOIN, then pExpr must be from the WHERE
111958	** clause, not an ON clause.
111959	**
111960	** (6) Either:
111961	**
	@@ -113858,10 +113970,63 @@
113858	}
113859	#endif /* !defined(SQLITE_UNTESTABLE) */
113860	}
113861	return target;
113862	}





















































113863
113864	/*
113865	** Check to see if pExpr is one of the indexed expressions on pParse->pIdxEpr.
113866	** If it is, then resolve the expression by reading from the index and
113867	** return the register into which the value has been read. If pExpr is
	@@ -113891,10 +114056,21 @@
113891	){
113892	/* Affinity mismatch on a generated column */
113893	continue;
113894	}
113895











113896	v = pParse->pVdbe;
113897	assert( v!=0 );
113898	if( p->bMaybeNullRow ){
113899	/* If the index is on a NULL row due to an outer join, then we
113900	** cannot extract the value from the index. The value must be
	@@ -115421,35 +115597,41 @@
115421	**
115422	** Additionally, if pExpr is a simple SQL value and the value is the
115423	** same as that currently bound to variable pVar, non-zero is returned.
115424	** Otherwise, if the values are not the same or if pExpr is not a simple
115425	** SQL value, zero is returned.



115426	*/
115427	static int exprCompareVariable(
115428	const Parse *pParse,
115429	const Expr *pVar,
115430	const Expr *pExpr
115431	){
115432	int res = 0;
115433	int iVar;
115434	sqlite3_value pL, pR = 0;
115435




115436	sqlite3ValueFromExpr(pParse->db, pExpr, SQLITE_UTF8, SQLITE_AFF_BLOB, &pR);
115437	if( pR ){
115438	iVar = pVar->iColumn;
115439	sqlite3VdbeSetVarmask(pParse->pVdbe, iVar);
115440	pL = sqlite3VdbeGetBoundValue(pParse->pReprepare, iVar, SQLITE_AFF_BLOB);
115441	if( pL ){
115442	if( sqlite3_value_type(pL)==SQLITE_TEXT ){
115443	sqlite3_value_text(pL); /* Make sure the encoding is UTF-8 */
115444	}
115445	res = 0==sqlite3MemCompare(pL, pR, 0);
115446	}
115447	sqlite3ValueFree(pR);
115448	sqlite3ValueFree(pL);
115449	}
115450
115451	return res;
115452	}
115453
115454	/*
115455	** Do a deep comparison of two expression trees. Return 0 if the two
	@@ -115471,16 +115653,14 @@
115471	** can be sure the expressions are the same. In the places where
115472	** this routine is used, it does not hurt to get an extra 2 - that
115473	** just might result in some slightly slower code. But returning
115474	** an incorrect 0 or 1 could lead to a malfunction.
115475	**
115476	** If pParse is not NULL then TK_VARIABLE terms in pA with bindings in
115477	** pParse->pReprepare can be matched against literals in pB. The
115478	** pParse->pVdbe->expmask bitmask is updated for each variable referenced.
115479	** If pParse is NULL (the normal case) then any TK_VARIABLE term in
115480	** Argument pParse should normally be NULL. If it is not NULL and pA or
115481	** pB causes a return value of 2.
115482	*/
115483	SQLITE_PRIVATE int sqlite3ExprCompare(
115484	const Parse *pParse,
115485	const Expr *pA,
115486	const Expr *pB,
	@@ -115488,12 +115668,12 @@
115488	){
115489	u32 combinedFlags;
115490	if( pA==0 \|\| pB==0 ){
115491	return pB==pA ? 0 : 2;
115492	}
115493	if( pParse && pA->op==TK_VARIABLE && exprCompareVariable(pParse, pA, pB) ){
115494	return 0;
115495	}
115496	combinedFlags = pA->flags \| pB->flags;
115497	if( combinedFlags & EP_IntValue ){
115498	if( (pA->flags&pB->flags&EP_IntValue)!=0 && pA->u.iValue==pB->u.iValue ){
115499	return 0;
	@@ -115683,23 +115863,75 @@
115683	return exprImpliesNotNull(pParse, p->pLeft, pNN, iTab, 1);
115684	}
115685	}
115686	return 0;
115687	}


















































115688
115689	/*
115690	** Return true if we can prove the pE2 will always be true if pE1 is
115691	** true. Return false if we cannot complete the proof or if pE2 might
115692	** be false. Examples:
115693	**
115694	** pE1: x==5 pE2: x==5 Result: true
115695	** pE1: x>0 pE2: x==5 Result: false
115696	** pE1: x=21 pE2: x=21 OR y=43 Result: true
115697	** pE1: x!=123 pE2: x IS NOT NULL Result: true
115698	** pE1: x!=?1 pE2: x IS NOT NULL Result: true
115699	** pE1: x IS NULL pE2: x IS NOT NULL Result: false
115700	** pE1: x IS ?2 pE2: x IS NOT NULL Result: false


115701	**
115702	** When comparing TK_COLUMN nodes between pE1 and pE2, if pE2 has
115703	** Expr.iTable<0 then assume a table number given by iTab.
115704	**
115705	** If pParse is not NULL, then the values of bound variables in pE1 are
	@@ -115729,10 +115961,13 @@
115729	if( pE2->op==TK_NOTNULL
115730	&& exprImpliesNotNull(pParse, pE1, pE2->pLeft, iTab, 0)
115731	){
115732	return 1;
115733	}



115734	return 0;
115735	}
115736
115737	/* This is a helper function to impliesNotNullRow(). In this routine,
115738	** set pWalker->eCode to one only if both of the input expressions
	@@ -120689,12 +120924,12 @@
120689	int nIdxCol = 1; /* Number of columns in stat4 records */
120690
120691	char zIndex; / Index name */
120692	Index pIdx; / Pointer to the index object */
120693	int nSample; /* Number of samples */
120694	int nByte; /* Bytes of space required */
120695	int i; /* Bytes of space required */
120696	tRowcnt pSpace; / Available allocated memory space */
120697	u8 pPtr; / Available memory as a u8 for easier manipulation */
120698
120699	zIndex = (char *)sqlite3_column_text(pStmt, 0);
120700	if( zIndex==0 ) continue;
	@@ -121140,19 +121375,10 @@
121140	rc = sqlite3Init(db, &zErrDyn);
121141	}
121142	sqlite3BtreeLeaveAll(db);
121143	assert( zErrDyn==0 \|\| rc!=SQLITE_OK );
121144	}
121145	#ifdef SQLITE_USER_AUTHENTICATION
121146	if( rc==SQLITE_OK && !REOPEN_AS_MEMDB(db) ){
121147	u8 newAuth = 0;
121148	rc = sqlite3UserAuthCheckLogin(db, zName, &newAuth);
121149	if( newAuth<db->auth.authLevel ){
121150	rc = SQLITE_AUTH_USER;
121151	}
121152	}
121153	#endif
121154	if( rc ){
121155	if( ALWAYS(!REOPEN_AS_MEMDB(db)) ){
121156	int iDb = db->nDb - 1;
121157	assert( iDb>=2 );
121158	if( db->aDb[iDb].pBt ){
	@@ -121646,15 +121872,11 @@
121646	sqlite3 db = pParse->db; / Database handle */
121647	char zDb = db->aDb[iDb].zDbSName; / Schema name of attached database */
121648	int rc; /* Auth callback return code */
121649
121650	if( db->init.busy ) return SQLITE_OK;
121651	rc = db->xAuth(db->pAuthArg, SQLITE_READ, zTab,zCol,zDb,pParse->zAuthContext
121652	#ifdef SQLITE_USER_AUTHENTICATION
121653	,db->auth.zAuthUser
121654	#endif
121655	);
121656	if( rc==SQLITE_DENY ){
121657	char *z = sqlite3_mprintf("%s.%s", zTab, zCol);
121658	if( db->nDb>2 \|\| iDb!=0 ) z = sqlite3_mprintf("%s.%z", zDb, z);
121659	sqlite3ErrorMsg(pParse, "access to %z is prohibited", z);
121660	pParse->rc = SQLITE_AUTH;
	@@ -121757,15 +121979,11 @@
121757	testcase( zArg1==0 );
121758	testcase( zArg2==0 );
121759	testcase( zArg3==0 );
121760	testcase( pParse->zAuthContext==0 );
121761
121762	rc = db->xAuth(db->pAuthArg, code, zArg1, zArg2, zArg3, pParse->zAuthContext
121763	#ifdef SQLITE_USER_AUTHENTICATION
121764	,db->auth.zAuthUser
121765	#endif
121766	);
121767	if( rc==SQLITE_DENY ){
121768	sqlite3ErrorMsg(pParse, "not authorized");
121769	pParse->rc = SQLITE_AUTH;
121770	}else if( rc!=SQLITE_OK && rc!=SQLITE_IGNORE ){
121771	rc = SQLITE_DENY;
	@@ -121994,21 +122212,10 @@
121994	sqlite3VdbeJumpHere(v, addrRewind);
121995	}
121996	}
121997	sqlite3VdbeAddOp0(v, OP_Halt);
121998
121999	#if SQLITE_USER_AUTHENTICATION && !defined(SQLITE_OMIT_SHARED_CACHE)
122000	if( pParse->nTableLock>0 && db->init.busy==0 ){
122001	sqlite3UserAuthInit(db);
122002	if( db->auth.authLevel<UAUTH_User ){
122003	sqlite3ErrorMsg(pParse, "user not authenticated");
122004	pParse->rc = SQLITE_AUTH_USER;
122005	return;
122006	}
122007	}
122008	#endif
122009
122010	/* The cookie mask contains one bit for each database file open.
122011	** (Bit 0 is for main, bit 1 is for temp, and so forth.) Bits are
122012	** set for each database that is used. Generate code to start a
122013	** transaction on each used database and to verify the schema cookie
122014	** on each used database.
	@@ -122133,20 +122340,10 @@
122133	sqlite3DbFree(db, zSql);
122134	memcpy(PARSE_TAIL(pParse), saveBuf, PARSE_TAIL_SZ);
122135	pParse->nested--;
122136	}
122137
122138	#if SQLITE_USER_AUTHENTICATION
122139	/*
122140	** Return TRUE if zTable is the name of the system table that stores the
122141	** list of users and their access credentials.
122142	*/
122143	SQLITE_PRIVATE int sqlite3UserAuthTable(const char *zTable){
122144	return sqlite3_stricmp(zTable, "sqlite_user")==0;
122145	}
122146	#endif
122147
122148	/*
122149	** Locate the in-memory structure that describes a particular database
122150	** table given the name of that table and (optionally) the name of the
122151	** database containing the table. Return NULL if not found.
122152	**
	@@ -122161,17 +122358,10 @@
122161	Table *p = 0;
122162	int i;
122163
122164	/* All mutexes are required for schema access. Make sure we hold them. */
122165	assert( zDatabase!=0 \|\| sqlite3BtreeHoldsAllMutexes(db) );
122166	#if SQLITE_USER_AUTHENTICATION
122167	/* Only the admin user is allowed to know that the sqlite_user table
122168	** exists */
122169	if( db->auth.authLevel<UAUTH_Admin && sqlite3UserAuthTable(zName)!=0 ){
122170	return 0;
122171	}
122172	#endif
122173	if( zDatabase ){
122174	for(i=0; i<db->nDb; i++){
122175	if( sqlite3StrICmp(zDatabase, db->aDb[i].zDbSName)==0 ) break;
122176	}
122177	if( i>=db->nDb ){
	@@ -125826,13 +126016,10 @@
125826
125827	assert( pTab!=0 );
125828	if( sqlite3StrNICmp(pTab->zName, "sqlite_", 7)==0
125829	&& db->init.busy==0
125830	&& pTblName!=0
125831	#if SQLITE_USER_AUTHENTICATION
125832	&& sqlite3UserAuthTable(pTab->zName)==0
125833	#endif
125834	){
125835	sqlite3ErrorMsg(pParse, "table %s may not be indexed", pTab->zName);
125836	goto exit_create_index;
125837	}
125838	#ifndef SQLITE_OMIT_VIEW
	@@ -128224,10 +128411,11 @@
128224	** 4) The table is a shadow table, the database connection is in
128225	** defensive mode, and the current sqlite3_prepare()
128226	** is for a top-level SQL statement.
128227	*/
128228	static int vtabIsReadOnly(Parse pParse, Table pTab){

128229	if( sqlite3GetVTable(pParse->db, pTab)->pMod->pModule->xUpdate==0 ){
128230	return 1;
128231	}
128232
128233	/* Within triggers:
	@@ -131716,11 +131904,17 @@
131716	#ifdef SQLITE_DEBUG
131717	/*
131718	** Implementation of fpdecode(x,y,z) function.
131719	**
131720	** x is a real number that is to be decoded. y is the precision.
131721	** z is the maximum real precision.






131722	*/
131723	static void fpdecodeFunc(
131724	sqlite3_context *context,
131725	int argc,
131726	sqlite3_value **argv
	@@ -131740,10 +131934,86 @@
131740	sqlite3_snprintf(sizeof(zBuf), zBuf, "NaN");
131741	}else{
131742	sqlite3_snprintf(sizeof(zBuf), zBuf, "%c%.*s/%d", s.sign, s.n, s.z, s.iDP);
131743	}
131744	sqlite3_result_text(context, zBuf, -1, SQLITE_TRANSIENT);












































































131745	}
131746	#endif /* SQLITE_DEBUG */
131747
131748	/*
131749	** All of the FuncDef structures in the aBuiltinFunc[] array above
	@@ -131777,13 +132047,10 @@
131777	#endif
131778	#ifndef SQLITE_OMIT_LOAD_EXTENSION
131779	SFUNCTION(load_extension, 1, 0, 0, loadExt ),
131780	SFUNCTION(load_extension, 2, 0, 0, loadExt ),
131781	#endif
131782	#if SQLITE_USER_AUTHENTICATION
131783	FUNCTION(sqlite_crypt, 2, 0, 0, sqlite3CryptFunc ),
131784	#endif
131785	#ifndef SQLITE_OMIT_COMPILEOPTION_DIAGS
131786	DFUNCTION(sqlite_compileoption_used,1, 0, 0, compileoptionusedFunc ),
131787	DFUNCTION(sqlite_compileoption_get, 1, 0, 0, compileoptiongetFunc ),
131788	#endif /* SQLITE_OMIT_COMPILEOPTION_DIAGS */
131789	INLINE_FUNC(unlikely, 1, INLINEFUNC_unlikely, SQLITE_FUNC_UNLIKELY),
	@@ -131805,11 +132072,12 @@
131805	FUNCTION(max, -1, 1, 1, minmaxFunc ),
131806	FUNCTION(max, 0, 1, 1, 0 ),
131807	WAGGREGATE(max, 1, 1, 1, minmaxStep, minMaxFinalize, minMaxValue, 0,
131808	SQLITE_FUNC_MINMAX\|SQLITE_FUNC_ANYORDER ),
131809	FUNCTION2(typeof, 1, 0, 0, typeofFunc, SQLITE_FUNC_TYPEOF),
131810	FUNCTION2(subtype, 1, 0, 0, subtypeFunc, SQLITE_FUNC_TYPEOF),

131811	FUNCTION2(length, 1, 0, 0, lengthFunc, SQLITE_FUNC_LENGTH),
131812	FUNCTION2(octet_length, 1, 0, 0, bytelengthFunc,SQLITE_FUNC_BYTELEN),
131813	FUNCTION(instr, 2, 0, 0, instrFunc ),
131814	FUNCTION(printf, -1, 0, 0, printfFunc ),
131815	FUNCTION(format, -1, 0, 0, printfFunc ),
	@@ -131816,10 +132084,11 @@
131816	FUNCTION(unicode, 1, 0, 0, unicodeFunc ),
131817	FUNCTION(char, -1, 0, 0, charFunc ),
131818	FUNCTION(abs, 1, 0, 0, absFunc ),
131819	#ifdef SQLITE_DEBUG
131820	FUNCTION(fpdecode, 3, 0, 0, fpdecodeFunc ),

131821	#endif
131822	#ifndef SQLITE_OMIT_FLOATING_POINT
131823	FUNCTION(round, 1, 0, 0, roundFunc ),
131824	FUNCTION(round, 2, 0, 0, roundFunc ),
131825	#endif
	@@ -131910,15 +132179,18 @@
131910	MFUNCTION(atanh, 1, atanh, math1Func ),
131911	#endif
131912	MFUNCTION(sqrt, 1, sqrt, math1Func ),
131913	MFUNCTION(radians, 1, degToRad, math1Func ),
131914	MFUNCTION(degrees, 1, radToDeg, math1Func ),
131915	FUNCTION(pi, 0, 0, 0, piFunc ),
131916	#endif /* SQLITE_ENABLE_MATH_FUNCTIONS */
131917	FUNCTION(sign, 1, 0, 0, signFunc ),
131918	INLINE_FUNC(coalesce, -1, INLINEFUNC_coalesce, 0 ),

131919	INLINE_FUNC(iif, 3, INLINEFUNC_iif, 0 ),


131920	};
131921	#ifndef SQLITE_OMIT_ALTERTABLE
131922	sqlite3AlterFunctions();
131923	#endif
131924	sqlite3WindowFunctions();
	@@ -140427,16 +140699,10 @@
140427	if( db->autoCommit==0 ){
140428	/* Foreign key support may not be enabled or disabled while not
140429	** in auto-commit mode. */
140430	mask &= ~(SQLITE_ForeignKeys);
140431	}
140432	#if SQLITE_USER_AUTHENTICATION
140433	if( db->auth.authLevel==UAUTH_User ){
140434	/* Do not allow non-admin users to modify the schema arbitrarily */
140435	mask &= ~(SQLITE_WriteSchema);
140436	}
140437	#endif
140438
140439	if( sqlite3GetBoolean(zRight, 0) ){
140440	if( (mask & SQLITE_WriteSchema)==0
140441	\|\| (db->flags & SQLITE_Defensive)==0
140442	){
	@@ -140568,11 +140834,12 @@
140568	pTab = sqliteHashData(k);
140569	if( pTab->nCol==0 ){
140570	char zSql = sqlite3MPrintf(db, "SELECTFROM\"%w\"", pTab->zName);
140571	if( zSql ){
140572	sqlite3_stmt *pDummy = 0;
140573	(void)sqlite3_prepare(db, zSql, -1, &pDummy, 0);

140574	(void)sqlite3_finalize(pDummy);
140575	sqlite3DbFree(db, zSql);
140576	}
140577	if( db->mallocFailed ){
140578	sqlite3ErrorMsg(db->pParse, "out of memory");
	@@ -141044,10 +141311,11 @@
141044	}
141045	aRoot[0] = cnt;
141046
141047	/* Make sure sufficient number of registers have been allocated */
141048	sqlite3TouchRegister(pParse, 8+cnt);

141049	sqlite3ClearTempRegCache(pParse);
141050
141051	/* Do the b-tree integrity checks */
141052	sqlite3VdbeAddOp4(v, OP_IntegrityCk, 1, cnt, 8, (char*)aRoot,P4_INTARRAY);
141053	sqlite3VdbeChangeP5(v, (u8)i);
	@@ -142668,18 +142936,11 @@
142668	encoding = (u8)meta[BTREE_TEXT_ENCODING-1] & 3;
142669	if( encoding==0 ) encoding = SQLITE_UTF8;
142670	#else
142671	encoding = SQLITE_UTF8;
142672	#endif
142673	if( db->nVdbeActive>0 && encoding!=ENC(db)
142674	&& (db->mDbFlags & DBFLAG_Vacuum)==0
142675	){
142676	rc = SQLITE_LOCKED;
142677	goto initone_error_out;
142678	}else{
142679	sqlite3SetTextEncoding(db, encoding);
142680	}
142681	}else{
142682	/* If opening an attached database, the encoding much match ENC(db) */
142683	if( (meta[BTREE_TEXT_ENCODING-1] & 3)!=ENC(db) ){
142684	sqlite3SetString(pzErrMsg, db, "attached databases must use the same"
142685	" text encoding as main database");
	@@ -143369,16 +143630,28 @@
143369	#endif
143370	*ppStmt = 0;
143371	if( !sqlite3SafetyCheckOk(db)\|\|zSql==0 ){
143372	return SQLITE_MISUSE_BKPT;
143373	}






143374	if( nBytes>=0 ){
143375	int sz;
143376	const char z = (const char)zSql;
143377	for(sz=0; sz<nBytes && (z[sz]!=0 \|\| z[sz+1]!=0); sz += 2){}
143378	nBytes = sz;





143379	}

143380	sqlite3_mutex_enter(db->mutex);
143381	zSql8 = sqlite3Utf16to8(db, zSql, nBytes, SQLITE_UTF16NATIVE);
143382	if( zSql8 ){
143383	rc = sqlite3LockAndPrepare(db, zSql8, -1, prepFlags, 0, ppStmt, &zTail8);
143384	}
	@@ -143388,11 +143661,11 @@
143388	** equivalent pointer into the UTF-16 string by counting the unicode
143389	** characters between zSql8 and zTail8, and then returning a pointer
143390	** the same number of characters into the UTF-16 string.
143391	*/
143392	int chars_parsed = sqlite3Utf8CharLen(zSql8, (int)(zTail8-zSql8));
143393	pzTail = (u8 )zSql + sqlite3Utf16ByteLen(zSql, chars_parsed);
143394	}
143395	sqlite3DbFree(db, zSql8);
143396	rc = sqlite3ApiExit(db, rc);
143397	sqlite3_mutex_leave(db->mutex);
143398	return rc;
	@@ -147361,36 +147634,36 @@
147361	return pExpr;
147362	}
147363	if( pSubst->isOuterJoin ){
147364	ExprSetProperty(pNew, EP_CanBeNull);
147365	}




















147366	if( ExprHasProperty(pExpr,EP_OuterON\|EP_InnerON) ){
147367	sqlite3SetJoinExpr(pNew, pExpr->w.iJoin,
147368	pExpr->flags & (EP_OuterON\|EP_InnerON));
147369	}
147370	sqlite3ExprDelete(db, pExpr);
147371	pExpr = pNew;
147372	if( pExpr->op==TK_TRUEFALSE ){
147373	pExpr->u.iValue = sqlite3ExprTruthValue(pExpr);
147374	pExpr->op = TK_INTEGER;
147375	ExprSetProperty(pExpr, EP_IntValue);
147376	}
147377
147378	/* Ensure that the expression now has an implicit collation sequence,
147379	** just as it did when it was a column of a view or sub-query. */
147380	{
147381	CollSeq *pNat = sqlite3ExprCollSeq(pSubst->pParse, pExpr);
147382	CollSeq *pColl = sqlite3ExprCollSeq(pSubst->pParse,
147383	pSubst->pCList->a[iColumn].pExpr
147384	);
147385	if( pNat!=pColl \|\| (pExpr->op!=TK_COLUMN && pExpr->op!=TK_COLLATE) ){
147386	pExpr = sqlite3ExprAddCollateString(pSubst->pParse, pExpr,
147387	(pColl ? pColl->zName : "BINARY")
147388	);
147389	}
147390	}
147391	ExprClearProperty(pExpr, EP_Collate);
147392	}
147393	}
147394	}else{
147395	if( pExpr->op==TK_IF_NULL_ROW && pExpr->iTable==pSubst->iTable ){
147396	pExpr->iTable = pSubst->iNewTable;
	@@ -148123,20 +148396,20 @@
148123	}
148124
148125	/* Transfer the FROM clause terms from the subquery into the
148126	** outer query.
148127	*/

148128	for(i=0; i<nSubSrc; i++){
148129	SrcItem *pItem = &pSrc->a[i+iFrom];
148130	assert( pItem->fg.isTabFunc==0 );
148131	assert( pItem->fg.isSubquery
148132	\|\| pItem->fg.fixedSchema
148133	\|\| pItem->u4.zDatabase==0 );
148134	if( pItem->fg.isUsing ) sqlite3IdListDelete(db, pItem->u3.pUsing);
148135	*pItem = pSubSrc->a[i];
148136	pItem->fg.jointype \|= ltorj;
148137	iNewParent = pSubSrc->a[i].iCursor;
148138	memset(&pSubSrc->a[i], 0, sizeof(pSubSrc->a[i]));
148139	}
148140	pSrc->a[iFrom].fg.jointype &= JT_LTORJ;
148141	pSrc->a[iFrom].fg.jointype \|= jointype \| ltorj;
148142
	@@ -148172,10 +148445,11 @@
148172	pSub->pOrderBy = 0;
148173	}
148174	pWhere = pSub->pWhere;
148175	pSub->pWhere = 0;
148176	if( isOuterJoin>0 ){

148177	sqlite3SetJoinExpr(pWhere, iNewParent, EP_OuterON);
148178	}
148179	if( pWhere ){
148180	if( pParent->pWhere ){
148181	pParent->pWhere = sqlite3PExpr(pParse, TK_AND, pWhere, pParent->pWhere);
	@@ -151271,11 +151545,11 @@
151271	sqlite3TreeViewSelect(0, p, 0);
151272	}
151273	#endif
151274	assert( pSubq->pSelect && (pSub->selFlags & SF_PushDown)!=0 );
151275	}else{
151276	TREETRACE(0x4000,pParse,p,("WHERE-lcause push-down not possible\n"));
151277	}
151278
151279	/* Convert unused result columns of the subquery into simple NULL
151280	** expressions, to avoid unneeded searching and computation.
151281	** tag-select-0440
	@@ -156980,10 +157254,11 @@
156980	assert( sParse.zErrMsg==0 );
156981	if( !pTab->aCol ){
156982	Table *pNew = sParse.pNewTable;
156983	Index *pIdx;
156984	pTab->aCol = pNew->aCol;

156985	sqlite3ExprListDelete(db, pNew->u.tab.pDfltList);
156986	pTab->nNVCol = pTab->nCol = pNew->nCol;
156987	pTab->tabFlags \|= pNew->tabFlags & (TF_WithoutRowid\|TF_NoVisibleRowid);
156988	pNew->nCol = 0;
156989	pNew->aCol = 0;
	@@ -158044,13 +158319,21 @@
158044	SQLITE_PRIVATE int sqlite3WhereExplainBloomFilter(
158045	const Parse pParse, / Parse context */
158046	const WhereInfo pWInfo, / WHERE clause */
158047	const WhereLevel pLevel / Bloom filter on this level */
158048	);







158049	#else
158050	# define sqlite3WhereExplainOneScan(u,v,w,x) 0
158051	# define sqlite3WhereExplainBloomFilter(u,v,w) 0

158052	#endif /* SQLITE_OMIT_EXPLAIN */
158053	#ifdef SQLITE_ENABLE_STMT_SCANSTATUS
158054	SQLITE_PRIVATE void sqlite3WhereAddScanStatus(
158055	Vdbe v, / Vdbe to add scanstatus entry to */
158056	SrcList pSrclist, / FROM clause pLvl reads data from */
	@@ -158248,42 +158531,42 @@
158248	}
158249	sqlite3_str_append(pStr, ")", 1);
158250	}
158251
158252	/*
158253	** This function is a no-op unless currently processing an EXPLAIN QUERY PLAN
158254	** command, or if stmt_scanstatus_v2() stats are enabled, or if SQLITE_DEBUG
158255	** was defined at compile-time. If it is not a no-op, a single OP_Explain
158256	** opcode is added to the output to describe the table scan strategy in pLevel.
158257	**
158258	** If an OP_Explain opcode is added to the VM, its address is returned.
158259	** Otherwise, if no OP_Explain is coded, zero is returned.
158260	*/
158261	SQLITE_PRIVATE int sqlite3WhereExplainOneScan(
158262	Parse pParse, / Parse context */

158263	SrcList pTabList, / Table list this loop refers to */
158264	WhereLevel pLevel, / Scan to write OP_Explain opcode for */
158265	u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
158266	){
158267	int ret = 0;
158268	#if !defined(SQLITE_DEBUG)
158269	if( sqlite3ParseToplevel(pParse)->explain==2 \|\| IS_STMT_SCANSTATUS(pParse->db) )
158270	#endif
158271	{


158272	SrcItem *pItem = &pTabList->a[pLevel->iFrom];
158273	Vdbe v = pParse->pVdbe; / VM being constructed */
158274	sqlite3 db = pParse->db; / Database handle */
158275	int isSearch; /* True for a SEARCH. False for SCAN. */
158276	WhereLoop pLoop; / The controlling WhereLoop object */
158277	u32 flags; /* Flags that describe this loop */

158278	char zMsg; / Text to add to EQP output */

158279	StrAccum str; /* EQP output string */
158280	char zBuf[100]; /* Initial space for EQP output string */
158281


158282	pLoop = pLevel->pWLoop;
158283	flags = pLoop->wsFlags;
158284	if( (flags&WHERE_MULTI_OR) \|\| (wctrlFlags&WHERE_OR_SUBCLAUSE) ) return 0;
158285
158286	isSearch = (flags&(WHERE_BTM_LIMIT\|WHERE_TOP_LIMIT))!=0
158287	\|\| ((flags&WHERE_VIRTUALTABLE)==0 && (pLoop->u.btree.nEq>0))
158288	\|\| (wctrlFlags&(WHERE_ORDERBY_MIN\|WHERE_ORDERBY_MAX));
158289
	@@ -158303,11 +158586,11 @@
158303	}
158304	}else if( flags & WHERE_PARTIALIDX ){
158305	zFmt = "AUTOMATIC PARTIAL COVERING INDEX";
158306	}else if( flags & WHERE_AUTO_INDEX ){
158307	zFmt = "AUTOMATIC COVERING INDEX";
158308	}else if( flags & WHERE_IDX_ONLY ){
158309	zFmt = "COVERING INDEX %s";
158310	}else{
158311	zFmt = "INDEX %s";
158312	}
158313	if( zFmt ){
	@@ -158355,15 +158638,54 @@
158355	sqlite3LogEstToInt(pLoop->nOut));
158356	}else{
158357	sqlite3_str_append(&str, " (~1 row)", 9);
158358	}
158359	#endif

158360	zMsg = sqlite3StrAccumFinish(&str);
158361	sqlite3ExplainBreakpoint("",zMsg);
158362	ret = sqlite3VdbeAddOp4(v, OP_Explain, sqlite3VdbeCurrentAddr(v),
158363	pParse->addrExplain, pLoop->rRun,
158364	zMsg, P4_DYNAMIC);






































158365	}
158366	return ret;
158367	}
158368
158369	/*
	@@ -158458,13 +158780,14 @@
158458	if( wsFlags & WHERE_INDEXED ){
158459	sqlite3VdbeScanStatusRange(v, addrExplain, -1, pLvl->iIdxCur);
158460	}
158461	}else{
158462	int addr;

158463	assert( pSrclist->a[pLvl->iFrom].fg.isSubquery );
158464	addr = pSrclist->a[pLvl->iFrom].u4.pSubq->addrFillSub;
158465	VdbeOp *pOp = sqlite3VdbeGetOp(v, addr-1);
158466	assert( sqlite3VdbeDb(v)->mallocFailed \|\| pOp->opcode==OP_InitCoroutine );
158467	assert( sqlite3VdbeDb(v)->mallocFailed \|\| pOp->p2>addr );
158468	sqlite3VdbeScanStatusRange(v, addrExplain, addr, pOp->p2-1);
158469	}
158470	}
	@@ -158713,10 +159036,11 @@
158713	if( pOrigLhs ){
158714	sqlite3ExprListDelete(db, pOrigLhs);
158715	pNew->pLeft->x.pList = pLhs;
158716	}
158717	pSelect->pEList = pRhs;

158718	if( pLhs && pLhs->nExpr==1 ){
158719	/* Take care here not to generate a TK_VECTOR containing only a
158720	** single value. Since the parser never creates such a vector, some
158721	** of the subroutines do not handle this case. */
158722	Expr *p = pLhs->a[0].pExpr;
	@@ -161260,24 +161584,29 @@
161260	}else if( op==TK_STRING ){
161261	assert( !ExprHasProperty(pRight, EP_IntValue) );
161262	z = (u8*)pRight->u.zToken;
161263	}
161264	if( z ){
161265
161266	/* Count the number of prefix characters prior to the first wildcard.
161267	** If the underlying database has a UTF16LE encoding, then only consider
161268	** ASCII characters. Note that the encoding of z[] is UTF8 - we are
161269	** dealing with only UTF8 here in this code, but the database engine
161270	** itself might be processing content using a different encoding. */
161271	cnt = 0;
161272	while( (c=z[cnt])!=0 && c!=wc[0] && c!=wc[1] && c!=wc[2] ){
161273	cnt++;
161274	if( c==wc[3] && z[cnt]!=0 ){
161275	cnt++;
161276	}else if( c>=0x80 && ENC(db)==SQLITE_UTF16LE ){
161277	cnt--;
161278	break;





161279	}
161280	}
161281
161282	/* The optimization is possible only if (1) the pattern does not begin
161283	** with a wildcard and if (2) the non-wildcard prefix does not end with
	@@ -161285,11 +161614,11 @@
161285	** a single escape character. The second condition is necessary so
161286	** that we can increment the prefix key to find an upper bound for the
161287	** range search. The third is because the caller assumes that the pattern
161288	** consists of at least one character after all escapes have been
161289	** removed. */
161290	if( (cnt>1 \|\| (cnt>0 && z[0]!=wc[3])) && 255!=(u8)z[cnt-1] ){
161291	Expr *pPrefix;
161292
161293	/* A "complete" match if the pattern ends with "" or "%" /
161294	*pisComplete = c==wc[0] && z[cnt+1]==0 && ENC(db)!=SQLITE_UTF16LE;
161295
	@@ -163780,11 +164109,11 @@
163780	\|\| pTerm->pExpr->w.iJoin != pSrc->iCursor
163781	){
163782	return 0;
163783	}
163784	if( (pSrc->fg.jointype & (JT_LEFT\|JT_RIGHT))!=0
163785	&& ExprHasProperty(pTerm->pExpr, EP_InnerON)
163786	){
163787	return 0;
163788	}
163789	return 1;
163790	}
	@@ -164577,13 +164906,15 @@
164577	** Whether or not an error is returned, it is the responsibility of the
164578	** caller to eventually free p->idxStr if p->needToFreeIdxStr indicates
164579	** that this is required.
164580	*/
164581	static int vtabBestIndex(Parse pParse, Table pTab, sqlite3_index_info *p){
164582	sqlite3_vtab *pVtab = sqlite3GetVTable(pParse->db, pTab)->pVtab;
164583	int rc;

164584


164585	whereTraceIndexInfoInputs(p, pTab);
164586	pParse->db->nSchemaLock++;
164587	rc = pVtab->pModule->xBestIndex(pVtab, p);
164588	pParse->db->nSchemaLock--;
164589	whereTraceIndexInfoOutputs(p, pTab);
	@@ -165271,11 +165602,11 @@
165271	return rc;
165272	}
165273	#endif /* SQLITE_ENABLE_STAT4 */
165274
165275
165276	#ifdef WHERETRACE_ENABLED
165277	/*
165278	** Print the content of a WhereTerm object
165279	*/
165280	SQLITE_PRIVATE void sqlite3WhereTermPrint(WhereTerm *pTerm, int iTerm){
165281	if( pTerm==0 ){
	@@ -165314,10 +165645,13 @@
165314	sqlite3DebugPrintf(" iParent=%d", pTerm->iParent);
165315	}
165316	sqlite3DebugPrintf("\n");
165317	sqlite3TreeViewExpr(0, pTerm->pExpr, 0);
165318	}



165319	}
165320	#endif
165321
165322	#ifdef WHERETRACE_ENABLED
165323	/*
	@@ -165522,11 +165856,11 @@
165522	** Return TRUE if X is a proper subset of Y but is of equal or less cost.
165523	** In other words, return true if all constraints of X are also part of Y
165524	** and Y has additional constraints that might speed the search that X lacks
165525	** but the cost of running X is not more than the cost of running Y.
165526	**
165527	** In other words, return true if the cost relationwship between X and Y
165528	** is inverted and needs to be adjusted.
165529	**
165530	** Case 1:
165531	**
165532	** (1a) X and Y use the same index.
	@@ -166500,11 +166834,10 @@
166500	pParse = pWC->pWInfo->pParse;
166501	while( pWhere->op==TK_AND ){
166502	if( !whereUsablePartialIndex(iTab,jointype,pWC,pWhere->pLeft) ) return 0;
166503	pWhere = pWhere->pRight;
166504	}
166505	if( pParse->db->flags & SQLITE_EnableQPSG ) pParse = 0;
166506	for(i=0, pTerm=pWC->a; i<pWC->nTerm; i++, pTerm++){
166507	Expr *pExpr;
166508	pExpr = pTerm->pExpr;
166509	if( (!ExprHasProperty(pExpr, EP_OuterON) \|\| pExpr->w.iJoin==iTab)
166510	&& ((jointype & JT_OUTER)==0 \|\| ExprHasProperty(pExpr, EP_OuterON))
	@@ -169140,10 +169473,11 @@
169140	hasRightJoin = (pWInfo->pTabList->a[0].fg.jointype & JT_LTORJ)!=0;
169141	for(i=pWInfo->nLevel-1; i>=1; i--){
169142	WhereTerm pTerm, pEnd;
169143	SrcItem *pItem;
169144	WhereLoop *pLoop;

169145	pLoop = pWInfo->a[i].pWLoop;
169146	pItem = &pWInfo->pTabList->a[pLoop->iTab];
169147	if( (pItem->fg.jointype & (JT_LEFT\|JT_RIGHT))!=JT_LEFT ) continue;
169148	if( (pWInfo->wctrlFlags & WHERE_WANT_DISTINCT)==0
169149	&& (pLoop->wsFlags & WHERE_ONEROW)==0
	@@ -169160,17 +169494,20 @@
169160	break;
169161	}
169162	}
169163	if( hasRightJoin
169164	&& ExprHasProperty(pTerm->pExpr, EP_InnerON)
169165	&& pTerm->pExpr->w.iJoin==pItem->iCursor
169166	){
169167	break; /* restriction (5) */
169168	}
169169	}
169170	if( pTerm<pEnd ) continue;
169171	WHERETRACE(0xffffffff, ("-> drop loop %c not used\n", pLoop->cId));



169172	notReady &= ~pLoop->maskSelf;
169173	for(pTerm=pWInfo->sWC.a; pTerm<pEnd; pTerm++){
169174	if( (pTerm->prereqAll & pLoop->maskSelf)!=0 ){
169175	pTerm->wtFlags \|= TERM_CODED;
169176	}
	@@ -169237,62 +169574,10 @@
169237	nSearch += pLoop->nOut;
169238	if( pWInfo->nOutStarDelta ) nSearch += pLoop->rStarDelta;
169239	}
169240	}
169241
169242	/*
169243	** Expression Node callback for sqlite3ExprCanReturnSubtype().
169244	**
169245	** Only a function call is able to return a subtype. So if the node
169246	** is not a function call, return WRC_Prune immediately.
169247	**
169248	** A function call is able to return a subtype if it has the
169249	** SQLITE_RESULT_SUBTYPE property.
169250	**
169251	** Assume that every function is able to pass-through a subtype from
169252	** one of its argument (using sqlite3_result_value()). Most functions
169253	** are not this way, but we don't have a mechanism to distinguish those
169254	** that are from those that are not, so assume they all work this way.
169255	** That means that if one of its arguments is another function and that
169256	** other function is able to return a subtype, then this function is
169257	** able to return a subtype.
169258	*/
169259	static int exprNodeCanReturnSubtype(Walker pWalker, Expr pExpr){
169260	int n;
169261	FuncDef *pDef;
169262	sqlite3 *db;
169263	if( pExpr->op!=TK_FUNCTION ){
169264	return WRC_Prune;
169265	}
169266	assert( ExprUseXList(pExpr) );
169267	db = pWalker->pParse->db;
169268	n = pExpr->x.pList ? pExpr->x.pList->nExpr : 0;
169269	pDef = sqlite3FindFunction(db, pExpr->u.zToken, n, ENC(db), 0);
169270	if( pDef==0 \|\| (pDef->funcFlags & SQLITE_RESULT_SUBTYPE)!=0 ){
169271	pWalker->eCode = 1;
169272	return WRC_Prune;
169273	}
169274	return WRC_Continue;
169275	}
169276
169277	/*
169278	** Return TRUE if expression pExpr is able to return a subtype.
169279	**
169280	** A TRUE return does not guarantee that a subtype will be returned.
169281	** It only indicates that a subtype return is possible. False positives
169282	** are acceptable as they only disable an optimization. False negatives,
169283	** on the other hand, can lead to incorrect answers.
169284	*/
169285	static int sqlite3ExprCanReturnSubtype(Parse pParse, Expr pExpr){
169286	Walker w;
169287	memset(&w, 0, sizeof(w));
169288	w.pParse = pParse;
169289	w.xExprCallback = exprNodeCanReturnSubtype;
169290	sqlite3WalkExpr(&w, pExpr);
169291	return w.eCode;
169292	}
169293
169294	/*
169295	** The index pIdx is used by a query and contains one or more expressions.
169296	** In other words pIdx is an index on an expression. iIdxCur is the cursor
169297	** number for the index and iDataCur is the cursor number for the corresponding
169298	** table.
	@@ -169322,16 +169607,10 @@
169322	pExpr = sqlite3ColumnExpr(pTab, &pTab->aCol[j]);
169323	}else{
169324	continue;
169325	}
169326	if( sqlite3ExprIsConstant(0,pExpr) ) continue;
169327	if( pExpr->op==TK_FUNCTION && sqlite3ExprCanReturnSubtype(pParse,pExpr) ){
169328	/* Functions that might set a subtype should not be replaced by the
169329	** value taken from an expression index since the index omits the
169330	** subtype. https://sqlite.org/forum/forumpost/68d284c86b082c3e */
169331	continue;
169332	}
169333	p = sqlite3DbMallocRaw(pParse->db, sizeof(IndexedExpr));
169334	if( p==0 ) break;
169335	p->pIENext = pParse->pIdxEpr;
169336	#ifdef WHERETRACE_ENABLED
169337	if( sqlite3WhereTrace & 0x200 ){
	@@ -170434,18 +170713,32 @@
170434	x = sqlite3TableColumnToIndex(pIdx, x);
170435	if( x>=0 ){
170436	pOp->p2 = x;
170437	pOp->p1 = pLevel->iIdxCur;
170438	OpcodeRewriteTrace(db, k, pOp);
170439	}else{
170440	/* Unable to translate the table reference into an index
170441	** reference. Verify that this is harmless - that the
170442	** table being referenced really is open.
170443	*/
170444	if( pLoop->wsFlags & WHERE_IDX_ONLY ){



170445	sqlite3ErrorMsg(pParse, "internal query planner error");
170446	pParse->rc = SQLITE_INTERNAL;















170447	}
170448	}
170449	}else if( pOp->opcode==OP_Rowid ){
170450	pOp->p1 = pLevel->iIdxCur;
170451	pOp->opcode = OP_IdxRowid;
	@@ -172149,10 +172442,11 @@
172149	for(pWin=pMWin; pWin; pWin=pWin->pNextWin){
172150	FuncDef *pFunc = pWin->pWFunc;
172151	int regArg;
172152	int nArg = pWin->bExprArgs ? 0 : windowArgCount(pWin);
172153	int i;

172154
172155	assert( bInverse==0 \|\| pWin->eStart!=TK_UNBOUNDED );
172156
172157	/* All OVER clauses in the same window function aggregate step must
172158	** be the same. */
	@@ -172164,10 +172458,22 @@
172164	}else{
172165	sqlite3VdbeAddOp3(v, OP_Column, pMWin->iEphCsr, pWin->iArgCol+i, reg+i);
172166	}
172167	}
172168	regArg = reg;












172169
172170	if( pMWin->regStartRowid==0
172171	&& (pFunc->funcFlags & SQLITE_FUNC_MINMAX)
172172	&& (pWin->eStart!=TK_UNBOUNDED)
172173	){
	@@ -172184,29 +172490,17 @@
172184	sqlite3VdbeAddOp1(v, OP_Delete, pWin->csrApp);
172185	sqlite3VdbeJumpHere(v, sqlite3VdbeCurrentAddr(v)-2);
172186	}
172187	sqlite3VdbeJumpHere(v, addrIsNull);
172188	}else if( pWin->regApp ){

172189	assert( pFunc->zName==nth_valueName
172190	\|\| pFunc->zName==first_valueName
172191	);
172192	assert( bInverse==0 \|\| bInverse==1 );
172193	sqlite3VdbeAddOp2(v, OP_AddImm, pWin->regApp+1-bInverse, 1);
172194	}else if( pFunc->xSFunc!=noopStepFunc ){
172195	int addrIf = 0;
172196	if( pWin->pFilter ){
172197	int regTmp;
172198	assert( ExprUseXList(pWin->pOwner) );
172199	assert( pWin->bExprArgs \|\| !nArg \|\|nArg==pWin->pOwner->x.pList->nExpr );
172200	assert( pWin->bExprArgs \|\| nArg \|\|pWin->pOwner->x.pList==0 );
172201	regTmp = sqlite3GetTempReg(pParse);
172202	sqlite3VdbeAddOp3(v, OP_Column, csr, pWin->iArgCol+nArg,regTmp);
172203	addrIf = sqlite3VdbeAddOp3(v, OP_IfNot, regTmp, 0, 1);
172204	VdbeCoverage(v);
172205	sqlite3ReleaseTempReg(pParse, regTmp);
172206	}
172207
172208	if( pWin->bExprArgs ){
172209	int iOp = sqlite3VdbeCurrentAddr(v);
172210	int iEnd;
172211
172212	assert( ExprUseXList(pWin->pOwner) );
	@@ -172233,12 +172527,13 @@
172233	sqlite3VdbeAppendP4(v, pFunc, P4_FUNCDEF);
172234	sqlite3VdbeChangeP5(v, (u8)nArg);
172235	if( pWin->bExprArgs ){
172236	sqlite3ReleaseTempRange(pParse, regArg, nArg);
172237	}
172238	if( addrIf ) sqlite3VdbeJumpHere(v, addrIf);
172239	}


172240	}
172241	}
172242
172243	/*
172244	** Values that may be passed as the second argument to windowCodeOp().
	@@ -173660,10 +173955,17 @@
173660	** Then the "b" IdList records the list "a,b,c".
173661	*/
173662	struct TrigEvent { int a; IdList * b; };
173663
173664	struct FrameBound { int eType; Expr *pExpr; };







173665
173666	/*
173667	** Disable lookaside memory allocation for objects that might be
173668	** shared across database connections.
173669	*/
	@@ -177553,11 +177855,15 @@
177553	}
177554	break;
177555	case 84: /* cmd ::= select */
177556	{
177557	SelectDest dest = {SRT_Output, 0, 0, 0, 0, 0, 0};
177558	sqlite3Select(pParse, yymsp[0].minor.yy555, &dest);




177559	sqlite3SelectDelete(pParse->db, yymsp[0].minor.yy555);
177560	}
177561	break;
177562	case 85: /* select ::= WITH wqlist selectnowith */
177563	{yymsp[-2].minor.yy555 = attachWithToSelect(pParse,yymsp[0].minor.yy555,yymsp[-1].minor.yy59);}
	@@ -178024,11 +178330,11 @@
178024	** that look like this: #1 #2 ... These terms refer to registers
178025	** in the virtual machine. #N is the N-th register. */
178026	Token t = yymsp[0].minor.yy0; /A-overwrites-X/
178027	assert( t.n>=2 );
178028	if( pParse->nested==0 ){
178029	sqlite3ErrorMsg(pParse, "near \"%T\": syntax error", &t);
178030	yymsp[0].minor.yy454 = 0;
178031	}else{
178032	yymsp[0].minor.yy454 = sqlite3PExpr(pParse, TK_REGISTER, 0, 0);
178033	if( yymsp[0].minor.yy454 ) sqlite3GetInt32(&t.z[1], &yymsp[0].minor.yy454->iTable);
178034	}
	@@ -178872,11 +179178,11 @@
178872	#define TOKEN yyminor
178873	/********** Begin %syntax_error code **************************************/
178874
178875	UNUSED_PARAMETER(yymajor); /* Silence some compiler warnings */
178876	if( TOKEN.z[0] ){
178877	sqlite3ErrorMsg(pParse, "near \"%T\": syntax error", &TOKEN);
178878	}else{
178879	sqlite3ErrorMsg(pParse, "incomplete input");
178880	}
178881	/********** End %syntax_error code ****************************************/
178882	sqlite3ParserARG_STORE /* Suppress warning about unused %extra_argument variable */
	@@ -180363,11 +180669,13 @@
180363	}
180364	if( pParse->zErrMsg \|\| (pParse->rc!=SQLITE_OK && pParse->rc!=SQLITE_DONE) ){
180365	if( pParse->zErrMsg==0 ){
180366	pParse->zErrMsg = sqlite3MPrintf(db, "%s", sqlite3ErrStr(pParse->rc));
180367	}
180368	sqlite3_log(pParse->rc, "%s in \"%s\"", pParse->zErrMsg, pParse->zTail);


180369	nErr++;
180370	}
180371	pParse->zTail = zSql;
180372	#ifndef SQLITE_OMIT_VIRTUALTABLE
180373	sqlite3_free(pParse->apVtabLock);
	@@ -181071,36 +181379,10 @@
181071	** all database files specified with a relative pathname.
181072	**
181073	** See also the "PRAGMA data_store_directory" SQL command.
181074	*/
181075	SQLITE_API char *sqlite3_data_directory = 0;
181076
181077	/*
181078	** Determine whether or not high-precision (long double) floating point
181079	** math works correctly on CPU currently running.
181080	*/
181081	static SQLITE_NOINLINE int hasHighPrecisionDouble(int rc){
181082	if( sizeof(LONGDOUBLE_TYPE)<=8 ){
181083	/* If the size of "long double" is not more than 8, then
181084	** high-precision math is not possible. */
181085	return 0;
181086	}else{
181087	/* Just because sizeof(long double)>8 does not mean that the underlying
181088	** hardware actually supports high-precision floating point. For example,
181089	** clearing the 0x100 bit in the floating-point control word on Intel
181090	** processors will make long double work like double, even though long
181091	** double takes up more space. The only way to determine if long double
181092	** actually works is to run an experiment. */
181093	LONGDOUBLE_TYPE a, b, c;
181094	rc++;
181095	a = 1.0+rc*0.1;
181096	b = 1.0e+18+rc*25.0;
181097	c = a+b;
181098	return b!=c;
181099	}
181100	}
181101
181102
181103	/*
181104	** Initialize SQLite.
181105	**
181106	** This routine must be called to initialize the memory allocation,
	@@ -181292,17 +181574,10 @@
181292	if( bRunExtraInit ){
181293	int SQLITE_EXTRA_INIT(const char*);
181294	rc = SQLITE_EXTRA_INIT(0);
181295	}
181296	#endif
181297
181298	/* Experimentally determine if high-precision floating point is
181299	** available. */
181300	#ifndef SQLITE_OMIT_WSD
181301	sqlite3Config.bUseLongDouble = hasHighPrecisionDouble(rc);
181302	#endif
181303
181304	return rc;
181305	}
181306
181307	/*
181308	** Undo the effects of sqlite3_initialize(). Must not be called while
	@@ -182369,14 +182644,10 @@
182369	#endif
182370
182371	sqlite3Error(db, SQLITE_OK); /* Deallocates any cached error strings. */
182372	sqlite3ValueFree(db->pErr);
182373	sqlite3CloseExtensions(db);
182374	#if SQLITE_USER_AUTHENTICATION
182375	sqlite3_free(db->auth.zAuthUser);
182376	sqlite3_free(db->auth.zAuthPW);
182377	#endif
182378
182379	db->eOpenState = SQLITE_STATE_ERROR;
182380
182381	/* The temp-database schema is allocated differently from the other schema
182382	** objects (using sqliteMalloc() directly, instead of sqlite3BtreeSchema()).
	@@ -183875,12 +184146,12 @@
183875	}
183876	oldLimit = db->aLimit[limitId];
183877	if( newLimit>=0 ){ /* IMP: R-52476-28732 */
183878	if( newLimit>aHardLimit[limitId] ){
183879	newLimit = aHardLimit[limitId]; /* IMP: R-51463-25634 */
183880	}else if( newLimit<1 && limitId==SQLITE_LIMIT_LENGTH ){
183881	newLimit = 1;
183882	}
183883	db->aLimit[limitId] = newLimit;
183884	}
183885	return oldLimit; /* IMP: R-53341-35419 */
183886	}
	@@ -184395,10 +184666,11 @@
184395	testcase( (1<<(flags&7))==0x04 ); /* READWRITE */
184396	testcase( (1<<(flags&7))==0x40 ); /* READWRITE \| CREATE */
184397	if( ((1<<(flags&7)) & 0x46)==0 ){
184398	rc = SQLITE_MISUSE_BKPT; /* IMP: R-18321-05872 */
184399	}else{

184400	rc = sqlite3ParseUri(zVfs, zFilename, &flags, &db->pVfs, &zOpen, &zErrMsg);
184401	}
184402	if( rc!=SQLITE_OK ){
184403	if( rc==SQLITE_NOMEM ) sqlite3OomFault(db);
184404	sqlite3ErrorWithMsg(db, rc, zErrMsg ? "%s" : 0, zErrMsg);
	@@ -185237,10 +185509,11 @@
185237	#ifndef SQLITE_OMIT_WINDOWFUNC
185238	sqlite3ShowWindow(0);
185239	sqlite3ShowWinFunc(0);
185240	#endif
185241	sqlite3ShowSelect(0);

185242	}
185243	#endif
185244	break;
185245	}
185246
	@@ -185549,28 +185822,10 @@
185549	*pI1 = rLogEst;
185550	*pU64 = sqlite3LogEstToInt(rLogEst);
185551	pI2 = sqlite3LogEst(pU64);
185552	break;
185553	}
185554
185555	#if !defined(SQLITE_OMIT_WSD)
185556	/* sqlite3_test_control(SQLITE_TESTCTRL_USELONGDOUBLE, int X);
185557	**
185558	** X<0 Make no changes to the bUseLongDouble. Just report value.
185559	** X==0 Disable bUseLongDouble
185560	** X==1 Enable bUseLongDouble
185561	** X>=2 Set bUseLongDouble to its default value for this platform
185562	*/
185563	case SQLITE_TESTCTRL_USELONGDOUBLE: {
185564	int b = va_arg(ap, int);
185565	if( b>=2 ) b = hasHighPrecisionDouble(b);
185566	if( b>=0 ) sqlite3Config.bUseLongDouble = b>0;
185567	rc = sqlite3Config.bUseLongDouble!=0;
185568	break;
185569	}
185570	#endif
185571
185572
185573	#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_WSD)
185574	/* sqlite3_test_control(SQLITE_TESTCTRL_TUNE, id, *piValue)
185575	**
185576	** If "id" is an integer between 1 and SQLITE_NTUNE then set the value
	@@ -185875,11 +186130,15 @@
185875	if( db->autoCommit==0 ){
185876	int iDb = sqlite3FindDbName(db, zDb);
185877	if( iDb==0 \|\| iDb>1 ){
185878	Btree *pBt = db->aDb[iDb].pBt;
185879	if( SQLITE_TXN_WRITE!=sqlite3BtreeTxnState(pBt) ){



185880	rc = sqlite3BtreeBeginTrans(pBt, 0, 0);

185881	if( rc==SQLITE_OK ){
185882	rc = sqlite3PagerSnapshotGet(sqlite3BtreePager(pBt), ppSnapshot);
185883	}
185884	}
185885	}
	@@ -189665,14 +189924,19 @@
189665
189666	assert_fts3_nc( p!=0 && p1!=0 && p2!=0 );
189667	if( *p1==POS_COLUMN ){
189668	p1++;
189669	p1 += fts3GetVarint32(p1, &iCol1);



189670	}
189671	if( *p2==POS_COLUMN ){
189672	p2++;
189673	p2 += fts3GetVarint32(p2, &iCol2);


189674	}
189675
189676	while( 1 ){
189677	if( iCol1==iCol2 ){
189678	char *pSave = p;
	@@ -192839,11 +193103,11 @@
192839	for(p=pExpr; p->pLeft; p=p->pLeft){
192840	assert( p->pRight->pPhrase->doclist.nList>0 );
192841	nTmp += p->pRight->pPhrase->doclist.nList;
192842	}
192843	nTmp += p->pPhrase->doclist.nList;
192844	aTmp = sqlite3_malloc64(nTmp*2);
192845	if( !aTmp ){
192846	*pRc = SQLITE_NOMEM;
192847	res = 0;
192848	}else{
192849	char *aPoslist = p->pPhrase->doclist.pList;
	@@ -193490,11 +193754,11 @@
193490	SQLITE_PRIVATE int sqlite3Fts3Corrupt(){
193491	return SQLITE_CORRUPT_VTAB;
193492	}
193493	#endif
193494
193495	#if !SQLITE_CORE
193496	/*
193497	** Initialize API pointer table, if required.
193498	*/
193499	#ifdef _WIN32
193500	__declspec(dllexport)
	@@ -194392,14 +194656,15 @@
194392	rc = pModule->xNext(pCursor, &zByte, &nByte, &iBegin, &iEnd, &iPos);
194393	if( rc==SQLITE_OK ){
194394	Fts3PhraseToken *pToken;
194395
194396	p = fts3ReallocOrFree(p, nSpace + ii*sizeof(Fts3PhraseToken));
194397	if( !p ) goto no_mem;
194398
194399	zTemp = fts3ReallocOrFree(zTemp, nTemp + nByte);
194400	if( !zTemp ) goto no_mem;



194401
194402	assert( nToken==ii );
194403	pToken = &((Fts3Phrase *)(&p[1]))->aToken[ii];
194404	memset(pToken, 0, sizeof(Fts3PhraseToken));
194405
	@@ -194410,53 +194675,51 @@
194410	pToken->isPrefix = (iEnd<nInput && zInput[iEnd]=='*');
194411	pToken->bFirst = (iBegin>0 && zInput[iBegin-1]=='^');
194412	nToken = ii+1;
194413	}
194414	}
194415
194416	pModule->xClose(pCursor);
194417	pCursor = 0;
194418	}
194419
194420	if( rc==SQLITE_DONE ){
194421	int jj;
194422	char *zBuf = 0;
194423
194424	p = fts3ReallocOrFree(p, nSpace + nToken*sizeof(Fts3PhraseToken) + nTemp);
194425	if( !p ) goto no_mem;



194426	memset(p, 0, (char )&(((Fts3Phrase )&p[1])->aToken[0])-(char *)p);
194427	p->eType = FTSQUERY_PHRASE;
194428	p->pPhrase = (Fts3Phrase *)&p[1];
194429	p->pPhrase->iColumn = pParse->iDefaultCol;
194430	p->pPhrase->nToken = nToken;
194431
194432	zBuf = (char *)&p->pPhrase->aToken[nToken];

194433	if( zTemp ){
194434	memcpy(zBuf, zTemp, nTemp);
194435	sqlite3_free(zTemp);
194436	}else{
194437	assert( nTemp==0 );
194438	}
194439
194440	for(jj=0; jj<p->pPhrase->nToken; jj++){
194441	p->pPhrase->aToken[jj].z = zBuf;
194442	zBuf += p->pPhrase->aToken[jj].n;
194443	}
194444	rc = SQLITE_OK;
194445	}
194446
194447	*ppExpr = p;
194448	return rc;
194449	no_mem:
194450
194451	if( pCursor ){
194452	pModule->xClose(pCursor);
194453	}
194454	sqlite3_free(zTemp);
194455	sqlite3_free(p);
194456	*ppExpr = 0;
194457	return SQLITE_NOMEM;



194458	}
194459
194460	/*
194461	** The output variable *ppExpr is populated with an allocated Fts3Expr
194462	** structure, or set to 0 if the end of the input buffer is reached.
	@@ -208867,11 +209130,13 @@
208867	int rawKey = 1;
208868	x = pParse->aBlob[iRoot];
208869	zPath++;
208870	if( zPath[0]=='"' ){
208871	zKey = zPath + 1;
208872	for(i=1; zPath[i] && zPath[i]!='"'; i++){}


208873	nKey = i-1;
208874	if( zPath[i] ){
208875	i++;
208876	}else{
208877	return JSON_LOOKUP_PATHERROR;
	@@ -217779,11 +218044,11 @@
217779	return sqlite3_create_function_v2(db, zQueryFunc, -1, SQLITE_ANY,
217780	(void *)pGeomCtx, geomCallback, 0, 0, rtreeFreeCallback
217781	);
217782	}
217783
217784	#if !SQLITE_CORE
217785	#ifdef _WIN32
217786	__declspec(dllexport)
217787	#endif
217788	SQLITE_API int sqlite3_rtree_init(
217789	sqlite3 *db,
	@@ -218370,11 +218635,11 @@
218370	}
218371
218372	return rc;
218373	}
218374
218375	#if !SQLITE_CORE
218376	#ifdef _WIN32
218377	__declspec(dllexport)
218378	#endif
218379	SQLITE_API int sqlite3_icu_init(
218380	sqlite3 *db,
	@@ -219628,10 +219893,31 @@
219628	struct RbuFrame {
219629	u32 iDbPage;
219630	u32 iWalFrame;
219631	};
219632





















219633	/*
219634	** RBU handle.
219635	**
219636	** nPhaseOneStep:
219637	** If the RBU database contains an rbu_count table, this value is set to
	@@ -219679,11 +219965,11 @@
219679	char zState; / Path to state db (or NULL if zRbu) */
219680	char zStateDb[5]; /* Db name for state ("stat" or "main") */
219681	int rc; /* Value returned by last rbu_step() call */
219682	char zErrmsg; / Error message if rc!=SQLITE_OK */
219683	int nStep; /* Rows processed for current object */
219684	int nProgress; /* Rows processed for all objects */
219685	RbuObjIter objiter; /* Iterator for skipping through tbl/idx */
219686	const char zVfsName; / Name of automatically created rbu vfs */
219687	rbu_file pTargetFd; / File handle open on target db */
219688	int nPagePerSector; /* Pages per sector for pTargetFd */
219689	i64 iOalSz;
	@@ -219796,11 +220082,11 @@
219796	unsigned char *zStart = z;
219797	while( (c = zValue[0x7f&*(z++)])>=0 ){
219798	v = (v<<6) + c;
219799	}
219800	z--;
219801	*pLen -= z - zStart;
219802	pz = (char)z;
219803	return v;
219804	}
219805
219806	#if RBU_ENABLE_DELTA_CKSUM
	@@ -219981,10 +220267,11 @@
219981	int nOut;
219982	int nOut2;
219983	char *aOut;
219984
219985	assert( argc==2 );

219986
219987	nOrig = sqlite3_value_bytes(argv[0]);
219988	aOrig = (const char*)sqlite3_value_blob(argv[0]);
219989	nDelta = sqlite3_value_bytes(argv[1]);
219990	aDelta = (const char*)sqlite3_value_blob(argv[1]);
	@@ -221560,17 +221847,17 @@
221560	nParen++;
221561	}
221562	else if( c==')' ){
221563	nParen--;
221564	if( nParen==0 ){
221565	int nSpan = &zSql[i] - pIter->aIdxCol[iIdxCol].zSpan;
221566	pIter->aIdxCol[iIdxCol++].nSpan = nSpan;
221567	i++;
221568	break;
221569	}
221570	}else if( c==',' && nParen==1 ){
221571	int nSpan = &zSql[i] - pIter->aIdxCol[iIdxCol].zSpan;
221572	pIter->aIdxCol[iIdxCol++].nSpan = nSpan;
221573	pIter->aIdxCol[iIdxCol].zSpan = &zSql[i+1];
221574	}else if( c=='"' \|\| c=='\'' \|\| c=='`' ){
221575	for(i++; 1; i++){
221576	if( zSql[i]==c ){
	@@ -222256,10 +222543,12 @@
222256	int i, sz;
222257	sz = (int)strlen(z)&0xffffff;
222258	for(i=sz-1; i>0 && z[i]!='/' && z[i]!='.'; i--){}
222259	if( z[i]=='.' && sz>i+4 ) memmove(&z[i+1], &z[sz-3], 4);
222260	}


222261	#endif
222262	}
222263
222264	/*
222265	** Return the current wal-index header checksum for the target database
	@@ -222840,11 +223129,11 @@
222840	"INSERT OR REPLACE INTO %s.rbu_state(k, v) VALUES "
222841	"(%d, %d), "
222842	"(%d, %Q), "
222843	"(%d, %Q), "
222844	"(%d, %d), "
222845	"(%d, %d), "
222846	"(%d, %lld), "
222847	"(%d, %lld), "
222848	"(%d, %lld), "
222849	"(%d, %lld), "
222850	"(%d, %Q) ",
	@@ -223198,10 +223487,11 @@
223198	char *zErrmsg = 0;
223199	int rc;
223200	sqlite3 *db = (rbuIsVacuum(p) ? p->dbRbu : p->dbMain);
223201
223202	assert( nVal==1 );

223203
223204	rc = prepareFreeAndCollectError(db, &pStmt, &zErrmsg,
223205	sqlite3_mprintf("SELECT count(*) FROM sqlite_schema "
223206	"WHERE type='index' AND tbl_name = %Q", sqlite3_value_text(apVal[0]))
223207	);
	@@ -223473,11 +223763,11 @@
223473	const char *zTarget,
223474	const char *zState
223475	){
223476	if( zTarget==0 ){ return rbuMisuseError(); }
223477	if( zState ){
223478	int n = strlen(zState);
223479	if( n>=7 && 0==memcmp("-vactmp", &zState[n-7], 7) ){
223480	return rbuMisuseError();
223481	}
223482	}
223483	/* TODO: Check that both arguments are non-NULL */
	@@ -223690,10 +223980,11 @@
223690	/*
223691	** Default xRename callback for RBU.
223692	*/
223693	static int xDefaultRename(void pArg, const char zOld, const char *zNew){
223694	int rc = SQLITE_OK;

223695	#if defined(_WIN32_WCE)
223696	{
223697	LPWSTR zWideOld;
223698	LPWSTR zWideNew;
223699
	@@ -224594,10 +224885,13 @@
224594
224595	/*
224596	** No-op.
224597	*/
224598	static int rbuVfsGetLastError(sqlite3_vfs pVfs, int a, char b){



224599	return 0;
224600	}
224601
224602	/*
224603	** Deregister and destroy an RBU vfs created by an earlier call to
	@@ -225650,11 +225944,17 @@
225650	** schema for the database file that is to be read. The default schema is
225651	** "main".
225652	**
225653	** The data field of sqlite_dbpage table can be updated. The new
225654	** value must be a BLOB which is the correct page size, otherwise the
225655	** update fails. Rows may not be deleted or inserted.






225656	*/
225657
225658	/* #include "sqliteInt.h" Requires access to internal data structures */
225659	#if (defined(SQLITE_ENABLE_DBPAGE_VTAB) \|\| defined(SQLITE_TEST)) \
225660	&& !defined(SQLITE_OMIT_VIRTUALTABLE)
	@@ -225673,17 +225973,18 @@
225673	};
225674
225675	struct DbpageTable {
225676	sqlite3_vtab base; /* Base class. Must be first */
225677	sqlite3 db; / The database */


225678	};
225679
225680	/* Columns */
225681	#define DBPAGE_COLUMN_PGNO 0
225682	#define DBPAGE_COLUMN_DATA 1
225683	#define DBPAGE_COLUMN_SCHEMA 2
225684
225685
225686
225687	/*
225688	** Connect to or create a dbpagevfs virtual table.
225689	*/
	@@ -225943,15 +226244,15 @@
225943	DbpageTable pTab = (DbpageTable )pVtab;
225944	Pgno pgno;
225945	DbPage *pDbPage = 0;
225946	int rc = SQLITE_OK;
225947	char *zErr = 0;
225948	const char *zSchema;
225949	int iDb;
225950	Btree *pBt;
225951	Pager *pPager;
225952	int szPage;

225953
225954	(void)pRowid;
225955	if( pTab->db->flags & SQLITE_Defensive ){
225956	zErr = "read-only";
225957	goto update_fail;
	@@ -225958,45 +226259,62 @@
225958	}
225959	if( argc==1 ){
225960	zErr = "cannot delete";
225961	goto update_fail;
225962	}
225963	pgno = sqlite3_value_int(argv[0]);
225964	if( sqlite3_value_type(argv[0])==SQLITE_NULL
225965	\|\| (Pgno)sqlite3_value_int(argv[1])!=pgno
225966	){
225967	zErr = "cannot insert";
225968	goto update_fail;
225969	}
225970	zSchema = (const char*)sqlite3_value_text(argv[4]);
225971	iDb = ALWAYS(zSchema) ? sqlite3FindDbName(pTab->db, zSchema) : -1;
225972	if( NEVER(iDb<0) ){
225973	zErr = "no such schema";
225974	goto update_fail;








225975	}
225976	pBt = pTab->db->aDb[iDb].pBt;
225977	if( NEVER(pgno<1) \|\| NEVER(pBt==0) \|\| NEVER(pgno>sqlite3BtreeLastPage(pBt)) ){
225978	zErr = "bad page number";
225979	goto update_fail;
225980	}
225981	szPage = sqlite3BtreeGetPageSize(pBt);
225982	if( sqlite3_value_type(argv[3])!=SQLITE_BLOB
225983	\|\| sqlite3_value_bytes(argv[3])!=szPage
225984	){
225985	zErr = "bad page value";
225986	goto update_fail;








225987	}
225988	pPager = sqlite3BtreePager(pBt);
225989	rc = sqlite3PagerGet(pPager, pgno, (DbPage**)&pDbPage, 0);
225990	if( rc==SQLITE_OK ){
225991	const void *pData = sqlite3_value_blob(argv[3]);
225992	assert( pData!=0 \|\| pTab->db->mallocFailed );
225993	if( pData
225994	&& (rc = sqlite3PagerWrite(pDbPage))==SQLITE_OK
225995	){
225996	memcpy(sqlite3PagerGetData(pDbPage), pData, szPage);
225997	}


225998	}
225999	sqlite3PagerUnref(pDbPage);
226000	return rc;
226001
226002	update_fail:
	@@ -226015,13 +226333,35 @@
226015	int i;
226016	for(i=0; i<db->nDb; i++){
226017	Btree *pBt = db->aDb[i].pBt;
226018	if( pBt ) (void)sqlite3BtreeBeginTrans(pBt, 1, 0);
226019	}























226020	return SQLITE_OK;
226021	}
226022
226023
226024	/*
226025	** Invoke this routine to register the "dbpage" virtual table module
226026	*/
226027	SQLITE_PRIVATE int sqlite3DbpageRegister(sqlite3 *db){
	@@ -226039,18 +226379,18 @@
226039	dbpageEof, /* xEof - check for end of scan */
226040	dbpageColumn, /* xColumn - read data */
226041	dbpageRowid, /* xRowid - read data */
226042	dbpageUpdate, /* xUpdate */
226043	dbpageBegin, /* xBegin */
226044	0, /* xSync */
226045	0, /* xCommit */
226046	0, /* xRollback */
226047	0, /* xFindMethod */
226048	0, /* xRename */
226049	0, /* xSavepoint */
226050	0, /* xRelease */
226051	0, /* xRollbackTo */
226052	0, /* xShadowName */
226053	0 /* xIntegrity */
226054	};
226055	return sqlite3_create_module(db, "sqlite_dbpage", &dbpage_module, 0);
226056	}
	@@ -226134,10 +226474,14 @@
226134	/*
226135	** An object of this type is used internally as an abstraction for
226136	** input data. Input data may be supplied either as a single large buffer
226137	** (e.g. sqlite3changeset_start()) or using a stream function (e.g.
226138	** sqlite3changeset_start_strm()).




226139	*/
226140	struct SessionInput {
226141	int bNoDiscard; /* If true, do not discard in InputBuffer() */
226142	int iCurrent; /* Offset in aData[] of current change */
226143	int iNext; /* Offset in aData[] of next change */
	@@ -227817,20 +228161,23 @@
227817	/* Figure out how large an allocation is required */
227818	nByte = sizeof(SessionChange);
227819	for(i=0; i<(pTab->nCol-pTab->bRowid); i++){
227820	sqlite3_value *p = 0;
227821	if( op!=SQLITE_INSERT ){
227822	TESTONLY(int trc = ) pSession->hook.xOld(pSession->hook.pCtx, i, &p);
227823	assert( trc==SQLITE_OK );

227824	}else if( pTab->abPK[i] ){
227825	TESTONLY(int trc = ) pSession->hook.xNew(pSession->hook.pCtx, i, &p);
227826	assert( trc==SQLITE_OK );
227827	}
227828
227829	/* This may fail if SQLite value p contains a utf-16 string that must
227830	** be converted to utf-8 and an OOM error occurs while doing so. */
227831	rc = sessionSerializeValue(0, p, &nByte);


227832	if( rc!=SQLITE_OK ) goto error_out;
227833	}
227834	if( pTab->bRowid ){
227835	nByte += 9; /* Size of rowid field - an integer */
227836	}
	@@ -231184,19 +231531,25 @@
231184	int rc = SQLITE_OK; /* Return code */
231185	const char zTab = 0; / Name of current table */
231186	int nTab = 0; /* Result of sqlite3Strlen30(zTab) */
231187	SessionApplyCtx sApply; /* changeset_apply() context object */
231188	int bPatchset;

231189
231190	assert( xConflict!=0 );






231191
231192	pIter->in.bNoDiscard = 1;
231193	memset(&sApply, 0, sizeof(sApply));
231194	sApply.bRebase = (ppRebase && pnRebase);
231195	sApply.bInvertConstraints = !!(flags & SQLITE_CHANGESETAPPLY_INVERT);
231196	sApply.bIgnoreNoop = !!(flags & SQLITE_CHANGESETAPPLY_IGNORENOOP);
231197	sqlite3_mutex_enter(sqlite3_db_mutex(db));
231198	if( (flags & SQLITE_CHANGESETAPPLY_NOSAVEPOINT)==0 ){
231199	rc = sqlite3_exec(db, "SAVEPOINT changeset_apply", 0, 0, 0);
231200	}
231201	if( rc==SQLITE_OK ){
231202	rc = sqlite3_exec(db, "PRAGMA defer_foreign_keys = 1", 0, 0, 0);
	@@ -231354,10 +231707,16 @@
231354	sqlite3_finalize(sApply.pDelete);
231355	sqlite3_finalize(sApply.pSelect);
231356	sqlite3_free((char)sApply.azCol); / cast works around VC++ bug */
231357	sqlite3_free((char*)sApply.constraints.aBuf);
231358	sqlite3_free((char*)sApply.rebase.aBuf);






231359	sqlite3_mutex_leave(sqlite3_db_mutex(db));
231360	return rc;
231361	}
231362
231363	/*
	@@ -231382,28 +231741,17 @@
231382	int flags
231383	){
231384	sqlite3_changeset_iter pIter; / Iterator to skip through changeset */
231385	int bInv = !!(flags & SQLITE_CHANGESETAPPLY_INVERT);
231386	int rc = sessionChangesetStart(&pIter, 0, 0, nChangeset, pChangeset, bInv, 1);
231387	u64 savedFlag = db->flags & SQLITE_FkNoAction;
231388
231389	if( flags & SQLITE_CHANGESETAPPLY_FKNOACTION ){
231390	db->flags \|= ((u64)SQLITE_FkNoAction);
231391	db->aDb[0].pSchema->schema_cookie -= 32;
231392	}
231393
231394	if( rc==SQLITE_OK ){
231395	rc = sessionChangesetApply(
231396	db, pIter, xFilter, xConflict, pCtx, ppRebase, pnRebase, flags
231397	);
231398	}
231399
231400	if( (flags & SQLITE_CHANGESETAPPLY_FKNOACTION) && savedFlag==0 ){
231401	assert( db->flags & SQLITE_FkNoAction );
231402	db->flags &= ~((u64)SQLITE_FkNoAction);
231403	db->aDb[0].pSchema->schema_cookie -= 32;
231404	}
231405	return rc;
231406	}
231407
231408	/*
231409	** Apply the changeset passed via pChangeset/nChangeset to the main database
	@@ -231720,10 +232068,13 @@
231720	if( op==SQLITE_INSERT \|\| (op==SQLITE_DELETE && pGrp->bPatch==0) ){
231721	/* Append the missing default column values to the record. */
231722	sessionAppendBlob(pOut, aRec, nRec, &rc);
231723	if( rc==SQLITE_OK && pTab->pDfltStmt==0 ){
231724	rc = sessionPrepareDfltStmt(pGrp->db, pTab, &pTab->pDfltStmt);



231725	}
231726	for(ii=nCol; rc==SQLITE_OK && ii<pTab->nCol; ii++){
231727	int eType = sqlite3_column_type(pTab->pDfltStmt, ii);
231728	sessionAppendByte(pOut, eType, &rc);
231729	switch( eType ){
	@@ -231736,10 +232087,11 @@
231736	double rVal = sqlite3_column_int64(pTab->pDfltStmt, ii);
231737	memcpy(&iVal, &rVal, sizeof(i64));
231738	}
231739	if( SQLITE_OK==sessionBufferGrow(pOut, 8, &rc) ){
231740	sessionPutI64(&pOut->aBuf[pOut->nBuf], iVal);

231741	}
231742	break;
231743	}
231744
231745	case SQLITE_BLOB:
	@@ -231874,10 +232226,12 @@
231874	SessionChange *pExist = 0;
231875	SessionChange **pp = 0;
231876	SessionTable *pTab = 0;
231877	u8 *aRec = &pIter->in.aData[pIter->in.iCurrent + 2];
231878	int nRec = (pIter->in.iNext - pIter->in.iCurrent) - 2;


231879
231880	/* Ensure that only changesets, or only patchsets, but not a mixture
231881	** of both, are being combined. It is an error to try to combine a
231882	** changeset and a patchset. */
231883	if( pGrp->pList==0 ){
	@@ -231952,10 +232306,11 @@
231952	){
231953	u8 *aRec;
231954	int nRec;
231955	int rc = SQLITE_OK;
231956

231957	while( SQLITE_ROW==(sessionChangesetNext(pIter, &aRec, &nRec, 0)) ){
231958	rc = sessionOneChangeToHash(pGrp, pIter, bRebase);
231959	if( rc!=SQLITE_OK ) break;
231960	}
231961
	@@ -232583,20 +232938,46 @@
232583	#endif /* SQLITE_ENABLE_SESSION && SQLITE_ENABLE_PREUPDATE_HOOK */
232584
232585	/************ End of sqlite3session.c ************************************/
232586	/************ Begin file fts5.c ******************************************/
232587
232588




















232589	#if !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_FTS5)
232590
232591	#if !defined(NDEBUG) && !defined(SQLITE_DEBUG)
232592	# define NDEBUG 1
232593	#endif
232594	#if defined(NDEBUG) && defined(SQLITE_DEBUG)
232595	# undef NDEBUG
232596	#endif
232597






232598	/*
232599	** 2014 May 31
232600	**
232601	** The author disclaims copyright to this source code. In place of
232602	** a legal notice, here is a blessing:
	@@ -232893,17 +233274,32 @@
232893	** This is used to access token iToken of phrase hit iIdx within the
232894	** current row. If iIdx is less than zero or greater than or equal to the
232895	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
232896	** output variable (*ppToken) is set to point to a buffer containing the
232897	** matching document token, and (*pnToken) to the size of that buffer in
232898	** bytes. This API is not available if the specified token matches a
232899	** prefix query term. In that case both output variables are always set
232900	** to 0.
232901	**
232902	** The output text is not a copy of the document text that was tokenized.
232903	** It is the output of the tokenizer module. For tokendata=1 tables, this
232904	** includes any embedded 0x00 and trailing data.

















232905	**
232906	** This API can be quite slow if used with an FTS5 table created with the
232907	** "detail=none" or "detail=column" option.
232908	**
232909	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
	@@ -232991,11 +233387,10 @@
232991	** CUSTOM TOKENIZERS
232992	**
232993	** Applications may also register custom tokenizer types. A tokenizer
232994	** is registered by providing fts5 with a populated instance of the
232995	** following structure. All structure methods must be defined, setting
232996	**
232997	** any member of the fts5_tokenizer struct to NULL leads to undefined
232998	** behaviour. The structure methods are expected to function as follows:
232999	**
233000	** xCreate:
233001	** This function is used to allocate and initialize a tokenizer instance.
	@@ -233560,10 +233955,11 @@
233560	u8 abUnindexed; / True for unindexed columns */
233561	int nPrefix; /* Number of prefix indexes */
233562	int aPrefix; / Sizes in bytes of nPrefix prefix indexes */
233563	int eContent; /* An FTS5_CONTENT value */
233564	int bContentlessDelete; /* "contentless_delete=" option (dflt==0) */

233565	char zContent; / content table */
233566	char zContentRowid; / "content_rowid=" option value */
233567	int bColumnsize; /* "columnsize=" option value (dflt==1) */
233568	int bTokendata; /* "tokendata=" option value (dflt==0) */
233569	int bLocale; /* "locale=" option value (dflt==0) */
	@@ -233582,11 +233978,12 @@
233582	int nUsermerge; /* 'usermerge' setting */
233583	int nHashSize; /* Bytes of memory for in-memory hash */
233584	char zRank; / Name of rank function */
233585	char zRankArgs; / Arguments to rank function */
233586	int bSecureDelete; /* 'secure-delete' */
233587	int nDeleteMerge; /* 'deletemerge' */

233588
233589	/* If non-NULL, points to sqlite3_vtab.base.zErrmsg. Often NULL. */
233590	char **pzErrmsg;
233591
233592	#ifdef SQLITE_DEBUG
	@@ -233598,13 +233995,14 @@
233598	** the expected version if the 'secure-delete' option has ever been
233599	** set on the table. */
233600	#define FTS5_CURRENT_VERSION 4
233601	#define FTS5_CURRENT_VERSION_SECUREDELETE 5
233602
233603	#define FTS5_CONTENT_NORMAL 0
233604	#define FTS5_CONTENT_NONE 1
233605	#define FTS5_CONTENT_EXTERNAL 2

233606
233607	#define FTS5_DETAIL_FULL 0
233608	#define FTS5_DETAIL_NONE 1
233609	#define FTS5_DETAIL_COLUMNS 2
233610
	@@ -233838,11 +234236,18 @@
233838	static int sqlite3Fts5StructureTest(Fts5Index, void);
233839
233840	/*
233841	** Used by xInstToken():
233842	*/
233843	static int sqlite3Fts5IterToken(Fts5IndexIter, i64, int, int, const char, int);







233844
233845	/*
233846	** Insert or remove data to or from the index. Each time a document is
233847	** added to or removed from the index, this function is called one or more
233848	** times.
	@@ -233972,20 +234377,17 @@
233972
233973	static Fts5Table sqlite3Fts5TableFromCsrid(Fts5Global, i64);
233974
233975	static int sqlite3Fts5FlushToDisk(Fts5Table*);
233976
233977	static int sqlite3Fts5ExtractText(
233978	Fts5Config *pConfig,
233979	sqlite3_value pVal, / Value to extract text from */
233980	int bContent, /* Loaded from content table */
233981	int pbResetTokenizer, / OUT: True if ClearLocale() required */
233982	const char *ppText, / OUT: Pointer to text buffer */
233983	int pnText / OUT: Size of (ppText) in bytes /
233984	);
233985
233986	static void sqlite3Fts5ClearLocale(Fts5Config *pConfig);






233987
233988	/*
233989	** End of interface to code in fts5.c.
233990	**************************************************************************/
233991
	@@ -234063,11 +234465,11 @@
234063
234064	static int sqlite3Fts5DropAll(Fts5Config*);
234065	static int sqlite3Fts5CreateTable(Fts5Config, const char, const char, int, char *);
234066
234067	static int sqlite3Fts5StorageDelete(Fts5Storage p, i64, sqlite3_value*, int);
234068	static int sqlite3Fts5StorageContentInsert(Fts5Storage p, sqlite3_value, i64);
234069	static int sqlite3Fts5StorageIndexInsert(Fts5Storage p, sqlite3_value*, i64);
234070
234071	static int sqlite3Fts5StorageIntegrity(Fts5Storage *p, int iArg);
234072
234073	static int sqlite3Fts5StorageStmt(Fts5Storage p, int eStmt, sqlite3_stmt, char*);
	@@ -237270,10 +237672,11 @@
237270	const char zArg, / Argument to parse */
237271	char *pzErr / OUT: Error message */
237272	){
237273	int rc = SQLITE_OK;
237274	int nCmd = (int)strlen(zCmd);

237275	if( sqlite3_strnicmp("prefix", zCmd, nCmd)==0 ){
237276	const int nByte = sizeof(int) * FTS5_MAX_PREFIX_INDEXES;
237277	const char *p;
237278	int bFirst = 1;
237279	if( pConfig->aPrefix==0 ){
	@@ -237388,10 +237791,20 @@
237388	}else{
237389	pConfig->bContentlessDelete = (zArg[0]=='1');
237390	}
237391	return rc;
237392	}










237393
237394	if( sqlite3_strnicmp("content_rowid", zCmd, nCmd)==0 ){
237395	if( pConfig->zContentRowid ){
237396	*pzErr = sqlite3_mprintf("multiple content_rowid=... directives");
237397	rc = SQLITE_ERROR;
	@@ -237506,11 +237919,12 @@
237506
237507	static int fts5ConfigParseColumn(
237508	Fts5Config *p,
237509	char *zCol,
237510	char *zArg,
237511	char **pzErr

237512	){
237513	int rc = SQLITE_OK;
237514	if( 0==sqlite3_stricmp(zCol, FTS5_RANK_NAME)
237515	\|\| 0==sqlite3_stricmp(zCol, FTS5_ROWID_NAME)
237516	){
	@@ -237517,10 +237931,11 @@
237517	*pzErr = sqlite3_mprintf("reserved fts5 column name: %s", zCol);
237518	rc = SQLITE_ERROR;
237519	}else if( zArg ){
237520	if( 0==sqlite3_stricmp(zArg, "unindexed") ){
237521	p->abUnindexed[p->nCol] = 1;

237522	}else{
237523	*pzErr = sqlite3_mprintf("unrecognized column option: %s", zArg);
237524	rc = SQLITE_ERROR;
237525	}
237526	}
	@@ -237537,15 +237952,30 @@
237537	int rc = SQLITE_OK;
237538	Fts5Buffer buf = {0, 0, 0};
237539
237540	sqlite3Fts5BufferAppendPrintf(&rc, &buf, "T.%Q", p->zContentRowid);
237541	if( p->eContent!=FTS5_CONTENT_NONE ){




237542	for(i=0; i<p->nCol; i++){
237543	if( p->eContent==FTS5_CONTENT_EXTERNAL ){
237544	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.%Q", p->azCol[i]);
237545	}else{
237546	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.c%d", i);











237547	}
237548	}
237549	}
237550
237551	assert( p->zContentExprlist==0 );
	@@ -237575,10 +238005,11 @@
237575	){
237576	int rc = SQLITE_OK; /* Return code */
237577	Fts5Config pRet; / New object to return */
237578	int i;
237579	sqlite3_int64 nByte;

237580
237581	ppOut = pRet = (Fts5Config)sqlite3_malloc(sizeof(Fts5Config));
237582	if( pRet==0 ) return SQLITE_NOMEM;
237583	memset(pRet, 0, sizeof(Fts5Config));
237584	pRet->pGlobal = pGlobal;
	@@ -237634,11 +238065,11 @@
237634	ALWAYS(zOne)?zOne:"",
237635	zTwo?zTwo:"",
237636	pzErr
237637	);
237638	}else{
237639	rc = fts5ConfigParseColumn(pRet, zOne, zTwo, pzErr);
237640	zOne = 0;
237641	}
237642	}
237643	}
237644
	@@ -237665,18 +238096,34 @@
237665	*pzErr = sqlite3_mprintf(
237666	"contentless_delete=1 is incompatible with columnsize=0"
237667	);
237668	rc = SQLITE_ERROR;
237669	}













237670
237671	/* If no zContent option was specified, fill in the default values. */
237672	if( rc==SQLITE_OK && pRet->zContent==0 ){
237673	const char *zTail = 0;
237674	assert( pRet->eContent==FTS5_CONTENT_NORMAL
237675	\|\| pRet->eContent==FTS5_CONTENT_NONE
237676	);
237677	if( pRet->eContent==FTS5_CONTENT_NORMAL ){



237678	zTail = "content";
237679	}else if( pRet->bColumnsize ){
237680	zTail = "docsize";
237681	}
237682
	@@ -238010,10 +238457,23 @@
238010	if( bVal<0 ){
238011	*pbBadkey = 1;
238012	}else{
238013	pConfig->bSecureDelete = (bVal ? 1 : 0);
238014	}













238015	}else{
238016	*pbBadkey = 1;
238017	}
238018	return rc;
238019	}
	@@ -241145,11 +241605,11 @@
241145	&& memcmp(pT->pTerm, pToken, pT->nQueryTerm)==0
241146	){
241147	int rc = sqlite3Fts5PoslistWriterAppend(
241148	&pExpr->apExprPhrase[i]->poslist, &p->aPopulator[i].writer, p->iOff
241149	);
241150	if( rc==SQLITE_OK && pExpr->pConfig->bTokendata && !pT->bPrefix ){
241151	int iCol = p->iOff>>32;
241152	int iTokOff = p->iOff & 0x7FFFFFFF;
241153	rc = sqlite3Fts5IndexIterWriteTokendata(
241154	pT->pIter, pToken, nToken, iRowid, iCol, iTokOff
241155	);
	@@ -241338,19 +241798,18 @@
241338	pPhrase = pExpr->apExprPhrase[iPhrase];
241339	if( iToken<0 \|\| iToken>=pPhrase->nTerm ){
241340	return SQLITE_RANGE;
241341	}
241342	pTerm = &pPhrase->aTerm[iToken];
241343	if( pTerm->bPrefix==0 ){
241344	if( pExpr->pConfig->bTokendata ){
241345	rc = sqlite3Fts5IterToken(
241346	pTerm->pIter, iRowid, iCol, iOff+iToken, ppOut, pnOut
241347	);
241348	}else{
241349	*ppOut = pTerm->pTerm;
241350	*pnOut = pTerm->nFullTerm;
241351	}
241352	}
241353	return rc;
241354	}
241355
241356	/*
	@@ -246847,10 +247306,15 @@
246847	if( nPercent>=pConfig->nDeleteMerge && nPercent>nBest ){
246848	iRet = ii;
246849	nBest = nPercent;
246850	}
246851	}





246852	}
246853	}
246854	return iRet;
246855	}
246856
	@@ -248156,10 +248620,387 @@
248156	fts5BufferFree(&tmp);
248157	memset(&out.p[out.n], 0, FTS5_DATA_ZERO_PADDING);
248158	*p1 = out;
248159	}
248160

























































































































































































































































































































































































248161	static void fts5SetupPrefixIter(
248162	Fts5Index p, / Index to read from */
248163	int bDesc, /* True for "ORDER BY rowid DESC" */
248164	int iIdx, /* Index to scan for data */
248165	u8 pToken, / Buffer containing prefix to match */
	@@ -248166,137 +249007,89 @@
248166	int nToken, /* Size of buffer pToken in bytes */
248167	Fts5Colset pColset, / Restrict matches to these columns */
248168	Fts5Iter *ppIter / OUT: New iterator */
248169	){
248170	Fts5Structure *pStruct;
248171	Fts5Buffer *aBuf;
248172	int nBuf = 32;
248173	int nMerge = 1;
248174
248175	void (xMerge)(Fts5Index, Fts5Buffer, int, Fts5Buffer);
248176	void (xAppend)(Fts5Index, u64, Fts5Iter, Fts5Buffer);












248177	if( p->pConfig->eDetail==FTS5_DETAIL_NONE ){
248178	xMerge = fts5MergeRowidLists;
248179	xAppend = fts5AppendRowid;
248180	}else{
248181	nMerge = FTS5_MERGE_NLIST-1;
248182	nBuf = nMerge8; / Sufficient to merge (16^8)==(2^32) lists */
248183	xMerge = fts5MergePrefixLists;
248184	xAppend = fts5AppendPoslist;
248185	}
248186
248187	aBuf = (Fts5Buffer)fts5IdxMalloc(p, sizeof(Fts5Buffer)nBuf);
248188	pStruct = fts5StructureRead(p);
248189	assert( p->rc!=SQLITE_OK \|\| (aBuf && pStruct) );
248190
248191	if( p->rc==SQLITE_OK ){
248192	const int flags = FTS5INDEX_QUERY_SCAN
248193	\| FTS5INDEX_QUERY_SKIPEMPTY
248194	\| FTS5INDEX_QUERY_NOOUTPUT;
248195	int i;
248196	i64 iLastRowid = 0;
248197	Fts5Iter p1 = 0; / Iterator used to gather data from index */
248198	Fts5Data *pData;
248199	Fts5Buffer doclist;
248200	int bNewTerm = 1;
248201
248202	memset(&doclist, 0, sizeof(doclist));
248203
248204	/* If iIdx is non-zero, then it is the number of a prefix-index for
248205	** prefixes 1 character longer than the prefix being queried for. That
248206	** index contains all the doclists required, except for the one
248207	** corresponding to the prefix itself. That one is extracted from the
248208	** main term index here. */
248209	if( iIdx!=0 ){
248210	int dummy = 0;
248211	const int f2 = FTS5INDEX_QUERY_SKIPEMPTY\|FTS5INDEX_QUERY_NOOUTPUT;
248212	pToken[0] = FTS5_MAIN_PREFIX;
248213	fts5MultiIterNew(p, pStruct, f2, pColset, pToken, nToken, -1, 0, &p1);
248214	fts5IterSetOutputCb(&p->rc, p1);
248215	for(;
248216	fts5MultiIterEof(p, p1)==0;
248217	fts5MultiIterNext2(p, p1, &dummy)
248218	){
248219	Fts5SegIter *pSeg = &p1->aSeg[ p1->aFirst[1].iFirst ];
248220	p1->xSetOutputs(p1, pSeg);
248221	if( p1->base.nData ){
248222	xAppend(p, (u64)p1->base.iRowid-(u64)iLastRowid, p1, &doclist);
248223	iLastRowid = p1->base.iRowid;
248224	}
248225	}
248226	fts5MultiIterFree(p1);
248227	}
248228
248229	pToken[0] = FTS5_MAIN_PREFIX + iIdx;
248230	fts5MultiIterNew(p, pStruct, flags, pColset, pToken, nToken, -1, 0, &p1);
248231	fts5IterSetOutputCb(&p->rc, p1);
248232
248233	for( /* no-op */ ;
248234	fts5MultiIterEof(p, p1)==0;
248235	fts5MultiIterNext2(p, p1, &bNewTerm)
248236	){
248237	Fts5SegIter *pSeg = &p1->aSeg[ p1->aFirst[1].iFirst ];
248238	int nTerm = pSeg->term.n;
248239	const u8 *pTerm = pSeg->term.p;
248240	p1->xSetOutputs(p1, pSeg);
248241
248242	assert_nc( memcmp(pToken, pTerm, MIN(nToken, nTerm))<=0 );
248243	if( bNewTerm ){
248244	if( nTerm<nToken \|\| memcmp(pToken, pTerm, nToken) ) break;
248245	}
248246
248247	if( p1->base.nData==0 ) continue;
248248	if( p1->base.iRowid<=iLastRowid && doclist.n>0 ){
248249	for(i=0; p->rc==SQLITE_OK && doclist.n; i++){
248250	int i1 = i*nMerge;
248251	int iStore;
248252	assert( i1+nMerge<=nBuf );
248253	for(iStore=i1; iStore<i1+nMerge; iStore++){
248254	if( aBuf[iStore].n==0 ){
248255	fts5BufferSwap(&doclist, &aBuf[iStore]);
248256	fts5BufferZero(&doclist);
248257	break;
248258	}
248259	}
248260	if( iStore==i1+nMerge ){
248261	xMerge(p, &doclist, nMerge, &aBuf[i1]);
248262	for(iStore=i1; iStore<i1+nMerge; iStore++){
248263	fts5BufferZero(&aBuf[iStore]);
248264	}
248265	}
248266	}
248267	iLastRowid = 0;
248268	}
248269
248270	xAppend(p, (u64)p1->base.iRowid-(u64)iLastRowid, p1, &doclist);
248271	iLastRowid = p1->base.iRowid;
248272	}
248273
248274	assert( (nBuf%nMerge)==0 );
248275	for(i=0; i<nBuf; i+=nMerge){
248276	int iFree;
248277	if( p->rc==SQLITE_OK ){
248278	xMerge(p, &doclist, nMerge, &aBuf[i]);
248279	}
248280	for(iFree=i; iFree<i+nMerge; iFree++){
248281	fts5BufferFree(&aBuf[iFree]);
248282	}
248283	}
248284	fts5MultiIterFree(p1);
248285
248286	pData = fts5IdxMalloc(p, sizeof(*pData)+doclist.n+FTS5_DATA_ZERO_PADDING);
248287	if( pData ){
248288	pData->p = (u8*)&pData[1];
248289	pData->nn = pData->szLeaf = doclist.n;
248290	if( doclist.n ) memcpy(pData->p, doclist.p, doclist.n);
248291	fts5MultiIterNew2(p, pData, bDesc, ppIter);
248292	}
248293	fts5BufferFree(&doclist);





248294	}
248295


248296	fts5StructureRelease(pStruct);
248297	sqlite3_free(aBuf);
248298	}
248299
248300
248301	/*
248302	** Indicate that all subsequent calls to sqlite3Fts5IndexWrite() pertain
	@@ -248546,42 +249339,10 @@
248546	static void fts5SegIterSetEOF(Fts5SegIter *pSeg){
248547	fts5DataRelease(pSeg->pLeaf);
248548	pSeg->pLeaf = 0;
248549	}
248550
248551	/*
248552	** Usually, a tokendata=1 iterator (struct Fts5TokenDataIter) accumulates an
248553	** array of these for each row it visits. Or, for an iterator used by an
248554	** "ORDER BY rank" query, it accumulates an array of these for the entire
248555	** query.
248556	**
248557	** Each instance in the array indicates the iterator (and therefore term)
248558	** associated with position iPos of rowid iRowid. This is used by the
248559	** xInstToken() API.
248560	*/
248561	struct Fts5TokenDataMap {
248562	i64 iRowid; /* Row this token is located in */
248563	i64 iPos; /* Position of token */
248564	int iIter; /* Iterator token was read from */
248565	};
248566
248567	/*
248568	** An object used to supplement Fts5Iter for tokendata=1 iterators.
248569	*/
248570	struct Fts5TokenDataIter {
248571	int nIter;
248572	int nIterAlloc;
248573
248574	int nMap;
248575	int nMapAlloc;
248576	Fts5TokenDataMap *aMap;
248577
248578	Fts5PoslistReader *aPoslistReader;
248579	int *aPoslistToIter;
248580	Fts5Iter *apIter[1];
248581	};
248582
248583	/*
248584	** This function appends iterator pAppend to Fts5TokenDataIter pIn and
248585	** returns the result.
248586	*/
248587	static Fts5TokenDataIter *fts5AppendTokendataIter(
	@@ -248614,58 +249375,10 @@
248614	assert( pRet==0 \|\| pRet->nIter<=pRet->nIterAlloc );
248615
248616	return pRet;
248617	}
248618
248619	/*
248620	** Delete an Fts5TokenDataIter structure and its contents.
248621	*/
248622	static void fts5TokendataIterDelete(Fts5TokenDataIter *pSet){
248623	if( pSet ){
248624	int ii;
248625	for(ii=0; ii<pSet->nIter; ii++){
248626	fts5MultiIterFree(pSet->apIter[ii]);
248627	}
248628	sqlite3_free(pSet->aPoslistReader);
248629	sqlite3_free(pSet->aMap);
248630	sqlite3_free(pSet);
248631	}
248632	}
248633
248634	/*
248635	** Append a mapping to the token-map belonging to object pT.
248636	*/
248637	static void fts5TokendataIterAppendMap(
248638	Fts5Index *p,
248639	Fts5TokenDataIter *pT,
248640	int iIter,
248641	i64 iRowid,
248642	i64 iPos
248643	){
248644	if( p->rc==SQLITE_OK ){
248645	if( pT->nMap==pT->nMapAlloc ){
248646	int nNew = pT->nMapAlloc ? pT->nMapAlloc*2 : 64;
248647	int nByte = nNew * sizeof(Fts5TokenDataMap);
248648	Fts5TokenDataMap *aNew;
248649
248650	aNew = (Fts5TokenDataMap*)sqlite3_realloc(pT->aMap, nByte);
248651	if( aNew==0 ){
248652	p->rc = SQLITE_NOMEM;
248653	return;
248654	}
248655
248656	pT->aMap = aNew;
248657	pT->nMapAlloc = nNew;
248658	}
248659
248660	pT->aMap[pT->nMap].iRowid = iRowid;
248661	pT->aMap[pT->nMap].iPos = iPos;
248662	pT->aMap[pT->nMap].iIter = iIter;
248663	pT->nMap++;
248664	}
248665	}
248666
248667	/*
248668	** The iterator passed as the only argument must be a tokendata=1 iterator
248669	** (pIter->pTokenDataIter!=0). This function sets the iterator output
248670	** variables (pIter->base.*) according to the contents of the current
248671	** row.
	@@ -248702,11 +249415,11 @@
248702	int eDetail = pIter->pIndex->pConfig->eDetail;
248703	pIter->base.bEof = 0;
248704	pIter->base.iRowid = iRowid;
248705
248706	if( nHit==1 && eDetail==FTS5_DETAIL_FULL ){
248707	fts5TokendataIterAppendMap(pIter->pIndex, pT, iMin, iRowid, -1);
248708	}else
248709	if( nHit>1 && eDetail!=FTS5_DETAIL_NONE ){
248710	int nReader = 0;
248711	int nByte = 0;
248712	i64 iPrev = 0;
	@@ -248955,10 +249668,11 @@
248955
248956	if( p->rc==SQLITE_OK ){
248957	pRet = fts5MultiIterAlloc(p, 0);
248958	}
248959	if( pRet ){

248960	pRet->pTokenDataIter = pSet;
248961	if( pSet ){
248962	fts5IterSetOutputsTokendata(pRet);
248963	}else{
248964	pRet->base.bEof = 1;
	@@ -248969,11 +249683,10 @@
248969
248970	fts5StructureRelease(pStruct);
248971	fts5BufferFree(&bSeek);
248972	return pRet;
248973	}
248974
248975
248976	/*
248977	** Open a new iterator to iterate though all rowid that match the
248978	** specified token or token prefix.
248979	*/
	@@ -248995,10 +249708,15 @@
248995	int iIdx = 0; /* Index to search */
248996	int iPrefixIdx = 0; /* +1 prefix index */
248997	int bTokendata = pConfig->bTokendata;
248998	if( nToken>0 ) memcpy(&buf.p[1], pToken, nToken);
248999





249000	if( flags & (FTS5INDEX_QUERY_NOTOKENDATA\|FTS5INDEX_QUERY_SCAN) ){
249001	bTokendata = 0;
249002	}
249003
249004	/* Figure out which index to search and set iIdx accordingly. If this
	@@ -249025,11 +249743,11 @@
249025	if( nIdxChar==nChar+1 ) iPrefixIdx = iIdx;
249026	}
249027	}
249028
249029	if( bTokendata && iIdx==0 ){
249030	buf.p[0] = '0';
249031	pRet = fts5SetupTokendataIter(p, buf.p, nToken+1, pColset);
249032	}else if( iIdx<=pConfig->nPrefix ){
249033	/* Straight index lookup */
249034	Fts5Structure *pStruct = fts5StructureRead(p);
249035	buf.p[0] = (u8)(FTS5_MAIN_PREFIX + iIdx);
	@@ -249038,11 +249756,11 @@
249038	pColset, buf.p, nToken+1, -1, 0, &pRet
249039	);
249040	fts5StructureRelease(pStruct);
249041	}
249042	}else{
249043	/* Scan multiple terms in the main index */
249044	int bDesc = (flags & FTS5INDEX_QUERY_DESC)!=0;
249045	fts5SetupPrefixIter(p, bDesc, iPrefixIdx, buf.p, nToken+1, pColset,&pRet);
249046	if( pRet==0 ){
249047	assert( p->rc!=SQLITE_OK );
249048	}else{
	@@ -249074,11 +249792,12 @@
249074	** Move to the next matching rowid.
249075	*/
249076	static int sqlite3Fts5IterNext(Fts5IndexIter *pIndexIter){
249077	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249078	assert( pIter->pIndex->rc==SQLITE_OK );
249079	if( pIter->pTokenDataIter ){

249080	fts5TokendataIterNext(pIter, 0, 0);
249081	}else{
249082	fts5MultiIterNext(pIter->pIndex, pIter, 0, 0);
249083	}
249084	return fts5IndexReturn(pIter->pIndex);
	@@ -249111,11 +249830,12 @@
249111	** definition of "at or after" depends on whether this iterator iterates
249112	** in ascending or descending rowid order.
249113	*/
249114	static int sqlite3Fts5IterNextFrom(Fts5IndexIter *pIndexIter, i64 iMatch){
249115	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249116	if( pIter->pTokenDataIter ){

249117	fts5TokendataIterNext(pIter, 1, iMatch);
249118	}else{
249119	fts5MultiIterNextFrom(pIter->pIndex, pIter, iMatch);
249120	}
249121	return fts5IndexReturn(pIter->pIndex);
	@@ -249129,32 +249849,87 @@
249129	const char z = (const char)fts5MultiIterTerm((Fts5Iter*)pIndexIter, &n);
249130	assert_nc( z \|\| n<=1 );
249131	*pn = n-1;
249132	return (z ? &z[1] : 0);
249133	}











































249134
249135	/*
249136	** This is used by xInstToken() to access the token at offset iOff, column
249137	** iCol of row iRowid. The token is returned via output variables *ppOut
249138	** and *pnOut. The iterator passed as the first argument must be a tokendata=1
249139	** iterator (pIter->pTokenDataIter!=0).


249140	*/
249141	static int sqlite3Fts5IterToken(
249142	Fts5IndexIter *pIndexIter,

249143	i64 iRowid,
249144	int iCol,
249145	int iOff,
249146	const char *ppOut, int pnOut
249147	){
249148	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249149	Fts5TokenDataIter *pT = pIter->pTokenDataIter;
249150	Fts5TokenDataMap *aMap = pT->aMap;
249151	i64 iPos = (((i64)iCol)<<32) + iOff;
249152
249153	int i1 = 0;
249154	int i2 = pT->nMap;
249155	int iTest = 0;










249156
249157	while( i2>i1 ){
249158	iTest = (i1 + i2) / 2;
249159
249160	if( aMap[iTest].iRowid<iRowid ){
	@@ -249174,13 +249949,19 @@
249174	}
249175	}
249176	}
249177
249178	if( i2>i1 ){
249179	Fts5Iter *pMap = pT->apIter[aMap[iTest].iIter];
249180	ppOut = (const char)pMap->aSeg[0].term.p+1;
249181	*pnOut = pMap->aSeg[0].term.n-1;






249182	}
249183
249184	return SQLITE_OK;
249185	}
249186
	@@ -249188,11 +249969,13 @@
249188	** Clear any existing entries from the token-map associated with the
249189	** iterator passed as the only argument.
249190	*/
249191	static void sqlite3Fts5IndexIterClearTokendata(Fts5IndexIter *pIndexIter){
249192	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249193	if( pIter && pIter->pTokenDataIter ){


249194	pIter->pTokenDataIter->nMap = 0;
249195	}
249196	}
249197
249198	/*
	@@ -249208,21 +249991,33 @@
249208	i64 iRowid, int iCol, int iOff
249209	){
249210	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249211	Fts5TokenDataIter *pT = pIter->pTokenDataIter;
249212	Fts5Index *p = pIter->pIndex;
249213	int ii;
249214
249215	assert( p->pConfig->eDetail!=FTS5_DETAIL_FULL );
249216	assert( pIter->pTokenDataIter );
249217
249218	for(ii=0; ii<pT->nIter; ii++){
249219	Fts5Buffer *pTerm = &pT->apIter[ii]->aSeg[0].term;
249220	if( nToken==pTerm->n-1 && memcmp(pToken, pTerm->p+1, nToken)==0 ) break;
249221	}
249222	if( ii<pT->nIter ){
249223	fts5TokendataIterAppendMap(p, pT, ii, iRowid, (((i64)iCol)<<32) + iOff);












249224	}
249225	return fts5IndexReturn(p);
249226	}
249227
249228	/*
	@@ -250771,11 +251566,11 @@
250771	return rc;
250772	}
250773
250774	/*
250775	** We must have a single struct=? constraint that will be passed through
250776	** into the xFilter method. If there is no valid stmt=? constraint,
250777	** then return an SQLITE_CONSTRAINT error.
250778	*/
250779	static int fts5structBestIndexMethod(
250780	sqlite3_vtab *tab,
250781	sqlite3_index_info *pIdxInfo
	@@ -251113,12 +251908,22 @@
251113	i64 iNextId; /* Used to allocate unique cursor ids */
251114	Fts5Auxiliary pAux; / First in list of all aux. functions */
251115	Fts5TokenizerModule pTok; / First in list of all tokenizer modules */
251116	Fts5TokenizerModule pDfltTok; / Default tokenizer module */
251117	Fts5Cursor pCsr; / First in list of all open cursors */

251118	};
251119









251120	/*
251121	** Each auxiliary function registered with the FTS5 module is represented
251122	** by an object of the following type. All such objects are stored as part
251123	** of the Fts5Global.pAux list.
251124	*/
	@@ -251277,16 +252082,10 @@
251277	#define FTS5CSR_REQUIRE_POSLIST 0x40
251278
251279	#define BitFlagAllTest(x,y) (((x) & (y))==(y))
251280	#define BitFlagTest(x,y) (((x) & (y))!=0)
251281
251282	/*
251283	** The subtype value and header bytes used by fts5_locale().
251284	*/
251285	#define FTS5_LOCALE_SUBTYPE ((unsigned int)'L')
251286	#define FTS5_LOCALE_HEADER "\x00\xE0\xB2\xEB"
251287
251288
251289	/*
251290	** Macros to Set(), Clear() and Test() cursor flags.
251291	*/
251292	#define CsrFlagSet(pCsr, flag) ((pCsr)->csrflags \|= (flag))
	@@ -251359,14 +252158,20 @@
251359	#else
251360	# define fts5CheckTransactionState(x,y,z)
251361	#endif
251362
251363	/*
251364	** Return true if pTab is a contentless table.


251365	*/
251366	static int fts5IsContentless(Fts5FullTable *pTab){
251367	return pTab->p.pConfig->eContent==FTS5_CONTENT_NONE;




251368	}
251369
251370	/*
251371	** Delete a virtual table handle allocated by fts5InitVtab().
251372	*/
	@@ -251653,10 +252458,11 @@
251653	){
251654	/* A MATCH operator or equivalent */
251655	if( p->usable==0 \|\| iCol<0 ){
251656	/* As there exists an unusable MATCH constraint this is an
251657	** unusable plan. Return SQLITE_CONSTRAINT. */

251658	return SQLITE_CONSTRAINT;
251659	}else{
251660	if( iCol==nCol+1 ){
251661	if( bSeenRank ) continue;
251662	idxStr[iIdxStr++] = 'r';
	@@ -252286,11 +253092,11 @@
252286	** Arrange for subsequent calls to sqlite3Fts5Tokenize() to use the locale
252287	** specified by pLocale/nLocale. The buffer indicated by pLocale must remain
252288	** valid until after the final call to sqlite3Fts5Tokenize() that will use
252289	** the locale.
252290	*/
252291	static void fts5SetLocale(
252292	Fts5Config *pConfig,
252293	const char *zLocale,
252294	int nLocale
252295	){
252296	Fts5TokenizerConfig *pT = &pConfig->t;
	@@ -252297,141 +253103,88 @@
252297	pT->pLocale = zLocale;
252298	pT->nLocale = nLocale;
252299	}
252300
252301	/*
252302	** Clear any locale configured by an earlier call to fts5SetLocale() or
252303	** sqlite3Fts5ExtractText().
252304	*/
252305	static void sqlite3Fts5ClearLocale(Fts5Config *pConfig){
252306	fts5SetLocale(pConfig, 0, 0);
252307	}
252308
252309	/*
252310	** This function is used to extract utf-8 text from an sqlite3_value. This
252311	** is usually done in order to tokenize it. For example, when:
252312	**
252313	** * a value is written to an fts5 table,
252314	** * a value is deleted from an FTS5_CONTENT_NORMAL table,
252315	** * a value containing a query expression is passed to xFilter()
252316	**
252317	** and so on.
252318	**
252319	** This function handles 2 cases:
252320	**
252321	** 1) Ordinary values. The text can be extracted from these using
252322	** sqlite3_value_text().
252323	**
252324	** 2) Combination text/locale blobs created by fts5_locale(). There
252325	** are several cases for these:
252326	**
252327	** * Blobs tagged with FTS5_LOCALE_SUBTYPE.
252328	** * Blobs read from the content table of a locale=1 external-content
252329	** table, and
252330	** * Blobs read from the content table of a locale=1 regular
252331	** content table.
252332	**
252333	** The first two cases above should have the 4 byte FTS5_LOCALE_HEADER
252334	** header. It is an error if a blob with the subtype or a blob read
252335	** from the content table of an external content table does not have
252336	** the required header. A blob read from the content table of a regular
252337	** locale=1 table does not have the header. This is to save space.
252338	**
252339	** If successful, SQLITE_OK is returned and output parameters (*ppText)
252340	** and (*pnText) are set to point to a buffer containing the extracted utf-8
252341	** text and its length in bytes, respectively. The buffer is not
252342	** nul-terminated. It has the same lifetime as the sqlite3_value object
252343	** from which it is extracted.
252344	**
252345	** Parameter bContent must be true if the value was read from an indexed
252346	** column (i.e. not UNINDEXED) of the on disk content.
252347	**
252348	** If pbResetTokenizer is not NULL and if case (2) is used, then
252349	** fts5SetLocale() is called to ensure subsequent sqlite3Fts5Tokenize() calls
252350	** use the locale. In this case (*pbResetTokenizer) is set to true before
252351	** returning, to indicate that the caller must call sqlite3Fts5ClearLocale()
252352	** to clear the locale after tokenizing the text.
252353	*/
252354	static int sqlite3Fts5ExtractText(
252355	Fts5Config *pConfig,
252356	sqlite3_value pVal, / Value to extract text from */
252357	int bContent, /* True if indexed table content */
252358	int pbResetTokenizer, / OUT: True if xSetLocale(NULL) required */
252359	const char *ppText, / OUT: Pointer to text buffer */
252360	int pnText / OUT: Size of (ppText) in bytes /
252361	){
252362	const char *pText = 0;
252363	int nText = 0;
252364	int rc = SQLITE_OK;
252365	int bDecodeBlob = 0;
252366
252367	assert( pbResetTokenizer==0 \|\| *pbResetTokenizer==0 );
252368	assert( bContent==0 \|\| pConfig->eContent!=FTS5_CONTENT_NONE );
252369	assert( bContent==0 \|\| sqlite3_value_subtype(pVal)==0 );
252370
252371	if( sqlite3_value_type(pVal)==SQLITE_BLOB ){
252372	if( sqlite3_value_subtype(pVal)==FTS5_LOCALE_SUBTYPE
252373	\|\| (bContent && pConfig->bLocale)
252374	){
252375	bDecodeBlob = 1;
252376	}
252377	}
252378
252379	if( bDecodeBlob ){
252380	const int SZHDR = sizeof(FTS5_LOCALE_HEADER)-1;
252381	const u8 *pBlob = sqlite3_value_blob(pVal);
252382	int nBlob = sqlite3_value_bytes(pVal);
252383
252384	/* Unless this blob was read from the %_content table of an
252385	** FTS5_CONTENT_NORMAL table, it should have the 4 byte fts5_locale()
252386	** header. Check for this. If it is not found, return an error. */
252387	if( (!bContent \|\| pConfig->eContent!=FTS5_CONTENT_NORMAL) ){
252388	if( nBlob<SZHDR \|\| memcmp(FTS5_LOCALE_HEADER, pBlob, SZHDR) ){
252389	rc = SQLITE_ERROR;
252390	}else{
252391	pBlob += 4;
252392	nBlob -= 4;
252393	}
252394	}
252395
252396	if( rc==SQLITE_OK ){
252397	int nLocale = 0;
252398
252399	for(nLocale=0; nLocale<nBlob; nLocale++){
252400	if( pBlob[nLocale]==0x00 ) break;
252401	}
252402	if( nLocale==nBlob \|\| nLocale==0 ){
252403	rc = SQLITE_ERROR;
252404	}else{
252405	pText = (const char*)&pBlob[nLocale+1];
252406	nText = nBlob-nLocale-1;
252407
252408	if( pbResetTokenizer ){
252409	fts5SetLocale(pConfig, (const char*)pBlob, nLocale);
252410	*pbResetTokenizer = 1;
252411	}
252412	}
252413	}
252414
252415	}else{
252416	pText = (const char*)sqlite3_value_text(pVal);
252417	nText = sqlite3_value_bytes(pVal);
252418	}
252419
252420	*ppText = pText;
252421	*pnText = nText;
252422	return rc;







252423	}
252424
252425	/*
252426	** Argument pVal is the text of a full-text search expression. It may or
252427	** may not have been wrapped by fts5_locale(). This function extracts
252428	** the text of the expression, and sets output variable (*pzText) to
252429	** point to a nul-terminated buffer containing the expression.
252430	**
252431	** If pVal was an fts5_locale() value, then fts5SetLocale() is called to
252432	** set the tokenizer to use the specified locale.
252433	**
252434	** If output variable (*pbFreeAndReset) is set to true, then the caller
252435	** is required to (a) call sqlite3Fts5ClearLocale() to reset the tokenizer
252436	** locale, and (b) call sqlite3_free() to free (*pzText).
252437	*/
	@@ -252439,28 +253192,26 @@
252439	Fts5Config pConfig, / Fts5 configuration */
252440	sqlite3_value pVal, / Value to extract expression text from */
252441	char *pzText, / OUT: nul-terminated buffer of text */
252442	int pbFreeAndReset / OUT: Free (pzText) and clear locale /
252443	){
252444	const char *zText = 0;
252445	int nText = 0;
252446	int rc = SQLITE_OK;
252447	int bReset = 0;
252448
252449	*pbFreeAndReset = 0;
252450	rc = sqlite3Fts5ExtractText(pConfig, pVal, 0, &bReset, &zText, &nText);
252451	if( rc==SQLITE_OK ){
252452	if( bReset ){
252453	pzText = sqlite3Fts5Mprintf(&rc, "%.s", nText, zText);
252454	if( rc!=SQLITE_OK ){
252455	sqlite3Fts5ClearLocale(pConfig);
252456	}else{
252457	*pbFreeAndReset = 1;
252458	}
252459	}else{
252460	pzText = (char)zText;
252461	}
252462	}
252463
252464	return rc;
252465	}
252466
	@@ -252493,10 +253244,11 @@
252493	sqlite3_value pRowidEq = 0; / rowid = ? expression (or NULL) */
252494	sqlite3_value pRowidLe = 0; / rowid <= ? expression (or NULL) */
252495	sqlite3_value pRowidGe = 0; / rowid >= ? expression (or NULL) */
252496	int iCol; /* Column on LHS of MATCH operator */
252497	char **pzErrmsg = pConfig->pzErrmsg;

252498	int i;
252499	int iIdxStr = 0;
252500	Fts5Expr *pExpr = 0;
252501
252502	assert( pConfig->bLock==0 );
	@@ -252528,10 +253280,13 @@
252528	int bInternal = 0;
252529
252530	rc = fts5ExtractExprText(pConfig, apVal[i], &zText, &bFreeAndReset);
252531	if( rc!=SQLITE_OK ) goto filter_out;
252532	if( zText==0 ) zText = "";



252533
252534	iCol = 0;
252535	do{
252536	iCol = iCol*10 + (idxStr[iIdxStr]-'0');
252537	iIdxStr++;
	@@ -252668,10 +253423,11 @@
252668	}
252669
252670	filter_out:
252671	sqlite3Fts5ExprFree(pExpr);
252672	pConfig->pzErrmsg = pzErrmsg;

252673	return rc;
252674	}
252675
252676	/*
252677	** This is the xEof method of the virtual table. SQLite calls this
	@@ -252808,11 +253564,11 @@
252808	}else{
252809	rc = sqlite3Fts5StorageDeleteAll(pTab->pStorage);
252810	}
252811	bLoadConfig = 1;
252812	}else if( 0==sqlite3_stricmp("rebuild", zCmd) ){
252813	if( pConfig->eContent==FTS5_CONTENT_NONE ){
252814	fts5SetVtabError(pTab,
252815	"'rebuild' may not be used with a contentless fts5 table"
252816	);
252817	rc = SQLITE_ERROR;
252818	}else{
	@@ -252877,17 +253633,78 @@
252877	sqlite3_value **apVal,
252878	i64 *piRowid
252879	){
252880	int rc = *pRc;
252881	if( rc==SQLITE_OK ){
252882	rc = sqlite3Fts5StorageContentInsert(pTab->pStorage, apVal, piRowid);
252883	}
252884	if( rc==SQLITE_OK ){
252885	rc = sqlite3Fts5StorageIndexInsert(pTab->pStorage, apVal, *piRowid);
252886	}
252887	*pRc = rc;
252888	}





























































252889
252890	/*
252891	** This function is the implementation of the xUpdate callback used by
252892	** FTS3 virtual tables. It is invoked by SQLite each time a row is to be
252893	** inserted, updated or deleted.
	@@ -252971,48 +253788,38 @@
252971	}
252972
252973	assert( eType0==SQLITE_INTEGER \|\| eType0==SQLITE_NULL );
252974	assert( nArg!=1 \|\| eType0==SQLITE_INTEGER );
252975
252976	/* Filter out attempts to run UPDATE or DELETE on contentless tables.
252977	** This is not suported. Except - they are both supported if the CREATE
252978	** VIRTUAL TABLE statement contained "contentless_delete=1". */
252979	if( eType0==SQLITE_INTEGER
252980	&& pConfig->eContent==FTS5_CONTENT_NONE
252981	&& pConfig->bContentlessDelete==0
252982	){
252983	pTab->p.base.zErrMsg = sqlite3_mprintf(
252984	"cannot %s contentless fts5 table: %s",
252985	(nArg>1 ? "UPDATE" : "DELETE from"), pConfig->zName
252986	);
252987	rc = SQLITE_ERROR;
252988	}
252989
252990	/* DELETE */
252991	else if( nArg==1 ){
252992	i64 iDel = sqlite3_value_int64(apVal[0]); /* Rowid to delete */
252993	rc = sqlite3Fts5StorageDelete(pTab->pStorage, iDel, 0, 0);
252994	bUpdateOrDelete = 1;









252995	}
252996
252997	/* INSERT or UPDATE */
252998	else{
252999	int eType1 = sqlite3_value_numeric_type(apVal[1]);
253000
253001	/* Ensure that no fts5_locale() values are written to locale=0 tables.
253002	** And that no blobs except fts5_locale() blobs are written to indexed
253003	** (i.e. not UNINDEXED) columns of locale=1 tables. */
253004	int ii;
253005	for(ii=0; ii<pConfig->nCol; ii++){
253006	if( sqlite3_value_type(apVal[ii+2])==SQLITE_BLOB ){
253007	int bSub = (sqlite3_value_subtype(apVal[ii+2])==FTS5_LOCALE_SUBTYPE);
253008	if( (pConfig->bLocale && !bSub && pConfig->abUnindexed[ii]==0)
253009	\|\| (pConfig->bLocale==0 && bSub)
253010	){
253011	if( pConfig->bLocale==0 ){
253012	fts5SetVtabError(pTab, "fts5_locale() requires locale=1");
253013	}
253014	rc = SQLITE_MISMATCH;
253015	goto update_out;
253016	}
253017	}
253018	}
	@@ -253028,39 +253835,59 @@
253028	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253029	}
253030
253031	/* UPDATE */
253032	else{
253033	i64 iOld = sqlite3_value_int64(apVal[0]); /* Old rowid */
253034	i64 iNew = sqlite3_value_int64(apVal[1]); /* New rowid */
253035	if( eType1!=SQLITE_INTEGER ){
253036	rc = SQLITE_MISMATCH;
253037	}else if( iOld!=iNew ){
253038	if( eConflict==SQLITE_REPLACE ){
253039	rc = sqlite3Fts5StorageDelete(pTab->pStorage, iOld, 0, 1);
253040	if( rc==SQLITE_OK ){
253041	rc = sqlite3Fts5StorageDelete(pTab->pStorage, iNew, 0, 0);
253042	}
253043	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253044	}else{
253045	rc = sqlite3Fts5StorageFindDeleteRow(pTab->pStorage, iOld);
253046	if( rc==SQLITE_OK ){
253047	rc = sqlite3Fts5StorageContentInsert(pTab->pStorage,apVal,pRowid);
253048	}
253049	if( rc==SQLITE_OK ){
253050	rc = sqlite3Fts5StorageDelete(pTab->pStorage, iOld, 0, 1);
253051	}
253052	if( rc==SQLITE_OK ){
253053	rc = sqlite3Fts5StorageIndexInsert(pTab->pStorage, apVal,*pRowid);
253054	}
253055	}
253056	}else{
253057	rc = sqlite3Fts5StorageDelete(pTab->pStorage, iOld, 0, 1);
253058	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253059	}
253060	bUpdateOrDelete = 1;
253061	sqlite3Fts5StorageReleaseDeleteRow(pTab->pStorage);




















253062	}
253063
253064	}
253065	}
253066
	@@ -253169,15 +253996,15 @@
253169	){
253170	Fts5Cursor pCsr = (Fts5Cursor)pCtx;
253171	Fts5Table pTab = (Fts5Table)(pCsr->base.pVtab);
253172	int rc = SQLITE_OK;
253173
253174	fts5SetLocale(pTab->pConfig, pLoc, nLoc);
253175	rc = sqlite3Fts5Tokenize(pTab->pConfig,
253176	FTS5_TOKENIZE_AUX, pText, nText, pUserData, xToken
253177	);
253178	fts5SetLocale(pTab->pConfig, 0, 0);
253179
253180	return rc;
253181	}
253182
253183	/*
	@@ -253200,10 +254027,53 @@
253200
253201	static int fts5ApiPhraseSize(Fts5Context *pCtx, int iPhrase){
253202	Fts5Cursor pCsr = (Fts5Cursor)pCtx;
253203	return sqlite3Fts5ExprPhraseSize(pCsr->pExpr, iPhrase);
253204	}











































253205
253206	static int fts5ApiColumnText(
253207	Fts5Context *pCtx,
253208	int iCol,
253209	const char **pz,
	@@ -253214,20 +254084,18 @@
253214	Fts5Table pTab = (Fts5Table)(pCsr->base.pVtab);
253215
253216	assert( pCsr->ePlan!=FTS5_PLAN_SPECIAL );
253217	if( iCol<0 \|\| iCol>=pTab->pConfig->nCol ){
253218	rc = SQLITE_RANGE;
253219	}else if( fts5IsContentless((Fts5FullTable*)(pCsr->base.pVtab)) ){
253220	*pz = 0;
253221	*pn = 0;
253222	}else{
253223	rc = fts5SeekCursor(pCsr, 0);
253224	if( rc==SQLITE_OK ){
253225	Fts5Config *pConfig = pTab->pConfig;
253226	int bContent = (pConfig->abUnindexed[iCol]==0);
253227	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
253228	sqlite3Fts5ExtractText(pConfig, pVal, bContent, 0, pz, pn);
253229	}
253230	}
253231	return rc;
253232	}
253233
	@@ -253249,11 +254117,11 @@
253249	int bLive = (pCsr->pSorter==0);
253250
253251	if( iPhrase<0 \|\| iPhrase>=sqlite3Fts5ExprPhraseCount(pCsr->pExpr) ){
253252	rc = SQLITE_RANGE;
253253	}else if( pConfig->eDetail!=FTS5_DETAIL_FULL
253254	&& pConfig->eContent==FTS5_CONTENT_NONE
253255	){
253256	*pa = 0;
253257	*pn = 0;
253258	return SQLITE_OK;
253259	}else if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_POSLIST) ){
	@@ -253265,21 +254133,19 @@
253265	if( aPopulator==0 ) rc = SQLITE_NOMEM;
253266	if( rc==SQLITE_OK ){
253267	rc = fts5SeekCursor(pCsr, 0);
253268	}
253269	for(i=0; i<pConfig->nCol && rc==SQLITE_OK; i++){
253270	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, i+1);
253271	const char *z = 0;
253272	int n = 0;
253273	int bReset = 0;
253274	rc = sqlite3Fts5ExtractText(pConfig, pVal, 1, &bReset, &z, &n);
253275	if( rc==SQLITE_OK ){
253276	rc = sqlite3Fts5ExprPopulatePoslists(
253277	pConfig, pCsr->pExpr, aPopulator, i, z, n
253278	);
253279	}
253280	if( bReset ) sqlite3Fts5ClearLocale(pConfig);
253281	}
253282	sqlite3_free(aPopulator);
253283
253284	if( pCsr->pSorter ){
253285	sqlite3Fts5ExprCheckPoslists(pCsr->pExpr, pCsr->pSorter->iRowid);
	@@ -253447,11 +254313,11 @@
253447
253448	if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_DOCSIZE) ){
253449	if( pConfig->bColumnsize ){
253450	i64 iRowid = fts5CursorRowid(pCsr);
253451	rc = sqlite3Fts5StorageDocsize(pTab->pStorage, iRowid, pCsr->aColumnSize);
253452	}else if( pConfig->zContent==0 ){
253453	int i;
253454	for(i=0; i<pConfig->nCol; i++){
253455	if( pConfig->abUnindexed[i]==0 ){
253456	pCsr->aColumnSize[i] = -1;
253457	}
	@@ -253461,21 +254327,18 @@
253461	rc = fts5SeekCursor(pCsr, 0);
253462	for(i=0; rc==SQLITE_OK && i<pConfig->nCol; i++){
253463	if( pConfig->abUnindexed[i]==0 ){
253464	const char *z = 0;
253465	int n = 0;
253466	int bReset = 0;
253467	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, i+1);
253468
253469	pCsr->aColumnSize[i] = 0;
253470	rc = sqlite3Fts5ExtractText(pConfig, pVal, 1, &bReset, &z, &n);
253471	if( rc==SQLITE_OK ){
253472	rc = sqlite3Fts5Tokenize(pConfig, FTS5_TOKENIZE_AUX,
253473	z, n, (void*)&pCsr->aColumnSize[i], fts5ColumnSizeCb
253474	);
253475	if( bReset ) sqlite3Fts5ClearLocale(pConfig);
253476	}

253477	}
253478	}
253479	}
253480	CsrFlagClear(pCsr, FTS5CSR_REQUIRE_DOCSIZE);
253481	}
	@@ -253738,46 +254601,23 @@
253738	assert( pCsr->ePlan!=FTS5_PLAN_SPECIAL );
253739	if( iCol<0 \|\| iCol>=pConfig->nCol ){
253740	rc = SQLITE_RANGE;
253741	}else if(
253742	pConfig->abUnindexed[iCol]==0
253743	&& pConfig->eContent!=FTS5_CONTENT_NONE
253744	&& pConfig->bLocale
253745	){
253746	rc = fts5SeekCursor(pCsr, 0);
253747	if( rc==SQLITE_OK ){
253748	/* Load the value into pVal. pVal is a locale/text pair iff:
253749	**
253750	** 1) It is an SQLITE_BLOB, and
253751	** 2) Either the subtype is FTS5_LOCALE_SUBTYPE, or else the
253752	** value was loaded from an FTS5_CONTENT_NORMAL table, and
253753	** 3) It does not begin with an 0x00 byte.
253754	*/
253755	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
253756	if( sqlite3_value_type(pVal)==SQLITE_BLOB ){
253757	const u8 pBlob = (const u8)sqlite3_value_blob(pVal);
253758	int nBlob = sqlite3_value_bytes(pVal);
253759	if( pConfig->eContent==FTS5_CONTENT_EXTERNAL ){
253760	const int SZHDR = sizeof(FTS5_LOCALE_HEADER)-1;
253761	if( nBlob<SZHDR \|\| memcmp(FTS5_LOCALE_HEADER, pBlob, SZHDR) ){
253762	rc = SQLITE_ERROR;
253763	}
253764	pBlob += 4;
253765	nBlob -= 4;
253766	}
253767	if( rc==SQLITE_OK ){
253768	int nLocale = 0;
253769	for(nLocale=0; nLocale<nBlob && pBlob[nLocale]!=0x00; nLocale++);
253770	if( nLocale==nBlob \|\| nLocale==0 ){
253771	rc = SQLITE_ERROR;
253772	}else{
253773	/* A locale/text pair */
253774	pzLocale = (const char)pBlob;
253775	*pnLocale = nLocale;
253776	}
253777	}
253778	}
253779	}
253780	}
253781
253782	return rc;
253783	}
	@@ -253994,61 +254834,10 @@
253994
253995	sqlite3_result_blob(pCtx, val.p, val.n, sqlite3_free);
253996	return rc;
253997	}
253998
253999	/*
254000	** Value pVal was read from column iCol of the FTS5 table. This function
254001	** returns it to the owner of pCtx via a call to an sqlite3_result_xxx()
254002	** function. This function deals with the same cases as
254003	** sqlite3Fts5ExtractText():
254004	**
254005	** 1) Ordinary values. These can be returned using sqlite3_result_value().
254006	**
254007	** 2) Blobs from fts5_locale(). The text is extracted from these and
254008	** returned via sqlite3_result_text(). The locale is discarded.
254009	*/
254010	static void fts5ExtractValueFromColumn(
254011	sqlite3_context *pCtx,
254012	Fts5Config *pConfig,
254013	int iCol,
254014	sqlite3_value *pVal
254015	){
254016	assert( pConfig->eContent!=FTS5_CONTENT_NONE );
254017
254018	if( pConfig->bLocale
254019	&& sqlite3_value_type(pVal)==SQLITE_BLOB
254020	&& pConfig->abUnindexed[iCol]==0
254021	){
254022	const int SZHDR = sizeof(FTS5_LOCALE_HEADER)-1;
254023	const u8 *pBlob = sqlite3_value_blob(pVal);
254024	int nBlob = sqlite3_value_bytes(pVal);
254025	int ii;
254026
254027	if( pConfig->eContent==FTS5_CONTENT_EXTERNAL ){
254028	if( nBlob<SZHDR \|\| memcmp(pBlob, FTS5_LOCALE_HEADER, SZHDR) ){
254029	sqlite3_result_error_code(pCtx, SQLITE_ERROR);
254030	return;
254031	}else{
254032	pBlob += 4;
254033	nBlob -= 4;
254034	}
254035	}
254036
254037	for(ii=0; ii<nBlob && pBlob[ii]; ii++);
254038	if( ii==0 \|\| ii==nBlob ){
254039	sqlite3_result_error_code(pCtx, SQLITE_ERROR);
254040	}else{
254041	const char pText = (const char)&pBlob[ii+1];
254042	sqlite3_result_text(pCtx, pText, nBlob-ii-1, SQLITE_TRANSIENT);
254043	}
254044	return;
254045	}
254046
254047	sqlite3_result_value(pCtx, pVal);
254048	}
254049
254050	/*
254051	** This is the xColumn method, called by SQLite to request a value from
254052	** the row that the supplied cursor currently points to.
254053	*/
254054	static int fts5ColumnMethod(
	@@ -254087,26 +254876,31 @@
254087	if( pCsr->pRank \|\| SQLITE_OK==(rc = fts5FindRankFunction(pCsr)) ){
254088	fts5ApiInvoke(pCsr->pRank, pCsr, pCtx, pCsr->nRankArg, pCsr->apRankArg);
254089	}
254090	}
254091	}else{
254092	/* A column created by the user containing values. */
254093	int bNochange = sqlite3_vtab_nochange(pCtx);
254094
254095	if( fts5IsContentless(pTab) ){
254096	if( bNochange && pConfig->bContentlessDelete ){
254097	fts5ResultError(pCtx, "cannot UPDATE a subset of "
254098	"columns on fts5 contentless-delete table: %s", pConfig->zName
254099	);
254100	}
254101	}else if( bNochange==0 \|\| pConfig->eContent!=FTS5_CONTENT_NORMAL ){
254102	pConfig->pzErrmsg = &pTab->p.base.zErrMsg;
254103	rc = fts5SeekCursor(pCsr, 1);
254104	if( rc==SQLITE_OK ){
254105	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
254106	fts5ExtractValueFromColumn(pCtx, pConfig, iCol, pVal);













254107	}

254108	pConfig->pzErrmsg = 0;
254109	}
254110	}
254111
254112	return rc;
	@@ -254625,11 +255419,11 @@
254625	int nArg, /* Number of args */
254626	sqlite3_value *apUnused / Function arguments */
254627	){
254628	assert( nArg==0 );
254629	UNUSED_PARAM2(nArg, apUnused);
254630	sqlite3_result_text(pCtx, "fts5: 2024-09-02 18:41:59 e6bec37ea1ca51e1d048941ce4c5211d8fc5c5e3556a1441f9c79b036843f9e3", -1, SQLITE_TRANSIENT);
254631	}
254632
254633	/*
254634	** Implementation of fts5_locale(LOCALE, TEXT) function.
254635	**
	@@ -254664,36 +255458,48 @@
254664	nText = sqlite3_value_bytes(apArg[1]);
254665
254666	if( zLocale==0 \|\| zLocale[0]=='\0' ){
254667	sqlite3_result_text(pCtx, zText, nText, SQLITE_TRANSIENT);
254668	}else{

254669	u8 *pBlob = 0;
254670	u8 *pCsr = 0;
254671	int nBlob = 0;
254672	const int nHdr = 4;
254673	assert( sizeof(FTS5_LOCALE_HEADER)==nHdr+1 );
254674
254675	nBlob = nHdr + nLocale + 1 + nText;
254676	pBlob = (u8*)sqlite3_malloc(nBlob);
254677	if( pBlob==0 ){
254678	sqlite3_result_error_nomem(pCtx);
254679	return;
254680	}
254681
254682	pCsr = pBlob;
254683	memcpy(pCsr, FTS5_LOCALE_HEADER, nHdr);
254684	pCsr += nHdr;
254685	memcpy(pCsr, zLocale, nLocale);
254686	pCsr += nLocale;
254687	(*pCsr++) = 0x00;
254688	if( zText ) memcpy(pCsr, zText, nText);
254689	assert( &pCsr[nText]==&pBlob[nBlob] );
254690
254691	sqlite3_result_blob(pCtx, pBlob, nBlob, sqlite3_free);
254692	sqlite3_result_subtype(pCtx, FTS5_LOCALE_SUBTYPE);
254693	}
254694	}














254695
254696	/*
254697	** Return true if zName is the extension on one of the shadow tables used
254698	** by this module.
254699	*/
	@@ -254789,10 +255595,20 @@
254789	pGlobal->api.xCreateFunction = fts5CreateAux;
254790	pGlobal->api.xCreateTokenizer = fts5CreateTokenizer;
254791	pGlobal->api.xFindTokenizer = fts5FindTokenizer;
254792	pGlobal->api.xCreateTokenizer_v2 = fts5CreateTokenizer_v2;
254793	pGlobal->api.xFindTokenizer_v2 = fts5FindTokenizer_v2;










254794	rc = sqlite3_create_module_v2(db, "fts5", &fts5Mod, p, fts5ModuleDestroy);
254795	if( rc==SQLITE_OK ) rc = sqlite3Fts5IndexInit(db);
254796	if( rc==SQLITE_OK ) rc = sqlite3Fts5ExprInit(pGlobal, db);
254797	if( rc==SQLITE_OK ) rc = sqlite3Fts5AuxInit(&pGlobal->api);
254798	if( rc==SQLITE_OK ) rc = sqlite3Fts5TokenizerInit(&pGlobal->api);
	@@ -254810,13 +255626,20 @@
254810	);
254811	}
254812	if( rc==SQLITE_OK ){
254813	rc = sqlite3_create_function(
254814	db, "fts5_locale", 2,
254815	SQLITE_UTF8\|SQLITE_INNOCUOUS\|SQLITE_RESULT_SUBTYPE,
254816	p, fts5LocaleFunc, 0, 0
254817	);







254818	}
254819	}
254820
254821	/* If SQLITE_FTS5_ENABLE_TEST_MI is defined, assume that the file
254822	** fts5_test_mi.c is compiled and linked into the executable. And call
	@@ -255015,24 +255838,39 @@
255015	);
255016	break;
255017
255018	case FTS5_STMT_INSERT_CONTENT:
255019	case FTS5_STMT_REPLACE_CONTENT: {
255020	int nCol = pC->nCol + 1;
255021	char *zBind;
255022	int i;
255023
255024	zBind = sqlite3_malloc64(1 + nCol*2);
255025	if( zBind ){
255026	for(i=0; i<nCol; i++){
255027	zBind[i*2] = '?';
255028	zBind[i*2 + 1] = ',';
255029	}
255030	zBind[i*2-1] = '\0';
255031	zSql = sqlite3_mprintf(azStmt[eStmt], pC->zDb, pC->zName, zBind);
255032	sqlite3_free(zBind);
255033	}
















255034	break;
255035	}
255036
255037	case FTS5_STMT_REPLACE_DOCSIZE:
255038	zSql = sqlite3_mprintf(azStmt[eStmt], pC->zDb, pC->zName,
	@@ -255214,23 +256052,37 @@
255214	p->aTotalSize = (i64*)&p[1];
255215	p->pConfig = pConfig;
255216	p->pIndex = pIndex;
255217
255218	if( bCreate ){
255219	if( pConfig->eContent==FTS5_CONTENT_NORMAL ){


255220	int nDefn = 32 + pConfig->nCol*10;
255221	char zDefn = sqlite3_malloc64(32 + (sqlite3_int64)pConfig->nCol 10);
255222	if( zDefn==0 ){
255223	rc = SQLITE_NOMEM;
255224	}else{
255225	int i;
255226	int iOff;
255227	sqlite3_snprintf(nDefn, zDefn, "id INTEGER PRIMARY KEY");
255228	iOff = (int)strlen(zDefn);
255229	for(i=0; i<pConfig->nCol; i++){
255230	sqlite3_snprintf(nDefn-iOff, &zDefn[iOff], ", c%d", i);
255231	iOff += (int)strlen(&zDefn[iOff]);












255232	}
255233	rc = sqlite3Fts5CreateTable(pConfig, "content", zDefn, 0, pzErr);
255234	}
255235	sqlite3_free(zDefn);
255236	}
	@@ -255379,33 +256231,43 @@
255379	for(iCol=1; rc==SQLITE_OK && iCol<=pConfig->nCol; iCol++){
255380	if( pConfig->abUnindexed[iCol-1]==0 ){
255381	sqlite3_value *pVal = 0;
255382	const char *pText = 0;
255383	int nText = 0;
255384	int bReset = 0;

255385
255386	assert( pSeek==0 \|\| apVal==0 );
255387	assert( pSeek!=0 \|\| apVal!=0 );
255388	if( pSeek ){
255389	pVal = sqlite3_column_value(pSeek, iCol);
255390	}else{
255391	pVal = apVal[iCol-1];
255392	}
255393
255394	rc = sqlite3Fts5ExtractText(
255395	pConfig, pVal, pSeek!=0, &bReset, &pText, &nText
255396	);








255397	if( rc==SQLITE_OK ){

255398	ctx.szCol = 0;
255399	rc = sqlite3Fts5Tokenize(pConfig, FTS5_TOKENIZE_DOCUMENT,
255400	pText, nText, (void*)&ctx, fts5StorageInsertCallback
255401	);
255402	p->aTotalSize[iCol-1] -= (i64)ctx.szCol;
255403	if( rc==SQLITE_OK && p->aTotalSize[iCol-1]<0 ){
255404	rc = FTS5_CORRUPT;
255405	}
255406	if( bReset ) sqlite3Fts5ClearLocale(pConfig);
255407	}
255408	}
255409	}
255410	if( rc==SQLITE_OK && p->nTotalRow<1 ){
255411	rc = FTS5_CORRUPT;
	@@ -255446,11 +256308,13 @@
255446	i64 iOrigin = 0;
255447	sqlite3_stmt *pLookup = 0;
255448	int rc = SQLITE_OK;
255449
255450	assert( p->pConfig->bContentlessDelete );
255451	assert( p->pConfig->eContent==FTS5_CONTENT_NONE );


255452
255453	/* Look up the origin of the document in the %_docsize table. Store
255454	** this in stack variable iOrigin. */
255455	rc = fts5StorageGetStmt(p, FTS5_STMT_LOOKUP_DOCSIZE, &pLookup, 0);
255456	if( rc==SQLITE_OK ){
	@@ -255570,10 +256434,16 @@
255570	}
255571
255572	if( rc==SQLITE_OK ){
255573	if( p->pConfig->bContentlessDelete ){
255574	rc = fts5StorageContentlessDelete(p, iDel);






255575	}else{
255576	rc = fts5StorageDeleteFromIndex(p, iDel, apVal, bSaveRow);
255577	}
255578	}
255579
	@@ -255586,11 +256456,13 @@
255586	rc = sqlite3_reset(pDel);
255587	}
255588	}
255589
255590	/* Delete the %_content record */
255591	if( pConfig->eContent==FTS5_CONTENT_NORMAL ){


255592	if( rc==SQLITE_OK ){
255593	rc = fts5StorageGetStmt(p, FTS5_STMT_DELETE_CONTENT, &pDel, 0);
255594	}
255595	if( rc==SQLITE_OK ){
255596	sqlite3_bind_int64(pDel, 1, iDel);
	@@ -255618,12 +256490,17 @@
255618	pConfig->zDb, pConfig->zName,
255619	pConfig->zDb, pConfig->zName
255620	);
255621	if( rc==SQLITE_OK && pConfig->bColumnsize ){
255622	rc = fts5ExecPrintf(pConfig->db, 0,
255623	"DELETE FROM %Q.'%q_docsize';",
255624	pConfig->zDb, pConfig->zName





255625	);
255626	}
255627
255628	/* Reinitialize the %_data table. This call creates the initial structure
255629	** and averages records. */
	@@ -255660,24 +256537,39 @@
255660	sqlite3Fts5BufferZero(&buf);
255661	rc = sqlite3Fts5IndexBeginWrite(p->pIndex, 0, iRowid);
255662	for(ctx.iCol=0; rc==SQLITE_OK && ctx.iCol<pConfig->nCol; ctx.iCol++){
255663	ctx.szCol = 0;
255664	if( pConfig->abUnindexed[ctx.iCol]==0 ){
255665	int bReset = 0; /* True if tokenizer locale must be reset */
255666	int nText = 0; /* Size of pText in bytes */
255667	const char pText = 0; / Pointer to buffer containing text value */



255668	sqlite3_value *pVal = sqlite3_column_value(pScan, ctx.iCol+1);













255669
255670	rc = sqlite3Fts5ExtractText(pConfig, pVal, 1, &bReset, &pText, &nText);
255671	if( rc==SQLITE_OK ){

255672	rc = sqlite3Fts5Tokenize(pConfig,
255673	FTS5_TOKENIZE_DOCUMENT,
255674	pText, nText,
255675	(void*)&ctx,
255676	fts5StorageInsertCallback
255677	);
255678	if( bReset ) sqlite3Fts5ClearLocale(pConfig);
255679	}
255680	}
255681	sqlite3Fts5BufferAppendVarint(&rc, &buf, ctx.szCol);
255682	p->aTotalSize[ctx.iCol] += (i64)ctx.szCol;
255683	}
	@@ -255740,53 +256632,75 @@
255740	/*
255741	** Insert a new row into the FTS content table.
255742	*/
255743	static int sqlite3Fts5StorageContentInsert(
255744	Fts5Storage *p,

255745	sqlite3_value **apVal,
255746	i64 *piRowid
255747	){
255748	Fts5Config *pConfig = p->pConfig;
255749	int rc = SQLITE_OK;
255750
255751	/* Insert the new row into the %_content table. */
255752	if( pConfig->eContent!=FTS5_CONTENT_NORMAL ){


255753	if( sqlite3_value_type(apVal[1])==SQLITE_INTEGER ){
255754	*piRowid = sqlite3_value_int64(apVal[1]);
255755	}else{
255756	rc = fts5StorageNewRowid(p, piRowid);
255757	}
255758	}else{
255759	sqlite3_stmt pInsert = 0; / Statement to write %_content table */
255760	int i; /* Counter variable */
255761	rc = fts5StorageGetStmt(p, FTS5_STMT_INSERT_CONTENT, &pInsert, 0);
255762	for(i=1; rc==SQLITE_OK && i<=pConfig->nCol+1; i++){
255763	sqlite3_value *pVal = apVal[i];
255764	if( sqlite3_value_nochange(pVal) && p->pSavedRow ){
255765	/* This is an UPDATE statement, and column (i-2) was not modified.
255766	** Retrieve the value from Fts5Storage.pSavedRow instead. */
255767	pVal = sqlite3_column_value(p->pSavedRow, i-1);
255768	}else if( sqlite3_value_subtype(pVal)==FTS5_LOCALE_SUBTYPE ){
255769	assert( pConfig->bLocale );
255770	assert( i>1 );
255771	if( pConfig->abUnindexed[i-2] ){
255772	/* At attempt to insert an fts5_locale() value into an UNINDEXED
255773	** column. Strip the locale away and just bind the text. */













255774	const char *pText = 0;

255775	int nText = 0;
255776	rc = sqlite3Fts5ExtractText(pConfig, pVal, 0, 0, &pText, &nText);
255777	sqlite3_bind_text(pInsert, i, pText, nText, SQLITE_TRANSIENT);
255778	}else{
255779	const u8 pBlob = (const u8)sqlite3_value_blob(pVal);
255780	int nBlob = sqlite3_value_bytes(pVal);
255781	assert( nBlob>4 );
255782	sqlite3_bind_blob(pInsert, i, pBlob+4, nBlob-4, SQLITE_TRANSIENT);
255783	}
255784	continue;
255785	}
255786
255787	rc = sqlite3_bind_value(pInsert, i, pVal);





255788	}
255789	if( rc==SQLITE_OK ){
255790	sqlite3_step(pInsert);
255791	rc = sqlite3_reset(pInsert);
255792	}
	@@ -255817,27 +256731,41 @@
255817	rc = sqlite3Fts5IndexBeginWrite(p->pIndex, 0, iRowid);
255818	}
255819	for(ctx.iCol=0; rc==SQLITE_OK && ctx.iCol<pConfig->nCol; ctx.iCol++){
255820	ctx.szCol = 0;
255821	if( pConfig->abUnindexed[ctx.iCol]==0 ){
255822	int bReset = 0; /* True if tokenizer locale must be reset */
255823	int nText = 0; /* Size of pText in bytes */
255824	const char pText = 0; / Pointer to buffer containing text value */



255825	sqlite3_value *pVal = apVal[ctx.iCol+2];
255826	int bDisk = 0;
255827	if( p->pSavedRow && sqlite3_value_nochange(pVal) ){
255828	pVal = sqlite3_column_value(p->pSavedRow, ctx.iCol+1);
255829	bDisk = 1;













255830	}
255831	rc = sqlite3Fts5ExtractText(pConfig, pVal, bDisk, &bReset, &pText,&nText);
255832	if( rc==SQLITE_OK ){
255833	assert( bReset==0 \|\| pConfig->bLocale );
255834	rc = sqlite3Fts5Tokenize(pConfig,
255835	FTS5_TOKENIZE_DOCUMENT, pText, nText, (void*)&ctx,
255836	fts5StorageInsertCallback
255837	);
255838	if( bReset ) sqlite3Fts5ClearLocale(pConfig);
255839	}
255840	}
255841	sqlite3Fts5BufferAppendVarint(&rc, &buf, ctx.szCol);
255842	p->aTotalSize[ctx.iCol] += (i64)ctx.szCol;
255843	}
	@@ -255998,41 +256926,65 @@
255998	}
255999	if( rc==SQLITE_OK && pConfig->eDetail==FTS5_DETAIL_NONE ){
256000	rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
256001	}
256002	for(i=0; rc==SQLITE_OK && i<pConfig->nCol; i++){
256003	if( pConfig->abUnindexed[i] ) continue;
256004	ctx.iCol = i;
256005	ctx.szCol = 0;
256006	if( pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
256007	rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
256008	}
256009	if( rc==SQLITE_OK ){
256010	int bReset = 0; /* True if tokenizer locale must be reset */
256011	int nText = 0; /* Size of pText in bytes */
256012	const char pText = 0; / Pointer to buffer containing text value */
256013
256014	rc = sqlite3Fts5ExtractText(pConfig,
256015	sqlite3_column_value(pScan, i+1), 1, &bReset, &pText, &nText
256016	);
256017	if( rc==SQLITE_OK ){

















256018	rc = sqlite3Fts5Tokenize(pConfig,
256019	FTS5_TOKENIZE_DOCUMENT,
256020	pText, nText,
256021	(void*)&ctx,
256022	fts5StorageIntegrityCallback
256023	);
256024	if( bReset ) sqlite3Fts5ClearLocale(pConfig);
256025	}
256026	}
256027	if( rc==SQLITE_OK && pConfig->bColumnsize && ctx.szCol!=aColSize[i] ){
256028	rc = FTS5_CORRUPT;
256029	}
256030	aTotalSize[i] += ctx.szCol;
256031	if( pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
256032	sqlite3Fts5TermsetFree(ctx.pTermset);
256033	ctx.pTermset = 0;







256034	}
256035	}
256036	sqlite3Fts5TermsetFree(ctx.pTermset);
256037	ctx.pTermset = 0;
256038
	@@ -256454,11 +257406,11 @@
256454
256455	#define READ_UTF8(zIn, zTerm, c) \
256456	c = *(zIn++); \
256457	if( c>=0xc0 ){ \
256458	c = sqlite3Utf8Trans1[c-0xc0]; \
256459	while( zIn!=zTerm && (*zIn & 0xc0)==0x80 ){ \
256460	c = (c<<6) + (0x3f & *(zIn++)); \
256461	} \
256462	if( c<0x80 \
256463	\|\| (c&0xFFFFF800)==0xD800 \
256464	\|\| (c&0xFFFFFFFE)==0xFFFE ){ c = 0xFFFD; } \
	@@ -257609,22 +258561,22 @@
257609	int rc = SQLITE_OK;
257610	char aBuf[32];
257611	char *zOut = aBuf;
257612	int ii;
257613	const unsigned char zIn = (const unsigned char)pText;
257614	const unsigned char *zEof = &zIn[nText];
257615	u32 iCode;
257616	int aStart[3]; /* Input offset of each character in aBuf[] */
257617
257618	UNUSED_PARAM(unusedFlags);
257619
257620	/* Populate aBuf[] with the characters for the first trigram. */
257621	for(ii=0; ii<3; ii++){
257622	do {
257623	aStart[ii] = zIn - (const unsigned char*)pText;

257624	READ_UTF8(zIn, zEof, iCode);
257625	if( iCode==0 ) return SQLITE_OK;
257626	if( p->bFold ) iCode = sqlite3Fts5UnicodeFold(iCode, p->iFoldParam);
257627	}while( iCode==0 );
257628	WRITE_UTF8(zOut, iCode);
257629	}
257630
	@@ -257641,12 +258593,15 @@
257641	const char *z1;
257642
257643	/* Read characters from the input up until the first non-diacritic */
257644	do {
257645	iNext = zIn - (const unsigned char*)pText;




257646	READ_UTF8(zIn, zEof, iCode);
257647	if( iCode==0 ) break;
257648	if( p->bFold ) iCode = sqlite3Fts5UnicodeFold(iCode, p->iFoldParam);
257649	}while( iCode==0 );
257650
257651	/* Pass the current trigram back to fts5 */
257652	rc = xToken(pCtx, 0, aBuf, zOut-aBuf, aStart[0], iNext);
	@@ -259679,11 +260634,11 @@
259679
259680	return sqlite3_create_module_v2(db, "fts5vocab", &fts5Vocab, p, 0);
259681	}
259682
259683
259684
259685	#endif /* !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_FTS5) */
259686
259687	/************ End of fts5.c **********************************************/
259688	/************ Begin file stmt.c ******************************************/
259689	/*
	@@ -260035,6 +260990,7 @@
260035	#endif /* !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_STMTVTAB) */
260036
260037	/************ End of stmt.c **********************************************/
260038	/* Return the source-id for this library */
260039	SQLITE_API const char *sqlite3_sourceid(void){ return SQLITE_SOURCE_ID; }

260040	/************************ End of sqlite3.c ****************************/
260041

	--- extsrc/sqlite3.c
	+++ extsrc/sqlite3.c
	@@ -1,8 +1,8 @@
1	/******************************************************************************
2	** This file is an amalgamation of many separate C source files from SQLite
3	** version 3.48.0. By combining all the individual C code files into this
4	** single large file, the entire code can be compiled as a single translation
5	** unit. This allows many compilers to do optimizations that would not be
6	** possible if the files were compiled separately. Performance improvements
7	** of 5% or more are commonly seen when SQLite is compiled as a single
8	** translation unit.
	@@ -16,12 +16,15 @@
16	** if you want a wrapper to interface SQLite with your choice of programming
17	** language. The code for the "sqlite3" command-line shell is also in a
18	** separate file. This file contains only code for the core SQLite library.
19	**
20	** The content in this amalgamation comes from Fossil check-in
21	** e2bae4143afd07de1ae55a6d2606a3b541a5 with changes in files:
22	**
23	**
24	*/
25	#ifndef SQLITE_AMALGAMATION
26	#define SQLITE_CORE 1
27	#define SQLITE_AMALGAMATION 1
28	#ifndef SQLITE_PRIVATE
29	# define SQLITE_PRIVATE static
30	#endif
	@@ -460,13 +463,13 @@
463	**
464	** See also: [sqlite3_libversion()],
465	** [sqlite3_libversion_number()], [sqlite3_sourceid()],
466	** [sqlite_version()] and [sqlite_source_id()].
467	*/
468	#define SQLITE_VERSION "3.48.0"
469	#define SQLITE_VERSION_NUMBER 3048000
470	#define SQLITE_SOURCE_ID "2024-12-09 20:46:36 e2bae4143afd07de1ae55a6d2606a3b541a5b94568aa41f6a96e5d1245471653"
471
472	/*
473	** CAPI3REF: Run-Time Library Version Numbers
474	** KEYWORDS: sqlite3_version sqlite3_sourceid
475	**
	@@ -966,10 +969,17 @@
969	**
970	** The SQLITE_IOCAP_BATCH_ATOMIC property means that the underlying
971	** filesystem supports doing multiple write operations atomically when those
972	** write operations are bracketed by [SQLITE_FCNTL_BEGIN_ATOMIC_WRITE] and
973	** [SQLITE_FCNTL_COMMIT_ATOMIC_WRITE].
974	**
975	** The SQLITE_IOCAP_SUBPAGE_READ property means that it is ok to read
976	** from the database file in amounts that are not a multiple of the
977	** page size and that do not begin at a page boundary. Without this
978	** property, SQLite is careful to only do full-page reads and write
979	** on aligned pages, with the one exception that it will do a sub-page
980	** read of the first page to access the database header.
981	*/
982	#define SQLITE_IOCAP_ATOMIC 0x00000001
983	#define SQLITE_IOCAP_ATOMIC512 0x00000002
984	#define SQLITE_IOCAP_ATOMIC1K 0x00000004
985	#define SQLITE_IOCAP_ATOMIC2K 0x00000008
	@@ -982,10 +992,11 @@
992	#define SQLITE_IOCAP_SEQUENTIAL 0x00000400
993	#define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN 0x00000800
994	#define SQLITE_IOCAP_POWERSAFE_OVERWRITE 0x00001000
995	#define SQLITE_IOCAP_IMMUTABLE 0x00002000
996	#define SQLITE_IOCAP_BATCH_ATOMIC 0x00004000
997	#define SQLITE_IOCAP_SUBPAGE_READ 0x00008000
998
999	/*
1000	** CAPI3REF: File Locking Levels
1001	**
1002	** SQLite uses one of these integer values as the second
	@@ -1128,10 +1139,11 @@
1139	** <li> [SQLITE_IOCAP_SEQUENTIAL]
1140	** <li> [SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN]
1141	** <li> [SQLITE_IOCAP_POWERSAFE_OVERWRITE]
1142	** <li> [SQLITE_IOCAP_IMMUTABLE]
1143	** <li> [SQLITE_IOCAP_BATCH_ATOMIC]
1144	** <li> [SQLITE_IOCAP_SUBPAGE_READ]
1145	** </ul>
1146	**
1147	** The SQLITE_IOCAP_ATOMIC property means that all writes of
1148	** any size are atomic. The SQLITE_IOCAP_ATOMICnnn values
1149	** mean that writes of blocks that are nnn bytes in size and
	@@ -1405,10 +1417,15 @@
1417	** The [SQLITE_FCNTL_WIN32_SET_HANDLE] opcode is used for debugging. This
1418	** opcode causes the xFileControl method to swap the file handle with the one
1419	** pointed to by the pArg argument. This capability is used during testing
1420	** and only needs to be supported when SQLITE_TEST is defined.
1421	**
1422	** <li>[[SQLITE_FCNTL_NULL_IO]]
1423	** The [SQLITE_FCNTL_NULL_IO] opcode sets the low-level file descriptor
1424	** or file handle for the [sqlite3_file] object such that it will no longer
1425	** read or write to the database file.
1426	**
1427	** <li>[[SQLITE_FCNTL_WAL_BLOCK]]
1428	** The [SQLITE_FCNTL_WAL_BLOCK] is a signal to the VFS layer that it might
1429	** be advantageous to block on the next WAL lock if the lock is not immediately
1430	** available. The WAL subsystem issues this signal during rare
1431	** circumstances in order to fix a problem with priority inversion.
	@@ -1558,10 +1575,11 @@
1575	#define SQLITE_FCNTL_RESERVE_BYTES 38
1576	#define SQLITE_FCNTL_CKPT_START 39
1577	#define SQLITE_FCNTL_EXTERNAL_READER 40
1578	#define SQLITE_FCNTL_CKSM_FILE 41
1579	#define SQLITE_FCNTL_RESET_CACHE 42
1580	#define SQLITE_FCNTL_NULL_IO 43
1581
1582	/* deprecated names */
1583	#define SQLITE_GET_LOCKPROXYFILE SQLITE_FCNTL_GET_LOCKPROXYFILE
1584	#define SQLITE_SET_LOCKPROXYFILE SQLITE_FCNTL_SET_LOCKPROXYFILE
1585	#define SQLITE_LAST_ERRNO SQLITE_FCNTL_LAST_ERRNO
	@@ -2936,14 +2954,18 @@
2954	**
2955	** ^These functions return the number of rows modified, inserted or
2956	** deleted by the most recently completed INSERT, UPDATE or DELETE
2957	** statement on the database connection specified by the only parameter.
2958	** The two functions are identical except for the type of the return value
2959	** and that if the number of rows modified by the most recent INSERT, UPDATE,
2960	** or DELETE is greater than the maximum value supported by type "int", then
2961	** the return value of sqlite3_changes() is undefined. ^Executing any other
2962	** type of SQL statement does not modify the value returned by these functions.
2963	** For the purposes of this interface, a CREATE TABLE AS SELECT statement
2964	** does not count as an INSERT, UPDATE or DELETE statement and hence the rows
2965	** added to the new table by the CREATE TABLE AS SELECT statement are not
2966	** counted.
2967	**
2968	** ^Only changes made directly by the INSERT, UPDATE or DELETE statement are
2969	** considered - auxiliary changes caused by [CREATE TRIGGER \| triggers],
2970	** [foreign key actions] or [REPLACE] constraint resolution are not counted.
2971	**
	@@ -4499,15 +4521,26 @@
4521	**
4522	** [[SQLITE_PREPARE_NO_VTAB]] <dt>SQLITE_PREPARE_NO_VTAB</dt>
4523	** <dd>The SQLITE_PREPARE_NO_VTAB flag causes the SQL compiler
4524	** to return an error (error code SQLITE_ERROR) if the statement uses
4525	** any virtual tables.
4526	**
4527	** [[SQLITE_PREPARE_DONT_LOG]] <dt>SQLITE_PREPARE_DONT_LOG</dt>
4528	** <dd>The SQLITE_PREPARE_DONT_LOG flag prevents SQL compiler
4529	** errors from being sent to the error log defined by
4530	** [SQLITE_CONFIG_LOG]. This can be used, for example, to do test
4531	** compiles to see if some SQL syntax is well-formed, without generating
4532	** messages on the global error log when it is not. If the test compile
4533	** fails, the sqlite3_prepare_v3() call returns the same error indications
4534	** with or without this flag; it just omits the call to [sqlite3_log()] that
4535	** logs the error.
4536	** </dl>
4537	*/
4538	#define SQLITE_PREPARE_PERSISTENT 0x01
4539	#define SQLITE_PREPARE_NORMALIZE 0x02
4540	#define SQLITE_PREPARE_NO_VTAB 0x04
4541	#define SQLITE_PREPARE_DONT_LOG 0x10
4542
4543	/*
4544	** CAPI3REF: Compiling An SQL Statement
4545	** KEYWORDS: {SQL statement compiler}
4546	** METHOD: sqlite3
	@@ -4536,17 +4569,21 @@
4569	** and sqlite3_prepare_v3()
4570	** interfaces use UTF-8, and sqlite3_prepare16(), sqlite3_prepare16_v2(),
4571	** and sqlite3_prepare16_v3() use UTF-16.
4572	**
4573	** ^If the nByte argument is negative, then zSql is read up to the
4574	** first zero terminator. ^If nByte is positive, then it is the maximum
4575	** number of bytes read from zSql. When nByte is positive, zSql is read
4576	** up to the first zero terminator or until the nByte bytes have been read,
4577	** whichever comes first. ^If nByte is zero, then no prepared
4578	** statement is generated.
4579	** If the caller knows that the supplied string is nul-terminated, then
4580	** there is a small performance advantage to passing an nByte parameter that
4581	** is the number of bytes in the input string <i>including</i>
4582	** the nul-terminator.
4583	** Note that nByte measure the length of the input in bytes, not
4584	** characters, even for the UTF-16 interfaces.
4585	**
4586	** ^If pzTail is not NULL then *pzTail is made to point to the first byte
4587	** past the end of the first SQL statement in zSql. These routines only
4588	** compile the first statement in zSql, so *pzTail is left pointing to
4589	** what remains uncompiled.
	@@ -5913,11 +5950,11 @@
5950	** The SQLITE_SUBTYPE flag indicates to SQLite that a function might call
5951	** [sqlite3_value_subtype()] to inspect the sub-types of its arguments.
5952	** This flag instructs SQLite to omit some corner-case optimizations that
5953	** might disrupt the operation of the [sqlite3_value_subtype()] function,
5954	** causing it to return zero rather than the correct subtype().
5955	** All SQL functions that invoke [sqlite3_value_subtype()] should have this
5956	** property. If the SQLITE_SUBTYPE property is omitted, then the return
5957	** value from [sqlite3_value_subtype()] might sometimes be zero even though
5958	** a non-zero subtype was specified by the function argument expression.
5959	**
5960	** [[SQLITE_RESULT_SUBTYPE]] <dt>SQLITE_RESULT_SUBTYPE</dt><dd>
	@@ -8678,11 +8715,11 @@
8715	#define SQLITE_TESTCTRL_EXTRA_SCHEMA_CHECKS 29
8716	#define SQLITE_TESTCTRL_SEEK_COUNT 30
8717	#define SQLITE_TESTCTRL_TRACEFLAGS 31
8718	#define SQLITE_TESTCTRL_TUNE 32
8719	#define SQLITE_TESTCTRL_LOGEST 33
8720	#define SQLITE_TESTCTRL_USELONGDOUBLE 34 /* NOT USED */
8721	#define SQLITE_TESTCTRL_LAST 34 /* Largest TESTCTRL */
8722
8723	/*
8724	** CAPI3REF: SQL Keyword Checking
8725	**
	@@ -9654,10 +9691,20 @@
9691	** threads may safely make multiple concurrent calls to sqlite3_backup_step().
9692	** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
9693	** APIs are not strictly speaking threadsafe. If they are invoked at the
9694	** same time as another thread is invoking sqlite3_backup_step() it is
9695	** possible that they return invalid values.
9696	**
9697	** <b>Alternatives To Using The Backup API</b>
9698	**
9699	** Other techniques for safely creating a consistent backup of an SQLite
9700	** database include:
9701	**
9702	** <ul>
9703	** <li> The [VACUUM INTO] command.
9704	** <li> The [sqlite3_rsync] utility program.
9705	** </ul>
9706	*/
9707	SQLITE_API sqlite3_backup *sqlite3_backup_init(
9708	sqlite3 pDest, / Destination database handle */
9709	const char zDestName, / Destination database name */
9710	sqlite3 pSource, / Source database handle */
	@@ -10852,10 +10899,18 @@
10899	** schema S in database connection D. ^On success, the
10900	** [sqlite3_snapshot_get(D,S,P)] interface writes a pointer to the newly
10901	** created [sqlite3_snapshot] object into *P and returns SQLITE_OK.
10902	** If there is not already a read-transaction open on schema S when
10903	** this function is called, one is opened automatically.
10904	**
10905	** If a read-transaction is opened by this function, then it is guaranteed
10906	** that the returned snapshot object may not be invalidated by a database
10907	** writer or checkpointer until after the read-transaction is closed. This
10908	** is not guaranteed if a read-transaction is already open when this
10909	** function is called. In that case, any subsequent write or checkpoint
10910	** operation on the database may invalidate the returned snapshot handle,
10911	** even while the read-transaction remains open.
10912	**
10913	** The following must be true for this function to succeed. If any of
10914	** the following statements are false when sqlite3_snapshot_get() is
10915	** called, SQLITE_ERROR is returned. The final value of *P is undefined
10916	** in this case.
	@@ -11172,11 +11227,11 @@
11227	#endif
11228
11229	#if 0
11230	} /* End of the 'extern "C"' block */
11231	#endif
11232	/* #endif for SQLITE3_H will be added by mksqlite3.tcl */
11233
11234	/****** Begin file sqlite3rtree.h *******/
11235	/*
11236	** 2010 August 30
11237	**
	@@ -13423,17 +13478,32 @@
13478	** This is used to access token iToken of phrase hit iIdx within the
13479	** current row. If iIdx is less than zero or greater than or equal to the
13480	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
13481	** output variable (*ppToken) is set to point to a buffer containing the
13482	** matching document token, and (*pnToken) to the size of that buffer in
13483	** bytes.


13484	**
13485	** The output text is not a copy of the document text that was tokenized.
13486	** It is the output of the tokenizer module. For tokendata=1 tables, this
13487	** includes any embedded 0x00 and trailing data.
13488	**
13489	** This API may be slow in some cases if the token identified by parameters
13490	** iIdx and iToken matched a prefix token in the query. In most cases, the
13491	** first call to this API for each prefix token in the query is forced
13492	** to scan the portion of the full-text index that matches the prefix
13493	** token to collect the extra data required by this API. If the prefix
13494	** token matches a large number of token instances in the document set,
13495	** this may be a performance problem.
13496	**
13497	** If the user knows in advance that a query may use this API for a
13498	** prefix token, FTS5 may be configured to collect all required data as part
13499	** of the initial querying of the full-text index, avoiding the second scan
13500	** entirely. This also causes prefix queries that do not use this API to
13501	** run more slowly and use more memory. FTS5 may be configured in this way
13502	** either on a per-table basis using the [FTS5 insttoken \| 'insttoken']
13503	** option, or on a per-query basis using the
13504	** [fts5_insttoken \| fts5_insttoken()] user function.
13505	**
13506	** This API can be quite slow if used with an FTS5 table created with the
13507	** "detail=none" or "detail=column" option.
13508	**
13509	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
	@@ -13521,11 +13591,10 @@
13591	** CUSTOM TOKENIZERS
13592	**
13593	** Applications may also register custom tokenizer types. A tokenizer
13594	** is registered by providing fts5 with a populated instance of the
13595	** following structure. All structure methods must be defined, setting

13596	** any member of the fts5_tokenizer struct to NULL leads to undefined
13597	** behaviour. The structure methods are expected to function as follows:
13598	**
13599	** xCreate:
13600	** This function is used to allocate and initialize a tokenizer instance.
	@@ -13865,10 +13934,11 @@
13934	#endif
13935
13936	#endif /* _FTS5_H */
13937
13938	/****** End of fts5.h *******/
13939	#endif /* SQLITE3_H */
13940
13941	/************ End of sqlite3.h *******************************************/
13942	/************ Continuing where we left off in sqliteInt.h ****************/
13943
13944	/*
	@@ -13910,10 +13980,11 @@
13980	** to count the size: 2^31-1 or 2147483647.
13981	*/
13982	#ifndef SQLITE_MAX_LENGTH
13983	# define SQLITE_MAX_LENGTH 1000000000
13984	#endif
13985	#define SQLITE_MIN_LENGTH 30 /* Minimum value for the length limit */
13986
13987	/*
13988	** This is the maximum number of
13989	**
13990	** * Columns in a table
	@@ -14809,10 +14880,11 @@
14880	#include <stdio.h>
14881	#include <stdlib.h>
14882	#include <string.h>
14883	#include <assert.h>
14884	#include <stddef.h>
14885	#include <ctype.h>
14886
14887	/*
14888	** Use a macro to replace memcpy() if compiled with SQLITE_INLINE_MEMCPY.
14889	** This allows better measurements of where memcpy() is used when running
14890	** cachegrind. But this macro version of memcpy() is very slow so it
	@@ -14831,11 +14903,10 @@
14903	#ifdef SQLITE_OMIT_FLOATING_POINT
14904	# define double sqlite_int64
14905	# define float sqlite_int64
14906	# define fabs(X) ((X)<0?-(X):(X))
14907	# define sqlite3IsOverflow(X) 0

14908	# ifndef SQLITE_BIG_DBL
14909	# define SQLITE_BIG_DBL (((sqlite3_int64)1)<<50)
14910	# endif
14911	# define SQLITE_OMIT_DATETIME_FUNCS 1
14912	# define SQLITE_OMIT_TRACE 1
	@@ -15006,13 +15077,10 @@
15077	# define INT8_TYPE int8_t
15078	# else
15079	# define INT8_TYPE signed char
15080	# endif
15081	#endif



15082	typedef sqlite_int64 i64; /* 8-byte signed integer */
15083	typedef sqlite_uint64 u64; /* 8-byte unsigned integer */
15084	typedef UINT32_TYPE u32; /* 4-byte unsigned integer */
15085	typedef UINT16_TYPE u16; /* 2-byte unsigned integer */
15086	typedef INT16_TYPE i16; /* 2-byte signed integer */
	@@ -16391,10 +16459,13 @@
16459	struct KeyInfo, / First argument to compare function */
16460	BtCursor pCursor / Space to write cursor structure */
16461	);
16462	SQLITE_PRIVATE BtCursor *sqlite3BtreeFakeValidCursor(void);
16463	SQLITE_PRIVATE int sqlite3BtreeCursorSize(void);
16464	#ifdef SQLITE_DEBUG
16465	SQLITE_PRIVATE int sqlite3BtreeClosesWithCursor(Btree,BtCursor);
16466	#endif
16467	SQLITE_PRIVATE void sqlite3BtreeCursorZero(BtCursor*);
16468	SQLITE_PRIVATE void sqlite3BtreeCursorHintFlags(BtCursor*, unsigned);
16469	#ifdef SQLITE_ENABLE_CURSOR_HINTS
16470	SQLITE_PRIVATE void sqlite3BtreeCursorHint(BtCursor*, int, ...);
16471	#endif
	@@ -17004,11 +17075,11 @@
17075
17076	/*
17077	** Additional non-public SQLITE_PREPARE_* flags
17078	*/
17079	#define SQLITE_PREPARE_SAVESQL 0x80 /* Preserve SQL text */
17080	#define SQLITE_PREPARE_MASK 0x1f /* Mask of public flags */
17081
17082	/*
17083	** Prototypes for the VDBE interface. See comments on the implementation
17084	** for a description of what each of these routines does.
17085	*/
	@@ -17719,51 +17790,15 @@
17790	struct FuncDefHash {
17791	FuncDef a[SQLITE_FUNC_HASH_SZ]; / Hash table for functions */
17792	};
17793	#define SQLITE_FUNC_HASH(C,L) (((C)+(L))%SQLITE_FUNC_HASH_SZ)
17794































17795	/*
17796	** typedef for the authorization callback function.
17797	*/
17798	typedef int (sqlite3_xauth)(void,int,const char,const char,const char*,
17799	const char*);





17800
17801	#ifndef SQLITE_OMIT_DEPRECATED
17802	/* This is an extra SQLITE_TRACE macro that indicates "legacy" tracing
17803	** in the style of sqlite3_trace()
17804	*/
	@@ -17920,13 +17955,10 @@
17955	sqlite3 pUnlockConnection; / Connection to watch for unlock */
17956	void pUnlockArg; / Argument to xUnlockNotify */
17957	void (xUnlockNotify)(void , int); / Unlock notify callback */
17958	sqlite3 pNextBlocked; / Next in list of all blocked connections */
17959	#endif



17960	};
17961
17962	/*
17963	** A macro to discover the encoding of a database.
17964	*/
	@@ -19221,11 +19253,11 @@
19253	#define EP_Quoted 0x4000000 /* TK_ID was originally quoted */
19254	#define EP_Static 0x8000000 /* Held in memory not obtained from malloc() */
19255	#define EP_IsTrue 0x10000000 /* Always has boolean value of TRUE */
19256	#define EP_IsFalse 0x20000000 /* Always has boolean value of FALSE */
19257	#define EP_FromDDL 0x40000000 /* Originates from sqlite_schema */
19258	#define EP_SubtArg 0x80000000 /* Is argument to SQLITE_SUBTYPE function */
19259
19260	/* The EP_Propagate mask is a set of properties that automatically propagate
19261	** upwards into parent nodes.
19262	*/
19263	#define EP_Propagate (EP_Collate\|EP_Subquery\|EP_HasFunc)
	@@ -19777,11 +19809,11 @@
19809	**
19810	** SRT_Set The result must be a single column. Store each
19811	** row of result as the key in table pDest->iSDParm.
19812	** Apply the affinity pDest->affSdst before storing
19813	** results. if pDest->iSDParm2 is positive, then it is
19814	** a register holding a Bloom filter for the IN operator
19815	** that should be populated in addition to the
19816	** pDest->iSDParm table. This SRT is used to
19817	** implement "IN (SELECT ...)".
19818	**
19819	** SRT_EphemTab Create an temporary table pDest->iSDParm and store
	@@ -20376,11 +20408,10 @@
20408	u8 bFullMutex; /* True to enable full mutexing */
20409	u8 bOpenUri; /* True to interpret filenames as URIs */
20410	u8 bUseCis; /* Use covering indices for full-scans */
20411	u8 bSmallMalloc; /* Avoid large memory allocations if true */
20412	u8 bExtraSchemaChecks; /* Verify type,name,tbl_name in schema */

20413	#ifdef SQLITE_DEBUG
20414	u8 bJsonSelfcheck; /* Double-check JSON parsing */
20415	#endif
20416	int mxStrlen; /* Maximum string length */
20417	int neverCorrupt; /* Database is always well-formed */
	@@ -20751,19 +20782,10 @@
20782	*/
20783	#if defined(SQLITE_ENABLE_FTS4) && !defined(SQLITE_ENABLE_FTS3)
20784	# define SQLITE_ENABLE_FTS3 1
20785	#endif
20786









20787	/*
20788	** The following macros mimic the standard library functions toupper(),
20789	** isspace(), isalnum(), isdigit() and isxdigit(), respectively. The
20790	** sqlite versions only work for ASCII characters, regardless of locale.
20791	*/
	@@ -21381,11 +21403,11 @@
21403	SQLITE_PRIVATE int sqlite3AtoF(const char z, double, int, u8);
21404	SQLITE_PRIVATE int sqlite3GetInt32(const char , int);
21405	SQLITE_PRIVATE int sqlite3GetUInt32(const char, u32);
21406	SQLITE_PRIVATE int sqlite3Atoi(const char*);
21407	#ifndef SQLITE_OMIT_UTF16
21408	SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *pData, int nByte, int nChar);
21409	#endif
21410	SQLITE_PRIVATE int sqlite3Utf8CharLen(const char *pData, int nByte);
21411	SQLITE_PRIVATE u32 sqlite3Utf8Read(const u8**);
21412	SQLITE_PRIVATE int sqlite3Utf8ReadLimited(const u8, int, u32);
21413	SQLITE_PRIVATE LogEst sqlite3LogEst(u64);
	@@ -22839,13 +22861,10 @@
22861	"UNLINK_AFTER_CLOSE",
22862	#endif
22863	#ifdef SQLITE_UNTESTABLE
22864	"UNTESTABLE",
22865	#endif



22866	#ifdef SQLITE_USE_ALLOCA
22867	"USE_ALLOCA",
22868	#endif
22869	#ifdef SQLITE_USE_FCNTL_TRACE
22870	"USE_FCNTL_TRACE",
	@@ -23117,11 +23136,10 @@
23136	SQLITE_THREADSAFE==1, /* bFullMutex */
23137	SQLITE_USE_URI, /* bOpenUri */
23138	SQLITE_ALLOW_COVERING_INDEX_SCAN, /* bUseCis */
23139	0, /* bSmallMalloc */
23140	1, /* bExtraSchemaChecks */

23141	#ifdef SQLITE_DEBUG
23142	0, /* bJsonSelfcheck */
23143	#endif
23144	0x7ffffffe, /* mxStrlen */
23145	0, /* neverCorrupt */
	@@ -23837,13 +23855,15 @@
23855	UnpackedRecord pNewUnpacked; / Unpacked version of new.* record */
23856	int iNewReg; /* Register for new.* values */
23857	int iBlobWrite; /* Value returned by preupdate_blobwrite() */
23858	i64 iKey1; /* First key value passed to hook */
23859	i64 iKey2; /* Second key value passed to hook */
23860	Mem oldipk; /* Memory cell holding "old" IPK value */
23861	Mem aNew; / Array of new.* values */
23862	Table pTab; / Schema object being updated */
23863	Index pPk; / PK index if pTab is WITHOUT ROWID */
23864	sqlite3_value *apDflt; / Array of default values, if required */
23865	};
23866
23867	/*
23868	** An instance of this object is used to pass an vector of values into
23869	** OP_VFilter, the xFilter method of a virtual table. The vector is the
	@@ -29285,20 +29305,33 @@
29305
29306	#ifndef NDEBUG
29307	/*
29308	** The sqlite3_mutex_held() and sqlite3_mutex_notheld() routine are
29309	** intended for use inside assert() statements.
29310	**
29311	** Because these routines raise false-positive alerts in TSAN, disable
29312	** them (make them always return 1) when compiling with TSAN.
29313	*/
29314	SQLITE_API int sqlite3_mutex_held(sqlite3_mutex *p){
29315	# if defined(__has_feature)
29316	# if __has_feature(thread_sanitizer)
29317	p = 0;
29318	# endif
29319	# endif
29320	assert( p==0 \|\| sqlite3GlobalConfig.mutex.xMutexHeld );
29321	return p==0 \|\| sqlite3GlobalConfig.mutex.xMutexHeld(p);
29322	}
29323	SQLITE_API int sqlite3_mutex_notheld(sqlite3_mutex *p){
29324	# if defined(__has_feature)
29325	# if __has_feature(thread_sanitizer)
29326	p = 0;
29327	# endif
29328	# endif
29329	assert( p==0 \|\| sqlite3GlobalConfig.mutex.xMutexNotheld );
29330	return p==0 \|\| sqlite3GlobalConfig.mutex.xMutexNotheld(p);
29331	}
29332	#endif /* NDEBUG */
29333
29334	#endif /* !defined(SQLITE_MUTEX_OMIT) */
29335
29336	/************ End of mutex.c *********************************************/
29337	/************ Begin file mutex_noop.c ************************************/
	@@ -32272,10 +32305,11 @@
32305	&& (ExprHasProperty(pExpr,EP_OuterON\|EP_InnerON) \|\| pExpr->w.iOfst<=0)
32306	){
32307	pExpr = pExpr->pLeft;
32308	}
32309	if( pExpr==0 ) return;
32310	if( ExprHasProperty(pExpr, EP_FromDDL) ) return;
32311	db->errByteOffset = pExpr->w.iOfst;
32312	}
32313
32314	/*
32315	** Enlarge the memory allocation on a StrAccum object so that it is
	@@ -33001,11 +33035,11 @@
33035	}
33036	if( pItem->fg.isCte ){
33037	sqlite3_str_appendf(&x, " CteUse=0x%p", pItem->u2.pCteUse);
33038	}
33039	if( pItem->fg.isOn \|\| (pItem->fg.isUsing==0 && pItem->u3.pOn!=0) ){
33040	sqlite3_str_appendf(&x, " isOn");
33041	}
33042	if( pItem->fg.isTabFunc ) sqlite3_str_appendf(&x, " isTabFunc");
33043	if( pItem->fg.isCorrelated ) sqlite3_str_appendf(&x, " isCorrelated");
33044	if( pItem->fg.isMaterialized ) sqlite3_str_appendf(&x, " isMaterialized");
33045	if( pItem->fg.viaCoroutine ) sqlite3_str_appendf(&x, " viaCoroutine");
	@@ -34085,10 +34119,14 @@
34119	**
34120	** This routines are given external linkage so that they will always be
34121	** accessible to the debugging, and to avoid warnings about unused
34122	** functions. But these routines only exist in debugging builds, so they
34123	** do not contaminate the interface.
34124	**
34125	** See Also:
34126	**
34127	** sqlite3ShowWhereTerm() in where.c
34128	*/
34129	SQLITE_PRIVATE void sqlite3ShowExpr(const Expr *p){ sqlite3TreeViewExpr(0,p,0); }
34130	SQLITE_PRIVATE void sqlite3ShowExprList(const ExprList *p){ sqlite3TreeViewExprList(0,p,0,0);}
34131	SQLITE_PRIVATE void sqlite3ShowIdList(const IdList *p){ sqlite3TreeViewIdList(0,p,0,0); }
34132	SQLITE_PRIVATE void sqlite3ShowSrcList(const SrcList *p){ sqlite3TreeViewSrcList(0,p); }
	@@ -34687,11 +34725,11 @@
34725	*/
34726	#define READ_UTF8(zIn, zTerm, c) \
34727	c = *(zIn++); \
34728	if( c>=0xc0 ){ \
34729	c = sqlite3Utf8Trans1[c-0xc0]; \
34730	while( zIn<zTerm && (*zIn & 0xc0)==0x80 ){ \
34731	c = (c<<6) + (0x3f & *(zIn++)); \
34732	} \
34733	if( c<0x80 \
34734	\|\| (c&0xFFFFF800)==0xD800 \
34735	\|\| (c&0xFFFFFFFE)==0xFFFE ){ c = 0xFFFD; } \
	@@ -35065,24 +35103,26 @@
35103	assert( m.z \|\| db->mallocFailed );
35104	return m.z;
35105	}
35106
35107	/*
35108	** zIn is a UTF-16 encoded unicode string at least nByte bytes long.
35109	** Return the number of bytes in the first nChar unicode characters
35110	** in pZ. nChar must be non-negative. Surrogate pairs count as a single
35111	** character.
35112	*/
35113	SQLITE_PRIVATE int sqlite3Utf16ByteLen(const void *zIn, int nByte, int nChar){
35114	int c;
35115	unsigned char const *z = zIn;
35116	unsigned char const *zEnd = &z[nByte-1];
35117	int n = 0;
35118
35119	if( SQLITE_UTF16NATIVE==SQLITE_UTF16LE ) z++;
35120	while( n<nChar && ALWAYS(z<=zEnd) ){
35121	c = z[0];
35122	z += 2;
35123	if( c>=0xd8 && c<0xdc && z<=zEnd && z[0]>=0xdc && z[0]<0xe0 ) z += 2;
35124	n++;
35125	}
35126	return (int)(z-(unsigned char const *)zIn)
35127	- (SQLITE_UTF16NATIVE==SQLITE_UTF16LE);
35128	}
	@@ -35659,10 +35699,12 @@
35699	int esign = 1; /* sign of exponent */
35700	int e = 0; /* exponent */
35701	int eValid = 1; /* True exponent is either not used or is well-formed */
35702	int nDigit = 0; /* Number of digits processed */
35703	int eType = 1; /* 1: pure integer, 2+: fractional -1 or less: bad UTF16 */
35704	u64 s2; /* round-tripped significand */
35705	double rr[2];
35706
35707	assert( enc==SQLITE_UTF8 \|\| enc==SQLITE_UTF16LE \|\| enc==SQLITE_UTF16BE );
35708	pResult = 0.0; / Default return value, in case of an error */
35709	if( length==0 ) return 0;
35710
	@@ -35761,81 +35803,65 @@
35803
35804	/* adjust exponent by d, and update sign */
35805	e = (e*esign) + d;
35806
35807	/* Try to adjust the exponent to make it smaller */
35808	while( e>0 && s<((LARGEST_UINT64-0x7ff)/10) ){
35809	s *= 10;
35810	e--;
35811	}
35812	while( e<0 && (s%10)==0 ){
35813	s /= 10;
35814	e++;
35815	}
35816
35817	rr[0] = (double)s;
35818	assert( sizeof(s2)==sizeof(rr[0]) );
35819	#ifdef SQLITE_DEBUG
35820	rr[1] = 18446744073709549568.0;
35821	memcpy(&s2, &rr[1], sizeof(s2));
35822	assert( s2==0x43efffffffffffffLL );
35823	#endif
35824	/* Largest double that can be safely converted to u64
35825	** vvvvvvvvvvvvvvvvvvvvvv */
35826	if( rr[0]<=18446744073709549568.0 ){
35827	s2 = (u64)rr[0];




















35828	rr[1] = s>=s2 ? (double)(s - s2) : -(double)(s2 - s);
35829	}else{
35830	rr[1] = 0.0;
35831	}
35832	assert( rr[1]<=1.0e-10rr[0] ); / Equal only when rr[0]==0.0 */
35833
35834	if( e>0 ){
35835	while( e>=100 ){
35836	e -= 100;
35837	dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
35838	}
35839	while( e>=10 ){
35840	e -= 10;
35841	dekkerMul2(rr, 1.0e+10, 0.0);
35842	}
35843	while( e>=1 ){
35844	e -= 1;
35845	dekkerMul2(rr, 1.0e+01, 0.0);
35846	}
35847	}else{
35848	while( e<=-100 ){
35849	e += 100;
35850	dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
35851	}
35852	while( e<=-10 ){
35853	e += 10;
35854	dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
35855	}
35856	while( e<=-1 ){
35857	e += 1;
35858	dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
35859	}
35860	}
35861	*pResult = rr[0]+rr[1];
35862	if( sqlite3IsNaN(pResult) ) pResult = 1e300*1e300;
35863	if( sign<0 ) pResult = -pResult;
35864	assert( !sqlite3IsNaN(*pResult) );
35865
35866	atof_return:
35867	/* return true if number and no extra non-whitespace characters after */
	@@ -36152,13 +36178,14 @@
36178	*/
36179	SQLITE_PRIVATE void sqlite3FpDecode(FpDecode *p, double r, int iRound, int mxRound){
36180	int i;
36181	u64 v;
36182	int e, exp = 0;
36183	double rr[2];
36184
36185	p->isSpecial = 0;
36186	p->z = p->zBuf;

36187	assert( mxRound>0 );
36188
36189	/* Convert negative numbers to positive. Deal with Infinity, 0.0, and
36190	** NaN. */
36191	if( r<0.0 ){
	@@ -36182,66 +36209,49 @@
36209	return;
36210	}
36211
36212	/* Multiply r by powers of ten until it lands somewhere in between
36213	** 1.0e+19 and 1.0e+17.
36214	**
36215	** Use Dekker-style double-double computation to increase the
36216	** precision.
36217	**
36218	** The error terms on constants like 1.0e+100 computed using the
36219	** decimal extension, for example as follows:
36220	**
36221	** SELECT decimal_exp(decimal_sub('1.0e+100',decimal(1.0e+100)));
36222	*/
36223	rr[0] = r;
36224	rr[1] = 0.0;
36225	if( rr[0]>9.223372036854774784e+18 ){
36226	while( rr[0]>9.223372036854774784e+118 ){
36227	exp += 100;
36228	dekkerMul2(rr, 1.0e-100, -1.99918998026028836196e-117);
36229	}
36230	while( rr[0]>9.223372036854774784e+28 ){
36231	exp += 10;
36232	dekkerMul2(rr, 1.0e-10, -3.6432197315497741579e-27);
36233	}
36234	while( rr[0]>9.223372036854774784e+18 ){
36235	exp += 1;
36236	dekkerMul2(rr, 1.0e-01, -5.5511151231257827021e-18);
36237	}
36238	}else{
36239	while( rr[0]<9.223372036854774784e-83 ){
36240	exp -= 100;
36241	dekkerMul2(rr, 1.0e+100, -1.5902891109759918046e+83);
36242	}
36243	while( rr[0]<9.223372036854774784e+07 ){
36244	exp -= 10;
36245	dekkerMul2(rr, 1.0e+10, 0.0);
36246	}
36247	while( rr[0]<9.22337203685477478e+17 ){
36248	exp -= 1;
36249	dekkerMul2(rr, 1.0e+01, 0.0);
36250	}
36251	}
36252	v = rr[1]<0.0 ? (u64)rr[0]-(u64)(-rr[1]) : (u64)rr[0]+(u64)rr[1];

















36253
36254	/* Extract significant digits. */
36255	i = sizeof(p->zBuf)-1;
36256	assert( v>0 );
36257	while( v ){ p->zBuf[i--] = (v%10) + '0'; v /= 10; }
	@@ -37008,108 +37018,10 @@
37018	i += pIn[i+1];
37019	}while( i<mx );
37020	return 0;
37021	}
37022


































































































37023	/************ End of util.c **********************************************/
37024	/************ Begin file hash.c ******************************************/
37025	/*
37026	** 2001 September 22
37027	**
	@@ -38787,11 +38699,11 @@
38699	# define F_SETLKW 7
38700	# endif
38701	# endif
38702	#else /* !SQLITE_WASI */
38703	# ifndef HAVE_FCHMOD
38704	# define HAVE_FCHMOD 1
38705	# endif
38706	#endif /* SQLITE_WASI */
38707
38708	#ifdef SQLITE_WASI
38709	# define osGetpid(X) (pid_t)1
	@@ -41038,58 +40950,37 @@
40950	** file by this or any other process. If such a lock is held, set *pResOut
40951	** to a non-zero value otherwise *pResOut is set to zero. The return value
40952	** is set to SQLITE_OK unless an I/O error occurs during lock checking.
40953	*/
40954	static int flockCheckReservedLock(sqlite3_file id, int pResOut){
40955	#ifdef SQLITE_DEBUG

40956	unixFile pFile = (unixFile)id;
40957	#else
40958	UNUSED_PARAMETER(id);
40959	#endif
40960
40961	SimulateIOError( return SQLITE_IOERR_CHECKRESERVEDLOCK; );
40962
40963	assert( pFile );
40964	assert( pFile->eFileLock<=SHARED_LOCK );
40965
40966	/* The flock VFS only ever takes exclusive locks (see function flockLock).
40967	** Therefore, if this connection is holding any lock at all, no other
40968	** connection may be holding a RESERVED lock. So set *pResOut to 0
40969	** in this case.
40970	**
40971	** Or, this connection may be holding no lock. In that case, set *pResOut to
40972	** 0 as well. The caller will then attempt to take an EXCLUSIVE lock on the
40973	** db in order to roll the hot journal back. If there is another connection
40974	** holding a lock, that attempt will fail and an SQLITE_BUSY returned to
40975	** the user. With other VFS, we try to avoid this, in order to allow a reader
40976	** to proceed while a writer is preparing its transaction. But that won't
40977	** work with the flock VFS - as it always takes EXCLUSIVE locks - so it is
40978	** not a problem in this case. */
40979	*pResOut = 0;
40980
40981	return SQLITE_OK;























40982	}
40983
40984	/*
40985	** Lock the file with the lock specified by parameter eFileLock - one
40986	** of the following:
	@@ -42582,10 +42473,15 @@
42473	int rc = osIoctl(pFile->h, F2FS_IOC_ABORT_VOLATILE_WRITE);
42474	return rc ? SQLITE_IOERR_ROLLBACK_ATOMIC : SQLITE_OK;
42475	}
42476	#endif /* __linux__ && SQLITE_ENABLE_BATCH_ATOMIC_WRITE */
42477
42478	case SQLITE_FCNTL_NULL_IO: {
42479	osClose(pFile->h);
42480	pFile->h = -1;
42481	return SQLITE_OK;
42482	}
42483	case SQLITE_FCNTL_LOCKSTATE: {
42484	(int)pArg = pFile->eFileLock;
42485	return SQLITE_OK;
42486	}
42487	case SQLITE_FCNTL_LAST_ERRNO: {
	@@ -42723,10 +42619,11 @@
42619
42620	/* Set the POWERSAFE_OVERWRITE flag if requested. */
42621	if( pFd->ctrlFlags & UNIXFILE_PSOW ){
42622	pFd->deviceCharacteristics \|= SQLITE_IOCAP_POWERSAFE_OVERWRITE;
42623	}
42624	pFd->deviceCharacteristics \|= SQLITE_IOCAP_SUBPAGE_READ;
42625
42626	pFd->sectorSize = SQLITE_DEFAULT_SECTOR_SIZE;
42627	}
42628	}
42629	#else
	@@ -42773,19 +42670,19 @@
42670	0;
42671	}else if( !strcmp(fsInfo.f_basetype, "qnx4") ){
42672	pFile->sectorSize = fsInfo.f_bsize;
42673	pFile->deviceCharacteristics =
42674	/* full bitset of atomics from max sector size and smaller */
42675	(((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2) \|
42676	SQLITE_IOCAP_SEQUENTIAL \| /* The ram filesystem has no write behind
42677	** so it is ordered */
42678	0;
42679	}else if( strstr(fsInfo.f_basetype, "dos") ){
42680	pFile->sectorSize = fsInfo.f_bsize;
42681	pFile->deviceCharacteristics =
42682	/* full bitset of atomics from max sector size and smaller */
42683	(((pFile->sectorSize / 512 * SQLITE_IOCAP_ATOMIC512) << 1) - 2) \|
42684	SQLITE_IOCAP_SEQUENTIAL \| /* The ram filesystem has no write behind
42685	** so it is ordered */
42686	0;
42687	}else{
42688	pFile->deviceCharacteristics =
	@@ -50462,10 +50359,15 @@
50359	OSTRACE(("FCNTL oldFile=%p, newFile=%p, rc=SQLITE_OK\n",
50360	hOldFile, pFile->h));
50361	return SQLITE_OK;
50362	}
50363	#endif
50364	case SQLITE_FCNTL_NULL_IO: {
50365	(void)osCloseHandle(pFile->h);
50366	pFile->h = NULL;
50367	return SQLITE_OK;
50368	}
50369	case SQLITE_FCNTL_TEMPFILENAME: {
50370	char *zTFile = 0;
50371	int rc = winGetTempname(pFile->pVfs, &zTFile);
50372	if( rc==SQLITE_OK ){
50373	(char*)pArg = zTFile;
	@@ -50523,11 +50425,11 @@
50425	/*
50426	** Return a vector of device characteristics.
50427	*/
50428	static int winDeviceCharacteristics(sqlite3_file *id){
50429	winFile p = (winFile)id;
50430	return SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN \| SQLITE_IOCAP_SUBPAGE_READ \|
50431	((p->ctrlFlags & WINFILE_PSOW)?SQLITE_IOCAP_POWERSAFE_OVERWRITE:0);
50432	}
50433
50434	/*
50435	** Windows will only let you create file view mappings
	@@ -51911,11 +51813,11 @@
51813	*/
51814	char zTmpname = 0; / For temporary filename, if necessary. */
51815
51816	int rc = SQLITE_OK; /* Function Return Code */
51817	#if !defined(NDEBUG) \|\| SQLITE_OS_WINCE
51818	int eType = flags&0x0FFF00; /* Type of file to open */
51819	#endif
51820
51821	int isExclusive = (flags & SQLITE_OPEN_EXCLUSIVE);
51822	int isDelete = (flags & SQLITE_OPEN_DELETEONCLOSE);
51823	int isCreate = (flags & SQLITE_OPEN_CREATE);
	@@ -58131,24 +58033,32 @@
58033	#ifdef SQLITE_DIRECT_OVERFLOW_READ
58034	/*
58035	** Return true if page pgno can be read directly from the database file
58036	** by the b-tree layer. This is the case if:
58037	**
58038	** (1) the database file is open
58039	** (2) the VFS for the database is able to do unaligned sub-page reads
58040	** (3) there are no dirty pages in the cache, and
58041	** (4) the desired page is not currently in the wal file.
58042	*/
58043	SQLITE_PRIVATE int sqlite3PagerDirectReadOk(Pager *pPager, Pgno pgno){
58044	assert( pPager!=0 );
58045	assert( pPager->fd!=0 );
58046	if( pPager->fd->pMethods==0 ) return 0; /* Case (1) */
58047	if( sqlite3PCacheIsDirty(pPager->pPCache) ) return 0; /* Failed (3) */
58048	#ifndef SQLITE_OMIT_WAL
58049	if( pPager->pWal ){
58050	u32 iRead = 0;
58051	(void)sqlite3WalFindFrame(pPager->pWal, pgno, &iRead);
58052	return iRead==0; /* Condition (4) */
58053	}
58054	#endif
58055	assert( pPager->fd->pMethods->xDeviceCharacteristics!=0 );
58056	if( (pPager->fd->pMethods->xDeviceCharacteristics(pPager->fd)
58057	& SQLITE_IOCAP_SUBPAGE_READ)==0 ){
58058	return 0; /* Case (2) */
58059	}
58060	return 1;
58061	}
58062	#endif
58063
58064	#ifndef SQLITE_OMIT_WAL
	@@ -65171,11 +65081,11 @@
65081	** 20: Salt-2, a different random integer changing with each ckpt
65082	** 24: Checksum-1 (first part of checksum for first 24 bytes of header).
65083	** 28: Checksum-2 (second part of checksum for first 24 bytes of header).
65084	**
65085	** Immediately following the wal-header are zero or more frames. Each
65086	** frame consists of a 24-byte frame-header followed by <page-size> bytes
65087	** of page data. The frame-header is six big-endian 32-bit unsigned
65088	** integer values, as follows:
65089	**
65090	** 0: Page number.
65091	** 4: For commit records, the size of the database image in pages
	@@ -65668,10 +65578,11 @@
65578	int nSehTry; /* Number of nested SEH_TRY{} blocks */
65579	u8 lockError; /* True if a locking error has occurred */
65580	#endif
65581	#ifdef SQLITE_ENABLE_SNAPSHOT
65582	WalIndexHdr pSnapshot; / Start transaction here if not NULL */
65583	int bGetSnapshot; /* Transaction opened for sqlite3_get_snapshot() */
65584	#endif
65585	#ifdef SQLITE_ENABLE_SETLK_TIMEOUT
65586	sqlite3 *db;
65587	#endif
65588	};
	@@ -67560,11 +67471,11 @@
67471	return SQLITE_IOERR_IN_PAGE;
67472	}
67473
67474	/*
67475	** Assert that the Wal.lockMask mask, which indicates the locks held
67476	** by the connection, is consistent with the Wal.readLock, Wal.writeLock
67477	** and Wal.ckptLock variables. To be used as:
67478	**
67479	** assert( walAssertLockmask(pWal) );
67480	*/
67481	static int walAssertLockmask(Wal *pWal){
	@@ -68224,11 +68135,11 @@
68135	assert( pWal->apWiData[0]!=0 );
68136	pInfo = walCkptInfo(pWal);
68137	SEH_INJECT_FAULT;
68138	if( !useWal && AtomicLoad(&pInfo->nBackfill)==pWal->hdr.mxFrame
68139	#ifdef SQLITE_ENABLE_SNAPSHOT
68140	&& ((pWal->bGetSnapshot==0 && pWal->pSnapshot==0) \|\| pWal->hdr.mxFrame==0)
68141	#endif
68142	){
68143	/* The WAL has been completely backfilled (or it is empty).
68144	** and can be safely ignored.
68145	*/
	@@ -69624,11 +69535,24 @@
69535	*/
69536	SQLITE_PRIVATE void sqlite3WalSnapshotOpen(
69537	Wal *pWal,
69538	sqlite3_snapshot *pSnapshot
69539	){
69540	if( pSnapshot && ((WalIndexHdr*)pSnapshot)->iVersion==0 ){
69541	/* iVersion==0 means that this is a call to sqlite3_snapshot_get(). In
69542	** this case set the bGetSnapshot flag so that if the call to
69543	** sqlite3_snapshot_get() is about to read transaction on this wal
69544	** file, it does not take read-lock 0 if the wal file has been completely
69545	** checkpointed. Taking read-lock 0 would work, but then it would be
69546	** possible for a subsequent writer to destroy the snapshot even while
69547	** this connection is holding its read-transaction open. This is contrary
69548	** to user expectations, so we avoid it by not taking read-lock 0. */
69549	pWal->bGetSnapshot = 1;
69550	}else{
69551	pWal->pSnapshot = (WalIndexHdr*)pSnapshot;
69552	pWal->bGetSnapshot = 0;
69553	}
69554	}
69555
69556	/*
69557	** Return a +ve value if snapshot p1 is newer than p2. A -ve value if
69558	** p1 is older than p2 and zero if p1 and p2 are the same snapshot.
	@@ -75504,10 +75428,29 @@
75428	** this routine.
75429	*/
75430	SQLITE_PRIVATE int sqlite3BtreeCursorSize(void){
75431	return ROUND8(sizeof(BtCursor));
75432	}
75433
75434	#ifdef SQLITE_DEBUG
75435	/*
75436	** Return true if and only if the Btree object will be automatically
75437	** closed with the BtCursor closes. This is used within assert() statements
75438	** only.
75439	*/
75440	SQLITE_PRIVATE int sqlite3BtreeClosesWithCursor(
75441	Btree pBtree, / the btree object */
75442	BtCursor pCur / Corresponding cursor */
75443	){
75444	BtShared *pBt = pBtree->pBt;
75445	if( (pBt->openFlags & BTREE_SINGLE)==0 ) return 0;
75446	if( pBt->pCursor!=pCur ) return 0;
75447	if( pCur->pNext!=0 ) return 0;
75448	if( pCur->pBtree!=pBtree ) return 0;
75449	return 1;
75450	}
75451	#endif
75452
75453	/*
75454	** Initialize memory that will be converted into a BtCursor object.
75455	**
75456	** The simple approach here would be to memset() the entire object
	@@ -84538,11 +84481,12 @@
84481	if( apVal==0 ){
84482	rc = SQLITE_NOMEM_BKPT;
84483	goto value_from_function_out;
84484	}
84485	for(i=0; i<nVal; i++){
84486	rc = sqlite3Stat4ValueFromExpr(pCtx->pParse, pList->a[i].pExpr, aff,
84487	&apVal[i]);
84488	if( apVal[i]==0 \|\| rc!=SQLITE_OK ) goto value_from_function_out;
84489	}
84490	}
84491
84492	pVal = valueNew(db, pCtx);
	@@ -89571,11 +89515,11 @@
89515	/* The following two functions are used only within testcase() to prove
89516	** test coverage. These functions do no exist for production builds.
89517	** We must use separate SQLITE_NOINLINE functions here, since otherwise
89518	** optimizer code movement causes gcov to become very confused.
89519	*/
89520	#if defined(SQLITE_COVERAGE_TEST) \|\| defined(SQLITE_DEBUG)
89521	static int SQLITE_NOINLINE doubleLt(double a, double b){ return a<b; }
89522	static int SQLITE_NOINLINE doubleEq(double a, double b){ return a==b; }
89523	#endif
89524
89525	/*
	@@ -89586,17 +89530,10 @@
89530	SQLITE_PRIVATE int sqlite3IntFloatCompare(i64 i, double r){
89531	if( sqlite3IsNaN(r) ){
89532	/* SQLite considers NaN to be a NULL. And all integer values are greater
89533	** than NULL */
89534	return 1;







89535	}else{
89536	i64 y;
89537	if( r<-9223372036854775808.0 ) return +1;
89538	if( r>=9223372036854775808.0 ) return -1;
89539	y = (i64)r;
	@@ -90595,17 +90532,25 @@
90532	db->xPreUpdateCallback(db->pPreUpdateArg, db, op, zDb, zTbl, iKey1, iKey2);
90533	db->pPreUpdate = 0;
90534	sqlite3DbFree(db, preupdate.aRecord);
90535	vdbeFreeUnpacked(db, preupdate.keyinfo.nKeyField+1, preupdate.pUnpacked);
90536	vdbeFreeUnpacked(db, preupdate.keyinfo.nKeyField+1, preupdate.pNewUnpacked);
90537	sqlite3VdbeMemRelease(&preupdate.oldipk);
90538	if( preupdate.aNew ){
90539	int i;
90540	for(i=0; i<pCsr->nField; i++){
90541	sqlite3VdbeMemRelease(&preupdate.aNew[i]);
90542	}
90543	sqlite3DbNNFreeNN(db, preupdate.aNew);
90544	}
90545	if( preupdate.apDflt ){
90546	int i;
90547	for(i=0; i<pTab->nCol; i++){
90548	sqlite3ValueFree(preupdate.apDflt[i]);
90549	}
90550	sqlite3DbFree(db, preupdate.apDflt);
90551	}
90552	}
90553	#endif /* SQLITE_ENABLE_PREUPDATE_HOOK */
90554
90555	/************ End of vdbeaux.c *******************************************/
90556	/************ Begin file vdbeapi.c ***************************************/
	@@ -92230,10 +92175,21 @@
92175	** A successful evaluation of this routine acquires the mutex on p.
92176	** the mutex is released if any kind of error occurs.
92177	**
92178	** The error code stored in database p->db is overwritten with the return
92179	** value in any case.
92180	**
92181	** (tag-20240917-01) If vdbeUnbind(p,(u32)(i-1)) returns SQLITE_OK,
92182	** that means all of the the following will be true:
92183	**
92184	** p!=0
92185	** p->pVar!=0
92186	** i>0
92187	** i<=p->nVar
92188	**
92189	** An assert() is normally added after vdbeUnbind() to help static analyzers
92190	** realize this.
92191	*/
92192	static int vdbeUnbind(Vdbe *p, unsigned int i){
92193	Mem *pVar;
92194	if( vdbeSafetyNotNull(p) ){
92195	return SQLITE_MISUSE_BKPT;
	@@ -92287,10 +92243,11 @@
92243	Mem *pVar;
92244	int rc;
92245
92246	rc = vdbeUnbind(p, (u32)(i-1));
92247	if( rc==SQLITE_OK ){
92248	assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92249	if( zData!=0 ){
92250	pVar = &p->aVar[i-1];
92251	rc = sqlite3VdbeMemSetStr(pVar, zData, nData, encoding, xDel);
92252	if( rc==SQLITE_OK && encoding!=0 ){
92253	rc = sqlite3VdbeChangeEncoding(pVar, ENC(p->db));
	@@ -92336,10 +92293,11 @@
92293	SQLITE_API int sqlite3_bind_double(sqlite3_stmt *pStmt, int i, double rValue){
92294	int rc;
92295	Vdbe p = (Vdbe )pStmt;
92296	rc = vdbeUnbind(p, (u32)(i-1));
92297	if( rc==SQLITE_OK ){
92298	assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92299	sqlite3VdbeMemSetDouble(&p->aVar[i-1], rValue);
92300	sqlite3_mutex_leave(p->db->mutex);
92301	}
92302	return rc;
92303	}
	@@ -92349,10 +92307,11 @@
92307	SQLITE_API int sqlite3_bind_int64(sqlite3_stmt *pStmt, int i, sqlite_int64 iValue){
92308	int rc;
92309	Vdbe p = (Vdbe )pStmt;
92310	rc = vdbeUnbind(p, (u32)(i-1));
92311	if( rc==SQLITE_OK ){
92312	assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92313	sqlite3VdbeMemSetInt64(&p->aVar[i-1], iValue);
92314	sqlite3_mutex_leave(p->db->mutex);
92315	}
92316	return rc;
92317	}
	@@ -92359,10 +92318,11 @@
92318	SQLITE_API int sqlite3_bind_null(sqlite3_stmt *pStmt, int i){
92319	int rc;
92320	Vdbe p = (Vdbe)pStmt;
92321	rc = vdbeUnbind(p, (u32)(i-1));
92322	if( rc==SQLITE_OK ){
92323	assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92324	sqlite3_mutex_leave(p->db->mutex);
92325	}
92326	return rc;
92327	}
92328	SQLITE_API int sqlite3_bind_pointer(
	@@ -92374,10 +92334,11 @@
92334	){
92335	int rc;
92336	Vdbe p = (Vdbe)pStmt;
92337	rc = vdbeUnbind(p, (u32)(i-1));
92338	if( rc==SQLITE_OK ){
92339	assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92340	sqlite3VdbeMemSetPointer(&p->aVar[i-1], pPtr, zPTtype, xDestructor);
92341	sqlite3_mutex_leave(p->db->mutex);
92342	}else if( xDestructor ){
92343	xDestructor(pPtr);
92344	}
	@@ -92455,10 +92416,11 @@
92416	SQLITE_API int sqlite3_bind_zeroblob(sqlite3_stmt *pStmt, int i, int n){
92417	int rc;
92418	Vdbe p = (Vdbe )pStmt;
92419	rc = vdbeUnbind(p, (u32)(i-1));
92420	if( rc==SQLITE_OK ){
92421	assert( p!=0 && p->aVar!=0 && i>0 && i<=p->nVar ); /* tag-20240917-01 */
92422	#ifndef SQLITE_OMIT_INCRBLOB
92423	sqlite3VdbeMemSetZeroBlob(&p->aVar[i-1], n);
92424	#else
92425	rc = sqlite3VdbeMemSetZeroBlob(&p->aVar[i-1], n);
92426	#endif
	@@ -92789,41 +92751,68 @@
92751	if( iIdx>=p->pCsr->nField \|\| iIdx<0 ){
92752	rc = SQLITE_RANGE;
92753	goto preupdate_old_out;
92754	}
92755






















92756	if( iIdx==p->pTab->iPKey ){
92757	*ppValue = pMem = &p->oldipk;
92758	sqlite3VdbeMemSetInt64(pMem, p->iKey1);
92759	}else{
92760
92761	/* If the old.* record has not yet been loaded into memory, do so now. */
92762	if( p->pUnpacked==0 ){
92763	u32 nRec;
92764	u8 *aRec;
92765
92766	assert( p->pCsr->eCurType==CURTYPE_BTREE );
92767	nRec = sqlite3BtreePayloadSize(p->pCsr->uc.pCursor);
92768	aRec = sqlite3DbMallocRaw(db, nRec);
92769	if( !aRec ) goto preupdate_old_out;
92770	rc = sqlite3BtreePayload(p->pCsr->uc.pCursor, 0, nRec, aRec);
92771	if( rc==SQLITE_OK ){
92772	p->pUnpacked = vdbeUnpackRecord(&p->keyinfo, nRec, aRec);
92773	if( !p->pUnpacked ) rc = SQLITE_NOMEM;
92774	}
92775	if( rc!=SQLITE_OK ){
92776	sqlite3DbFree(db, aRec);
92777	goto preupdate_old_out;
92778	}
92779	p->aRecord = aRec;
92780	}
92781
92782	pMem = *ppValue = &p->pUnpacked->aMem[iIdx];
92783	if( iIdx>=p->pUnpacked->nField ){
92784	/* This occurs when the table has been extended using ALTER TABLE
92785	** ADD COLUMN. The value to return is the default value of the column. */
92786	Column *pCol = &p->pTab->aCol[iIdx];
92787	if( pCol->iDflt>0 ){
92788	if( p->apDflt==0 ){
92789	int nByte = sizeof(sqlite3_value)p->pTab->nCol;
92790	p->apDflt = (sqlite3_value**)sqlite3DbMallocZero(db, nByte);
92791	if( p->apDflt==0 ) goto preupdate_old_out;
92792	}
92793	if( p->apDflt[iIdx]==0 ){
92794	sqlite3_value *pVal = 0;
92795	Expr *pDflt;
92796	assert( p->pTab!=0 && IsOrdinaryTable(p->pTab) );
92797	pDflt = p->pTab->u.tab.pDfltList->a[pCol->iDflt-1].pExpr;
92798	rc = sqlite3ValueFromExpr(db, pDflt, ENC(db), pCol->affinity, &pVal);
92799	if( rc==SQLITE_OK && pVal==0 ){
92800	rc = SQLITE_CORRUPT_BKPT;
92801	}
92802	p->apDflt[iIdx] = pVal;
92803	}
92804	*ppValue = p->apDflt[iIdx];
92805	}else{
92806	ppValue = (sqlite3_value )columnNullValue();
92807	}
92808	}else if( p->pTab->aCol[iIdx].affinity==SQLITE_AFF_REAL ){
92809	if( pMem->flags & (MEM_Int\|MEM_IntReal) ){
92810	testcase( pMem->flags & MEM_Int );
92811	testcase( pMem->flags & MEM_IntReal );
92812	sqlite3VdbeMemRealify(pMem);
92813	}
92814	}
92815	}
92816
92817	preupdate_old_out:
92818	sqlite3Error(db, rc);
	@@ -93367,10 +93356,108 @@
93356	** commenting and indentation practices when changing or adding code.
93357	*/
93358	/* #include "sqliteInt.h" */
93359	/* #include "vdbeInt.h" */
93360
93361	/*
93362	** High-resolution hardware timer used for debugging and testing only.
93363	*/
93364	#if defined(VDBE_PROFILE) \
93365	\|\| defined(SQLITE_PERFORMANCE_TRACE) \
93366	\|\| defined(SQLITE_ENABLE_STMT_SCANSTATUS)
93367	/************ Include hwtime.h in the middle of vdbe.c *******************/
93368	/************ Begin file hwtime.h ****************************************/
93369	/*
93370	** 2008 May 27
93371	**
93372	** The author disclaims copyright to this source code. In place of
93373	** a legal notice, here is a blessing:
93374	**
93375	** May you do good and not evil.
93376	** May you find forgiveness for yourself and forgive others.
93377	** May you share freely, never taking more than you give.
93378	**
93379	******************************************************************************
93380	**
93381	** This file contains inline asm code for retrieving "high-performance"
93382	** counters for x86 and x86_64 class CPUs.
93383	*/
93384	#ifndef SQLITE_HWTIME_H
93385	#define SQLITE_HWTIME_H
93386
93387	/*
93388	** The following routine only works on Pentium-class (or newer) processors.
93389	** It uses the RDTSC opcode to read the cycle count value out of the
93390	** processor and returns that value. This can be used for high-res
93391	** profiling.
93392	*/
93393	#if !defined(__STRICT_ANSI__) && \
93394	(defined(__GNUC__) \|\| defined(_MSC_VER)) && \
93395	(defined(i386) \|\| defined(__i386__) \|\| defined(_M_IX86))
93396
93397	#if defined(__GNUC__)
93398
93399	__inline__ sqlite_uint64 sqlite3Hwtime(void){
93400	unsigned int lo, hi;
93401	__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
93402	return (sqlite_uint64)hi << 32 \| lo;
93403	}
93404
93405	#elif defined(_MSC_VER)
93406
93407	__declspec(naked) __inline sqlite_uint64 __cdecl sqlite3Hwtime(void){
93408	__asm {
93409	rdtsc
93410	ret ; return value at EDX:EAX
93411	}
93412	}
93413
93414	#endif
93415
93416	#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__x86_64__))
93417
93418	__inline__ sqlite_uint64 sqlite3Hwtime(void){
93419	unsigned int lo, hi;
93420	__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
93421	return (sqlite_uint64)hi << 32 \| lo;
93422	}
93423
93424	#elif !defined(__STRICT_ANSI__) && (defined(__GNUC__) && defined(__ppc__))
93425
93426	__inline__ sqlite_uint64 sqlite3Hwtime(void){
93427	unsigned long long retval;
93428	unsigned long junk;
93429	__asm__ __volatile__ ("\n\
93430	1: mftbu %1\n\
93431	mftb %L0\n\
93432	mftbu %0\n\
93433	cmpw %0,%1\n\
93434	bne 1b"
93435	: "=r" (retval), "=r" (junk));
93436	return retval;
93437	}
93438
93439	#else
93440
93441	/*
93442	** asm() is needed for hardware timing support. Without asm(),
93443	** disable the sqlite3Hwtime() routine.
93444	**
93445	** sqlite3Hwtime() is only used for some obscure debugging
93446	** and analysis configurations, not in any deliverable, so this
93447	** should not be a great loss.
93448	*/
93449	SQLITE_PRIVATE sqlite_uint64 sqlite3Hwtime(void){ return ((sqlite_uint64)0); }
93450
93451	#endif
93452
93453	#endif /* !defined(SQLITE_HWTIME_H) */
93454
93455	/************ End of hwtime.h ********************************************/
93456	/************ Continuing where we left off in vdbe.c *********************/
93457	#endif
93458
93459	/*
93460	** Invoke this macro on memory cells just prior to changing the
93461	** value of the cell. This macro verifies that shallow copies are
93462	** not misused. A shallow copy of a string or blob just copies a
93463	** pointer to the string or blob, not the content. If the original
	@@ -97875,12 +97962,17 @@
97962	0, pCx->uc.pCursor);
97963	pCx->isTable = 1;
97964	}
97965	}
97966	pCx->isOrdered = (pOp->p5!=BTREE_UNORDERED);
97967	assert( p->apCsr[pOp->p1]==pCx );
97968	if( rc ){
97969	assert( !sqlite3BtreeClosesWithCursor(pCx->ub.pBtx, pCx->uc.pCursor) );
97970	sqlite3BtreeClose(pCx->ub.pBtx);
97971	p->apCsr[pOp->p1] = 0; /* Not required; helps with static analysis */
97972	}else{
97973	assert( sqlite3BtreeClosesWithCursor(pCx->ub.pBtx, pCx->uc.pCursor) );
97974	}
97975	}
97976	}
97977	if( rc ) goto abort_due_to_error;
97978	pCx->nullRow = 1;
	@@ -102394,11 +102486,11 @@
102486	** element.
102487	**
102488	** As with all opcodes, the meanings of the parameters for OP_Explain
102489	** are subject to change from one release to the next. Applications
102490	** should not attempt to interpret or use any of the information
102491	** contained in the OP_Explain opcode. The information provided by this
102492	** opcode is intended for testing and debugging use only.
102493	*/
102494	default: { /* This is really OP_Noop, OP_Explain */
102495	assert( pOp->opcode==OP_Noop \|\| pOp->opcode==OP_Explain );
102496
	@@ -107490,11 +107582,11 @@
107582	** non-VIEW candidate plus multiple VIEW candidates. In other
107583	** words non-VIEW candidate terms take precedence over VIEWs.
107584	*/
107585	if( cntTab==0
107586	\|\| (cntTab==1
107587	&& pMatch!=0
107588	&& ALWAYS(pMatch->pSTab!=0)
107589	&& (pMatch->pSTab->tabFlags & TF_Ephemeral)!=0
107590	&& (pTab->tabFlags & TF_Ephemeral)==0)
107591	){
107592	cntTab = 1;
	@@ -108123,12 +108215,12 @@
108215	}
108216
108217	/* Resolve function names
108218	*/
108219	case TK_FUNCTION: {
108220	ExprList pList; / The argument list */
108221	int n; /* Number of arguments */
108222	int no_such_func = 0; /* True if no such function exists */
108223	int wrong_num_args = 0; /* True if wrong number of arguments */
108224	int is_agg = 0; /* True if is an aggregate function */
108225	const char zId; / The function name. */
108226	FuncDef pDef; / Information about the function */
	@@ -108137,10 +108229,12 @@
108229	#ifndef SQLITE_OMIT_WINDOWFUNC
108230	Window *pWin = (IsWindowFunc(pExpr) ? pExpr->y.pWin : 0);
108231	#endif
108232	assert( !ExprHasProperty(pExpr, EP_xIsSelect\|EP_IntValue) );
108233	assert( pExpr->pLeft==0 \|\| pExpr->pLeft->op==TK_ORDER );
108234	pList = pExpr->x.pList;
108235	n = pList ? pList->nExpr : 0;
108236	zId = pExpr->u.zToken;
108237	pDef = sqlite3FindFunction(pParse->db, zId, n, enc, 0);
108238	if( pDef==0 ){
108239	pDef = sqlite3FindFunction(pParse->db, zId, -2, enc, 0);
108240	if( pDef==0 ){
	@@ -108185,10 +108279,28 @@
108279	pExpr->op = TK_NULL;
108280	return WRC_Prune;
108281	}
108282	}
108283	#endif
108284
108285	/* If the function may call sqlite3_value_subtype(), then set the
108286	** EP_SubtArg flag on all of its argument expressions. This prevents
108287	** where.c from replacing the expression with a value read from an
108288	** index on the same expression, which will not have the correct
108289	** subtype. Also set the flag if the function expression itself is
108290	** an EP_SubtArg expression. In this case subtypes are required as
108291	** the function may return a value with a subtype back to its
108292	** caller using sqlite3_result_value(). */
108293	if( (pDef->funcFlags & SQLITE_SUBTYPE)
108294	\|\| ExprHasProperty(pExpr, EP_SubtArg)
108295	){
108296	int ii;
108297	for(ii=0; ii<n; ii++){
108298	ExprSetProperty(pList->a[ii].pExpr, EP_SubtArg);
108299	}
108300	}
108301
108302	if( pDef->funcFlags & (SQLITE_FUNC_CONSTANT\|SQLITE_FUNC_SLOCHNG) ){
108303	/* For the purposes of the EP_ConstFunc flag, date and time
108304	** functions and other functions that change slowly are considered
108305	** constant because they are constant for the duration of one query.
108306	** This allows them to be factored out of inner loops. */
	@@ -111951,11 +112063,11 @@
112063	**
112064	** (4) If pSrc is the right operand of a LEFT JOIN, then...
112065	** (4a) pExpr must come from an ON clause..
112066	** (4b) and specifically the ON clause associated with the LEFT JOIN.
112067	**
112068	** (5) If pSrc is the right operand of a LEFT JOIN or the left
112069	** operand of a RIGHT JOIN, then pExpr must be from the WHERE
112070	** clause, not an ON clause.
112071	**
112072	** (6) Either:
112073	**
	@@ -113858,10 +113970,63 @@
113970	}
113971	#endif /* !defined(SQLITE_UNTESTABLE) */
113972	}
113973	return target;
113974	}
113975
113976	/*
113977	** Expression Node callback for sqlite3ExprCanReturnSubtype().
113978	**
113979	** Only a function call is able to return a subtype. So if the node
113980	** is not a function call, return WRC_Prune immediately.
113981	**
113982	** A function call is able to return a subtype if it has the
113983	** SQLITE_RESULT_SUBTYPE property.
113984	**
113985	** Assume that every function is able to pass-through a subtype from
113986	** one of its argument (using sqlite3_result_value()). Most functions
113987	** are not this way, but we don't have a mechanism to distinguish those
113988	** that are from those that are not, so assume they all work this way.
113989	** That means that if one of its arguments is another function and that
113990	** other function is able to return a subtype, then this function is
113991	** able to return a subtype.
113992	*/
113993	static int exprNodeCanReturnSubtype(Walker pWalker, Expr pExpr){
113994	int n;
113995	FuncDef *pDef;
113996	sqlite3 *db;
113997	if( pExpr->op!=TK_FUNCTION ){
113998	return WRC_Prune;
113999	}
114000	assert( ExprUseXList(pExpr) );
114001	db = pWalker->pParse->db;
114002	n = ALWAYS(pExpr->x.pList) ? pExpr->x.pList->nExpr : 0;
114003	pDef = sqlite3FindFunction(db, pExpr->u.zToken, n, ENC(db), 0);
114004	if( NEVER(pDef==0) \|\| (pDef->funcFlags & SQLITE_RESULT_SUBTYPE)!=0 ){
114005	pWalker->eCode = 1;
114006	return WRC_Prune;
114007	}
114008	return WRC_Continue;
114009	}
114010
114011	/*
114012	** Return TRUE if expression pExpr is able to return a subtype.
114013	**
114014	** A TRUE return does not guarantee that a subtype will be returned.
114015	** It only indicates that a subtype return is possible. False positives
114016	** are acceptable as they only disable an optimization. False negatives,
114017	** on the other hand, can lead to incorrect answers.
114018	*/
114019	static int sqlite3ExprCanReturnSubtype(Parse pParse, Expr pExpr){
114020	Walker w;
114021	memset(&w, 0, sizeof(w));
114022	w.pParse = pParse;
114023	w.xExprCallback = exprNodeCanReturnSubtype;
114024	sqlite3WalkExpr(&w, pExpr);
114025	return w.eCode;
114026	}
114027
114028
114029	/*
114030	** Check to see if pExpr is one of the indexed expressions on pParse->pIdxEpr.
114031	** If it is, then resolve the expression by reading from the index and
114032	** return the register into which the value has been read. If pExpr is
	@@ -113891,10 +114056,21 @@
114056	){
114057	/* Affinity mismatch on a generated column */
114058	continue;
114059	}
114060
114061
114062	/* Functions that might set a subtype should not be replaced by the
114063	** value taken from an expression index if they are themselves an
114064	** argument to another scalar function or aggregate.
114065	** https://sqlite.org/forum/forumpost/68d284c86b082c3e */
114066	if( ExprHasProperty(pExpr, EP_SubtArg)
114067	&& sqlite3ExprCanReturnSubtype(pParse, pExpr)
114068	){
114069	continue;
114070	}
114071
114072	v = pParse->pVdbe;
114073	assert( v!=0 );
114074	if( p->bMaybeNullRow ){
114075	/* If the index is on a NULL row due to an outer join, then we
114076	** cannot extract the value from the index. The value must be
	@@ -115421,35 +115597,41 @@
115597	**
115598	** Additionally, if pExpr is a simple SQL value and the value is the
115599	** same as that currently bound to variable pVar, non-zero is returned.
115600	** Otherwise, if the values are not the same or if pExpr is not a simple
115601	** SQL value, zero is returned.
115602	**
115603	** If the SQLITE_EnableQPSG flag is set on the database connection, then
115604	** this routine always returns false.
115605	*/
115606	static SQLITE_NOINLINE int exprCompareVariable(
115607	const Parse *pParse,
115608	const Expr *pVar,
115609	const Expr *pExpr
115610	){
115611	int res = 2;
115612	int iVar;
115613	sqlite3_value pL, pR = 0;
115614
115615	if( pExpr->op==TK_VARIABLE && pVar->iColumn==pExpr->iColumn ){
115616	return 0;
115617	}
115618	if( (pParse->db->flags & SQLITE_EnableQPSG)!=0 ) return 2;
115619	sqlite3ValueFromExpr(pParse->db, pExpr, SQLITE_UTF8, SQLITE_AFF_BLOB, &pR);
115620	if( pR ){
115621	iVar = pVar->iColumn;
115622	sqlite3VdbeSetVarmask(pParse->pVdbe, iVar);
115623	pL = sqlite3VdbeGetBoundValue(pParse->pReprepare, iVar, SQLITE_AFF_BLOB);
115624	if( pL ){
115625	if( sqlite3_value_type(pL)==SQLITE_TEXT ){
115626	sqlite3_value_text(pL); /* Make sure the encoding is UTF-8 */
115627	}
115628	res = sqlite3MemCompare(pL, pR, 0) ? 2 : 0;
115629	}
115630	sqlite3ValueFree(pR);
115631	sqlite3ValueFree(pL);
115632	}

115633	return res;
115634	}
115635
115636	/*
115637	** Do a deep comparison of two expression trees. Return 0 if the two
	@@ -115471,16 +115653,14 @@
115653	** can be sure the expressions are the same. In the places where
115654	** this routine is used, it does not hurt to get an extra 2 - that
115655	** just might result in some slightly slower code. But returning
115656	** an incorrect 0 or 1 could lead to a malfunction.
115657	**
115658	** If pParse is not NULL and SQLITE_EnableQPSG is off then TK_VARIABLE
115659	** terms in pA with bindings in pParse->pReprepare can be matched against
115660	** literals in pB. The pParse->pVdbe->expmask bitmask is updated for
115661	** each variable referenced.


115662	*/
115663	SQLITE_PRIVATE int sqlite3ExprCompare(
115664	const Parse *pParse,
115665	const Expr *pA,
115666	const Expr *pB,
	@@ -115488,12 +115668,12 @@
115668	){
115669	u32 combinedFlags;
115670	if( pA==0 \|\| pB==0 ){
115671	return pB==pA ? 0 : 2;
115672	}
115673	if( pParse && pA->op==TK_VARIABLE ){
115674	return exprCompareVariable(pParse, pA, pB);
115675	}
115676	combinedFlags = pA->flags \| pB->flags;
115677	if( combinedFlags & EP_IntValue ){
115678	if( (pA->flags&pB->flags&EP_IntValue)!=0 && pA->u.iValue==pB->u.iValue ){
115679	return 0;
	@@ -115683,23 +115863,75 @@
115863	return exprImpliesNotNull(pParse, p->pLeft, pNN, iTab, 1);
115864	}
115865	}
115866	return 0;
115867	}
115868
115869	/*
115870	** Return true if the boolean value of the expression is always either
115871	** FALSE or NULL.
115872	*/
115873	static int sqlite3ExprIsNotTrue(Expr *pExpr){
115874	int v;
115875	if( pExpr->op==TK_NULL ) return 1;
115876	if( pExpr->op==TK_TRUEFALSE && sqlite3ExprTruthValue(pExpr)==0 ) return 1;
115877	v = 1;
115878	if( sqlite3ExprIsInteger(pExpr, &v, 0) && v==0 ) return 1;
115879	return 0;
115880	}
115881
115882	/*
115883	** Return true if the expression is one of the following:
115884	**
115885	** CASE WHEN x THEN y END
115886	** CASE WHEN x THEN y ELSE NULL END
115887	** CASE WHEN x THEN y ELSE false END
115888	** iif(x,y)
115889	** iif(x,y,NULL)
115890	** iif(x,y,false)
115891	*/
115892	static int sqlite3ExprIsIIF(sqlite3 db, const Expr pExpr){
115893	ExprList *pList;
115894	if( pExpr->op==TK_FUNCTION ){
115895	const char *z = pExpr->u.zToken;
115896	FuncDef *pDef;
115897	if( (z[0]!='i' && z[0]!='I') ) return 0;
115898	if( pExpr->x.pList==0 ) return 0;
115899	pDef = sqlite3FindFunction(db, z, pExpr->x.pList->nExpr, ENC(db), 0);
115900	#ifdef SQLITE_ENABLE_UNKNOWN_SQL_FUNCTION
115901	if( pDef==0 ) return 0;
115902	#else
115903	if( NEVER(pDef==0) ) return 0;
115904	#endif
115905	if( (pDef->funcFlags & SQLITE_FUNC_INLINE)==0 ) return 0;
115906	if( SQLITE_PTR_TO_INT(pDef->pUserData)!=INLINEFUNC_iif ) return 0;
115907	}else if( pExpr->op==TK_CASE ){
115908	if( pExpr->pLeft!=0 ) return 0;
115909	}else{
115910	return 0;
115911	}
115912	pList = pExpr->x.pList;
115913	assert( pList!=0 );
115914	if( pList->nExpr==2 ) return 1;
115915	if( pList->nExpr==3 && sqlite3ExprIsNotTrue(pList->a[2].pExpr) ) return 1;
115916	return 0;
115917	}
115918
115919	/*
115920	** Return true if we can prove the pE2 will always be true if pE1 is
115921	** true. Return false if we cannot complete the proof or if pE2 might
115922	** be false. Examples:
115923	**
115924	** pE1: x==5 pE2: x==5 Result: true
115925	** pE1: x>0 pE2: x==5 Result: false
115926	** pE1: x=21 pE2: x=21 OR y=43 Result: true
115927	** pE1: x!=123 pE2: x IS NOT NULL Result: true
115928	** pE1: x!=?1 pE2: x IS NOT NULL Result: true
115929	** pE1: x IS NULL pE2: x IS NOT NULL Result: false
115930	** pE1: x IS ?2 pE2: x IS NOT NULL Result: false
115931	** pE1: iif(x,y) pE2: x Result: true
115932	** PE1: iif(x,y,0) pE2: x Result: true
115933	**
115934	** When comparing TK_COLUMN nodes between pE1 and pE2, if pE2 has
115935	** Expr.iTable<0 then assume a table number given by iTab.
115936	**
115937	** If pParse is not NULL, then the values of bound variables in pE1 are
	@@ -115729,10 +115961,13 @@
115961	if( pE2->op==TK_NOTNULL
115962	&& exprImpliesNotNull(pParse, pE1, pE2->pLeft, iTab, 0)
115963	){
115964	return 1;
115965	}
115966	if( sqlite3ExprIsIIF(pParse->db, pE1) ){
115967	return sqlite3ExprImpliesExpr(pParse,pE1->x.pList->a[0].pExpr,pE2,iTab);
115968	}
115969	return 0;
115970	}
115971
115972	/* This is a helper function to impliesNotNullRow(). In this routine,
115973	** set pWalker->eCode to one only if both of the input expressions
	@@ -120689,12 +120924,12 @@
120924	int nIdxCol = 1; /* Number of columns in stat4 records */
120925
120926	char zIndex; / Index name */
120927	Index pIdx; / Pointer to the index object */
120928	int nSample; /* Number of samples */
120929	i64 nByte; /* Bytes of space required */
120930	i64 i; /* Bytes of space required */
120931	tRowcnt pSpace; / Available allocated memory space */
120932	u8 pPtr; / Available memory as a u8 for easier manipulation */
120933
120934	zIndex = (char *)sqlite3_column_text(pStmt, 0);
120935	if( zIndex==0 ) continue;
	@@ -121140,19 +121375,10 @@
121375	rc = sqlite3Init(db, &zErrDyn);
121376	}
121377	sqlite3BtreeLeaveAll(db);
121378	assert( zErrDyn==0 \|\| rc!=SQLITE_OK );
121379	}









121380	if( rc ){
121381	if( ALWAYS(!REOPEN_AS_MEMDB(db)) ){
121382	int iDb = db->nDb - 1;
121383	assert( iDb>=2 );
121384	if( db->aDb[iDb].pBt ){
	@@ -121646,15 +121872,11 @@
121872	sqlite3 db = pParse->db; / Database handle */
121873	char zDb = db->aDb[iDb].zDbSName; / Schema name of attached database */
121874	int rc; /* Auth callback return code */
121875
121876	if( db->init.busy ) return SQLITE_OK;
121877	rc = db->xAuth(db->pAuthArg, SQLITE_READ, zTab,zCol,zDb,pParse->zAuthContext);




121878	if( rc==SQLITE_DENY ){
121879	char *z = sqlite3_mprintf("%s.%s", zTab, zCol);
121880	if( db->nDb>2 \|\| iDb!=0 ) z = sqlite3_mprintf("%s.%z", zDb, z);
121881	sqlite3ErrorMsg(pParse, "access to %z is prohibited", z);
121882	pParse->rc = SQLITE_AUTH;
	@@ -121757,15 +121979,11 @@
121979	testcase( zArg1==0 );
121980	testcase( zArg2==0 );
121981	testcase( zArg3==0 );
121982	testcase( pParse->zAuthContext==0 );
121983
121984	rc = db->xAuth(db->pAuthArg,code,zArg1,zArg2,zArg3,pParse->zAuthContext);




121985	if( rc==SQLITE_DENY ){
121986	sqlite3ErrorMsg(pParse, "not authorized");
121987	pParse->rc = SQLITE_AUTH;
121988	}else if( rc!=SQLITE_OK && rc!=SQLITE_IGNORE ){
121989	rc = SQLITE_DENY;
	@@ -121994,21 +122212,10 @@
122212	sqlite3VdbeJumpHere(v, addrRewind);
122213	}
122214	}
122215	sqlite3VdbeAddOp0(v, OP_Halt);
122216











122217	/* The cookie mask contains one bit for each database file open.
122218	** (Bit 0 is for main, bit 1 is for temp, and so forth.) Bits are
122219	** set for each database that is used. Generate code to start a
122220	** transaction on each used database and to verify the schema cookie
122221	** on each used database.
	@@ -122133,20 +122340,10 @@
122340	sqlite3DbFree(db, zSql);
122341	memcpy(PARSE_TAIL(pParse), saveBuf, PARSE_TAIL_SZ);
122342	pParse->nested--;
122343	}
122344










122345	/*
122346	** Locate the in-memory structure that describes a particular database
122347	** table given the name of that table and (optionally) the name of the
122348	** database containing the table. Return NULL if not found.
122349	**
	@@ -122161,17 +122358,10 @@
122358	Table *p = 0;
122359	int i;
122360
122361	/* All mutexes are required for schema access. Make sure we hold them. */
122362	assert( zDatabase!=0 \|\| sqlite3BtreeHoldsAllMutexes(db) );







122363	if( zDatabase ){
122364	for(i=0; i<db->nDb; i++){
122365	if( sqlite3StrICmp(zDatabase, db->aDb[i].zDbSName)==0 ) break;
122366	}
122367	if( i>=db->nDb ){
	@@ -125826,13 +126016,10 @@
126016
126017	assert( pTab!=0 );
126018	if( sqlite3StrNICmp(pTab->zName, "sqlite_", 7)==0
126019	&& db->init.busy==0
126020	&& pTblName!=0



126021	){
126022	sqlite3ErrorMsg(pParse, "table %s may not be indexed", pTab->zName);
126023	goto exit_create_index;
126024	}
126025	#ifndef SQLITE_OMIT_VIEW
	@@ -128224,10 +128411,11 @@
128411	** 4) The table is a shadow table, the database connection is in
128412	** defensive mode, and the current sqlite3_prepare()
128413	** is for a top-level SQL statement.
128414	*/
128415	static int vtabIsReadOnly(Parse pParse, Table pTab){
128416	assert( IsVirtual(pTab) );
128417	if( sqlite3GetVTable(pParse->db, pTab)->pMod->pModule->xUpdate==0 ){
128418	return 1;
128419	}
128420
128421	/* Within triggers:
	@@ -131716,11 +131904,17 @@
131904	#ifdef SQLITE_DEBUG
131905	/*
131906	** Implementation of fpdecode(x,y,z) function.
131907	**
131908	** x is a real number that is to be decoded. y is the precision.
131909	** z is the maximum real precision. Return a string that shows the
131910	** results of the sqlite3FpDecode() function.
131911	**
131912	** Used for testing and debugging only, specifically testing and debugging
131913	** of the sqlite3FpDecode() function. This SQL function does not appear
131914	** in production builds. This function is not an API and is subject to
131915	** modification or removal in future versions of SQLite.
131916	*/
131917	static void fpdecodeFunc(
131918	sqlite3_context *context,
131919	int argc,
131920	sqlite3_value **argv
	@@ -131740,10 +131934,86 @@
131934	sqlite3_snprintf(sizeof(zBuf), zBuf, "NaN");
131935	}else{
131936	sqlite3_snprintf(sizeof(zBuf), zBuf, "%c%.*s/%d", s.sign, s.n, s.z, s.iDP);
131937	}
131938	sqlite3_result_text(context, zBuf, -1, SQLITE_TRANSIENT);
131939	}
131940	#endif /* SQLITE_DEBUG */
131941
131942	#ifdef SQLITE_DEBUG
131943	/*
131944	** Implementation of parseuri(uri,flags) function.
131945	**
131946	** Required Arguments:
131947	** "uri" The URI to parse.
131948	** "flags" Bitmask of flags, as if to sqlite3_open_v2().
131949	**
131950	** Additional arguments beyond the first two make calls to
131951	** sqlite3_uri_key() for integers and sqlite3_uri_parameter for
131952	** anything else.
131953	**
131954	** The result is a string showing the results of calling sqlite3ParseUri().
131955	**
131956	** Used for testing and debugging only, specifically testing and debugging
131957	** of the sqlite3ParseUri() function. This SQL function does not appear
131958	** in production builds. This function is not an API and is subject to
131959	** modification or removal in future versions of SQLite.
131960	*/
131961	static void parseuriFunc(
131962	sqlite3_context *ctx,
131963	int argc,
131964	sqlite3_value **argv
131965	){
131966	sqlite3_str *pResult;
131967	const char *zVfs;
131968	const char *zUri;
131969	unsigned int flgs;
131970	int rc;
131971	sqlite3_vfs *pVfs = 0;
131972	char *zFile = 0;
131973	char *zErr = 0;
131974
131975	if( argc<2 ) return;
131976	pVfs = sqlite3_vfs_find(0);
131977	assert( pVfs );
131978	zVfs = pVfs->zName;
131979	zUri = (const char*)sqlite3_value_text(argv[0]);
131980	if( zUri==0 ) return;
131981	flgs = (unsigned int)sqlite3_value_int(argv[1]);
131982	rc = sqlite3ParseUri(zVfs, zUri, &flgs, &pVfs, &zFile, &zErr);
131983	pResult = sqlite3_str_new(0);
131984	if( pResult ){
131985	int i;
131986	sqlite3_str_appendf(pResult, "rc=%d", rc);
131987	sqlite3_str_appendf(pResult, ", flags=0x%x", flgs);
131988	sqlite3_str_appendf(pResult, ", vfs=%Q", pVfs ? pVfs->zName: 0);
131989	sqlite3_str_appendf(pResult, ", err=%Q", zErr);
131990	sqlite3_str_appendf(pResult, ", file=%Q", zFile);
131991	if( zFile ){
131992	const char *z = zFile;
131993	z += sqlite3Strlen30(z)+1;
131994	while( z[0] ){
131995	sqlite3_str_appendf(pResult, ", %Q", z);
131996	z += sqlite3Strlen30(z)+1;
131997	}
131998	for(i=2; i<argc; i++){
131999	const char *zArg;
132000	if( sqlite3_value_type(argv[i])==SQLITE_INTEGER ){
132001	int k = sqlite3_value_int(argv[i]);
132002	sqlite3_str_appendf(pResult, ", '%d:%q'",k,sqlite3_uri_key(zFile, k));
132003	}else if( (zArg = (const char*)sqlite3_value_text(argv[i]))!=0 ){
132004	sqlite3_str_appendf(pResult, ", '%q:%q'",
132005	zArg, sqlite3_uri_parameter(zFile,zArg));
132006	}else{
132007	sqlite3_str_appendf(pResult, ", NULL");
132008	}
132009	}
132010	}
132011	sqlite3_result_text(ctx, sqlite3_str_finish(pResult), -1, sqlite3_free);
132012	}
132013	sqlite3_free_filename(zFile);
132014	sqlite3_free(zErr);
132015	}
132016	#endif /* SQLITE_DEBUG */
132017
132018	/*
132019	** All of the FuncDef structures in the aBuiltinFunc[] array above
	@@ -131777,13 +132047,10 @@
132047	#endif
132048	#ifndef SQLITE_OMIT_LOAD_EXTENSION
132049	SFUNCTION(load_extension, 1, 0, 0, loadExt ),
132050	SFUNCTION(load_extension, 2, 0, 0, loadExt ),
132051	#endif



132052	#ifndef SQLITE_OMIT_COMPILEOPTION_DIAGS
132053	DFUNCTION(sqlite_compileoption_used,1, 0, 0, compileoptionusedFunc ),
132054	DFUNCTION(sqlite_compileoption_get, 1, 0, 0, compileoptiongetFunc ),
132055	#endif /* SQLITE_OMIT_COMPILEOPTION_DIAGS */
132056	INLINE_FUNC(unlikely, 1, INLINEFUNC_unlikely, SQLITE_FUNC_UNLIKELY),
	@@ -131805,11 +132072,12 @@
132072	FUNCTION(max, -1, 1, 1, minmaxFunc ),
132073	FUNCTION(max, 0, 1, 1, 0 ),
132074	WAGGREGATE(max, 1, 1, 1, minmaxStep, minMaxFinalize, minMaxValue, 0,
132075	SQLITE_FUNC_MINMAX\|SQLITE_FUNC_ANYORDER ),
132076	FUNCTION2(typeof, 1, 0, 0, typeofFunc, SQLITE_FUNC_TYPEOF),
132077	FUNCTION2(subtype, 1, 0, 0, subtypeFunc,
132078	SQLITE_FUNC_TYPEOF\|SQLITE_SUBTYPE),
132079	FUNCTION2(length, 1, 0, 0, lengthFunc, SQLITE_FUNC_LENGTH),
132080	FUNCTION2(octet_length, 1, 0, 0, bytelengthFunc,SQLITE_FUNC_BYTELEN),
132081	FUNCTION(instr, 2, 0, 0, instrFunc ),
132082	FUNCTION(printf, -1, 0, 0, printfFunc ),
132083	FUNCTION(format, -1, 0, 0, printfFunc ),
	@@ -131816,10 +132084,11 @@
132084	FUNCTION(unicode, 1, 0, 0, unicodeFunc ),
132085	FUNCTION(char, -1, 0, 0, charFunc ),
132086	FUNCTION(abs, 1, 0, 0, absFunc ),
132087	#ifdef SQLITE_DEBUG
132088	FUNCTION(fpdecode, 3, 0, 0, fpdecodeFunc ),
132089	FUNCTION(parseuri, -1, 0, 0, parseuriFunc ),
132090	#endif
132091	#ifndef SQLITE_OMIT_FLOATING_POINT
132092	FUNCTION(round, 1, 0, 0, roundFunc ),
132093	FUNCTION(round, 2, 0, 0, roundFunc ),
132094	#endif
	@@ -131910,15 +132179,18 @@
132179	MFUNCTION(atanh, 1, atanh, math1Func ),
132180	#endif
132181	MFUNCTION(sqrt, 1, sqrt, math1Func ),
132182	MFUNCTION(radians, 1, degToRad, math1Func ),
132183	MFUNCTION(degrees, 1, radToDeg, math1Func ),
132184	MFUNCTION(pi, 0, 0, piFunc ),
132185	#endif /* SQLITE_ENABLE_MATH_FUNCTIONS */
132186	FUNCTION(sign, 1, 0, 0, signFunc ),
132187	INLINE_FUNC(coalesce, -1, INLINEFUNC_coalesce, 0 ),
132188	INLINE_FUNC(iif, 2, INLINEFUNC_iif, 0 ),
132189	INLINE_FUNC(iif, 3, INLINEFUNC_iif, 0 ),
132190	INLINE_FUNC(if, 2, INLINEFUNC_iif, 0 ),
132191	INLINE_FUNC(if, 3, INLINEFUNC_iif, 0 ),
132192	};
132193	#ifndef SQLITE_OMIT_ALTERTABLE
132194	sqlite3AlterFunctions();
132195	#endif
132196	sqlite3WindowFunctions();
	@@ -140427,16 +140699,10 @@
140699	if( db->autoCommit==0 ){
140700	/* Foreign key support may not be enabled or disabled while not
140701	** in auto-commit mode. */
140702	mask &= ~(SQLITE_ForeignKeys);
140703	}






140704
140705	if( sqlite3GetBoolean(zRight, 0) ){
140706	if( (mask & SQLITE_WriteSchema)==0
140707	\|\| (db->flags & SQLITE_Defensive)==0
140708	){
	@@ -140568,11 +140834,12 @@
140834	pTab = sqliteHashData(k);
140835	if( pTab->nCol==0 ){
140836	char zSql = sqlite3MPrintf(db, "SELECTFROM\"%w\"", pTab->zName);
140837	if( zSql ){
140838	sqlite3_stmt *pDummy = 0;
140839	(void)sqlite3_prepare_v3(db, zSql, -1, SQLITE_PREPARE_DONT_LOG,
140840	&pDummy, 0);
140841	(void)sqlite3_finalize(pDummy);
140842	sqlite3DbFree(db, zSql);
140843	}
140844	if( db->mallocFailed ){
140845	sqlite3ErrorMsg(db->pParse, "out of memory");
	@@ -141044,10 +141311,11 @@
141311	}
141312	aRoot[0] = cnt;
141313
141314	/* Make sure sufficient number of registers have been allocated */
141315	sqlite3TouchRegister(pParse, 8+cnt);
141316	sqlite3VdbeAddOp3(v, OP_Null, 0, 8, 8+cnt);
141317	sqlite3ClearTempRegCache(pParse);
141318
141319	/* Do the b-tree integrity checks */
141320	sqlite3VdbeAddOp4(v, OP_IntegrityCk, 1, cnt, 8, (char*)aRoot,P4_INTARRAY);
141321	sqlite3VdbeChangeP5(v, (u8)i);
	@@ -142668,18 +142936,11 @@
142936	encoding = (u8)meta[BTREE_TEXT_ENCODING-1] & 3;
142937	if( encoding==0 ) encoding = SQLITE_UTF8;
142938	#else
142939	encoding = SQLITE_UTF8;
142940	#endif
142941	sqlite3SetTextEncoding(db, encoding);







142942	}else{
142943	/* If opening an attached database, the encoding much match ENC(db) */
142944	if( (meta[BTREE_TEXT_ENCODING-1] & 3)!=ENC(db) ){
142945	sqlite3SetString(pzErrMsg, db, "attached databases must use the same"
142946	" text encoding as main database");
	@@ -143369,16 +143630,28 @@
143630	#endif
143631	*ppStmt = 0;
143632	if( !sqlite3SafetyCheckOk(db)\|\|zSql==0 ){
143633	return SQLITE_MISUSE_BKPT;
143634	}
143635
143636	/* Make sure nBytes is non-negative and correct. It should be the
143637	** number of bytes until the end of the input buffer or until the first
143638	** U+0000 character. If the input nBytes is odd, convert it into
143639	** an even number. If the input nBytes is negative, then the input
143640	** must be terminated by at least one U+0000 character */
143641	if( nBytes>=0 ){
143642	int sz;
143643	const char z = (const char)zSql;
143644	for(sz=0; sz<nBytes && (z[sz]!=0 \|\| z[sz+1]!=0); sz += 2){}
143645	nBytes = sz;
143646	}else{
143647	int sz;
143648	const char z = (const char)zSql;
143649	for(sz=0; z[sz]!=0 \|\| z[sz+1]!=0; sz += 2){}
143650	nBytes = sz;
143651	}
143652
143653	sqlite3_mutex_enter(db->mutex);
143654	zSql8 = sqlite3Utf16to8(db, zSql, nBytes, SQLITE_UTF16NATIVE);
143655	if( zSql8 ){
143656	rc = sqlite3LockAndPrepare(db, zSql8, -1, prepFlags, 0, ppStmt, &zTail8);
143657	}
	@@ -143388,11 +143661,11 @@
143661	** equivalent pointer into the UTF-16 string by counting the unicode
143662	** characters between zSql8 and zTail8, and then returning a pointer
143663	** the same number of characters into the UTF-16 string.
143664	*/
143665	int chars_parsed = sqlite3Utf8CharLen(zSql8, (int)(zTail8-zSql8));
143666	pzTail = (u8 )zSql + sqlite3Utf16ByteLen(zSql, nBytes, chars_parsed);
143667	}
143668	sqlite3DbFree(db, zSql8);
143669	rc = sqlite3ApiExit(db, rc);
143670	sqlite3_mutex_leave(db->mutex);
143671	return rc;
	@@ -147361,36 +147634,36 @@
147634	return pExpr;
147635	}
147636	if( pSubst->isOuterJoin ){
147637	ExprSetProperty(pNew, EP_CanBeNull);
147638	}
147639	if( pNew->op==TK_TRUEFALSE ){
147640	pNew->u.iValue = sqlite3ExprTruthValue(pNew);
147641	pNew->op = TK_INTEGER;
147642	ExprSetProperty(pNew, EP_IntValue);
147643	}
147644
147645	/* Ensure that the expression now has an implicit collation sequence,
147646	** just as it did when it was a column of a view or sub-query. */
147647	{
147648	CollSeq *pNat = sqlite3ExprCollSeq(pSubst->pParse, pNew);
147649	CollSeq *pColl = sqlite3ExprCollSeq(pSubst->pParse,
147650	pSubst->pCList->a[iColumn].pExpr
147651	);
147652	if( pNat!=pColl \|\| (pNew->op!=TK_COLUMN && pNew->op!=TK_COLLATE) ){
147653	pNew = sqlite3ExprAddCollateString(pSubst->pParse, pNew,
147654	(pColl ? pColl->zName : "BINARY")
147655	);
147656	}
147657	}
147658	ExprClearProperty(pNew, EP_Collate);
147659	if( ExprHasProperty(pExpr,EP_OuterON\|EP_InnerON) ){
147660	sqlite3SetJoinExpr(pNew, pExpr->w.iJoin,
147661	pExpr->flags & (EP_OuterON\|EP_InnerON));
147662	}
147663	sqlite3ExprDelete(db, pExpr);
147664	pExpr = pNew;




















147665	}
147666	}
147667	}else{
147668	if( pExpr->op==TK_IF_NULL_ROW && pExpr->iTable==pSubst->iTable ){
147669	pExpr->iTable = pSubst->iNewTable;
	@@ -148123,20 +148396,20 @@
148396	}
148397
148398	/* Transfer the FROM clause terms from the subquery into the
148399	** outer query.
148400	*/
148401	iNewParent = pSubSrc->a[0].iCursor;
148402	for(i=0; i<nSubSrc; i++){
148403	SrcItem *pItem = &pSrc->a[i+iFrom];
148404	assert( pItem->fg.isTabFunc==0 );
148405	assert( pItem->fg.isSubquery
148406	\|\| pItem->fg.fixedSchema
148407	\|\| pItem->u4.zDatabase==0 );
148408	if( pItem->fg.isUsing ) sqlite3IdListDelete(db, pItem->u3.pUsing);
148409	*pItem = pSubSrc->a[i];
148410	pItem->fg.jointype \|= ltorj;

148411	memset(&pSubSrc->a[i], 0, sizeof(pSubSrc->a[i]));
148412	}
148413	pSrc->a[iFrom].fg.jointype &= JT_LTORJ;
148414	pSrc->a[iFrom].fg.jointype \|= jointype \| ltorj;
148415
	@@ -148172,10 +148445,11 @@
148445	pSub->pOrderBy = 0;
148446	}
148447	pWhere = pSub->pWhere;
148448	pSub->pWhere = 0;
148449	if( isOuterJoin>0 ){
148450	assert( pSubSrc->nSrc==1 );
148451	sqlite3SetJoinExpr(pWhere, iNewParent, EP_OuterON);
148452	}
148453	if( pWhere ){
148454	if( pParent->pWhere ){
148455	pParent->pWhere = sqlite3PExpr(pParse, TK_AND, pWhere, pParent->pWhere);
	@@ -151271,11 +151545,11 @@
151545	sqlite3TreeViewSelect(0, p, 0);
151546	}
151547	#endif
151548	assert( pSubq->pSelect && (pSub->selFlags & SF_PushDown)!=0 );
151549	}else{
151550	TREETRACE(0x4000,pParse,p,("WHERE-clause push-down not possible\n"));
151551	}
151552
151553	/* Convert unused result columns of the subquery into simple NULL
151554	** expressions, to avoid unneeded searching and computation.
151555	** tag-select-0440
	@@ -156980,10 +157254,11 @@
157254	assert( sParse.zErrMsg==0 );
157255	if( !pTab->aCol ){
157256	Table *pNew = sParse.pNewTable;
157257	Index *pIdx;
157258	pTab->aCol = pNew->aCol;
157259	assert( IsOrdinaryTable(pNew) );
157260	sqlite3ExprListDelete(db, pNew->u.tab.pDfltList);
157261	pTab->nNVCol = pTab->nCol = pNew->nCol;
157262	pTab->tabFlags \|= pNew->tabFlags & (TF_WithoutRowid\|TF_NoVisibleRowid);
157263	pNew->nCol = 0;
157264	pNew->aCol = 0;
	@@ -158044,13 +158319,21 @@
158319	SQLITE_PRIVATE int sqlite3WhereExplainBloomFilter(
158320	const Parse pParse, / Parse context */
158321	const WhereInfo pWInfo, / WHERE clause */
158322	const WhereLevel pLevel / Bloom filter on this level */
158323	);
158324	SQLITE_PRIVATE void sqlite3WhereAddExplainText(
158325	Parse pParse, / Parse context */
158326	int addr,
158327	SrcList pTabList, / Table list this loop refers to */
158328	WhereLevel pLevel, / Scan to write OP_Explain opcode for */
158329	u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
158330	);
158331	#else
158332	# define sqlite3WhereExplainOneScan(u,v,w,x) 0
158333	# define sqlite3WhereExplainBloomFilter(u,v,w) 0
158334	# define sqlite3WhereAddExplainText(u,v,w,x,y)
158335	#endif /* SQLITE_OMIT_EXPLAIN */
158336	#ifdef SQLITE_ENABLE_STMT_SCANSTATUS
158337	SQLITE_PRIVATE void sqlite3WhereAddScanStatus(
158338	Vdbe v, / Vdbe to add scanstatus entry to */
158339	SrcList pSrclist, / FROM clause pLvl reads data from */
	@@ -158248,42 +158531,42 @@
158531	}
158532	sqlite3_str_append(pStr, ")", 1);
158533	}
158534
158535	/*
158536	** This function sets the P4 value of an existing OP_Explain opcode to
158537	** text describing the loop in pLevel. If the OP_Explain opcode already has
158538	** a P4 value, it is freed before it is overwritten.




158539	*/
158540	SQLITE_PRIVATE void sqlite3WhereAddExplainText(
158541	Parse pParse, / Parse context */
158542	int addr, /* Address of OP_Explain opcode */
158543	SrcList pTabList, / Table list this loop refers to */
158544	WhereLevel pLevel, / Scan to write OP_Explain opcode for */
158545	u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
158546	){

158547	#if !defined(SQLITE_DEBUG)
158548	if( sqlite3ParseToplevel(pParse)->explain==2 \|\| IS_STMT_SCANSTATUS(pParse->db) )
158549	#endif
158550	{
158551	VdbeOp *pOp = sqlite3VdbeGetOp(pParse->pVdbe, addr);
158552
158553	SrcItem *pItem = &pTabList->a[pLevel->iFrom];

158554	sqlite3 db = pParse->db; / Database handle */
158555	int isSearch; /* True for a SEARCH. False for SCAN. */
158556	WhereLoop pLoop; / The controlling WhereLoop object */
158557	u32 flags; /* Flags that describe this loop */
158558	#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_EXPLAIN)
158559	char zMsg; / Text to add to EQP output */
158560	#endif
158561	StrAccum str; /* EQP output string */
158562	char zBuf[100]; /* Initial space for EQP output string */
158563
158564	if( db->mallocFailed ) return;
158565
158566	pLoop = pLevel->pWLoop;
158567	flags = pLoop->wsFlags;

158568
158569	isSearch = (flags&(WHERE_BTM_LIMIT\|WHERE_TOP_LIMIT))!=0
158570	\|\| ((flags&WHERE_VIRTUALTABLE)==0 && (pLoop->u.btree.nEq>0))
158571	\|\| (wctrlFlags&(WHERE_ORDERBY_MIN\|WHERE_ORDERBY_MAX));
158572
	@@ -158303,11 +158586,11 @@
158586	}
158587	}else if( flags & WHERE_PARTIALIDX ){
158588	zFmt = "AUTOMATIC PARTIAL COVERING INDEX";
158589	}else if( flags & WHERE_AUTO_INDEX ){
158590	zFmt = "AUTOMATIC COVERING INDEX";
158591	}else if( flags & (WHERE_IDX_ONLY\|WHERE_EXPRIDX) ){
158592	zFmt = "COVERING INDEX %s";
158593	}else{
158594	zFmt = "INDEX %s";
158595	}
158596	if( zFmt ){
	@@ -158355,15 +158638,54 @@
158638	sqlite3LogEstToInt(pLoop->nOut));
158639	}else{
158640	sqlite3_str_append(&str, " (~1 row)", 9);
158641	}
158642	#endif
158643	#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_EXPLAIN)
158644	zMsg = sqlite3StrAccumFinish(&str);
158645	sqlite3ExplainBreakpoint("",zMsg);
158646	#endif
158647
158648	assert( pOp->opcode==OP_Explain );
158649	assert( pOp->p4type==P4_DYNAMIC \|\| pOp->p4.z==0 );
158650	sqlite3DbFree(db, pOp->p4.z);
158651	pOp->p4type = P4_DYNAMIC;
158652	pOp->p4.z = sqlite3StrAccumFinish(&str);
158653	}
158654	}
158655
158656
158657	/*
158658	** This function is a no-op unless currently processing an EXPLAIN QUERY PLAN
158659	** command, or if stmt_scanstatus_v2() stats are enabled, or if SQLITE_DEBUG
158660	** was defined at compile-time. If it is not a no-op, a single OP_Explain
158661	** opcode is added to the output to describe the table scan strategy in pLevel.
158662	**
158663	** If an OP_Explain opcode is added to the VM, its address is returned.
158664	** Otherwise, if no OP_Explain is coded, zero is returned.
158665	*/
158666	SQLITE_PRIVATE int sqlite3WhereExplainOneScan(
158667	Parse pParse, / Parse context */
158668	SrcList pTabList, / Table list this loop refers to */
158669	WhereLevel pLevel, / Scan to write OP_Explain opcode for */
158670	u16 wctrlFlags /* Flags passed to sqlite3WhereBegin() */
158671	){
158672	int ret = 0;
158673	#if !defined(SQLITE_DEBUG)
158674	if( sqlite3ParseToplevel(pParse)->explain==2 \|\| IS_STMT_SCANSTATUS(pParse->db) )
158675	#endif
158676	{
158677	if( (pLevel->pWLoop->wsFlags & WHERE_MULTI_OR)==0
158678	&& (wctrlFlags & WHERE_OR_SUBCLAUSE)==0
158679	){
158680	Vdbe *v = pParse->pVdbe;
158681	int addr = sqlite3VdbeCurrentAddr(v);
158682	ret = sqlite3VdbeAddOp3(
158683	v, OP_Explain, addr, pParse->addrExplain, pLevel->pWLoop->rRun
158684	);
158685	sqlite3WhereAddExplainText(pParse, addr, pTabList, pLevel, wctrlFlags);
158686	}
158687	}
158688	return ret;
158689	}
158690
158691	/*
	@@ -158458,13 +158780,14 @@
158780	if( wsFlags & WHERE_INDEXED ){
158781	sqlite3VdbeScanStatusRange(v, addrExplain, -1, pLvl->iIdxCur);
158782	}
158783	}else{
158784	int addr;
158785	VdbeOp *pOp;
158786	assert( pSrclist->a[pLvl->iFrom].fg.isSubquery );
158787	addr = pSrclist->a[pLvl->iFrom].u4.pSubq->addrFillSub;
158788	pOp = sqlite3VdbeGetOp(v, addr-1);
158789	assert( sqlite3VdbeDb(v)->mallocFailed \|\| pOp->opcode==OP_InitCoroutine );
158790	assert( sqlite3VdbeDb(v)->mallocFailed \|\| pOp->p2>addr );
158791	sqlite3VdbeScanStatusRange(v, addrExplain, addr, pOp->p2-1);
158792	}
158793	}
	@@ -158713,10 +159036,11 @@
159036	if( pOrigLhs ){
159037	sqlite3ExprListDelete(db, pOrigLhs);
159038	pNew->pLeft->x.pList = pLhs;
159039	}
159040	pSelect->pEList = pRhs;
159041	pSelect->selId = ++pParse->nSelect; /* Req'd for SubrtnSig validity */
159042	if( pLhs && pLhs->nExpr==1 ){
159043	/* Take care here not to generate a TK_VECTOR containing only a
159044	** single value. Since the parser never creates such a vector, some
159045	** of the subroutines do not handle this case. */
159046	Expr *p = pLhs->a[0].pExpr;
	@@ -161260,24 +161584,29 @@
161584	}else if( op==TK_STRING ){
161585	assert( !ExprHasProperty(pRight, EP_IntValue) );
161586	z = (u8*)pRight->u.zToken;
161587	}
161588	if( z ){
161589	/* Count the number of prefix bytes prior to the first wildcard.
161590	** or U+fffd character. If the underlying database has a UTF16LE
161591	** encoding, then only consider ASCII characters. Note that the
161592	** encoding of z[] is UTF8 - we are dealing with only UTF8 here in
161593	** this code, but the database engine itself might be processing
161594	** content using a different encoding. */
161595	cnt = 0;
161596	while( (c=z[cnt])!=0 && c!=wc[0] && c!=wc[1] && c!=wc[2] ){
161597	cnt++;
161598	if( c==wc[3] && z[cnt]>0 && z[cnt]<0x80 ){
161599	cnt++;
161600	}else if( c>=0x80 ){
161601	const u8 *z2 = z+cnt-1;
161602	if( sqlite3Utf8Read(&z2)==0xfffd \|\| ENC(db)==SQLITE_UTF16LE ){
161603	cnt--;
161604	break;
161605	}else{
161606	cnt = (int)(z2-z);
161607	}
161608	}
161609	}
161610
161611	/* The optimization is possible only if (1) the pattern does not begin
161612	** with a wildcard and if (2) the non-wildcard prefix does not end with
	@@ -161285,11 +161614,11 @@
161614	** a single escape character. The second condition is necessary so
161615	** that we can increment the prefix key to find an upper bound for the
161616	** range search. The third is because the caller assumes that the pattern
161617	** consists of at least one character after all escapes have been
161618	** removed. */
161619	if( (cnt>1 \|\| (cnt>0 && z[0]!=wc[3])) && ALWAYS(255!=(u8)z[cnt-1]) ){
161620	Expr *pPrefix;
161621
161622	/* A "complete" match if the pattern ends with "" or "%" /
161623	*pisComplete = c==wc[0] && z[cnt+1]==0 && ENC(db)!=SQLITE_UTF16LE;
161624
	@@ -163780,11 +164109,11 @@
164109	\|\| pTerm->pExpr->w.iJoin != pSrc->iCursor
164110	){
164111	return 0;
164112	}
164113	if( (pSrc->fg.jointype & (JT_LEFT\|JT_RIGHT))!=0
164114	&& NEVER(ExprHasProperty(pTerm->pExpr, EP_InnerON))
164115	){
164116	return 0;
164117	}
164118	return 1;
164119	}
	@@ -164577,13 +164906,15 @@
164906	** Whether or not an error is returned, it is the responsibility of the
164907	** caller to eventually free p->idxStr if p->needToFreeIdxStr indicates
164908	** that this is required.
164909	*/
164910	static int vtabBestIndex(Parse pParse, Table pTab, sqlite3_index_info *p){

164911	int rc;
164912	sqlite3_vtab *pVtab;
164913
164914	assert( IsVirtual(pTab) );
164915	pVtab = sqlite3GetVTable(pParse->db, pTab)->pVtab;
164916	whereTraceIndexInfoInputs(p, pTab);
164917	pParse->db->nSchemaLock++;
164918	rc = pVtab->pModule->xBestIndex(pVtab, p);
164919	pParse->db->nSchemaLock--;
164920	whereTraceIndexInfoOutputs(p, pTab);
	@@ -165271,11 +165602,11 @@
165602	return rc;
165603	}
165604	#endif /* SQLITE_ENABLE_STAT4 */
165605
165606
165607	#if defined(WHERETRACE_ENABLED) \|\| defined(SQLITE_DEBUG)
165608	/*
165609	** Print the content of a WhereTerm object
165610	*/
165611	SQLITE_PRIVATE void sqlite3WhereTermPrint(WhereTerm *pTerm, int iTerm){
165612	if( pTerm==0 ){
	@@ -165314,10 +165645,13 @@
165645	sqlite3DebugPrintf(" iParent=%d", pTerm->iParent);
165646	}
165647	sqlite3DebugPrintf("\n");
165648	sqlite3TreeViewExpr(0, pTerm->pExpr, 0);
165649	}
165650	}
165651	SQLITE_PRIVATE void sqlite3ShowWhereTerm(WhereTerm *pTerm){
165652	sqlite3WhereTermPrint(pTerm, 0);
165653	}
165654	#endif
165655
165656	#ifdef WHERETRACE_ENABLED
165657	/*
	@@ -165522,11 +165856,11 @@
165856	** Return TRUE if X is a proper subset of Y but is of equal or less cost.
165857	** In other words, return true if all constraints of X are also part of Y
165858	** and Y has additional constraints that might speed the search that X lacks
165859	** but the cost of running X is not more than the cost of running Y.
165860	**
165861	** In other words, return true if the cost relationship between X and Y
165862	** is inverted and needs to be adjusted.
165863	**
165864	** Case 1:
165865	**
165866	** (1a) X and Y use the same index.
	@@ -166500,11 +166834,10 @@
166834	pParse = pWC->pWInfo->pParse;
166835	while( pWhere->op==TK_AND ){
166836	if( !whereUsablePartialIndex(iTab,jointype,pWC,pWhere->pLeft) ) return 0;
166837	pWhere = pWhere->pRight;
166838	}

166839	for(i=0, pTerm=pWC->a; i<pWC->nTerm; i++, pTerm++){
166840	Expr *pExpr;
166841	pExpr = pTerm->pExpr;
166842	if( (!ExprHasProperty(pExpr, EP_OuterON) \|\| pExpr->w.iJoin==iTab)
166843	&& ((jointype & JT_OUTER)==0 \|\| ExprHasProperty(pExpr, EP_OuterON))
	@@ -169140,10 +169473,11 @@
169473	hasRightJoin = (pWInfo->pTabList->a[0].fg.jointype & JT_LTORJ)!=0;
169474	for(i=pWInfo->nLevel-1; i>=1; i--){
169475	WhereTerm pTerm, pEnd;
169476	SrcItem *pItem;
169477	WhereLoop *pLoop;
169478	Bitmask m1;
169479	pLoop = pWInfo->a[i].pWLoop;
169480	pItem = &pWInfo->pTabList->a[pLoop->iTab];
169481	if( (pItem->fg.jointype & (JT_LEFT\|JT_RIGHT))!=JT_LEFT ) continue;
169482	if( (pWInfo->wctrlFlags & WHERE_WANT_DISTINCT)==0
169483	&& (pLoop->wsFlags & WHERE_ONEROW)==0
	@@ -169160,17 +169494,20 @@
169494	break;
169495	}
169496	}
169497	if( hasRightJoin
169498	&& ExprHasProperty(pTerm->pExpr, EP_InnerON)
169499	&& NEVER(pTerm->pExpr->w.iJoin==pItem->iCursor)
169500	){
169501	break; /* restriction (5) */
169502	}
169503	}
169504	if( pTerm<pEnd ) continue;
169505	WHERETRACE(0xffffffff,("-> omit unused FROM-clause term %c\n",pLoop->cId));
169506	m1 = MASKBIT(i)-1;
169507	testcase( ((pWInfo->revMask>>1) & ~m1)!=0 );
169508	pWInfo->revMask = (m1 & pWInfo->revMask) \| ((pWInfo->revMask>>1) & ~m1);
169509	notReady &= ~pLoop->maskSelf;
169510	for(pTerm=pWInfo->sWC.a; pTerm<pEnd; pTerm++){
169511	if( (pTerm->prereqAll & pLoop->maskSelf)!=0 ){
169512	pTerm->wtFlags \|= TERM_CODED;
169513	}
	@@ -169237,62 +169574,10 @@
169574	nSearch += pLoop->nOut;
169575	if( pWInfo->nOutStarDelta ) nSearch += pLoop->rStarDelta;
169576	}
169577	}
169578




















































169579	/*
169580	** The index pIdx is used by a query and contains one or more expressions.
169581	** In other words pIdx is an index on an expression. iIdxCur is the cursor
169582	** number for the index and iDataCur is the cursor number for the corresponding
169583	** table.
	@@ -169322,16 +169607,10 @@
169607	pExpr = sqlite3ColumnExpr(pTab, &pTab->aCol[j]);
169608	}else{
169609	continue;
169610	}
169611	if( sqlite3ExprIsConstant(0,pExpr) ) continue;






169612	p = sqlite3DbMallocRaw(pParse->db, sizeof(IndexedExpr));
169613	if( p==0 ) break;
169614	p->pIENext = pParse->pIdxEpr;
169615	#ifdef WHERETRACE_ENABLED
169616	if( sqlite3WhereTrace & 0x200 ){
	@@ -170434,18 +170713,32 @@
170713	x = sqlite3TableColumnToIndex(pIdx, x);
170714	if( x>=0 ){
170715	pOp->p2 = x;
170716	pOp->p1 = pLevel->iIdxCur;
170717	OpcodeRewriteTrace(db, k, pOp);
170718	}else if( pLoop->wsFlags & (WHERE_IDX_ONLY\|WHERE_EXPRIDX) ){




170719	if( pLoop->wsFlags & WHERE_IDX_ONLY ){
170720	/* An error. pLoop is supposed to be a covering index loop,
170721	** and yet the VM code refers to a column of the table that
170722	** is not part of the index. */
170723	sqlite3ErrorMsg(pParse, "internal query planner error");
170724	pParse->rc = SQLITE_INTERNAL;
170725	}else{
170726	/* The WHERE_EXPRIDX flag is set by the planner when it is likely
170727	** that pLoop is a covering index loop, but it is not possible
170728	** to be 100% sure. In this case, any OP_Explain opcode
170729	** corresponding to this loop describes the index as a "COVERING
170730	** INDEX". But, pOp proves that pLoop is not actually a covering
170731	** index loop. So clear the WHERE_EXPRIDX flag and rewrite the
170732	** text that accompanies the OP_Explain opcode, if any. */
170733	pLoop->wsFlags &= ~WHERE_EXPRIDX;
170734	sqlite3WhereAddExplainText(pParse,
170735	pLevel->addrBody-1,
170736	pTabList,
170737	pLevel,
170738	pWInfo->wctrlFlags
170739	);
170740	}
170741	}
170742	}else if( pOp->opcode==OP_Rowid ){
170743	pOp->p1 = pLevel->iIdxCur;
170744	pOp->opcode = OP_IdxRowid;
	@@ -172149,10 +172442,11 @@
172442	for(pWin=pMWin; pWin; pWin=pWin->pNextWin){
172443	FuncDef *pFunc = pWin->pWFunc;
172444	int regArg;
172445	int nArg = pWin->bExprArgs ? 0 : windowArgCount(pWin);
172446	int i;
172447	int addrIf = 0;
172448
172449	assert( bInverse==0 \|\| pWin->eStart!=TK_UNBOUNDED );
172450
172451	/* All OVER clauses in the same window function aggregate step must
172452	** be the same. */
	@@ -172164,10 +172458,22 @@
172458	}else{
172459	sqlite3VdbeAddOp3(v, OP_Column, pMWin->iEphCsr, pWin->iArgCol+i, reg+i);
172460	}
172461	}
172462	regArg = reg;
172463
172464	if( pWin->pFilter ){
172465	int regTmp;
172466	assert( ExprUseXList(pWin->pOwner) );
172467	assert( pWin->bExprArgs \|\| !nArg \|\|nArg==pWin->pOwner->x.pList->nExpr );
172468	assert( pWin->bExprArgs \|\| nArg \|\|pWin->pOwner->x.pList==0 );
172469	regTmp = sqlite3GetTempReg(pParse);
172470	sqlite3VdbeAddOp3(v, OP_Column, csr, pWin->iArgCol+nArg,regTmp);
172471	addrIf = sqlite3VdbeAddOp3(v, OP_IfNot, regTmp, 0, 1);
172472	VdbeCoverage(v);
172473	sqlite3ReleaseTempReg(pParse, regTmp);
172474	}
172475
172476	if( pMWin->regStartRowid==0
172477	&& (pFunc->funcFlags & SQLITE_FUNC_MINMAX)
172478	&& (pWin->eStart!=TK_UNBOUNDED)
172479	){
	@@ -172184,29 +172490,17 @@
172490	sqlite3VdbeAddOp1(v, OP_Delete, pWin->csrApp);
172491	sqlite3VdbeJumpHere(v, sqlite3VdbeCurrentAddr(v)-2);
172492	}
172493	sqlite3VdbeJumpHere(v, addrIsNull);
172494	}else if( pWin->regApp ){
172495	assert( pWin->pFilter==0 );
172496	assert( pFunc->zName==nth_valueName
172497	\|\| pFunc->zName==first_valueName
172498	);
172499	assert( bInverse==0 \|\| bInverse==1 );
172500	sqlite3VdbeAddOp2(v, OP_AddImm, pWin->regApp+1-bInverse, 1);
172501	}else if( pFunc->xSFunc!=noopStepFunc ){













172502	if( pWin->bExprArgs ){
172503	int iOp = sqlite3VdbeCurrentAddr(v);
172504	int iEnd;
172505
172506	assert( ExprUseXList(pWin->pOwner) );
	@@ -172233,12 +172527,13 @@
172527	sqlite3VdbeAppendP4(v, pFunc, P4_FUNCDEF);
172528	sqlite3VdbeChangeP5(v, (u8)nArg);
172529	if( pWin->bExprArgs ){
172530	sqlite3ReleaseTempRange(pParse, regArg, nArg);
172531	}

172532	}
172533
172534	if( addrIf ) sqlite3VdbeJumpHere(v, addrIf);
172535	}
172536	}
172537
172538	/*
172539	** Values that may be passed as the second argument to windowCodeOp().
	@@ -173660,10 +173955,17 @@
173955	** Then the "b" IdList records the list "a,b,c".
173956	*/
173957	struct TrigEvent { int a; IdList * b; };
173958
173959	struct FrameBound { int eType; Expr *pExpr; };
173960
173961	/*
173962	** Generate a syntax error
173963	*/
173964	static void parserSyntaxError(Parse pParse, Token p){
173965	sqlite3ErrorMsg(pParse, "near \"%T\": syntax error", p);
173966	}
173967
173968	/*
173969	** Disable lookaside memory allocation for objects that might be
173970	** shared across database connections.
173971	*/
	@@ -177553,11 +177855,15 @@
177855	}
177856	break;
177857	case 84: /* cmd ::= select */
177858	{
177859	SelectDest dest = {SRT_Output, 0, 0, 0, 0, 0, 0};
177860	if( (pParse->db->mDbFlags & DBFLAG_EncodingFixed)!=0
177861	\|\| sqlite3ReadSchema(pParse)==SQLITE_OK
177862	){
177863	sqlite3Select(pParse, yymsp[0].minor.yy555, &dest);
177864	}
177865	sqlite3SelectDelete(pParse->db, yymsp[0].minor.yy555);
177866	}
177867	break;
177868	case 85: /* select ::= WITH wqlist selectnowith */
177869	{yymsp[-2].minor.yy555 = attachWithToSelect(pParse,yymsp[0].minor.yy555,yymsp[-1].minor.yy59);}
	@@ -178024,11 +178330,11 @@
178330	** that look like this: #1 #2 ... These terms refer to registers
178331	** in the virtual machine. #N is the N-th register. */
178332	Token t = yymsp[0].minor.yy0; /A-overwrites-X/
178333	assert( t.n>=2 );
178334	if( pParse->nested==0 ){
178335	parserSyntaxError(pParse, &t);
178336	yymsp[0].minor.yy454 = 0;
178337	}else{
178338	yymsp[0].minor.yy454 = sqlite3PExpr(pParse, TK_REGISTER, 0, 0);
178339	if( yymsp[0].minor.yy454 ) sqlite3GetInt32(&t.z[1], &yymsp[0].minor.yy454->iTable);
178340	}
	@@ -178872,11 +179178,11 @@
179178	#define TOKEN yyminor
179179	/********** Begin %syntax_error code **************************************/
179180
179181	UNUSED_PARAMETER(yymajor); /* Silence some compiler warnings */
179182	if( TOKEN.z[0] ){
179183	parserSyntaxError(pParse, &TOKEN);
179184	}else{
179185	sqlite3ErrorMsg(pParse, "incomplete input");
179186	}
179187	/********** End %syntax_error code ****************************************/
179188	sqlite3ParserARG_STORE /* Suppress warning about unused %extra_argument variable */
	@@ -180363,11 +180669,13 @@
180669	}
180670	if( pParse->zErrMsg \|\| (pParse->rc!=SQLITE_OK && pParse->rc!=SQLITE_DONE) ){
180671	if( pParse->zErrMsg==0 ){
180672	pParse->zErrMsg = sqlite3MPrintf(db, "%s", sqlite3ErrStr(pParse->rc));
180673	}
180674	if( (pParse->prepFlags & SQLITE_PREPARE_DONT_LOG)==0 ){
180675	sqlite3_log(pParse->rc, "%s in \"%s\"", pParse->zErrMsg, pParse->zTail);
180676	}
180677	nErr++;
180678	}
180679	pParse->zTail = zSql;
180680	#ifndef SQLITE_OMIT_VIRTUALTABLE
180681	sqlite3_free(pParse->apVtabLock);
	@@ -181071,36 +181379,10 @@
181379	** all database files specified with a relative pathname.
181380	**
181381	** See also the "PRAGMA data_store_directory" SQL command.
181382	*/
181383	SQLITE_API char *sqlite3_data_directory = 0;


























181384
181385	/*
181386	** Initialize SQLite.
181387	**
181388	** This routine must be called to initialize the memory allocation,
	@@ -181292,17 +181574,10 @@
181574	if( bRunExtraInit ){
181575	int SQLITE_EXTRA_INIT(const char*);
181576	rc = SQLITE_EXTRA_INIT(0);
181577	}
181578	#endif







181579	return rc;
181580	}
181581
181582	/*
181583	** Undo the effects of sqlite3_initialize(). Must not be called while
	@@ -182369,14 +182644,10 @@
182644	#endif
182645
182646	sqlite3Error(db, SQLITE_OK); /* Deallocates any cached error strings. */
182647	sqlite3ValueFree(db->pErr);
182648	sqlite3CloseExtensions(db);




182649
182650	db->eOpenState = SQLITE_STATE_ERROR;
182651
182652	/* The temp-database schema is allocated differently from the other schema
182653	** objects (using sqliteMalloc() directly, instead of sqlite3BtreeSchema()).
	@@ -183875,12 +184146,12 @@
184146	}
184147	oldLimit = db->aLimit[limitId];
184148	if( newLimit>=0 ){ /* IMP: R-52476-28732 */
184149	if( newLimit>aHardLimit[limitId] ){
184150	newLimit = aHardLimit[limitId]; /* IMP: R-51463-25634 */
184151	}else if( newLimit<SQLITE_MIN_LENGTH && limitId==SQLITE_LIMIT_LENGTH ){
184152	newLimit = SQLITE_MIN_LENGTH;
184153	}
184154	db->aLimit[limitId] = newLimit;
184155	}
184156	return oldLimit; /* IMP: R-53341-35419 */
184157	}
	@@ -184395,10 +184666,11 @@
184666	testcase( (1<<(flags&7))==0x04 ); /* READWRITE */
184667	testcase( (1<<(flags&7))==0x40 ); /* READWRITE \| CREATE */
184668	if( ((1<<(flags&7)) & 0x46)==0 ){
184669	rc = SQLITE_MISUSE_BKPT; /* IMP: R-18321-05872 */
184670	}else{
184671	if( zFilename==0 ) zFilename = ":memory:";
184672	rc = sqlite3ParseUri(zVfs, zFilename, &flags, &db->pVfs, &zOpen, &zErrMsg);
184673	}
184674	if( rc!=SQLITE_OK ){
184675	if( rc==SQLITE_NOMEM ) sqlite3OomFault(db);
184676	sqlite3ErrorWithMsg(db, rc, zErrMsg ? "%s" : 0, zErrMsg);
	@@ -185237,10 +185509,11 @@
185509	#ifndef SQLITE_OMIT_WINDOWFUNC
185510	sqlite3ShowWindow(0);
185511	sqlite3ShowWinFunc(0);
185512	#endif
185513	sqlite3ShowSelect(0);
185514	sqlite3ShowWhereTerm(0);
185515	}
185516	#endif
185517	break;
185518	}
185519
	@@ -185549,28 +185822,10 @@
185822	*pI1 = rLogEst;
185823	*pU64 = sqlite3LogEstToInt(rLogEst);
185824	pI2 = sqlite3LogEst(pU64);
185825	break;
185826	}


















185827
185828	#if defined(SQLITE_DEBUG) && !defined(SQLITE_OMIT_WSD)
185829	/* sqlite3_test_control(SQLITE_TESTCTRL_TUNE, id, *piValue)
185830	**
185831	** If "id" is an integer between 1 and SQLITE_NTUNE then set the value
	@@ -185875,11 +186130,15 @@
186130	if( db->autoCommit==0 ){
186131	int iDb = sqlite3FindDbName(db, zDb);
186132	if( iDb==0 \|\| iDb>1 ){
186133	Btree *pBt = db->aDb[iDb].pBt;
186134	if( SQLITE_TXN_WRITE!=sqlite3BtreeTxnState(pBt) ){
186135	Pager *pPager = sqlite3BtreePager(pBt);
186136	i64 dummy = 0;
186137	sqlite3PagerSnapshotOpen(pPager, (sqlite3_snapshot*)&dummy);
186138	rc = sqlite3BtreeBeginTrans(pBt, 0, 0);
186139	sqlite3PagerSnapshotOpen(pPager, 0);
186140	if( rc==SQLITE_OK ){
186141	rc = sqlite3PagerSnapshotGet(sqlite3BtreePager(pBt), ppSnapshot);
186142	}
186143	}
186144	}
	@@ -189665,14 +189924,19 @@
189924
189925	assert_fts3_nc( p!=0 && p1!=0 && p2!=0 );
189926	if( *p1==POS_COLUMN ){
189927	p1++;
189928	p1 += fts3GetVarint32(p1, &iCol1);
189929	/* iCol1==0 indicates corruption. Column 0 does not have a POS_COLUMN
189930	** entry, so this is actually end-of-doclist. */
189931	if( iCol1==0 ) return 0;
189932	}
189933	if( *p2==POS_COLUMN ){
189934	p2++;
189935	p2 += fts3GetVarint32(p2, &iCol2);
189936	/* As above, iCol2==0 indicates corruption. */
189937	if( iCol2==0 ) return 0;
189938	}
189939
189940	while( 1 ){
189941	if( iCol1==iCol2 ){
189942	char *pSave = p;
	@@ -192839,11 +193103,11 @@
193103	for(p=pExpr; p->pLeft; p=p->pLeft){
193104	assert( p->pRight->pPhrase->doclist.nList>0 );
193105	nTmp += p->pRight->pPhrase->doclist.nList;
193106	}
193107	nTmp += p->pPhrase->doclist.nList;
193108	aTmp = sqlite3_malloc64(nTmp*2 + FTS3_VARINT_MAX);
193109	if( !aTmp ){
193110	*pRc = SQLITE_NOMEM;
193111	res = 0;
193112	}else{
193113	char *aPoslist = p->pPhrase->doclist.pList;
	@@ -193490,11 +193754,11 @@
193754	SQLITE_PRIVATE int sqlite3Fts3Corrupt(){
193755	return SQLITE_CORRUPT_VTAB;
193756	}
193757	#endif
193758
193759	#if !defined(SQLITE_CORE)
193760	/*
193761	** Initialize API pointer table, if required.
193762	*/
193763	#ifdef _WIN32
193764	__declspec(dllexport)
	@@ -194392,14 +194656,15 @@
194656	rc = pModule->xNext(pCursor, &zByte, &nByte, &iBegin, &iEnd, &iPos);
194657	if( rc==SQLITE_OK ){
194658	Fts3PhraseToken *pToken;
194659
194660	p = fts3ReallocOrFree(p, nSpace + ii*sizeof(Fts3PhraseToken));


194661	zTemp = fts3ReallocOrFree(zTemp, nTemp + nByte);
194662	if( !zTemp \|\| !p ){
194663	rc = SQLITE_NOMEM;
194664	goto getnextstring_out;
194665	}
194666
194667	assert( nToken==ii );
194668	pToken = &((Fts3Phrase *)(&p[1]))->aToken[ii];
194669	memset(pToken, 0, sizeof(Fts3PhraseToken));
194670
	@@ -194410,53 +194675,51 @@
194675	pToken->isPrefix = (iEnd<nInput && zInput[iEnd]=='*');
194676	pToken->bFirst = (iBegin>0 && zInput[iBegin-1]=='^');
194677	nToken = ii+1;
194678	}
194679	}



194680	}
194681
194682	if( rc==SQLITE_DONE ){
194683	int jj;
194684	char *zBuf = 0;
194685
194686	p = fts3ReallocOrFree(p, nSpace + nToken*sizeof(Fts3PhraseToken) + nTemp);
194687	if( !p ){
194688	rc = SQLITE_NOMEM;
194689	goto getnextstring_out;
194690	}
194691	memset(p, 0, (char )&(((Fts3Phrase )&p[1])->aToken[0])-(char *)p);
194692	p->eType = FTSQUERY_PHRASE;
194693	p->pPhrase = (Fts3Phrase *)&p[1];
194694	p->pPhrase->iColumn = pParse->iDefaultCol;
194695	p->pPhrase->nToken = nToken;
194696
194697	zBuf = (char *)&p->pPhrase->aToken[nToken];
194698	assert( nTemp==0 \|\| zTemp );
194699	if( zTemp ){
194700	memcpy(zBuf, zTemp, nTemp);



194701	}
194702
194703	for(jj=0; jj<p->pPhrase->nToken; jj++){
194704	p->pPhrase->aToken[jj].z = zBuf;
194705	zBuf += p->pPhrase->aToken[jj].n;
194706	}
194707	rc = SQLITE_OK;
194708	}
194709
194710	getnextstring_out:



194711	if( pCursor ){
194712	pModule->xClose(pCursor);
194713	}
194714	sqlite3_free(zTemp);
194715	if( rc!=SQLITE_OK ){
194716	sqlite3_free(p);
194717	p = 0;
194718	}
194719	*ppExpr = p;
194720	return rc;
194721	}
194722
194723	/*
194724	** The output variable *ppExpr is populated with an allocated Fts3Expr
194725	** structure, or set to 0 if the end of the input buffer is reached.
	@@ -208867,11 +209130,13 @@
209130	int rawKey = 1;
209131	x = pParse->aBlob[iRoot];
209132	zPath++;
209133	if( zPath[0]=='"' ){
209134	zKey = zPath + 1;
209135	for(i=1; zPath[i] && zPath[i]!='"'; i++){
209136	if( zPath[i]=='\\' && zPath[i+1]!=0 ) i++;
209137	}
209138	nKey = i-1;
209139	if( zPath[i] ){
209140	i++;
209141	}else{
209142	return JSON_LOOKUP_PATHERROR;
	@@ -217779,11 +218044,11 @@
218044	return sqlite3_create_function_v2(db, zQueryFunc, -1, SQLITE_ANY,
218045	(void *)pGeomCtx, geomCallback, 0, 0, rtreeFreeCallback
218046	);
218047	}
218048
218049	#ifndef SQLITE_CORE
218050	#ifdef _WIN32
218051	__declspec(dllexport)
218052	#endif
218053	SQLITE_API int sqlite3_rtree_init(
218054	sqlite3 *db,
	@@ -218370,11 +218635,11 @@
218635	}
218636
218637	return rc;
218638	}
218639
218640	#ifndef SQLITE_CORE
218641	#ifdef _WIN32
218642	__declspec(dllexport)
218643	#endif
218644	SQLITE_API int sqlite3_icu_init(
218645	sqlite3 *db,
	@@ -219628,10 +219893,31 @@
219893	struct RbuFrame {
219894	u32 iDbPage;
219895	u32 iWalFrame;
219896	};
219897
219898	#ifndef UNUSED_PARAMETER
219899	/*
219900	** The following macros are used to suppress compiler warnings and to
219901	** make it clear to human readers when a function parameter is deliberately
219902	** left unused within the body of a function. This usually happens when
219903	** a function is called via a function pointer. For example the
219904	** implementation of an SQL aggregate step callback may not use the
219905	** parameter indicating the number of arguments passed to the aggregate,
219906	** if it knows that this is enforced elsewhere.
219907	**
219908	** When a function parameter is not used at all within the body of a function,
219909	** it is generally named "NotUsed" or "NotUsed2" to make things even clearer.
219910	** However, these macros may also be used to suppress warnings related to
219911	** parameters that may or may not be used depending on compilation options.
219912	** For example those parameters only used in assert() statements. In these
219913	** cases the parameters are named as per the usual conventions.
219914	*/
219915	#define UNUSED_PARAMETER(x) (void)(x)
219916	#define UNUSED_PARAMETER2(x,y) UNUSED_PARAMETER(x),UNUSED_PARAMETER(y)
219917	#endif
219918
219919	/*
219920	** RBU handle.
219921	**
219922	** nPhaseOneStep:
219923	** If the RBU database contains an rbu_count table, this value is set to
	@@ -219679,11 +219965,11 @@
219965	char zState; / Path to state db (or NULL if zRbu) */
219966	char zStateDb[5]; /* Db name for state ("stat" or "main") */
219967	int rc; /* Value returned by last rbu_step() call */
219968	char zErrmsg; / Error message if rc!=SQLITE_OK */
219969	int nStep; /* Rows processed for current object */
219970	sqlite3_int64 nProgress; /* Rows processed for all objects */
219971	RbuObjIter objiter; /* Iterator for skipping through tbl/idx */
219972	const char zVfsName; / Name of automatically created rbu vfs */
219973	rbu_file pTargetFd; / File handle open on target db */
219974	int nPagePerSector; /* Pages per sector for pTargetFd */
219975	i64 iOalSz;
	@@ -219796,11 +220082,11 @@
220082	unsigned char *zStart = z;
220083	while( (c = zValue[0x7f&*(z++)])>=0 ){
220084	v = (v<<6) + c;
220085	}
220086	z--;
220087	*pLen -= (int)(z - zStart);
220088	pz = (char)z;
220089	return v;
220090	}
220091
220092	#if RBU_ENABLE_DELTA_CKSUM
	@@ -219981,10 +220267,11 @@
220267	int nOut;
220268	int nOut2;
220269	char *aOut;
220270
220271	assert( argc==2 );
220272	UNUSED_PARAMETER(argc);
220273
220274	nOrig = sqlite3_value_bytes(argv[0]);
220275	aOrig = (const char*)sqlite3_value_blob(argv[0]);
220276	nDelta = sqlite3_value_bytes(argv[1]);
220277	aDelta = (const char*)sqlite3_value_blob(argv[1]);
	@@ -221560,17 +221847,17 @@
221847	nParen++;
221848	}
221849	else if( c==')' ){
221850	nParen--;
221851	if( nParen==0 ){
221852	int nSpan = (int)(&zSql[i] - pIter->aIdxCol[iIdxCol].zSpan);
221853	pIter->aIdxCol[iIdxCol++].nSpan = nSpan;
221854	i++;
221855	break;
221856	}
221857	}else if( c==',' && nParen==1 ){
221858	int nSpan = (int)(&zSql[i] - pIter->aIdxCol[iIdxCol].zSpan);
221859	pIter->aIdxCol[iIdxCol++].nSpan = nSpan;
221860	pIter->aIdxCol[iIdxCol].zSpan = &zSql[i+1];
221861	}else if( c=='"' \|\| c=='\'' \|\| c=='`' ){
221862	for(i++; 1; i++){
221863	if( zSql[i]==c ){
	@@ -222256,10 +222543,12 @@
222543	int i, sz;
222544	sz = (int)strlen(z)&0xffffff;
222545	for(i=sz-1; i>0 && z[i]!='/' && z[i]!='.'; i--){}
222546	if( z[i]=='.' && sz>i+4 ) memmove(&z[i+1], &z[sz-3], 4);
222547	}
222548	#else
222549	UNUSED_PARAMETER2(zBase,z);
222550	#endif
222551	}
222552
222553	/*
222554	** Return the current wal-index header checksum for the target database
	@@ -222840,11 +223129,11 @@
223129	"INSERT OR REPLACE INTO %s.rbu_state(k, v) VALUES "
223130	"(%d, %d), "
223131	"(%d, %Q), "
223132	"(%d, %Q), "
223133	"(%d, %d), "
223134	"(%d, %lld), "
223135	"(%d, %lld), "
223136	"(%d, %lld), "
223137	"(%d, %lld), "
223138	"(%d, %lld), "
223139	"(%d, %Q) ",
	@@ -223198,10 +223487,11 @@
223487	char *zErrmsg = 0;
223488	int rc;
223489	sqlite3 *db = (rbuIsVacuum(p) ? p->dbRbu : p->dbMain);
223490
223491	assert( nVal==1 );
223492	UNUSED_PARAMETER(nVal);
223493
223494	rc = prepareFreeAndCollectError(db, &pStmt, &zErrmsg,
223495	sqlite3_mprintf("SELECT count(*) FROM sqlite_schema "
223496	"WHERE type='index' AND tbl_name = %Q", sqlite3_value_text(apVal[0]))
223497	);
	@@ -223473,11 +223763,11 @@
223763	const char *zTarget,
223764	const char *zState
223765	){
223766	if( zTarget==0 ){ return rbuMisuseError(); }
223767	if( zState ){
223768	size_t n = strlen(zState);
223769	if( n>=7 && 0==memcmp("-vactmp", &zState[n-7], 7) ){
223770	return rbuMisuseError();
223771	}
223772	}
223773	/* TODO: Check that both arguments are non-NULL */
	@@ -223690,10 +223980,11 @@
223980	/*
223981	** Default xRename callback for RBU.
223982	*/
223983	static int xDefaultRename(void pArg, const char zOld, const char *zNew){
223984	int rc = SQLITE_OK;
223985	UNUSED_PARAMETER(pArg);
223986	#if defined(_WIN32_WCE)
223987	{
223988	LPWSTR zWideOld;
223989	LPWSTR zWideNew;
223990
	@@ -224594,10 +224885,13 @@
224885
224886	/*
224887	** No-op.
224888	*/
224889	static int rbuVfsGetLastError(sqlite3_vfs pVfs, int a, char b){
224890	UNUSED_PARAMETER(pVfs);
224891	UNUSED_PARAMETER(a);
224892	UNUSED_PARAMETER(b);
224893	return 0;
224894	}
224895
224896	/*
224897	** Deregister and destroy an RBU vfs created by an earlier call to
	@@ -225650,11 +225944,17 @@
225944	** schema for the database file that is to be read. The default schema is
225945	** "main".
225946	**
225947	** The data field of sqlite_dbpage table can be updated. The new
225948	** value must be a BLOB which is the correct page size, otherwise the
225949	** update fails. INSERT operations also work, and operate as if they
225950	** where REPLACE. The size of the database can be extended by INSERT-ing
225951	** new pages on the end.
225952	**
225953	** Rows may not be deleted. However, doing an INSERT to page number N
225954	** with NULL page data causes the N-th page and all subsequent pages to be
225955	** deleted and the database to be truncated.
225956	*/
225957
225958	/* #include "sqliteInt.h" Requires access to internal data structures */
225959	#if (defined(SQLITE_ENABLE_DBPAGE_VTAB) \|\| defined(SQLITE_TEST)) \
225960	&& !defined(SQLITE_OMIT_VIRTUALTABLE)
	@@ -225673,17 +225973,18 @@
225973	};
225974
225975	struct DbpageTable {
225976	sqlite3_vtab base; /* Base class. Must be first */
225977	sqlite3 db; / The database */
225978	int iDbTrunc; /* Database to truncate */
225979	Pgno pgnoTrunc; /* Size to truncate to */
225980	};
225981
225982	/* Columns */
225983	#define DBPAGE_COLUMN_PGNO 0
225984	#define DBPAGE_COLUMN_DATA 1
225985	#define DBPAGE_COLUMN_SCHEMA 2

225986
225987
225988	/*
225989	** Connect to or create a dbpagevfs virtual table.
225990	*/
	@@ -225943,15 +226244,15 @@
226244	DbpageTable pTab = (DbpageTable )pVtab;
226245	Pgno pgno;
226246	DbPage *pDbPage = 0;
226247	int rc = SQLITE_OK;
226248	char *zErr = 0;

226249	int iDb;
226250	Btree *pBt;
226251	Pager *pPager;
226252	int szPage;
226253	int isInsert;
226254
226255	(void)pRowid;
226256	if( pTab->db->flags & SQLITE_Defensive ){
226257	zErr = "read-only";
226258	goto update_fail;
	@@ -225958,45 +226259,62 @@
226259	}
226260	if( argc==1 ){
226261	zErr = "cannot delete";
226262	goto update_fail;
226263	}
226264	if( sqlite3_value_type(argv[0])==SQLITE_NULL ){
226265	pgno = (Pgno)sqlite3_value_int(argv[2]);
226266	isInsert = 1;
226267	}else{
226268	pgno = sqlite3_value_int(argv[0]);
226269	if( (Pgno)sqlite3_value_int(argv[1])!=pgno ){
226270	zErr = "cannot insert";
226271	goto update_fail;
226272	}
226273	isInsert = 0;
226274	}
226275	if( sqlite3_value_type(argv[4])==SQLITE_NULL ){
226276	iDb = 0;
226277	}else{
226278	const char zSchema = (const char)sqlite3_value_text(argv[4]);
226279	iDb = sqlite3FindDbName(pTab->db, zSchema);
226280	if( iDb<0 ){
226281	zErr = "no such schema";
226282	goto update_fail;
226283	}
226284	}
226285	pBt = pTab->db->aDb[iDb].pBt;
226286	if( pgno<1 \|\| NEVER(pBt==0) ){
226287	zErr = "bad page number";
226288	goto update_fail;
226289	}
226290	szPage = sqlite3BtreeGetPageSize(pBt);
226291	if( sqlite3_value_type(argv[3])!=SQLITE_BLOB
226292	\|\| sqlite3_value_bytes(argv[3])!=szPage
226293	){
226294	if( sqlite3_value_type(argv[3])==SQLITE_NULL && isInsert && pgno>1 ){
226295	/* "INSERT INTO dbpage($PGNO,NULL)" causes page number $PGNO and
226296	** all subsequent pages to be deleted. */
226297	pTab->iDbTrunc = iDb;
226298	pgno--;
226299	pTab->pgnoTrunc = pgno;
226300	}else{
226301	zErr = "bad page value";
226302	goto update_fail;
226303	}
226304	}
226305	pPager = sqlite3BtreePager(pBt);
226306	rc = sqlite3PagerGet(pPager, pgno, (DbPage**)&pDbPage, 0);
226307	if( rc==SQLITE_OK ){
226308	const void *pData = sqlite3_value_blob(argv[3]);
226309	if( (rc = sqlite3PagerWrite(pDbPage))==SQLITE_OK && pData ){
226310	unsigned char *aPage = sqlite3PagerGetData(pDbPage);
226311	memcpy(aPage, pData, szPage);
226312	pTab->pgnoTrunc = 0;

226313	}
226314	}else{
226315	pTab->pgnoTrunc = 0;
226316	}
226317	sqlite3PagerUnref(pDbPage);
226318	return rc;
226319
226320	update_fail:
	@@ -226015,13 +226333,35 @@
226333	int i;
226334	for(i=0; i<db->nDb; i++){
226335	Btree *pBt = db->aDb[i].pBt;
226336	if( pBt ) (void)sqlite3BtreeBeginTrans(pBt, 1, 0);
226337	}
226338	pTab->pgnoTrunc = 0;
226339	return SQLITE_OK;
226340	}
226341
226342	/* Invoke sqlite3PagerTruncate() as necessary, just prior to COMMIT
226343	*/
226344	static int dbpageSync(sqlite3_vtab *pVtab){
226345	DbpageTable pTab = (DbpageTable )pVtab;
226346	if( pTab->pgnoTrunc>0 ){
226347	Btree *pBt = pTab->db->aDb[pTab->iDbTrunc].pBt;
226348	Pager *pPager = sqlite3BtreePager(pBt);
226349	sqlite3PagerTruncateImage(pPager, pTab->pgnoTrunc);
226350	}
226351	pTab->pgnoTrunc = 0;
226352	return SQLITE_OK;
226353	}
226354
226355	/* Cancel any pending truncate.
226356	*/
226357	static int dbpageRollbackTo(sqlite3_vtab *pVtab, int notUsed1){
226358	DbpageTable pTab = (DbpageTable )pVtab;
226359	pTab->pgnoTrunc = 0;
226360	(void)notUsed1;
226361	return SQLITE_OK;
226362	}

226363
226364	/*
226365	** Invoke this routine to register the "dbpage" virtual table module
226366	*/
226367	SQLITE_PRIVATE int sqlite3DbpageRegister(sqlite3 *db){
	@@ -226039,18 +226379,18 @@
226379	dbpageEof, /* xEof - check for end of scan */
226380	dbpageColumn, /* xColumn - read data */
226381	dbpageRowid, /* xRowid - read data */
226382	dbpageUpdate, /* xUpdate */
226383	dbpageBegin, /* xBegin */
226384	dbpageSync, /* xSync */
226385	0, /* xCommit */
226386	0, /* xRollback */
226387	0, /* xFindMethod */
226388	0, /* xRename */
226389	0, /* xSavepoint */
226390	0, /* xRelease */
226391	dbpageRollbackTo, /* xRollbackTo */
226392	0, /* xShadowName */
226393	0 /* xIntegrity */
226394	};
226395	return sqlite3_create_module(db, "sqlite_dbpage", &dbpage_module, 0);
226396	}
	@@ -226134,10 +226474,14 @@
226474	/*
226475	** An object of this type is used internally as an abstraction for
226476	** input data. Input data may be supplied either as a single large buffer
226477	** (e.g. sqlite3changeset_start()) or using a stream function (e.g.
226478	** sqlite3changeset_start_strm()).
226479	**
226480	** bNoDiscard:
226481	** If true, then the only time data is discarded is as a result of explicit
226482	** sessionDiscardData() calls. Not within every sessionInputBuffer() call.
226483	*/
226484	struct SessionInput {
226485	int bNoDiscard; /* If true, do not discard in InputBuffer() */
226486	int iCurrent; /* Offset in aData[] of current change */
226487	int iNext; /* Offset in aData[] of next change */
	@@ -227817,20 +228161,23 @@
228161	/* Figure out how large an allocation is required */
228162	nByte = sizeof(SessionChange);
228163	for(i=0; i<(pTab->nCol-pTab->bRowid); i++){
228164	sqlite3_value *p = 0;
228165	if( op!=SQLITE_INSERT ){
228166	/* This may fail if the column has a non-NULL default and was added
228167	** using ALTER TABLE ADD COLUMN after this record was created. */
228168	rc = pSession->hook.xOld(pSession->hook.pCtx, i, &p);
228169	}else if( pTab->abPK[i] ){
228170	TESTONLY(int trc = ) pSession->hook.xNew(pSession->hook.pCtx, i, &p);
228171	assert( trc==SQLITE_OK );
228172	}
228173
228174	if( rc==SQLITE_OK ){
228175	/* This may fail if SQLite value p contains a utf-16 string that must
228176	** be converted to utf-8 and an OOM error occurs while doing so. */
228177	rc = sessionSerializeValue(0, p, &nByte);
228178	}
228179	if( rc!=SQLITE_OK ) goto error_out;
228180	}
228181	if( pTab->bRowid ){
228182	nByte += 9; /* Size of rowid field - an integer */
228183	}
	@@ -231184,19 +231531,25 @@
231531	int rc = SQLITE_OK; /* Return code */
231532	const char zTab = 0; / Name of current table */
231533	int nTab = 0; /* Result of sqlite3Strlen30(zTab) */
231534	SessionApplyCtx sApply; /* changeset_apply() context object */
231535	int bPatchset;
231536	u64 savedFlag = db->flags & SQLITE_FkNoAction;
231537
231538	assert( xConflict!=0 );
231539
231540	sqlite3_mutex_enter(sqlite3_db_mutex(db));
231541	if( flags & SQLITE_CHANGESETAPPLY_FKNOACTION ){
231542	db->flags \|= ((u64)SQLITE_FkNoAction);
231543	db->aDb[0].pSchema->schema_cookie -= 32;
231544	}
231545
231546	pIter->in.bNoDiscard = 1;
231547	memset(&sApply, 0, sizeof(sApply));
231548	sApply.bRebase = (ppRebase && pnRebase);
231549	sApply.bInvertConstraints = !!(flags & SQLITE_CHANGESETAPPLY_INVERT);
231550	sApply.bIgnoreNoop = !!(flags & SQLITE_CHANGESETAPPLY_IGNORENOOP);

231551	if( (flags & SQLITE_CHANGESETAPPLY_NOSAVEPOINT)==0 ){
231552	rc = sqlite3_exec(db, "SAVEPOINT changeset_apply", 0, 0, 0);
231553	}
231554	if( rc==SQLITE_OK ){
231555	rc = sqlite3_exec(db, "PRAGMA defer_foreign_keys = 1", 0, 0, 0);
	@@ -231354,10 +231707,16 @@
231707	sqlite3_finalize(sApply.pDelete);
231708	sqlite3_finalize(sApply.pSelect);
231709	sqlite3_free((char)sApply.azCol); / cast works around VC++ bug */
231710	sqlite3_free((char*)sApply.constraints.aBuf);
231711	sqlite3_free((char*)sApply.rebase.aBuf);
231712
231713	if( (flags & SQLITE_CHANGESETAPPLY_FKNOACTION) && savedFlag==0 ){
231714	assert( db->flags & SQLITE_FkNoAction );
231715	db->flags &= ~((u64)SQLITE_FkNoAction);
231716	db->aDb[0].pSchema->schema_cookie -= 32;
231717	}
231718	sqlite3_mutex_leave(sqlite3_db_mutex(db));
231719	return rc;
231720	}
231721
231722	/*
	@@ -231382,28 +231741,17 @@
231741	int flags
231742	){
231743	sqlite3_changeset_iter pIter; / Iterator to skip through changeset */
231744	int bInv = !!(flags & SQLITE_CHANGESETAPPLY_INVERT);
231745	int rc = sessionChangesetStart(&pIter, 0, 0, nChangeset, pChangeset, bInv, 1);






231746
231747	if( rc==SQLITE_OK ){
231748	rc = sessionChangesetApply(
231749	db, pIter, xFilter, xConflict, pCtx, ppRebase, pnRebase, flags
231750	);
231751	}
231752





231753	return rc;
231754	}
231755
231756	/*
231757	** Apply the changeset passed via pChangeset/nChangeset to the main database
	@@ -231720,10 +232068,13 @@
232068	if( op==SQLITE_INSERT \|\| (op==SQLITE_DELETE && pGrp->bPatch==0) ){
232069	/* Append the missing default column values to the record. */
232070	sessionAppendBlob(pOut, aRec, nRec, &rc);
232071	if( rc==SQLITE_OK && pTab->pDfltStmt==0 ){
232072	rc = sessionPrepareDfltStmt(pGrp->db, pTab, &pTab->pDfltStmt);
232073	if( rc==SQLITE_OK && SQLITE_ROW!=sqlite3_step(pTab->pDfltStmt) ){
232074	rc = sqlite3_errcode(pGrp->db);
232075	}
232076	}
232077	for(ii=nCol; rc==SQLITE_OK && ii<pTab->nCol; ii++){
232078	int eType = sqlite3_column_type(pTab->pDfltStmt, ii);
232079	sessionAppendByte(pOut, eType, &rc);
232080	switch( eType ){
	@@ -231736,10 +232087,11 @@
232087	double rVal = sqlite3_column_int64(pTab->pDfltStmt, ii);
232088	memcpy(&iVal, &rVal, sizeof(i64));
232089	}
232090	if( SQLITE_OK==sessionBufferGrow(pOut, 8, &rc) ){
232091	sessionPutI64(&pOut->aBuf[pOut->nBuf], iVal);
232092	pOut->nBuf += 8;
232093	}
232094	break;
232095	}
232096
232097	case SQLITE_BLOB:
	@@ -231874,10 +232226,12 @@
232226	SessionChange *pExist = 0;
232227	SessionChange **pp = 0;
232228	SessionTable *pTab = 0;
232229	u8 *aRec = &pIter->in.aData[pIter->in.iCurrent + 2];
232230	int nRec = (pIter->in.iNext - pIter->in.iCurrent) - 2;
232231
232232	assert( nRec>0 );
232233
232234	/* Ensure that only changesets, or only patchsets, but not a mixture
232235	** of both, are being combined. It is an error to try to combine a
232236	** changeset and a patchset. */
232237	if( pGrp->pList==0 ){
	@@ -231952,10 +232306,11 @@
232306	){
232307	u8 *aRec;
232308	int nRec;
232309	int rc = SQLITE_OK;
232310
232311	pIter->in.bNoDiscard = 1;
232312	while( SQLITE_ROW==(sessionChangesetNext(pIter, &aRec, &nRec, 0)) ){
232313	rc = sessionOneChangeToHash(pGrp, pIter, bRebase);
232314	if( rc!=SQLITE_OK ) break;
232315	}
232316
	@@ -232583,20 +232938,46 @@
232938	#endif /* SQLITE_ENABLE_SESSION && SQLITE_ENABLE_PREUPDATE_HOOK */
232939
232940	/************ End of sqlite3session.c ************************************/
232941	/************ Begin file fts5.c ******************************************/
232942
232943	/*
232944	** This, the "fts5.c" source file, is a composite file that is itself
232945	** assembled from the following files:
232946	**
232947	** fts5.h
232948	** fts5Int.h
232949	** fts5parse.h <--- Generated from fts5parse.y by Lemon
232950	** fts5parse.c <--- Generated from fts5parse.y by Lemon
232951	** fts5_aux.c
232952	** fts5_buffer.c
232953	** fts5_config.c
232954	** fts5_expr.c
232955	** fts5_hash.c
232956	** fts5_index.c
232957	** fts5_main.c
232958	** fts5_storage.c
232959	** fts5_tokenize.c
232960	** fts5_unicode2.c
232961	** fts5_varint.c
232962	** fts5_vocab.c
232963	*/
232964	#if !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_FTS5)
232965
232966	#if !defined(NDEBUG) && !defined(SQLITE_DEBUG)
232967	# define NDEBUG 1
232968	#endif
232969	#if defined(NDEBUG) && defined(SQLITE_DEBUG)
232970	# undef NDEBUG
232971	#endif
232972
232973	#ifdef HAVE_STDINT_H
232974	/* #include <stdint.h> */
232975	#endif
232976	#ifdef HAVE_INTTYPES_H
232977	/* #include <inttypes.h> */
232978	#endif
232979	/*
232980	** 2014 May 31
232981	**
232982	** The author disclaims copyright to this source code. In place of
232983	** a legal notice, here is a blessing:
	@@ -232893,17 +233274,32 @@
233274	** This is used to access token iToken of phrase hit iIdx within the
233275	** current row. If iIdx is less than zero or greater than or equal to the
233276	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
233277	** output variable (*ppToken) is set to point to a buffer containing the
233278	** matching document token, and (*pnToken) to the size of that buffer in
233279	** bytes.


233280	**
233281	** The output text is not a copy of the document text that was tokenized.
233282	** It is the output of the tokenizer module. For tokendata=1 tables, this
233283	** includes any embedded 0x00 and trailing data.
233284	**
233285	** This API may be slow in some cases if the token identified by parameters
233286	** iIdx and iToken matched a prefix token in the query. In most cases, the
233287	** first call to this API for each prefix token in the query is forced
233288	** to scan the portion of the full-text index that matches the prefix
233289	** token to collect the extra data required by this API. If the prefix
233290	** token matches a large number of token instances in the document set,
233291	** this may be a performance problem.
233292	**
233293	** If the user knows in advance that a query may use this API for a
233294	** prefix token, FTS5 may be configured to collect all required data as part
233295	** of the initial querying of the full-text index, avoiding the second scan
233296	** entirely. This also causes prefix queries that do not use this API to
233297	** run more slowly and use more memory. FTS5 may be configured in this way
233298	** either on a per-table basis using the [FTS5 insttoken \| 'insttoken']
233299	** option, or on a per-query basis using the
233300	** [fts5_insttoken \| fts5_insttoken()] user function.
233301	**
233302	** This API can be quite slow if used with an FTS5 table created with the
233303	** "detail=none" or "detail=column" option.
233304	**
233305	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
	@@ -232991,11 +233387,10 @@
233387	** CUSTOM TOKENIZERS
233388	**
233389	** Applications may also register custom tokenizer types. A tokenizer
233390	** is registered by providing fts5 with a populated instance of the
233391	** following structure. All structure methods must be defined, setting

233392	** any member of the fts5_tokenizer struct to NULL leads to undefined
233393	** behaviour. The structure methods are expected to function as follows:
233394	**
233395	** xCreate:
233396	** This function is used to allocate and initialize a tokenizer instance.
	@@ -233560,10 +233955,11 @@
233955	u8 abUnindexed; / True for unindexed columns */
233956	int nPrefix; /* Number of prefix indexes */
233957	int aPrefix; / Sizes in bytes of nPrefix prefix indexes */
233958	int eContent; /* An FTS5_CONTENT value */
233959	int bContentlessDelete; /* "contentless_delete=" option (dflt==0) */
233960	int bContentlessUnindexed; /* "contentless_unindexed=" option (dflt=0) */
233961	char zContent; / content table */
233962	char zContentRowid; / "content_rowid=" option value */
233963	int bColumnsize; /* "columnsize=" option value (dflt==1) */
233964	int bTokendata; /* "tokendata=" option value (dflt==0) */
233965	int bLocale; /* "locale=" option value (dflt==0) */
	@@ -233582,11 +233978,12 @@
233978	int nUsermerge; /* 'usermerge' setting */
233979	int nHashSize; /* Bytes of memory for in-memory hash */
233980	char zRank; / Name of rank function */
233981	char zRankArgs; / Arguments to rank function */
233982	int bSecureDelete; /* 'secure-delete' */
233983	int nDeleteMerge; /* 'deletemerge' */
233984	int bPrefixInsttoken; /* 'prefix-insttoken' */
233985
233986	/* If non-NULL, points to sqlite3_vtab.base.zErrmsg. Often NULL. */
233987	char **pzErrmsg;
233988
233989	#ifdef SQLITE_DEBUG
	@@ -233598,13 +233995,14 @@
233995	** the expected version if the 'secure-delete' option has ever been
233996	** set on the table. */
233997	#define FTS5_CURRENT_VERSION 4
233998	#define FTS5_CURRENT_VERSION_SECUREDELETE 5
233999
234000	#define FTS5_CONTENT_NORMAL 0
234001	#define FTS5_CONTENT_NONE 1
234002	#define FTS5_CONTENT_EXTERNAL 2
234003	#define FTS5_CONTENT_UNINDEXED 3
234004
234005	#define FTS5_DETAIL_FULL 0
234006	#define FTS5_DETAIL_NONE 1
234007	#define FTS5_DETAIL_COLUMNS 2
234008
	@@ -233838,11 +234236,18 @@
234236	static int sqlite3Fts5StructureTest(Fts5Index, void);
234237
234238	/*
234239	** Used by xInstToken():
234240	*/
234241	static int sqlite3Fts5IterToken(
234242	Fts5IndexIter *pIndexIter,
234243	const char *pToken, int nToken,
234244	i64 iRowid,
234245	int iCol,
234246	int iOff,
234247	const char *ppOut, int pnOut
234248	);
234249
234250	/*
234251	** Insert or remove data to or from the index. Each time a document is
234252	** added to or removed from the index, this function is called one or more
234253	** times.
	@@ -233972,20 +234377,17 @@
234377
234378	static Fts5Table sqlite3Fts5TableFromCsrid(Fts5Global, i64);
234379
234380	static int sqlite3Fts5FlushToDisk(Fts5Table*);
234381









234382	static void sqlite3Fts5ClearLocale(Fts5Config *pConfig);
234383	static void sqlite3Fts5SetLocale(Fts5Config pConfig, const char pLoc, int nLoc);
234384
234385	static int sqlite3Fts5IsLocaleValue(Fts5Config pConfig, sqlite3_value pVal);
234386	static int sqlite3Fts5DecodeLocaleValue(sqlite3_value *pVal,
234387	const char *ppText, int pnText, const char *ppLoc, int pnLoc
234388	);
234389
234390	/*
234391	** End of interface to code in fts5.c.
234392	**************************************************************************/
234393
	@@ -234063,11 +234465,11 @@
234465
234466	static int sqlite3Fts5DropAll(Fts5Config*);
234467	static int sqlite3Fts5CreateTable(Fts5Config, const char, const char, int, char *);
234468
234469	static int sqlite3Fts5StorageDelete(Fts5Storage p, i64, sqlite3_value*, int);
234470	static int sqlite3Fts5StorageContentInsert(Fts5Storage p, int, sqlite3_value, i64);
234471	static int sqlite3Fts5StorageIndexInsert(Fts5Storage p, sqlite3_value*, i64);
234472
234473	static int sqlite3Fts5StorageIntegrity(Fts5Storage *p, int iArg);
234474
234475	static int sqlite3Fts5StorageStmt(Fts5Storage p, int eStmt, sqlite3_stmt, char*);
	@@ -237270,10 +237672,11 @@
237672	const char zArg, / Argument to parse */
237673	char *pzErr / OUT: Error message */
237674	){
237675	int rc = SQLITE_OK;
237676	int nCmd = (int)strlen(zCmd);
237677
237678	if( sqlite3_strnicmp("prefix", zCmd, nCmd)==0 ){
237679	const int nByte = sizeof(int) * FTS5_MAX_PREFIX_INDEXES;
237680	const char *p;
237681	int bFirst = 1;
237682	if( pConfig->aPrefix==0 ){
	@@ -237388,10 +237791,20 @@
237791	}else{
237792	pConfig->bContentlessDelete = (zArg[0]=='1');
237793	}
237794	return rc;
237795	}
237796
237797	if( sqlite3_strnicmp("contentless_unindexed", zCmd, nCmd)==0 ){
237798	if( (zArg[0]!='0' && zArg[0]!='1') \|\| zArg[1]!='\0' ){
237799	*pzErr = sqlite3_mprintf("malformed contentless_delete=... directive");
237800	rc = SQLITE_ERROR;
237801	}else{
237802	pConfig->bContentlessUnindexed = (zArg[0]=='1');
237803	}
237804	return rc;
237805	}
237806
237807	if( sqlite3_strnicmp("content_rowid", zCmd, nCmd)==0 ){
237808	if( pConfig->zContentRowid ){
237809	*pzErr = sqlite3_mprintf("multiple content_rowid=... directives");
237810	rc = SQLITE_ERROR;
	@@ -237506,11 +237919,12 @@
237919
237920	static int fts5ConfigParseColumn(
237921	Fts5Config *p,
237922	char *zCol,
237923	char *zArg,
237924	char **pzErr,
237925	int *pbUnindexed
237926	){
237927	int rc = SQLITE_OK;
237928	if( 0==sqlite3_stricmp(zCol, FTS5_RANK_NAME)
237929	\|\| 0==sqlite3_stricmp(zCol, FTS5_ROWID_NAME)
237930	){
	@@ -237517,10 +237931,11 @@
237931	*pzErr = sqlite3_mprintf("reserved fts5 column name: %s", zCol);
237932	rc = SQLITE_ERROR;
237933	}else if( zArg ){
237934	if( 0==sqlite3_stricmp(zArg, "unindexed") ){
237935	p->abUnindexed[p->nCol] = 1;
237936	*pbUnindexed = 1;
237937	}else{
237938	*pzErr = sqlite3_mprintf("unrecognized column option: %s", zArg);
237939	rc = SQLITE_ERROR;
237940	}
237941	}
	@@ -237537,15 +237952,30 @@
237952	int rc = SQLITE_OK;
237953	Fts5Buffer buf = {0, 0, 0};
237954
237955	sqlite3Fts5BufferAppendPrintf(&rc, &buf, "T.%Q", p->zContentRowid);
237956	if( p->eContent!=FTS5_CONTENT_NONE ){
237957	assert( p->eContent==FTS5_CONTENT_EXTERNAL
237958	\|\| p->eContent==FTS5_CONTENT_NORMAL
237959	\|\| p->eContent==FTS5_CONTENT_UNINDEXED
237960	);
237961	for(i=0; i<p->nCol; i++){
237962	if( p->eContent==FTS5_CONTENT_EXTERNAL ){
237963	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.%Q", p->azCol[i]);
237964	}else if( p->eContent==FTS5_CONTENT_NORMAL \|\| p->abUnindexed[i] ){
237965	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.c%d", i);
237966	}else{
237967	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", NULL");
237968	}
237969	}
237970	}
237971	if( p->eContent==FTS5_CONTENT_NORMAL && p->bLocale ){
237972	for(i=0; i<p->nCol; i++){
237973	if( p->abUnindexed[i]==0 ){
237974	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", T.l%d", i);
237975	}else{
237976	sqlite3Fts5BufferAppendPrintf(&rc, &buf, ", NULL");
237977	}
237978	}
237979	}
237980
237981	assert( p->zContentExprlist==0 );
	@@ -237575,10 +238005,11 @@
238005	){
238006	int rc = SQLITE_OK; /* Return code */
238007	Fts5Config pRet; / New object to return */
238008	int i;
238009	sqlite3_int64 nByte;
238010	int bUnindexed = 0; /* True if there are one or more UNINDEXED */
238011
238012	ppOut = pRet = (Fts5Config)sqlite3_malloc(sizeof(Fts5Config));
238013	if( pRet==0 ) return SQLITE_NOMEM;
238014	memset(pRet, 0, sizeof(Fts5Config));
238015	pRet->pGlobal = pGlobal;
	@@ -237634,11 +238065,11 @@
238065	ALWAYS(zOne)?zOne:"",
238066	zTwo?zTwo:"",
238067	pzErr
238068	);
238069	}else{
238070	rc = fts5ConfigParseColumn(pRet, zOne, zTwo, pzErr, &bUnindexed);
238071	zOne = 0;
238072	}
238073	}
238074	}
238075
	@@ -237665,18 +238096,34 @@
238096	*pzErr = sqlite3_mprintf(
238097	"contentless_delete=1 is incompatible with columnsize=0"
238098	);
238099	rc = SQLITE_ERROR;
238100	}
238101
238102	/* We only allow contentless_unindexed=1 if the table is actually a
238103	** contentless one.
238104	*/
238105	if( rc==SQLITE_OK
238106	&& pRet->bContentlessUnindexed
238107	&& pRet->eContent!=FTS5_CONTENT_NONE
238108	){
238109	*pzErr = sqlite3_mprintf(
238110	"contentless_unindexed=1 requires a contentless table"
238111	);
238112	rc = SQLITE_ERROR;
238113	}
238114
238115	/* If no zContent option was specified, fill in the default values. */
238116	if( rc==SQLITE_OK && pRet->zContent==0 ){
238117	const char *zTail = 0;
238118	assert( pRet->eContent==FTS5_CONTENT_NORMAL
238119	\|\| pRet->eContent==FTS5_CONTENT_NONE
238120	);
238121	if( pRet->eContent==FTS5_CONTENT_NORMAL ){
238122	zTail = "content";
238123	}else if( bUnindexed && pRet->bContentlessUnindexed ){
238124	pRet->eContent = FTS5_CONTENT_UNINDEXED;
238125	zTail = "content";
238126	}else if( pRet->bColumnsize ){
238127	zTail = "docsize";
238128	}
238129
	@@ -238010,10 +238457,23 @@
238457	if( bVal<0 ){
238458	*pbBadkey = 1;
238459	}else{
238460	pConfig->bSecureDelete = (bVal ? 1 : 0);
238461	}
238462	}
238463
238464	else if( 0==sqlite3_stricmp(zKey, "insttoken") ){
238465	int bVal = -1;
238466	if( SQLITE_INTEGER==sqlite3_value_numeric_type(pVal) ){
238467	bVal = sqlite3_value_int(pVal);
238468	}
238469	if( bVal<0 ){
238470	*pbBadkey = 1;
238471	}else{
238472	pConfig->bPrefixInsttoken = (bVal ? 1 : 0);
238473	}
238474
238475	}else{
238476	*pbBadkey = 1;
238477	}
238478	return rc;
238479	}
	@@ -241145,11 +241605,11 @@
241605	&& memcmp(pT->pTerm, pToken, pT->nQueryTerm)==0
241606	){
241607	int rc = sqlite3Fts5PoslistWriterAppend(
241608	&pExpr->apExprPhrase[i]->poslist, &p->aPopulator[i].writer, p->iOff
241609	);
241610	if( rc==SQLITE_OK && (pExpr->pConfig->bTokendata \|\| pT->bPrefix) ){
241611	int iCol = p->iOff>>32;
241612	int iTokOff = p->iOff & 0x7FFFFFFF;
241613	rc = sqlite3Fts5IndexIterWriteTokendata(
241614	pT->pIter, pToken, nToken, iRowid, iCol, iTokOff
241615	);
	@@ -241338,19 +241798,18 @@
241798	pPhrase = pExpr->apExprPhrase[iPhrase];
241799	if( iToken<0 \|\| iToken>=pPhrase->nTerm ){
241800	return SQLITE_RANGE;
241801	}
241802	pTerm = &pPhrase->aTerm[iToken];
241803	if( pExpr->pConfig->bTokendata \|\| pTerm->bPrefix ){
241804	rc = sqlite3Fts5IterToken(
241805	pTerm->pIter, pTerm->pTerm, pTerm->nQueryTerm,
241806	iRowid, iCol, iOff+iToken, ppOut, pnOut
241807	);
241808	}else{
241809	*ppOut = pTerm->pTerm;
241810	*pnOut = pTerm->nFullTerm;

241811	}
241812	return rc;
241813	}
241814
241815	/*
	@@ -246847,10 +247306,15 @@
247306	if( nPercent>=pConfig->nDeleteMerge && nPercent>nBest ){
247307	iRet = ii;
247308	nBest = nPercent;
247309	}
247310	}
247311
247312	/* If pLvl is already the input level to an ongoing merge, look no
247313	** further for a merge candidate. The caller should be allowed to
247314	** continue merging from pLvl first. */
247315	if( pLvl->nMerge ) break;
247316	}
247317	}
247318	return iRet;
247319	}
247320
	@@ -248156,10 +248620,387 @@
248620	fts5BufferFree(&tmp);
248621	memset(&out.p[out.n], 0, FTS5_DATA_ZERO_PADDING);
248622	*p1 = out;
248623	}
248624
248625
248626	/*
248627	** Iterate through a range of entries in the FTS index, invoking the xVisit
248628	** callback for each of them.
248629	**
248630	** Parameter pToken points to an nToken buffer containing an FTS index term
248631	** (i.e. a document term with the preceding 1 byte index identifier -
248632	** FTS5_MAIN_PREFIX or similar). If bPrefix is true, then the call visits
248633	** all entries for terms that have pToken/nToken as a prefix. If bPrefix
248634	** is false, then only entries with pToken/nToken as the entire key are
248635	** visited.
248636	**
248637	** If the current table is a tokendata=1 table, then if bPrefix is true then
248638	** each index term is treated separately. However, if bPrefix is false, then
248639	** all index terms corresponding to pToken/nToken are collapsed into a single
248640	** term before the callback is invoked.
248641	**
248642	** The callback invoked for each entry visited is specified by paramter xVisit.
248643	** Each time it is invoked, it is passed a pointer to the Fts5Index object,
248644	** a copy of the 7th paramter to this function (pCtx) and a pointer to the
248645	** iterator that indicates the current entry. If the current entry is the
248646	** first with a new term (i.e. different from that of the previous entry,
248647	** including the very first term), then the final two parameters are passed
248648	** a pointer to the term and its size in bytes, respectively. If the current
248649	** entry is not the first associated with its term, these two parameters
248650	** are passed 0.
248651	**
248652	** If parameter pColset is not NULL, then it is used to filter entries before
248653	** the callback is invoked.
248654	*/
248655	static int fts5VisitEntries(
248656	Fts5Index p, / Fts5 index object */
248657	Fts5Colset pColset, / Columns filter to apply, or NULL */
248658	u8 pToken, / Buffer containing token */
248659	int nToken, /* Size of buffer pToken in bytes */
248660	int bPrefix, /* True for a prefix scan */
248661	void (xVisit)(Fts5Index, void pCtx, Fts5Iter pIter, const u8*, int),
248662	void pCtx / Passed as second argument to xVisit() */
248663	){
248664	const int flags = (bPrefix ? FTS5INDEX_QUERY_SCAN : 0)
248665	\| FTS5INDEX_QUERY_SKIPEMPTY
248666	\| FTS5INDEX_QUERY_NOOUTPUT;
248667	Fts5Iter p1 = 0; / Iterator used to gather data from index */
248668	int bNewTerm = 1;
248669	Fts5Structure *pStruct = fts5StructureRead(p);
248670
248671	fts5MultiIterNew(p, pStruct, flags, pColset, pToken, nToken, -1, 0, &p1);
248672	fts5IterSetOutputCb(&p->rc, p1);
248673	for( /* no-op */ ;
248674	fts5MultiIterEof(p, p1)==0;
248675	fts5MultiIterNext2(p, p1, &bNewTerm)
248676	){
248677	Fts5SegIter *pSeg = &p1->aSeg[ p1->aFirst[1].iFirst ];
248678	int nNew = 0;
248679	const u8 *pNew = 0;
248680
248681	p1->xSetOutputs(p1, pSeg);
248682	if( p->rc ) break;
248683
248684	if( bNewTerm ){
248685	nNew = pSeg->term.n;
248686	pNew = pSeg->term.p;
248687	if( nNew<nToken \|\| memcmp(pToken, pNew, nToken) ) break;
248688	}
248689
248690	xVisit(p, pCtx, p1, pNew, nNew);
248691	}
248692	fts5MultiIterFree(p1);
248693
248694	fts5StructureRelease(pStruct);
248695	return p->rc;
248696	}
248697
248698
248699	/*
248700	** Usually, a tokendata=1 iterator (struct Fts5TokenDataIter) accumulates an
248701	** array of these for each row it visits (so all iRowid fields are the same).
248702	** Or, for an iterator used by an "ORDER BY rank" query, it accumulates an
248703	** array of these for the entire query (in which case iRowid fields may take
248704	** a variety of values).
248705	**
248706	** Each instance in the array indicates the iterator (and therefore term)
248707	** associated with position iPos of rowid iRowid. This is used by the
248708	** xInstToken() API.
248709	**
248710	** iRowid:
248711	** Rowid for the current entry.
248712	**
248713	** iPos:
248714	** Position of current entry within row. In the usual ((iCol<<32)+iOff)
248715	** format (e.g. see macros FTS5_POS2COLUMN() and FTS5_POS2OFFSET()).
248716	**
248717	** iIter:
248718	** If the Fts5TokenDataIter iterator that the entry is part of is
248719	** actually an iterator (i.e. with nIter>0, not just a container for
248720	** Fts5TokenDataMap structures), then this variable is an index into
248721	** the apIter[] array. The corresponding term is that which the iterator
248722	** at apIter[iIter] currently points to.
248723	**
248724	** Or, if the Fts5TokenDataIter iterator is just a container object
248725	** (nIter==0), then iIter is an index into the term.p[] buffer where
248726	** the term is stored.
248727	**
248728	** nByte:
248729	** In the case where iIter is an index into term.p[], this variable
248730	** is the size of the term in bytes. If iIter is an index into apIter[],
248731	** this variable is unused.
248732	*/
248733	struct Fts5TokenDataMap {
248734	i64 iRowid; /* Row this token is located in */
248735	i64 iPos; /* Position of token */
248736	int iIter; /* Iterator token was read from */
248737	int nByte; /* Length of token in bytes (or 0) */
248738	};
248739
248740	/*
248741	** An object used to supplement Fts5Iter for tokendata=1 iterators.
248742	**
248743	** This object serves two purposes. The first is as a container for an array
248744	** of Fts5TokenDataMap structures, which are used to find the token required
248745	** when the xInstToken() API is used. This is done by the nMapAlloc, nMap and
248746	** aMap[] variables.
248747	*/
248748	struct Fts5TokenDataIter {
248749	int nMapAlloc; /* Allocated size of aMap[] in entries */
248750	int nMap; /* Number of valid entries in aMap[] */
248751	Fts5TokenDataMap aMap; / Array of (rowid+pos -> token) mappings */
248752
248753	/* The following are used for prefix-queries only. */
248754	Fts5Buffer terms;
248755
248756	/* The following are used for other full-token tokendata queries only. */
248757	int nIter;
248758	int nIterAlloc;
248759	Fts5PoslistReader *aPoslistReader;
248760	int *aPoslistToIter;
248761	Fts5Iter *apIter[1];
248762	};
248763
248764	/*
248765	** The two input arrays - a1[] and a2[] - are in sorted order. This function
248766	** merges the two arrays together and writes the result to output array
248767	** aOut[]. aOut[] is guaranteed to be large enough to hold the result.
248768	**
248769	** Duplicate entries are copied into the output. So the size of the output
248770	** array is always (n1+n2) entries.
248771	*/
248772	static void fts5TokendataMerge(
248773	Fts5TokenDataMap a1, int n1, / Input array 1 */
248774	Fts5TokenDataMap a2, int n2, / Input array 2 */
248775	Fts5TokenDataMap aOut / Output array */
248776	){
248777	int i1 = 0;
248778	int i2 = 0;
248779
248780	assert( n1>=0 && n2>=0 );
248781	while( i1<n1 \|\| i2<n2 ){
248782	Fts5TokenDataMap *pOut = &aOut[i1+i2];
248783	if( i2>=n2 \|\| (i1<n1 && (
248784	a1[i1].iRowid<a2[i2].iRowid
248785	\|\| (a1[i1].iRowid==a2[i2].iRowid && a1[i1].iPos<=a2[i2].iPos)
248786	))){
248787	memcpy(pOut, &a1[i1], sizeof(Fts5TokenDataMap));
248788	i1++;
248789	}else{
248790	memcpy(pOut, &a2[i2], sizeof(Fts5TokenDataMap));
248791	i2++;
248792	}
248793	}
248794	}
248795
248796
248797	/*
248798	** Append a mapping to the token-map belonging to object pT.
248799	*/
248800	static void fts5TokendataIterAppendMap(
248801	Fts5Index *p,
248802	Fts5TokenDataIter *pT,
248803	int iIter,
248804	int nByte,
248805	i64 iRowid,
248806	i64 iPos
248807	){
248808	if( p->rc==SQLITE_OK ){
248809	if( pT->nMap==pT->nMapAlloc ){
248810	int nNew = pT->nMapAlloc ? pT->nMapAlloc*2 : 64;
248811	int nAlloc = nNew * sizeof(Fts5TokenDataMap);
248812	Fts5TokenDataMap *aNew;
248813
248814	aNew = (Fts5TokenDataMap*)sqlite3_realloc(pT->aMap, nAlloc);
248815	if( aNew==0 ){
248816	p->rc = SQLITE_NOMEM;
248817	return;
248818	}
248819
248820	pT->aMap = aNew;
248821	pT->nMapAlloc = nNew;
248822	}
248823
248824	pT->aMap[pT->nMap].iRowid = iRowid;
248825	pT->aMap[pT->nMap].iPos = iPos;
248826	pT->aMap[pT->nMap].iIter = iIter;
248827	pT->aMap[pT->nMap].nByte = nByte;
248828	pT->nMap++;
248829	}
248830	}
248831
248832	/*
248833	** Sort the contents of the pT->aMap[] array.
248834	**
248835	** The sorting algorithm requries a malloc(). If this fails, an error code
248836	** is left in Fts5Index.rc before returning.
248837	*/
248838	static void fts5TokendataIterSortMap(Fts5Index p, Fts5TokenDataIter pT){
248839	Fts5TokenDataMap *aTmp = 0;
248840	int nByte = pT->nMap * sizeof(Fts5TokenDataMap);
248841
248842	aTmp = (Fts5TokenDataMap*)sqlite3Fts5MallocZero(&p->rc, nByte);
248843	if( aTmp ){
248844	Fts5TokenDataMap *a1 = pT->aMap;
248845	Fts5TokenDataMap *a2 = aTmp;
248846	i64 nHalf;
248847
248848	for(nHalf=1; nHalf<pT->nMap; nHalf=nHalf*2){
248849	int i1;
248850	for(i1=0; i1<pT->nMap; i1+=(nHalf*2)){
248851	int n1 = MIN(nHalf, pT->nMap-i1);
248852	int n2 = MIN(nHalf, pT->nMap-i1-n1);
248853	fts5TokendataMerge(&a1[i1], n1, &a1[i1+n1], n2, &a2[i1]);
248854	}
248855	SWAPVAL(Fts5TokenDataMap*, a1, a2);
248856	}
248857
248858	if( a1!=pT->aMap ){
248859	memcpy(pT->aMap, a1, pT->nMap*sizeof(Fts5TokenDataMap));
248860	}
248861	sqlite3_free(aTmp);
248862
248863	#ifdef SQLITE_DEBUG
248864	{
248865	int ii;
248866	for(ii=1; ii<pT->nMap; ii++){
248867	Fts5TokenDataMap *p1 = &pT->aMap[ii-1];
248868	Fts5TokenDataMap *p2 = &pT->aMap[ii];
248869	assert( p1->iRowid<p2->iRowid
248870	\|\| (p1->iRowid==p2->iRowid && p1->iPos<=p2->iPos)
248871	);
248872	}
248873	}
248874	#endif
248875	}
248876	}
248877
248878	/*
248879	** Delete an Fts5TokenDataIter structure and its contents.
248880	*/
248881	static void fts5TokendataIterDelete(Fts5TokenDataIter *pSet){
248882	if( pSet ){
248883	int ii;
248884	for(ii=0; ii<pSet->nIter; ii++){
248885	fts5MultiIterFree(pSet->apIter[ii]);
248886	}
248887	fts5BufferFree(&pSet->terms);
248888	sqlite3_free(pSet->aPoslistReader);
248889	sqlite3_free(pSet->aMap);
248890	sqlite3_free(pSet);
248891	}
248892	}
248893
248894
248895	/*
248896	** fts5VisitEntries() context object used by fts5SetupPrefixIterTokendata()
248897	** to pass data to prefixIterSetupTokendataCb().
248898	*/
248899	typedef struct TokendataSetupCtx TokendataSetupCtx;
248900	struct TokendataSetupCtx {
248901	Fts5TokenDataIter pT; / Object being populated with mappings */
248902	int iTermOff; /* Offset of current term in terms.p[] */
248903	int nTermByte; /* Size of current term in bytes */
248904	};
248905
248906	/*
248907	** fts5VisitEntries() callback used by fts5SetupPrefixIterTokendata(). This
248908	** callback adds an entry to the Fts5TokenDataIter.aMap[] array for each
248909	** position in the current position-list. It doesn't matter that some of
248910	** these may be out of order - they will be sorted later.
248911	*/
248912	static void prefixIterSetupTokendataCb(
248913	Fts5Index *p,
248914	void *pCtx,
248915	Fts5Iter *p1,
248916	const u8 *pNew,
248917	int nNew
248918	){
248919	TokendataSetupCtx pSetup = (TokendataSetupCtx)pCtx;
248920	int iPosOff = 0;
248921	i64 iPos = 0;
248922
248923	if( pNew ){
248924	pSetup->nTermByte = nNew-1;
248925	pSetup->iTermOff = pSetup->pT->terms.n;
248926	fts5BufferAppendBlob(&p->rc, &pSetup->pT->terms, nNew-1, pNew+1);
248927	}
248928
248929	while( 0==sqlite3Fts5PoslistNext64(
248930	p1->base.pData, p1->base.nData, &iPosOff, &iPos
248931	) ){
248932	fts5TokendataIterAppendMap(p,
248933	pSetup->pT, pSetup->iTermOff, pSetup->nTermByte, p1->base.iRowid, iPos
248934	);
248935	}
248936	}
248937
248938
248939	/*
248940	** Context object passed by fts5SetupPrefixIter() to fts5VisitEntries().
248941	*/
248942	typedef struct PrefixSetupCtx PrefixSetupCtx;
248943	struct PrefixSetupCtx {
248944	void (xMerge)(Fts5Index, Fts5Buffer, int, Fts5Buffer);
248945	void (xAppend)(Fts5Index, u64, Fts5Iter, Fts5Buffer);
248946	i64 iLastRowid;
248947	int nMerge;
248948	Fts5Buffer *aBuf;
248949	int nBuf;
248950	Fts5Buffer doclist;
248951	TokendataSetupCtx *pTokendata;
248952	};
248953
248954	/*
248955	** fts5VisitEntries() callback used by fts5SetupPrefixIter()
248956	*/
248957	static void prefixIterSetupCb(
248958	Fts5Index *p,
248959	void *pCtx,
248960	Fts5Iter *p1,
248961	const u8 *pNew,
248962	int nNew
248963	){
248964	PrefixSetupCtx pSetup = (PrefixSetupCtx)pCtx;
248965	const int nMerge = pSetup->nMerge;
248966
248967	if( p1->base.nData>0 ){
248968	if( p1->base.iRowid<=pSetup->iLastRowid && pSetup->doclist.n>0 ){
248969	int i;
248970	for(i=0; p->rc==SQLITE_OK && pSetup->doclist.n; i++){
248971	int i1 = i*nMerge;
248972	int iStore;
248973	assert( i1+nMerge<=pSetup->nBuf );
248974	for(iStore=i1; iStore<i1+nMerge; iStore++){
248975	if( pSetup->aBuf[iStore].n==0 ){
248976	fts5BufferSwap(&pSetup->doclist, &pSetup->aBuf[iStore]);
248977	fts5BufferZero(&pSetup->doclist);
248978	break;
248979	}
248980	}
248981	if( iStore==i1+nMerge ){
248982	pSetup->xMerge(p, &pSetup->doclist, nMerge, &pSetup->aBuf[i1]);
248983	for(iStore=i1; iStore<i1+nMerge; iStore++){
248984	fts5BufferZero(&pSetup->aBuf[iStore]);
248985	}
248986	}
248987	}
248988	pSetup->iLastRowid = 0;
248989	}
248990
248991	pSetup->xAppend(
248992	p, (u64)p1->base.iRowid-(u64)pSetup->iLastRowid, p1, &pSetup->doclist
248993	);
248994	pSetup->iLastRowid = p1->base.iRowid;
248995	}
248996
248997	if( pSetup->pTokendata ){
248998	prefixIterSetupTokendataCb(p, (void*)pSetup->pTokendata, p1, pNew, nNew);
248999	}
249000	}
249001
249002	static void fts5SetupPrefixIter(
249003	Fts5Index p, / Index to read from */
249004	int bDesc, /* True for "ORDER BY rowid DESC" */
249005	int iIdx, /* Index to scan for data */
249006	u8 pToken, / Buffer containing prefix to match */
	@@ -248166,137 +249007,89 @@
249007	int nToken, /* Size of buffer pToken in bytes */
249008	Fts5Colset pColset, / Restrict matches to these columns */
249009	Fts5Iter *ppIter / OUT: New iterator */
249010	){
249011	Fts5Structure *pStruct;
249012	PrefixSetupCtx s;
249013	TokendataSetupCtx s2;

249014
249015	memset(&s, 0, sizeof(s));
249016	memset(&s2, 0, sizeof(s2));
249017
249018	s.nMerge = 1;
249019	s.iLastRowid = 0;
249020	s.nBuf = 32;
249021	if( iIdx==0
249022	&& p->pConfig->eDetail==FTS5_DETAIL_FULL
249023	&& p->pConfig->bPrefixInsttoken
249024	){
249025	s.pTokendata = &s2;
249026	s2.pT = (Fts5TokenDataIter)fts5IdxMalloc(p, sizeof(s2.pT));
249027	}
249028
249029	if( p->pConfig->eDetail==FTS5_DETAIL_NONE ){
249030	s.xMerge = fts5MergeRowidLists;
249031	s.xAppend = fts5AppendRowid;
249032	}else{
249033	s.nMerge = FTS5_MERGE_NLIST-1;
249034	s.nBuf = s.nMerge8; / Sufficient to merge (16^8)==(2^32) lists */
249035	s.xMerge = fts5MergePrefixLists;
249036	s.xAppend = fts5AppendPoslist;
249037	}
249038
249039	s.aBuf = (Fts5Buffer)fts5IdxMalloc(p, sizeof(Fts5Buffer)s.nBuf);
249040	pStruct = fts5StructureRead(p);
249041	assert( p->rc!=SQLITE_OK \|\| (s.aBuf && pStruct) );
249042
249043	if( p->rc==SQLITE_OK ){
249044	void pCtx = (void)&s;


249045	int i;


249046	Fts5Data *pData;




249047
249048	/* If iIdx is non-zero, then it is the number of a prefix-index for
249049	** prefixes 1 character longer than the prefix being queried for. That
249050	** index contains all the doclists required, except for the one
249051	** corresponding to the prefix itself. That one is extracted from the
249052	** main term index here. */
249053	if( iIdx!=0 ){


249054	pToken[0] = FTS5_MAIN_PREFIX;
249055	fts5VisitEntries(p, pColset, pToken, nToken, 0, prefixIterSetupCb, pCtx);













249056	}
249057
249058	pToken[0] = FTS5_MAIN_PREFIX + iIdx;
249059	fts5VisitEntries(p, pColset, pToken, nToken, 1, prefixIterSetupCb, pCtx);
249060
249061	assert( (s.nBuf%s.nMerge)==0 );
249062	for(i=0; i<s.nBuf; i+=s.nMerge){










































249063	int iFree;
249064	if( p->rc==SQLITE_OK ){
249065	s.xMerge(p, &s.doclist, s.nMerge, &s.aBuf[i]);
249066	}
249067	for(iFree=i; iFree<i+s.nMerge; iFree++){
249068	fts5BufferFree(&s.aBuf[iFree]);
249069	}
249070	}

249071
249072	pData = fts5IdxMalloc(p, sizeof(*pData)+s.doclist.n+FTS5_DATA_ZERO_PADDING);
249073	if( pData ){
249074	pData->p = (u8*)&pData[1];
249075	pData->nn = pData->szLeaf = s.doclist.n;
249076	if( s.doclist.n ) memcpy(pData->p, s.doclist.p, s.doclist.n);
249077	fts5MultiIterNew2(p, pData, bDesc, ppIter);
249078	}
249079
249080	if( p->rc==SQLITE_OK && s.pTokendata ){
249081	fts5TokendataIterSortMap(p, s2.pT);
249082	(*ppIter)->pTokenDataIter = s2.pT;
249083	s2.pT = 0;
249084	}
249085	}
249086
249087	fts5TokendataIterDelete(s2.pT);
249088	fts5BufferFree(&s.doclist);
249089	fts5StructureRelease(pStruct);
249090	sqlite3_free(s.aBuf);
249091	}
249092
249093
249094	/*
249095	** Indicate that all subsequent calls to sqlite3Fts5IndexWrite() pertain
	@@ -248546,42 +249339,10 @@
249339	static void fts5SegIterSetEOF(Fts5SegIter *pSeg){
249340	fts5DataRelease(pSeg->pLeaf);
249341	pSeg->pLeaf = 0;
249342	}
249343
































249344	/*
249345	** This function appends iterator pAppend to Fts5TokenDataIter pIn and
249346	** returns the result.
249347	*/
249348	static Fts5TokenDataIter *fts5AppendTokendataIter(
	@@ -248614,58 +249375,10 @@
249375	assert( pRet==0 \|\| pRet->nIter<=pRet->nIterAlloc );
249376
249377	return pRet;
249378	}
249379
















































249380	/*
249381	** The iterator passed as the only argument must be a tokendata=1 iterator
249382	** (pIter->pTokenDataIter!=0). This function sets the iterator output
249383	** variables (pIter->base.*) according to the contents of the current
249384	** row.
	@@ -248702,11 +249415,11 @@
249415	int eDetail = pIter->pIndex->pConfig->eDetail;
249416	pIter->base.bEof = 0;
249417	pIter->base.iRowid = iRowid;
249418
249419	if( nHit==1 && eDetail==FTS5_DETAIL_FULL ){
249420	fts5TokendataIterAppendMap(pIter->pIndex, pT, iMin, 0, iRowid, -1);
249421	}else
249422	if( nHit>1 && eDetail!=FTS5_DETAIL_NONE ){
249423	int nReader = 0;
249424	int nByte = 0;
249425	i64 iPrev = 0;
	@@ -248955,10 +249668,11 @@
249668
249669	if( p->rc==SQLITE_OK ){
249670	pRet = fts5MultiIterAlloc(p, 0);
249671	}
249672	if( pRet ){
249673	pRet->nSeg = 0;
249674	pRet->pTokenDataIter = pSet;
249675	if( pSet ){
249676	fts5IterSetOutputsTokendata(pRet);
249677	}else{
249678	pRet->base.bEof = 1;
	@@ -248969,11 +249683,10 @@
249683
249684	fts5StructureRelease(pStruct);
249685	fts5BufferFree(&bSeek);
249686	return pRet;
249687	}

249688
249689	/*
249690	** Open a new iterator to iterate though all rowid that match the
249691	** specified token or token prefix.
249692	*/
	@@ -248995,10 +249708,15 @@
249708	int iIdx = 0; /* Index to search */
249709	int iPrefixIdx = 0; /* +1 prefix index */
249710	int bTokendata = pConfig->bTokendata;
249711	if( nToken>0 ) memcpy(&buf.p[1], pToken, nToken);
249712
249713	/* The NOTOKENDATA flag is set when each token in a tokendata=1 table
249714	** should be treated individually, instead of merging all those with
249715	** a common prefix into a single entry. This is used, for example, by
249716	** queries performed as part of an integrity-check, or by the fts5vocab
249717	** module. */
249718	if( flags & (FTS5INDEX_QUERY_NOTOKENDATA\|FTS5INDEX_QUERY_SCAN) ){
249719	bTokendata = 0;
249720	}
249721
249722	/* Figure out which index to search and set iIdx accordingly. If this
	@@ -249025,11 +249743,11 @@
249743	if( nIdxChar==nChar+1 ) iPrefixIdx = iIdx;
249744	}
249745	}
249746
249747	if( bTokendata && iIdx==0 ){
249748	buf.p[0] = FTS5_MAIN_PREFIX;
249749	pRet = fts5SetupTokendataIter(p, buf.p, nToken+1, pColset);
249750	}else if( iIdx<=pConfig->nPrefix ){
249751	/* Straight index lookup */
249752	Fts5Structure *pStruct = fts5StructureRead(p);
249753	buf.p[0] = (u8)(FTS5_MAIN_PREFIX + iIdx);
	@@ -249038,11 +249756,11 @@
249756	pColset, buf.p, nToken+1, -1, 0, &pRet
249757	);
249758	fts5StructureRelease(pStruct);
249759	}
249760	}else{
249761	/* Scan multiple terms in the main index for a prefix query. */
249762	int bDesc = (flags & FTS5INDEX_QUERY_DESC)!=0;
249763	fts5SetupPrefixIter(p, bDesc, iPrefixIdx, buf.p, nToken+1, pColset,&pRet);
249764	if( pRet==0 ){
249765	assert( p->rc!=SQLITE_OK );
249766	}else{
	@@ -249074,11 +249792,12 @@
249792	** Move to the next matching rowid.
249793	*/
249794	static int sqlite3Fts5IterNext(Fts5IndexIter *pIndexIter){
249795	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249796	assert( pIter->pIndex->rc==SQLITE_OK );
249797	if( pIter->nSeg==0 ){
249798	assert( pIter->pTokenDataIter );
249799	fts5TokendataIterNext(pIter, 0, 0);
249800	}else{
249801	fts5MultiIterNext(pIter->pIndex, pIter, 0, 0);
249802	}
249803	return fts5IndexReturn(pIter->pIndex);
	@@ -249111,11 +249830,12 @@
249830	** definition of "at or after" depends on whether this iterator iterates
249831	** in ascending or descending rowid order.
249832	*/
249833	static int sqlite3Fts5IterNextFrom(Fts5IndexIter *pIndexIter, i64 iMatch){
249834	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249835	if( pIter->nSeg==0 ){
249836	assert( pIter->pTokenDataIter );
249837	fts5TokendataIterNext(pIter, 1, iMatch);
249838	}else{
249839	fts5MultiIterNextFrom(pIter->pIndex, pIter, iMatch);
249840	}
249841	return fts5IndexReturn(pIter->pIndex);
	@@ -249129,32 +249849,87 @@
249849	const char z = (const char)fts5MultiIterTerm((Fts5Iter*)pIndexIter, &n);
249850	assert_nc( z \|\| n<=1 );
249851	*pn = n-1;
249852	return (z ? &z[1] : 0);
249853	}
249854
249855	/*
249856	** pIter is a prefix query. This function populates pIter->pTokenDataIter
249857	** with an Fts5TokenDataIter object containing mappings for all rows
249858	** matched by the query.
249859	*/
249860	static int fts5SetupPrefixIterTokendata(
249861	Fts5Iter *pIter,
249862	const char pToken, / Token prefix to search for */
249863	int nToken /* Size of pToken in bytes */
249864	){
249865	Fts5Index *p = pIter->pIndex;
249866	Fts5Buffer token = {0, 0, 0};
249867	TokendataSetupCtx ctx;
249868
249869	memset(&ctx, 0, sizeof(ctx));
249870
249871	fts5BufferGrow(&p->rc, &token, nToken+1);
249872	ctx.pT = (Fts5TokenDataIter)sqlite3Fts5MallocZero(&p->rc, sizeof(ctx.pT));
249873
249874	if( p->rc==SQLITE_OK ){
249875
249876	/* Fill in the token prefix to search for */
249877	token.p[0] = FTS5_MAIN_PREFIX;
249878	memcpy(&token.p[1], pToken, nToken);
249879	token.n = nToken+1;
249880
249881	fts5VisitEntries(
249882	p, 0, token.p, token.n, 1, prefixIterSetupTokendataCb, (void*)&ctx
249883	);
249884
249885	fts5TokendataIterSortMap(p, ctx.pT);
249886	}
249887
249888	if( p->rc==SQLITE_OK ){
249889	pIter->pTokenDataIter = ctx.pT;
249890	}else{
249891	fts5TokendataIterDelete(ctx.pT);
249892	}
249893	fts5BufferFree(&token);
249894
249895	return fts5IndexReturn(p);
249896	}
249897
249898	/*
249899	** This is used by xInstToken() to access the token at offset iOff, column
249900	** iCol of row iRowid. The token is returned via output variables *ppOut
249901	** and *pnOut. The iterator passed as the first argument must be a tokendata=1
249902	** iterator (pIter->pTokenDataIter!=0).
249903	**
249904	** pToken/nToken:
249905	*/
249906	static int sqlite3Fts5IterToken(
249907	Fts5IndexIter *pIndexIter,
249908	const char *pToken, int nToken,
249909	i64 iRowid,
249910	int iCol,
249911	int iOff,
249912	const char *ppOut, int pnOut
249913	){
249914	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249915	Fts5TokenDataIter *pT = pIter->pTokenDataIter;

249916	i64 iPos = (((i64)iCol)<<32) + iOff;
249917	Fts5TokenDataMap *aMap = 0;
249918	int i1 = 0;
249919	int i2 = 0;
249920	int iTest = 0;
249921
249922	assert( pT \|\| (pToken && pIter->nSeg>0) );
249923	if( pT==0 ){
249924	int rc = fts5SetupPrefixIterTokendata(pIter, pToken, nToken);
249925	if( rc!=SQLITE_OK ) return rc;
249926	pT = pIter->pTokenDataIter;
249927	}
249928
249929	i2 = pT->nMap;
249930	aMap = pT->aMap;
249931
249932	while( i2>i1 ){
249933	iTest = (i1 + i2) / 2;
249934
249935	if( aMap[iTest].iRowid<iRowid ){
	@@ -249174,13 +249949,19 @@
249949	}
249950	}
249951	}
249952
249953	if( i2>i1 ){
249954	if( pIter->nSeg==0 ){
249955	Fts5Iter *pMap = pT->apIter[aMap[iTest].iIter];
249956	ppOut = (const char)pMap->aSeg[0].term.p+1;
249957	*pnOut = pMap->aSeg[0].term.n-1;
249958	}else{
249959	Fts5TokenDataMap *p = &aMap[iTest];
249960	ppOut = (const char)&pT->terms.p[p->iIter];
249961	*pnOut = aMap[iTest].nByte;
249962	}
249963	}
249964
249965	return SQLITE_OK;
249966	}
249967
	@@ -249188,11 +249969,13 @@
249969	** Clear any existing entries from the token-map associated with the
249970	** iterator passed as the only argument.
249971	*/
249972	static void sqlite3Fts5IndexIterClearTokendata(Fts5IndexIter *pIndexIter){
249973	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249974	if( pIter && pIter->pTokenDataIter
249975	&& (pIter->nSeg==0 \|\| pIter->pIndex->pConfig->eDetail!=FTS5_DETAIL_FULL)
249976	){
249977	pIter->pTokenDataIter->nMap = 0;
249978	}
249979	}
249980
249981	/*
	@@ -249208,21 +249991,33 @@
249991	i64 iRowid, int iCol, int iOff
249992	){
249993	Fts5Iter pIter = (Fts5Iter)pIndexIter;
249994	Fts5TokenDataIter *pT = pIter->pTokenDataIter;
249995	Fts5Index *p = pIter->pIndex;
249996	i64 iPos = (((i64)iCol)<<32) + iOff;
249997
249998	assert( p->pConfig->eDetail!=FTS5_DETAIL_FULL );
249999	assert( pIter->pTokenDataIter \|\| pIter->nSeg>0 );
250000	if( pIter->nSeg>0 ){
250001	/* This is a prefix term iterator. */
250002	if( pT==0 ){
250003	pT = (Fts5TokenDataIter)sqlite3Fts5MallocZero(&p->rc, sizeof(pT));
250004	pIter->pTokenDataIter = pT;
250005	}
250006	if( pT ){
250007	fts5TokendataIterAppendMap(p, pT, pT->terms.n, nToken, iRowid, iPos);
250008	fts5BufferAppendBlob(&p->rc, &pT->terms, nToken, (const u8*)pToken);
250009	}
250010	}else{
250011	int ii;
250012	for(ii=0; ii<pT->nIter; ii++){
250013	Fts5Buffer *pTerm = &pT->apIter[ii]->aSeg[0].term;
250014	if( nToken==pTerm->n-1 && memcmp(pToken, pTerm->p+1, nToken)==0 ) break;
250015	}
250016	if( ii<pT->nIter ){
250017	fts5TokendataIterAppendMap(p, pT, ii, 0, iRowid, iPos);
250018	}
250019	}
250020	return fts5IndexReturn(p);
250021	}
250022
250023	/*
	@@ -250771,11 +251566,11 @@
251566	return rc;
251567	}
251568
251569	/*
251570	** We must have a single struct=? constraint that will be passed through
251571	** into the xFilter method. If there is no valid struct=? constraint,
251572	** then return an SQLITE_CONSTRAINT error.
251573	*/
251574	static int fts5structBestIndexMethod(
251575	sqlite3_vtab *tab,
251576	sqlite3_index_info *pIdxInfo
	@@ -251113,12 +251908,22 @@
251908	i64 iNextId; /* Used to allocate unique cursor ids */
251909	Fts5Auxiliary pAux; / First in list of all aux. functions */
251910	Fts5TokenizerModule pTok; / First in list of all tokenizer modules */
251911	Fts5TokenizerModule pDfltTok; / Default tokenizer module */
251912	Fts5Cursor pCsr; / First in list of all open cursors */
251913	u32 aLocaleHdr[4];
251914	};
251915
251916	/*
251917	** Size of header on fts5_locale() values. And macro to access a buffer
251918	** containing a copy of the header from an Fts5Config pointer.
251919	*/
251920	#define FTS5_LOCALE_HDR_SIZE ((int)sizeof( ((Fts5Global*)0)->aLocaleHdr ))
251921	#define FTS5_LOCALE_HDR(pConfig) ((const u8*)(pConfig->pGlobal->aLocaleHdr))
251922
251923	#define FTS5_INSTTOKEN_SUBTYPE 73
251924
251925	/*
251926	** Each auxiliary function registered with the FTS5 module is represented
251927	** by an object of the following type. All such objects are stored as part
251928	** of the Fts5Global.pAux list.
251929	*/
	@@ -251277,16 +252082,10 @@
252082	#define FTS5CSR_REQUIRE_POSLIST 0x40
252083
252084	#define BitFlagAllTest(x,y) (((x) & (y))==(y))
252085	#define BitFlagTest(x,y) (((x) & (y))!=0)
252086






252087
252088	/*
252089	** Macros to Set(), Clear() and Test() cursor flags.
252090	*/
252091	#define CsrFlagSet(pCsr, flag) ((pCsr)->csrflags \|= (flag))
	@@ -251359,14 +252158,20 @@
252158	#else
252159	# define fts5CheckTransactionState(x,y,z)
252160	#endif
252161
252162	/*
252163	** Return true if pTab is a contentless table. If parameter bIncludeUnindexed
252164	** is true, this includes contentless tables that store UNINDEXED columns
252165	** only.
252166	*/
252167	static int fts5IsContentless(Fts5FullTable *pTab, int bIncludeUnindexed){
252168	int eContent = pTab->p.pConfig->eContent;
252169	return (
252170	eContent==FTS5_CONTENT_NONE
252171	\|\| (bIncludeUnindexed && eContent==FTS5_CONTENT_UNINDEXED)
252172	);
252173	}
252174
252175	/*
252176	** Delete a virtual table handle allocated by fts5InitVtab().
252177	*/
	@@ -251653,10 +252458,11 @@
252458	){
252459	/* A MATCH operator or equivalent */
252460	if( p->usable==0 \|\| iCol<0 ){
252461	/* As there exists an unusable MATCH constraint this is an
252462	** unusable plan. Return SQLITE_CONSTRAINT. */
252463	idxStr[iIdxStr] = 0;
252464	return SQLITE_CONSTRAINT;
252465	}else{
252466	if( iCol==nCol+1 ){
252467	if( bSeenRank ) continue;
252468	idxStr[iIdxStr++] = 'r';
	@@ -252286,11 +253092,11 @@
253092	** Arrange for subsequent calls to sqlite3Fts5Tokenize() to use the locale
253093	** specified by pLocale/nLocale. The buffer indicated by pLocale must remain
253094	** valid until after the final call to sqlite3Fts5Tokenize() that will use
253095	** the locale.
253096	*/
253097	static void sqlite3Fts5SetLocale(
253098	Fts5Config *pConfig,
253099	const char *zLocale,
253100	int nLocale
253101	){
253102	Fts5TokenizerConfig *pT = &pConfig->t;
	@@ -252297,141 +253103,88 @@
253103	pT->pLocale = zLocale;
253104	pT->nLocale = nLocale;
253105	}
253106
253107	/*
253108	** Clear any locale configured by an earlier call to sqlite3Fts5SetLocale().

253109	*/
253110	static void sqlite3Fts5ClearLocale(Fts5Config *pConfig){
253111	sqlite3Fts5SetLocale(pConfig, 0, 0);
253112	}
253113
253114	/*
253115	** Return true if the value passed as the only argument is an
253116	** fts5_locale() value.









































253117	*/
253118	static int sqlite3Fts5IsLocaleValue(Fts5Config pConfig, sqlite3_value pVal){
253119	int ret = 0;















253120	if( sqlite3_value_type(pVal)==SQLITE_BLOB ){
253121	/* Call sqlite3_value_bytes() after sqlite3_value_blob() in this case.
253122	** If the blob was created using zeroblob(), then sqlite3_value_blob()
253123	** may call malloc(). If this malloc() fails, then the values returned
253124	** by both value_blob() and value_bytes() will be 0. If value_bytes() were
253125	** called first, then the NULL pointer returned by value_blob() might
253126	** be dereferenced. */



253127	const u8 *pBlob = sqlite3_value_blob(pVal);
253128	int nBlob = sqlite3_value_bytes(pVal);
253129	if( nBlob>FTS5_LOCALE_HDR_SIZE
253130	&& 0==memcmp(pBlob, FTS5_LOCALE_HDR(pConfig), FTS5_LOCALE_HDR_SIZE)
253131	){
253132	ret = 1;
253133	}
253134	}
253135	return ret;
253136	}
253137
253138	/*
253139	** Value pVal is guaranteed to be an fts5_locale() value, according to
253140	** sqlite3Fts5IsLocaleValue(). This function extracts the text and locale
253141	** from the value and returns them separately.
253142	**
253143	** If successful, SQLITE_OK is returned and (ppText) and (ppLoc) set
253144	** to point to buffers containing the text and locale, as utf-8,
253145	** respectively. In this case output parameters (pnText) and (pnLoc) are
253146	** set to the sizes in bytes of these two buffers.
253147	**
253148	** Or, if an error occurs, then an SQLite error code is returned. The final
253149	** value of the four output parameters is undefined in this case.
253150	*/
253151	static int sqlite3Fts5DecodeLocaleValue(
253152	sqlite3_value *pVal,
253153	const char **ppText,
253154	int *pnText,
253155	const char **ppLoc,
253156	int *pnLoc
253157	){
253158	const char *p = sqlite3_value_blob(pVal);
253159	int n = sqlite3_value_bytes(pVal);
253160	int nLoc = 0;
253161
253162	assert( sqlite3_value_type(pVal)==SQLITE_BLOB );
253163	assert( n>FTS5_LOCALE_HDR_SIZE );
253164
253165	for(nLoc=FTS5_LOCALE_HDR_SIZE; p[nLoc]; nLoc++){
253166	if( nLoc==(n-1) ){
253167	return SQLITE_MISMATCH;
253168	}
253169	}
253170	*ppLoc = &p[FTS5_LOCALE_HDR_SIZE];
253171	*pnLoc = nLoc - FTS5_LOCALE_HDR_SIZE;
253172
253173	*ppText = &p[nLoc+1];
253174	*pnText = n - nLoc - 1;
253175	return SQLITE_OK;
253176	}
253177
253178	/*
253179	** Argument pVal is the text of a full-text search expression. It may or
253180	** may not have been wrapped by fts5_locale(). This function extracts
253181	** the text of the expression, and sets output variable (*pzText) to
253182	** point to a nul-terminated buffer containing the expression.
253183	**
253184	** If pVal was an fts5_locale() value, then sqlite3Fts5SetLocale() is called
253185	** to set the tokenizer to use the specified locale.
253186	**
253187	** If output variable (*pbFreeAndReset) is set to true, then the caller
253188	** is required to (a) call sqlite3Fts5ClearLocale() to reset the tokenizer
253189	** locale, and (b) call sqlite3_free() to free (*pzText).
253190	*/
	@@ -252439,28 +253192,26 @@
253192	Fts5Config pConfig, / Fts5 configuration */
253193	sqlite3_value pVal, / Value to extract expression text from */
253194	char *pzText, / OUT: nul-terminated buffer of text */
253195	int pbFreeAndReset / OUT: Free (pzText) and clear locale /
253196	){


253197	int rc = SQLITE_OK;
253198
253199	if( sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
253200	const char *pText = 0;
253201	int nText = 0;
253202	const char *pLoc = 0;
253203	int nLoc = 0;
253204	rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
253205	pzText = sqlite3Fts5Mprintf(&rc, "%.s", nText, pText);
253206	if( rc==SQLITE_OK ){
253207	sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
253208	}
253209	*pbFreeAndReset = 1;
253210	}else{
253211	pzText = (char)sqlite3_value_text(pVal);
253212	*pbFreeAndReset = 0;
253213	}
253214
253215	return rc;
253216	}
253217
	@@ -252493,10 +253244,11 @@
253244	sqlite3_value pRowidEq = 0; / rowid = ? expression (or NULL) */
253245	sqlite3_value pRowidLe = 0; / rowid <= ? expression (or NULL) */
253246	sqlite3_value pRowidGe = 0; / rowid >= ? expression (or NULL) */
253247	int iCol; /* Column on LHS of MATCH operator */
253248	char **pzErrmsg = pConfig->pzErrmsg;
253249	int bPrefixInsttoken = pConfig->bPrefixInsttoken;
253250	int i;
253251	int iIdxStr = 0;
253252	Fts5Expr *pExpr = 0;
253253
253254	assert( pConfig->bLock==0 );
	@@ -252528,10 +253280,13 @@
253280	int bInternal = 0;
253281
253282	rc = fts5ExtractExprText(pConfig, apVal[i], &zText, &bFreeAndReset);
253283	if( rc!=SQLITE_OK ) goto filter_out;
253284	if( zText==0 ) zText = "";
253285	if( sqlite3_value_subtype(apVal[i])==FTS5_INSTTOKEN_SUBTYPE ){
253286	pConfig->bPrefixInsttoken = 1;
253287	}
253288
253289	iCol = 0;
253290	do{
253291	iCol = iCol*10 + (idxStr[iIdxStr]-'0');
253292	iIdxStr++;
	@@ -252668,10 +253423,11 @@
253423	}
253424
253425	filter_out:
253426	sqlite3Fts5ExprFree(pExpr);
253427	pConfig->pzErrmsg = pzErrmsg;
253428	pConfig->bPrefixInsttoken = bPrefixInsttoken;
253429	return rc;
253430	}
253431
253432	/*
253433	** This is the xEof method of the virtual table. SQLite calls this
	@@ -252808,11 +253564,11 @@
253564	}else{
253565	rc = sqlite3Fts5StorageDeleteAll(pTab->pStorage);
253566	}
253567	bLoadConfig = 1;
253568	}else if( 0==sqlite3_stricmp("rebuild", zCmd) ){
253569	if( fts5IsContentless(pTab, 1) ){
253570	fts5SetVtabError(pTab,
253571	"'rebuild' may not be used with a contentless fts5 table"
253572	);
253573	rc = SQLITE_ERROR;
253574	}else{
	@@ -252877,17 +253633,78 @@
253633	sqlite3_value **apVal,
253634	i64 *piRowid
253635	){
253636	int rc = *pRc;
253637	if( rc==SQLITE_OK ){
253638	rc = sqlite3Fts5StorageContentInsert(pTab->pStorage, 0, apVal, piRowid);
253639	}
253640	if( rc==SQLITE_OK ){
253641	rc = sqlite3Fts5StorageIndexInsert(pTab->pStorage, apVal, *piRowid);
253642	}
253643	*pRc = rc;
253644	}
253645
253646	/*
253647	**
253648	** This function is called when the user attempts an UPDATE on a contentless
253649	** table. Parameter bRowidModified is true if the UPDATE statement modifies
253650	** the rowid value. Parameter apVal[] contains the new values for each user
253651	** defined column of the fts5 table. pConfig is the configuration object of the
253652	** table being updated (guaranteed to be contentless). The contentless_delete=1
253653	** and contentless_unindexed=1 options may or may not be set.
253654	**
253655	** This function returns SQLITE_OK if the UPDATE can go ahead, or an SQLite
253656	** error code if it cannot. In this case an error message is also loaded into
253657	** pConfig. Output parameter (*pbContent) is set to true if the caller should
253658	** update the %_content table only - not the FTS index or any other shadow
253659	** table. This occurs when an UPDATE modifies only UNINDEXED columns of the
253660	** table.
253661	**
253662	** An UPDATE may proceed if:
253663	**
253664	** * The only columns modified are UNINDEXED columns, or
253665	**
253666	** * The contentless_delete=1 option was specified and all of the indexed
253667	** columns (not a subset) have been modified.
253668	*/
253669	static int fts5ContentlessUpdate(
253670	Fts5Config *pConfig,
253671	sqlite3_value **apVal,
253672	int bRowidModified,
253673	int *pbContent
253674	){
253675	int ii;
253676	int bSeenIndex = 0; /* Have seen modified indexed column */
253677	int bSeenIndexNC = 0; /* Have seen unmodified indexed column */
253678	int rc = SQLITE_OK;
253679
253680	for(ii=0; ii<pConfig->nCol; ii++){
253681	if( pConfig->abUnindexed[ii]==0 ){
253682	if( sqlite3_value_nochange(apVal[ii]) ){
253683	bSeenIndexNC++;
253684	}else{
253685	bSeenIndex++;
253686	}
253687	}
253688	}
253689
253690	if( bSeenIndex==0 && bRowidModified==0 ){
253691	*pbContent = 1;
253692	}else{
253693	if( bSeenIndexNC \|\| pConfig->bContentlessDelete==0 ){
253694	rc = SQLITE_ERROR;
253695	sqlite3Fts5ConfigErrmsg(pConfig,
253696	(pConfig->bContentlessDelete ?
253697	"%s a subset of columns on fts5 contentless-delete table: %s" :
253698	"%s contentless fts5 table: %s")
253699	, "cannot UPDATE", pConfig->zName
253700	);
253701	}
253702	}
253703
253704	return rc;
253705	}
253706
253707	/*
253708	** This function is the implementation of the xUpdate callback used by
253709	** FTS3 virtual tables. It is invoked by SQLite each time a row is to be
253710	** inserted, updated or deleted.
	@@ -252971,48 +253788,38 @@
253788	}
253789
253790	assert( eType0==SQLITE_INTEGER \|\| eType0==SQLITE_NULL );
253791	assert( nArg!=1 \|\| eType0==SQLITE_INTEGER );
253792














253793	/* DELETE */
253794	if( nArg==1 ){
253795	/* It is only possible to DELETE from a contentless table if the
253796	** contentless_delete=1 flag is set. */
253797	if( fts5IsContentless(pTab, 1) && pConfig->bContentlessDelete==0 ){
253798	fts5SetVtabError(pTab,
253799	"cannot DELETE from contentless fts5 table: %s", pConfig->zName
253800	);
253801	rc = SQLITE_ERROR;
253802	}else{
253803	i64 iDel = sqlite3_value_int64(apVal[0]); /* Rowid to delete */
253804	rc = sqlite3Fts5StorageDelete(pTab->pStorage, iDel, 0, 0);
253805	bUpdateOrDelete = 1;
253806	}
253807	}
253808
253809	/* INSERT or UPDATE */
253810	else{
253811	int eType1 = sqlite3_value_numeric_type(apVal[1]);
253812
253813	/* It is an error to write an fts5_locale() value to a table without
253814	** the locale=1 option. */
253815	if( pConfig->bLocale==0 ){
253816	int ii;
253817	for(ii=0; ii<pConfig->nCol; ii++){
253818	sqlite3_value *pVal = apVal[ii+2];
253819	if( sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
253820	fts5SetVtabError(pTab, "fts5_locale() requires locale=1");





253821	rc = SQLITE_MISMATCH;
253822	goto update_out;
253823	}
253824	}
253825	}
	@@ -253028,39 +253835,59 @@
253835	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253836	}
253837
253838	/* UPDATE */
253839	else{
253840	Fts5Storage *pStorage = pTab->pStorage;
253841	i64 iOld = sqlite3_value_int64(apVal[0]); /* Old rowid */
253842	i64 iNew = sqlite3_value_int64(apVal[1]); /* New rowid */
253843	int bContent = 0; /* Content only update */
253844
253845	/* If this is a contentless table (including contentless_unindexed=1
253846	** tables), check if the UPDATE may proceed. */
253847	if( fts5IsContentless(pTab, 1) ){
253848	rc = fts5ContentlessUpdate(pConfig, &apVal[2], iOld!=iNew, &bContent);
253849	if( rc!=SQLITE_OK ) goto update_out;
253850	}
253851
253852	if( eType1!=SQLITE_INTEGER ){
253853	rc = SQLITE_MISMATCH;
253854	}else if( iOld!=iNew ){
253855	assert( bContent==0 );
253856	if( eConflict==SQLITE_REPLACE ){
253857	rc = sqlite3Fts5StorageDelete(pStorage, iOld, 0, 1);
253858	if( rc==SQLITE_OK ){
253859	rc = sqlite3Fts5StorageDelete(pStorage, iNew, 0, 0);
253860	}
253861	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253862	}else{
253863	rc = sqlite3Fts5StorageFindDeleteRow(pStorage, iOld);
253864	if( rc==SQLITE_OK ){
253865	rc = sqlite3Fts5StorageContentInsert(pStorage, 0, apVal, pRowid);
253866	}
253867	if( rc==SQLITE_OK ){
253868	rc = sqlite3Fts5StorageDelete(pStorage, iOld, 0, 0);
253869	}
253870	if( rc==SQLITE_OK ){
253871	rc = sqlite3Fts5StorageIndexInsert(pStorage, apVal, *pRowid);
253872	}
253873	}
253874	}else if( bContent ){
253875	/* This occurs when an UPDATE on a contentless table affects only
253876	** UNINDEXED columns. This is a no-op for contentless_unindexed=0
253877	** tables, or a write to the %_content table only for =1 tables. */
253878	assert( fts5IsContentless(pTab, 1) );
253879	rc = sqlite3Fts5StorageFindDeleteRow(pStorage, iOld);
253880	if( rc==SQLITE_OK ){
253881	rc = sqlite3Fts5StorageContentInsert(pStorage, 1, apVal, pRowid);
253882	}
253883	}else{
253884	rc = sqlite3Fts5StorageDelete(pStorage, iOld, 0, 1);
253885	fts5StorageInsert(&rc, pTab, apVal, pRowid);
253886	}
253887	bUpdateOrDelete = 1;
253888	sqlite3Fts5StorageReleaseDeleteRow(pStorage);
253889	}
253890
253891	}
253892	}
253893
	@@ -253169,15 +253996,15 @@
253996	){
253997	Fts5Cursor pCsr = (Fts5Cursor)pCtx;
253998	Fts5Table pTab = (Fts5Table)(pCsr->base.pVtab);
253999	int rc = SQLITE_OK;
254000
254001	sqlite3Fts5SetLocale(pTab->pConfig, pLoc, nLoc);
254002	rc = sqlite3Fts5Tokenize(pTab->pConfig,
254003	FTS5_TOKENIZE_AUX, pText, nText, pUserData, xToken
254004	);
254005	sqlite3Fts5SetLocale(pTab->pConfig, 0, 0);
254006
254007	return rc;
254008	}
254009
254010	/*
	@@ -253200,10 +254027,53 @@
254027
254028	static int fts5ApiPhraseSize(Fts5Context *pCtx, int iPhrase){
254029	Fts5Cursor pCsr = (Fts5Cursor)pCtx;
254030	return sqlite3Fts5ExprPhraseSize(pCsr->pExpr, iPhrase);
254031	}
254032
254033	/*
254034	** Argument pStmt is an SQL statement of the type used by Fts5Cursor. This
254035	** function extracts the text value of column iCol of the current row.
254036	** Additionally, if there is an associated locale, it invokes
254037	** sqlite3Fts5SetLocale() to configure the tokenizer. In all cases the caller
254038	** should invoke sqlite3Fts5ClearLocale() to clear the locale at some point
254039	** after this function returns.
254040	**
254041	** If successful, (*ppText) is set to point to a buffer containing the text
254042	** value as utf-8 and SQLITE_OK returned. (*pnText) is set to the size of that
254043	** buffer in bytes. It is not guaranteed to be nul-terminated. If an error
254044	** occurs, an SQLite error code is returned. The final values of the two
254045	** output parameters are undefined in this case.
254046	*/
254047	static int fts5TextFromStmt(
254048	Fts5Config *pConfig,
254049	sqlite3_stmt *pStmt,
254050	int iCol,
254051	const char **ppText,
254052	int *pnText
254053	){
254054	sqlite3_value *pVal = sqlite3_column_value(pStmt, iCol+1);
254055	const char *pLoc = 0;
254056	int nLoc = 0;
254057	int rc = SQLITE_OK;
254058
254059	if( pConfig->bLocale
254060	&& pConfig->eContent==FTS5_CONTENT_EXTERNAL
254061	&& sqlite3Fts5IsLocaleValue(pConfig, pVal)
254062	){
254063	rc = sqlite3Fts5DecodeLocaleValue(pVal, ppText, pnText, &pLoc, &nLoc);
254064	}else{
254065	ppText = (const char)sqlite3_value_text(pVal);
254066	*pnText = sqlite3_value_bytes(pVal);
254067	if( pConfig->bLocale && pConfig->eContent==FTS5_CONTENT_NORMAL ){
254068	pLoc = (const char*)sqlite3_column_text(pStmt, iCol+1+pConfig->nCol);
254069	nLoc = sqlite3_column_bytes(pStmt, iCol+1+pConfig->nCol);
254070	}
254071	}
254072	sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
254073	return rc;
254074	}
254075
254076	static int fts5ApiColumnText(
254077	Fts5Context *pCtx,
254078	int iCol,
254079	const char **pz,
	@@ -253214,20 +254084,18 @@
254084	Fts5Table pTab = (Fts5Table)(pCsr->base.pVtab);
254085
254086	assert( pCsr->ePlan!=FTS5_PLAN_SPECIAL );
254087	if( iCol<0 \|\| iCol>=pTab->pConfig->nCol ){
254088	rc = SQLITE_RANGE;
254089	}else if( fts5IsContentless((Fts5FullTable*)(pCsr->base.pVtab), 0) ){
254090	*pz = 0;
254091	*pn = 0;
254092	}else{
254093	rc = fts5SeekCursor(pCsr, 0);
254094	if( rc==SQLITE_OK ){
254095	rc = fts5TextFromStmt(pTab->pConfig, pCsr->pStmt, iCol, pz, pn);
254096	sqlite3Fts5ClearLocale(pTab->pConfig);


254097	}
254098	}
254099	return rc;
254100	}
254101
	@@ -253249,11 +254117,11 @@
254117	int bLive = (pCsr->pSorter==0);
254118
254119	if( iPhrase<0 \|\| iPhrase>=sqlite3Fts5ExprPhraseCount(pCsr->pExpr) ){
254120	rc = SQLITE_RANGE;
254121	}else if( pConfig->eDetail!=FTS5_DETAIL_FULL
254122	&& fts5IsContentless((Fts5FullTable*)pCsr->base.pVtab, 1)
254123	){
254124	*pa = 0;
254125	*pn = 0;
254126	return SQLITE_OK;
254127	}else if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_POSLIST) ){
	@@ -253265,21 +254133,19 @@
254133	if( aPopulator==0 ) rc = SQLITE_NOMEM;
254134	if( rc==SQLITE_OK ){
254135	rc = fts5SeekCursor(pCsr, 0);
254136	}
254137	for(i=0; i<pConfig->nCol && rc==SQLITE_OK; i++){

254138	const char *z = 0;
254139	int n = 0;
254140	rc = fts5TextFromStmt(pConfig, pCsr->pStmt, i, &z, &n);

254141	if( rc==SQLITE_OK ){
254142	rc = sqlite3Fts5ExprPopulatePoslists(
254143	pConfig, pCsr->pExpr, aPopulator, i, z, n
254144	);
254145	}
254146	sqlite3Fts5ClearLocale(pConfig);
254147	}
254148	sqlite3_free(aPopulator);
254149
254150	if( pCsr->pSorter ){
254151	sqlite3Fts5ExprCheckPoslists(pCsr->pExpr, pCsr->pSorter->iRowid);
	@@ -253447,11 +254313,11 @@
254313
254314	if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_DOCSIZE) ){
254315	if( pConfig->bColumnsize ){
254316	i64 iRowid = fts5CursorRowid(pCsr);
254317	rc = sqlite3Fts5StorageDocsize(pTab->pStorage, iRowid, pCsr->aColumnSize);
254318	}else if( !pConfig->zContent \|\| pConfig->eContent==FTS5_CONTENT_UNINDEXED ){
254319	int i;
254320	for(i=0; i<pConfig->nCol; i++){
254321	if( pConfig->abUnindexed[i]==0 ){
254322	pCsr->aColumnSize[i] = -1;
254323	}
	@@ -253461,21 +254327,18 @@
254327	rc = fts5SeekCursor(pCsr, 0);
254328	for(i=0; rc==SQLITE_OK && i<pConfig->nCol; i++){
254329	if( pConfig->abUnindexed[i]==0 ){
254330	const char *z = 0;
254331	int n = 0;



254332	pCsr->aColumnSize[i] = 0;
254333	rc = fts5TextFromStmt(pConfig, pCsr->pStmt, i, &z, &n);
254334	if( rc==SQLITE_OK ){
254335	rc = sqlite3Fts5Tokenize(pConfig, FTS5_TOKENIZE_AUX,
254336	z, n, (void*)&pCsr->aColumnSize[i], fts5ColumnSizeCb
254337	);

254338	}
254339	sqlite3Fts5ClearLocale(pConfig);
254340	}
254341	}
254342	}
254343	CsrFlagClear(pCsr, FTS5CSR_REQUIRE_DOCSIZE);
254344	}
	@@ -253738,46 +254601,23 @@
254601	assert( pCsr->ePlan!=FTS5_PLAN_SPECIAL );
254602	if( iCol<0 \|\| iCol>=pConfig->nCol ){
254603	rc = SQLITE_RANGE;
254604	}else if(
254605	pConfig->abUnindexed[iCol]==0
254606	&& 0==fts5IsContentless((Fts5FullTable*)pCsr->base.pVtab, 1)
254607	&& pConfig->bLocale
254608	){
254609	rc = fts5SeekCursor(pCsr, 0);
254610	if( rc==SQLITE_OK ){
254611	const char *zDummy = 0;
254612	int nDummy = 0;
254613	rc = fts5TextFromStmt(pConfig, pCsr->pStmt, iCol, &zDummy, &nDummy);
254614	if( rc==SQLITE_OK ){
254615	*pzLocale = pConfig->t.pLocale;
254616	*pnLocale = pConfig->t.nLocale;
254617	}
254618	sqlite3Fts5ClearLocale(pConfig);























254619	}
254620	}
254621
254622	return rc;
254623	}
	@@ -253994,61 +254834,10 @@
254834
254835	sqlite3_result_blob(pCtx, val.p, val.n, sqlite3_free);
254836	return rc;
254837	}
254838



















































254839	/*
254840	** This is the xColumn method, called by SQLite to request a value from
254841	** the row that the supplied cursor currently points to.
254842	*/
254843	static int fts5ColumnMethod(
	@@ -254087,26 +254876,31 @@
254876	if( pCsr->pRank \|\| SQLITE_OK==(rc = fts5FindRankFunction(pCsr)) ){
254877	fts5ApiInvoke(pCsr->pRank, pCsr, pCtx, pCsr->nRankArg, pCsr->apRankArg);
254878	}
254879	}
254880	}else{
254881	if( !sqlite3_vtab_nochange(pCtx) && pConfig->eContent!=FTS5_CONTENT_NONE ){









254882	pConfig->pzErrmsg = &pTab->p.base.zErrMsg;
254883	rc = fts5SeekCursor(pCsr, 1);
254884	if( rc==SQLITE_OK ){
254885	sqlite3_value *pVal = sqlite3_column_value(pCsr->pStmt, iCol+1);
254886	if( pConfig->bLocale
254887	&& pConfig->eContent==FTS5_CONTENT_EXTERNAL
254888	&& sqlite3Fts5IsLocaleValue(pConfig, pVal)
254889	){
254890	const char *z = 0;
254891	int n = 0;
254892	rc = fts5TextFromStmt(pConfig, pCsr->pStmt, iCol, &z, &n);
254893	if( rc==SQLITE_OK ){
254894	sqlite3_result_text(pCtx, z, n, SQLITE_TRANSIENT);
254895	}
254896	sqlite3Fts5ClearLocale(pConfig);
254897	}else{
254898	sqlite3_result_value(pCtx, pVal);
254899	}
254900	}
254901
254902	pConfig->pzErrmsg = 0;
254903	}
254904	}
254905
254906	return rc;
	@@ -254625,11 +255419,11 @@
255419	int nArg, /* Number of args */
255420	sqlite3_value *apUnused / Function arguments */
255421	){
255422	assert( nArg==0 );
255423	UNUSED_PARAM2(nArg, apUnused);
255424	sqlite3_result_text(pCtx, "fts5: 2024-12-09 20:46:36 e2bae4143afd07de1ae55a6d2606a3b541a5b94568aa41f6a96e5d1245471653", -1, SQLITE_TRANSIENT);
255425	}
255426
255427	/*
255428	** Implementation of fts5_locale(LOCALE, TEXT) function.
255429	**
	@@ -254664,36 +255458,48 @@
255458	nText = sqlite3_value_bytes(apArg[1]);
255459
255460	if( zLocale==0 \|\| zLocale[0]=='\0' ){
255461	sqlite3_result_text(pCtx, zText, nText, SQLITE_TRANSIENT);
255462	}else{
255463	Fts5Global p = (Fts5Global)sqlite3_user_data(pCtx);
255464	u8 *pBlob = 0;
255465	u8 *pCsr = 0;
255466	int nBlob = 0;


255467
255468	nBlob = FTS5_LOCALE_HDR_SIZE + nLocale + 1 + nText;
255469	pBlob = (u8*)sqlite3_malloc(nBlob);
255470	if( pBlob==0 ){
255471	sqlite3_result_error_nomem(pCtx);
255472	return;
255473	}
255474
255475	pCsr = pBlob;
255476	memcpy(pCsr, (const u8*)p->aLocaleHdr, FTS5_LOCALE_HDR_SIZE);
255477	pCsr += FTS5_LOCALE_HDR_SIZE;
255478	memcpy(pCsr, zLocale, nLocale);
255479	pCsr += nLocale;
255480	(*pCsr++) = 0x00;
255481	if( zText ) memcpy(pCsr, zText, nText);
255482	assert( &pCsr[nText]==&pBlob[nBlob] );
255483
255484	sqlite3_result_blob(pCtx, pBlob, nBlob, sqlite3_free);

255485	}
255486	}
255487
255488	/*
255489	** Implementation of fts5_insttoken() function.
255490	*/
255491	static void fts5InsttokenFunc(
255492	sqlite3_context pCtx, / Function call context */
255493	int nArg, /* Number of args */
255494	sqlite3_value *apArg / Function arguments */
255495	){
255496	assert( nArg==1 );
255497	(void)nArg;
255498	sqlite3_result_value(pCtx, apArg[0]);
255499	sqlite3_result_subtype(pCtx, FTS5_INSTTOKEN_SUBTYPE);
255500	}
255501
255502	/*
255503	** Return true if zName is the extension on one of the shadow tables used
255504	** by this module.
255505	*/
	@@ -254789,10 +255595,20 @@
255595	pGlobal->api.xCreateFunction = fts5CreateAux;
255596	pGlobal->api.xCreateTokenizer = fts5CreateTokenizer;
255597	pGlobal->api.xFindTokenizer = fts5FindTokenizer;
255598	pGlobal->api.xCreateTokenizer_v2 = fts5CreateTokenizer_v2;
255599	pGlobal->api.xFindTokenizer_v2 = fts5FindTokenizer_v2;
255600
255601	/* Initialize pGlobal->aLocaleHdr[] to a 128-bit pseudo-random vector.
255602	** The constants below were generated randomly. */
255603	sqlite3_randomness(sizeof(pGlobal->aLocaleHdr), pGlobal->aLocaleHdr);
255604	pGlobal->aLocaleHdr[0] ^= 0xF924976D;
255605	pGlobal->aLocaleHdr[1] ^= 0x16596E13;
255606	pGlobal->aLocaleHdr[2] ^= 0x7C80BEAA;
255607	pGlobal->aLocaleHdr[3] ^= 0x9B03A67F;
255608	assert( sizeof(pGlobal->aLocaleHdr)==16 );
255609
255610	rc = sqlite3_create_module_v2(db, "fts5", &fts5Mod, p, fts5ModuleDestroy);
255611	if( rc==SQLITE_OK ) rc = sqlite3Fts5IndexInit(db);
255612	if( rc==SQLITE_OK ) rc = sqlite3Fts5ExprInit(pGlobal, db);
255613	if( rc==SQLITE_OK ) rc = sqlite3Fts5AuxInit(&pGlobal->api);
255614	if( rc==SQLITE_OK ) rc = sqlite3Fts5TokenizerInit(&pGlobal->api);
	@@ -254810,13 +255626,20 @@
255626	);
255627	}
255628	if( rc==SQLITE_OK ){
255629	rc = sqlite3_create_function(
255630	db, "fts5_locale", 2,
255631	SQLITE_UTF8\|SQLITE_INNOCUOUS\|SQLITE_RESULT_SUBTYPE\|SQLITE_SUBTYPE,
255632	p, fts5LocaleFunc, 0, 0
255633	);
255634	}
255635	if( rc==SQLITE_OK ){
255636	rc = sqlite3_create_function(
255637	db, "fts5_insttoken", 1,
255638	SQLITE_UTF8\|SQLITE_INNOCUOUS\|SQLITE_RESULT_SUBTYPE,
255639	p, fts5InsttokenFunc, 0, 0
255640	);
255641	}
255642	}
255643
255644	/* If SQLITE_FTS5_ENABLE_TEST_MI is defined, assume that the file
255645	** fts5_test_mi.c is compiled and linked into the executable. And call
	@@ -255015,24 +255838,39 @@
255838	);
255839	break;
255840
255841	case FTS5_STMT_INSERT_CONTENT:
255842	case FTS5_STMT_REPLACE_CONTENT: {
255843	char *zBind = 0;

255844	int i;
255845
255846	assert( pC->eContent==FTS5_CONTENT_NORMAL
255847	\|\| pC->eContent==FTS5_CONTENT_UNINDEXED
255848	);
255849
255850	/* Add bindings for the "c*" columns - those that store the actual
255851	** table content. If eContent==NORMAL, then there is one binding
255852	** for each column. Or, if eContent==UNINDEXED, then there are only
255853	** bindings for the UNINDEXED columns. */
255854	for(i=0; rc==SQLITE_OK && i<(pC->nCol+1); i++){
255855	if( !i \|\| pC->eContent==FTS5_CONTENT_NORMAL \|\| pC->abUnindexed[i-1] ){
255856	zBind = sqlite3Fts5Mprintf(&rc, "%z%s?%d", zBind, zBind?",":"",i+1);
255857	}
255858	}
255859
255860	/* Add bindings for any "l*" columns. Only non-UNINDEXED columns
255861	** require these. */
255862	if( pC->bLocale && pC->eContent==FTS5_CONTENT_NORMAL ){
255863	for(i=0; rc==SQLITE_OK && i<pC->nCol; i++){
255864	if( pC->abUnindexed[i]==0 ){
255865	zBind = sqlite3Fts5Mprintf(&rc, "%z,?%d", zBind, pC->nCol+i+2);
255866	}
255867	}
255868	}
255869
255870	zSql = sqlite3Fts5Mprintf(&rc, azStmt[eStmt], pC->zDb, pC->zName,zBind);
255871	sqlite3_free(zBind);
255872	break;
255873	}
255874
255875	case FTS5_STMT_REPLACE_DOCSIZE:
255876	zSql = sqlite3_mprintf(azStmt[eStmt], pC->zDb, pC->zName,
	@@ -255214,23 +256052,37 @@
256052	p->aTotalSize = (i64*)&p[1];
256053	p->pConfig = pConfig;
256054	p->pIndex = pIndex;
256055
256056	if( bCreate ){
256057	if( pConfig->eContent==FTS5_CONTENT_NORMAL
256058	\|\| pConfig->eContent==FTS5_CONTENT_UNINDEXED
256059	){
256060	int nDefn = 32 + pConfig->nCol*10;
256061	char zDefn = sqlite3_malloc64(32 + (sqlite3_int64)pConfig->nCol 20);
256062	if( zDefn==0 ){
256063	rc = SQLITE_NOMEM;
256064	}else{
256065	int i;
256066	int iOff;
256067	sqlite3_snprintf(nDefn, zDefn, "id INTEGER PRIMARY KEY");
256068	iOff = (int)strlen(zDefn);
256069	for(i=0; i<pConfig->nCol; i++){
256070	if( pConfig->eContent==FTS5_CONTENT_NORMAL
256071	\|\| pConfig->abUnindexed[i]
256072	){
256073	sqlite3_snprintf(nDefn-iOff, &zDefn[iOff], ", c%d", i);
256074	iOff += (int)strlen(&zDefn[iOff]);
256075	}
256076	}
256077	if( pConfig->bLocale ){
256078	for(i=0; i<pConfig->nCol; i++){
256079	if( pConfig->abUnindexed[i]==0 ){
256080	sqlite3_snprintf(nDefn-iOff, &zDefn[iOff], ", l%d", i);
256081	iOff += (int)strlen(&zDefn[iOff]);
256082	}
256083	}
256084	}
256085	rc = sqlite3Fts5CreateTable(pConfig, "content", zDefn, 0, pzErr);
256086	}
256087	sqlite3_free(zDefn);
256088	}
	@@ -255379,33 +256231,43 @@
256231	for(iCol=1; rc==SQLITE_OK && iCol<=pConfig->nCol; iCol++){
256232	if( pConfig->abUnindexed[iCol-1]==0 ){
256233	sqlite3_value *pVal = 0;
256234	const char *pText = 0;
256235	int nText = 0;
256236	const char *pLoc = 0;
256237	int nLoc = 0;
256238
256239	assert( pSeek==0 \|\| apVal==0 );
256240	assert( pSeek!=0 \|\| apVal!=0 );
256241	if( pSeek ){
256242	pVal = sqlite3_column_value(pSeek, iCol);
256243	}else{
256244	pVal = apVal[iCol-1];
256245	}
256246
256247	if( pConfig->bLocale && sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
256248	rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
256249	}else{
256250	pText = (const char*)sqlite3_value_text(pVal);
256251	nText = sqlite3_value_bytes(pVal);
256252	if( pConfig->bLocale && pSeek ){
256253	pLoc = (const char*)sqlite3_column_text(pSeek, iCol + pConfig->nCol);
256254	nLoc = sqlite3_column_bytes(pSeek, iCol + pConfig->nCol);
256255	}
256256	}
256257
256258	if( rc==SQLITE_OK ){
256259	sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
256260	ctx.szCol = 0;
256261	rc = sqlite3Fts5Tokenize(pConfig, FTS5_TOKENIZE_DOCUMENT,
256262	pText, nText, (void*)&ctx, fts5StorageInsertCallback
256263	);
256264	p->aTotalSize[iCol-1] -= (i64)ctx.szCol;
256265	if( rc==SQLITE_OK && p->aTotalSize[iCol-1]<0 ){
256266	rc = FTS5_CORRUPT;
256267	}
256268	sqlite3Fts5ClearLocale(pConfig);
256269	}
256270	}
256271	}
256272	if( rc==SQLITE_OK && p->nTotalRow<1 ){
256273	rc = FTS5_CORRUPT;
	@@ -255446,11 +256308,13 @@
256308	i64 iOrigin = 0;
256309	sqlite3_stmt *pLookup = 0;
256310	int rc = SQLITE_OK;
256311
256312	assert( p->pConfig->bContentlessDelete );
256313	assert( p->pConfig->eContent==FTS5_CONTENT_NONE
256314	\|\| p->pConfig->eContent==FTS5_CONTENT_UNINDEXED
256315	);
256316
256317	/* Look up the origin of the document in the %_docsize table. Store
256318	** this in stack variable iOrigin. */
256319	rc = fts5StorageGetStmt(p, FTS5_STMT_LOOKUP_DOCSIZE, &pLookup, 0);
256320	if( rc==SQLITE_OK ){
	@@ -255570,10 +256434,16 @@
256434	}
256435
256436	if( rc==SQLITE_OK ){
256437	if( p->pConfig->bContentlessDelete ){
256438	rc = fts5StorageContentlessDelete(p, iDel);
256439	if( rc==SQLITE_OK
256440	&& bSaveRow
256441	&& p->pConfig->eContent==FTS5_CONTENT_UNINDEXED
256442	){
256443	rc = sqlite3Fts5StorageFindDeleteRow(p, iDel);
256444	}
256445	}else{
256446	rc = fts5StorageDeleteFromIndex(p, iDel, apVal, bSaveRow);
256447	}
256448	}
256449
	@@ -255586,11 +256456,13 @@
256456	rc = sqlite3_reset(pDel);
256457	}
256458	}
256459
256460	/* Delete the %_content record */
256461	if( pConfig->eContent==FTS5_CONTENT_NORMAL
256462	\|\| pConfig->eContent==FTS5_CONTENT_UNINDEXED
256463	){
256464	if( rc==SQLITE_OK ){
256465	rc = fts5StorageGetStmt(p, FTS5_STMT_DELETE_CONTENT, &pDel, 0);
256466	}
256467	if( rc==SQLITE_OK ){
256468	sqlite3_bind_int64(pDel, 1, iDel);
	@@ -255618,12 +256490,17 @@
256490	pConfig->zDb, pConfig->zName,
256491	pConfig->zDb, pConfig->zName
256492	);
256493	if( rc==SQLITE_OK && pConfig->bColumnsize ){
256494	rc = fts5ExecPrintf(pConfig->db, 0,
256495	"DELETE FROM %Q.'%q_docsize';", pConfig->zDb, pConfig->zName
256496	);
256497	}
256498
256499	if( rc==SQLITE_OK && pConfig->eContent==FTS5_CONTENT_UNINDEXED ){
256500	rc = fts5ExecPrintf(pConfig->db, 0,
256501	"DELETE FROM %Q.'%q_content';", pConfig->zDb, pConfig->zName
256502	);
256503	}
256504
256505	/* Reinitialize the %_data table. This call creates the initial structure
256506	** and averages records. */
	@@ -255660,24 +256537,39 @@
256537	sqlite3Fts5BufferZero(&buf);
256538	rc = sqlite3Fts5IndexBeginWrite(p->pIndex, 0, iRowid);
256539	for(ctx.iCol=0; rc==SQLITE_OK && ctx.iCol<pConfig->nCol; ctx.iCol++){
256540	ctx.szCol = 0;
256541	if( pConfig->abUnindexed[ctx.iCol]==0 ){

256542	int nText = 0; /* Size of pText in bytes */
256543	const char pText = 0; / Pointer to buffer containing text value */
256544	int nLoc = 0; /* Size of pLoc in bytes */
256545	const char pLoc = 0; / Pointer to buffer containing text value */
256546
256547	sqlite3_value *pVal = sqlite3_column_value(pScan, ctx.iCol+1);
256548	if( pConfig->eContent==FTS5_CONTENT_EXTERNAL
256549	&& sqlite3Fts5IsLocaleValue(pConfig, pVal)
256550	){
256551	rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
256552	}else{
256553	pText = (const char*)sqlite3_value_text(pVal);
256554	nText = sqlite3_value_bytes(pVal);
256555	if( pConfig->bLocale ){
256556	int iCol = ctx.iCol + 1 + pConfig->nCol;
256557	pLoc = (const char*)sqlite3_column_text(pScan, iCol);
256558	nLoc = sqlite3_column_bytes(pScan, iCol);
256559	}
256560	}
256561

256562	if( rc==SQLITE_OK ){
256563	sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
256564	rc = sqlite3Fts5Tokenize(pConfig,
256565	FTS5_TOKENIZE_DOCUMENT,
256566	pText, nText,
256567	(void*)&ctx,
256568	fts5StorageInsertCallback
256569	);
256570	sqlite3Fts5ClearLocale(pConfig);
256571	}
256572	}
256573	sqlite3Fts5BufferAppendVarint(&rc, &buf, ctx.szCol);
256574	p->aTotalSize[ctx.iCol] += (i64)ctx.szCol;
256575	}
	@@ -255740,53 +256632,75 @@
256632	/*
256633	** Insert a new row into the FTS content table.
256634	*/
256635	static int sqlite3Fts5StorageContentInsert(
256636	Fts5Storage *p,
256637	int bReplace, /* True to use REPLACE instead of INSERT */
256638	sqlite3_value **apVal,
256639	i64 *piRowid
256640	){
256641	Fts5Config *pConfig = p->pConfig;
256642	int rc = SQLITE_OK;
256643
256644	/* Insert the new row into the %_content table. */
256645	if( pConfig->eContent!=FTS5_CONTENT_NORMAL
256646	&& pConfig->eContent!=FTS5_CONTENT_UNINDEXED
256647	){
256648	if( sqlite3_value_type(apVal[1])==SQLITE_INTEGER ){
256649	*piRowid = sqlite3_value_int64(apVal[1]);
256650	}else{
256651	rc = fts5StorageNewRowid(p, piRowid);
256652	}
256653	}else{
256654	sqlite3_stmt pInsert = 0; / Statement to write %_content table */
256655	int i; /* Counter variable */
256656
256657	assert( FTS5_STMT_INSERT_CONTENT+1==FTS5_STMT_REPLACE_CONTENT );
256658	assert( bReplace==0 \|\| bReplace==1 );
256659	rc = fts5StorageGetStmt(p, FTS5_STMT_INSERT_CONTENT+bReplace, &pInsert, 0);
256660	if( pInsert ) sqlite3_clear_bindings(pInsert);
256661
256662	/* Bind the rowid value */
256663	sqlite3_bind_value(pInsert, 1, apVal[1]);
256664
256665	/* Loop through values for user-defined columns. i=2 is the leftmost
256666	** user-defined column. As is column 1 of pSavedRow. */
256667	for(i=2; rc==SQLITE_OK && i<=pConfig->nCol+1; i++){
256668	int bUnindexed = pConfig->abUnindexed[i-2];
256669	if( pConfig->eContent==FTS5_CONTENT_NORMAL \|\| bUnindexed ){
256670	sqlite3_value *pVal = apVal[i];
256671
256672	if( sqlite3_value_nochange(pVal) && p->pSavedRow ){
256673	/* This is an UPDATE statement, and user-defined column (i-2) was not
256674	** modified. Retrieve the value from Fts5Storage.pSavedRow. */
256675	pVal = sqlite3_column_value(p->pSavedRow, i-1);
256676	if( pConfig->bLocale && bUnindexed==0 ){
256677	sqlite3_bind_value(pInsert, pConfig->nCol + i,
256678	sqlite3_column_value(p->pSavedRow, pConfig->nCol + i - 1)
256679	);
256680	}
256681	}else if( sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
256682	const char *pText = 0;
256683	const char *pLoc = 0;
256684	int nText = 0;
256685	int nLoc = 0;
256686	assert( pConfig->bLocale );
256687
256688	rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
256689	if( rc==SQLITE_OK ){
256690	sqlite3_bind_text(pInsert, i, pText, nText, SQLITE_TRANSIENT);
256691	if( bUnindexed==0 ){
256692	int iLoc = pConfig->nCol + i;
256693	sqlite3_bind_text(pInsert, iLoc, pLoc, nLoc, SQLITE_TRANSIENT);
256694	}
256695	}
256696
256697	continue;
256698	}
256699
256700	rc = sqlite3_bind_value(pInsert, i, pVal);
256701	}
256702	}
256703	if( rc==SQLITE_OK ){
256704	sqlite3_step(pInsert);
256705	rc = sqlite3_reset(pInsert);
256706	}
	@@ -255817,27 +256731,41 @@
256731	rc = sqlite3Fts5IndexBeginWrite(p->pIndex, 0, iRowid);
256732	}
256733	for(ctx.iCol=0; rc==SQLITE_OK && ctx.iCol<pConfig->nCol; ctx.iCol++){
256734	ctx.szCol = 0;
256735	if( pConfig->abUnindexed[ctx.iCol]==0 ){

256736	int nText = 0; /* Size of pText in bytes */
256737	const char pText = 0; / Pointer to buffer containing text value */
256738	int nLoc = 0; /* Size of pText in bytes */
256739	const char pLoc = 0; / Pointer to buffer containing text value */
256740
256741	sqlite3_value *pVal = apVal[ctx.iCol+2];

256742	if( p->pSavedRow && sqlite3_value_nochange(pVal) ){
256743	pVal = sqlite3_column_value(p->pSavedRow, ctx.iCol+1);
256744	if( pConfig->eContent==FTS5_CONTENT_NORMAL && pConfig->bLocale ){
256745	int iCol = ctx.iCol + 1 + pConfig->nCol;
256746	pLoc = (const char*)sqlite3_column_text(p->pSavedRow, iCol);
256747	nLoc = sqlite3_column_bytes(p->pSavedRow, iCol);
256748	}
256749	}else{
256750	pVal = apVal[ctx.iCol+2];
256751	}
256752
256753	if( pConfig->bLocale && sqlite3Fts5IsLocaleValue(pConfig, pVal) ){
256754	rc = sqlite3Fts5DecodeLocaleValue(pVal, &pText, &nText, &pLoc, &nLoc);
256755	}else{
256756	pText = (const char*)sqlite3_value_text(pVal);
256757	nText = sqlite3_value_bytes(pVal);
256758	}
256759
256760	if( rc==SQLITE_OK ){
256761	sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
256762	rc = sqlite3Fts5Tokenize(pConfig,
256763	FTS5_TOKENIZE_DOCUMENT, pText, nText, (void*)&ctx,
256764	fts5StorageInsertCallback
256765	);
256766	sqlite3Fts5ClearLocale(pConfig);
256767	}
256768	}
256769	sqlite3Fts5BufferAppendVarint(&rc, &buf, ctx.szCol);
256770	p->aTotalSize[ctx.iCol] += (i64)ctx.szCol;
256771	}
	@@ -255998,41 +256926,65 @@
256926	}
256927	if( rc==SQLITE_OK && pConfig->eDetail==FTS5_DETAIL_NONE ){
256928	rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
256929	}
256930	for(i=0; rc==SQLITE_OK && i<pConfig->nCol; i++){
256931	if( pConfig->abUnindexed[i]==0 ){
256932	const char *pText = 0;
256933	int nText = 0;
256934	const char *pLoc = 0;
256935	int nLoc = 0;
256936	sqlite3_value *pVal = sqlite3_column_value(pScan, i+1);
256937
256938	if( pConfig->eContent==FTS5_CONTENT_EXTERNAL
256939	&& sqlite3Fts5IsLocaleValue(pConfig, pVal)
256940	){
256941	rc = sqlite3Fts5DecodeLocaleValue(
256942	pVal, &pText, &nText, &pLoc, &nLoc
256943	);
256944	}else{
256945	if( pConfig->eContent==FTS5_CONTENT_NORMAL && pConfig->bLocale ){
256946	int iCol = i + 1 + pConfig->nCol;
256947	pLoc = (const char*)sqlite3_column_text(pScan, iCol);
256948	nLoc = sqlite3_column_bytes(pScan, iCol);
256949	}
256950	pText = (const char*)sqlite3_value_text(pVal);
256951	nText = sqlite3_value_bytes(pVal);
256952	}
256953
256954	ctx.iCol = i;
256955	ctx.szCol = 0;
256956
256957	if( rc==SQLITE_OK && pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
256958	rc = sqlite3Fts5TermsetNew(&ctx.pTermset);
256959	}
256960
256961	if( rc==SQLITE_OK ){
256962	sqlite3Fts5SetLocale(pConfig, pLoc, nLoc);
256963	rc = sqlite3Fts5Tokenize(pConfig,
256964	FTS5_TOKENIZE_DOCUMENT,
256965	pText, nText,
256966	(void*)&ctx,
256967	fts5StorageIntegrityCallback
256968	);
256969	sqlite3Fts5ClearLocale(pConfig);
256970	}
256971
256972	/* If this is not a columnsize=0 database, check that the number
256973	** of tokens in the value matches the aColSize[] value read from
256974	** the %_docsize table. */
256975	if( rc==SQLITE_OK
256976	&& pConfig->bColumnsize
256977	&& ctx.szCol!=aColSize[i]
256978	){
256979	rc = FTS5_CORRUPT;
256980	}
256981	aTotalSize[i] += ctx.szCol;
256982	if( pConfig->eDetail==FTS5_DETAIL_COLUMNS ){
256983	sqlite3Fts5TermsetFree(ctx.pTermset);
256984	ctx.pTermset = 0;
256985	}
256986	}
256987	}
256988	sqlite3Fts5TermsetFree(ctx.pTermset);
256989	ctx.pTermset = 0;
256990
	@@ -256454,11 +257406,11 @@
257406
257407	#define READ_UTF8(zIn, zTerm, c) \
257408	c = *(zIn++); \
257409	if( c>=0xc0 ){ \
257410	c = sqlite3Utf8Trans1[c-0xc0]; \
257411	while( zIn<zTerm && (*zIn & 0xc0)==0x80 ){ \
257412	c = (c<<6) + (0x3f & *(zIn++)); \
257413	} \
257414	if( c<0x80 \
257415	\|\| (c&0xFFFFF800)==0xD800 \
257416	\|\| (c&0xFFFFFFFE)==0xFFFE ){ c = 0xFFFD; } \
	@@ -257609,22 +258561,22 @@
258561	int rc = SQLITE_OK;
258562	char aBuf[32];
258563	char *zOut = aBuf;
258564	int ii;
258565	const unsigned char zIn = (const unsigned char)pText;
258566	const unsigned char *zEof = (zIn ? &zIn[nText] : 0);
258567	u32 iCode = 0;
258568	int aStart[3]; /* Input offset of each character in aBuf[] */
258569
258570	UNUSED_PARAM(unusedFlags);
258571
258572	/* Populate aBuf[] with the characters for the first trigram. */
258573	for(ii=0; ii<3; ii++){
258574	do {
258575	aStart[ii] = zIn - (const unsigned char*)pText;
258576	if( zIn>=zEof ) return SQLITE_OK;
258577	READ_UTF8(zIn, zEof, iCode);

258578	if( p->bFold ) iCode = sqlite3Fts5UnicodeFold(iCode, p->iFoldParam);
258579	}while( iCode==0 );
258580	WRITE_UTF8(zOut, iCode);
258581	}
258582
	@@ -257641,12 +258593,15 @@
258593	const char *z1;
258594
258595	/* Read characters from the input up until the first non-diacritic */
258596	do {
258597	iNext = zIn - (const unsigned char*)pText;
258598	if( zIn>=zEof ){
258599	iCode = 0;
258600	break;
258601	}
258602	READ_UTF8(zIn, zEof, iCode);

258603	if( p->bFold ) iCode = sqlite3Fts5UnicodeFold(iCode, p->iFoldParam);
258604	}while( iCode==0 );
258605
258606	/* Pass the current trigram back to fts5 */
258607	rc = xToken(pCtx, 0, aBuf, zOut-aBuf, aStart[0], iNext);
	@@ -259679,11 +260634,11 @@
260634
260635	return sqlite3_create_module_v2(db, "fts5vocab", &fts5Vocab, p, 0);
260636	}
260637
260638
260639	/* Here ends the fts5.c composite file. */
260640	#endif /* !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_FTS5) */
260641
260642	/************ End of fts5.c **********************************************/
260643	/************ Begin file stmt.c ******************************************/
260644	/*
	@@ -260035,6 +260990,7 @@
260990	#endif /* !defined(SQLITE_CORE) \|\| defined(SQLITE_ENABLE_STMTVTAB) */
260991
260992	/************ End of stmt.c **********************************************/
260993	/* Return the source-id for this library */
260994	SQLITE_API const char *sqlite3_sourceid(void){ return SQLITE_SOURCE_ID; }
260995	#endif /* SQLITE_AMALGAMATION */
260996	/************************ End of sqlite3.c ****************************/
260997

M extsrc/sqlite3.h

+80 -13

		--- extsrc/sqlite3.h
		+++ extsrc/sqlite3.h
		@@ -144,13 +144,13 @@
144	144	**
145	145	** See also: [sqlite3_libversion()],
146	146	** [sqlite3_libversion_number()], [sqlite3_sourceid()],
147	147	** [sqlite_version()] and [sqlite_source_id()].
148	148	*/
149		-#define SQLITE_VERSION "3.47.0"
150		-#define SQLITE_VERSION_NUMBER 3047000
151		-#define SQLITE_SOURCE_ID "2024-09-02 21:59:31 7891a266c4425722ae8b9231397ef9e42e2432be9e6b70632dfaf9ff15300d2c"
	149	+#define SQLITE_VERSION "3.48.0"
	150	+#define SQLITE_VERSION_NUMBER 3048000
	151	+#define SQLITE_SOURCE_ID "2024-12-09 20:46:36 e2bae4143afd07de1ae55a6d2606a3b541a5b94568aa41f6a96e5d1245471653"
152	152
153	153	/*
154	154	** CAPI3REF: Run-Time Library Version Numbers
155	155	** KEYWORDS: sqlite3_version sqlite3_sourceid
156	156	**
		@@ -650,10 +650,17 @@
650	650	**
651	651	** The SQLITE_IOCAP_BATCH_ATOMIC property means that the underlying
652	652	** filesystem supports doing multiple write operations atomically when those
653	653	** write operations are bracketed by [SQLITE_FCNTL_BEGIN_ATOMIC_WRITE] and
654	654	** [SQLITE_FCNTL_COMMIT_ATOMIC_WRITE].
	655	+**
	656	+** The SQLITE_IOCAP_SUBPAGE_READ property means that it is ok to read
	657	+** from the database file in amounts that are not a multiple of the
	658	+** page size and that do not begin at a page boundary. Without this
	659	+** property, SQLite is careful to only do full-page reads and write
	660	+** on aligned pages, with the one exception that it will do a sub-page
	661	+** read of the first page to access the database header.
655	662	*/
656	663	#define SQLITE_IOCAP_ATOMIC 0x00000001
657	664	#define SQLITE_IOCAP_ATOMIC512 0x00000002
658	665	#define SQLITE_IOCAP_ATOMIC1K 0x00000004
659	666	#define SQLITE_IOCAP_ATOMIC2K 0x00000008
		@@ -666,10 +673,11 @@
666	673	#define SQLITE_IOCAP_SEQUENTIAL 0x00000400
667	674	#define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN 0x00000800
668	675	#define SQLITE_IOCAP_POWERSAFE_OVERWRITE 0x00001000
669	676	#define SQLITE_IOCAP_IMMUTABLE 0x00002000
670	677	#define SQLITE_IOCAP_BATCH_ATOMIC 0x00004000
	678	+#define SQLITE_IOCAP_SUBPAGE_READ 0x00008000
671	679
672	680	/*
673	681	** CAPI3REF: File Locking Levels
674	682	**
675	683	** SQLite uses one of these integer values as the second
		@@ -812,10 +820,11 @@
812	820	** <li> [SQLITE_IOCAP_SEQUENTIAL]
813	821	** <li> [SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN]
814	822	** <li> [SQLITE_IOCAP_POWERSAFE_OVERWRITE]
815	823	** <li> [SQLITE_IOCAP_IMMUTABLE]
816	824	** <li> [SQLITE_IOCAP_BATCH_ATOMIC]
	825	+** <li> [SQLITE_IOCAP_SUBPAGE_READ]
817	826	** </ul>
818	827	**
819	828	** The SQLITE_IOCAP_ATOMIC property means that all writes of
820	829	** any size are atomic. The SQLITE_IOCAP_ATOMICnnn values
821	830	** mean that writes of blocks that are nnn bytes in size and
		@@ -1089,10 +1098,15 @@
1089	1098	** The [SQLITE_FCNTL_WIN32_SET_HANDLE] opcode is used for debugging. This
1090	1099	** opcode causes the xFileControl method to swap the file handle with the one
1091	1100	** pointed to by the pArg argument. This capability is used during testing
1092	1101	** and only needs to be supported when SQLITE_TEST is defined.
1093	1102	**
	1103	+** <li>[[SQLITE_FCNTL_NULL_IO]]
	1104	+** The [SQLITE_FCNTL_NULL_IO] opcode sets the low-level file descriptor
	1105	+** or file handle for the [sqlite3_file] object such that it will no longer
	1106	+** read or write to the database file.
	1107	+**
1094	1108	** <li>[[SQLITE_FCNTL_WAL_BLOCK]]
1095	1109	** The [SQLITE_FCNTL_WAL_BLOCK] is a signal to the VFS layer that it might
1096	1110	** be advantageous to block on the next WAL lock if the lock is not immediately
1097	1111	** available. The WAL subsystem issues this signal during rare
1098	1112	** circumstances in order to fix a problem with priority inversion.
		@@ -1242,10 +1256,11 @@
1242	1256	#define SQLITE_FCNTL_RESERVE_BYTES 38
1243	1257	#define SQLITE_FCNTL_CKPT_START 39
1244	1258	#define SQLITE_FCNTL_EXTERNAL_READER 40
1245	1259	#define SQLITE_FCNTL_CKSM_FILE 41
1246	1260	#define SQLITE_FCNTL_RESET_CACHE 42
	1261	+#define SQLITE_FCNTL_NULL_IO 43
1247	1262
1248	1263	/* deprecated names */
1249	1264	#define SQLITE_GET_LOCKPROXYFILE SQLITE_FCNTL_GET_LOCKPROXYFILE
1250	1265	#define SQLITE_SET_LOCKPROXYFILE SQLITE_FCNTL_SET_LOCKPROXYFILE
1251	1266	#define SQLITE_LAST_ERRNO SQLITE_FCNTL_LAST_ERRNO
		@@ -2620,14 +2635,18 @@
2620	2635	**
2621	2636	** ^These functions return the number of rows modified, inserted or
2622	2637	** deleted by the most recently completed INSERT, UPDATE or DELETE
2623	2638	** statement on the database connection specified by the only parameter.
2624	2639	** The two functions are identical except for the type of the return value
2625		-** and that if the number of rows modified by the most recent INSERT, UPDATE
	2640	+** and that if the number of rows modified by the most recent INSERT, UPDATE,
2626	2641	** or DELETE is greater than the maximum value supported by type "int", then
2627	2642	** the return value of sqlite3_changes() is undefined. ^Executing any other
2628	2643	** type of SQL statement does not modify the value returned by these functions.
	2644	+** For the purposes of this interface, a CREATE TABLE AS SELECT statement
	2645	+** does not count as an INSERT, UPDATE or DELETE statement and hence the rows
	2646	+** added to the new table by the CREATE TABLE AS SELECT statement are not
	2647	+** counted.
2629	2648	**
2630	2649	** ^Only changes made directly by the INSERT, UPDATE or DELETE statement are
2631	2650	** considered - auxiliary changes caused by [CREATE TRIGGER \| triggers],
2632	2651	** [foreign key actions] or [REPLACE] constraint resolution are not counted.
2633	2652	**
		@@ -4183,15 +4202,26 @@
4183	4202	**
4184	4203	** [[SQLITE_PREPARE_NO_VTAB]] <dt>SQLITE_PREPARE_NO_VTAB</dt>
4185	4204	** <dd>The SQLITE_PREPARE_NO_VTAB flag causes the SQL compiler
4186	4205	** to return an error (error code SQLITE_ERROR) if the statement uses
4187	4206	** any virtual tables.
	4207	+**
	4208	+** [[SQLITE_PREPARE_DONT_LOG]] <dt>SQLITE_PREPARE_DONT_LOG</dt>
	4209	+** <dd>The SQLITE_PREPARE_DONT_LOG flag prevents SQL compiler
	4210	+** errors from being sent to the error log defined by
	4211	+** [SQLITE_CONFIG_LOG]. This can be used, for example, to do test
	4212	+** compiles to see if some SQL syntax is well-formed, without generating
	4213	+** messages on the global error log when it is not. If the test compile
	4214	+** fails, the sqlite3_prepare_v3() call returns the same error indications
	4215	+** with or without this flag; it just omits the call to [sqlite3_log()] that
	4216	+** logs the error.
4188	4217	** </dl>
4189	4218	*/
4190	4219	#define SQLITE_PREPARE_PERSISTENT 0x01
4191	4220	#define SQLITE_PREPARE_NORMALIZE 0x02
4192	4221	#define SQLITE_PREPARE_NO_VTAB 0x04
	4222	+#define SQLITE_PREPARE_DONT_LOG 0x10
4193	4223
4194	4224	/*
4195	4225	** CAPI3REF: Compiling An SQL Statement
4196	4226	** KEYWORDS: {SQL statement compiler}
4197	4227	** METHOD: sqlite3
		@@ -4220,17 +4250,21 @@
4220	4250	** and sqlite3_prepare_v3()
4221	4251	** interfaces use UTF-8, and sqlite3_prepare16(), sqlite3_prepare16_v2(),
4222	4252	** and sqlite3_prepare16_v3() use UTF-16.
4223	4253	**
4224	4254	** ^If the nByte argument is negative, then zSql is read up to the
4225		-** first zero terminator. ^If nByte is positive, then it is the
4226		-** number of bytes read from zSql. ^If nByte is zero, then no prepared
	4255	+** first zero terminator. ^If nByte is positive, then it is the maximum
	4256	+** number of bytes read from zSql. When nByte is positive, zSql is read
	4257	+** up to the first zero terminator or until the nByte bytes have been read,
	4258	+** whichever comes first. ^If nByte is zero, then no prepared
4227	4259	** statement is generated.
4228	4260	** If the caller knows that the supplied string is nul-terminated, then
4229	4261	** there is a small performance advantage to passing an nByte parameter that
4230	4262	** is the number of bytes in the input string <i>including</i>
4231	4263	** the nul-terminator.
	4264	+** Note that nByte measure the length of the input in bytes, not
	4265	+** characters, even for the UTF-16 interfaces.
4232	4266	**
4233	4267	** ^If pzTail is not NULL then *pzTail is made to point to the first byte
4234	4268	** past the end of the first SQL statement in zSql. These routines only
4235	4269	** compile the first statement in zSql, so *pzTail is left pointing to
4236	4270	** what remains uncompiled.
		@@ -5597,11 +5631,11 @@
5597	5631	** The SQLITE_SUBTYPE flag indicates to SQLite that a function might call
5598	5632	** [sqlite3_value_subtype()] to inspect the sub-types of its arguments.
5599	5633	** This flag instructs SQLite to omit some corner-case optimizations that
5600	5634	** might disrupt the operation of the [sqlite3_value_subtype()] function,
5601	5635	** causing it to return zero rather than the correct subtype().
5602		-** SQL functions that invokes [sqlite3_value_subtype()] should have this
	5636	+** All SQL functions that invoke [sqlite3_value_subtype()] should have this
5603	5637	** property. If the SQLITE_SUBTYPE property is omitted, then the return
5604	5638	** value from [sqlite3_value_subtype()] might sometimes be zero even though
5605	5639	** a non-zero subtype was specified by the function argument expression.
5606	5640	**
5607	5641	** [[SQLITE_RESULT_SUBTYPE]] <dt>SQLITE_RESULT_SUBTYPE</dt><dd>
		@@ -8362,11 +8396,11 @@
8362	8396	#define SQLITE_TESTCTRL_EXTRA_SCHEMA_CHECKS 29
8363	8397	#define SQLITE_TESTCTRL_SEEK_COUNT 30
8364	8398	#define SQLITE_TESTCTRL_TRACEFLAGS 31
8365	8399	#define SQLITE_TESTCTRL_TUNE 32
8366	8400	#define SQLITE_TESTCTRL_LOGEST 33
8367		-#define SQLITE_TESTCTRL_USELONGDOUBLE 34
	8401	+#define SQLITE_TESTCTRL_USELONGDOUBLE 34 /* NOT USED */
8368	8402	#define SQLITE_TESTCTRL_LAST 34 /* Largest TESTCTRL */
8369	8403
8370	8404	/*
8371	8405	** CAPI3REF: SQL Keyword Checking
8372	8406	**
		@@ -9338,10 +9372,20 @@
9338	9372	** threads may safely make multiple concurrent calls to sqlite3_backup_step().
9339	9373	** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
9340	9374	** APIs are not strictly speaking threadsafe. If they are invoked at the
9341	9375	** same time as another thread is invoking sqlite3_backup_step() it is
9342	9376	** possible that they return invalid values.
	9377	+**
	9378	+** <b>Alternatives To Using The Backup API</b>
	9379	+**
	9380	+** Other techniques for safely creating a consistent backup of an SQLite
	9381	+** database include:
	9382	+**
	9383	+** <ul>
	9384	+** <li> The [VACUUM INTO] command.
	9385	+** <li> The [sqlite3_rsync] utility program.
	9386	+** </ul>
9343	9387	*/
9344	9388	SQLITE_API sqlite3_backup *sqlite3_backup_init(
9345	9389	sqlite3 pDest, / Destination database handle */
9346	9390	const char zDestName, / Destination database name */
9347	9391	sqlite3 pSource, / Source database handle */
		@@ -10536,10 +10580,18 @@
10536	10580	** schema S in database connection D. ^On success, the
10537	10581	** [sqlite3_snapshot_get(D,S,P)] interface writes a pointer to the newly
10538	10582	** created [sqlite3_snapshot] object into *P and returns SQLITE_OK.
10539	10583	** If there is not already a read-transaction open on schema S when
10540	10584	** this function is called, one is opened automatically.
	10585	+**
	10586	+** If a read-transaction is opened by this function, then it is guaranteed
	10587	+** that the returned snapshot object may not be invalidated by a database
	10588	+** writer or checkpointer until after the read-transaction is closed. This
	10589	+** is not guaranteed if a read-transaction is already open when this
	10590	+** function is called. In that case, any subsequent write or checkpoint
	10591	+** operation on the database may invalidate the returned snapshot handle,
	10592	+** even while the read-transaction remains open.
10541	10593	**
10542	10594	** The following must be true for this function to succeed. If any of
10543	10595	** the following statements are false when sqlite3_snapshot_get() is
10544	10596	** called, SQLITE_ERROR is returned. The final value of *P is undefined
10545	10597	** in this case.
		@@ -10856,11 +10908,11 @@
10856	10908	#endif
10857	10909
10858	10910	#ifdef __cplusplus
10859	10911	} /* End of the 'extern "C"' block */
10860	10912	#endif
10861		-#endif /* SQLITE3_H */
	10913	+/* #endif for SQLITE3_H will be added by mksqlite3.tcl */
10862	10914
10863	10915	/****** Begin file sqlite3rtree.h *******/
10864	10916	/*
10865	10917	** 2010 August 30
10866	10918	**
		@@ -13107,17 +13159,32 @@
13107	13159	** This is used to access token iToken of phrase hit iIdx within the
13108	13160	** current row. If iIdx is less than zero or greater than or equal to the
13109	13161	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
13110	13162	** output variable (*ppToken) is set to point to a buffer containing the
13111	13163	** matching document token, and (*pnToken) to the size of that buffer in
13112		-** bytes. This API is not available if the specified token matches a
13113		-** prefix query term. In that case both output variables are always set
13114		-** to 0.
	13164	+** bytes.
13115	13165	**
13116	13166	** The output text is not a copy of the document text that was tokenized.
13117	13167	** It is the output of the tokenizer module. For tokendata=1 tables, this
13118	13168	** includes any embedded 0x00 and trailing data.
	13169	+**
	13170	+** This API may be slow in some cases if the token identified by parameters
	13171	+** iIdx and iToken matched a prefix token in the query. In most cases, the
	13172	+** first call to this API for each prefix token in the query is forced
	13173	+** to scan the portion of the full-text index that matches the prefix
	13174	+** token to collect the extra data required by this API. If the prefix
	13175	+** token matches a large number of token instances in the document set,
	13176	+** this may be a performance problem.
	13177	+**
	13178	+** If the user knows in advance that a query may use this API for a
	13179	+** prefix token, FTS5 may be configured to collect all required data as part
	13180	+** of the initial querying of the full-text index, avoiding the second scan
	13181	+** entirely. This also causes prefix queries that do not use this API to
	13182	+** run more slowly and use more memory. FTS5 may be configured in this way
	13183	+** either on a per-table basis using the [FTS5 insttoken \| 'insttoken']
	13184	+** option, or on a per-query basis using the
	13185	+** [fts5_insttoken \| fts5_insttoken()] user function.
13119	13186	**
13120	13187	** This API can be quite slow if used with an FTS5 table created with the
13121	13188	** "detail=none" or "detail=column" option.
13122	13189	**
13123	13190	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
		@@ -13205,11 +13272,10 @@
13205	13272	** CUSTOM TOKENIZERS
13206	13273	**
13207	13274	** Applications may also register custom tokenizer types. A tokenizer
13208	13275	** is registered by providing fts5 with a populated instance of the
13209	13276	** following structure. All structure methods must be defined, setting
13210		-**
13211	13277	** any member of the fts5_tokenizer struct to NULL leads to undefined
13212	13278	** behaviour. The structure methods are expected to function as follows:
13213	13279	**
13214	13280	** xCreate:
13215	13281	** This function is used to allocate and initialize a tokenizer instance.
		@@ -13549,5 +13615,6 @@
13549	13615	#endif
13550	13616
13551	13617	#endif /* _FTS5_H */
13552	13618
13553	13619	/****** End of fts5.h *******/
	13620	+#endif /* SQLITE3_H */
13554	13621

	--- extsrc/sqlite3.h
	+++ extsrc/sqlite3.h
	@@ -144,13 +144,13 @@
144	**
145	** See also: [sqlite3_libversion()],
146	** [sqlite3_libversion_number()], [sqlite3_sourceid()],
147	** [sqlite_version()] and [sqlite_source_id()].
148	*/
149	#define SQLITE_VERSION "3.47.0"
150	#define SQLITE_VERSION_NUMBER 3047000
151	#define SQLITE_SOURCE_ID "2024-09-02 21:59:31 7891a266c4425722ae8b9231397ef9e42e2432be9e6b70632dfaf9ff15300d2c"
152
153	/*
154	** CAPI3REF: Run-Time Library Version Numbers
155	** KEYWORDS: sqlite3_version sqlite3_sourceid
156	**
	@@ -650,10 +650,17 @@
650	**
651	** The SQLITE_IOCAP_BATCH_ATOMIC property means that the underlying
652	** filesystem supports doing multiple write operations atomically when those
653	** write operations are bracketed by [SQLITE_FCNTL_BEGIN_ATOMIC_WRITE] and
654	** [SQLITE_FCNTL_COMMIT_ATOMIC_WRITE].







655	*/
656	#define SQLITE_IOCAP_ATOMIC 0x00000001
657	#define SQLITE_IOCAP_ATOMIC512 0x00000002
658	#define SQLITE_IOCAP_ATOMIC1K 0x00000004
659	#define SQLITE_IOCAP_ATOMIC2K 0x00000008
	@@ -666,10 +673,11 @@
666	#define SQLITE_IOCAP_SEQUENTIAL 0x00000400
667	#define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN 0x00000800
668	#define SQLITE_IOCAP_POWERSAFE_OVERWRITE 0x00001000
669	#define SQLITE_IOCAP_IMMUTABLE 0x00002000
670	#define SQLITE_IOCAP_BATCH_ATOMIC 0x00004000

671
672	/*
673	** CAPI3REF: File Locking Levels
674	**
675	** SQLite uses one of these integer values as the second
	@@ -812,10 +820,11 @@
812	** <li> [SQLITE_IOCAP_SEQUENTIAL]
813	** <li> [SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN]
814	** <li> [SQLITE_IOCAP_POWERSAFE_OVERWRITE]
815	** <li> [SQLITE_IOCAP_IMMUTABLE]
816	** <li> [SQLITE_IOCAP_BATCH_ATOMIC]

817	** </ul>
818	**
819	** The SQLITE_IOCAP_ATOMIC property means that all writes of
820	** any size are atomic. The SQLITE_IOCAP_ATOMICnnn values
821	** mean that writes of blocks that are nnn bytes in size and
	@@ -1089,10 +1098,15 @@
1089	** The [SQLITE_FCNTL_WIN32_SET_HANDLE] opcode is used for debugging. This
1090	** opcode causes the xFileControl method to swap the file handle with the one
1091	** pointed to by the pArg argument. This capability is used during testing
1092	** and only needs to be supported when SQLITE_TEST is defined.
1093	**





1094	** <li>[[SQLITE_FCNTL_WAL_BLOCK]]
1095	** The [SQLITE_FCNTL_WAL_BLOCK] is a signal to the VFS layer that it might
1096	** be advantageous to block on the next WAL lock if the lock is not immediately
1097	** available. The WAL subsystem issues this signal during rare
1098	** circumstances in order to fix a problem with priority inversion.
	@@ -1242,10 +1256,11 @@
1242	#define SQLITE_FCNTL_RESERVE_BYTES 38
1243	#define SQLITE_FCNTL_CKPT_START 39
1244	#define SQLITE_FCNTL_EXTERNAL_READER 40
1245	#define SQLITE_FCNTL_CKSM_FILE 41
1246	#define SQLITE_FCNTL_RESET_CACHE 42

1247
1248	/* deprecated names */
1249	#define SQLITE_GET_LOCKPROXYFILE SQLITE_FCNTL_GET_LOCKPROXYFILE
1250	#define SQLITE_SET_LOCKPROXYFILE SQLITE_FCNTL_SET_LOCKPROXYFILE
1251	#define SQLITE_LAST_ERRNO SQLITE_FCNTL_LAST_ERRNO
	@@ -2620,14 +2635,18 @@
2620	**
2621	** ^These functions return the number of rows modified, inserted or
2622	** deleted by the most recently completed INSERT, UPDATE or DELETE
2623	** statement on the database connection specified by the only parameter.
2624	** The two functions are identical except for the type of the return value
2625	** and that if the number of rows modified by the most recent INSERT, UPDATE
2626	** or DELETE is greater than the maximum value supported by type "int", then
2627	** the return value of sqlite3_changes() is undefined. ^Executing any other
2628	** type of SQL statement does not modify the value returned by these functions.




2629	**
2630	** ^Only changes made directly by the INSERT, UPDATE or DELETE statement are
2631	** considered - auxiliary changes caused by [CREATE TRIGGER \| triggers],
2632	** [foreign key actions] or [REPLACE] constraint resolution are not counted.
2633	**
	@@ -4183,15 +4202,26 @@
4183	**
4184	** [[SQLITE_PREPARE_NO_VTAB]] <dt>SQLITE_PREPARE_NO_VTAB</dt>
4185	** <dd>The SQLITE_PREPARE_NO_VTAB flag causes the SQL compiler
4186	** to return an error (error code SQLITE_ERROR) if the statement uses
4187	** any virtual tables.










4188	** </dl>
4189	*/
4190	#define SQLITE_PREPARE_PERSISTENT 0x01
4191	#define SQLITE_PREPARE_NORMALIZE 0x02
4192	#define SQLITE_PREPARE_NO_VTAB 0x04

4193
4194	/*
4195	** CAPI3REF: Compiling An SQL Statement
4196	** KEYWORDS: {SQL statement compiler}
4197	** METHOD: sqlite3
	@@ -4220,17 +4250,21 @@
4220	** and sqlite3_prepare_v3()
4221	** interfaces use UTF-8, and sqlite3_prepare16(), sqlite3_prepare16_v2(),
4222	** and sqlite3_prepare16_v3() use UTF-16.
4223	**
4224	** ^If the nByte argument is negative, then zSql is read up to the
4225	** first zero terminator. ^If nByte is positive, then it is the
4226	** number of bytes read from zSql. ^If nByte is zero, then no prepared


4227	** statement is generated.
4228	** If the caller knows that the supplied string is nul-terminated, then
4229	** there is a small performance advantage to passing an nByte parameter that
4230	** is the number of bytes in the input string <i>including</i>
4231	** the nul-terminator.


4232	**
4233	** ^If pzTail is not NULL then *pzTail is made to point to the first byte
4234	** past the end of the first SQL statement in zSql. These routines only
4235	** compile the first statement in zSql, so *pzTail is left pointing to
4236	** what remains uncompiled.
	@@ -5597,11 +5631,11 @@
5597	** The SQLITE_SUBTYPE flag indicates to SQLite that a function might call
5598	** [sqlite3_value_subtype()] to inspect the sub-types of its arguments.
5599	** This flag instructs SQLite to omit some corner-case optimizations that
5600	** might disrupt the operation of the [sqlite3_value_subtype()] function,
5601	** causing it to return zero rather than the correct subtype().
5602	** SQL functions that invokes [sqlite3_value_subtype()] should have this
5603	** property. If the SQLITE_SUBTYPE property is omitted, then the return
5604	** value from [sqlite3_value_subtype()] might sometimes be zero even though
5605	** a non-zero subtype was specified by the function argument expression.
5606	**
5607	** [[SQLITE_RESULT_SUBTYPE]] <dt>SQLITE_RESULT_SUBTYPE</dt><dd>
	@@ -8362,11 +8396,11 @@
8362	#define SQLITE_TESTCTRL_EXTRA_SCHEMA_CHECKS 29
8363	#define SQLITE_TESTCTRL_SEEK_COUNT 30
8364	#define SQLITE_TESTCTRL_TRACEFLAGS 31
8365	#define SQLITE_TESTCTRL_TUNE 32
8366	#define SQLITE_TESTCTRL_LOGEST 33
8367	#define SQLITE_TESTCTRL_USELONGDOUBLE 34
8368	#define SQLITE_TESTCTRL_LAST 34 /* Largest TESTCTRL */
8369
8370	/*
8371	** CAPI3REF: SQL Keyword Checking
8372	**
	@@ -9338,10 +9372,20 @@
9338	** threads may safely make multiple concurrent calls to sqlite3_backup_step().
9339	** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
9340	** APIs are not strictly speaking threadsafe. If they are invoked at the
9341	** same time as another thread is invoking sqlite3_backup_step() it is
9342	** possible that they return invalid values.










9343	*/
9344	SQLITE_API sqlite3_backup *sqlite3_backup_init(
9345	sqlite3 pDest, / Destination database handle */
9346	const char zDestName, / Destination database name */
9347	sqlite3 pSource, / Source database handle */
	@@ -10536,10 +10580,18 @@
10536	** schema S in database connection D. ^On success, the
10537	** [sqlite3_snapshot_get(D,S,P)] interface writes a pointer to the newly
10538	** created [sqlite3_snapshot] object into *P and returns SQLITE_OK.
10539	** If there is not already a read-transaction open on schema S when
10540	** this function is called, one is opened automatically.








10541	**
10542	** The following must be true for this function to succeed. If any of
10543	** the following statements are false when sqlite3_snapshot_get() is
10544	** called, SQLITE_ERROR is returned. The final value of *P is undefined
10545	** in this case.
	@@ -10856,11 +10908,11 @@
10856	#endif
10857
10858	#ifdef __cplusplus
10859	} /* End of the 'extern "C"' block */
10860	#endif
10861	#endif /* SQLITE3_H */
10862
10863	/****** Begin file sqlite3rtree.h *******/
10864	/*
10865	** 2010 August 30
10866	**
	@@ -13107,17 +13159,32 @@
13107	** This is used to access token iToken of phrase hit iIdx within the
13108	** current row. If iIdx is less than zero or greater than or equal to the
13109	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
13110	** output variable (*ppToken) is set to point to a buffer containing the
13111	** matching document token, and (*pnToken) to the size of that buffer in
13112	** bytes. This API is not available if the specified token matches a
13113	** prefix query term. In that case both output variables are always set
13114	** to 0.
13115	**
13116	** The output text is not a copy of the document text that was tokenized.
13117	** It is the output of the tokenizer module. For tokendata=1 tables, this
13118	** includes any embedded 0x00 and trailing data.

















13119	**
13120	** This API can be quite slow if used with an FTS5 table created with the
13121	** "detail=none" or "detail=column" option.
13122	**
13123	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
	@@ -13205,11 +13272,10 @@
13205	** CUSTOM TOKENIZERS
13206	**
13207	** Applications may also register custom tokenizer types. A tokenizer
13208	** is registered by providing fts5 with a populated instance of the
13209	** following structure. All structure methods must be defined, setting
13210	**
13211	** any member of the fts5_tokenizer struct to NULL leads to undefined
13212	** behaviour. The structure methods are expected to function as follows:
13213	**
13214	** xCreate:
13215	** This function is used to allocate and initialize a tokenizer instance.
	@@ -13549,5 +13615,6 @@
13549	#endif
13550
13551	#endif /* _FTS5_H */
13552
13553	/****** End of fts5.h *******/

13554

	--- extsrc/sqlite3.h
	+++ extsrc/sqlite3.h
	@@ -144,13 +144,13 @@
144	**
145	** See also: [sqlite3_libversion()],
146	** [sqlite3_libversion_number()], [sqlite3_sourceid()],
147	** [sqlite_version()] and [sqlite_source_id()].
148	*/
149	#define SQLITE_VERSION "3.48.0"
150	#define SQLITE_VERSION_NUMBER 3048000
151	#define SQLITE_SOURCE_ID "2024-12-09 20:46:36 e2bae4143afd07de1ae55a6d2606a3b541a5b94568aa41f6a96e5d1245471653"
152
153	/*
154	** CAPI3REF: Run-Time Library Version Numbers
155	** KEYWORDS: sqlite3_version sqlite3_sourceid
156	**
	@@ -650,10 +650,17 @@
650	**
651	** The SQLITE_IOCAP_BATCH_ATOMIC property means that the underlying
652	** filesystem supports doing multiple write operations atomically when those
653	** write operations are bracketed by [SQLITE_FCNTL_BEGIN_ATOMIC_WRITE] and
654	** [SQLITE_FCNTL_COMMIT_ATOMIC_WRITE].
655	**
656	** The SQLITE_IOCAP_SUBPAGE_READ property means that it is ok to read
657	** from the database file in amounts that are not a multiple of the
658	** page size and that do not begin at a page boundary. Without this
659	** property, SQLite is careful to only do full-page reads and write
660	** on aligned pages, with the one exception that it will do a sub-page
661	** read of the first page to access the database header.
662	*/
663	#define SQLITE_IOCAP_ATOMIC 0x00000001
664	#define SQLITE_IOCAP_ATOMIC512 0x00000002
665	#define SQLITE_IOCAP_ATOMIC1K 0x00000004
666	#define SQLITE_IOCAP_ATOMIC2K 0x00000008
	@@ -666,10 +673,11 @@
673	#define SQLITE_IOCAP_SEQUENTIAL 0x00000400
674	#define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN 0x00000800
675	#define SQLITE_IOCAP_POWERSAFE_OVERWRITE 0x00001000
676	#define SQLITE_IOCAP_IMMUTABLE 0x00002000
677	#define SQLITE_IOCAP_BATCH_ATOMIC 0x00004000
678	#define SQLITE_IOCAP_SUBPAGE_READ 0x00008000
679
680	/*
681	** CAPI3REF: File Locking Levels
682	**
683	** SQLite uses one of these integer values as the second
	@@ -812,10 +820,11 @@
820	** <li> [SQLITE_IOCAP_SEQUENTIAL]
821	** <li> [SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN]
822	** <li> [SQLITE_IOCAP_POWERSAFE_OVERWRITE]
823	** <li> [SQLITE_IOCAP_IMMUTABLE]
824	** <li> [SQLITE_IOCAP_BATCH_ATOMIC]
825	** <li> [SQLITE_IOCAP_SUBPAGE_READ]
826	** </ul>
827	**
828	** The SQLITE_IOCAP_ATOMIC property means that all writes of
829	** any size are atomic. The SQLITE_IOCAP_ATOMICnnn values
830	** mean that writes of blocks that are nnn bytes in size and
	@@ -1089,10 +1098,15 @@
1098	** The [SQLITE_FCNTL_WIN32_SET_HANDLE] opcode is used for debugging. This
1099	** opcode causes the xFileControl method to swap the file handle with the one
1100	** pointed to by the pArg argument. This capability is used during testing
1101	** and only needs to be supported when SQLITE_TEST is defined.
1102	**
1103	** <li>[[SQLITE_FCNTL_NULL_IO]]
1104	** The [SQLITE_FCNTL_NULL_IO] opcode sets the low-level file descriptor
1105	** or file handle for the [sqlite3_file] object such that it will no longer
1106	** read or write to the database file.
1107	**
1108	** <li>[[SQLITE_FCNTL_WAL_BLOCK]]
1109	** The [SQLITE_FCNTL_WAL_BLOCK] is a signal to the VFS layer that it might
1110	** be advantageous to block on the next WAL lock if the lock is not immediately
1111	** available. The WAL subsystem issues this signal during rare
1112	** circumstances in order to fix a problem with priority inversion.
	@@ -1242,10 +1256,11 @@
1256	#define SQLITE_FCNTL_RESERVE_BYTES 38
1257	#define SQLITE_FCNTL_CKPT_START 39
1258	#define SQLITE_FCNTL_EXTERNAL_READER 40
1259	#define SQLITE_FCNTL_CKSM_FILE 41
1260	#define SQLITE_FCNTL_RESET_CACHE 42
1261	#define SQLITE_FCNTL_NULL_IO 43
1262
1263	/* deprecated names */
1264	#define SQLITE_GET_LOCKPROXYFILE SQLITE_FCNTL_GET_LOCKPROXYFILE
1265	#define SQLITE_SET_LOCKPROXYFILE SQLITE_FCNTL_SET_LOCKPROXYFILE
1266	#define SQLITE_LAST_ERRNO SQLITE_FCNTL_LAST_ERRNO
	@@ -2620,14 +2635,18 @@
2635	**
2636	** ^These functions return the number of rows modified, inserted or
2637	** deleted by the most recently completed INSERT, UPDATE or DELETE
2638	** statement on the database connection specified by the only parameter.
2639	** The two functions are identical except for the type of the return value
2640	** and that if the number of rows modified by the most recent INSERT, UPDATE,
2641	** or DELETE is greater than the maximum value supported by type "int", then
2642	** the return value of sqlite3_changes() is undefined. ^Executing any other
2643	** type of SQL statement does not modify the value returned by these functions.
2644	** For the purposes of this interface, a CREATE TABLE AS SELECT statement
2645	** does not count as an INSERT, UPDATE or DELETE statement and hence the rows
2646	** added to the new table by the CREATE TABLE AS SELECT statement are not
2647	** counted.
2648	**
2649	** ^Only changes made directly by the INSERT, UPDATE or DELETE statement are
2650	** considered - auxiliary changes caused by [CREATE TRIGGER \| triggers],
2651	** [foreign key actions] or [REPLACE] constraint resolution are not counted.
2652	**
	@@ -4183,15 +4202,26 @@
4202	**
4203	** [[SQLITE_PREPARE_NO_VTAB]] <dt>SQLITE_PREPARE_NO_VTAB</dt>
4204	** <dd>The SQLITE_PREPARE_NO_VTAB flag causes the SQL compiler
4205	** to return an error (error code SQLITE_ERROR) if the statement uses
4206	** any virtual tables.
4207	**
4208	** [[SQLITE_PREPARE_DONT_LOG]] <dt>SQLITE_PREPARE_DONT_LOG</dt>
4209	** <dd>The SQLITE_PREPARE_DONT_LOG flag prevents SQL compiler
4210	** errors from being sent to the error log defined by
4211	** [SQLITE_CONFIG_LOG]. This can be used, for example, to do test
4212	** compiles to see if some SQL syntax is well-formed, without generating
4213	** messages on the global error log when it is not. If the test compile
4214	** fails, the sqlite3_prepare_v3() call returns the same error indications
4215	** with or without this flag; it just omits the call to [sqlite3_log()] that
4216	** logs the error.
4217	** </dl>
4218	*/
4219	#define SQLITE_PREPARE_PERSISTENT 0x01
4220	#define SQLITE_PREPARE_NORMALIZE 0x02
4221	#define SQLITE_PREPARE_NO_VTAB 0x04
4222	#define SQLITE_PREPARE_DONT_LOG 0x10
4223
4224	/*
4225	** CAPI3REF: Compiling An SQL Statement
4226	** KEYWORDS: {SQL statement compiler}
4227	** METHOD: sqlite3
	@@ -4220,17 +4250,21 @@
4250	** and sqlite3_prepare_v3()
4251	** interfaces use UTF-8, and sqlite3_prepare16(), sqlite3_prepare16_v2(),
4252	** and sqlite3_prepare16_v3() use UTF-16.
4253	**
4254	** ^If the nByte argument is negative, then zSql is read up to the
4255	** first zero terminator. ^If nByte is positive, then it is the maximum
4256	** number of bytes read from zSql. When nByte is positive, zSql is read
4257	** up to the first zero terminator or until the nByte bytes have been read,
4258	** whichever comes first. ^If nByte is zero, then no prepared
4259	** statement is generated.
4260	** If the caller knows that the supplied string is nul-terminated, then
4261	** there is a small performance advantage to passing an nByte parameter that
4262	** is the number of bytes in the input string <i>including</i>
4263	** the nul-terminator.
4264	** Note that nByte measure the length of the input in bytes, not
4265	** characters, even for the UTF-16 interfaces.
4266	**
4267	** ^If pzTail is not NULL then *pzTail is made to point to the first byte
4268	** past the end of the first SQL statement in zSql. These routines only
4269	** compile the first statement in zSql, so *pzTail is left pointing to
4270	** what remains uncompiled.
	@@ -5597,11 +5631,11 @@
5631	** The SQLITE_SUBTYPE flag indicates to SQLite that a function might call
5632	** [sqlite3_value_subtype()] to inspect the sub-types of its arguments.
5633	** This flag instructs SQLite to omit some corner-case optimizations that
5634	** might disrupt the operation of the [sqlite3_value_subtype()] function,
5635	** causing it to return zero rather than the correct subtype().
5636	** All SQL functions that invoke [sqlite3_value_subtype()] should have this
5637	** property. If the SQLITE_SUBTYPE property is omitted, then the return
5638	** value from [sqlite3_value_subtype()] might sometimes be zero even though
5639	** a non-zero subtype was specified by the function argument expression.
5640	**
5641	** [[SQLITE_RESULT_SUBTYPE]] <dt>SQLITE_RESULT_SUBTYPE</dt><dd>
	@@ -8362,11 +8396,11 @@
8396	#define SQLITE_TESTCTRL_EXTRA_SCHEMA_CHECKS 29
8397	#define SQLITE_TESTCTRL_SEEK_COUNT 30
8398	#define SQLITE_TESTCTRL_TRACEFLAGS 31
8399	#define SQLITE_TESTCTRL_TUNE 32
8400	#define SQLITE_TESTCTRL_LOGEST 33
8401	#define SQLITE_TESTCTRL_USELONGDOUBLE 34 /* NOT USED */
8402	#define SQLITE_TESTCTRL_LAST 34 /* Largest TESTCTRL */
8403
8404	/*
8405	** CAPI3REF: SQL Keyword Checking
8406	**
	@@ -9338,10 +9372,20 @@
9372	** threads may safely make multiple concurrent calls to sqlite3_backup_step().
9373	** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
9374	** APIs are not strictly speaking threadsafe. If they are invoked at the
9375	** same time as another thread is invoking sqlite3_backup_step() it is
9376	** possible that they return invalid values.
9377	**
9378	** <b>Alternatives To Using The Backup API</b>
9379	**
9380	** Other techniques for safely creating a consistent backup of an SQLite
9381	** database include:
9382	**
9383	** <ul>
9384	** <li> The [VACUUM INTO] command.
9385	** <li> The [sqlite3_rsync] utility program.
9386	** </ul>
9387	*/
9388	SQLITE_API sqlite3_backup *sqlite3_backup_init(
9389	sqlite3 pDest, / Destination database handle */
9390	const char zDestName, / Destination database name */
9391	sqlite3 pSource, / Source database handle */
	@@ -10536,10 +10580,18 @@
10580	** schema S in database connection D. ^On success, the
10581	** [sqlite3_snapshot_get(D,S,P)] interface writes a pointer to the newly
10582	** created [sqlite3_snapshot] object into *P and returns SQLITE_OK.
10583	** If there is not already a read-transaction open on schema S when
10584	** this function is called, one is opened automatically.
10585	**
10586	** If a read-transaction is opened by this function, then it is guaranteed
10587	** that the returned snapshot object may not be invalidated by a database
10588	** writer or checkpointer until after the read-transaction is closed. This
10589	** is not guaranteed if a read-transaction is already open when this
10590	** function is called. In that case, any subsequent write or checkpoint
10591	** operation on the database may invalidate the returned snapshot handle,
10592	** even while the read-transaction remains open.
10593	**
10594	** The following must be true for this function to succeed. If any of
10595	** the following statements are false when sqlite3_snapshot_get() is
10596	** called, SQLITE_ERROR is returned. The final value of *P is undefined
10597	** in this case.
	@@ -10856,11 +10908,11 @@
10908	#endif
10909
10910	#ifdef __cplusplus
10911	} /* End of the 'extern "C"' block */
10912	#endif
10913	/* #endif for SQLITE3_H will be added by mksqlite3.tcl */
10914
10915	/****** Begin file sqlite3rtree.h *******/
10916	/*
10917	** 2010 August 30
10918	**
	@@ -13107,17 +13159,32 @@
13159	** This is used to access token iToken of phrase hit iIdx within the
13160	** current row. If iIdx is less than zero or greater than or equal to the
13161	** value returned by xInstCount(), SQLITE_RANGE is returned. Otherwise,
13162	** output variable (*ppToken) is set to point to a buffer containing the
13163	** matching document token, and (*pnToken) to the size of that buffer in
13164	** bytes.


13165	**
13166	** The output text is not a copy of the document text that was tokenized.
13167	** It is the output of the tokenizer module. For tokendata=1 tables, this
13168	** includes any embedded 0x00 and trailing data.
13169	**
13170	** This API may be slow in some cases if the token identified by parameters
13171	** iIdx and iToken matched a prefix token in the query. In most cases, the
13172	** first call to this API for each prefix token in the query is forced
13173	** to scan the portion of the full-text index that matches the prefix
13174	** token to collect the extra data required by this API. If the prefix
13175	** token matches a large number of token instances in the document set,
13176	** this may be a performance problem.
13177	**
13178	** If the user knows in advance that a query may use this API for a
13179	** prefix token, FTS5 may be configured to collect all required data as part
13180	** of the initial querying of the full-text index, avoiding the second scan
13181	** entirely. This also causes prefix queries that do not use this API to
13182	** run more slowly and use more memory. FTS5 may be configured in this way
13183	** either on a per-table basis using the [FTS5 insttoken \| 'insttoken']
13184	** option, or on a per-query basis using the
13185	** [fts5_insttoken \| fts5_insttoken()] user function.
13186	**
13187	** This API can be quite slow if used with an FTS5 table created with the
13188	** "detail=none" or "detail=column" option.
13189	**
13190	** xColumnLocale(pFts5, iIdx, pzLocale, pnLocale)
	@@ -13205,11 +13272,10 @@
13272	** CUSTOM TOKENIZERS
13273	**
13274	** Applications may also register custom tokenizer types. A tokenizer
13275	** is registered by providing fts5 with a populated instance of the
13276	** following structure. All structure methods must be defined, setting

13277	** any member of the fts5_tokenizer struct to NULL leads to undefined
13278	** behaviour. The structure methods are expected to function as follows:
13279	**
13280	** xCreate:
13281	** This function is used to allocate and initialize a tokenizer instance.
	@@ -13549,5 +13615,6 @@
13615	#endif
13616
13617	#endif /* _FTS5_H */
13618
13619	/****** End of fts5.h *******/
13620	#endif /* SQLITE3_H */
13621

M fossil.1

+22 -17

		--- fossil.1
		+++ fossil.1
		@@ -1,6 +1,6 @@
1		-.TH FOSSIL "1" "July 2021" "https://fossil-scm.org" "User Commands"
	1	+.TH FOSSIL "1" "Oct 2024" "https://fossil-scm.org" "User Commands"
2	2	.SH NAME
3	3	fossil \- Distributed Version Control System
4	4	.SH SYNOPSIS
5	5	.B fossil
6	6	\fIhelp\fR
		@@ -14,26 +14,31 @@
14	14	Fossil is a distributed version control system (DVCS) with built-in
15	15	forum, wiki, ticket tracker, CGI/HTTP interface, and HTTP server.
16	16
17	17	.SH Common COMMANDs:
18	18
19		-add cat diff ls revert timeline
20		-.br
21		-addremove changes extras merge rm ui
22		-.br
23		-all chat finfo mv settings undo
24		-.br
25		-amend clean gdiff open sql unversioned
26		-.br
27		-annotate clone grep pull stash update
28		-.br
29		-bisect commit help push status version
30		-.br
31		-blame dbstat info rebuild sync
32		-.br
33		-branch delete init remote tag
34		-.br
	19	+add cherrypick grep push sync
	20	+.br
	21	+addremove clean help rebuild tag
	22	+.br
	23	+all clone info remote timeline
	24	+.br
	25	+amend commit init repack tree
	26	+.br
	27	+annotate dbstat ls revert ui
	28	+.br
	29	+bisect delete merge rm undo
	30	+.br
	31	+blame describe merge-base settings unversioned
	32	+.br
	33	+branch diff mv sql update
	34	+.br
	35	+cat extras open ssl-config version
	36	+.br
	37	+changes finfo patch stash xdiff
	38	+.br
	39	+chat gdiff pull status
35	40
36	41	.SH FEATURES
37	42
38	43	Features as described on the fossil home page.
39	44
40	45

	--- fossil.1
	+++ fossil.1
	@@ -1,6 +1,6 @@
1	.TH FOSSIL "1" "July 2021" "https://fossil-scm.org" "User Commands"
2	.SH NAME
3	fossil \- Distributed Version Control System
4	.SH SYNOPSIS
5	.B fossil
6	\fIhelp\fR
	@@ -14,26 +14,31 @@
14	Fossil is a distributed version control system (DVCS) with built-in
15	forum, wiki, ticket tracker, CGI/HTTP interface, and HTTP server.
16
17	.SH Common COMMANDs:
18
19	add cat diff ls revert timeline
20	.br
21	addremove changes extras merge rm ui
22	.br
23	all chat finfo mv settings undo
24	.br
25	amend clean gdiff open sql unversioned
26	.br
27	annotate clone grep pull stash update
28	.br
29	bisect commit help push status version
30	.br
31	blame dbstat info rebuild sync
32	.br
33	branch delete init remote tag
34	.br





35
36	.SH FEATURES
37
38	Features as described on the fossil home page.
39
40

	--- fossil.1
	+++ fossil.1
	@@ -1,6 +1,6 @@
1	.TH FOSSIL "1" "Oct 2024" "https://fossil-scm.org" "User Commands"
2	.SH NAME
3	fossil \- Distributed Version Control System
4	.SH SYNOPSIS
5	.B fossil
6	\fIhelp\fR
	@@ -14,26 +14,31 @@
14	Fossil is a distributed version control system (DVCS) with built-in
15	forum, wiki, ticket tracker, CGI/HTTP interface, and HTTP server.
16
17	.SH Common COMMANDs:
18
19	add cherrypick grep push sync
20	.br
21	addremove clean help rebuild tag
22	.br
23	all clone info remote timeline
24	.br
25	amend commit init repack tree
26	.br
27	annotate dbstat ls revert ui
28	.br
29	bisect delete merge rm undo
30	.br
31	blame describe merge-base settings unversioned
32	.br
33	branch diff mv sql update
34	.br
35	cat extras open ssl-config version
36	.br
37	changes finfo patch stash xdiff
38	.br
39	chat gdiff pull status
40
41	.SH FEATURES
42
43	Features as described on the fossil home page.
44
45

M skins/darkmode/css.txt

+4 -1

		--- skins/darkmode/css.txt
		+++ skins/darkmode/css.txt
		@@ -98,14 +98,17 @@
98	98	}
99	99	.fileage tr:hover,
100	100	div.filetreeline:hover {
101	101	background-color: #333;
102	102	}
	103	+div.file-change-line button {
	104	+ background-color: #484848
	105	+}
103	106	.button,
104	107	button {
105	108	color: #aaa;
106		- background-color: #444;
	109	+ background-color: #484848;
107	110	border-radius: 5px;
108	111	border: 0
109	112	}
110	113	.button:hover,
111	114	button:hover {
112	115

	--- skins/darkmode/css.txt
	+++ skins/darkmode/css.txt
	@@ -98,14 +98,17 @@
98	}
99	.fileage tr:hover,
100	div.filetreeline:hover {
101	background-color: #333;
102	}



103	.button,
104	button {
105	color: #aaa;
106	background-color: #444;
107	border-radius: 5px;
108	border: 0
109	}
110	.button:hover,
111	button:hover {
112

	--- skins/darkmode/css.txt
	+++ skins/darkmode/css.txt
	@@ -98,14 +98,17 @@
98	}
99	.fileage tr:hover,
100	div.filetreeline:hover {
101	background-color: #333;
102	}
103	div.file-change-line button {
104	background-color: #484848
105	}
106	.button,
107	button {
108	color: #aaa;
109	background-color: #484848;
110	border-radius: 5px;
111	border: 0
112	}
113	.button:hover,
114	button:hover {
115

M skins/default/css.txt

+7 -6

		--- skins/default/css.txt
		+++ skins/default/css.txt
		@@ -42,10 +42,11 @@
42	42	.artifact h1.page-title,
43	43	.dir h1.page-title,
44	44	.doc h1.page-title,
45	45	.wiki h1.page-title {
46	46	display: block; /* …for potentially long doc titles… */
	47	+ color: #444;
47	48	}
48	49	.artifact .title > .page-title,
49	50	.dir .title > .page-title,
50	51	.doc .title > .page-title,
51	52	.wiki .title > .page-title {
		@@ -725,16 +726,16 @@
725	726	margin-left: 20pt; /* special case for MD in forum; need less indent */
726	727	}
727	728
728	729	/* Fossil UI uses these, but in sufficiently constrained ways that we
729	730	* don't have to be nearly as careful to avoid an overreach. */
730		- .doc > .content h1, .artifact h1, .dir h1, .fileedit h1, .wiki h1 { margin-left: 10pt; }
731		- .doc > .content h2, .artifact h2, .dir h2, .fileedit h2, .wiki h2 { margin-left: 20pt; }
732		- .doc > .content h3, .artifact h3, .dir h3, .fileedit h3, .wiki h3 { margin-left: 30pt; }
733		- .doc > .content h4, .artifact h4, .dir h4, .fileedit h4, .wiki h4 { margin-left: 40pt; }
734		- .doc > .content h5, .artifact h5, .dir h5, .fileedit h5, .wiki h5 { margin-left: 50pt; }
735		- .doc > .content hr, .artifact hr, .dir hr, .fileedit hr, .wiki hr { margin-left: 10pt; }
	731	+ .doc > .content h1, .artifact .content h1, .dir .content h1, .fileedit .content h1, .wiki .content h1 { margin-left: 10pt; }
	732	+ .doc > .content h2, .artifact .content h2, .dir .content h2, .fileedit .content h2, .wiki .content h2 { margin-left: 20pt; }
	733	+ .doc > .content h3, .artifact .content h3, .dir .content h3, .fileedit .content h3, .wiki .content h3 { margin-left: 30pt; }
	734	+ .doc > .content h4, .artifact .content h4, .dir .content h4, .fileedit .content h4, .wiki .content h4 { margin-left: 40pt; }
	735	+ .doc > .content h5, .artifact .content h5, .dir .content h5, .fileedit .content h5, .wiki .content h5 { margin-left: 50pt; }
	736	+ .doc > .content hr, .artifact .content hr, .dir .content hr, .fileedit .content hr, .wiki .content hr { margin-left: 10pt; }
736	737
737	738	/* Don't need to be nearly as careful with tags Fossil UI doesn't use. */
738	739	.doc dd, .artifact dd, .dir dd, .fileedit dd, .wikiedit dd { margin-left: 30pt; margin-bottom: 1em; }
739	740	.doc dl, .artifact dl, .dir dl, .fileedit dl, .wikiedit dl { margin-left: 60pt; }
740	741	.doc dt, .artifact dt, .dir dt, .fileedit dt, .wikiedit dt { margin-left: 10pt; }
741	742

	--- skins/default/css.txt
	+++ skins/default/css.txt
	@@ -42,10 +42,11 @@
42	.artifact h1.page-title,
43	.dir h1.page-title,
44	.doc h1.page-title,
45	.wiki h1.page-title {
46	display: block; /* …for potentially long doc titles… */

47	}
48	.artifact .title > .page-title,
49	.dir .title > .page-title,
50	.doc .title > .page-title,
51	.wiki .title > .page-title {
	@@ -725,16 +726,16 @@
725	margin-left: 20pt; /* special case for MD in forum; need less indent */
726	}
727
728	/* Fossil UI uses these, but in sufficiently constrained ways that we
729	* don't have to be nearly as careful to avoid an overreach. */
730	.doc > .content h1, .artifact h1, .dir h1, .fileedit h1, .wiki h1 { margin-left: 10pt; }
731	.doc > .content h2, .artifact h2, .dir h2, .fileedit h2, .wiki h2 { margin-left: 20pt; }
732	.doc > .content h3, .artifact h3, .dir h3, .fileedit h3, .wiki h3 { margin-left: 30pt; }
733	.doc > .content h4, .artifact h4, .dir h4, .fileedit h4, .wiki h4 { margin-left: 40pt; }
734	.doc > .content h5, .artifact h5, .dir h5, .fileedit h5, .wiki h5 { margin-left: 50pt; }
735	.doc > .content hr, .artifact hr, .dir hr, .fileedit hr, .wiki hr { margin-left: 10pt; }
736
737	/* Don't need to be nearly as careful with tags Fossil UI doesn't use. */
738	.doc dd, .artifact dd, .dir dd, .fileedit dd, .wikiedit dd { margin-left: 30pt; margin-bottom: 1em; }
739	.doc dl, .artifact dl, .dir dl, .fileedit dl, .wikiedit dl { margin-left: 60pt; }
740	.doc dt, .artifact dt, .dir dt, .fileedit dt, .wikiedit dt { margin-left: 10pt; }
741

	--- skins/default/css.txt
	+++ skins/default/css.txt
	@@ -42,10 +42,11 @@
42	.artifact h1.page-title,
43	.dir h1.page-title,
44	.doc h1.page-title,
45	.wiki h1.page-title {
46	display: block; /* …for potentially long doc titles… */
47	color: #444;
48	}
49	.artifact .title > .page-title,
50	.dir .title > .page-title,
51	.doc .title > .page-title,
52	.wiki .title > .page-title {
	@@ -725,16 +726,16 @@
726	margin-left: 20pt; /* special case for MD in forum; need less indent */
727	}
728
729	/* Fossil UI uses these, but in sufficiently constrained ways that we
730	* don't have to be nearly as careful to avoid an overreach. */
731	.doc > .content h1, .artifact .content h1, .dir .content h1, .fileedit .content h1, .wiki .content h1 { margin-left: 10pt; }
732	.doc > .content h2, .artifact .content h2, .dir .content h2, .fileedit .content h2, .wiki .content h2 { margin-left: 20pt; }
733	.doc > .content h3, .artifact .content h3, .dir .content h3, .fileedit .content h3, .wiki .content h3 { margin-left: 30pt; }
734	.doc > .content h4, .artifact .content h4, .dir .content h4, .fileedit .content h4, .wiki .content h4 { margin-left: 40pt; }
735	.doc > .content h5, .artifact .content h5, .dir .content h5, .fileedit .content h5, .wiki .content h5 { margin-left: 50pt; }
736	.doc > .content hr, .artifact .content hr, .dir .content hr, .fileedit .content hr, .wiki .content hr { margin-left: 10pt; }
737
738	/* Don't need to be nearly as careful with tags Fossil UI doesn't use. */
739	.doc dd, .artifact dd, .dir dd, .fileedit dd, .wikiedit dd { margin-left: 30pt; margin-bottom: 1em; }
740	.doc dl, .artifact dl, .dir dl, .fileedit dl, .wikiedit dl { margin-left: 60pt; }
741	.doc dt, .artifact dt, .dir dt, .fileedit dt, .wikiedit dt { margin-left: 10pt; }
742

M src/add.c

+1 -1

		--- src/add.c
		+++ src/add.c
		@@ -353,11 +353,11 @@
353	353	** for files to be excluded. Example: '.o,.obj,*.exe' If the --ignore
354	354	** option does not appear on the command line then the "ignore-glob" setting
355	355	** is used. If the --clean option does not appear on the command line then
356	356	** the "clean-glob" setting is used.
357	357	**
358		-** If files are attempted to be added explicitly on the command line which
	358	+** When attempting to explicitly add files on the commandline, and if those
359	359	** match "ignore-glob", a confirmation is asked first. This can be prevented
360	360	** using the -f\|--force option.
361	361	**
362	362	** The --case-sensitive option determines whether or not filenames should
363	363	** be treated case sensitive or not. If the option is not given, the default
364	364

	--- src/add.c
	+++ src/add.c
	@@ -353,11 +353,11 @@
353	** for files to be excluded. Example: '.o,.obj,*.exe' If the --ignore
354	** option does not appear on the command line then the "ignore-glob" setting
355	** is used. If the --clean option does not appear on the command line then
356	** the "clean-glob" setting is used.
357	**
358	** If files are attempted to be added explicitly on the command line which
359	** match "ignore-glob", a confirmation is asked first. This can be prevented
360	** using the -f\|--force option.
361	**
362	** The --case-sensitive option determines whether or not filenames should
363	** be treated case sensitive or not. If the option is not given, the default
364

	--- src/add.c
	+++ src/add.c
	@@ -353,11 +353,11 @@
353	** for files to be excluded. Example: '.o,.obj,*.exe' If the --ignore
354	** option does not appear on the command line then the "ignore-glob" setting
355	** is used. If the --clean option does not appear on the command line then
356	** the "clean-glob" setting is used.
357	**
358	** When attempting to explicitly add files on the commandline, and if those
359	** match "ignore-glob", a confirmation is asked first. This can be prevented
360	** using the -f\|--force option.
361	**
362	** The --case-sensitive option determines whether or not filenames should
363	** be treated case sensitive or not. If the option is not given, the default
364

M src/allrepo.c

+23 -6

		--- src/allrepo.c
		+++ src/allrepo.c
		@@ -50,11 +50,10 @@
50	50	for(i=iStart; i<g.argc; i++){
51	51	blob_appendf(pExtra, " %s", g.argv[i]);
52	52	}
53	53	}
54	54
55		-
56	55	/*
57	56	** COMMAND: all
58	57	**
59	58	** Usage: %fossil all SUBCOMMAND ...
60	59	**
		@@ -429,44 +428,62 @@
429	428	"add cache changes clean dbstat extras fts-config git ignore "
430	429	"info list ls pull push rebuild remote "
431	430	"server settings sync ui unset whatis");
432	431	}
433	432	verify_all_options();
434		- db_multi_exec("CREATE TEMP TABLE repolist(name,tag);");
	433	+ db_multi_exec(
	434	+ "CREATE TEMP TABLE repolist(\n"
	435	+ " name TEXT, -- Filename\n"
	436	+ " tag TEXT, -- Key for the GLOBAL_CONFIG table entry\n"
	437	+ " inode TEXT -- Unique identifier for this file\n"
	438	+ ");\n"
	439	+
	440	+ /* The seenFile() table holds inode names for entries that have
	441	+ ** already been processed. */
	442	+ "CREATE TEMP TABLE seenFile(x TEXT COLLATE nocase);\n"
	443	+
	444	+ /* The toDel() table holds the "tag" for entries that need to be
	445	+ ** deleted because they are redundant or no longer exist */
	446	+ "CREATE TEMP TABLE toDel(x TEXT);\n"
	447	+ );
	448	+ sqlite3_create_function(g.db, "inode", 1, SQLITE_UTF8, 0,
	449	+ file_inode_sql_func, 0, 0);
435	450	if( useCheckouts ){
436	451	db_multi_exec(
437	452	"INSERT INTO repolist "
438		- "SELECT DISTINCT substr(name, 7), name COLLATE nocase"
	453	+ "SELECT substr(name, 7), name, inode(substr(name,7))"
439	454	" FROM global_config"
440	455	" WHERE substr(name, 1, 6)=='ckout:'"
441	456	" ORDER BY 1"
442	457	);
443	458	}else{
444	459	db_multi_exec(
445	460	"INSERT INTO repolist "
446		- "SELECT DISTINCT substr(name, 6), name COLLATE nocase"
	461	+ "SELECT substr(name, 6), name, inode(substr(name,6))"
447	462	" FROM global_config"
448	463	" WHERE substr(name, 1, 5)=='repo:'"
449	464	" ORDER BY 1"
450	465	);
451	466	}
452		- db_multi_exec("CREATE TEMP TABLE toDel(x TEXT)");
453		- db_prepare(&q, "SELECT name, tag FROM repolist ORDER BY 1");
	467	+ db_prepare(&q,"SELECT name, tag, inode FROM repolist ORDER BY 1");
454	468	while( db_step(&q)==SQLITE_ROW ){
455	469	int rc;
456	470	const char *zFilename = db_column_text(&q, 0);
	471	+ const char *zInode = db_column_text(&q,2);
457	472	#if !USE_SEE
458	473	if( sqlite3_strglob("*.efossil", zFilename)==0 ) continue;
459	474	#endif
460	475	if( file_access(zFilename, F_OK)
461	476	\|\| !file_is_canonical(zFilename)
462	477	\|\| (useCheckouts && file_isdir(zFilename, ExtFILE)!=1)
	478	+ \|\| db_exists("SELECT 1 FROM temp.seenFile where x=%Q", zInode)
463	479	){
464	480	db_multi_exec("INSERT INTO toDel VALUES(%Q)", db_column_text(&q, 1));
465	481	nToDel++;
466	482	continue;
467	483	}
	484	+ db_multi_exec("INSERT INTO seenFile(x) VALUES(%Q)", zInode);
468	485	if( zCmd[0]=='l' ){
469	486	fossil_print("%s\n", zFilename);
470	487	continue;
471	488	}else if( showFile ){
472	489	fossil_print("%s: %s\n", useCheckouts ? "check-out" : "repository",
473	490

	--- src/allrepo.c
	+++ src/allrepo.c
	@@ -50,11 +50,10 @@
50	for(i=iStart; i<g.argc; i++){
51	blob_appendf(pExtra, " %s", g.argv[i]);
52	}
53	}
54
55
56	/*
57	** COMMAND: all
58	**
59	** Usage: %fossil all SUBCOMMAND ...
60	**
	@@ -429,44 +428,62 @@
429	"add cache changes clean dbstat extras fts-config git ignore "
430	"info list ls pull push rebuild remote "
431	"server settings sync ui unset whatis");
432	}
433	verify_all_options();
434	db_multi_exec("CREATE TEMP TABLE repolist(name,tag);");
















435	if( useCheckouts ){
436	db_multi_exec(
437	"INSERT INTO repolist "
438	"SELECT DISTINCT substr(name, 7), name COLLATE nocase"
439	" FROM global_config"
440	" WHERE substr(name, 1, 6)=='ckout:'"
441	" ORDER BY 1"
442	);
443	}else{
444	db_multi_exec(
445	"INSERT INTO repolist "
446	"SELECT DISTINCT substr(name, 6), name COLLATE nocase"
447	" FROM global_config"
448	" WHERE substr(name, 1, 5)=='repo:'"
449	" ORDER BY 1"
450	);
451	}
452	db_multi_exec("CREATE TEMP TABLE toDel(x TEXT)");
453	db_prepare(&q, "SELECT name, tag FROM repolist ORDER BY 1");
454	while( db_step(&q)==SQLITE_ROW ){
455	int rc;
456	const char *zFilename = db_column_text(&q, 0);

457	#if !USE_SEE
458	if( sqlite3_strglob("*.efossil", zFilename)==0 ) continue;
459	#endif
460	if( file_access(zFilename, F_OK)
461	\|\| !file_is_canonical(zFilename)
462	\|\| (useCheckouts && file_isdir(zFilename, ExtFILE)!=1)

463	){
464	db_multi_exec("INSERT INTO toDel VALUES(%Q)", db_column_text(&q, 1));
465	nToDel++;
466	continue;
467	}

468	if( zCmd[0]=='l' ){
469	fossil_print("%s\n", zFilename);
470	continue;
471	}else if( showFile ){
472	fossil_print("%s: %s\n", useCheckouts ? "check-out" : "repository",
473

	--- src/allrepo.c
	+++ src/allrepo.c
	@@ -50,11 +50,10 @@
50	for(i=iStart; i<g.argc; i++){
51	blob_appendf(pExtra, " %s", g.argv[i]);
52	}
53	}
54

55	/*
56	** COMMAND: all
57	**
58	** Usage: %fossil all SUBCOMMAND ...
59	**
	@@ -429,44 +428,62 @@
428	"add cache changes clean dbstat extras fts-config git ignore "
429	"info list ls pull push rebuild remote "
430	"server settings sync ui unset whatis");
431	}
432	verify_all_options();
433	db_multi_exec(
434	"CREATE TEMP TABLE repolist(\n"
435	" name TEXT, -- Filename\n"
436	" tag TEXT, -- Key for the GLOBAL_CONFIG table entry\n"
437	" inode TEXT -- Unique identifier for this file\n"
438	");\n"
439
440	/* The seenFile() table holds inode names for entries that have
441	** already been processed. */
442	"CREATE TEMP TABLE seenFile(x TEXT COLLATE nocase);\n"
443
444	/* The toDel() table holds the "tag" for entries that need to be
445	** deleted because they are redundant or no longer exist */
446	"CREATE TEMP TABLE toDel(x TEXT);\n"
447	);
448	sqlite3_create_function(g.db, "inode", 1, SQLITE_UTF8, 0,
449	file_inode_sql_func, 0, 0);
450	if( useCheckouts ){
451	db_multi_exec(
452	"INSERT INTO repolist "
453	"SELECT substr(name, 7), name, inode(substr(name,7))"
454	" FROM global_config"
455	" WHERE substr(name, 1, 6)=='ckout:'"
456	" ORDER BY 1"
457	);
458	}else{
459	db_multi_exec(
460	"INSERT INTO repolist "
461	"SELECT substr(name, 6), name, inode(substr(name,6))"
462	" FROM global_config"
463	" WHERE substr(name, 1, 5)=='repo:'"
464	" ORDER BY 1"
465	);
466	}
467	db_prepare(&q,"SELECT name, tag, inode FROM repolist ORDER BY 1");

468	while( db_step(&q)==SQLITE_ROW ){
469	int rc;
470	const char *zFilename = db_column_text(&q, 0);
471	const char *zInode = db_column_text(&q,2);
472	#if !USE_SEE
473	if( sqlite3_strglob("*.efossil", zFilename)==0 ) continue;
474	#endif
475	if( file_access(zFilename, F_OK)
476	\|\| !file_is_canonical(zFilename)
477	\|\| (useCheckouts && file_isdir(zFilename, ExtFILE)!=1)
478	\|\| db_exists("SELECT 1 FROM temp.seenFile where x=%Q", zInode)
479	){
480	db_multi_exec("INSERT INTO toDel VALUES(%Q)", db_column_text(&q, 1));
481	nToDel++;
482	continue;
483	}
484	db_multi_exec("INSERT INTO seenFile(x) VALUES(%Q)", zInode);
485	if( zCmd[0]=='l' ){
486	fossil_print("%s\n", zFilename);
487	continue;
488	}else if( showFile ){
489	fossil_print("%s: %s\n", useCheckouts ? "check-out" : "repository",
490

M src/attach.c

+8 -1

		--- src/attach.c
		+++ src/attach.c
		@@ -634,11 +634,12 @@
634	634	/*
635	635	** Output HTML to show a list of attachments.
636	636	*/
637	637	void attachment_list(
638	638	const char zTarget, / Object that things are attached to */
639		- const char zHeader / Header to display with attachments */
	639	+ const char zHeader, / Header to display with attachments */
	640	+ int fHorizontalRule /* Insert <hr> separator above header */
640	641	){
641	642	int cnt = 0;
642	643	Stmt q;
643	644	db_prepare(&q,
644	645	"SELECT datetime(mtime,toLocal()), filename, user,"
		@@ -654,11 +655,16 @@
654	655	const char *zUser = db_column_text(&q, 2);
655	656	const char *zUuid = db_column_text(&q, 3);
656	657	const char *zSrc = db_column_text(&q, 4);
657	658	const char *zDispUser = zUser && zUser[0] ? zUser : "anonymous";
658	659	if( cnt==0 ){
	660	+ @ <section class='attachlist'>
	661	+ if( fHorizontalRule ){
	662	+ @ <hr>
	663	+ }
659	664	@ %s(zHeader)
	665	+ @ <ul>
660	666	}
661	667	cnt++;
662	668	@ <li>
663	669	@ %z(href("%R/artifact/%!S",zSrc))%h(zFile)</a>
664	670	@ [<a href="%R/attachdownload/%t(zFile)?page=%t(zTarget)&file=%t(zFile)">download</a>]
		@@ -667,10 +673,11 @@
667	673	@ [%z(href("%R/ainfo/%!S",zUuid))details</a>]
668	674	@ </li>
669	675	}
670	676	if( cnt ){
671	677	@ </ul>
	678	+ @ </section>
672	679	}
673	680	db_finalize(&q);
674	681
675	682	}
676	683
677	684

	--- src/attach.c
	+++ src/attach.c
	@@ -634,11 +634,12 @@
634	/*
635	** Output HTML to show a list of attachments.
636	*/
637	void attachment_list(
638	const char zTarget, / Object that things are attached to */
639	const char zHeader / Header to display with attachments */

640	){
641	int cnt = 0;
642	Stmt q;
643	db_prepare(&q,
644	"SELECT datetime(mtime,toLocal()), filename, user,"
	@@ -654,11 +655,16 @@
654	const char *zUser = db_column_text(&q, 2);
655	const char *zUuid = db_column_text(&q, 3);
656	const char *zSrc = db_column_text(&q, 4);
657	const char *zDispUser = zUser && zUser[0] ? zUser : "anonymous";
658	if( cnt==0 ){




659	@ %s(zHeader)

660	}
661	cnt++;
662	@ <li>
663	@ %z(href("%R/artifact/%!S",zSrc))%h(zFile)</a>
664	@ [<a href="%R/attachdownload/%t(zFile)?page=%t(zTarget)&file=%t(zFile)">download</a>]
	@@ -667,10 +673,11 @@
667	@ [%z(href("%R/ainfo/%!S",zUuid))details</a>]
668	@ </li>
669	}
670	if( cnt ){
671	@ </ul>

672	}
673	db_finalize(&q);
674
675	}
676
677

	--- src/attach.c
	+++ src/attach.c
	@@ -634,11 +634,12 @@
634	/*
635	** Output HTML to show a list of attachments.
636	*/
637	void attachment_list(
638	const char zTarget, / Object that things are attached to */
639	const char zHeader, / Header to display with attachments */
640	int fHorizontalRule /* Insert <hr> separator above header */
641	){
642	int cnt = 0;
643	Stmt q;
644	db_prepare(&q,
645	"SELECT datetime(mtime,toLocal()), filename, user,"
	@@ -654,11 +655,16 @@
655	const char *zUser = db_column_text(&q, 2);
656	const char *zUuid = db_column_text(&q, 3);
657	const char *zSrc = db_column_text(&q, 4);
658	const char *zDispUser = zUser && zUser[0] ? zUser : "anonymous";
659	if( cnt==0 ){
660	@ <section class='attachlist'>
661	if( fHorizontalRule ){
662	@ <hr>
663	}
664	@ %s(zHeader)
665	@ <ul>
666	}
667	cnt++;
668	@ <li>
669	@ %z(href("%R/artifact/%!S",zSrc))%h(zFile)</a>
670	@ [<a href="%R/attachdownload/%t(zFile)?page=%t(zTarget)&file=%t(zFile)">download</a>]
	@@ -667,10 +673,11 @@
673	@ [%z(href("%R/ainfo/%!S",zUuid))details</a>]
674	@ </li>
675	}
676	if( cnt ){
677	@ </ul>
678	@ </section>
679	}
680	db_finalize(&q);
681
682	}
683
684

M src/bisect.c

+12 -4

		--- src/bisect.c
		+++ src/bisect.c
		@@ -543,14 +543,16 @@
543	543	**
544	544	** > fossil bisect vlist\|ls\|status ?-a\|--all?
545	545	**
546	546	** List the versions in between the inner-most "bad" and "good".
547	547	**
548		-** > fossil bisect ui
	548	+** > fossil bisect ui ?HOST@USER:PATH?
549	549	**
550	550	** Like "fossil ui" except start on a timeline that shows only the
551		-** check-ins that are part of the current bisect.
	551	+** check-ins that are part of the current bisect. If the optional
	552	+** fourth term is added, then information is shown for the bisect that
	553	+** occurred in the PATH directory by USER on remote machine HOST.
552	554	**
553	555	** > fossil bisect undo
554	556	**
555	557	** Undo the most recent "good", "bad", or "skip" command.
556	558	*/
		@@ -719,17 +721,23 @@
719	721	}
720	722	}else if( strncmp(zCmd, "reset", n)==0 ){
721	723	bisect_reset();
722	724	}else if( strcmp(zCmd, "ui")==0 ){
723	725	char *newArgv[8];
	726	+ verify_all_options();
724	727	newArgv[0] = g.argv[0];
725	728	newArgv[1] = "ui";
726	729	newArgv[2] = "--page";
727	730	newArgv[3] = "timeline?bisect";
728		- newArgv[4] = 0;
	731	+ if( g.argc==4 ){
	732	+ newArgv[4] = g.argv[3];
	733	+ g.argc = 5;
	734	+ }else{
	735	+ g.argc = 4;
	736	+ }
	737	+ newArgv[g.argc] = 0;
729	738	g.argv = newArgv;
730		- g.argc = 4;
731	739	cmd_webserver();
732	740	}else if( strncmp(zCmd, "vlist", n)==0
733	741	\|\| strncmp(zCmd, "ls", n)==0
734	742	\|\| strncmp(zCmd, "status", n)==0
735	743	){
736	744

	--- src/bisect.c
	+++ src/bisect.c
	@@ -543,14 +543,16 @@
543	**
544	** > fossil bisect vlist\|ls\|status ?-a\|--all?
545	**
546	** List the versions in between the inner-most "bad" and "good".
547	**
548	** > fossil bisect ui
549	**
550	** Like "fossil ui" except start on a timeline that shows only the
551	** check-ins that are part of the current bisect.


552	**
553	** > fossil bisect undo
554	**
555	** Undo the most recent "good", "bad", or "skip" command.
556	*/
	@@ -719,17 +721,23 @@
719	}
720	}else if( strncmp(zCmd, "reset", n)==0 ){
721	bisect_reset();
722	}else if( strcmp(zCmd, "ui")==0 ){
723	char *newArgv[8];

724	newArgv[0] = g.argv[0];
725	newArgv[1] = "ui";
726	newArgv[2] = "--page";
727	newArgv[3] = "timeline?bisect";
728	newArgv[4] = 0;






729	g.argv = newArgv;
730	g.argc = 4;
731	cmd_webserver();
732	}else if( strncmp(zCmd, "vlist", n)==0
733	\|\| strncmp(zCmd, "ls", n)==0
734	\|\| strncmp(zCmd, "status", n)==0
735	){
736

	--- src/bisect.c
	+++ src/bisect.c
	@@ -543,14 +543,16 @@
543	**
544	** > fossil bisect vlist\|ls\|status ?-a\|--all?
545	**
546	** List the versions in between the inner-most "bad" and "good".
547	**
548	** > fossil bisect ui ?HOST@USER:PATH?
549	**
550	** Like "fossil ui" except start on a timeline that shows only the
551	** check-ins that are part of the current bisect. If the optional
552	** fourth term is added, then information is shown for the bisect that
553	** occurred in the PATH directory by USER on remote machine HOST.
554	**
555	** > fossil bisect undo
556	**
557	** Undo the most recent "good", "bad", or "skip" command.
558	*/
	@@ -719,17 +721,23 @@
721	}
722	}else if( strncmp(zCmd, "reset", n)==0 ){
723	bisect_reset();
724	}else if( strcmp(zCmd, "ui")==0 ){
725	char *newArgv[8];
726	verify_all_options();
727	newArgv[0] = g.argv[0];
728	newArgv[1] = "ui";
729	newArgv[2] = "--page";
730	newArgv[3] = "timeline?bisect";
731	if( g.argc==4 ){
732	newArgv[4] = g.argv[3];
733	g.argc = 5;
734	}else{
735	g.argc = 4;
736	}
737	newArgv[g.argc] = 0;
738	g.argv = newArgv;

739	cmd_webserver();
740	}else if( strncmp(zCmd, "vlist", n)==0
741	\|\| strncmp(zCmd, "ls", n)==0
742	\|\| strncmp(zCmd, "status", n)==0
743	){
744

M src/blob.c

+49 -1

		--- src/blob.c
		+++ src/blob.c
		@@ -665,11 +665,12 @@
665	665	pBlob->nUsed = dehttpize(pBlob->aData);
666	666	}
667	667
668	668	/*
669	669	** Extract N bytes from blob pFrom and use it to initialize blob pTo.
670		-** Return the actual number of bytes extracted.
	670	+** Return the actual number of bytes extracted. The cursor position
	671	+** is advanced by the number of bytes extracted.
671	672	**
672	673	** After this call completes, pTo will be an ephemeral blob.
673	674	*/
674	675	int blob_extract(Blob pFrom, int N, Blob pTo){
675	676	blob_is_init(pFrom);
		@@ -687,10 +688,57 @@
687	688	pTo->iCursor = 0;
688	689	pTo->xRealloc = blobReallocStatic;
689	690	pFrom->iCursor += N;
690	691	return N;
691	692	}
	693	+
	694	+/*
	695	+ Extract N lines** of text from blob pFrom beginning at the current
	696	+** cursor position and use that text to initialize blob pTo. Unlike the
	697	+** blob_extract() routine, the cursor position is unchanged.
	698	+**
	699	+** pTo is assumed to be uninitialized.
	700	+**
	701	+** After this call completes, pTo will be an ephemeral blob.
	702	+*/
	703	+int blob_extract_lines(Blob pFrom, int N, Blob pTo){
	704	+ int i;
	705	+ int mx;
	706	+ int iStart;
	707	+ int n;
	708	+ const char *z;
	709	+
	710	+ blob_zero(pTo);
	711	+ z = pFrom->aData;
	712	+ i = pFrom->iCursor;
	713	+ mx = pFrom->nUsed;
	714	+ while( N>0 ){
	715	+ while( i<mx && z[i]!='\n' ){ i++; }
	716	+ if( i>=mx ) break;
	717	+ i++;
	718	+ N--;
	719	+ }
	720	+ iStart = pFrom->iCursor;
	721	+ n = blob_extract(pFrom, i-pFrom->iCursor, pTo);
	722	+ pFrom->iCursor = iStart;
	723	+ return n;
	724	+}
	725	+
	726	+/*
	727	+** Return the number of lines of text in the blob. If the last
	728	+** line is incomplete (if it does not have a \n at the end) then
	729	+** it still counts.
	730	+*/
	731	+int blob_linecount(Blob *p){
	732	+ int n = 0;
	733	+ int i;
	734	+ for(i=0; i<p->nUsed; i++){
	735	+ if( p->aData[i]=='\n' ) n++;
	736	+ }
	737	+ if( p->nUsed>0 && p->aData[p->nUsed-1]!='\n' ) n++;
	738	+ return n;
	739	+}
692	740
693	741	/*
694	742	** Rewind the cursor on a blob back to the beginning.
695	743	*/
696	744	void blob_rewind(Blob *p){
697	745

	--- src/blob.c
	+++ src/blob.c
	@@ -665,11 +665,12 @@
665	pBlob->nUsed = dehttpize(pBlob->aData);
666	}
667
668	/*
669	** Extract N bytes from blob pFrom and use it to initialize blob pTo.
670	** Return the actual number of bytes extracted.

671	**
672	** After this call completes, pTo will be an ephemeral blob.
673	*/
674	int blob_extract(Blob pFrom, int N, Blob pTo){
675	blob_is_init(pFrom);
	@@ -687,10 +688,57 @@
687	pTo->iCursor = 0;
688	pTo->xRealloc = blobReallocStatic;
689	pFrom->iCursor += N;
690	return N;
691	}















































692
693	/*
694	** Rewind the cursor on a blob back to the beginning.
695	*/
696	void blob_rewind(Blob *p){
697

	--- src/blob.c
	+++ src/blob.c
	@@ -665,11 +665,12 @@
665	pBlob->nUsed = dehttpize(pBlob->aData);
666	}
667
668	/*
669	** Extract N bytes from blob pFrom and use it to initialize blob pTo.
670	** Return the actual number of bytes extracted. The cursor position
671	** is advanced by the number of bytes extracted.
672	**
673	** After this call completes, pTo will be an ephemeral blob.
674	*/
675	int blob_extract(Blob pFrom, int N, Blob pTo){
676	blob_is_init(pFrom);
	@@ -687,10 +688,57 @@
688	pTo->iCursor = 0;
689	pTo->xRealloc = blobReallocStatic;
690	pFrom->iCursor += N;
691	return N;
692	}
693
694	/*
695	Extract N lines** of text from blob pFrom beginning at the current
696	** cursor position and use that text to initialize blob pTo. Unlike the
697	** blob_extract() routine, the cursor position is unchanged.
698	**
699	** pTo is assumed to be uninitialized.
700	**
701	** After this call completes, pTo will be an ephemeral blob.
702	*/
703	int blob_extract_lines(Blob pFrom, int N, Blob pTo){
704	int i;
705	int mx;
706	int iStart;
707	int n;
708	const char *z;
709
710	blob_zero(pTo);
711	z = pFrom->aData;
712	i = pFrom->iCursor;
713	mx = pFrom->nUsed;
714	while( N>0 ){
715	while( i<mx && z[i]!='\n' ){ i++; }
716	if( i>=mx ) break;
717	i++;
718	N--;
719	}
720	iStart = pFrom->iCursor;
721	n = blob_extract(pFrom, i-pFrom->iCursor, pTo);
722	pFrom->iCursor = iStart;
723	return n;
724	}
725
726	/*
727	** Return the number of lines of text in the blob. If the last
728	** line is incomplete (if it does not have a \n at the end) then
729	** it still counts.
730	*/
731	int blob_linecount(Blob *p){
732	int n = 0;
733	int i;
734	for(i=0; i<p->nUsed; i++){
735	if( p->aData[i]=='\n' ) n++;
736	}
737	if( p->nUsed>0 && p->aData[p->nUsed-1]!='\n' ) n++;
738	return n;
739	}
740
741	/*
742	** Rewind the cursor on a blob back to the beginning.
743	*/
744	void blob_rewind(Blob *p){
745

M src/branch.c

+4 -3

		--- src/branch.c
		+++ src/branch.c
		@@ -640,12 +640,12 @@
640	640	** -M\|--unmerged List branches not merged into the current branch
641	641	** -p List only private branches
642	642	** -r Reverse the sort order
643	643	** -t Show recently changed branches first
644	644	** --self List only branches where you participate
645		-** --username USER List only branches where USER participate
646		-** --users N List up to N users partipiating
	645	+** --username USER List only branches where USER participates
	646	+** --users N List up to N users participating
647	647	**
648	648	** The current branch is marked with an asterisk. Private branches are
649	649	** marked with a hash sign.
650	650	**
651	651	** If GLOB is given, show only branches matching the pattern.
		@@ -660,11 +660,12 @@
660	660	** Create a new branch BRANCH-NAME off of check-in BASIS.
661	661	**
662	662	** Options:
663	663	** --private Branch is private (i.e., remains local)
664	664	** --bgcolor COLOR Use COLOR instead of automatic background
665		-** --nosign Do not sign contents on this branch
	665	+** --nosign Do not sign the manifest for the check-in
	666	+** that creates this branch
666	667	** --nosync Do not auto-sync prior to creating the branch
667	668	** --date-override DATE DATE to use instead of 'now'
668	669	** --user-override USER USER to use instead of the current default
669	670	**
670	671	** DATE may be "now" or "YYYY-MM-DDTHH:MM:SS.SSS". If in
671	672

	--- src/branch.c
	+++ src/branch.c
	@@ -640,12 +640,12 @@
640	** -M\|--unmerged List branches not merged into the current branch
641	** -p List only private branches
642	** -r Reverse the sort order
643	** -t Show recently changed branches first
644	** --self List only branches where you participate
645	** --username USER List only branches where USER participate
646	** --users N List up to N users partipiating
647	**
648	** The current branch is marked with an asterisk. Private branches are
649	** marked with a hash sign.
650	**
651	** If GLOB is given, show only branches matching the pattern.
	@@ -660,11 +660,12 @@
660	** Create a new branch BRANCH-NAME off of check-in BASIS.
661	**
662	** Options:
663	** --private Branch is private (i.e., remains local)
664	** --bgcolor COLOR Use COLOR instead of automatic background
665	** --nosign Do not sign contents on this branch

666	** --nosync Do not auto-sync prior to creating the branch
667	** --date-override DATE DATE to use instead of 'now'
668	** --user-override USER USER to use instead of the current default
669	**
670	** DATE may be "now" or "YYYY-MM-DDTHH:MM:SS.SSS". If in
671

	--- src/branch.c
	+++ src/branch.c
	@@ -640,12 +640,12 @@
640	** -M\|--unmerged List branches not merged into the current branch
641	** -p List only private branches
642	** -r Reverse the sort order
643	** -t Show recently changed branches first
644	** --self List only branches where you participate
645	** --username USER List only branches where USER participates
646	** --users N List up to N users participating
647	**
648	** The current branch is marked with an asterisk. Private branches are
649	** marked with a hash sign.
650	**
651	** If GLOB is given, show only branches matching the pattern.
	@@ -660,11 +660,12 @@
660	** Create a new branch BRANCH-NAME off of check-in BASIS.
661	**
662	** Options:
663	** --private Branch is private (i.e., remains local)
664	** --bgcolor COLOR Use COLOR instead of automatic background
665	** --nosign Do not sign the manifest for the check-in
666	** that creates this branch
667	** --nosync Do not auto-sync prior to creating the branch
668	** --date-override DATE DATE to use instead of 'now'
669	** --user-override USER USER to use instead of the current default
670	**
671	** DATE may be "now" or "YYYY-MM-DDTHH:MM:SS.SSS". If in
672

M src/browse.c

-17

		--- src/browse.c
		+++ src/browse.c
		@@ -1258,22 +1258,5 @@
1258	1258	@ </table></div>
1259	1259	db_finalize(&q1);
1260	1260	db_finalize(&q2);
1261	1261	style_finish_page();
1262	1262	}
1263		-
1264		-/*
1265		-** WEBPAGE: files
1266		-**
1267		-** Show files as a flat table. If the ci=LABEL query parameter is provided,
1268		-** then show all the files in the specified check-in. Without the ci= query
1269		-** parameter show all files across all check-ins.
1270		-**
1271		-** Query parameters:
1272		-**
1273		-** name=PATH Directory to display. Optional
1274		-** ci=LABEL Show only files in this check-in. Optional.
1275		-** re=REGEXP Show only files matching REGEXP. Optional.
1276		-*/
1277		-void files_page(void){
1278		- return;
1279		-}
1280	1263

	--- src/browse.c
	+++ src/browse.c
	@@ -1258,22 +1258,5 @@
1258	@ </table></div>
1259	db_finalize(&q1);
1260	db_finalize(&q2);
1261	style_finish_page();
1262	}
1263
1264	/*
1265	** WEBPAGE: files
1266	**
1267	** Show files as a flat table. If the ci=LABEL query parameter is provided,
1268	** then show all the files in the specified check-in. Without the ci= query
1269	** parameter show all files across all check-ins.
1270	**
1271	** Query parameters:
1272	**
1273	** name=PATH Directory to display. Optional
1274	** ci=LABEL Show only files in this check-in. Optional.
1275	** re=REGEXP Show only files matching REGEXP. Optional.
1276	*/
1277	void files_page(void){
1278	return;
1279	}
1280

	--- src/browse.c
	+++ src/browse.c
	@@ -1258,22 +1258,5 @@
1258	@ </table></div>
1259	db_finalize(&q1);
1260	db_finalize(&q2);
1261	style_finish_page();
1262	}

















1263

M src/builtin.c

+3 -1

		--- src/builtin.c
		+++ src/builtin.c
		@@ -718,11 +718,13 @@
718	718	** the final one! */
719	719	} fjs[] = {
720	720	/* This list ordering isn't strictly important. */
721	721	{"confirmer", 0, 0},
722	722	{"copybutton", 0, "dom\0"},
723		- {"diff", 0, "dom\0fetch\0"},
	723	+ {"diff", 0, "dom\0fetch\0storage\0"
	724	+ /* maintenance note: "diff" needs "storage" for storing the the
	725	+ ** sbs-sync-scroll toggle. */},
724	726	{"dom", 0, 0},
725	727	{"fetch", 0, 0},
726	728	{"numbered-lines", 0, "popupwidget\0copybutton\0"},
727	729	{"pikchr", 0, "dom\0"},
728	730	{"popupwidget", 0, "dom\0"},
729	731

	--- src/builtin.c
	+++ src/builtin.c
	@@ -718,11 +718,13 @@
718	** the final one! */
719	} fjs[] = {
720	/* This list ordering isn't strictly important. */
721	{"confirmer", 0, 0},
722	{"copybutton", 0, "dom\0"},
723	{"diff", 0, "dom\0fetch\0"},


724	{"dom", 0, 0},
725	{"fetch", 0, 0},
726	{"numbered-lines", 0, "popupwidget\0copybutton\0"},
727	{"pikchr", 0, "dom\0"},
728	{"popupwidget", 0, "dom\0"},
729

	--- src/builtin.c
	+++ src/builtin.c
	@@ -718,11 +718,13 @@
718	** the final one! */
719	} fjs[] = {
720	/* This list ordering isn't strictly important. */
721	{"confirmer", 0, 0},
722	{"copybutton", 0, "dom\0"},
723	{"diff", 0, "dom\0fetch\0storage\0"
724	/* maintenance note: "diff" needs "storage" for storing the the
725	** sbs-sync-scroll toggle. */},
726	{"dom", 0, 0},
727	{"fetch", 0, 0},
728	{"numbered-lines", 0, "popupwidget\0copybutton\0"},
729	{"pikchr", 0, "dom\0"},
730	{"popupwidget", 0, "dom\0"},
731

M src/chat.c

+9 -3

		--- src/chat.c
		+++ src/chat.c
		@@ -1137,12 +1137,11 @@
1137	1137	);
1138	1138	fossil_free(zBranch);
1139	1139	fossil_free(zUuid);
1140	1140	}else if( zType[0]=='w' ){
1141	1141	/* Wiki page changes */
1142		- char *zUuid;
1143		- zUuid = db_text(0, "SELECT uuid FROM blob WHERE rid=%d", rid);
	1142	+ char *zUuid = rid_to_uuid(rid);
1144	1143	wiki_hyperlink_override(zUuid);
1145	1144	if( zMsg[0]=='-' ){
1146	1145	zRes = mprintf("Delete wiki page <a href='%R/whistory?name=%t'>%h</a>",
1147	1146	zMsg+1, zMsg+1);
1148	1147	}else if( zMsg[0]=='+' ){
		@@ -1155,10 +1154,17 @@
1155	1154	}else{
1156	1155	zRes = mprintf("%W", zMsg);
1157	1156	}
1158	1157	wiki_hyperlink_override(0);
1159	1158	fossil_free(zUuid);
	1159	+ }else if( zType[0]=='f' ){
	1160	+ /* Forum changes */
	1161	+ char *zUuid = rid_to_uuid(rid);
	1162	+ zRes = mprintf( "%W (artifact: <a href='%R/info/%S'>%S</a>, "
	1163	+ "user: <a href='%R/timeline?u=%t&c=%S'>%h</a>)",
	1164	+ zMsg, zUuid, zUuid, zUser, zUuid, zUser);
	1165	+ fossil_free(zUuid);
1160	1166	}else{
1161	1167	/* Anything else */
1162	1168	zRes = mprintf("%W", zMsg);
1163	1169	}
1164	1170	if( zRes ){
		@@ -1196,11 +1202,11 @@
1196	1202	**
1197	1203	** --all Download all chat content. Normally only
1198	1204	** previously undownloaded content is retrieved.
1199	1205	** --debug Additional debugging output
1200	1206	** --out DATABASE Store CHAT table in separate database file
1201		-** DATABASE rather that adding to local clone
	1207	+** DATABASE rather than adding to local clone
1202	1208	** --unsafe Allow the use of unencrypted http://
1203	1209	**
1204	1210	** > fossil chat send [ARGUMENTS]
1205	1211	**
1206	1212	** This command sends a new message to the chatroom. The message
1207	1213

	--- src/chat.c
	+++ src/chat.c
	@@ -1137,12 +1137,11 @@
1137	);
1138	fossil_free(zBranch);
1139	fossil_free(zUuid);
1140	}else if( zType[0]=='w' ){
1141	/* Wiki page changes */
1142	char *zUuid;
1143	zUuid = db_text(0, "SELECT uuid FROM blob WHERE rid=%d", rid);
1144	wiki_hyperlink_override(zUuid);
1145	if( zMsg[0]=='-' ){
1146	zRes = mprintf("Delete wiki page <a href='%R/whistory?name=%t'>%h</a>",
1147	zMsg+1, zMsg+1);
1148	}else if( zMsg[0]=='+' ){
	@@ -1155,10 +1154,17 @@
1155	}else{
1156	zRes = mprintf("%W", zMsg);
1157	}
1158	wiki_hyperlink_override(0);
1159	fossil_free(zUuid);







1160	}else{
1161	/* Anything else */
1162	zRes = mprintf("%W", zMsg);
1163	}
1164	if( zRes ){
	@@ -1196,11 +1202,11 @@
1196	**
1197	** --all Download all chat content. Normally only
1198	** previously undownloaded content is retrieved.
1199	** --debug Additional debugging output
1200	** --out DATABASE Store CHAT table in separate database file
1201	** DATABASE rather that adding to local clone
1202	** --unsafe Allow the use of unencrypted http://
1203	**
1204	** > fossil chat send [ARGUMENTS]
1205	**
1206	** This command sends a new message to the chatroom. The message
1207

	--- src/chat.c
	+++ src/chat.c
	@@ -1137,12 +1137,11 @@
1137	);
1138	fossil_free(zBranch);
1139	fossil_free(zUuid);
1140	}else if( zType[0]=='w' ){
1141	/* Wiki page changes */
1142	char *zUuid = rid_to_uuid(rid);

1143	wiki_hyperlink_override(zUuid);
1144	if( zMsg[0]=='-' ){
1145	zRes = mprintf("Delete wiki page <a href='%R/whistory?name=%t'>%h</a>",
1146	zMsg+1, zMsg+1);
1147	}else if( zMsg[0]=='+' ){
	@@ -1155,10 +1154,17 @@
1154	}else{
1155	zRes = mprintf("%W", zMsg);
1156	}
1157	wiki_hyperlink_override(0);
1158	fossil_free(zUuid);
1159	}else if( zType[0]=='f' ){
1160	/* Forum changes */
1161	char *zUuid = rid_to_uuid(rid);
1162	zRes = mprintf( "%W (artifact: <a href='%R/info/%S'>%S</a>, "
1163	"user: <a href='%R/timeline?u=%t&c=%S'>%h</a>)",
1164	zMsg, zUuid, zUuid, zUser, zUuid, zUser);
1165	fossil_free(zUuid);
1166	}else{
1167	/* Anything else */
1168	zRes = mprintf("%W", zMsg);
1169	}
1170	if( zRes ){
	@@ -1196,11 +1202,11 @@
1202	**
1203	** --all Download all chat content. Normally only
1204	** previously undownloaded content is retrieved.
1205	** --debug Additional debugging output
1206	** --out DATABASE Store CHAT table in separate database file
1207	** DATABASE rather than adding to local clone
1208	** --unsafe Allow the use of unencrypted http://
1209	**
1210	** > fossil chat send [ARGUMENTS]
1211	**
1212	** This command sends a new message to the chatroom. The message
1213

M src/checkin.c

+114 -14

		--- src/checkin.c
		+++ src/checkin.c
		@@ -430,10 +430,11 @@
430	430	**
431	431	** The "fossil changes --extra" command is equivalent to "fossil extras".
432	432	**
433	433	** General options:
434	434	** --abs-paths Display absolute pathnames
	435	+** -b\|--brief Show a single keyword for the status
435	436	** --rel-paths Display pathnames relative to the current working
436	437	** directory
437	438	** --hash Verify file status using hashing rather than
438	439	** relying on file mtimes
439	440	** --case-sensitive BOOL Override case-sensitive setting
		@@ -493,10 +494,48 @@
493	494	unsigned scanFlags = 0;
494	495	unsigned flags = 0;
495	496	int vid, i;
496	497
497	498	fossil_pledge("stdio rpath wpath cpath fattr id flock tty chown");
	499	+
	500	+ if( find_option("brief","b",0) ){
	501	+ /* The --brief or -b option is special. It cannot be used with any
	502	+ ** other options. It outputs a single keyword which indicates the
	503	+ ** fossil status, for use by shell scripts. The output might be
	504	+ ** one of:
	505	+ **
	506	+ ** clean The current working directory is within an
	507	+ ** unmodified fossil check-out.
	508	+ **
	509	+ ** dirty The pwd is within a fossil check-out that has
	510	+ ** uncommitted changes
	511	+ **
	512	+ ** none The pwd is not within a fossil check-out.
	513	+ */
	514	+ int chnged;
	515	+ if( g.argc>2 ){
	516	+ fossil_fatal("No other arguments or options may occur with --brief");
	517	+ }
	518	+ if( db_open_local(0)==0 ){
	519	+ fossil_print("none\n");
	520	+ return;
	521	+ }
	522	+ vid = db_lget_int("checkout", 0);
	523	+ vfile_check_signature(vid, 0);
	524	+ chnged = db_int(0,
	525	+ "SELECT 1 FROM vfile"
	526	+ " WHERE vid=%d"
	527	+ " AND (chnged>0 OR deleted OR rid==0)",
	528	+ vid
	529	+ );
	530	+ if( chnged ){
	531	+ fossil_print("dirty\n");
	532	+ }else{
	533	+ fossil_print("clean\n");
	534	+ }
	535	+ return;
	536	+ }
498	537
499	538	/* Load affirmative flag options. */
500	539	for( i=0; i<count(flagDefs); ++i ){
501	540	if( (command==CHANGES \|\| !(flagDefs[i].mask & C_CLASSIFY))
502	541	&& find_option(flagDefs[i].option, 0, 0) ){
		@@ -528,11 +567,11 @@
528	567	}
529	568
530	569	/* Confirm current working directory is within check-out. */
531	570	db_must_be_within_tree();
532	571
533		- /* Get check-out version. l*/
	572	+ /* Get check-out version. */
534	573	vid = db_lget_int("checkout", 0);
535	574
536	575	/* Relative path flag determination is done by a shared function. */
537	576	if( determine_cwd_relative_option() ){
538	577	flags \|= C_RELPATH;
		@@ -928,12 +967,12 @@
928	967	/*
929	968	** COMMAND: tree
930	969	**
931	970	** Usage: %fossil tree ?OPTIONS? ?PATHS ...?
932	971	**
933		-** List all files in the current check-out in after the fashion of the
934		-** "tree" command. If PATHS is included, only the named files
	972	+** List all files in the current check-out much like the "tree"
	973	+** command does. If PATHS is included, only the named files
935	974	** (or their children if directories) are shown.
936	975	**
937	976	** Options:
938	977	** -r VERSION The specific check-in to list
939	978	** -R\|--repository REPO Extract info from repository REPO
		@@ -2279,11 +2318,11 @@
2279	2318	**
2280	2319	** A check-in is not permitted to fork unless the --allow-fork option
2281	2320	** appears. An empty check-in (i.e. with nothing changed) is not
2282	2321	** allowed unless the --allow-empty option appears. A check-in may not
2283	2322	** be older than its ancestor unless the --allow-older option appears.
2284		-** If any of files in the check-in appear to contain unresolved merge
	2323	+** If any files in the check-in appear to contain unresolved merge
2285	2324	** conflicts, the check-in will not be allowed unless the
2286	2325	** --allow-conflict option is present. In addition, the entire
2287	2326	** check-in process may be aborted if a file contains content that
2288	2327	** appears to be binary, Unicode text, or text with CR/LF line endings
2289	2328	** unless the interactive user chooses to proceed. If there is no
		@@ -2319,11 +2358,11 @@
2319	2358	** than relying on file mtimes
2320	2359	** --ignore-clock-skew If a clock skew is detected, ignore it and
2321	2360	** behave as if the user had entered 'yes' to
2322	2361	** the question of whether to proceed despite
2323	2362	** the skew.
2324		-** --ignore-oversize Do not warning the user about oversized files
	2363	+** --ignore-oversize Do not warn the user about oversized files
2325	2364	** --integrate Close all merged-in branches
2326	2365	** -m\|--comment COMMENT-TEXT Use COMMENT-TEXT as commit comment
2327	2366	** -M\|--message-file FILE Read the commit comment from given file
2328	2367	** --mimetype MIMETYPE Mimetype of check-in comment
2329	2368	** -n\|--dry-run If given, display instead of run actions
		@@ -2395,10 +2434,12 @@
2395	2434	Blob ans; /* Answer to continuation prompts */
2396	2435	char cReply; /* First character of ans */
2397	2436	int bRecheck = 0; /* Repeat fork and closed-branch checks*/
2398	2437	int bIgnoreSkew = 0; /* --ignore-clock-skew flag */
2399	2438	int mxSize;
	2439	+ char zCurBranch = 0; / The current branch name of checkout */
	2440	+ char zNewBranch = 0; / The branch name after update */
2400	2441
2401	2442	memset(&sCiInfo, 0, sizeof(sCiInfo));
2402	2443	url_proxy_options();
2403	2444	/* --sha1sum is an undocumented alias for --hash for backwards compatiblity */
2404	2445	useHash = find_option("hash",0,0)!=0 \|\| find_option("sha1sum",0,0)!=0;
		@@ -2469,10 +2510,55 @@
2469	2510	}
2470	2511	}else{
2471	2512	privateParent = content_is_private(vid);
2472	2513	}
2473	2514
	2515	+ user_select();
	2516	+ /*
	2517	+ ** Check that the user exists.
	2518	+ */
	2519	+ if( !db_exists("SELECT 1 FROM user WHERE login=%Q", g.zLogin) ){
	2520	+ fossil_fatal("no such user: %s", g.zLogin);
	2521	+ }
	2522	+
	2523	+ /*
	2524	+ ** Detect if the branch name has changed from the parent check-in
	2525	+ ** and prompt if necessary
	2526	+ **/
	2527	+ zCurBranch = db_text(0,
	2528	+ " SELECT value FROM tagxref AS tx"
	2529	+ " WHERE rid=(SELECT pid"
	2530	+ " FROM tagxref LEFT JOIN event ON srcid=objid"
	2531	+ " LEFT JOIN plink ON rid=cid"
	2532	+ " WHERE rid=%d AND tagxref.tagid=%d"
	2533	+ " AND srcid!=origid"
	2534	+ " AND tagtype=2 AND coalesce(euser,user)!=%Q)"
	2535	+ " AND tx.tagid=%d",
	2536	+ vid, TAG_BRANCH, g.zLogin, TAG_BRANCH
	2537	+ );
	2538	+ if( zCurBranch!=0 && zCurBranch[0]!=0
	2539	+ && forceFlag==0
	2540	+ && noPrompt==0
	2541	+ ){
	2542	+ zNewBranch = branch_of_rid(vid);
	2543	+ fossil_warning(
	2544	+ "WARNING: The parent check-in [%.10s] has been moved from branch\n"
	2545	+ " '%s' over to branch '%s'.",
	2546	+ rid_to_uuid(vid), zCurBranch, zNewBranch
	2547	+ );
	2548	+ prompt_user("Commit anyway? (y/N) ", &ans);
	2549	+ cReply = blob_str(&ans)[0];
	2550	+ blob_reset(&ans);
	2551	+ if( cReply!='y' && cReply!='Y' ){
	2552	+ fossil_fatal("Abandoning commit because branch has changed");
	2553	+ }
	2554	+ fossil_free(zNewBranch);
	2555	+ fossil_free(zCurBranch);
	2556	+ zCurBranch = branch_of_rid(vid);
	2557	+ }
	2558	+ if( zCurBranch==0 ) zCurBranch = branch_of_rid(vid);
	2559	+
2474	2560	/* Track the "private" status */
2475	2561	g.markPrivate = privateFlag \|\| privateParent;
2476	2562	if( privateFlag && !privateParent ){
2477	2563	/* Apply default branch name ("private") and color ("orange") if not
2478	2564	** specified otherwise on the command-line, and if the parent is not
		@@ -2521,11 +2607,11 @@
2521	2607	**
2522	2608	** If the remote repository sent an avoid-delta-manifests pragma on
2523	2609	** the autosync above, then also try to avoid deltas, unless the
2524	2610	** --delta option is specified. The remote repo will send the
2525	2611	** avoid-delta-manifests pragma if it has its "forbid-delta-manifests"
2526		- ** setting is enabled.
	2612	+ ** setting enabled.
2527	2613	*/
2528	2614	if( !db_get_boolean("seen-delta-manifest",0)
2529	2615	\|\| db_get_boolean("forbid-delta-manifests",0)
2530	2616	\|\| g.bAvoidDeltaManifests
2531	2617	){
		@@ -2600,18 +2686,10 @@
2600	2686	"'%s' was renamed to '%s'", zFrom, zTo, zFrom, zTo);
2601	2687	}
2602	2688	db_finalize(&q);
2603	2689	}
2604	2690
2605		- user_select();
2606		- /*
2607		- ** Check that the user exists.
2608		- */
2609		- if( !db_exists("SELECT 1 FROM user WHERE login=%Q", g.zLogin) ){
2610		- fossil_fatal("no such user: %s", g.zLogin);
2611		- }
2612		-
2613	2691	hasChanges = unsaved_changes(useHash ? CKSIG_HASH : 0);
2614	2692	db_begin_transaction();
2615	2693	db_record_repository_filename(0);
2616	2694	if( hasChanges==0 && !isAMerge && !allowEmpty && !forceFlag ){
2617	2695	fossil_fatal("nothing has changed; use --allow-empty to override");
		@@ -2674,10 +2752,32 @@
2674	2752	" WHERE tagid=%d AND rid=%d AND tagtype>0"
2675	2753	" AND value=%Q", TAG_BRANCH, vid, sCiInfo.zBranch))
2676	2754	){
2677	2755	fossil_fatal("cannot commit against a closed leaf");
2678	2756	}
	2757	+
	2758	+ /* Require confirmation to continue with the check-in if the branch
	2759	+ ** has changed and the committer did not provide the same branch
	2760	+ */
	2761	+ zNewBranch = branch_of_rid(vid);
	2762	+ if( fossil_strcmp(zCurBranch, zNewBranch)!=0
	2763	+ && fossil_strcmp(sCiInfo.zBranch, zNewBranch)!=0
	2764	+ && forceFlag==0
	2765	+ && noPrompt==0
	2766	+ ){
	2767	+ fossil_warning("parent check-in [%.10s] branch changed from '%s' to '%s'",
	2768	+ rid_to_uuid(vid), zCurBranch, zNewBranch);
	2769	+ prompt_user("continue (y/N)? ", &ans);
	2770	+ cReply = blob_str(&ans)[0];
	2771	+ blob_reset(&ans);
	2772	+ if( cReply!='y' && cReply!='Y' ){
	2773	+ fossil_fatal("Abandoning commit because branch has changed");
	2774	+ }
	2775	+ fossil_free(zCurBranch);
	2776	+ zCurBranch = branch_of_rid(vid);
	2777	+ }
	2778	+ fossil_free(zNewBranch);
2679	2779
2680	2780	/* Always exit the loop on the second pass */
2681	2781	if( bRecheck ) break;
2682	2782
2683	2783
2684	2784

	--- src/checkin.c
	+++ src/checkin.c
	@@ -430,10 +430,11 @@
430	**
431	** The "fossil changes --extra" command is equivalent to "fossil extras".
432	**
433	** General options:
434	** --abs-paths Display absolute pathnames

435	** --rel-paths Display pathnames relative to the current working
436	** directory
437	** --hash Verify file status using hashing rather than
438	** relying on file mtimes
439	** --case-sensitive BOOL Override case-sensitive setting
	@@ -493,10 +494,48 @@
493	unsigned scanFlags = 0;
494	unsigned flags = 0;
495	int vid, i;
496
497	fossil_pledge("stdio rpath wpath cpath fattr id flock tty chown");






































498
499	/* Load affirmative flag options. */
500	for( i=0; i<count(flagDefs); ++i ){
501	if( (command==CHANGES \|\| !(flagDefs[i].mask & C_CLASSIFY))
502	&& find_option(flagDefs[i].option, 0, 0) ){
	@@ -528,11 +567,11 @@
528	}
529
530	/* Confirm current working directory is within check-out. */
531	db_must_be_within_tree();
532
533	/* Get check-out version. l*/
534	vid = db_lget_int("checkout", 0);
535
536	/* Relative path flag determination is done by a shared function. */
537	if( determine_cwd_relative_option() ){
538	flags \|= C_RELPATH;
	@@ -928,12 +967,12 @@
928	/*
929	** COMMAND: tree
930	**
931	** Usage: %fossil tree ?OPTIONS? ?PATHS ...?
932	**
933	** List all files in the current check-out in after the fashion of the
934	** "tree" command. If PATHS is included, only the named files
935	** (or their children if directories) are shown.
936	**
937	** Options:
938	** -r VERSION The specific check-in to list
939	** -R\|--repository REPO Extract info from repository REPO
	@@ -2279,11 +2318,11 @@
2279	**
2280	** A check-in is not permitted to fork unless the --allow-fork option
2281	** appears. An empty check-in (i.e. with nothing changed) is not
2282	** allowed unless the --allow-empty option appears. A check-in may not
2283	** be older than its ancestor unless the --allow-older option appears.
2284	** If any of files in the check-in appear to contain unresolved merge
2285	** conflicts, the check-in will not be allowed unless the
2286	** --allow-conflict option is present. In addition, the entire
2287	** check-in process may be aborted if a file contains content that
2288	** appears to be binary, Unicode text, or text with CR/LF line endings
2289	** unless the interactive user chooses to proceed. If there is no
	@@ -2319,11 +2358,11 @@
2319	** than relying on file mtimes
2320	** --ignore-clock-skew If a clock skew is detected, ignore it and
2321	** behave as if the user had entered 'yes' to
2322	** the question of whether to proceed despite
2323	** the skew.
2324	** --ignore-oversize Do not warning the user about oversized files
2325	** --integrate Close all merged-in branches
2326	** -m\|--comment COMMENT-TEXT Use COMMENT-TEXT as commit comment
2327	** -M\|--message-file FILE Read the commit comment from given file
2328	** --mimetype MIMETYPE Mimetype of check-in comment
2329	** -n\|--dry-run If given, display instead of run actions
	@@ -2395,10 +2434,12 @@
2395	Blob ans; /* Answer to continuation prompts */
2396	char cReply; /* First character of ans */
2397	int bRecheck = 0; /* Repeat fork and closed-branch checks*/
2398	int bIgnoreSkew = 0; /* --ignore-clock-skew flag */
2399	int mxSize;


2400
2401	memset(&sCiInfo, 0, sizeof(sCiInfo));
2402	url_proxy_options();
2403	/* --sha1sum is an undocumented alias for --hash for backwards compatiblity */
2404	useHash = find_option("hash",0,0)!=0 \|\| find_option("sha1sum",0,0)!=0;
	@@ -2469,10 +2510,55 @@
2469	}
2470	}else{
2471	privateParent = content_is_private(vid);
2472	}
2473













































2474	/* Track the "private" status */
2475	g.markPrivate = privateFlag \|\| privateParent;
2476	if( privateFlag && !privateParent ){
2477	/* Apply default branch name ("private") and color ("orange") if not
2478	** specified otherwise on the command-line, and if the parent is not
	@@ -2521,11 +2607,11 @@
2521	**
2522	** If the remote repository sent an avoid-delta-manifests pragma on
2523	** the autosync above, then also try to avoid deltas, unless the
2524	** --delta option is specified. The remote repo will send the
2525	** avoid-delta-manifests pragma if it has its "forbid-delta-manifests"
2526	** setting is enabled.
2527	*/
2528	if( !db_get_boolean("seen-delta-manifest",0)
2529	\|\| db_get_boolean("forbid-delta-manifests",0)
2530	\|\| g.bAvoidDeltaManifests
2531	){
	@@ -2600,18 +2686,10 @@
2600	"'%s' was renamed to '%s'", zFrom, zTo, zFrom, zTo);
2601	}
2602	db_finalize(&q);
2603	}
2604
2605	user_select();
2606	/*
2607	** Check that the user exists.
2608	*/
2609	if( !db_exists("SELECT 1 FROM user WHERE login=%Q", g.zLogin) ){
2610	fossil_fatal("no such user: %s", g.zLogin);
2611	}
2612
2613	hasChanges = unsaved_changes(useHash ? CKSIG_HASH : 0);
2614	db_begin_transaction();
2615	db_record_repository_filename(0);
2616	if( hasChanges==0 && !isAMerge && !allowEmpty && !forceFlag ){
2617	fossil_fatal("nothing has changed; use --allow-empty to override");
	@@ -2674,10 +2752,32 @@
2674	" WHERE tagid=%d AND rid=%d AND tagtype>0"
2675	" AND value=%Q", TAG_BRANCH, vid, sCiInfo.zBranch))
2676	){
2677	fossil_fatal("cannot commit against a closed leaf");
2678	}






















2679
2680	/* Always exit the loop on the second pass */
2681	if( bRecheck ) break;
2682
2683
2684

	--- src/checkin.c
	+++ src/checkin.c
	@@ -430,10 +430,11 @@
430	**
431	** The "fossil changes --extra" command is equivalent to "fossil extras".
432	**
433	** General options:
434	** --abs-paths Display absolute pathnames
435	** -b\|--brief Show a single keyword for the status
436	** --rel-paths Display pathnames relative to the current working
437	** directory
438	** --hash Verify file status using hashing rather than
439	** relying on file mtimes
440	** --case-sensitive BOOL Override case-sensitive setting
	@@ -493,10 +494,48 @@
494	unsigned scanFlags = 0;
495	unsigned flags = 0;
496	int vid, i;
497
498	fossil_pledge("stdio rpath wpath cpath fattr id flock tty chown");
499
500	if( find_option("brief","b",0) ){
501	/* The --brief or -b option is special. It cannot be used with any
502	** other options. It outputs a single keyword which indicates the
503	** fossil status, for use by shell scripts. The output might be
504	** one of:
505	**
506	** clean The current working directory is within an
507	** unmodified fossil check-out.
508	**
509	** dirty The pwd is within a fossil check-out that has
510	** uncommitted changes
511	**
512	** none The pwd is not within a fossil check-out.
513	*/
514	int chnged;
515	if( g.argc>2 ){
516	fossil_fatal("No other arguments or options may occur with --brief");
517	}
518	if( db_open_local(0)==0 ){
519	fossil_print("none\n");
520	return;
521	}
522	vid = db_lget_int("checkout", 0);
523	vfile_check_signature(vid, 0);
524	chnged = db_int(0,
525	"SELECT 1 FROM vfile"
526	" WHERE vid=%d"
527	" AND (chnged>0 OR deleted OR rid==0)",
528	vid
529	);
530	if( chnged ){
531	fossil_print("dirty\n");
532	}else{
533	fossil_print("clean\n");
534	}
535	return;
536	}
537
538	/* Load affirmative flag options. */
539	for( i=0; i<count(flagDefs); ++i ){
540	if( (command==CHANGES \|\| !(flagDefs[i].mask & C_CLASSIFY))
541	&& find_option(flagDefs[i].option, 0, 0) ){
	@@ -528,11 +567,11 @@
567	}
568
569	/* Confirm current working directory is within check-out. */
570	db_must_be_within_tree();
571
572	/* Get check-out version. */
573	vid = db_lget_int("checkout", 0);
574
575	/* Relative path flag determination is done by a shared function. */
576	if( determine_cwd_relative_option() ){
577	flags \|= C_RELPATH;
	@@ -928,12 +967,12 @@
967	/*
968	** COMMAND: tree
969	**
970	** Usage: %fossil tree ?OPTIONS? ?PATHS ...?
971	**
972	** List all files in the current check-out much like the "tree"
973	** command does. If PATHS is included, only the named files
974	** (or their children if directories) are shown.
975	**
976	** Options:
977	** -r VERSION The specific check-in to list
978	** -R\|--repository REPO Extract info from repository REPO
	@@ -2279,11 +2318,11 @@
2318	**
2319	** A check-in is not permitted to fork unless the --allow-fork option
2320	** appears. An empty check-in (i.e. with nothing changed) is not
2321	** allowed unless the --allow-empty option appears. A check-in may not
2322	** be older than its ancestor unless the --allow-older option appears.
2323	** If any files in the check-in appear to contain unresolved merge
2324	** conflicts, the check-in will not be allowed unless the
2325	** --allow-conflict option is present. In addition, the entire
2326	** check-in process may be aborted if a file contains content that
2327	** appears to be binary, Unicode text, or text with CR/LF line endings
2328	** unless the interactive user chooses to proceed. If there is no
	@@ -2319,11 +2358,11 @@
2358	** than relying on file mtimes
2359	** --ignore-clock-skew If a clock skew is detected, ignore it and
2360	** behave as if the user had entered 'yes' to
2361	** the question of whether to proceed despite
2362	** the skew.
2363	** --ignore-oversize Do not warn the user about oversized files
2364	** --integrate Close all merged-in branches
2365	** -m\|--comment COMMENT-TEXT Use COMMENT-TEXT as commit comment
2366	** -M\|--message-file FILE Read the commit comment from given file
2367	** --mimetype MIMETYPE Mimetype of check-in comment
2368	** -n\|--dry-run If given, display instead of run actions
	@@ -2395,10 +2434,12 @@
2434	Blob ans; /* Answer to continuation prompts */
2435	char cReply; /* First character of ans */
2436	int bRecheck = 0; /* Repeat fork and closed-branch checks*/
2437	int bIgnoreSkew = 0; /* --ignore-clock-skew flag */
2438	int mxSize;
2439	char zCurBranch = 0; / The current branch name of checkout */
2440	char zNewBranch = 0; / The branch name after update */
2441
2442	memset(&sCiInfo, 0, sizeof(sCiInfo));
2443	url_proxy_options();
2444	/* --sha1sum is an undocumented alias for --hash for backwards compatiblity */
2445	useHash = find_option("hash",0,0)!=0 \|\| find_option("sha1sum",0,0)!=0;
	@@ -2469,10 +2510,55 @@
2510	}
2511	}else{
2512	privateParent = content_is_private(vid);
2513	}
2514
2515	user_select();
2516	/*
2517	** Check that the user exists.
2518	*/
2519	if( !db_exists("SELECT 1 FROM user WHERE login=%Q", g.zLogin) ){
2520	fossil_fatal("no such user: %s", g.zLogin);
2521	}
2522
2523	/*
2524	** Detect if the branch name has changed from the parent check-in
2525	** and prompt if necessary
2526	**/
2527	zCurBranch = db_text(0,
2528	" SELECT value FROM tagxref AS tx"
2529	" WHERE rid=(SELECT pid"
2530	" FROM tagxref LEFT JOIN event ON srcid=objid"
2531	" LEFT JOIN plink ON rid=cid"
2532	" WHERE rid=%d AND tagxref.tagid=%d"
2533	" AND srcid!=origid"
2534	" AND tagtype=2 AND coalesce(euser,user)!=%Q)"
2535	" AND tx.tagid=%d",
2536	vid, TAG_BRANCH, g.zLogin, TAG_BRANCH
2537	);
2538	if( zCurBranch!=0 && zCurBranch[0]!=0
2539	&& forceFlag==0
2540	&& noPrompt==0
2541	){
2542	zNewBranch = branch_of_rid(vid);
2543	fossil_warning(
2544	"WARNING: The parent check-in [%.10s] has been moved from branch\n"
2545	" '%s' over to branch '%s'.",
2546	rid_to_uuid(vid), zCurBranch, zNewBranch
2547	);
2548	prompt_user("Commit anyway? (y/N) ", &ans);
2549	cReply = blob_str(&ans)[0];
2550	blob_reset(&ans);
2551	if( cReply!='y' && cReply!='Y' ){
2552	fossil_fatal("Abandoning commit because branch has changed");
2553	}
2554	fossil_free(zNewBranch);
2555	fossil_free(zCurBranch);
2556	zCurBranch = branch_of_rid(vid);
2557	}
2558	if( zCurBranch==0 ) zCurBranch = branch_of_rid(vid);
2559
2560	/* Track the "private" status */
2561	g.markPrivate = privateFlag \|\| privateParent;
2562	if( privateFlag && !privateParent ){
2563	/* Apply default branch name ("private") and color ("orange") if not
2564	** specified otherwise on the command-line, and if the parent is not
	@@ -2521,11 +2607,11 @@
2607	**
2608	** If the remote repository sent an avoid-delta-manifests pragma on
2609	** the autosync above, then also try to avoid deltas, unless the
2610	** --delta option is specified. The remote repo will send the
2611	** avoid-delta-manifests pragma if it has its "forbid-delta-manifests"
2612	** setting enabled.
2613	*/
2614	if( !db_get_boolean("seen-delta-manifest",0)
2615	\|\| db_get_boolean("forbid-delta-manifests",0)
2616	\|\| g.bAvoidDeltaManifests
2617	){
	@@ -2600,18 +2686,10 @@
2686	"'%s' was renamed to '%s'", zFrom, zTo, zFrom, zTo);
2687	}
2688	db_finalize(&q);
2689	}
2690








2691	hasChanges = unsaved_changes(useHash ? CKSIG_HASH : 0);
2692	db_begin_transaction();
2693	db_record_repository_filename(0);
2694	if( hasChanges==0 && !isAMerge && !allowEmpty && !forceFlag ){
2695	fossil_fatal("nothing has changed; use --allow-empty to override");
	@@ -2674,10 +2752,32 @@
2752	" WHERE tagid=%d AND rid=%d AND tagtype>0"
2753	" AND value=%Q", TAG_BRANCH, vid, sCiInfo.zBranch))
2754	){
2755	fossil_fatal("cannot commit against a closed leaf");
2756	}
2757
2758	/* Require confirmation to continue with the check-in if the branch
2759	** has changed and the committer did not provide the same branch
2760	*/
2761	zNewBranch = branch_of_rid(vid);
2762	if( fossil_strcmp(zCurBranch, zNewBranch)!=0
2763	&& fossil_strcmp(sCiInfo.zBranch, zNewBranch)!=0
2764	&& forceFlag==0
2765	&& noPrompt==0
2766	){
2767	fossil_warning("parent check-in [%.10s] branch changed from '%s' to '%s'",
2768	rid_to_uuid(vid), zCurBranch, zNewBranch);
2769	prompt_user("continue (y/N)? ", &ans);
2770	cReply = blob_str(&ans)[0];
2771	blob_reset(&ans);
2772	if( cReply!='y' && cReply!='Y' ){
2773	fossil_fatal("Abandoning commit because branch has changed");
2774	}
2775	fossil_free(zCurBranch);
2776	zCurBranch = branch_of_rid(vid);
2777	}
2778	fossil_free(zNewBranch);
2779
2780	/* Always exit the loop on the second pass */
2781	if( bRecheck ) break;
2782
2783
2784

M src/comformat.c

+221 -71

		--- src/comformat.c
		+++ src/comformat.c
		@@ -31,10 +31,117 @@
31	31	#define COMMENT_PRINT_ORIG_BREAK ((u32)0x00000010) /* Break before original. */
32	32	#define COMMENT_PRINT_DEFAULT (COMMENT_PRINT_LEGACY) /* Defaults. */
33	33	#define COMMENT_PRINT_UNSET (-1) /* Not initialized. */
34	34	#endif
35	35
	36	+/******* Code copied from SQLite src/shell.c.in on 2024-09-30 ********/
	37	+/* Lookup table to estimate the number of columns consumed by a Unicode
	38	+** character.
	39	+*/
	40	+static const struct {
	41	+ unsigned char w; /* Width of the character in columns */
	42	+ int iFirst; /* First character in a span having this width */
	43	+} aUWidth[] = {
	44	+ /* {1, 0x00000}, */
	45	+ {0, 0x00300}, {1, 0x00370}, {0, 0x00483}, {1, 0x00487}, {0, 0x00488},
	46	+ {1, 0x0048a}, {0, 0x00591}, {1, 0x005be}, {0, 0x005bf}, {1, 0x005c0},
	47	+ {0, 0x005c1}, {1, 0x005c3}, {0, 0x005c4}, {1, 0x005c6}, {0, 0x005c7},
	48	+ {1, 0x005c8}, {0, 0x00600}, {1, 0x00604}, {0, 0x00610}, {1, 0x00616},
	49	+ {0, 0x0064b}, {1, 0x0065f}, {0, 0x00670}, {1, 0x00671}, {0, 0x006d6},
	50	+ {1, 0x006e5}, {0, 0x006e7}, {1, 0x006e9}, {0, 0x006ea}, {1, 0x006ee},
	51	+ {0, 0x0070f}, {1, 0x00710}, {0, 0x00711}, {1, 0x00712}, {0, 0x00730},
	52	+ {1, 0x0074b}, {0, 0x007a6}, {1, 0x007b1}, {0, 0x007eb}, {1, 0x007f4},
	53	+ {0, 0x00901}, {1, 0x00903}, {0, 0x0093c}, {1, 0x0093d}, {0, 0x00941},
	54	+ {1, 0x00949}, {0, 0x0094d}, {1, 0x0094e}, {0, 0x00951}, {1, 0x00955},
	55	+ {0, 0x00962}, {1, 0x00964}, {0, 0x00981}, {1, 0x00982}, {0, 0x009bc},
	56	+ {1, 0x009bd}, {0, 0x009c1}, {1, 0x009c5}, {0, 0x009cd}, {1, 0x009ce},
	57	+ {0, 0x009e2}, {1, 0x009e4}, {0, 0x00a01}, {1, 0x00a03}, {0, 0x00a3c},
	58	+ {1, 0x00a3d}, {0, 0x00a41}, {1, 0x00a43}, {0, 0x00a47}, {1, 0x00a49},
	59	+ {0, 0x00a4b}, {1, 0x00a4e}, {0, 0x00a70}, {1, 0x00a72}, {0, 0x00a81},
	60	+ {1, 0x00a83}, {0, 0x00abc}, {1, 0x00abd}, {0, 0x00ac1}, {1, 0x00ac6},
	61	+ {0, 0x00ac7}, {1, 0x00ac9}, {0, 0x00acd}, {1, 0x00ace}, {0, 0x00ae2},
	62	+ {1, 0x00ae4}, {0, 0x00b01}, {1, 0x00b02}, {0, 0x00b3c}, {1, 0x00b3d},
	63	+ {0, 0x00b3f}, {1, 0x00b40}, {0, 0x00b41}, {1, 0x00b44}, {0, 0x00b4d},
	64	+ {1, 0x00b4e}, {0, 0x00b56}, {1, 0x00b57}, {0, 0x00b82}, {1, 0x00b83},
	65	+ {0, 0x00bc0}, {1, 0x00bc1}, {0, 0x00bcd}, {1, 0x00bce}, {0, 0x00c3e},
	66	+ {1, 0x00c41}, {0, 0x00c46}, {1, 0x00c49}, {0, 0x00c4a}, {1, 0x00c4e},
	67	+ {0, 0x00c55}, {1, 0x00c57}, {0, 0x00cbc}, {1, 0x00cbd}, {0, 0x00cbf},
	68	+ {1, 0x00cc0}, {0, 0x00cc6}, {1, 0x00cc7}, {0, 0x00ccc}, {1, 0x00cce},
	69	+ {0, 0x00ce2}, {1, 0x00ce4}, {0, 0x00d41}, {1, 0x00d44}, {0, 0x00d4d},
	70	+ {1, 0x00d4e}, {0, 0x00dca}, {1, 0x00dcb}, {0, 0x00dd2}, {1, 0x00dd5},
	71	+ {0, 0x00dd6}, {1, 0x00dd7}, {0, 0x00e31}, {1, 0x00e32}, {0, 0x00e34},
	72	+ {1, 0x00e3b}, {0, 0x00e47}, {1, 0x00e4f}, {0, 0x00eb1}, {1, 0x00eb2},
	73	+ {0, 0x00eb4}, {1, 0x00eba}, {0, 0x00ebb}, {1, 0x00ebd}, {0, 0x00ec8},
	74	+ {1, 0x00ece}, {0, 0x00f18}, {1, 0x00f1a}, {0, 0x00f35}, {1, 0x00f36},
	75	+ {0, 0x00f37}, {1, 0x00f38}, {0, 0x00f39}, {1, 0x00f3a}, {0, 0x00f71},
	76	+ {1, 0x00f7f}, {0, 0x00f80}, {1, 0x00f85}, {0, 0x00f86}, {1, 0x00f88},
	77	+ {0, 0x00f90}, {1, 0x00f98}, {0, 0x00f99}, {1, 0x00fbd}, {0, 0x00fc6},
	78	+ {1, 0x00fc7}, {0, 0x0102d}, {1, 0x01031}, {0, 0x01032}, {1, 0x01033},
	79	+ {0, 0x01036}, {1, 0x01038}, {0, 0x01039}, {1, 0x0103a}, {0, 0x01058},
	80	+ {1, 0x0105a}, {2, 0x01100}, {0, 0x01160}, {1, 0x01200}, {0, 0x0135f},
	81	+ {1, 0x01360}, {0, 0x01712}, {1, 0x01715}, {0, 0x01732}, {1, 0x01735},
	82	+ {0, 0x01752}, {1, 0x01754}, {0, 0x01772}, {1, 0x01774}, {0, 0x017b4},
	83	+ {1, 0x017b6}, {0, 0x017b7}, {1, 0x017be}, {0, 0x017c6}, {1, 0x017c7},
	84	+ {0, 0x017c9}, {1, 0x017d4}, {0, 0x017dd}, {1, 0x017de}, {0, 0x0180b},
	85	+ {1, 0x0180e}, {0, 0x018a9}, {1, 0x018aa}, {0, 0x01920}, {1, 0x01923},
	86	+ {0, 0x01927}, {1, 0x01929}, {0, 0x01932}, {1, 0x01933}, {0, 0x01939},
	87	+ {1, 0x0193c}, {0, 0x01a17}, {1, 0x01a19}, {0, 0x01b00}, {1, 0x01b04},
	88	+ {0, 0x01b34}, {1, 0x01b35}, {0, 0x01b36}, {1, 0x01b3b}, {0, 0x01b3c},
	89	+ {1, 0x01b3d}, {0, 0x01b42}, {1, 0x01b43}, {0, 0x01b6b}, {1, 0x01b74},
	90	+ {0, 0x01dc0}, {1, 0x01dcb}, {0, 0x01dfe}, {1, 0x01e00}, {0, 0x0200b},
	91	+ {1, 0x02010}, {0, 0x0202a}, {1, 0x0202f}, {0, 0x02060}, {1, 0x02064},
	92	+ {0, 0x0206a}, {1, 0x02070}, {0, 0x020d0}, {1, 0x020f0}, {2, 0x02329},
	93	+ {1, 0x0232b}, {2, 0x02e80}, {0, 0x0302a}, {2, 0x03030}, {1, 0x0303f},
	94	+ {2, 0x03040}, {0, 0x03099}, {2, 0x0309b}, {1, 0x0a4d0}, {0, 0x0a806},
	95	+ {1, 0x0a807}, {0, 0x0a80b}, {1, 0x0a80c}, {0, 0x0a825}, {1, 0x0a827},
	96	+ {2, 0x0ac00}, {1, 0x0d7a4}, {2, 0x0f900}, {1, 0x0fb00}, {0, 0x0fb1e},
	97	+ {1, 0x0fb1f}, {0, 0x0fe00}, {2, 0x0fe10}, {1, 0x0fe1a}, {0, 0x0fe20},
	98	+ {1, 0x0fe24}, {2, 0x0fe30}, {1, 0x0fe70}, {0, 0x0feff}, {2, 0x0ff00},
	99	+ {1, 0x0ff61}, {2, 0x0ffe0}, {1, 0x0ffe7}, {0, 0x0fff9}, {1, 0x0fffc},
	100	+ {0, 0x10a01}, {1, 0x10a04}, {0, 0x10a05}, {1, 0x10a07}, {0, 0x10a0c},
	101	+ {1, 0x10a10}, {0, 0x10a38}, {1, 0x10a3b}, {0, 0x10a3f}, {1, 0x10a40},
	102	+ {0, 0x1d167}, {1, 0x1d16a}, {0, 0x1d173}, {1, 0x1d183}, {0, 0x1d185},
	103	+ {1, 0x1d18c}, {0, 0x1d1aa}, {1, 0x1d1ae}, {0, 0x1d242}, {1, 0x1d245},
	104	+ {2, 0x20000}, {1, 0x2fffe}, {2, 0x30000}, {1, 0x3fffe}, {0, 0xe0001},
	105	+ {1, 0xe0002}, {0, 0xe0020}, {1, 0xe0080}, {0, 0xe0100}, {1, 0xe01f0}
	106	+};
	107	+
	108	+/*
	109	+** Return an estimate of the width, in columns, for the single Unicode
	110	+** character c. For normal characters, the answer is always 1. But the
	111	+** estimate might be 0 or 2 for zero-width and double-width characters.
	112	+**
	113	+** Different display devices display unicode using different widths. So
	114	+** it is impossible to know that true display width with 100% accuracy.
	115	+** Inaccuracies in the width estimates might cause columns to be misaligned.
	116	+** Unfortunately, there is nothing we can do about that.
	117	+*/
	118	+static int cli_wcwidth(int c){
	119	+ int iFirst, iLast;
	120	+
	121	+ /* Fast path for common characters */
	122	+ if( c<=0x300 ) return 1;
	123	+
	124	+ /* The general case */
	125	+ iFirst = 0;
	126	+ iLast = sizeof(aUWidth)/sizeof(aUWidth[0]) - 1;
	127	+ while( iFirst<iLast-1 ){
	128	+ int iMid = (iFirst+iLast)/2;
	129	+ int cMid = aUWidth[iMid].iFirst;
	130	+ if( cMid < c ){
	131	+ iFirst = iMid;
	132	+ }else if( cMid > c ){
	133	+ iLast = iMid - 1;
	134	+ }else{
	135	+ return aUWidth[iMid].w;
	136	+ }
	137	+ }
	138	+ if( aUWidth[iLast].iFirst > c ) return aUWidth[iFirst].w;
	139	+ return aUWidth[iLast].w;
	140	+}
	141	+/***** End of code copied from SQLite ***********************************/
	142	+
36	143	/*
37	144	** This is the previous value used by most external callers when they
38	145	** needed to specify a default maximum line length to be used with the
39	146	** comment_print() function.
40	147	*/
		@@ -108,62 +215,99 @@
108	215	** algorithm, the NUL character is treated the same as a spacing character.
109	216	*/
110	217	static int comment_next_space(
111	218	const char zLine, / [in] The comment line being printed. */
112	219	int index, /* [in] The current character index being handled. */
113		- int distUTF8 / [out] Distance to next space in UTF-8 sequences. */
	220	+ int maxChars, /* [in] Optimization hint to abort before space found. */
	221	+ int sumWidth / [out] Summated width of all characters to next space. */
114	222	){
115		- int nextIndex = index + 1;
116		- int fNonASCII=0;
	223	+ int cchUTF8, utf32, wcwidth = 0;
	224	+ int nextIndex = index;
117	225	for(;;){
118		- char c = zLine[nextIndex];
119		- if( (c&0x80)==0x80 ) fNonASCII=1;
120		- if( c==0 \|\| fossil_isspace(c) ){
121		- if( distUTF8 ){
122		- if( fNonASCII!=0 ){
123		- *distUTF8 = strlen_utf8(&zLine[index], nextIndex-index);
124		- }else{
125		- *distUTF8 = nextIndex-index;
126		- }
127		- }
	226	+ char_info_utf8(&zLine[nextIndex],&cchUTF8,&utf32);
	227	+ nextIndex += cchUTF8;
	228	+ wcwidth += cli_wcwidth(utf32);
	229	+ if( zLine[nextIndex]==0 \|\| fossil_isspace(zLine[nextIndex]) \|\|
	230	+ wcwidth>maxChars ){
	231	+ *sumWidth = wcwidth;
128	232	return nextIndex;
129	233	}
130		- nextIndex++;
131	234	}
132	235	return 0; /* NOT REACHED */
133	236	}
134	237
135	238	/*
136		-** Count the number of UTF-8 sequences in a string. Incomplete, ill-formed and
137		-** overlong sequences are counted as one sequence. The invalid lead bytes 0xC0
138		-** to 0xC1 and 0xF5 to 0xF7 are allowed to initiate (ill-formed) 2- and 4-byte
139		-** sequences, respectively, the other invalid lead bytes 0xF8 to 0xFF are
140		-** treated as invalid 1-byte sequences (as lone trail bytes).
141		-** Combining characters and East Asian Wide and Fullwidth characters are counted
142		-** as one, so this function does not calculate the effective "display width".
	239	+** Return information about the next (single- or multi-byte) character in the
	240	+** specified UTF-8 string: The number of UTF-8 code units (in this case: bytes)
	241	+** and the decoded UTF-32 code point. Incomplete, ill-formed and overlong
	242	+** sequences are consumed together as one invalid code point. The invalid lead
	243	+** bytes 0xC0 to 0xC1 and 0xF5 to 0xF7 are allowed to initiate (ill-formed) 2-
	244	+** and 4-byte sequences, respectively, the other invalid lead bytes 0xF8 to 0xFF
	245	+** are treated as invalid 1-byte sequences (as lone trail bytes), all resulting
	246	+** in one invalid code point. Invalid UTF-8 sequences encoding a non-scalar code
	247	+** point (UTF-16 surrogates U+D800 to U+DFFF) are allowed.
143	248	*/
144		-int strlen_utf8(const char *zString, int lengthBytes){
145		- int i; /* Counted bytes. */
146		- int lengthUTF8; /* Counted UTF-8 sequences. */
147		-#if 0
148		- assert( lengthBytes>=0 );
	249	+void char_info_utf8(
	250	+ const char *z,
	251	+ int *pCchUTF8,
	252	+ int *pUtf32
	253	+){
	254	+ int i = 0; /* Counted bytes. */
	255	+ int cchUTF8 = 1; /* Code units consumed. */
	256	+ int maxUTF8 = 1; /* Expected sequence length. */
	257	+ char c = z[i++];
	258	+ if( (c&0x80)==0x00 ){ /* 7-bit ASCII character. */
	259	+ *pCchUTF8 = 1;
	260	+ *pUtf32 = (int)z[0];
	261	+ return;
	262	+ }
	263	+ else if( (c&0xe0)==0xc0 ) maxUTF8 = 2; /* UTF-8 lead byte 110vvvvv */
	264	+ else if( (c&0xf0)==0xe0 ) maxUTF8 = 3; /* UTF-8 lead byte 1110vvvv */
	265	+ else if( (c&0xf8)==0xf0 ) maxUTF8 = 4; /* UTF-8 lead byte 11110vvv */
	266	+ while( cchUTF8<maxUTF8 &&
	267	+ (z[i]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
	268	+ cchUTF8++;
	269	+ i++;
	270	+ }
	271	+ *pCchUTF8 = cchUTF8;
	272	+ if( cchUTF8!=maxUTF8 \|\| /* Incomplete UTF-8 sequence. */
	273	+ ( cchUTF8==1 && (c&0x80)==0x80 )){ /* Lone UTF-8 trail byte. */
	274	+ pUtf32 = 0xfffd; / U+FFFD Replacement Character */
	275	+#ifdef FOSSIL_DEBUG
	276	+ assert( pUtf32!=0xfffd ); / Invalid UTF-8 sequence. */
149	277	#endif
150		- for(i=0, lengthUTF8=0; i<lengthBytes; i++, lengthUTF8++){
151		- char c = zString[i];
152		- int cchUTF8=1; /* Code units consumed. */
153		- int maxUTF8=1; /* Expected sequence length. */
154		- if( (c&0xe0)==0xc0 )maxUTF8=2; /* UTF-8 lead byte 110vvvvv */
155		- else if( (c&0xf0)==0xe0 )maxUTF8=3; /* UTF-8 lead byte 1110vvvv */
156		- else if( (c&0xf8)==0xf0 )maxUTF8=4; /* UTF-8 lead byte 11110vvv */
157		- while( cchUTF8<maxUTF8 &&
158		- i<lengthBytes-1 &&
159		- (zString[i+1]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
160		- cchUTF8++;
161		- i++;
162		- }
163		- }
164		- return lengthUTF8;
	278	+ return;
	279	+ }
	280	+ switch( cchUTF8 ){
	281	+ case 4:
	282	+ *pUtf32 =
	283	+ ( (z[0] & 0x0f)<<18 ) \|
	284	+ ( (z[1] & 0x3f)<<12 ) \|
	285	+ ( (z[2] & 0x3f)<< 6 ) \|
	286	+ ( (z[4] & 0x3f)<< 0 ) ;
	287	+ break;
	288	+ case 3:
	289	+ *pUtf32 =
	290	+ ( (z[0] & 0x0f)<<12 ) \|
	291	+ ( (z[1] & 0x3f)<< 6 ) \|
	292	+ ( (z[2] & 0x3f)<< 0 ) ;
	293	+ break;
	294	+ case 2:
	295	+ *pUtf32 =
	296	+ ( (z[0] & 0x1f)<< 6 ) \|
	297	+ ( (z[1] & 0x3f)<< 0 ) ;
	298	+ break;
	299	+ default:
	300	+ pUtf32 = 0xfffd; / U+FFFD Replacement Character */
	301	+ break;
	302	+ }
	303	+#ifdef FOSSIL_DEBUG
	304	+ assert(
	305	+ pUtf32>=0 && pUtf32<=0x10ffff && /* Valid range U+0000 to U+10FFFF. */
	306	+ pUtf32<0xd800 && pUtf32>0xdfff /* Non-scalar (UTF-16 surrogates). */
	307	+ );
	308	+#endif
165	309	}
166	310
167	311	/*
168	312	** This function is called when printing a logical comment line to calculate
169	313	** the necessary indenting. The caller needs to emit the indenting spaces.
		@@ -206,11 +350,10 @@
206	350	int pLineCnt, / [in/out] Pointer to the total line count. */
207	351	const char *pzLine / [out] Pointer to the end of the logical line. */
208	352	){
209	353	int index = 0, charCnt = 0, lineCnt = 0, maxChars, i;
210	354	char zBuf[400]; int iBuf=0; /* Output buffer and counter. */
211		- int cchUTF8, maxUTF8; /* Helper variables to count UTF-8 sequences. */
212	355	if( !zLine ) return;
213	356	if( lineChars<=0 ) return;
214	357	#if 0
215	358	assert( indent<sizeof(zBuf)-5 ); /* See following comments to explain */
216	359	assert( origIndent<sizeof(zBuf)-5 ); /* these limits. */
		@@ -229,10 +372,11 @@
229	372	/* Limit line indent to fit output buffer. */
230	373	origIndent = sizeof(zBuf)-6;
231	374	}
232	375	maxChars = lineChars;
233	376	for(;;){
	377	+ int cchUTF8, utf32;
234	378	int useChars = 1;
235	379	char c = zLine[index];
236	380	/* Flush the output buffer if there's no space left for at least one more
237	381	** (potentially 4-byte) UTF-8 sequence, one level of indentation spaces,
238	382	** a new line, and a terminating NULL. */
		@@ -260,43 +404,47 @@
260	404	if( c=='\n' ){
261	405	lineCnt++;
262	406	charCnt = 0;
263	407	useChars = 0;
264	408	}else if( c=='\t' ){
265		- int distUTF8;
266		- int nextIndex = comment_next_space(zLine, index, &distUTF8);
267		- if( nextIndex<=0 \|\| distUTF8>maxChars ){
	409	+ int sumWidth;
	410	+ int nextIndex = comment_next_space(zLine, index, maxChars, &sumWidth);
	411	+ if( nextIndex<=0 \|\| sumWidth>maxChars ){
268	412	break;
269	413	}
270	414	charCnt++;
271	415	useChars = COMMENT_TAB_WIDTH;
272	416	if( maxChars<useChars ){
273	417	zBuf[iBuf++] = ' ';
274	418	break;
275	419	}
276	420	}else if( wordBreak && fossil_isspace(c) ){
277		- int distUTF8;
278		- int nextIndex = comment_next_space(zLine, index, &distUTF8);
279		- if( nextIndex<=0 \|\| distUTF8>=maxChars ){
	421	+ int sumWidth;
	422	+ int nextIndex = comment_next_space(zLine, index, maxChars, &sumWidth);
	423	+ if( nextIndex<=0 \|\| sumWidth>=maxChars ){
280	424	break;
281	425	}
282	426	charCnt++;
283	427	}else{
284	428	charCnt++;
285	429	}
286	430	assert( c!='\n' \|\| charCnt==0 );
287	431	zBuf[iBuf++] = c;
288		- /* Skip over UTF-8 sequences, see comment on strlen_utf8() for details. */
289		- cchUTF8=1; /* Code units consumed. */
290		- maxUTF8=1; /* Expected sequence length. */
291		- if( (c&0xe0)==0xc0 )maxUTF8=2; /* UTF-8 lead byte 110vvvvv */
292		- else if( (c&0xf0)==0xe0 )maxUTF8=3; /* UTF-8 lead byte 1110vvvv */
293		- else if( (c&0xf8)==0xf0 )maxUTF8=4; /* UTF-8 lead byte 11110vvv */
294		- while( cchUTF8<maxUTF8 &&
295		- (zLine[index]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
296		- cchUTF8++;
297		- zBuf[iBuf++] = zLine[index++];
	432	+ char_info_utf8(&zLine[index-1],&cchUTF8,&utf32);
	433	+ if( cchUTF8>1 ){
	434	+ int wcwidth;
	435	+ wcwidth = cli_wcwidth(utf32);
	436	+ if( wcwidth>maxChars && lineChars>=wcwidth ){ /* rollback */
	437	+ index--;
	438	+ iBuf--;
	439	+ zBuf[iBuf] = 0;
	440	+ break;
	441	+ }
	442	+ for( ; cchUTF8>1; cchUTF8-- ){
	443	+ zBuf[iBuf++] = zLine[index++];
	444	+ }
	445	+ useChars += wcwidth - 1;
298	446	}
299	447	maxChars -= useChars;
300	448	if( maxChars<=0 ) break;
301	449	if( c=='\n' ) break;
302	450	}
		@@ -338,11 +486,10 @@
338	486	int si, sk, i, k, kc;
339	487	int doIndent = 0;
340	488	char *zBuf;
341	489	char zBuffer[400];
342	490	int lineCnt = 0;
343		- int cchUTF8, maxUTF8; /* Helper variables to count UTF-8 sequences. */
344	491
345	492	if( width<0 ){
346	493	comment_set_maxchars(indent, &maxChars);
347	494	}
348	495	if( zText==0 ) zText = "(NULL)";
		@@ -364,25 +511,25 @@
364	511	}
365	512	if( zBuf!=zBuffer) fossil_free(zBuf);
366	513	return lineCnt;
367	514	}
368	515	for(sk=si=i=k=kc=0; zText[i] && kc<maxChars; i++){
	516	+ int cchUTF8, utf32;
369	517	char c = zText[i];
370	518	kc++; /* Count complete UTF-8 sequences. */
371		- /* Skip over UTF-8 sequences, see comment on strlen_utf8() for details. */
372		- cchUTF8=1; /* Code units consumed. */
373		- maxUTF8=1; /* Expected sequence length. */
374		- if( (c&0xe0)==0xc0 )maxUTF8=2; /* UTF-8 lead byte 110vvvvv */
375		- else if( (c&0xf0)==0xe0 )maxUTF8=3; /* UTF-8 lead byte 1110vvvv */
376		- else if( (c&0xf8)==0xf0 )maxUTF8=4; /* UTF-8 lead byte 11110vvv */
377		- if( maxUTF8>1 ){
378		- zBuf[k++] = c;
379		- while( cchUTF8<maxUTF8 &&
380		- (zText[i+1]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
381		- cchUTF8++;
	519	+ char_info_utf8(&zText[i],&cchUTF8,&utf32);
	520	+ if( cchUTF8>1 ){
	521	+ int wcwidth;
	522	+ wcwidth = cli_wcwidth(utf32);
	523	+ if( kc+wcwidth-1>maxChars && maxChars>=wcwidth ){ /* rollback */
	524	+ kc--;
	525	+ break;
	526	+ }
	527	+ for( i--; cchUTF8>0; cchUTF8-- ){
382	528	zBuf[k++] = zText[++i];
383	529	}
	530	+ kc += wcwidth - 1;
384	531	}
385	532	else if( fossil_isspace(c) ){
386	533	si = i;
387	534	sk = k;
388	535	if( k==0 \|\| zBuf[k-1]!=' ' ){
		@@ -497,11 +644,13 @@
497	644	zLine = zText;
498	645	for(;;){
499	646	comment_print_line(zOrigText, zLine, indent, zLine>zText ? indent : 0,
500	647	maxChars, trimCrLf, trimSpace, wordBreak, origBreak,
501	648	&lineCnt, &zLine);
502		- if( !zLine \|\| !zLine[0] ) break;
	649	+ if( zLine==0 ) break;
	650	+ while( fossil_isspace(zLine[0]) ) zLine++;
	651	+ if( zLine[0]==0 ) break;
503	652	}
504	653	return lineCnt;
505	654	}
506	655
507	656	/*
		@@ -597,10 +746,11 @@
597	746	if( zIndent ){
598	747	indent = atoi(zIndent);
599	748	}else{
600	749	indent = -1; /* automatic */
601	750	}
	751	+ verify_all_options();
602	752	if( g.argc!=4 && g.argc!=5 ){
603	753	usage("?OPTIONS? PREFIX TEXT ?ORIGTEXT?");
604	754	}
605	755	zPrefix = g.argv[2];
606	756	zText = g.argv[3];
607	757

	--- src/comformat.c
	+++ src/comformat.c
	@@ -31,10 +31,117 @@
31	#define COMMENT_PRINT_ORIG_BREAK ((u32)0x00000010) /* Break before original. */
32	#define COMMENT_PRINT_DEFAULT (COMMENT_PRINT_LEGACY) /* Defaults. */
33	#define COMMENT_PRINT_UNSET (-1) /* Not initialized. */
34	#endif
35











































































































36	/*
37	** This is the previous value used by most external callers when they
38	** needed to specify a default maximum line length to be used with the
39	** comment_print() function.
40	*/
	@@ -108,62 +215,99 @@
108	** algorithm, the NUL character is treated the same as a spacing character.
109	*/
110	static int comment_next_space(
111	const char zLine, / [in] The comment line being printed. */
112	int index, /* [in] The current character index being handled. */
113	int distUTF8 / [out] Distance to next space in UTF-8 sequences. */

114	){
115	int nextIndex = index + 1;
116	int fNonASCII=0;
117	for(;;){
118	char c = zLine[nextIndex];
119	if( (c&0x80)==0x80 ) fNonASCII=1;
120	if( c==0 \|\| fossil_isspace(c) ){
121	if( distUTF8 ){
122	if( fNonASCII!=0 ){
123	*distUTF8 = strlen_utf8(&zLine[index], nextIndex-index);
124	}else{
125	*distUTF8 = nextIndex-index;
126	}
127	}
128	return nextIndex;
129	}
130	nextIndex++;
131	}
132	return 0; /* NOT REACHED */
133	}
134
135	/*
136	** Count the number of UTF-8 sequences in a string. Incomplete, ill-formed and
137	** overlong sequences are counted as one sequence. The invalid lead bytes 0xC0
138	** to 0xC1 and 0xF5 to 0xF7 are allowed to initiate (ill-formed) 2- and 4-byte
139	** sequences, respectively, the other invalid lead bytes 0xF8 to 0xFF are
140	** treated as invalid 1-byte sequences (as lone trail bytes).
141	** Combining characters and East Asian Wide and Fullwidth characters are counted
142	** as one, so this function does not calculate the effective "display width".


143	*/
144	int strlen_utf8(const char *zString, int lengthBytes){
145	int i; /* Counted bytes. */
146	int lengthUTF8; /* Counted UTF-8 sequences. */
147	#if 0
148	assert( lengthBytes>=0 );























149	#endif
150	for(i=0, lengthUTF8=0; i<lengthBytes; i++, lengthUTF8++){
151	char c = zString[i];
152	int cchUTF8=1; /* Code units consumed. */
153	int maxUTF8=1; /* Expected sequence length. */
154	if( (c&0xe0)==0xc0 )maxUTF8=2; /* UTF-8 lead byte 110vvvvv */
155	else if( (c&0xf0)==0xe0 )maxUTF8=3; /* UTF-8 lead byte 1110vvvv */
156	else if( (c&0xf8)==0xf0 )maxUTF8=4; /* UTF-8 lead byte 11110vvv */
157	while( cchUTF8<maxUTF8 &&
158	i<lengthBytes-1 &&
159	(zString[i+1]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
160	cchUTF8++;
161	i++;
162	}
163	}
164	return lengthUTF8;
















165	}
166
167	/*
168	** This function is called when printing a logical comment line to calculate
169	** the necessary indenting. The caller needs to emit the indenting spaces.
	@@ -206,11 +350,10 @@
206	int pLineCnt, / [in/out] Pointer to the total line count. */
207	const char *pzLine / [out] Pointer to the end of the logical line. */
208	){
209	int index = 0, charCnt = 0, lineCnt = 0, maxChars, i;
210	char zBuf[400]; int iBuf=0; /* Output buffer and counter. */
211	int cchUTF8, maxUTF8; /* Helper variables to count UTF-8 sequences. */
212	if( !zLine ) return;
213	if( lineChars<=0 ) return;
214	#if 0
215	assert( indent<sizeof(zBuf)-5 ); /* See following comments to explain */
216	assert( origIndent<sizeof(zBuf)-5 ); /* these limits. */
	@@ -229,10 +372,11 @@
229	/* Limit line indent to fit output buffer. */
230	origIndent = sizeof(zBuf)-6;
231	}
232	maxChars = lineChars;
233	for(;;){

234	int useChars = 1;
235	char c = zLine[index];
236	/* Flush the output buffer if there's no space left for at least one more
237	** (potentially 4-byte) UTF-8 sequence, one level of indentation spaces,
238	** a new line, and a terminating NULL. */
	@@ -260,43 +404,47 @@
260	if( c=='\n' ){
261	lineCnt++;
262	charCnt = 0;
263	useChars = 0;
264	}else if( c=='\t' ){
265	int distUTF8;
266	int nextIndex = comment_next_space(zLine, index, &distUTF8);
267	if( nextIndex<=0 \|\| distUTF8>maxChars ){
268	break;
269	}
270	charCnt++;
271	useChars = COMMENT_TAB_WIDTH;
272	if( maxChars<useChars ){
273	zBuf[iBuf++] = ' ';
274	break;
275	}
276	}else if( wordBreak && fossil_isspace(c) ){
277	int distUTF8;
278	int nextIndex = comment_next_space(zLine, index, &distUTF8);
279	if( nextIndex<=0 \|\| distUTF8>=maxChars ){
280	break;
281	}
282	charCnt++;
283	}else{
284	charCnt++;
285	}
286	assert( c!='\n' \|\| charCnt==0 );
287	zBuf[iBuf++] = c;
288	/* Skip over UTF-8 sequences, see comment on strlen_utf8() for details. */
289	cchUTF8=1; /* Code units consumed. */
290	maxUTF8=1; /* Expected sequence length. */
291	if( (c&0xe0)==0xc0 )maxUTF8=2; /* UTF-8 lead byte 110vvvvv */
292	else if( (c&0xf0)==0xe0 )maxUTF8=3; /* UTF-8 lead byte 1110vvvv */
293	else if( (c&0xf8)==0xf0 )maxUTF8=4; /* UTF-8 lead byte 11110vvv */
294	while( cchUTF8<maxUTF8 &&
295	(zLine[index]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
296	cchUTF8++;
297	zBuf[iBuf++] = zLine[index++];




298	}
299	maxChars -= useChars;
300	if( maxChars<=0 ) break;
301	if( c=='\n' ) break;
302	}
	@@ -338,11 +486,10 @@
338	int si, sk, i, k, kc;
339	int doIndent = 0;
340	char *zBuf;
341	char zBuffer[400];
342	int lineCnt = 0;
343	int cchUTF8, maxUTF8; /* Helper variables to count UTF-8 sequences. */
344
345	if( width<0 ){
346	comment_set_maxchars(indent, &maxChars);
347	}
348	if( zText==0 ) zText = "(NULL)";
	@@ -364,25 +511,25 @@
364	}
365	if( zBuf!=zBuffer) fossil_free(zBuf);
366	return lineCnt;
367	}
368	for(sk=si=i=k=kc=0; zText[i] && kc<maxChars; i++){

369	char c = zText[i];
370	kc++; /* Count complete UTF-8 sequences. */
371	/* Skip over UTF-8 sequences, see comment on strlen_utf8() for details. */
372	cchUTF8=1; /* Code units consumed. */
373	maxUTF8=1; /* Expected sequence length. */
374	if( (c&0xe0)==0xc0 )maxUTF8=2; /* UTF-8 lead byte 110vvvvv */
375	else if( (c&0xf0)==0xe0 )maxUTF8=3; /* UTF-8 lead byte 1110vvvv */
376	else if( (c&0xf8)==0xf0 )maxUTF8=4; /* UTF-8 lead byte 11110vvv */
377	if( maxUTF8>1 ){
378	zBuf[k++] = c;
379	while( cchUTF8<maxUTF8 &&
380	(zText[i+1]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
381	cchUTF8++;
382	zBuf[k++] = zText[++i];
383	}

384	}
385	else if( fossil_isspace(c) ){
386	si = i;
387	sk = k;
388	if( k==0 \|\| zBuf[k-1]!=' ' ){
	@@ -497,11 +644,13 @@
497	zLine = zText;
498	for(;;){
499	comment_print_line(zOrigText, zLine, indent, zLine>zText ? indent : 0,
500	maxChars, trimCrLf, trimSpace, wordBreak, origBreak,
501	&lineCnt, &zLine);
502	if( !zLine \|\| !zLine[0] ) break;


503	}
504	return lineCnt;
505	}
506
507	/*
	@@ -597,10 +746,11 @@
597	if( zIndent ){
598	indent = atoi(zIndent);
599	}else{
600	indent = -1; /* automatic */
601	}

602	if( g.argc!=4 && g.argc!=5 ){
603	usage("?OPTIONS? PREFIX TEXT ?ORIGTEXT?");
604	}
605	zPrefix = g.argv[2];
606	zText = g.argv[3];
607

	--- src/comformat.c
	+++ src/comformat.c
	@@ -31,10 +31,117 @@
31	#define COMMENT_PRINT_ORIG_BREAK ((u32)0x00000010) /* Break before original. */
32	#define COMMENT_PRINT_DEFAULT (COMMENT_PRINT_LEGACY) /* Defaults. */
33	#define COMMENT_PRINT_UNSET (-1) /* Not initialized. */
34	#endif
35
36	/******* Code copied from SQLite src/shell.c.in on 2024-09-30 ********/
37	/* Lookup table to estimate the number of columns consumed by a Unicode
38	** character.
39	*/
40	static const struct {
41	unsigned char w; /* Width of the character in columns */
42	int iFirst; /* First character in a span having this width */
43	} aUWidth[] = {
44	/* {1, 0x00000}, */
45	{0, 0x00300}, {1, 0x00370}, {0, 0x00483}, {1, 0x00487}, {0, 0x00488},
46	{1, 0x0048a}, {0, 0x00591}, {1, 0x005be}, {0, 0x005bf}, {1, 0x005c0},
47	{0, 0x005c1}, {1, 0x005c3}, {0, 0x005c4}, {1, 0x005c6}, {0, 0x005c7},
48	{1, 0x005c8}, {0, 0x00600}, {1, 0x00604}, {0, 0x00610}, {1, 0x00616},
49	{0, 0x0064b}, {1, 0x0065f}, {0, 0x00670}, {1, 0x00671}, {0, 0x006d6},
50	{1, 0x006e5}, {0, 0x006e7}, {1, 0x006e9}, {0, 0x006ea}, {1, 0x006ee},
51	{0, 0x0070f}, {1, 0x00710}, {0, 0x00711}, {1, 0x00712}, {0, 0x00730},
52	{1, 0x0074b}, {0, 0x007a6}, {1, 0x007b1}, {0, 0x007eb}, {1, 0x007f4},
53	{0, 0x00901}, {1, 0x00903}, {0, 0x0093c}, {1, 0x0093d}, {0, 0x00941},
54	{1, 0x00949}, {0, 0x0094d}, {1, 0x0094e}, {0, 0x00951}, {1, 0x00955},
55	{0, 0x00962}, {1, 0x00964}, {0, 0x00981}, {1, 0x00982}, {0, 0x009bc},
56	{1, 0x009bd}, {0, 0x009c1}, {1, 0x009c5}, {0, 0x009cd}, {1, 0x009ce},
57	{0, 0x009e2}, {1, 0x009e4}, {0, 0x00a01}, {1, 0x00a03}, {0, 0x00a3c},
58	{1, 0x00a3d}, {0, 0x00a41}, {1, 0x00a43}, {0, 0x00a47}, {1, 0x00a49},
59	{0, 0x00a4b}, {1, 0x00a4e}, {0, 0x00a70}, {1, 0x00a72}, {0, 0x00a81},
60	{1, 0x00a83}, {0, 0x00abc}, {1, 0x00abd}, {0, 0x00ac1}, {1, 0x00ac6},
61	{0, 0x00ac7}, {1, 0x00ac9}, {0, 0x00acd}, {1, 0x00ace}, {0, 0x00ae2},
62	{1, 0x00ae4}, {0, 0x00b01}, {1, 0x00b02}, {0, 0x00b3c}, {1, 0x00b3d},
63	{0, 0x00b3f}, {1, 0x00b40}, {0, 0x00b41}, {1, 0x00b44}, {0, 0x00b4d},
64	{1, 0x00b4e}, {0, 0x00b56}, {1, 0x00b57}, {0, 0x00b82}, {1, 0x00b83},
65	{0, 0x00bc0}, {1, 0x00bc1}, {0, 0x00bcd}, {1, 0x00bce}, {0, 0x00c3e},
66	{1, 0x00c41}, {0, 0x00c46}, {1, 0x00c49}, {0, 0x00c4a}, {1, 0x00c4e},
67	{0, 0x00c55}, {1, 0x00c57}, {0, 0x00cbc}, {1, 0x00cbd}, {0, 0x00cbf},
68	{1, 0x00cc0}, {0, 0x00cc6}, {1, 0x00cc7}, {0, 0x00ccc}, {1, 0x00cce},
69	{0, 0x00ce2}, {1, 0x00ce4}, {0, 0x00d41}, {1, 0x00d44}, {0, 0x00d4d},
70	{1, 0x00d4e}, {0, 0x00dca}, {1, 0x00dcb}, {0, 0x00dd2}, {1, 0x00dd5},
71	{0, 0x00dd6}, {1, 0x00dd7}, {0, 0x00e31}, {1, 0x00e32}, {0, 0x00e34},
72	{1, 0x00e3b}, {0, 0x00e47}, {1, 0x00e4f}, {0, 0x00eb1}, {1, 0x00eb2},
73	{0, 0x00eb4}, {1, 0x00eba}, {0, 0x00ebb}, {1, 0x00ebd}, {0, 0x00ec8},
74	{1, 0x00ece}, {0, 0x00f18}, {1, 0x00f1a}, {0, 0x00f35}, {1, 0x00f36},
75	{0, 0x00f37}, {1, 0x00f38}, {0, 0x00f39}, {1, 0x00f3a}, {0, 0x00f71},
76	{1, 0x00f7f}, {0, 0x00f80}, {1, 0x00f85}, {0, 0x00f86}, {1, 0x00f88},
77	{0, 0x00f90}, {1, 0x00f98}, {0, 0x00f99}, {1, 0x00fbd}, {0, 0x00fc6},
78	{1, 0x00fc7}, {0, 0x0102d}, {1, 0x01031}, {0, 0x01032}, {1, 0x01033},
79	{0, 0x01036}, {1, 0x01038}, {0, 0x01039}, {1, 0x0103a}, {0, 0x01058},
80	{1, 0x0105a}, {2, 0x01100}, {0, 0x01160}, {1, 0x01200}, {0, 0x0135f},
81	{1, 0x01360}, {0, 0x01712}, {1, 0x01715}, {0, 0x01732}, {1, 0x01735},
82	{0, 0x01752}, {1, 0x01754}, {0, 0x01772}, {1, 0x01774}, {0, 0x017b4},
83	{1, 0x017b6}, {0, 0x017b7}, {1, 0x017be}, {0, 0x017c6}, {1, 0x017c7},
84	{0, 0x017c9}, {1, 0x017d4}, {0, 0x017dd}, {1, 0x017de}, {0, 0x0180b},
85	{1, 0x0180e}, {0, 0x018a9}, {1, 0x018aa}, {0, 0x01920}, {1, 0x01923},
86	{0, 0x01927}, {1, 0x01929}, {0, 0x01932}, {1, 0x01933}, {0, 0x01939},
87	{1, 0x0193c}, {0, 0x01a17}, {1, 0x01a19}, {0, 0x01b00}, {1, 0x01b04},
88	{0, 0x01b34}, {1, 0x01b35}, {0, 0x01b36}, {1, 0x01b3b}, {0, 0x01b3c},
89	{1, 0x01b3d}, {0, 0x01b42}, {1, 0x01b43}, {0, 0x01b6b}, {1, 0x01b74},
90	{0, 0x01dc0}, {1, 0x01dcb}, {0, 0x01dfe}, {1, 0x01e00}, {0, 0x0200b},
91	{1, 0x02010}, {0, 0x0202a}, {1, 0x0202f}, {0, 0x02060}, {1, 0x02064},
92	{0, 0x0206a}, {1, 0x02070}, {0, 0x020d0}, {1, 0x020f0}, {2, 0x02329},
93	{1, 0x0232b}, {2, 0x02e80}, {0, 0x0302a}, {2, 0x03030}, {1, 0x0303f},
94	{2, 0x03040}, {0, 0x03099}, {2, 0x0309b}, {1, 0x0a4d0}, {0, 0x0a806},
95	{1, 0x0a807}, {0, 0x0a80b}, {1, 0x0a80c}, {0, 0x0a825}, {1, 0x0a827},
96	{2, 0x0ac00}, {1, 0x0d7a4}, {2, 0x0f900}, {1, 0x0fb00}, {0, 0x0fb1e},
97	{1, 0x0fb1f}, {0, 0x0fe00}, {2, 0x0fe10}, {1, 0x0fe1a}, {0, 0x0fe20},
98	{1, 0x0fe24}, {2, 0x0fe30}, {1, 0x0fe70}, {0, 0x0feff}, {2, 0x0ff00},
99	{1, 0x0ff61}, {2, 0x0ffe0}, {1, 0x0ffe7}, {0, 0x0fff9}, {1, 0x0fffc},
100	{0, 0x10a01}, {1, 0x10a04}, {0, 0x10a05}, {1, 0x10a07}, {0, 0x10a0c},
101	{1, 0x10a10}, {0, 0x10a38}, {1, 0x10a3b}, {0, 0x10a3f}, {1, 0x10a40},
102	{0, 0x1d167}, {1, 0x1d16a}, {0, 0x1d173}, {1, 0x1d183}, {0, 0x1d185},
103	{1, 0x1d18c}, {0, 0x1d1aa}, {1, 0x1d1ae}, {0, 0x1d242}, {1, 0x1d245},
104	{2, 0x20000}, {1, 0x2fffe}, {2, 0x30000}, {1, 0x3fffe}, {0, 0xe0001},
105	{1, 0xe0002}, {0, 0xe0020}, {1, 0xe0080}, {0, 0xe0100}, {1, 0xe01f0}
106	};
107
108	/*
109	** Return an estimate of the width, in columns, for the single Unicode
110	** character c. For normal characters, the answer is always 1. But the
111	** estimate might be 0 or 2 for zero-width and double-width characters.
112	**
113	** Different display devices display unicode using different widths. So
114	** it is impossible to know that true display width with 100% accuracy.
115	** Inaccuracies in the width estimates might cause columns to be misaligned.
116	** Unfortunately, there is nothing we can do about that.
117	*/
118	static int cli_wcwidth(int c){
119	int iFirst, iLast;
120
121	/* Fast path for common characters */
122	if( c<=0x300 ) return 1;
123
124	/* The general case */
125	iFirst = 0;
126	iLast = sizeof(aUWidth)/sizeof(aUWidth[0]) - 1;
127	while( iFirst<iLast-1 ){
128	int iMid = (iFirst+iLast)/2;
129	int cMid = aUWidth[iMid].iFirst;
130	if( cMid < c ){
131	iFirst = iMid;
132	}else if( cMid > c ){
133	iLast = iMid - 1;
134	}else{
135	return aUWidth[iMid].w;
136	}
137	}
138	if( aUWidth[iLast].iFirst > c ) return aUWidth[iFirst].w;
139	return aUWidth[iLast].w;
140	}
141	/***** End of code copied from SQLite ***********************************/
142
143	/*
144	** This is the previous value used by most external callers when they
145	** needed to specify a default maximum line length to be used with the
146	** comment_print() function.
147	*/
	@@ -108,62 +215,99 @@
215	** algorithm, the NUL character is treated the same as a spacing character.
216	*/
217	static int comment_next_space(
218	const char zLine, / [in] The comment line being printed. */
219	int index, /* [in] The current character index being handled. */
220	int maxChars, /* [in] Optimization hint to abort before space found. */
221	int sumWidth / [out] Summated width of all characters to next space. */
222	){
223	int cchUTF8, utf32, wcwidth = 0;
224	int nextIndex = index;
225	for(;;){
226	char_info_utf8(&zLine[nextIndex],&cchUTF8,&utf32);
227	nextIndex += cchUTF8;
228	wcwidth += cli_wcwidth(utf32);
229	if( zLine[nextIndex]==0 \|\| fossil_isspace(zLine[nextIndex]) \|\|
230	wcwidth>maxChars ){
231	*sumWidth = wcwidth;




232	return nextIndex;
233	}

234	}
235	return 0; /* NOT REACHED */
236	}
237
238	/*
239	** Return information about the next (single- or multi-byte) character in the
240	** specified UTF-8 string: The number of UTF-8 code units (in this case: bytes)
241	** and the decoded UTF-32 code point. Incomplete, ill-formed and overlong
242	** sequences are consumed together as one invalid code point. The invalid lead
243	** bytes 0xC0 to 0xC1 and 0xF5 to 0xF7 are allowed to initiate (ill-formed) 2-
244	** and 4-byte sequences, respectively, the other invalid lead bytes 0xF8 to 0xFF
245	** are treated as invalid 1-byte sequences (as lone trail bytes), all resulting
246	** in one invalid code point. Invalid UTF-8 sequences encoding a non-scalar code
247	** point (UTF-16 surrogates U+D800 to U+DFFF) are allowed.
248	*/
249	void char_info_utf8(
250	const char *z,
251	int *pCchUTF8,
252	int *pUtf32
253	){
254	int i = 0; /* Counted bytes. */
255	int cchUTF8 = 1; /* Code units consumed. */
256	int maxUTF8 = 1; /* Expected sequence length. */
257	char c = z[i++];
258	if( (c&0x80)==0x00 ){ /* 7-bit ASCII character. */
259	*pCchUTF8 = 1;
260	*pUtf32 = (int)z[0];
261	return;
262	}
263	else if( (c&0xe0)==0xc0 ) maxUTF8 = 2; /* UTF-8 lead byte 110vvvvv */
264	else if( (c&0xf0)==0xe0 ) maxUTF8 = 3; /* UTF-8 lead byte 1110vvvv */
265	else if( (c&0xf8)==0xf0 ) maxUTF8 = 4; /* UTF-8 lead byte 11110vvv */
266	while( cchUTF8<maxUTF8 &&
267	(z[i]&0xc0)==0x80 ){ /* UTF-8 trail byte 10vvvvvv */
268	cchUTF8++;
269	i++;
270	}
271	*pCchUTF8 = cchUTF8;
272	if( cchUTF8!=maxUTF8 \|\| /* Incomplete UTF-8 sequence. */
273	( cchUTF8==1 && (c&0x80)==0x80 )){ /* Lone UTF-8 trail byte. */
274	pUtf32 = 0xfffd; / U+FFFD Replacement Character */
275	#ifdef FOSSIL_DEBUG
276	assert( pUtf32!=0xfffd ); / Invalid UTF-8 sequence. */
277	#endif
278	return;
279	}
280	switch( cchUTF8 ){
281	case 4:
282	*pUtf32 =
283	( (z[0] & 0x0f)<<18 ) \|
284	( (z[1] & 0x3f)<<12 ) \|
285	( (z[2] & 0x3f)<< 6 ) \|
286	( (z[4] & 0x3f)<< 0 ) ;
287	break;
288	case 3:
289	*pUtf32 =
290	( (z[0] & 0x0f)<<12 ) \|
291	( (z[1] & 0x3f)<< 6 ) \|
292	( (z[2] & 0x3f)<< 0 ) ;
293	break;
294	case 2:
295	*pUtf32 =
296	( (z[0] & 0x1f)<< 6 ) \|
297	( (z[1] & 0x3f)<< 0 ) ;
298	break;
299	default:
300	pUtf32 = 0xfffd; / U+FFFD Replacement Character */
301	break;
302	}
303	#ifdef FOSSIL_DEBUG
304	assert(
305	pUtf32>=0 && pUtf32<=0x10ffff && /* Valid range U+0000 to U+10FFFF. */
306	pUtf32<0xd800 && pUtf32>0xdfff /* Non-scalar (UTF-16 surrogates). */
307	);
308	#endif
309	}
310
311	/*
312	** This function is called when printing a logical comment line to calculate
313	** the necessary indenting. The caller needs to emit the indenting spaces.
	@@ -206,11 +350,10 @@
350	int pLineCnt, / [in/out] Pointer to the total line count. */
351	const char *pzLine / [out] Pointer to the end of the logical line. */
352	){
353	int index = 0, charCnt = 0, lineCnt = 0, maxChars, i;
354	char zBuf[400]; int iBuf=0; /* Output buffer and counter. */

355	if( !zLine ) return;
356	if( lineChars<=0 ) return;
357	#if 0
358	assert( indent<sizeof(zBuf)-5 ); /* See following comments to explain */
359	assert( origIndent<sizeof(zBuf)-5 ); /* these limits. */
	@@ -229,10 +372,11 @@
372	/* Limit line indent to fit output buffer. */
373	origIndent = sizeof(zBuf)-6;
374	}
375	maxChars = lineChars;
376	for(;;){
377	int cchUTF8, utf32;
378	int useChars = 1;
379	char c = zLine[index];
380	/* Flush the output buffer if there's no space left for at least one more
381	** (potentially 4-byte) UTF-8 sequence, one level of indentation spaces,
382	** a new line, and a terminating NULL. */
	@@ -260,43 +404,47 @@
404	if( c=='\n' ){
405	lineCnt++;
406	charCnt = 0;
407	useChars = 0;
408	}else if( c=='\t' ){
409	int sumWidth;
410	int nextIndex = comment_next_space(zLine, index, maxChars, &sumWidth);
411	if( nextIndex<=0 \|\| sumWidth>maxChars ){
412	break;
413	}
414	charCnt++;
415	useChars = COMMENT_TAB_WIDTH;
416	if( maxChars<useChars ){
417	zBuf[iBuf++] = ' ';
418	break;
419	}
420	}else if( wordBreak && fossil_isspace(c) ){
421	int sumWidth;
422	int nextIndex = comment_next_space(zLine, index, maxChars, &sumWidth);
423	if( nextIndex<=0 \|\| sumWidth>=maxChars ){
424	break;
425	}
426	charCnt++;
427	}else{
428	charCnt++;
429	}
430	assert( c!='\n' \|\| charCnt==0 );
431	zBuf[iBuf++] = c;
432	char_info_utf8(&zLine[index-1],&cchUTF8,&utf32);
433	if( cchUTF8>1 ){
434	int wcwidth;
435	wcwidth = cli_wcwidth(utf32);
436	if( wcwidth>maxChars && lineChars>=wcwidth ){ /* rollback */
437	index--;
438	iBuf--;
439	zBuf[iBuf] = 0;
440	break;
441	}
442	for( ; cchUTF8>1; cchUTF8-- ){
443	zBuf[iBuf++] = zLine[index++];
444	}
445	useChars += wcwidth - 1;
446	}
447	maxChars -= useChars;
448	if( maxChars<=0 ) break;
449	if( c=='\n' ) break;
450	}
	@@ -338,11 +486,10 @@
486	int si, sk, i, k, kc;
487	int doIndent = 0;
488	char *zBuf;
489	char zBuffer[400];
490	int lineCnt = 0;

491
492	if( width<0 ){
493	comment_set_maxchars(indent, &maxChars);
494	}
495	if( zText==0 ) zText = "(NULL)";
	@@ -364,25 +511,25 @@
511	}
512	if( zBuf!=zBuffer) fossil_free(zBuf);
513	return lineCnt;
514	}
515	for(sk=si=i=k=kc=0; zText[i] && kc<maxChars; i++){
516	int cchUTF8, utf32;
517	char c = zText[i];
518	kc++; /* Count complete UTF-8 sequences. */
519	char_info_utf8(&zText[i],&cchUTF8,&utf32);
520	if( cchUTF8>1 ){
521	int wcwidth;
522	wcwidth = cli_wcwidth(utf32);
523	if( kc+wcwidth-1>maxChars && maxChars>=wcwidth ){ /* rollback */
524	kc--;
525	break;
526	}
527	for( i--; cchUTF8>0; cchUTF8-- ){


528	zBuf[k++] = zText[++i];
529	}
530	kc += wcwidth - 1;
531	}
532	else if( fossil_isspace(c) ){
533	si = i;
534	sk = k;
535	if( k==0 \|\| zBuf[k-1]!=' ' ){
	@@ -497,11 +644,13 @@
644	zLine = zText;
645	for(;;){
646	comment_print_line(zOrigText, zLine, indent, zLine>zText ? indent : 0,
647	maxChars, trimCrLf, trimSpace, wordBreak, origBreak,
648	&lineCnt, &zLine);
649	if( zLine==0 ) break;
650	while( fossil_isspace(zLine[0]) ) zLine++;
651	if( zLine[0]==0 ) break;
652	}
653	return lineCnt;
654	}
655
656	/*
	@@ -597,10 +746,11 @@
746	if( zIndent ){
747	indent = atoi(zIndent);
748	}else{
749	indent = -1; /* automatic */
750	}
751	verify_all_options();
752	if( g.argc!=4 && g.argc!=5 ){
753	usage("?OPTIONS? PREFIX TEXT ?ORIGTEXT?");
754	}
755	zPrefix = g.argv[2];
756	zText = g.argv[3];
757

M src/configure.c

+28 -9

		--- src/configure.c
		+++ src/configure.c
		@@ -118,24 +118,15 @@
118	118	{ "adunit-omit-if-user", CONFIGSET_SKIN },
119	119	{ "default-csp", CONFIGSET_SKIN },
120	120	{ "sitemap-extra", CONFIGSET_SKIN },
121	121	{ "safe-html", CONFIGSET_SKIN },
122	122
123		-#ifdef FOSSIL_ENABLE_TH1_DOCS
124		- { "th1-docs", CONFIGSET_TH1 },
125		-#endif
126	123	#ifdef FOSSIL_ENABLE_TH1_HOOKS
127	124	{ "th1-hooks", CONFIGSET_TH1 },
128	125	#endif
129		- { "th1-setup", CONFIGSET_TH1 },
130	126	{ "th1-uri-regexp", CONFIGSET_TH1 },
131	127
132		-#ifdef FOSSIL_ENABLE_TCL
133		- { "tcl", CONFIGSET_TH1 },
134		- { "tcl-setup", CONFIGSET_TH1 },
135		-#endif
136		-
137	128	{ "project-name", CONFIGSET_PROJ },
138	129	{ "short-project-name", CONFIGSET_PROJ },
139	130	{ "project-description", CONFIGSET_PROJ },
140	131	{ "index-page", CONFIGSET_PROJ },
141	132	{ "manifest", CONFIGSET_PROJ },
		@@ -239,17 +230,35 @@
239	230	**
240	231	** "Safe" in the previous paragraph means the permission is granted to
241	232	** export the property. In other words, the requesting side has presented
242	233	** login credentials and has sufficient capabilities to access the requested
243	234	** information.
	235	+**
	236	+** Settings which are specifically flagged as sensitive will (as of
	237	+** 2024-10-15) cause this function to return 0, regardless of user
	238	+** permissions. As an example, if the th1-setup setting were not
	239	+** sensitive then a malicious repo admin could set that to include
	240	+** arbitrary TCL code and affect users who configure fossil with the
	241	+** --with-tcl flag.
244	242	*/
245	243	int configure_is_exportable(const char *zName){
246	244	int i;
247	245	int n = strlen(zName);
	246	+ Setting *pSet;
248	247	if( n>2 && zName[0]=='\'' && zName[n-1]=='\'' ){
	248	+ char * zCpy;
249	249	zName++;
250	250	n -= 2;
	251	+ zCpy = fossil_strndup(zName, (ssize_t)n);
	252	+ pSet = db_find_setting(zCpy, 0);
	253	+ fossil_free(zCpy);
	254	+ }else{
	255	+ pSet = db_find_setting(zName, 0);
	256	+ }
	257	+ if( pSet && pSet->sensitive ){
	258	+ /* https://fossil-scm.org/forum/forumpost/6179500deadf6ec7 */
	259	+ return 0;
251	260	}
252	261	for(i=0; i<count(aConfig); i++){
253	262	if( strncmp(zName, aConfig[i].zName, n)==0 && aConfig[i].zName[n]==0 ){
254	263	int m = aConfig[i].groupMask;
255	264	if( !g.perm.Admin ){
		@@ -414,10 +423,15 @@
414	423	if( nToken>=count(azToken)-1 ) break;
415	424	}
416	425	if( nToken<2 ) return;
417	426	if( aType[ii].zName[0]=='/' ){
418	427	thisMask = configure_is_exportable(azToken[1]);
	428	+ if( 0==thisMask ){
	429	+ fossil_warning("Skipping non-exportable setting: %s = %s",
	430	+ azToken[1], nToken>3 ? azToken[3] : "?");
	431	+ /* Will be skipped below */
	432	+ }
419	433	}else{
420	434	thisMask = configure_is_exportable(aType[ii].zName);
421	435	}
422	436	if( (thisMask & groupMask)==0 ) return;
423	437	if( (thisMask & checkMask)!=0 ){
		@@ -681,10 +695,15 @@
681	695	}
682	696	db_prepare(&q, "SELECT mtime, quote(name), quote(value) FROM config"
683	697	" WHERE name=:name AND mtime>=%lld", iStart);
684	698	for(ii=0; ii<count(aConfig); ii++){
685	699	if( (aConfig[ii].groupMask & groupMask)!=0 && aConfig[ii].zName[0]!='@' ){
	700	+ const Setting * pSet = db_find_setting(aConfig[ii].zName, 0);
	701	+ if( pSet && pSet->sensitive ){
	702	+ /* https://fossil-scm.org/forum/forumpost/6179500deadf6ec7 */
	703	+ continue;
	704	+ }
686	705	db_bind_text(&q, ":name", aConfig[ii].zName);
687	706	while( db_step(&q)==SQLITE_ROW ){
688	707	blob_appendf(&rec,"%s %s value %s",
689	708	db_column_text(&q, 0),
690	709	db_column_text(&q, 1),
691	710

	--- src/configure.c
	+++ src/configure.c
	@@ -118,24 +118,15 @@
118	{ "adunit-omit-if-user", CONFIGSET_SKIN },
119	{ "default-csp", CONFIGSET_SKIN },
120	{ "sitemap-extra", CONFIGSET_SKIN },
121	{ "safe-html", CONFIGSET_SKIN },
122
123	#ifdef FOSSIL_ENABLE_TH1_DOCS
124	{ "th1-docs", CONFIGSET_TH1 },
125	#endif
126	#ifdef FOSSIL_ENABLE_TH1_HOOKS
127	{ "th1-hooks", CONFIGSET_TH1 },
128	#endif
129	{ "th1-setup", CONFIGSET_TH1 },
130	{ "th1-uri-regexp", CONFIGSET_TH1 },
131
132	#ifdef FOSSIL_ENABLE_TCL
133	{ "tcl", CONFIGSET_TH1 },
134	{ "tcl-setup", CONFIGSET_TH1 },
135	#endif
136
137	{ "project-name", CONFIGSET_PROJ },
138	{ "short-project-name", CONFIGSET_PROJ },
139	{ "project-description", CONFIGSET_PROJ },
140	{ "index-page", CONFIGSET_PROJ },
141	{ "manifest", CONFIGSET_PROJ },
	@@ -239,17 +230,35 @@
239	**
240	** "Safe" in the previous paragraph means the permission is granted to
241	** export the property. In other words, the requesting side has presented
242	** login credentials and has sufficient capabilities to access the requested
243	** information.







244	*/
245	int configure_is_exportable(const char *zName){
246	int i;
247	int n = strlen(zName);

248	if( n>2 && zName[0]=='\'' && zName[n-1]=='\'' ){

249	zName++;
250	n -= 2;









251	}
252	for(i=0; i<count(aConfig); i++){
253	if( strncmp(zName, aConfig[i].zName, n)==0 && aConfig[i].zName[n]==0 ){
254	int m = aConfig[i].groupMask;
255	if( !g.perm.Admin ){
	@@ -414,10 +423,15 @@
414	if( nToken>=count(azToken)-1 ) break;
415	}
416	if( nToken<2 ) return;
417	if( aType[ii].zName[0]=='/' ){
418	thisMask = configure_is_exportable(azToken[1]);





419	}else{
420	thisMask = configure_is_exportable(aType[ii].zName);
421	}
422	if( (thisMask & groupMask)==0 ) return;
423	if( (thisMask & checkMask)!=0 ){
	@@ -681,10 +695,15 @@
681	}
682	db_prepare(&q, "SELECT mtime, quote(name), quote(value) FROM config"
683	" WHERE name=:name AND mtime>=%lld", iStart);
684	for(ii=0; ii<count(aConfig); ii++){
685	if( (aConfig[ii].groupMask & groupMask)!=0 && aConfig[ii].zName[0]!='@' ){





686	db_bind_text(&q, ":name", aConfig[ii].zName);
687	while( db_step(&q)==SQLITE_ROW ){
688	blob_appendf(&rec,"%s %s value %s",
689	db_column_text(&q, 0),
690	db_column_text(&q, 1),
691

	--- src/configure.c
	+++ src/configure.c
	@@ -118,24 +118,15 @@
118	{ "adunit-omit-if-user", CONFIGSET_SKIN },
119	{ "default-csp", CONFIGSET_SKIN },
120	{ "sitemap-extra", CONFIGSET_SKIN },
121	{ "safe-html", CONFIGSET_SKIN },
122



123	#ifdef FOSSIL_ENABLE_TH1_HOOKS
124	{ "th1-hooks", CONFIGSET_TH1 },
125	#endif

126	{ "th1-uri-regexp", CONFIGSET_TH1 },
127





128	{ "project-name", CONFIGSET_PROJ },
129	{ "short-project-name", CONFIGSET_PROJ },
130	{ "project-description", CONFIGSET_PROJ },
131	{ "index-page", CONFIGSET_PROJ },
132	{ "manifest", CONFIGSET_PROJ },
	@@ -239,17 +230,35 @@
230	**
231	** "Safe" in the previous paragraph means the permission is granted to
232	** export the property. In other words, the requesting side has presented
233	** login credentials and has sufficient capabilities to access the requested
234	** information.
235	**
236	** Settings which are specifically flagged as sensitive will (as of
237	** 2024-10-15) cause this function to return 0, regardless of user
238	** permissions. As an example, if the th1-setup setting were not
239	** sensitive then a malicious repo admin could set that to include
240	** arbitrary TCL code and affect users who configure fossil with the
241	** --with-tcl flag.
242	*/
243	int configure_is_exportable(const char *zName){
244	int i;
245	int n = strlen(zName);
246	Setting *pSet;
247	if( n>2 && zName[0]=='\'' && zName[n-1]=='\'' ){
248	char * zCpy;
249	zName++;
250	n -= 2;
251	zCpy = fossil_strndup(zName, (ssize_t)n);
252	pSet = db_find_setting(zCpy, 0);
253	fossil_free(zCpy);
254	}else{
255	pSet = db_find_setting(zName, 0);
256	}
257	if( pSet && pSet->sensitive ){
258	/* https://fossil-scm.org/forum/forumpost/6179500deadf6ec7 */
259	return 0;
260	}
261	for(i=0; i<count(aConfig); i++){
262	if( strncmp(zName, aConfig[i].zName, n)==0 && aConfig[i].zName[n]==0 ){
263	int m = aConfig[i].groupMask;
264	if( !g.perm.Admin ){
	@@ -414,10 +423,15 @@
423	if( nToken>=count(azToken)-1 ) break;
424	}
425	if( nToken<2 ) return;
426	if( aType[ii].zName[0]=='/' ){
427	thisMask = configure_is_exportable(azToken[1]);
428	if( 0==thisMask ){
429	fossil_warning("Skipping non-exportable setting: %s = %s",
430	azToken[1], nToken>3 ? azToken[3] : "?");
431	/* Will be skipped below */
432	}
433	}else{
434	thisMask = configure_is_exportable(aType[ii].zName);
435	}
436	if( (thisMask & groupMask)==0 ) return;
437	if( (thisMask & checkMask)!=0 ){
	@@ -681,10 +695,15 @@
695	}
696	db_prepare(&q, "SELECT mtime, quote(name), quote(value) FROM config"
697	" WHERE name=:name AND mtime>=%lld", iStart);
698	for(ii=0; ii<count(aConfig); ii++){
699	if( (aConfig[ii].groupMask & groupMask)!=0 && aConfig[ii].zName[0]!='@' ){
700	const Setting * pSet = db_find_setting(aConfig[ii].zName, 0);
701	if( pSet && pSet->sensitive ){
702	/* https://fossil-scm.org/forum/forumpost/6179500deadf6ec7 */
703	continue;
704	}
705	db_bind_text(&q, ":name", aConfig[ii].zName);
706	while( db_step(&q)==SQLITE_ROW ){
707	blob_appendf(&rec,"%s %s value %s",
708	db_column_text(&q, 0),
709	db_column_text(&q, 1),
710

M src/db.c

+13 -14

		--- src/db.c
		+++ src/db.c
		@@ -1557,10 +1557,12 @@
1557	1557	sqlite3_create_function(db, "url_nouser", 1, SQLITE_UTF8, 0,
1558	1558	url_nouser_func,0,0);
1559	1559	sqlite3_create_function(db, "chat_msg_from_event", 4,
1560	1560	SQLITE_UTF8 \| SQLITE_INNOCUOUS, 0,
1561	1561	chat_msg_from_event, 0, 0);
	1562	+ sqlite3_create_function(db, "inode", 1, SQLITE_UTF8, 0,
	1563	+ file_inode_sql_func,0,0);
1562	1564
1563	1565	}
1564	1566
1565	1567	#if USE_SEE
1566	1568	/*
		@@ -4080,11 +4082,11 @@
4080	4082	** specified then that version is checked out. Otherwise the most recent
4081	4083	** check-in on the main branch (usually "trunk") is used.
4082	4084	**
4083	4085	** REPOSITORY can be the filename for a repository that already exists on the
4084	4086	** local machine or it can be a URI for a remote repository. If REPOSITORY
4085		-** is a URI in one of the formats recognized by the [[clone]] command, then
	4087	+** is a URI in one of the formats recognized by the [[clone]] command, the
4086	4088	** remote repo is first cloned, then the clone is opened. The clone will be
4087	4089	** stored in the current directory, or in DIR if the "--repodir DIR" option
4088	4090	** is used. The name of the clone will be taken from the last term of the URI.
4089	4091	** For "http:" and "https:" URIs, you can append an extra term to the end of
4090	4092	** the URI to get any repository name you like. For example:
		@@ -4681,17 +4683,10 @@
4681	4683	** to obtain a check-in lock during auto-sync, the server will
4682	4684	** send the "pragma avoid-delta-manifests" statement in its reply,
4683	4685	** which will cause the client to avoid generating a delta
4684	4686	** manifest.
4685	4687	*/
4686		-/*
4687		-** SETTING: forum-close-policy boolean default=off
4688		-** If true, forum moderators may close/re-open forum posts, and reply
4689		-** to closed posts. If false, only administrators may do so. Note that
4690		-** this only affects the forum web UI, not post-closing tags which
4691		-** arrive via the command-line or from synchronization with a remote.
4692		-*/
4693	4688	/*
4694	4689	** SETTING: gdiff-command width=40 default=gdiff sensitive
4695	4690	** The value is an external command to run when performing a graphical
4696	4691	** diff. If undefined, text diff will be used.
4697	4692	*/
		@@ -4841,15 +4836,19 @@
4841	4836	** URL of the HTTP proxy. If undefined or "system", the "http_proxy"
4842	4837	** environment variable is consulted. If "off", a direct HTTP connection is
4843	4838	** used.
4844	4839	*/
4845	4840	/*
4846		-** SETTING: redirect-to-https default=0 width=-1
4847		-** Specifies whether or not to redirect http:// requests to
4848		-** https:// URIs. A value of 0 (the default) means not to
	4841	+** SETTING: redirect-to-https default=0 width=2
	4842	+** Specifies whether or not to redirect unencrypted "http://" requests to
	4843	+** encrypted "https://" URIs. A value of 0 (the default) means do not
4849	4844	** redirect, 1 means to redirect only the /login page, and 2
4850	4845	** means to always redirect.
	4846	+**
	4847	+** For security, a value of 2 is recommended. The default value is 0
	4848	+** because not all sites are TLS-capable. But you should definitely enable
	4849	+** TLS and change this setting to 2 for all public-facing repositories.
4851	4850	*/
4852	4851	/*
4853	4852	** SETTING: relative-paths boolean default=on
4854	4853	** When showing changes and extras, report paths relative
4855	4854	** to the current working directory.
		@@ -4967,11 +4966,11 @@
4967	4966	** If enabled, special TH1 commands will be called before and
4968	4967	** after any Fossil command or web page.
4969	4968	*/
4970	4969	#endif
4971	4970	/*
4972		-** SETTING: th1-setup width=40 block-text
	4971	+** SETTING: th1-setup width=40 block-text sensitive
4973	4972	** This is the setup script to be evaluated after creating
4974	4973	** and initializing the TH1 interpreter. By default, this
4975	4974	** is empty and no extra setup is performed.
4976	4975	*/
4977	4976	/*
		@@ -5358,11 +5357,11 @@
5358	5357	}
5359	5358
5360	5359	/*
5361	5360	** Compute a "fingerprint" on the repository. A fingerprint is used
5362	5361	** to verify that that the repository has not been replaced by a clone
5363		-** of the same repository. More precisely, a fingerprint are used to
	5362	+** of the same repository. More precisely, a fingerprint is used to
5364	5363	** verify that the mapping between SHA3 hashes and RID values is unchanged.
5365	5364	**
5366	5365	** The check-out database ("localdb") stores RID values. When associating
5367	5366	** a check-out database against a repository database, it is useful to verify
5368	5367	** the fingerprint so that we know tha the RID values in the check-out
		@@ -5423,11 +5422,11 @@
5423	5422	** COMMAND: test-fingerprint
5424	5423	**
5425	5424	** Usage: %fossil test-fingerprint ?RCVID?
5426	5425	**
5427	5426	** Display the repository fingerprint using the supplied RCVID or
5428		-** using the latest RCVID if not is given on the command line.
	5427	+** using the latest RCVID if none is given on the command line.
5429	5428	** Show both the legacy and the newer version of the fingerprint,
5430	5429	** and the currently stored fingerprint if there is one.
5431	5430	*/
5432	5431	void test_fingerprint(void){
5433	5432	int rcvid = 0;
5434	5433

	--- src/db.c
	+++ src/db.c
	@@ -1557,10 +1557,12 @@
1557	sqlite3_create_function(db, "url_nouser", 1, SQLITE_UTF8, 0,
1558	url_nouser_func,0,0);
1559	sqlite3_create_function(db, "chat_msg_from_event", 4,
1560	SQLITE_UTF8 \| SQLITE_INNOCUOUS, 0,
1561	chat_msg_from_event, 0, 0);


1562
1563	}
1564
1565	#if USE_SEE
1566	/*
	@@ -4080,11 +4082,11 @@
4080	** specified then that version is checked out. Otherwise the most recent
4081	** check-in on the main branch (usually "trunk") is used.
4082	**
4083	** REPOSITORY can be the filename for a repository that already exists on the
4084	** local machine or it can be a URI for a remote repository. If REPOSITORY
4085	** is a URI in one of the formats recognized by the [[clone]] command, then
4086	** remote repo is first cloned, then the clone is opened. The clone will be
4087	** stored in the current directory, or in DIR if the "--repodir DIR" option
4088	** is used. The name of the clone will be taken from the last term of the URI.
4089	** For "http:" and "https:" URIs, you can append an extra term to the end of
4090	** the URI to get any repository name you like. For example:
	@@ -4681,17 +4683,10 @@
4681	** to obtain a check-in lock during auto-sync, the server will
4682	** send the "pragma avoid-delta-manifests" statement in its reply,
4683	** which will cause the client to avoid generating a delta
4684	** manifest.
4685	*/
4686	/*
4687	** SETTING: forum-close-policy boolean default=off
4688	** If true, forum moderators may close/re-open forum posts, and reply
4689	** to closed posts. If false, only administrators may do so. Note that
4690	** this only affects the forum web UI, not post-closing tags which
4691	** arrive via the command-line or from synchronization with a remote.
4692	*/
4693	/*
4694	** SETTING: gdiff-command width=40 default=gdiff sensitive
4695	** The value is an external command to run when performing a graphical
4696	** diff. If undefined, text diff will be used.
4697	*/
	@@ -4841,15 +4836,19 @@
4841	** URL of the HTTP proxy. If undefined or "system", the "http_proxy"
4842	** environment variable is consulted. If "off", a direct HTTP connection is
4843	** used.
4844	*/
4845	/*
4846	** SETTING: redirect-to-https default=0 width=-1
4847	** Specifies whether or not to redirect http:// requests to
4848	** https:// URIs. A value of 0 (the default) means not to
4849	** redirect, 1 means to redirect only the /login page, and 2
4850	** means to always redirect.




4851	*/
4852	/*
4853	** SETTING: relative-paths boolean default=on
4854	** When showing changes and extras, report paths relative
4855	** to the current working directory.
	@@ -4967,11 +4966,11 @@
4967	** If enabled, special TH1 commands will be called before and
4968	** after any Fossil command or web page.
4969	*/
4970	#endif
4971	/*
4972	** SETTING: th1-setup width=40 block-text
4973	** This is the setup script to be evaluated after creating
4974	** and initializing the TH1 interpreter. By default, this
4975	** is empty and no extra setup is performed.
4976	*/
4977	/*
	@@ -5358,11 +5357,11 @@
5358	}
5359
5360	/*
5361	** Compute a "fingerprint" on the repository. A fingerprint is used
5362	** to verify that that the repository has not been replaced by a clone
5363	** of the same repository. More precisely, a fingerprint are used to
5364	** verify that the mapping between SHA3 hashes and RID values is unchanged.
5365	**
5366	** The check-out database ("localdb") stores RID values. When associating
5367	** a check-out database against a repository database, it is useful to verify
5368	** the fingerprint so that we know tha the RID values in the check-out
	@@ -5423,11 +5422,11 @@
5423	** COMMAND: test-fingerprint
5424	**
5425	** Usage: %fossil test-fingerprint ?RCVID?
5426	**
5427	** Display the repository fingerprint using the supplied RCVID or
5428	** using the latest RCVID if not is given on the command line.
5429	** Show both the legacy and the newer version of the fingerprint,
5430	** and the currently stored fingerprint if there is one.
5431	*/
5432	void test_fingerprint(void){
5433	int rcvid = 0;
5434

	--- src/db.c
	+++ src/db.c
	@@ -1557,10 +1557,12 @@
1557	sqlite3_create_function(db, "url_nouser", 1, SQLITE_UTF8, 0,
1558	url_nouser_func,0,0);
1559	sqlite3_create_function(db, "chat_msg_from_event", 4,
1560	SQLITE_UTF8 \| SQLITE_INNOCUOUS, 0,
1561	chat_msg_from_event, 0, 0);
1562	sqlite3_create_function(db, "inode", 1, SQLITE_UTF8, 0,
1563	file_inode_sql_func,0,0);
1564
1565	}
1566
1567	#if USE_SEE
1568	/*
	@@ -4080,11 +4082,11 @@
4082	** specified then that version is checked out. Otherwise the most recent
4083	** check-in on the main branch (usually "trunk") is used.
4084	**
4085	** REPOSITORY can be the filename for a repository that already exists on the
4086	** local machine or it can be a URI for a remote repository. If REPOSITORY
4087	** is a URI in one of the formats recognized by the [[clone]] command, the
4088	** remote repo is first cloned, then the clone is opened. The clone will be
4089	** stored in the current directory, or in DIR if the "--repodir DIR" option
4090	** is used. The name of the clone will be taken from the last term of the URI.
4091	** For "http:" and "https:" URIs, you can append an extra term to the end of
4092	** the URI to get any repository name you like. For example:
	@@ -4681,17 +4683,10 @@
4683	** to obtain a check-in lock during auto-sync, the server will
4684	** send the "pragma avoid-delta-manifests" statement in its reply,
4685	** which will cause the client to avoid generating a delta
4686	** manifest.
4687	*/







4688	/*
4689	** SETTING: gdiff-command width=40 default=gdiff sensitive
4690	** The value is an external command to run when performing a graphical
4691	** diff. If undefined, text diff will be used.
4692	*/
	@@ -4841,15 +4836,19 @@
4836	** URL of the HTTP proxy. If undefined or "system", the "http_proxy"
4837	** environment variable is consulted. If "off", a direct HTTP connection is
4838	** used.
4839	*/
4840	/*
4841	** SETTING: redirect-to-https default=0 width=2
4842	** Specifies whether or not to redirect unencrypted "http://" requests to
4843	** encrypted "https://" URIs. A value of 0 (the default) means do not
4844	** redirect, 1 means to redirect only the /login page, and 2
4845	** means to always redirect.
4846	**
4847	** For security, a value of 2 is recommended. The default value is 0
4848	** because not all sites are TLS-capable. But you should definitely enable
4849	** TLS and change this setting to 2 for all public-facing repositories.
4850	*/
4851	/*
4852	** SETTING: relative-paths boolean default=on
4853	** When showing changes and extras, report paths relative
4854	** to the current working directory.
	@@ -4967,11 +4966,11 @@
4966	** If enabled, special TH1 commands will be called before and
4967	** after any Fossil command or web page.
4968	*/
4969	#endif
4970	/*
4971	** SETTING: th1-setup width=40 block-text sensitive
4972	** This is the setup script to be evaluated after creating
4973	** and initializing the TH1 interpreter. By default, this
4974	** is empty and no extra setup is performed.
4975	*/
4976	/*
	@@ -5358,11 +5357,11 @@
5357	}
5358
5359	/*
5360	** Compute a "fingerprint" on the repository. A fingerprint is used
5361	** to verify that that the repository has not been replaced by a clone
5362	** of the same repository. More precisely, a fingerprint is used to
5363	** verify that the mapping between SHA3 hashes and RID values is unchanged.
5364	**
5365	** The check-out database ("localdb") stores RID values. When associating
5366	** a check-out database against a repository database, it is useful to verify
5367	** the fingerprint so that we know tha the RID values in the check-out
	@@ -5423,11 +5422,11 @@
5422	** COMMAND: test-fingerprint
5423	**
5424	** Usage: %fossil test-fingerprint ?RCVID?
5425	**
5426	** Display the repository fingerprint using the supplied RCVID or
5427	** using the latest RCVID if none is given on the command line.
5428	** Show both the legacy and the newer version of the fingerprint,
5429	** and the currently stored fingerprint if there is one.
5430	*/
5431	void test_fingerprint(void){
5432	int rcvid = 0;
5433

M src/default.css

+52 -10

		--- src/default.css
		+++ src/default.css
		@@ -762,10 +762,19 @@
762	762	border-bottom: 3px solid gold;
763	763	}
764	764	body.tkt div.content ol.tkt-changes > li:target > ol {
765	765	border-left: 1px solid gold;
766	766	}
	767	+body.cpage-info .file-change-line,
	768	+body.cpage-vdiff .file-change-line {
	769	+ margin-top: 16px;
	770	+ margin-bottom: 16px;
	771	+ margin-right: 1em /* keep it from nudging right up against the scrollbar-reveal zone */;
	772	+ display: flex;
	773	+ flex-direction: row;
	774	+ justify-content: space-between;
	775	+}
767	776
768	777	span.modpending {
769	778	color: #b03800;
770	779	font-style: italic;
771	780	}
		@@ -919,10 +928,17 @@
919	928	padding: 1px;
920	929	}
921	930	div.forum_body p {
922	931	margin-top: 0;
923	932	}
	933	+div.forum-editor-widget{
	934	+ display: flex;
	935	+ flex-direction: column;
	936	+}
	937	+div.forum-editor-widget > textarea {
	938	+ max-width: initial;
	939	+}
924	940	td.form_label {
925	941	vertical-align: top;
926	942	text-align: right;
927	943	}
928	944	.debug {
		@@ -982,10 +998,17 @@
982	998	** in the thread view. */
983	999	}
984	1000	.forum div > form {
985	1001	margin: 0.5em 0;
986	1002	display: inline-block;
	1003	+}
	1004	+body.cpage-forumedit div > form,
	1005	+body.cpage-forume2 div > form{
	1006	+ width: 100%;
	1007	+}
	1008	+.forum div > form > * {
	1009	+ margin-bottom: 0.35em;
987	1010	}
988	1011	.forum-post-collapser {
989	1012	/* Common style for the bottom-of-post and right-of-post
990	1013	expand/collapse widgets. */
991	1014	font-size: 0.8em;
		@@ -1681,22 +1704,45 @@
1681	1704	DOM structure:
1682	1705	<DIV.pikchr-wrapper>
1683	1706	<DIV.pikchr-svg>
1684	1707	<SVG.pikchr>...</SVG>
1685	1708	</DIV.pikchr-svg>
1686		- <PRE.pikchr-src>...</PRE>
	1709	+ <DIV.pikchr-src>
	1710	+ <PRE>pikchr source code</PRE>
	1711	+ <SPAN class='hidden'><A>link to open pikchr in /pikchrshow</A></SPAN>
	1712	+ <!-- ^^^ is unhidden and activated by JS code -->
	1713	+ </DIV.pikchr-src>
1687	1714	</DIV.pikchr-wrapper>
1688	1715
1689	1716	************************************************************/
1690	1717	div.pikchr-wrapper {/outer wrapper elem for a pikchr construct/}
1691	1718	div.pikchr-svg {/wrapper for SVG.pikchr element/}
1692	1719	svg.pikchr {/pikchr SVG/
1693	1720	width: 100%/necessary for SOME SVGs for Chrome!/;
1694	1721	}
1695		-pre.pikchr-src {/source code view for a pikchr (see fossil.pikchr.js)/
	1722	+
	1723	+div.pikchr-src {
	1724	+ /Wrapper for source code view of a pikchr (see fossil.pikchr.js)/
	1725	+ display: flex;
	1726	+ flex-direction: column;
	1727	+}
	1728	+div.pikchr-src > pre {
	1729	+ /Source code for a pikchr/
1696	1730	box-sizing: border-box;
1697	1731	text-align: left;
	1732	+}
	1733	+div.pikchr-src > span {
	1734	+ /Wrapper for a link to open a pikchr in /pikchrshow/
	1735	+ margin-top: 0.5em;
	1736	+ margin-bottom: 0.5em;
	1737	+ font-size: 85%;
	1738	+}
	1739	+div.pikchr-src > span::before {
	1740	+ content: "[";
	1741	+}
	1742	+div.pikchr-src > span::after {
	1743	+ content: "]";
1698	1744	}
1699	1745	/* The .source-inline class tells the .source class that the
1700	1746	source view, when enabled, should be "inline" (same position
1701	1747	as the graphic), else the sources are shifted to the left as
1702	1748	if they were "plain text". */
		@@ -1712,14 +1758,14 @@
1712	1758	still-seemingly-legitimate browsers don't support grid mode. */
1713	1759	}
1714	1760	div.pikchr-wrapper.center > div.pikchr-svg {
1715	1761	width: 100%/necessary for Chrome!/;
1716	1762	}
1717		-div.pikchr-wrapper.center:not(.source) > pre.pikchr-src,
	1763	+div.pikchr-wrapper.center:not(.source) > div.pikchr-src,
1718	1764	div.pikchr-wrapper.center:not(.source) > div.pikchr-svg,
1719	1765	/* ^^^ Centered non-source-view elements */
1720		-div.pikchr-wrapper.center.source.source-inline > pre.pikchr-src,
	1766	+div.pikchr-wrapper.center.source.source-inline div.pikchr-src,
1721	1767	div.pikchr-wrapper.center.source.source-inline > div.pikchr-svg
1722	1768	/* ^^^ Centered inline-source-view elements */{
1723	1769	display:inline-block/allows parent text-align to do the alignment/;
1724	1770	/* ^^^^ Browser incompatibility: inline-block causes the centered
1725	1771	pikchr to shrink to the point of illegiblity in Chrome. The
		@@ -1743,14 +1789,14 @@
1743	1789	padding: 4em;
1744	1790	}
1745	1791
1746	1792	/* For pikchr-wrapper.source mode, toggle pre.pikchr-src and
1747	1793	svg.pikchr visibility... */
1748		-div.pikchr-wrapper.source > pre.pikchr-src {
	1794	+div.pikchr-wrapper.source > div.pikchr-src {
1749	1795	/* Source code ^^^^^^^ is visible, else it is hidden */
1750	1796	}
1751		-div.pikchr-wrapper:not(.source) > pre.pikchr-src {
	1797	+div.pikchr-wrapper:not(.source) > div.pikchr-src {
1752	1798	/* Hide sources when image is being shown. */
1753	1799	position: absolute !important;
1754	1800	opacity: 0 !important;
1755	1801	pointer-events: none !important;
1756	1802	display: none !important;
		@@ -1795,14 +1841,10 @@
1795	1841	}
1796	1842	body.fossil-dark-style .settings-icon {
1797	1843	filter: invert(100%);
1798	1844	}
1799	1845
1800		-input[type="checkbox"].diff-toggle {
1801		- float: right;
1802		-}
1803		-
1804	1846	body.branch .brlist > table > tbody > tr:hover:not(.selected),
1805	1847	body.branch .brlist > table > tbody > tr.selected {
1806	1848	background-color: #ffc;
1807	1849	}
1808	1850	body.branch .brlist > table > tbody td:first-child > input {
1809	1851

	--- src/default.css
	+++ src/default.css
	@@ -762,10 +762,19 @@
762	border-bottom: 3px solid gold;
763	}
764	body.tkt div.content ol.tkt-changes > li:target > ol {
765	border-left: 1px solid gold;
766	}









767
768	span.modpending {
769	color: #b03800;
770	font-style: italic;
771	}
	@@ -919,10 +928,17 @@
919	padding: 1px;
920	}
921	div.forum_body p {
922	margin-top: 0;
923	}







924	td.form_label {
925	vertical-align: top;
926	text-align: right;
927	}
928	.debug {
	@@ -982,10 +998,17 @@
982	** in the thread view. */
983	}
984	.forum div > form {
985	margin: 0.5em 0;
986	display: inline-block;







987	}
988	.forum-post-collapser {
989	/* Common style for the bottom-of-post and right-of-post
990	expand/collapse widgets. */
991	font-size: 0.8em;
	@@ -1681,22 +1704,45 @@
1681	DOM structure:
1682	<DIV.pikchr-wrapper>
1683	<DIV.pikchr-svg>
1684	<SVG.pikchr>...</SVG>
1685	</DIV.pikchr-svg>
1686	<PRE.pikchr-src>...</PRE>




1687	</DIV.pikchr-wrapper>
1688
1689	************************************************************/
1690	div.pikchr-wrapper {/outer wrapper elem for a pikchr construct/}
1691	div.pikchr-svg {/wrapper for SVG.pikchr element/}
1692	svg.pikchr {/pikchr SVG/
1693	width: 100%/necessary for SOME SVGs for Chrome!/;
1694	}
1695	pre.pikchr-src {/source code view for a pikchr (see fossil.pikchr.js)/







1696	box-sizing: border-box;
1697	text-align: left;












1698	}
1699	/* The .source-inline class tells the .source class that the
1700	source view, when enabled, should be "inline" (same position
1701	as the graphic), else the sources are shifted to the left as
1702	if they were "plain text". */
	@@ -1712,14 +1758,14 @@
1712	still-seemingly-legitimate browsers don't support grid mode. */
1713	}
1714	div.pikchr-wrapper.center > div.pikchr-svg {
1715	width: 100%/necessary for Chrome!/;
1716	}
1717	div.pikchr-wrapper.center:not(.source) > pre.pikchr-src,
1718	div.pikchr-wrapper.center:not(.source) > div.pikchr-svg,
1719	/* ^^^ Centered non-source-view elements */
1720	div.pikchr-wrapper.center.source.source-inline > pre.pikchr-src,
1721	div.pikchr-wrapper.center.source.source-inline > div.pikchr-svg
1722	/* ^^^ Centered inline-source-view elements */{
1723	display:inline-block/allows parent text-align to do the alignment/;
1724	/* ^^^^ Browser incompatibility: inline-block causes the centered
1725	pikchr to shrink to the point of illegiblity in Chrome. The
	@@ -1743,14 +1789,14 @@
1743	padding: 4em;
1744	}
1745
1746	/* For pikchr-wrapper.source mode, toggle pre.pikchr-src and
1747	svg.pikchr visibility... */
1748	div.pikchr-wrapper.source > pre.pikchr-src {
1749	/* Source code ^^^^^^^ is visible, else it is hidden */
1750	}
1751	div.pikchr-wrapper:not(.source) > pre.pikchr-src {
1752	/* Hide sources when image is being shown. */
1753	position: absolute !important;
1754	opacity: 0 !important;
1755	pointer-events: none !important;
1756	display: none !important;
	@@ -1795,14 +1841,10 @@
1795	}
1796	body.fossil-dark-style .settings-icon {
1797	filter: invert(100%);
1798	}
1799
1800	input[type="checkbox"].diff-toggle {
1801	float: right;
1802	}
1803
1804	body.branch .brlist > table > tbody > tr:hover:not(.selected),
1805	body.branch .brlist > table > tbody > tr.selected {
1806	background-color: #ffc;
1807	}
1808	body.branch .brlist > table > tbody td:first-child > input {
1809

	--- src/default.css
	+++ src/default.css
	@@ -762,10 +762,19 @@
762	border-bottom: 3px solid gold;
763	}
764	body.tkt div.content ol.tkt-changes > li:target > ol {
765	border-left: 1px solid gold;
766	}
767	body.cpage-info .file-change-line,
768	body.cpage-vdiff .file-change-line {
769	margin-top: 16px;
770	margin-bottom: 16px;
771	margin-right: 1em /* keep it from nudging right up against the scrollbar-reveal zone */;
772	display: flex;
773	flex-direction: row;
774	justify-content: space-between;
775	}
776
777	span.modpending {
778	color: #b03800;
779	font-style: italic;
780	}
	@@ -919,10 +928,17 @@
928	padding: 1px;
929	}
930	div.forum_body p {
931	margin-top: 0;
932	}
933	div.forum-editor-widget{
934	display: flex;
935	flex-direction: column;
936	}
937	div.forum-editor-widget > textarea {
938	max-width: initial;
939	}
940	td.form_label {
941	vertical-align: top;
942	text-align: right;
943	}
944	.debug {
	@@ -982,10 +998,17 @@
998	** in the thread view. */
999	}
1000	.forum div > form {
1001	margin: 0.5em 0;
1002	display: inline-block;
1003	}
1004	body.cpage-forumedit div > form,
1005	body.cpage-forume2 div > form{
1006	width: 100%;
1007	}
1008	.forum div > form > * {
1009	margin-bottom: 0.35em;
1010	}
1011	.forum-post-collapser {
1012	/* Common style for the bottom-of-post and right-of-post
1013	expand/collapse widgets. */
1014	font-size: 0.8em;
	@@ -1681,22 +1704,45 @@
1704	DOM structure:
1705	<DIV.pikchr-wrapper>
1706	<DIV.pikchr-svg>
1707	<SVG.pikchr>...</SVG>
1708	</DIV.pikchr-svg>
1709	<DIV.pikchr-src>
1710	<PRE>pikchr source code</PRE>
1711	<SPAN class='hidden'><A>link to open pikchr in /pikchrshow</A></SPAN>
1712	<!-- ^^^ is unhidden and activated by JS code -->
1713	</DIV.pikchr-src>
1714	</DIV.pikchr-wrapper>
1715
1716	************************************************************/
1717	div.pikchr-wrapper {/outer wrapper elem for a pikchr construct/}
1718	div.pikchr-svg {/wrapper for SVG.pikchr element/}
1719	svg.pikchr {/pikchr SVG/
1720	width: 100%/necessary for SOME SVGs for Chrome!/;
1721	}
1722
1723	div.pikchr-src {
1724	/Wrapper for source code view of a pikchr (see fossil.pikchr.js)/
1725	display: flex;
1726	flex-direction: column;
1727	}
1728	div.pikchr-src > pre {
1729	/Source code for a pikchr/
1730	box-sizing: border-box;
1731	text-align: left;
1732	}
1733	div.pikchr-src > span {
1734	/Wrapper for a link to open a pikchr in /pikchrshow/
1735	margin-top: 0.5em;
1736	margin-bottom: 0.5em;
1737	font-size: 85%;
1738	}
1739	div.pikchr-src > span::before {
1740	content: "[";
1741	}
1742	div.pikchr-src > span::after {
1743	content: "]";
1744	}
1745	/* The .source-inline class tells the .source class that the
1746	source view, when enabled, should be "inline" (same position
1747	as the graphic), else the sources are shifted to the left as
1748	if they were "plain text". */
	@@ -1712,14 +1758,14 @@
1758	still-seemingly-legitimate browsers don't support grid mode. */
1759	}
1760	div.pikchr-wrapper.center > div.pikchr-svg {
1761	width: 100%/necessary for Chrome!/;
1762	}
1763	div.pikchr-wrapper.center:not(.source) > div.pikchr-src,
1764	div.pikchr-wrapper.center:not(.source) > div.pikchr-svg,
1765	/* ^^^ Centered non-source-view elements */
1766	div.pikchr-wrapper.center.source.source-inline div.pikchr-src,
1767	div.pikchr-wrapper.center.source.source-inline > div.pikchr-svg
1768	/* ^^^ Centered inline-source-view elements */{
1769	display:inline-block/allows parent text-align to do the alignment/;
1770	/* ^^^^ Browser incompatibility: inline-block causes the centered
1771	pikchr to shrink to the point of illegiblity in Chrome. The
	@@ -1743,14 +1789,14 @@
1789	padding: 4em;
1790	}
1791
1792	/* For pikchr-wrapper.source mode, toggle pre.pikchr-src and
1793	svg.pikchr visibility... */
1794	div.pikchr-wrapper.source > div.pikchr-src {
1795	/* Source code ^^^^^^^ is visible, else it is hidden */
1796	}
1797	div.pikchr-wrapper:not(.source) > div.pikchr-src {
1798	/* Hide sources when image is being shown. */
1799	position: absolute !important;
1800	opacity: 0 !important;
1801	pointer-events: none !important;
1802	display: none !important;
	@@ -1795,14 +1841,10 @@
1841	}
1842	body.fossil-dark-style .settings-icon {
1843	filter: invert(100%);
1844	}
1845




1846	body.branch .brlist > table > tbody > tr:hover:not(.selected),
1847	body.branch .brlist > table > tbody > tr.selected {
1848	background-color: #ffc;
1849	}
1850	body.branch .brlist > table > tbody td:first-child > input {
1851

M src/default.css

+52 -10

		--- src/default.css
		+++ src/default.css
		@@ -762,10 +762,19 @@
762	762	border-bottom: 3px solid gold;
763	763	}
764	764	body.tkt div.content ol.tkt-changes > li:target > ol {
765	765	border-left: 1px solid gold;
766	766	}
	767	+body.cpage-info .file-change-line,
	768	+body.cpage-vdiff .file-change-line {
	769	+ margin-top: 16px;
	770	+ margin-bottom: 16px;
	771	+ margin-right: 1em /* keep it from nudging right up against the scrollbar-reveal zone */;
	772	+ display: flex;
	773	+ flex-direction: row;
	774	+ justify-content: space-between;
	775	+}
767	776
768	777	span.modpending {
769	778	color: #b03800;
770	779	font-style: italic;
771	780	}
		@@ -919,10 +928,17 @@
919	928	padding: 1px;
920	929	}
921	930	div.forum_body p {
922	931	margin-top: 0;
923	932	}
	933	+div.forum-editor-widget{
	934	+ display: flex;
	935	+ flex-direction: column;
	936	+}
	937	+div.forum-editor-widget > textarea {
	938	+ max-width: initial;
	939	+}
924	940	td.form_label {
925	941	vertical-align: top;
926	942	text-align: right;
927	943	}
928	944	.debug {
		@@ -982,10 +998,17 @@
982	998	** in the thread view. */
983	999	}
984	1000	.forum div > form {
985	1001	margin: 0.5em 0;
986	1002	display: inline-block;
	1003	+}
	1004	+body.cpage-forumedit div > form,
	1005	+body.cpage-forume2 div > form{
	1006	+ width: 100%;
	1007	+}
	1008	+.forum div > form > * {
	1009	+ margin-bottom: 0.35em;
987	1010	}
988	1011	.forum-post-collapser {
989	1012	/* Common style for the bottom-of-post and right-of-post
990	1013	expand/collapse widgets. */
991	1014	font-size: 0.8em;
		@@ -1681,22 +1704,45 @@
1681	1704	DOM structure:
1682	1705	<DIV.pikchr-wrapper>
1683	1706	<DIV.pikchr-svg>
1684	1707	<SVG.pikchr>...</SVG>
1685	1708	</DIV.pikchr-svg>
1686		- <PRE.pikchr-src>...</PRE>
	1709	+ <DIV.pikchr-src>
	1710	+ <PRE>pikchr source code</PRE>
	1711	+ <SPAN class='hidden'><A>link to open pikchr in /pikchrshow</A></SPAN>
	1712	+ <!-- ^^^ is unhidden and activated by JS code -->
	1713	+ </DIV.pikchr-src>
1687	1714	</DIV.pikchr-wrapper>
1688	1715
1689	1716	************************************************************/
1690	1717	div.pikchr-wrapper {/outer wrapper elem for a pikchr construct/}
1691	1718	div.pikchr-svg {/wrapper for SVG.pikchr element/}
1692	1719	svg.pikchr {/pikchr SVG/
1693	1720	width: 100%/necessary for SOME SVGs for Chrome!/;
1694	1721	}
1695		-pre.pikchr-src {/source code view for a pikchr (see fossil.pikchr.js)/
	1722	+
	1723	+div.pikchr-src {
	1724	+ /Wrapper for source code view of a pikchr (see fossil.pikchr.js)/
	1725	+ display: flex;
	1726	+ flex-direction: column;
	1727	+}
	1728	+div.pikchr-src > pre {
	1729	+ /Source code for a pikchr/
1696	1730	box-sizing: border-box;
1697	1731	text-align: left;
	1732	+}
	1733	+div.pikchr-src > span {
	1734	+ /Wrapper for a link to open a pikchr in /pikchrshow/
	1735	+ margin-top: 0.5em;
	1736	+ margin-bottom: 0.5em;
	1737	+ font-size: 85%;
	1738	+}
	1739	+div.pikchr-src > span::before {
	1740	+ content: "[";
	1741	+}
	1742	+div.pikchr-src > span::after {
	1743	+ content: "]";
1698	1744	}
1699	1745	/* The .source-inline class tells the .source class that the
1700	1746	source view, when enabled, should be "inline" (same position
1701	1747	as the graphic), else the sources are shifted to the left as
1702	1748	if they were "plain text". */
		@@ -1712,14 +1758,14 @@
1712	1758	still-seemingly-legitimate browsers don't support grid mode. */
1713	1759	}
1714	1760	div.pikchr-wrapper.center > div.pikchr-svg {
1715	1761	width: 100%/necessary for Chrome!/;
1716	1762	}
1717		-div.pikchr-wrapper.center:not(.source) > pre.pikchr-src,
	1763	+div.pikchr-wrapper.center:not(.source) > div.pikchr-src,
1718	1764	div.pikchr-wrapper.center:not(.source) > div.pikchr-svg,
1719	1765	/* ^^^ Centered non-source-view elements */
1720		-div.pikchr-wrapper.center.source.source-inline > pre.pikchr-src,
	1766	+div.pikchr-wrapper.center.source.source-inline div.pikchr-src,
1721	1767	div.pikchr-wrapper.center.source.source-inline > div.pikchr-svg
1722	1768	/* ^^^ Centered inline-source-view elements */{
1723	1769	display:inline-block/allows parent text-align to do the alignment/;
1724	1770	/* ^^^^ Browser incompatibility: inline-block causes the centered
1725	1771	pikchr to shrink to the point of illegiblity in Chrome. The
		@@ -1743,14 +1789,14 @@
1743	1789	padding: 4em;
1744	1790	}
1745	1791
1746	1792	/* For pikchr-wrapper.source mode, toggle pre.pikchr-src and
1747	1793	svg.pikchr visibility... */
1748		-div.pikchr-wrapper.source > pre.pikchr-src {
	1794	+div.pikchr-wrapper.source > div.pikchr-src {
1749	1795	/* Source code ^^^^^^^ is visible, else it is hidden */
1750	1796	}
1751		-div.pikchr-wrapper:not(.source) > pre.pikchr-src {
	1797	+div.pikchr-wrapper:not(.source) > div.pikchr-src {
1752	1798	/* Hide sources when image is being shown. */
1753	1799	position: absolute !important;
1754	1800	opacity: 0 !important;
1755	1801	pointer-events: none !important;
1756	1802	display: none !important;
		@@ -1795,14 +1841,10 @@
1795	1841	}
1796	1842	body.fossil-dark-style .settings-icon {
1797	1843	filter: invert(100%);
1798	1844	}
1799	1845
1800		-input[type="checkbox"].diff-toggle {
1801		- float: right;
1802		-}
1803		-
1804	1846	body.branch .brlist > table > tbody > tr:hover:not(.selected),
1805	1847	body.branch .brlist > table > tbody > tr.selected {
1806	1848	background-color: #ffc;
1807	1849	}
1808	1850	body.branch .brlist > table > tbody td:first-child > input {
1809	1851

	--- src/default.css
	+++ src/default.css
	@@ -762,10 +762,19 @@
762	border-bottom: 3px solid gold;
763	}
764	body.tkt div.content ol.tkt-changes > li:target > ol {
765	border-left: 1px solid gold;
766	}









767
768	span.modpending {
769	color: #b03800;
770	font-style: italic;
771	}
	@@ -919,10 +928,17 @@
919	padding: 1px;
920	}
921	div.forum_body p {
922	margin-top: 0;
923	}







924	td.form_label {
925	vertical-align: top;
926	text-align: right;
927	}
928	.debug {
	@@ -982,10 +998,17 @@
982	** in the thread view. */
983	}
984	.forum div > form {
985	margin: 0.5em 0;
986	display: inline-block;







987	}
988	.forum-post-collapser {
989	/* Common style for the bottom-of-post and right-of-post
990	expand/collapse widgets. */
991	font-size: 0.8em;
	@@ -1681,22 +1704,45 @@
1681	DOM structure:
1682	<DIV.pikchr-wrapper>
1683	<DIV.pikchr-svg>
1684	<SVG.pikchr>...</SVG>
1685	</DIV.pikchr-svg>
1686	<PRE.pikchr-src>...</PRE>




1687	</DIV.pikchr-wrapper>
1688
1689	************************************************************/
1690	div.pikchr-wrapper {/outer wrapper elem for a pikchr construct/}
1691	div.pikchr-svg {/wrapper for SVG.pikchr element/}
1692	svg.pikchr {/pikchr SVG/
1693	width: 100%/necessary for SOME SVGs for Chrome!/;
1694	}
1695	pre.pikchr-src {/source code view for a pikchr (see fossil.pikchr.js)/







1696	box-sizing: border-box;
1697	text-align: left;












1698	}
1699	/* The .source-inline class tells the .source class that the
1700	source view, when enabled, should be "inline" (same position
1701	as the graphic), else the sources are shifted to the left as
1702	if they were "plain text". */
	@@ -1712,14 +1758,14 @@
1712	still-seemingly-legitimate browsers don't support grid mode. */
1713	}
1714	div.pikchr-wrapper.center > div.pikchr-svg {
1715	width: 100%/necessary for Chrome!/;
1716	}
1717	div.pikchr-wrapper.center:not(.source) > pre.pikchr-src,
1718	div.pikchr-wrapper.center:not(.source) > div.pikchr-svg,
1719	/* ^^^ Centered non-source-view elements */
1720	div.pikchr-wrapper.center.source.source-inline > pre.pikchr-src,
1721	div.pikchr-wrapper.center.source.source-inline > div.pikchr-svg
1722	/* ^^^ Centered inline-source-view elements */{
1723	display:inline-block/allows parent text-align to do the alignment/;
1724	/* ^^^^ Browser incompatibility: inline-block causes the centered
1725	pikchr to shrink to the point of illegiblity in Chrome. The
	@@ -1743,14 +1789,14 @@
1743	padding: 4em;
1744	}
1745
1746	/* For pikchr-wrapper.source mode, toggle pre.pikchr-src and
1747	svg.pikchr visibility... */
1748	div.pikchr-wrapper.source > pre.pikchr-src {
1749	/* Source code ^^^^^^^ is visible, else it is hidden */
1750	}
1751	div.pikchr-wrapper:not(.source) > pre.pikchr-src {
1752	/* Hide sources when image is being shown. */
1753	position: absolute !important;
1754	opacity: 0 !important;
1755	pointer-events: none !important;
1756	display: none !important;
	@@ -1795,14 +1841,10 @@
1795	}
1796	body.fossil-dark-style .settings-icon {
1797	filter: invert(100%);
1798	}
1799
1800	input[type="checkbox"].diff-toggle {
1801	float: right;
1802	}
1803
1804	body.branch .brlist > table > tbody > tr:hover:not(.selected),
1805	body.branch .brlist > table > tbody > tr.selected {
1806	background-color: #ffc;
1807	}
1808	body.branch .brlist > table > tbody td:first-child > input {
1809

	--- src/default.css
	+++ src/default.css
	@@ -762,10 +762,19 @@
762	border-bottom: 3px solid gold;
763	}
764	body.tkt div.content ol.tkt-changes > li:target > ol {
765	border-left: 1px solid gold;
766	}
767	body.cpage-info .file-change-line,
768	body.cpage-vdiff .file-change-line {
769	margin-top: 16px;
770	margin-bottom: 16px;
771	margin-right: 1em /* keep it from nudging right up against the scrollbar-reveal zone */;
772	display: flex;
773	flex-direction: row;
774	justify-content: space-between;
775	}
776
777	span.modpending {
778	color: #b03800;
779	font-style: italic;
780	}
	@@ -919,10 +928,17 @@
928	padding: 1px;
929	}
930	div.forum_body p {
931	margin-top: 0;
932	}
933	div.forum-editor-widget{
934	display: flex;
935	flex-direction: column;
936	}
937	div.forum-editor-widget > textarea {
938	max-width: initial;
939	}
940	td.form_label {
941	vertical-align: top;
942	text-align: right;
943	}
944	.debug {
	@@ -982,10 +998,17 @@
998	** in the thread view. */
999	}
1000	.forum div > form {
1001	margin: 0.5em 0;
1002	display: inline-block;
1003	}
1004	body.cpage-forumedit div > form,
1005	body.cpage-forume2 div > form{
1006	width: 100%;
1007	}
1008	.forum div > form > * {
1009	margin-bottom: 0.35em;
1010	}
1011	.forum-post-collapser {
1012	/* Common style for the bottom-of-post and right-of-post
1013	expand/collapse widgets. */
1014	font-size: 0.8em;
	@@ -1681,22 +1704,45 @@
1704	DOM structure:
1705	<DIV.pikchr-wrapper>
1706	<DIV.pikchr-svg>
1707	<SVG.pikchr>...</SVG>
1708	</DIV.pikchr-svg>
1709	<DIV.pikchr-src>
1710	<PRE>pikchr source code</PRE>
1711	<SPAN class='hidden'><A>link to open pikchr in /pikchrshow</A></SPAN>
1712	<!-- ^^^ is unhidden and activated by JS code -->
1713	</DIV.pikchr-src>
1714	</DIV.pikchr-wrapper>
1715
1716	************************************************************/
1717	div.pikchr-wrapper {/outer wrapper elem for a pikchr construct/}
1718	div.pikchr-svg {/wrapper for SVG.pikchr element/}
1719	svg.pikchr {/pikchr SVG/
1720	width: 100%/necessary for SOME SVGs for Chrome!/;
1721	}
1722
1723	div.pikchr-src {
1724	/Wrapper for source code view of a pikchr (see fossil.pikchr.js)/
1725	display: flex;
1726	flex-direction: column;
1727	}
1728	div.pikchr-src > pre {
1729	/Source code for a pikchr/
1730	box-sizing: border-box;
1731	text-align: left;
1732	}
1733	div.pikchr-src > span {
1734	/Wrapper for a link to open a pikchr in /pikchrshow/
1735	margin-top: 0.5em;
1736	margin-bottom: 0.5em;
1737	font-size: 85%;
1738	}
1739	div.pikchr-src > span::before {
1740	content: "[";
1741	}
1742	div.pikchr-src > span::after {
1743	content: "]";
1744	}
1745	/* The .source-inline class tells the .source class that the
1746	source view, when enabled, should be "inline" (same position
1747	as the graphic), else the sources are shifted to the left as
1748	if they were "plain text". */
	@@ -1712,14 +1758,14 @@
1758	still-seemingly-legitimate browsers don't support grid mode. */
1759	}
1760	div.pikchr-wrapper.center > div.pikchr-svg {
1761	width: 100%/necessary for Chrome!/;
1762	}
1763	div.pikchr-wrapper.center:not(.source) > div.pikchr-src,
1764	div.pikchr-wrapper.center:not(.source) > div.pikchr-svg,
1765	/* ^^^ Centered non-source-view elements */
1766	div.pikchr-wrapper.center.source.source-inline div.pikchr-src,
1767	div.pikchr-wrapper.center.source.source-inline > div.pikchr-svg
1768	/* ^^^ Centered inline-source-view elements */{
1769	display:inline-block/allows parent text-align to do the alignment/;
1770	/* ^^^^ Browser incompatibility: inline-block causes the centered
1771	pikchr to shrink to the point of illegiblity in Chrome. The
	@@ -1743,14 +1789,14 @@
1789	padding: 4em;
1790	}
1791
1792	/* For pikchr-wrapper.source mode, toggle pre.pikchr-src and
1793	svg.pikchr visibility... */
1794	div.pikchr-wrapper.source > div.pikchr-src {
1795	/* Source code ^^^^^^^ is visible, else it is hidden */
1796	}
1797	div.pikchr-wrapper:not(.source) > div.pikchr-src {
1798	/* Hide sources when image is being shown. */
1799	position: absolute !important;
1800	opacity: 0 !important;
1801	pointer-events: none !important;
1802	display: none !important;
	@@ -1795,14 +1841,10 @@
1841	}
1842	body.fossil-dark-style .settings-icon {
1843	filter: invert(100%);
1844	}
1845




1846	body.branch .brlist > table > tbody > tr:hover:not(.selected),
1847	body.branch .brlist > table > tbody > tr.selected {
1848	background-color: #ffc;
1849	}
1850	body.branch .brlist > table > tbody td:first-child > input {
1851

M src/delta.c

+50 -48

		--- src/delta.c
		+++ src/delta.c
		@@ -225,59 +225,61 @@
225	225	** of four bytes.
226	226	*/
227	227	static unsigned int checksum(const char *zIn, size_t N){
228	228	static const int byteOrderTest = 1;
229	229	const unsigned char z = (const unsigned char )zIn;
230		- const unsigned char zEnd = (const unsigned char)&zIn[N&~3];
231	230	unsigned sum = 0;
232		- assert( (z - (const unsigned char)0)%4==0 ); / Four-byte alignment */
233		- if( 0==(char)&byteOrderTest ){
234		- /* This is a big-endian machine */
235		- while( z<zEnd ){
236		- sum += (unsigned)z;
237		- z += 4;
238		- }
239		- }else{
240		- /* A little-endian machine */
241		-#if GCC_VERSION>=4003000
242		- while( z<zEnd ){
243		- sum += __builtin_bswap32((unsigned)z);
244		- z += 4;
245		- }
246		-#elif defined(_MSC_VER) && _MSC_VER>=1300
247		- while( z<zEnd ){
248		- sum += _byteswap_ulong((unsigned)z);
249		- z += 4;
250		- }
251		-#else
252		- unsigned sum0 = 0;
253		- unsigned sum1 = 0;
254		- unsigned sum2 = 0;
255		- while(N >= 16){
256		- sum0 += ((unsigned)z[0] + z[4] + z[8] + z[12]);
257		- sum1 += ((unsigned)z[1] + z[5] + z[9] + z[13]);
258		- sum2 += ((unsigned)z[2] + z[6] + z[10]+ z[14]);
259		- sum += ((unsigned)z[3] + z[7] + z[11]+ z[15]);
260		- z += 16;
261		- N -= 16;
262		- }
263		- while(N >= 4){
264		- sum0 += z[0];
265		- sum1 += z[1];
266		- sum2 += z[2];
267		- sum += z[3];
268		- z += 4;
269		- N -= 4;
270		- }
271		- sum += (sum2 << 8) + (sum1 << 16) + (sum0 << 24);
272		-#endif
273		- }
274		- switch(N&3){
275		- case 3: sum += (z[2] << 8);
276		- case 2: sum += (z[1] << 16);
277		- case 1: sum += (z[0] << 24);
278		- default: ;
	231	+ if( N>0 ){
	232	+ const unsigned char zEnd = (const unsigned char)&zIn[N&~3];
	233	+ assert( (z - (const unsigned char)0)%4==0 ); / Four-byte alignment */
	234	+ if( 0==(char)&byteOrderTest ){
	235	+ /* This is a big-endian machine */
	236	+ while( z<zEnd ){
	237	+ sum += (unsigned)z;
	238	+ z += 4;
	239	+ }
	240	+ }else{
	241	+ /* A little-endian machine */
	242	+ #if GCC_VERSION>=4003000
	243	+ while( z<zEnd ){
	244	+ sum += __builtin_bswap32((unsigned)z);
	245	+ z += 4;
	246	+ }
	247	+ #elif defined(_MSC_VER) && _MSC_VER>=1300
	248	+ while( z<zEnd ){
	249	+ sum += _byteswap_ulong((unsigned)z);
	250	+ z += 4;
	251	+ }
	252	+ #else
	253	+ unsigned sum0 = 0;
	254	+ unsigned sum1 = 0;
	255	+ unsigned sum2 = 0;
	256	+ while(N >= 16){
	257	+ sum0 += ((unsigned)z[0] + z[4] + z[8] + z[12]);
	258	+ sum1 += ((unsigned)z[1] + z[5] + z[9] + z[13]);
	259	+ sum2 += ((unsigned)z[2] + z[6] + z[10]+ z[14]);
	260	+ sum += ((unsigned)z[3] + z[7] + z[11]+ z[15]);
	261	+ z += 16;
	262	+ N -= 16;
	263	+ }
	264	+ while(N >= 4){
	265	+ sum0 += z[0];
	266	+ sum1 += z[1];
	267	+ sum2 += z[2];
	268	+ sum += z[3];
	269	+ z += 4;
	270	+ N -= 4;
	271	+ }
	272	+ sum += (sum2 << 8) + (sum1 << 16) + (sum0 << 24);
	273	+ #endif
	274	+ }
	275	+ switch(N&3){
	276	+ case 3: sum += (z[2] << 8);
	277	+ case 2: sum += (z[1] << 16);
	278	+ case 1: sum += (z[0] << 24);
	279	+ default: ;
	280	+ }
279	281	}
280	282	return sum;
281	283	}
282	284
283	285	/*
284	286

	--- src/delta.c
	+++ src/delta.c
	@@ -225,59 +225,61 @@
225	** of four bytes.
226	*/
227	static unsigned int checksum(const char *zIn, size_t N){
228	static const int byteOrderTest = 1;
229	const unsigned char z = (const unsigned char )zIn;
230	const unsigned char zEnd = (const unsigned char)&zIn[N&~3];
231	unsigned sum = 0;
232	assert( (z - (const unsigned char)0)%4==0 ); / Four-byte alignment */
233	if( 0==(char)&byteOrderTest ){
234	/* This is a big-endian machine */
235	while( z<zEnd ){
236	sum += (unsigned)z;
237	z += 4;
238	}
239	}else{
240	/* A little-endian machine */
241	#if GCC_VERSION>=4003000
242	while( z<zEnd ){
243	sum += __builtin_bswap32((unsigned)z);
244	z += 4;
245	}
246	#elif defined(_MSC_VER) && _MSC_VER>=1300
247	while( z<zEnd ){
248	sum += _byteswap_ulong((unsigned)z);
249	z += 4;
250	}
251	#else
252	unsigned sum0 = 0;
253	unsigned sum1 = 0;
254	unsigned sum2 = 0;
255	while(N >= 16){
256	sum0 += ((unsigned)z[0] + z[4] + z[8] + z[12]);
257	sum1 += ((unsigned)z[1] + z[5] + z[9] + z[13]);
258	sum2 += ((unsigned)z[2] + z[6] + z[10]+ z[14]);
259	sum += ((unsigned)z[3] + z[7] + z[11]+ z[15]);
260	z += 16;
261	N -= 16;
262	}
263	while(N >= 4){
264	sum0 += z[0];
265	sum1 += z[1];
266	sum2 += z[2];
267	sum += z[3];
268	z += 4;
269	N -= 4;
270	}
271	sum += (sum2 << 8) + (sum1 << 16) + (sum0 << 24);
272	#endif
273	}
274	switch(N&3){
275	case 3: sum += (z[2] << 8);
276	case 2: sum += (z[1] << 16);
277	case 1: sum += (z[0] << 24);
278	default: ;



279	}
280	return sum;
281	}
282
283	/*
284

	--- src/delta.c
	+++ src/delta.c
	@@ -225,59 +225,61 @@
225	** of four bytes.
226	*/
227	static unsigned int checksum(const char *zIn, size_t N){
228	static const int byteOrderTest = 1;
229	const unsigned char z = (const unsigned char )zIn;

230	unsigned sum = 0;
231	if( N>0 ){
232	const unsigned char zEnd = (const unsigned char)&zIn[N&~3];
233	assert( (z - (const unsigned char)0)%4==0 ); / Four-byte alignment */
234	if( 0==(char)&byteOrderTest ){
235	/* This is a big-endian machine */
236	while( z<zEnd ){
237	sum += (unsigned)z;
238	z += 4;
239	}
240	}else{
241	/* A little-endian machine */
242	#if GCC_VERSION>=4003000
243	while( z<zEnd ){
244	sum += __builtin_bswap32((unsigned)z);
245	z += 4;
246	}
247	#elif defined(_MSC_VER) && _MSC_VER>=1300
248	while( z<zEnd ){
249	sum += _byteswap_ulong((unsigned)z);
250	z += 4;
251	}
252	#else
253	unsigned sum0 = 0;
254	unsigned sum1 = 0;
255	unsigned sum2 = 0;
256	while(N >= 16){
257	sum0 += ((unsigned)z[0] + z[4] + z[8] + z[12]);
258	sum1 += ((unsigned)z[1] + z[5] + z[9] + z[13]);
259	sum2 += ((unsigned)z[2] + z[6] + z[10]+ z[14]);
260	sum += ((unsigned)z[3] + z[7] + z[11]+ z[15]);
261	z += 16;
262	N -= 16;
263	}
264	while(N >= 4){
265	sum0 += z[0];
266	sum1 += z[1];
267	sum2 += z[2];
268	sum += z[3];
269	z += 4;
270	N -= 4;
271	}
272	sum += (sum2 << 8) + (sum1 << 16) + (sum0 << 24);
273	#endif
274	}
275	switch(N&3){
276	case 3: sum += (z[2] << 8);
277	case 2: sum += (z[1] << 16);
278	case 1: sum += (z[0] << 24);
279	default: ;
280	}
281	}
282	return sum;
283	}
284
285	/*
286

M src/diff.c

+146 -7

		--- src/diff.c
		+++ src/diff.c
		@@ -50,10 +50,11 @@
50	50	#define DIFF_RAW 0x00040000 /* Raw triples - for debugging */
51	51	#define DIFF_TCL 0x00080000 /* For the --tk option */
52	52	#define DIFF_INCBINARY 0x00100000 /* The --diff-binary option */
53	53	#define DIFF_SHOW_VERS 0x00200000 /* Show compared versions */
54	54	#define DIFF_DARKMODE 0x00400000 /* Use dark mode for HTML */
	55	+#define DIFF_BY_TOKEN 0x01000000 /* Split on tokens, not lines */
55	56
56	57	/*
57	58	** Per file information that may influence output.
58	59	*/
59	60	#define DIFF_FILE_ADDED 0x40000000 /* Added or rename destination */
		@@ -319,10 +320,113 @@
319	320
320	321	/* Return results */
321	322	*pnLine = nLine;
322	323	return a;
323	324	}
	325	+
	326	+/*
	327	+** Character classes for the purpose of tokenization.
	328	+**
	329	+** 1 - alphanumeric
	330	+** 2 - whitespace
	331	+** 3 - punctuation
	332	+*/
	333	+static char aTCharClass[256] = {
	334	+ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
	335	+ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
	336	+ 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
	337	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3,
	338	+
	339	+ 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	340	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3,
	341	+ 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	342	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3,
	343	+
	344	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	345	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	346	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	347	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	348	+
	349	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	350	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	351	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	352	+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
	353	+};
	354	+
	355	+/*
	356	+** Count the number of tokens in the given string.
	357	+*/
	358	+static int count_tokens(const unsigned char *p, int n){
	359	+ int nToken = 0;
	360	+ int iPrev = 0;
	361	+ int i;
	362	+ for(i=0; i<n; i++){
	363	+ char x = aTCharClass[p[i]];
	364	+ if( x!=iPrev ){
	365	+ iPrev = x;
	366	+ nToken++;
	367	+ }
	368	+ }
	369	+ return nToken;
	370	+}
	371	+
	372	+/*
	373	+** Return an array of DLine objects containing a pointer to the
	374	+** start of each token and a hash of that token. The lower
	375	+** bits of the hash store the length of each token.
	376	+**
	377	+** This is like break_into_lines() except that it works with tokens
	378	+** instead of lines. A token is:
	379	+**
	380	+** * A contiguous sequence of alphanumeric characters.
	381	+** * A contiguous sequence of whitespace
	382	+** * A contiguous sequence of punctuation characters.
	383	+**
	384	+** Return 0 if the file is binary or contains a line that is
	385	+** too long.
	386	+*/
	387	+static DLine *break_into_tokens(
	388	+ const char *z,
	389	+ int n,
	390	+ int *pnToken,
	391	+ u64 diffFlags
	392	+){
	393	+ int nToken, i, k;
	394	+ u64 h, h2;
	395	+ DLine *a;
	396	+ unsigned char p = (unsigned char)z;
	397	+
	398	+ nToken = count_tokens(p, n);
	399	+ a = fossil_malloc( sizeof(a[0])*(nToken+1) );
	400	+ memset(a, 0, sizeof(a[0])*(nToken+1));
	401	+ if( n==0 ){
	402	+ *pnToken = 0;
	403	+ return a;
	404	+ }
	405	+ i = 0;
	406	+ while( n>0 ){
	407	+ char x = aTCharClass[*p];
	408	+ h = 0xcbf29ce484222325LL;
	409	+ for(k=1; k<n && aTCharClass[p[k]]==x; k++){
	410	+ h ^= p[k];
	411	+ h *= 0x100000001b3LL;
	412	+ }
	413	+ a[i].z = (char*)p;
	414	+ a[i].n = k;
	415	+ a[i].h = h = ((h%281474976710597LL)<<LENGTH_MASK_SZ) \| k;
	416	+ h2 = h % nToken;
	417	+ a[i].iNext = a[h2].iHash;
	418	+ a[h2].iHash = i+1;
	419	+ p += k; n -= k;
	420	+ i++;
	421	+ };
	422	+ assert( i==nToken );
	423	+
	424	+ /* Return results */
	425	+ *pnToken = nToken;
	426	+ return a;
	427	+}
324	428
325	429	/*
326	430	** Return zero if two DLine elements are identical.
327	431	*/
328	432	static int compare_dline(const DLine pA, const DLine pB){
		@@ -2462,11 +2566,11 @@
2462	2566	int span; /* combined width of the input sequences */
2463	2567	int cutoff = 4; /* Max hash chain entries to follow */
2464	2568	int nextCutoff = -1; /* Value of cutoff for next iteration */
2465	2569
2466	2570	span = (iE1 - iS1) + (iE2 - iS2);
2467		- bestScore = -10000;
	2571	+ bestScore = -9223300000*(sqlite3_int64)1000000000;
2468	2572	score = 0;
2469	2573	iSXb = iSXp = iS1;
2470	2574	iEXb = iEXp = iS1;
2471	2575	iSYb = iSYp = iS2;
2472	2576	iEYb = iEYp = iS2;
		@@ -2997,14 +3101,21 @@
2997	3101	if( (pCfg->diffFlags & DIFF_IGNORE_ALLWS)==DIFF_IGNORE_ALLWS ){
2998	3102	c.xDiffer = compare_dline_ignore_allws;
2999	3103	}else{
3000	3104	c.xDiffer = compare_dline;
3001	3105	}
3002		- c.aFrom = break_into_lines(blob_str(pA_Blob), blob_size(pA_Blob),
3003		- &c.nFrom, pCfg->diffFlags);
3004		- c.aTo = break_into_lines(blob_str(pB_Blob), blob_size(pB_Blob),
3005		- &c.nTo, pCfg->diffFlags);
	3106	+ if( pCfg->diffFlags & DIFF_BY_TOKEN ){
	3107	+ c.aFrom = break_into_tokens(blob_str(pA_Blob), blob_size(pA_Blob),
	3108	+ &c.nFrom, pCfg->diffFlags);
	3109	+ c.aTo = break_into_tokens(blob_str(pB_Blob), blob_size(pB_Blob),
	3110	+ &c.nTo, pCfg->diffFlags);
	3111	+ }else{
	3112	+ c.aFrom = break_into_lines(blob_str(pA_Blob), blob_size(pA_Blob),
	3113	+ &c.nFrom, pCfg->diffFlags);
	3114	+ c.aTo = break_into_lines(blob_str(pB_Blob), blob_size(pB_Blob),
	3115	+ &c.nTo, pCfg->diffFlags);
	3116	+ }
3006	3117	if( c.aFrom==0 \|\| c.aTo==0 ){
3007	3118	fossil_free(c.aFrom);
3008	3119	fossil_free(c.aTo);
3009	3120	if( pOut ){
3010	3121	diff_errmsg(pOut, DIFF_CANNOT_COMPUTE_BINARY, pCfg->diffFlags);
		@@ -3035,10 +3146,26 @@
3035	3146	}
3036	3147	}
3037	3148	if( (pCfg->diffFlags & DIFF_NOOPT)==0 ){
3038	3149	diff_optimize(&c);
3039	3150	}
	3151	+ if( (pCfg->diffFlags & DIFF_BY_TOKEN)!=0 ){
	3152	+ /* Convert token counts into byte counts. */
	3153	+ int i;
	3154	+ int iA = 0;
	3155	+ int iB = 0;
	3156	+ for(i=0; c.aEdit[i] \|\| c.aEdit[i+1] \|\| c.aEdit[i+2]; i+=3){
	3157	+ int k, sum;
	3158	+ for(k=0, sum=0; k<c.aEdit[i]; k++) sum += c.aFrom[iA++].n;
	3159	+ iB += c.aEdit[i];
	3160	+ c.aEdit[i] = sum;
	3161	+ for(k=0, sum=0; k<c.aEdit[i+1]; k++) sum += c.aFrom[iA++].n;
	3162	+ c.aEdit[i+1] = sum;
	3163	+ for(k=0, sum=0; k<c.aEdit[i+2]; k++) sum += c.aTo[iB++].n;
	3164	+ c.aEdit[i+2] = sum;
	3165	+ }
	3166	+ }
3040	3167
3041	3168	if( pOut ){
3042	3169	if( pCfg->diffFlags & DIFF_NUMSTAT ){
3043	3170	int nDel = 0, nIns = 0, i;
3044	3171	for(i=0; c.aEdit[i] \|\| c.aEdit[i+1] \|\| c.aEdit[i+2]; i+=3){
		@@ -3049,11 +3176,11 @@
3049	3176	g.diffCnt[2] += nDel;
3050	3177	if( nIns+nDel ){
3051	3178	g.diffCnt[0]++;
3052	3179	blob_appendf(pOut, "%10d %10d", nIns, nDel);
3053	3180	}
3054		- }else if( pCfg->diffFlags & DIFF_RAW ){
	3181	+ }else if( pCfg->diffFlags & (DIFF_RAW\|DIFF_BY_TOKEN) ){
3055	3182	const int *R = c.aEdit;
3056	3183	unsigned int r;
3057	3184	for(r=0; R[r] \|\| R[r+1] \|\| R[r+2]; r += 3){
3058	3185	blob_appendf(pOut, " copy %6d delete %6d insert %6d\n",
3059	3186	R[r], R[r+1], R[r+2]);
		@@ -3100,20 +3227,29 @@
3100	3227	** Initialize the DiffConfig object using command-line options.
3101	3228	**
3102	3229	** Process diff-related command-line options and return an appropriate
3103	3230	** "diffFlags" integer.
3104	3231	**
	3232	+** -b\|--browser Show the diff output in a web-browser
3105	3233	** --brief Show filenames only DIFF_BRIEF
	3234	+** --by Shorthand for "--browser -y"
3106	3235	** -c\|--context N N lines of context. nContext
	3236	+** --dark Use dark mode for Tcl/Tk and HTML output
3107	3237	** --html Format for HTML DIFF_HTML
	3238	+** -i\|--internal Use built-in diff, not an external tool
3108	3239	** --invert Invert the diff DIFF_INVERT
	3240	+** --json Output formatted as JSON
3109	3241	** -n\|--linenum Show line numbers DIFF_LINENO
	3242	+** -N\|--new-file Alias for --verbose
3110	3243	** --noopt Disable optimization DIFF_NOOPT
3111	3244	** --numstat Show change counts DIFF_NUMSTAT
3112	3245	** --strip-trailing-cr Strip trailing CR DIFF_STRIP_EOLCR
	3246	+** --tcl Tcl-formatted output used internally by --tk
3113	3247	** --unified Unified diff. ~DIFF_SIDEBYSIDE
	3248	+** -v\|--verbose Show complete text of added or deleted files
3114	3249	** -w\|--ignore-all-space Ignore all whitespaces DIFF_IGNORE_ALLWS
	3250	+** --webpage Format output as a stand-alone HTML webpage
3115	3251	** -W\|--width N N character lines. wColumn
3116	3252	** -y\|--side-by-side Side-by-side diff. DIFF_SIDEBYSIDE
3117	3253	** -Z\|--ignore-trailing-space Ignore eol-whitespaces DIFF_IGNORE_EOLWS
3118	3254	*/
3119	3255	void diff_options(DiffConfig *pCfg, int isGDiff, int bUnifiedTextOnly){
		@@ -3157,10 +3293,13 @@
3157	3293
3158	3294	/* Undocumented and unsupported flags used for development
3159	3295	** debugging and analysis: */
3160	3296	if( find_option("debug",0,0)!=0 ) diffFlags \|= DIFF_DEBUG;
3161	3297	if( find_option("raw",0,0)!=0 ) diffFlags \|= DIFF_RAW;
	3298	+ if( find_option("bytoken",0,0)!=0 ){
	3299	+ diffFlags = DIFF_RAW\|DIFF_BY_TOKEN;
	3300	+ }
3162	3301	}
3163	3302	if( (z = find_option("context","c",1))!=0 ){
3164	3303	char *zEnd;
3165	3304	f = (int)strtol(z, &zEnd, 10);
3166	3305	if( zEnd[0]==0 && errno!=ERANGE ){
		@@ -3771,11 +3910,11 @@
3771	3910	int szHash; /* Display size of a version hash */
3772	3911	Blob treename; /* Name of file to be annotated */
3773	3912	char zFilename; / Name of file to be annotated */
3774	3913
3775	3914	bBlame = g.argv[1][0]!='a';
3776		- zRevision = find_option("r","revision",1);
	3915	+ zRevision = find_option("revision","r",1);
3777	3916	zLimit = find_option("limit","n",1);
3778	3917	zOrig = find_option("origin","o",1);
3779	3918	showLog = find_option("log","l",0)!=0;
3780	3919	if( find_option("ignore-trailing-space","Z",0)!=0 ){
3781	3920	annFlags = DIFF_IGNORE_EOLWS;
3782	3921

	--- src/diff.c
	+++ src/diff.c
	@@ -50,10 +50,11 @@
50	#define DIFF_RAW 0x00040000 /* Raw triples - for debugging */
51	#define DIFF_TCL 0x00080000 /* For the --tk option */
52	#define DIFF_INCBINARY 0x00100000 /* The --diff-binary option */
53	#define DIFF_SHOW_VERS 0x00200000 /* Show compared versions */
54	#define DIFF_DARKMODE 0x00400000 /* Use dark mode for HTML */

55
56	/*
57	** Per file information that may influence output.
58	*/
59	#define DIFF_FILE_ADDED 0x40000000 /* Added or rename destination */
	@@ -319,10 +320,113 @@
319
320	/* Return results */
321	*pnLine = nLine;
322	return a;
323	}







































































































324
325	/*
326	** Return zero if two DLine elements are identical.
327	*/
328	static int compare_dline(const DLine pA, const DLine pB){
	@@ -2462,11 +2566,11 @@
2462	int span; /* combined width of the input sequences */
2463	int cutoff = 4; /* Max hash chain entries to follow */
2464	int nextCutoff = -1; /* Value of cutoff for next iteration */
2465
2466	span = (iE1 - iS1) + (iE2 - iS2);
2467	bestScore = -10000;
2468	score = 0;
2469	iSXb = iSXp = iS1;
2470	iEXb = iEXp = iS1;
2471	iSYb = iSYp = iS2;
2472	iEYb = iEYp = iS2;
	@@ -2997,14 +3101,21 @@
2997	if( (pCfg->diffFlags & DIFF_IGNORE_ALLWS)==DIFF_IGNORE_ALLWS ){
2998	c.xDiffer = compare_dline_ignore_allws;
2999	}else{
3000	c.xDiffer = compare_dline;
3001	}
3002	c.aFrom = break_into_lines(blob_str(pA_Blob), blob_size(pA_Blob),
3003	&c.nFrom, pCfg->diffFlags);
3004	c.aTo = break_into_lines(blob_str(pB_Blob), blob_size(pB_Blob),
3005	&c.nTo, pCfg->diffFlags);







3006	if( c.aFrom==0 \|\| c.aTo==0 ){
3007	fossil_free(c.aFrom);
3008	fossil_free(c.aTo);
3009	if( pOut ){
3010	diff_errmsg(pOut, DIFF_CANNOT_COMPUTE_BINARY, pCfg->diffFlags);
	@@ -3035,10 +3146,26 @@
3035	}
3036	}
3037	if( (pCfg->diffFlags & DIFF_NOOPT)==0 ){
3038	diff_optimize(&c);
3039	}
















3040
3041	if( pOut ){
3042	if( pCfg->diffFlags & DIFF_NUMSTAT ){
3043	int nDel = 0, nIns = 0, i;
3044	for(i=0; c.aEdit[i] \|\| c.aEdit[i+1] \|\| c.aEdit[i+2]; i+=3){
	@@ -3049,11 +3176,11 @@
3049	g.diffCnt[2] += nDel;
3050	if( nIns+nDel ){
3051	g.diffCnt[0]++;
3052	blob_appendf(pOut, "%10d %10d", nIns, nDel);
3053	}
3054	}else if( pCfg->diffFlags & DIFF_RAW ){
3055	const int *R = c.aEdit;
3056	unsigned int r;
3057	for(r=0; R[r] \|\| R[r+1] \|\| R[r+2]; r += 3){
3058	blob_appendf(pOut, " copy %6d delete %6d insert %6d\n",
3059	R[r], R[r+1], R[r+2]);
	@@ -3100,20 +3227,29 @@
3100	** Initialize the DiffConfig object using command-line options.
3101	**
3102	** Process diff-related command-line options and return an appropriate
3103	** "diffFlags" integer.
3104	**

3105	** --brief Show filenames only DIFF_BRIEF

3106	** -c\|--context N N lines of context. nContext

3107	** --html Format for HTML DIFF_HTML

3108	** --invert Invert the diff DIFF_INVERT

3109	** -n\|--linenum Show line numbers DIFF_LINENO

3110	** --noopt Disable optimization DIFF_NOOPT
3111	** --numstat Show change counts DIFF_NUMSTAT
3112	** --strip-trailing-cr Strip trailing CR DIFF_STRIP_EOLCR

3113	** --unified Unified diff. ~DIFF_SIDEBYSIDE

3114	** -w\|--ignore-all-space Ignore all whitespaces DIFF_IGNORE_ALLWS

3115	** -W\|--width N N character lines. wColumn
3116	** -y\|--side-by-side Side-by-side diff. DIFF_SIDEBYSIDE
3117	** -Z\|--ignore-trailing-space Ignore eol-whitespaces DIFF_IGNORE_EOLWS
3118	*/
3119	void diff_options(DiffConfig *pCfg, int isGDiff, int bUnifiedTextOnly){
	@@ -3157,10 +3293,13 @@
3157
3158	/* Undocumented and unsupported flags used for development
3159	** debugging and analysis: */
3160	if( find_option("debug",0,0)!=0 ) diffFlags \|= DIFF_DEBUG;
3161	if( find_option("raw",0,0)!=0 ) diffFlags \|= DIFF_RAW;



3162	}
3163	if( (z = find_option("context","c",1))!=0 ){
3164	char *zEnd;
3165	f = (int)strtol(z, &zEnd, 10);
3166	if( zEnd[0]==0 && errno!=ERANGE ){
	@@ -3771,11 +3910,11 @@
3771	int szHash; /* Display size of a version hash */
3772	Blob treename; /* Name of file to be annotated */
3773	char zFilename; / Name of file to be annotated */
3774
3775	bBlame = g.argv[1][0]!='a';
3776	zRevision = find_option("r","revision",1);
3777	zLimit = find_option("limit","n",1);
3778	zOrig = find_option("origin","o",1);
3779	showLog = find_option("log","l",0)!=0;
3780	if( find_option("ignore-trailing-space","Z",0)!=0 ){
3781	annFlags = DIFF_IGNORE_EOLWS;
3782

	--- src/diff.c
	+++ src/diff.c
	@@ -50,10 +50,11 @@
50	#define DIFF_RAW 0x00040000 /* Raw triples - for debugging */
51	#define DIFF_TCL 0x00080000 /* For the --tk option */
52	#define DIFF_INCBINARY 0x00100000 /* The --diff-binary option */
53	#define DIFF_SHOW_VERS 0x00200000 /* Show compared versions */
54	#define DIFF_DARKMODE 0x00400000 /* Use dark mode for HTML */
55	#define DIFF_BY_TOKEN 0x01000000 /* Split on tokens, not lines */
56
57	/*
58	** Per file information that may influence output.
59	*/
60	#define DIFF_FILE_ADDED 0x40000000 /* Added or rename destination */
	@@ -319,10 +320,113 @@
320
321	/* Return results */
322	*pnLine = nLine;
323	return a;
324	}
325
326	/*
327	** Character classes for the purpose of tokenization.
328	**
329	** 1 - alphanumeric
330	** 2 - whitespace
331	** 3 - punctuation
332	*/
333	static char aTCharClass[256] = {
334	2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
335	2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
336	2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
337	1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3,
338
339	3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
340	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3,
341	3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
342	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3,
343
344	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
345	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
346	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
347	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
348
349	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
350	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
351	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
352	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
353	};
354
355	/*
356	** Count the number of tokens in the given string.
357	*/
358	static int count_tokens(const unsigned char *p, int n){
359	int nToken = 0;
360	int iPrev = 0;
361	int i;
362	for(i=0; i<n; i++){
363	char x = aTCharClass[p[i]];
364	if( x!=iPrev ){
365	iPrev = x;
366	nToken++;
367	}
368	}
369	return nToken;
370	}
371
372	/*
373	** Return an array of DLine objects containing a pointer to the
374	** start of each token and a hash of that token. The lower
375	** bits of the hash store the length of each token.
376	**
377	** This is like break_into_lines() except that it works with tokens
378	** instead of lines. A token is:
379	**
380	** * A contiguous sequence of alphanumeric characters.
381	** * A contiguous sequence of whitespace
382	** * A contiguous sequence of punctuation characters.
383	**
384	** Return 0 if the file is binary or contains a line that is
385	** too long.
386	*/
387	static DLine *break_into_tokens(
388	const char *z,
389	int n,
390	int *pnToken,
391	u64 diffFlags
392	){
393	int nToken, i, k;
394	u64 h, h2;
395	DLine *a;
396	unsigned char p = (unsigned char)z;
397
398	nToken = count_tokens(p, n);
399	a = fossil_malloc( sizeof(a[0])*(nToken+1) );
400	memset(a, 0, sizeof(a[0])*(nToken+1));
401	if( n==0 ){
402	*pnToken = 0;
403	return a;
404	}
405	i = 0;
406	while( n>0 ){
407	char x = aTCharClass[*p];
408	h = 0xcbf29ce484222325LL;
409	for(k=1; k<n && aTCharClass[p[k]]==x; k++){
410	h ^= p[k];
411	h *= 0x100000001b3LL;
412	}
413	a[i].z = (char*)p;
414	a[i].n = k;
415	a[i].h = h = ((h%281474976710597LL)<<LENGTH_MASK_SZ) \| k;
416	h2 = h % nToken;
417	a[i].iNext = a[h2].iHash;
418	a[h2].iHash = i+1;
419	p += k; n -= k;
420	i++;
421	};
422	assert( i==nToken );
423
424	/* Return results */
425	*pnToken = nToken;
426	return a;
427	}
428
429	/*
430	** Return zero if two DLine elements are identical.
431	*/
432	static int compare_dline(const DLine pA, const DLine pB){
	@@ -2462,11 +2566,11 @@
2566	int span; /* combined width of the input sequences */
2567	int cutoff = 4; /* Max hash chain entries to follow */
2568	int nextCutoff = -1; /* Value of cutoff for next iteration */
2569
2570	span = (iE1 - iS1) + (iE2 - iS2);
2571	bestScore = -9223300000*(sqlite3_int64)1000000000;
2572	score = 0;
2573	iSXb = iSXp = iS1;
2574	iEXb = iEXp = iS1;
2575	iSYb = iSYp = iS2;
2576	iEYb = iEYp = iS2;
	@@ -2997,14 +3101,21 @@
3101	if( (pCfg->diffFlags & DIFF_IGNORE_ALLWS)==DIFF_IGNORE_ALLWS ){
3102	c.xDiffer = compare_dline_ignore_allws;
3103	}else{
3104	c.xDiffer = compare_dline;
3105	}
3106	if( pCfg->diffFlags & DIFF_BY_TOKEN ){
3107	c.aFrom = break_into_tokens(blob_str(pA_Blob), blob_size(pA_Blob),
3108	&c.nFrom, pCfg->diffFlags);
3109	c.aTo = break_into_tokens(blob_str(pB_Blob), blob_size(pB_Blob),
3110	&c.nTo, pCfg->diffFlags);
3111	}else{
3112	c.aFrom = break_into_lines(blob_str(pA_Blob), blob_size(pA_Blob),
3113	&c.nFrom, pCfg->diffFlags);
3114	c.aTo = break_into_lines(blob_str(pB_Blob), blob_size(pB_Blob),
3115	&c.nTo, pCfg->diffFlags);
3116	}
3117	if( c.aFrom==0 \|\| c.aTo==0 ){
3118	fossil_free(c.aFrom);
3119	fossil_free(c.aTo);
3120	if( pOut ){
3121	diff_errmsg(pOut, DIFF_CANNOT_COMPUTE_BINARY, pCfg->diffFlags);
	@@ -3035,10 +3146,26 @@
3146	}
3147	}
3148	if( (pCfg->diffFlags & DIFF_NOOPT)==0 ){
3149	diff_optimize(&c);
3150	}
3151	if( (pCfg->diffFlags & DIFF_BY_TOKEN)!=0 ){
3152	/* Convert token counts into byte counts. */
3153	int i;
3154	int iA = 0;
3155	int iB = 0;
3156	for(i=0; c.aEdit[i] \|\| c.aEdit[i+1] \|\| c.aEdit[i+2]; i+=3){
3157	int k, sum;
3158	for(k=0, sum=0; k<c.aEdit[i]; k++) sum += c.aFrom[iA++].n;
3159	iB += c.aEdit[i];
3160	c.aEdit[i] = sum;
3161	for(k=0, sum=0; k<c.aEdit[i+1]; k++) sum += c.aFrom[iA++].n;
3162	c.aEdit[i+1] = sum;
3163	for(k=0, sum=0; k<c.aEdit[i+2]; k++) sum += c.aTo[iB++].n;
3164	c.aEdit[i+2] = sum;
3165	}
3166	}
3167
3168	if( pOut ){
3169	if( pCfg->diffFlags & DIFF_NUMSTAT ){
3170	int nDel = 0, nIns = 0, i;
3171	for(i=0; c.aEdit[i] \|\| c.aEdit[i+1] \|\| c.aEdit[i+2]; i+=3){
	@@ -3049,11 +3176,11 @@
3176	g.diffCnt[2] += nDel;
3177	if( nIns+nDel ){
3178	g.diffCnt[0]++;
3179	blob_appendf(pOut, "%10d %10d", nIns, nDel);
3180	}
3181	}else if( pCfg->diffFlags & (DIFF_RAW\|DIFF_BY_TOKEN) ){
3182	const int *R = c.aEdit;
3183	unsigned int r;
3184	for(r=0; R[r] \|\| R[r+1] \|\| R[r+2]; r += 3){
3185	blob_appendf(pOut, " copy %6d delete %6d insert %6d\n",
3186	R[r], R[r+1], R[r+2]);
	@@ -3100,20 +3227,29 @@
3227	** Initialize the DiffConfig object using command-line options.
3228	**
3229	** Process diff-related command-line options and return an appropriate
3230	** "diffFlags" integer.
3231	**
3232	** -b\|--browser Show the diff output in a web-browser
3233	** --brief Show filenames only DIFF_BRIEF
3234	** --by Shorthand for "--browser -y"
3235	** -c\|--context N N lines of context. nContext
3236	** --dark Use dark mode for Tcl/Tk and HTML output
3237	** --html Format for HTML DIFF_HTML
3238	** -i\|--internal Use built-in diff, not an external tool
3239	** --invert Invert the diff DIFF_INVERT
3240	** --json Output formatted as JSON
3241	** -n\|--linenum Show line numbers DIFF_LINENO
3242	** -N\|--new-file Alias for --verbose
3243	** --noopt Disable optimization DIFF_NOOPT
3244	** --numstat Show change counts DIFF_NUMSTAT
3245	** --strip-trailing-cr Strip trailing CR DIFF_STRIP_EOLCR
3246	** --tcl Tcl-formatted output used internally by --tk
3247	** --unified Unified diff. ~DIFF_SIDEBYSIDE
3248	** -v\|--verbose Show complete text of added or deleted files
3249	** -w\|--ignore-all-space Ignore all whitespaces DIFF_IGNORE_ALLWS
3250	** --webpage Format output as a stand-alone HTML webpage
3251	** -W\|--width N N character lines. wColumn
3252	** -y\|--side-by-side Side-by-side diff. DIFF_SIDEBYSIDE
3253	** -Z\|--ignore-trailing-space Ignore eol-whitespaces DIFF_IGNORE_EOLWS
3254	*/
3255	void diff_options(DiffConfig *pCfg, int isGDiff, int bUnifiedTextOnly){
	@@ -3157,10 +3293,13 @@
3293
3294	/* Undocumented and unsupported flags used for development
3295	** debugging and analysis: */
3296	if( find_option("debug",0,0)!=0 ) diffFlags \|= DIFF_DEBUG;
3297	if( find_option("raw",0,0)!=0 ) diffFlags \|= DIFF_RAW;
3298	if( find_option("bytoken",0,0)!=0 ){
3299	diffFlags = DIFF_RAW\|DIFF_BY_TOKEN;
3300	}
3301	}
3302	if( (z = find_option("context","c",1))!=0 ){
3303	char *zEnd;
3304	f = (int)strtol(z, &zEnd, 10);
3305	if( zEnd[0]==0 && errno!=ERANGE ){
	@@ -3771,11 +3910,11 @@
3910	int szHash; /* Display size of a version hash */
3911	Blob treename; /* Name of file to be annotated */
3912	char zFilename; / Name of file to be annotated */
3913
3914	bBlame = g.argv[1][0]!='a';
3915	zRevision = find_option("revision","r",1);
3916	zLimit = find_option("limit","n",1);
3917	zOrig = find_option("origin","o",1);
3918	showLog = find_option("log","l",0)!=0;
3919	if( find_option("ignore-trailing-space","Z",0)!=0 ){
3920	annFlags = DIFF_IGNORE_EOLWS;
3921

M src/diff.tcl

+1 -1

		--- src/diff.tcl
		+++ src/diff.tcl
		@@ -1,11 +1,11 @@
1	1	# The "diff --tk" command outputs prepends a "set fossilcmd {...}" line
2	2	# to this file, then runs this file using "tclsh" in order to display the
3	3	# graphical diff in a separate window. A typical "set fossilcmd" line
4	4	# looks like this:
5	5	#
6		-# set fossilcmd {\| "./fossil" diff --html -y -i -v}
	6	+# set fossilcmd {\| "./fossil" diff --tcl -i -v}
7	7	#
8	8	# This header comment is stripped off by the "mkbuiltin.c" program.
9	9	#
10	10	set prog {
11	11	package require Tk
12	12

	--- src/diff.tcl
	+++ src/diff.tcl
	@@ -1,11 +1,11 @@
1	# The "diff --tk" command outputs prepends a "set fossilcmd {...}" line
2	# to this file, then runs this file using "tclsh" in order to display the
3	# graphical diff in a separate window. A typical "set fossilcmd" line
4	# looks like this:
5	#
6	# set fossilcmd {\| "./fossil" diff --html -y -i -v}
7	#
8	# This header comment is stripped off by the "mkbuiltin.c" program.
9	#
10	set prog {
11	package require Tk
12

	--- src/diff.tcl
	+++ src/diff.tcl
	@@ -1,11 +1,11 @@
1	# The "diff --tk" command outputs prepends a "set fossilcmd {...}" line
2	# to this file, then runs this file using "tclsh" in order to display the
3	# graphical diff in a separate window. A typical "set fossilcmd" line
4	# looks like this:
5	#
6	# set fossilcmd {\| "./fossil" diff --tcl -i -v}
7	#
8	# This header comment is stripped off by the "mkbuiltin.c" program.
9	#
10	set prog {
11	package require Tk
12

M src/diffcmd.c

+8 -8

		--- src/diffcmd.c
		+++ src/diffcmd.c
		@@ -249,19 +249,19 @@
249	249	@ margin: 0 0 0 0;
250	250	@ line-height: inherit;
251	251	@ font-size: inherit;
252	252	@ }
253	253	@ td.diffln {
254		-@ width: 1px;
	254	+@ width: fit-content;
255	255	@ text-align: right;
256	256	@ padding: 0 1em 0 0;
257	257	@ }
258	258	@ td.difflne {
259	259	@ padding-bottom: 0.4em;
260	260	@ }
261	261	@ td.diffsep {
262		-@ width: 1px;
	262	+@ width: fit-content;
263	263	@ padding: 0 0.3em 0 1em;
264	264	@ line-height: inherit;
265	265	@ font-size: inherit;
266	266	@ }
267	267	@ td.diffsep pre {
		@@ -379,19 +379,19 @@
379	379	@ margin: 0 0 0 0;
380	380	@ line-height: inherit;
381	381	@ font-size: inherit;
382	382	@ }
383	383	@ td.diffln {
384		-@ width: 1px;
	384	+@ width: fit-content;
385	385	@ text-align: right;
386	386	@ padding: 0 1em 0 0;
387	387	@ }
388	388	@ td.difflne {
389	389	@ padding-bottom: 0.4em;
390	390	@ }
391	391	@ td.diffsep {
392		-@ width: 1px;
	392	+@ width: fit-content;
393	393	@ padding: 0 0.3em 0 1em;
394	394	@ line-height: inherit;
395	395	@ font-size: inherit;
396	396	@ }
397	397	@ td.diffsep pre {
		@@ -1266,11 +1266,11 @@
1266	1266	** -n\|--linenum Show line numbers
1267	1267	** -N\|--new-file Alias for --verbose
1268	1268	** --numstat Show only the number of added and deleted lines
1269	1269	** -y\|--side-by-side Side-by-side diff
1270	1270	** --strip-trailing-cr Strip trailing CR
1271		-** --tcl Tcl-formated output used internally by --tk
	1271	+** --tcl Tcl-formatted output used internally by --tk
1272	1272	** --tclsh PATH Tcl/Tk shell used for --tk (default: "tclsh")
1273	1273	** --tk Launch a Tcl/Tk GUI for display
1274	1274	** --to VERSION Select VERSION as target for the diff
1275	1275	** --undo Diff against the "undo" buffer
1276	1276	** --unified Unified diff
		@@ -1312,11 +1312,12 @@
1312	1312	zFrom = mprintf("root:%s", zBranch);
1313	1313	}
1314	1314	if( zCheckin!=0 && ( zFrom!=0 \|\| zTo!=0 ) ){
1315	1315	fossil_fatal("cannot use --checkin together with --from or --to");
1316	1316	}
1317		- g.diffCnt[0] = g.diffCnt[1] = g.diffCnt[2] = 0;
	1317	+ diff_options(&DCfg, isGDiff, 0);
	1318	+ determine_exec_relative_option(1);
1318	1319	if( 0==zCheckin ){
1319	1320	if( zTo==0 \|\| againstUndo ){
1320	1321	db_must_be_within_tree();
1321	1322	}else if( zFrom==0 ){
1322	1323	fossil_fatal("must use --from if --to is present");
		@@ -1324,13 +1325,12 @@
1324	1325	db_find_and_open_repository(0, 0);
1325	1326	}
1326	1327	}else{
1327	1328	db_find_and_open_repository(0, 0);
1328	1329	}
1329		- diff_options(&DCfg, isGDiff, 0);
1330		- determine_exec_relative_option(1);
1331	1330	verify_all_options();
	1331	+ g.diffCnt[0] = g.diffCnt[1] = g.diffCnt[2] = 0;
1332	1332	if( g.argc>=3 ){
1333	1333	int i;
1334	1334	Blob fname;
1335	1335	pFileDir = fossil_malloc( sizeof(pFileDir) (g.argc-1) );
1336	1336	memset(pFileDir, 0, sizeof(pFileDir) (g.argc-1));
1337	1337

	--- src/diffcmd.c
	+++ src/diffcmd.c
	@@ -249,19 +249,19 @@
249	@ margin: 0 0 0 0;
250	@ line-height: inherit;
251	@ font-size: inherit;
252	@ }
253	@ td.diffln {
254	@ width: 1px;
255	@ text-align: right;
256	@ padding: 0 1em 0 0;
257	@ }
258	@ td.difflne {
259	@ padding-bottom: 0.4em;
260	@ }
261	@ td.diffsep {
262	@ width: 1px;
263	@ padding: 0 0.3em 0 1em;
264	@ line-height: inherit;
265	@ font-size: inherit;
266	@ }
267	@ td.diffsep pre {
	@@ -379,19 +379,19 @@
379	@ margin: 0 0 0 0;
380	@ line-height: inherit;
381	@ font-size: inherit;
382	@ }
383	@ td.diffln {
384	@ width: 1px;
385	@ text-align: right;
386	@ padding: 0 1em 0 0;
387	@ }
388	@ td.difflne {
389	@ padding-bottom: 0.4em;
390	@ }
391	@ td.diffsep {
392	@ width: 1px;
393	@ padding: 0 0.3em 0 1em;
394	@ line-height: inherit;
395	@ font-size: inherit;
396	@ }
397	@ td.diffsep pre {
	@@ -1266,11 +1266,11 @@
1266	** -n\|--linenum Show line numbers
1267	** -N\|--new-file Alias for --verbose
1268	** --numstat Show only the number of added and deleted lines
1269	** -y\|--side-by-side Side-by-side diff
1270	** --strip-trailing-cr Strip trailing CR
1271	** --tcl Tcl-formated output used internally by --tk
1272	** --tclsh PATH Tcl/Tk shell used for --tk (default: "tclsh")
1273	** --tk Launch a Tcl/Tk GUI for display
1274	** --to VERSION Select VERSION as target for the diff
1275	** --undo Diff against the "undo" buffer
1276	** --unified Unified diff
	@@ -1312,11 +1312,12 @@
1312	zFrom = mprintf("root:%s", zBranch);
1313	}
1314	if( zCheckin!=0 && ( zFrom!=0 \|\| zTo!=0 ) ){
1315	fossil_fatal("cannot use --checkin together with --from or --to");
1316	}
1317	g.diffCnt[0] = g.diffCnt[1] = g.diffCnt[2] = 0;

1318	if( 0==zCheckin ){
1319	if( zTo==0 \|\| againstUndo ){
1320	db_must_be_within_tree();
1321	}else if( zFrom==0 ){
1322	fossil_fatal("must use --from if --to is present");
	@@ -1324,13 +1325,12 @@
1324	db_find_and_open_repository(0, 0);
1325	}
1326	}else{
1327	db_find_and_open_repository(0, 0);
1328	}
1329	diff_options(&DCfg, isGDiff, 0);
1330	determine_exec_relative_option(1);
1331	verify_all_options();

1332	if( g.argc>=3 ){
1333	int i;
1334	Blob fname;
1335	pFileDir = fossil_malloc( sizeof(pFileDir) (g.argc-1) );
1336	memset(pFileDir, 0, sizeof(pFileDir) (g.argc-1));
1337

	--- src/diffcmd.c
	+++ src/diffcmd.c
	@@ -249,19 +249,19 @@
249	@ margin: 0 0 0 0;
250	@ line-height: inherit;
251	@ font-size: inherit;
252	@ }
253	@ td.diffln {
254	@ width: fit-content;
255	@ text-align: right;
256	@ padding: 0 1em 0 0;
257	@ }
258	@ td.difflne {
259	@ padding-bottom: 0.4em;
260	@ }
261	@ td.diffsep {
262	@ width: fit-content;
263	@ padding: 0 0.3em 0 1em;
264	@ line-height: inherit;
265	@ font-size: inherit;
266	@ }
267	@ td.diffsep pre {
	@@ -379,19 +379,19 @@
379	@ margin: 0 0 0 0;
380	@ line-height: inherit;
381	@ font-size: inherit;
382	@ }
383	@ td.diffln {
384	@ width: fit-content;
385	@ text-align: right;
386	@ padding: 0 1em 0 0;
387	@ }
388	@ td.difflne {
389	@ padding-bottom: 0.4em;
390	@ }
391	@ td.diffsep {
392	@ width: fit-content;
393	@ padding: 0 0.3em 0 1em;
394	@ line-height: inherit;
395	@ font-size: inherit;
396	@ }
397	@ td.diffsep pre {
	@@ -1266,11 +1266,11 @@
1266	** -n\|--linenum Show line numbers
1267	** -N\|--new-file Alias for --verbose
1268	** --numstat Show only the number of added and deleted lines
1269	** -y\|--side-by-side Side-by-side diff
1270	** --strip-trailing-cr Strip trailing CR
1271	** --tcl Tcl-formatted output used internally by --tk
1272	** --tclsh PATH Tcl/Tk shell used for --tk (default: "tclsh")
1273	** --tk Launch a Tcl/Tk GUI for display
1274	** --to VERSION Select VERSION as target for the diff
1275	** --undo Diff against the "undo" buffer
1276	** --unified Unified diff
	@@ -1312,11 +1312,12 @@
1312	zFrom = mprintf("root:%s", zBranch);
1313	}
1314	if( zCheckin!=0 && ( zFrom!=0 \|\| zTo!=0 ) ){
1315	fossil_fatal("cannot use --checkin together with --from or --to");
1316	}
1317	diff_options(&DCfg, isGDiff, 0);
1318	determine_exec_relative_option(1);
1319	if( 0==zCheckin ){
1320	if( zTo==0 \|\| againstUndo ){
1321	db_must_be_within_tree();
1322	}else if( zFrom==0 ){
1323	fossil_fatal("must use --from if --to is present");
	@@ -1324,13 +1325,12 @@
1325	db_find_and_open_repository(0, 0);
1326	}
1327	}else{
1328	db_find_and_open_repository(0, 0);
1329	}


1330	verify_all_options();
1331	g.diffCnt[0] = g.diffCnt[1] = g.diffCnt[2] = 0;
1332	if( g.argc>=3 ){
1333	int i;
1334	Blob fname;
1335	pFileDir = fossil_malloc( sizeof(pFileDir) (g.argc-1) );
1336	memset(pFileDir, 0, sizeof(pFileDir) (g.argc-1));
1337

M src/dispatch.c

+26 -1

		--- src/dispatch.c
		+++ src/dispatch.c
		@@ -1307,17 +1307,42 @@
1307	1307	}
1308	1308
1309	1309	/*
1310	1310	** Return a pointer to the setting information array.
1311	1311	**
1312		-** This routine provides access to the aSetting2[] array which is created
	1312	+** This routine provides access to the aSetting[] array which is created
1313	1313	** by the mkindex utility program and included with <page_index.h>.
1314	1314	*/
1315	1315	const Setting setting_info(int pnCount){
1316	1316	if( pnCount ) *pnCount = (int)(sizeof(aSetting)/sizeof(aSetting[0])) - 1;
1317	1317	return aSetting;
1318	1318	}
	1319	+
	1320	+/*
	1321	+** Return a pointer to a specific Setting entry for the setting named
	1322	+** in the argument. Or return NULL if no such setting exists.
	1323	+**
	1324	+** The pointer returned points into the middle of the global aSetting[]
	1325	+** array that is generated by mkindex. Use setting_info() to fetch the
	1326	+** whole array. Use this routine to fetch a specific entry.
	1327	+*/
	1328	+const Setting setting_find(const char zName){
	1329	+ int iFirst = 0;
	1330	+ int iLast = ArraySize(aSetting)-1;
	1331	+ while( iFirst<=iLast ){
	1332	+ int iCur = (iFirst+iLast)/2;
	1333	+ int c = strcmp(aSetting[iCur].name, zName);
	1334	+ if( c<0 ){
	1335	+ iFirst = iCur+1;
	1336	+ }else if( c>0 ){
	1337	+ iLast = iCur-1;
	1338	+ }else{
	1339	+ return &aSetting[iCur];
	1340	+ }
	1341	+ }
	1342	+ return 0;
	1343	+}
1319	1344
1320	1345	/*****************************************************************************
1321	1346	** A virtual table for accessing the information in aCommand[], and
1322	1347	** especially the help-text
1323	1348	*/
1324	1349

	--- src/dispatch.c
	+++ src/dispatch.c
	@@ -1307,17 +1307,42 @@
1307	}
1308
1309	/*
1310	** Return a pointer to the setting information array.
1311	**
1312	** This routine provides access to the aSetting2[] array which is created
1313	** by the mkindex utility program and included with <page_index.h>.
1314	*/
1315	const Setting setting_info(int pnCount){
1316	if( pnCount ) *pnCount = (int)(sizeof(aSetting)/sizeof(aSetting[0])) - 1;
1317	return aSetting;
1318	}

























1319
1320	/*****************************************************************************
1321	** A virtual table for accessing the information in aCommand[], and
1322	** especially the help-text
1323	*/
1324

	--- src/dispatch.c
	+++ src/dispatch.c
	@@ -1307,17 +1307,42 @@
1307	}
1308
1309	/*
1310	** Return a pointer to the setting information array.
1311	**
1312	** This routine provides access to the aSetting[] array which is created
1313	** by the mkindex utility program and included with <page_index.h>.
1314	*/
1315	const Setting setting_info(int pnCount){
1316	if( pnCount ) *pnCount = (int)(sizeof(aSetting)/sizeof(aSetting[0])) - 1;
1317	return aSetting;
1318	}
1319
1320	/*
1321	** Return a pointer to a specific Setting entry for the setting named
1322	** in the argument. Or return NULL if no such setting exists.
1323	**
1324	** The pointer returned points into the middle of the global aSetting[]
1325	** array that is generated by mkindex. Use setting_info() to fetch the
1326	** whole array. Use this routine to fetch a specific entry.
1327	*/
1328	const Setting setting_find(const char zName){
1329	int iFirst = 0;
1330	int iLast = ArraySize(aSetting)-1;
1331	while( iFirst<=iLast ){
1332	int iCur = (iFirst+iLast)/2;
1333	int c = strcmp(aSetting[iCur].name, zName);
1334	if( c<0 ){
1335	iFirst = iCur+1;
1336	}else if( c>0 ){
1337	iLast = iCur-1;
1338	}else{
1339	return &aSetting[iCur];
1340	}
1341	}
1342	return 0;
1343	}
1344
1345	/*****************************************************************************
1346	** A virtual table for accessing the information in aCommand[], and
1347	** especially the help-text
1348	*/
1349

M src/doc.c

		--- src/doc.c
		+++ src/doc.c
		@@ -305,10 +305,12 @@
305	305	{ "xlsx", 4, "application/vnd.openxmlformats-"
306	306	"officedocument.spreadsheetml.sheet"},
307	307	{ "xlw", 3, "application/vnd.ms-excel" },
308	308	{ "xml", 3, "text/xml" },
309	309	{ "xpm", 3, "image/x-xpixmap" },
	310	+ { "xsl", 3, "text/xml" },
	311	+ { "xslt", 4, "text/xml" },
310	312	{ "xwd", 3, "image/x-xwindowdump" },
311	313	{ "xyz", 3, "chemical/x-pdb" },
312	314	{ "zip", 3, "application/zip" },
313	315	};
314	316
315	317

	--- src/doc.c
	+++ src/doc.c
	@@ -305,10 +305,12 @@
305	{ "xlsx", 4, "application/vnd.openxmlformats-"
306	"officedocument.spreadsheetml.sheet"},
307	{ "xlw", 3, "application/vnd.ms-excel" },
308	{ "xml", 3, "text/xml" },
309	{ "xpm", 3, "image/x-xpixmap" },


310	{ "xwd", 3, "image/x-xwindowdump" },
311	{ "xyz", 3, "chemical/x-pdb" },
312	{ "zip", 3, "application/zip" },
313	};
314
315

	--- src/doc.c
	+++ src/doc.c
	@@ -305,10 +305,12 @@
305	{ "xlsx", 4, "application/vnd.openxmlformats-"
306	"officedocument.spreadsheetml.sheet"},
307	{ "xlw", 3, "application/vnd.ms-excel" },
308	{ "xml", 3, "text/xml" },
309	{ "xpm", 3, "image/x-xpixmap" },
310	{ "xsl", 3, "text/xml" },
311	{ "xslt", 4, "text/xml" },
312	{ "xwd", 3, "image/x-xwindowdump" },
313	{ "xyz", 3, "chemical/x-pdb" },
314	{ "zip", 3, "application/zip" },
315	};
316
317

M src/event.c

+1 -1

		--- src/event.c
		+++ src/event.c
		@@ -229,11 +229,11 @@
229	229	}
230	230	zFullId = db_text(0, "SELECT SUBSTR(tagname,7)"
231	231	" FROM tag"
232	232	" WHERE tagname GLOB 'event-%q*'",
233	233	zId);
234		- attachment_list(zFullId, "<hr><h2>Attachments:</h2><ul>");
	234	+ attachment_list(zFullId, "<h2>Attachments:</h2>", 1);
235	235	document_emit_js();
236	236	style_finish_page();
237	237	manifest_destroy(pTNote);
238	238	}
239	239
240	240

	--- src/event.c
	+++ src/event.c
	@@ -229,11 +229,11 @@
229	}
230	zFullId = db_text(0, "SELECT SUBSTR(tagname,7)"
231	" FROM tag"
232	" WHERE tagname GLOB 'event-%q*'",
233	zId);
234	attachment_list(zFullId, "<hr><h2>Attachments:</h2><ul>");
235	document_emit_js();
236	style_finish_page();
237	manifest_destroy(pTNote);
238	}
239
240

	--- src/event.c
	+++ src/event.c
	@@ -229,11 +229,11 @@
229	}
230	zFullId = db_text(0, "SELECT SUBSTR(tagname,7)"
231	" FROM tag"
232	" WHERE tagname GLOB 'event-%q*'",
233	zId);
234	attachment_list(zFullId, "<h2>Attachments:</h2>", 1);
235	document_emit_js();
236	style_finish_page();
237	manifest_destroy(pTNote);
238	}
239
240

M src/file.c

+99 -9

		--- src/file.c
		+++ src/file.c
		@@ -1302,25 +1302,37 @@
1302	1302	}
1303	1303	}
1304	1304
1305	1305	/*
1306	1306	** Compute a canonical pathname for a file or directory.
1307		-** Make the name absolute if it is relative.
1308		-** Remove redundant / characters
1309		-** Remove all /./ path elements.
1310		-** Convert /A/../ to just /
	1307	+**
	1308	+** * Make the name absolute if it is relative.
	1309	+** * Remove redundant / characters
	1310	+** * Remove all /./ path elements.
	1311	+** * Convert /A/../ to just /
	1312	+** * On windows, add the drive letter prefix.
	1313	+**
1311	1314	** If the slash parameter is non-zero, the trailing slash, if any,
1312	1315	** is retained.
1313	1316	**
1314	1317	** See also: file_canonical_name_dup()
1315	1318	*/
1316	1319	void file_canonical_name(const char zOrigName, Blob pOut, int slash){
	1320	+ char zPwd[2000];
1317	1321	blob_zero(pOut);
1318	1322	if( file_is_absolute_path(zOrigName) ){
1319		- blob_appendf(pOut, "%/", zOrigName);
	1323	+#if defined(_WIN32)
	1324	+ if( fossil_isdirsep(zOrigName[0]) ){
	1325	+ /* Add the drive letter to the full pathname */
	1326	+ file_getcwd(zPwd, sizeof(zPwd)-strlen(zOrigName));
	1327	+ blob_appendf(pOut, "%.2s%/", zPwd, zOrigName);
	1328	+ }else
	1329	+#endif
	1330	+ {
	1331	+ blob_appendf(pOut, "%/", zOrigName);
	1332	+ }
1320	1333	}else{
1321		- char zPwd[2000];
1322	1334	file_getcwd(zPwd, sizeof(zPwd)-strlen(zOrigName));
1323	1335	if( zPwd[0]=='/' && strlen(zPwd)==1 ){
1324	1336	/* when on '/', don't add an extra '/' */
1325	1337	if( zOrigName[0]=='.' && strlen(zOrigName)==1 ){
1326	1338	/* '.' when on '/' mean '/' */
		@@ -1372,10 +1384,11 @@
1372	1384	** just fossil_strdup(). But for case-insenstiive but "case preserving"
1373	1385	** filesystems, such as on MacOS or Windows, we want the filename to be
1374	1386	** in the preserved casing. That's what this routine does.
1375	1387	*/
1376	1388	char file_case_preferred_name(const char zDir, const char *zPath){
	1389	+#ifndef _WIN32 /* Call win32_file_case_preferred_name() on Windows. */
1377	1390	DIR *d;
1378	1391	int i;
1379	1392	char *zResult = 0;
1380	1393	void *zNative = 0;
1381	1394
		@@ -1407,10 +1420,13 @@
1407	1420	closedir(d);
1408	1421	}
1409	1422	fossil_path_free(zNative);
1410	1423	if( zResult==0 ) zResult = fossil_strdup(zPath);
1411	1424	return zResult;
	1425	+#else /* !_WIN32 */
	1426	+ return win32_file_case_preferred_name(zDir,zPath);
	1427	+#endif /* !_WIN32 */
1412	1428	}
1413	1429
1414	1430	/*
1415	1431	** COMMAND: test-case-filename
1416	1432	**
		@@ -1637,13 +1653,16 @@
1637	1653	fossil_print("filenames_are_case_sensitive() = %d\n",
1638	1654	filenames_are_case_sensitive());
1639	1655	if( zAllow ){
1640	1656	g.allowSymlinks = !is_false(zAllow);
1641	1657	}
1642		- if( zRoot==0 ) zRoot = g.zLocalRoot;
	1658	+ if( zRoot==0 ) zRoot = g.zLocalRoot==0 ? "" : g.zLocalRoot;
1643	1659	fossil_print("db_allow_symlinks() = %d\n", db_allow_symlinks());
1644	1660	fossil_print("local-root = [%s]\n", zRoot);
	1661	+ if( g.db==0 ) sqlite3_open(":memory:", &g.db);
	1662	+ sqlite3_create_function(g.db, "inode", 1, SQLITE_UTF8, 0,
	1663	+ file_inode_sql_func, 0, 0);
1645	1664	for(i=2; i<g.argc; i++){
1646	1665	char *z;
1647	1666	emitFileStat(g.argv[i], slashFlag, resetFlag);
1648	1667	z = file_canonical_name_dup(g.argv[i]);
1649	1668	fossil_print(" file_canonical_name = %s\n", z);
		@@ -1653,10 +1672,13 @@
1653	1672	}else{
1654	1673	int n = file_nondir_objects_on_path(zRoot, z);
1655	1674	fossil_print("%.*s\n", n, z);
1656	1675	}
1657	1676	fossil_free(z);
	1677	+ z = db_text(0, "SELECT inode(%Q)", g.argv[i]);
	1678	+ fossil_print(" file_inode_sql_func = \"%s\"\n", z);
	1679	+ fossil_free(z);
1658	1680	}
1659	1681	}
1660	1682
1661	1683	/*
1662	1684	** COMMAND: test-canonical-name
		@@ -1695,13 +1717,15 @@
1695	1717	** Canonical names are full pathnames using "/" not "\" and which
1696	1718	** contain no "/./" or "/../" terms.
1697	1719	*/
1698	1720	int file_is_canonical(const char *z){
1699	1721	int i;
1700		- if( z[0]!='/'
	1722	+ if(
1701	1723	#if defined(_WIN32) \|\| defined(__CYGWIN__)
1702		- && (!fossil_isupper(z[0]) \|\| z[1]!=':' \|\| z[2]!='/')
	1724	+ !fossil_isupper(z[0]) \|\| z[1]!=':' \|\| !fossil_isdirsep(z[2])
	1725	+#else
	1726	+ z[0]!='/'
1703	1727	#endif
1704	1728	) return 0;
1705	1729
1706	1730	for(i=0; z[i]; i++){
1707	1731	if( z[i]=='\\' ) return 0;
		@@ -2949,18 +2973,84 @@
2949	2973	** Returns 1 if the given directory contains a file named .fslckout, 2
2950	2974	** if it contains a file named _FOSSIL_, else returns 0.
2951	2975	*/
2952	2976	int dir_has_ckout_db(const char *zDir){
2953	2977	int rc = 0;
	2978	+ i64 sz;
2954	2979	char * zCkoutDb = mprintf("%//.fslckout", zDir);
2955	2980	if(file_isfile(zCkoutDb, ExtFILE)){
2956	2981	rc = 1;
2957	2982	}else{
2958	2983	fossil_free(zCkoutDb);
2959	2984	zCkoutDb = mprintf("%//_FOSSIL_", zDir);
2960	2985	if(file_isfile(zCkoutDb, ExtFILE)){
2961	2986	rc = 2;
2962	2987	}
	2988	+ }
	2989	+ if( rc && ((sz = file_size(zCkoutDb, ExtFILE))<1024 \|\| (sz%512)!=0) ){
	2990	+ rc = 0;
2963	2991	}
2964	2992	fossil_free(zCkoutDb);
2965	2993	return rc;
2966	2994	}
	2995	+
	2996	+/*
	2997	+** This is the implementation of inode(FILENAME) SQL function.
	2998	+**
	2999	+** dev_inode(FILENAME) returns a string. If FILENAME exists and is
	3000	+** a regular file, then the return string is of the form:
	3001	+**
	3002	+** DEV/INODE
	3003	+**
	3004	+** Where DEV and INODE are the device number and inode number for
	3005	+** the file. On Windows, the volume serial number (DEV) and file
	3006	+** identifier (INODE) are used to compute the value, see comments
	3007	+** on the win32_file_id() function.
	3008	+**
	3009	+** If FILENAME does not exist, then the return is an empty string.
	3010	+**
	3011	+** The value of inode() can be used to eliminate files from a list
	3012	+** that have duplicates because they have differing names due to links.
	3013	+**
	3014	+** Code that wants to use this SQL function needs to first register
	3015	+** it using a call such as the following:
	3016	+**
	3017	+** sqlite3_create_function(g.db, "inode", 1, SQLITE_UTF8, 0,
	3018	+** file_inode_sql_func, 0, 0);
	3019	+*/
	3020	+void file_inode_sql_func(
	3021	+ sqlite3_context *context,
	3022	+ int argc,
	3023	+ sqlite3_value **argv
	3024	+){
	3025	+ const char *zFilename;
	3026	+ assert( argc==1 );
	3027	+ zFilename = (const char*)sqlite3_value_text(argv[0]);
	3028	+ if( zFilename==0 \|\| zFilename[0]==0 \|\| file_access(zFilename,F_OK) ){
	3029	+ sqlite3_result_text(context, "", 0, SQLITE_STATIC);
	3030	+ return;
	3031	+ }
	3032	+#if defined(_WIN32)
	3033	+ {
	3034	+ char *zFileId = win32_file_id(zFilename);
	3035	+ if( zFileId ){
	3036	+ sqlite3_result_text(context, zFileId, -1, fossil_free);
	3037	+ }else{
	3038	+ sqlite3_result_text(context, "", 0, SQLITE_STATIC);
	3039	+ }
	3040	+ }
	3041	+#else
	3042	+ {
	3043	+ struct stat buf;
	3044	+ int rc;
	3045	+ memset(&buf, 0, sizeof(buf));
	3046	+ rc = stat(zFilename, &buf);
	3047	+ if( rc ){
	3048	+ sqlite3_result_text(context, "", 0, SQLITE_STATIC);
	3049	+ }else{
	3050	+ sqlite3_result_text(context,
	3051	+ mprintf("%lld/%lld", (i64)buf.st_dev, (i64)buf.st_ino), -1,
	3052	+ fossil_free);
	3053	+ }
	3054	+ }
	3055	+#endif
	3056	+}
2967	3057

	--- src/file.c
	+++ src/file.c
	@@ -1302,25 +1302,37 @@
1302	}
1303	}
1304
1305	/*
1306	** Compute a canonical pathname for a file or directory.
1307	** Make the name absolute if it is relative.
1308	** Remove redundant / characters
1309	** Remove all /./ path elements.
1310	** Convert /A/../ to just /



1311	** If the slash parameter is non-zero, the trailing slash, if any,
1312	** is retained.
1313	**
1314	** See also: file_canonical_name_dup()
1315	*/
1316	void file_canonical_name(const char zOrigName, Blob pOut, int slash){

1317	blob_zero(pOut);
1318	if( file_is_absolute_path(zOrigName) ){
1319	blob_appendf(pOut, "%/", zOrigName);









1320	}else{
1321	char zPwd[2000];
1322	file_getcwd(zPwd, sizeof(zPwd)-strlen(zOrigName));
1323	if( zPwd[0]=='/' && strlen(zPwd)==1 ){
1324	/* when on '/', don't add an extra '/' */
1325	if( zOrigName[0]=='.' && strlen(zOrigName)==1 ){
1326	/* '.' when on '/' mean '/' */
	@@ -1372,10 +1384,11 @@
1372	** just fossil_strdup(). But for case-insenstiive but "case preserving"
1373	** filesystems, such as on MacOS or Windows, we want the filename to be
1374	** in the preserved casing. That's what this routine does.
1375	*/
1376	char file_case_preferred_name(const char zDir, const char *zPath){

1377	DIR *d;
1378	int i;
1379	char *zResult = 0;
1380	void *zNative = 0;
1381
	@@ -1407,10 +1420,13 @@
1407	closedir(d);
1408	}
1409	fossil_path_free(zNative);
1410	if( zResult==0 ) zResult = fossil_strdup(zPath);
1411	return zResult;



1412	}
1413
1414	/*
1415	** COMMAND: test-case-filename
1416	**
	@@ -1637,13 +1653,16 @@
1637	fossil_print("filenames_are_case_sensitive() = %d\n",
1638	filenames_are_case_sensitive());
1639	if( zAllow ){
1640	g.allowSymlinks = !is_false(zAllow);
1641	}
1642	if( zRoot==0 ) zRoot = g.zLocalRoot;
1643	fossil_print("db_allow_symlinks() = %d\n", db_allow_symlinks());
1644	fossil_print("local-root = [%s]\n", zRoot);



1645	for(i=2; i<g.argc; i++){
1646	char *z;
1647	emitFileStat(g.argv[i], slashFlag, resetFlag);
1648	z = file_canonical_name_dup(g.argv[i]);
1649	fossil_print(" file_canonical_name = %s\n", z);
	@@ -1653,10 +1672,13 @@
1653	}else{
1654	int n = file_nondir_objects_on_path(zRoot, z);
1655	fossil_print("%.*s\n", n, z);
1656	}
1657	fossil_free(z);



1658	}
1659	}
1660
1661	/*
1662	** COMMAND: test-canonical-name
	@@ -1695,13 +1717,15 @@
1695	** Canonical names are full pathnames using "/" not "\" and which
1696	** contain no "/./" or "/../" terms.
1697	*/
1698	int file_is_canonical(const char *z){
1699	int i;
1700	if( z[0]!='/'
1701	#if defined(_WIN32) \|\| defined(__CYGWIN__)
1702	&& (!fossil_isupper(z[0]) \|\| z[1]!=':' \|\| z[2]!='/')


1703	#endif
1704	) return 0;
1705
1706	for(i=0; z[i]; i++){
1707	if( z[i]=='\\' ) return 0;
	@@ -2949,18 +2973,84 @@
2949	** Returns 1 if the given directory contains a file named .fslckout, 2
2950	** if it contains a file named _FOSSIL_, else returns 0.
2951	*/
2952	int dir_has_ckout_db(const char *zDir){
2953	int rc = 0;

2954	char * zCkoutDb = mprintf("%//.fslckout", zDir);
2955	if(file_isfile(zCkoutDb, ExtFILE)){
2956	rc = 1;
2957	}else{
2958	fossil_free(zCkoutDb);
2959	zCkoutDb = mprintf("%//_FOSSIL_", zDir);
2960	if(file_isfile(zCkoutDb, ExtFILE)){
2961	rc = 2;
2962	}



2963	}
2964	fossil_free(zCkoutDb);
2965	return rc;
2966	}






























































2967

	--- src/file.c
	+++ src/file.c
	@@ -1302,25 +1302,37 @@
1302	}
1303	}
1304
1305	/*
1306	** Compute a canonical pathname for a file or directory.
1307	**
1308	** * Make the name absolute if it is relative.
1309	** * Remove redundant / characters
1310	** * Remove all /./ path elements.
1311	** * Convert /A/../ to just /
1312	** * On windows, add the drive letter prefix.
1313	**
1314	** If the slash parameter is non-zero, the trailing slash, if any,
1315	** is retained.
1316	**
1317	** See also: file_canonical_name_dup()
1318	*/
1319	void file_canonical_name(const char zOrigName, Blob pOut, int slash){
1320	char zPwd[2000];
1321	blob_zero(pOut);
1322	if( file_is_absolute_path(zOrigName) ){
1323	#if defined(_WIN32)
1324	if( fossil_isdirsep(zOrigName[0]) ){
1325	/* Add the drive letter to the full pathname */
1326	file_getcwd(zPwd, sizeof(zPwd)-strlen(zOrigName));
1327	blob_appendf(pOut, "%.2s%/", zPwd, zOrigName);
1328	}else
1329	#endif
1330	{
1331	blob_appendf(pOut, "%/", zOrigName);
1332	}
1333	}else{

1334	file_getcwd(zPwd, sizeof(zPwd)-strlen(zOrigName));
1335	if( zPwd[0]=='/' && strlen(zPwd)==1 ){
1336	/* when on '/', don't add an extra '/' */
1337	if( zOrigName[0]=='.' && strlen(zOrigName)==1 ){
1338	/* '.' when on '/' mean '/' */
	@@ -1372,10 +1384,11 @@
1384	** just fossil_strdup(). But for case-insenstiive but "case preserving"
1385	** filesystems, such as on MacOS or Windows, we want the filename to be
1386	** in the preserved casing. That's what this routine does.
1387	*/
1388	char file_case_preferred_name(const char zDir, const char *zPath){
1389	#ifndef _WIN32 /* Call win32_file_case_preferred_name() on Windows. */
1390	DIR *d;
1391	int i;
1392	char *zResult = 0;
1393	void *zNative = 0;
1394
	@@ -1407,10 +1420,13 @@
1420	closedir(d);
1421	}
1422	fossil_path_free(zNative);
1423	if( zResult==0 ) zResult = fossil_strdup(zPath);
1424	return zResult;
1425	#else /* !_WIN32 */
1426	return win32_file_case_preferred_name(zDir,zPath);
1427	#endif /* !_WIN32 */
1428	}
1429
1430	/*
1431	** COMMAND: test-case-filename
1432	**
	@@ -1637,13 +1653,16 @@
1653	fossil_print("filenames_are_case_sensitive() = %d\n",
1654	filenames_are_case_sensitive());
1655	if( zAllow ){
1656	g.allowSymlinks = !is_false(zAllow);
1657	}
1658	if( zRoot==0 ) zRoot = g.zLocalRoot==0 ? "" : g.zLocalRoot;
1659	fossil_print("db_allow_symlinks() = %d\n", db_allow_symlinks());
1660	fossil_print("local-root = [%s]\n", zRoot);
1661	if( g.db==0 ) sqlite3_open(":memory:", &g.db);
1662	sqlite3_create_function(g.db, "inode", 1, SQLITE_UTF8, 0,
1663	file_inode_sql_func, 0, 0);
1664	for(i=2; i<g.argc; i++){
1665	char *z;
1666	emitFileStat(g.argv[i], slashFlag, resetFlag);
1667	z = file_canonical_name_dup(g.argv[i]);
1668	fossil_print(" file_canonical_name = %s\n", z);
	@@ -1653,10 +1672,13 @@
1672	}else{
1673	int n = file_nondir_objects_on_path(zRoot, z);
1674	fossil_print("%.*s\n", n, z);
1675	}
1676	fossil_free(z);
1677	z = db_text(0, "SELECT inode(%Q)", g.argv[i]);
1678	fossil_print(" file_inode_sql_func = \"%s\"\n", z);
1679	fossil_free(z);
1680	}
1681	}
1682
1683	/*
1684	** COMMAND: test-canonical-name
	@@ -1695,13 +1717,15 @@
1717	** Canonical names are full pathnames using "/" not "\" and which
1718	** contain no "/./" or "/../" terms.
1719	*/
1720	int file_is_canonical(const char *z){
1721	int i;
1722	if(
1723	#if defined(_WIN32) \|\| defined(__CYGWIN__)
1724	!fossil_isupper(z[0]) \|\| z[1]!=':' \|\| !fossil_isdirsep(z[2])
1725	#else
1726	z[0]!='/'
1727	#endif
1728	) return 0;
1729
1730	for(i=0; z[i]; i++){
1731	if( z[i]=='\\' ) return 0;
	@@ -2949,18 +2973,84 @@
2973	** Returns 1 if the given directory contains a file named .fslckout, 2
2974	** if it contains a file named _FOSSIL_, else returns 0.
2975	*/
2976	int dir_has_ckout_db(const char *zDir){
2977	int rc = 0;
2978	i64 sz;
2979	char * zCkoutDb = mprintf("%//.fslckout", zDir);
2980	if(file_isfile(zCkoutDb, ExtFILE)){
2981	rc = 1;
2982	}else{
2983	fossil_free(zCkoutDb);
2984	zCkoutDb = mprintf("%//_FOSSIL_", zDir);
2985	if(file_isfile(zCkoutDb, ExtFILE)){
2986	rc = 2;
2987	}
2988	}
2989	if( rc && ((sz = file_size(zCkoutDb, ExtFILE))<1024 \|\| (sz%512)!=0) ){
2990	rc = 0;
2991	}
2992	fossil_free(zCkoutDb);
2993	return rc;
2994	}
2995
2996	/*
2997	** This is the implementation of inode(FILENAME) SQL function.
2998	**
2999	** dev_inode(FILENAME) returns a string. If FILENAME exists and is
3000	** a regular file, then the return string is of the form:
3001	**
3002	** DEV/INODE
3003	**
3004	** Where DEV and INODE are the device number and inode number for
3005	** the file. On Windows, the volume serial number (DEV) and file
3006	** identifier (INODE) are used to compute the value, see comments
3007	** on the win32_file_id() function.
3008	**
3009	** If FILENAME does not exist, then the return is an empty string.
3010	**
3011	** The value of inode() can be used to eliminate files from a list
3012	** that have duplicates because they have differing names due to links.
3013	**
3014	** Code that wants to use this SQL function needs to first register
3015	** it using a call such as the following:
3016	**
3017	** sqlite3_create_function(g.db, "inode", 1, SQLITE_UTF8, 0,
3018	** file_inode_sql_func, 0, 0);
3019	*/
3020	void file_inode_sql_func(
3021	sqlite3_context *context,
3022	int argc,
3023	sqlite3_value **argv
3024	){
3025	const char *zFilename;
3026	assert( argc==1 );
3027	zFilename = (const char*)sqlite3_value_text(argv[0]);
3028	if( zFilename==0 \|\| zFilename[0]==0 \|\| file_access(zFilename,F_OK) ){
3029	sqlite3_result_text(context, "", 0, SQLITE_STATIC);
3030	return;
3031	}
3032	#if defined(_WIN32)
3033	{
3034	char *zFileId = win32_file_id(zFilename);
3035	if( zFileId ){
3036	sqlite3_result_text(context, zFileId, -1, fossil_free);
3037	}else{
3038	sqlite3_result_text(context, "", 0, SQLITE_STATIC);
3039	}
3040	}
3041	#else
3042	{
3043	struct stat buf;
3044	int rc;
3045	memset(&buf, 0, sizeof(buf));
3046	rc = stat(zFilename, &buf);
3047	if( rc ){
3048	sqlite3_result_text(context, "", 0, SQLITE_STATIC);
3049	}else{
3050	sqlite3_result_text(context,
3051	mprintf("%lld/%lld", (i64)buf.st_dev, (i64)buf.st_ino), -1,
3052	fossil_free);
3053	}
3054	}
3055	#endif
3056	}
3057

M src/finfo.c

+13 -5

		--- src/finfo.c
		+++ src/finfo.c
		@@ -185,11 +185,11 @@
185	185	zLimit = find_option("limit","n",1);
186	186	zWidth = find_option("width","W",1);
187	187	iLimit = zLimit ? atoi(zLimit) : -1;
188	188	zOffset = find_option("offset",0,1);
189	189	iOffset = zOffset ? atoi(zOffset) : 0;
190		- iBrief = (find_option("brief","b",0) == 0);
	190	+ iBrief = find_option("brief","b",0) != 0;
191	191	if( iLimit==0 ){
192	192	iLimit = -1;
193	193	}
194	194	if( zWidth ){
195	195	iWidth = atoi(zWidth);
		@@ -228,11 +228,11 @@
228	228	" ORDER BY event.mtime DESC LIMIT %d OFFSET %d",
229	229	TAG_BRANCH, zFilename, filename_collation(),
230	230	iLimit, iOffset
231	231	);
232	232	blob_zero(&line);
233		- if( iBrief ){
	233	+ if( iBrief == 0 ){
234	234	fossil_print("History for %s\n", blob_str(&fname));
235	235	}
236	236	while( db_step(&q)==SQLITE_ROW ){
237	237	const char *zFileUuid = db_column_text(&q, 0);
238	238	const char *zCiUuid = db_column_text(&q,1);
		@@ -240,11 +240,11 @@
240	240	const char *zCom = db_column_text(&q, 3);
241	241	const char *zUser = db_column_text(&q, 4);
242	242	const char *zBr = db_column_text(&q, 5);
243	243	char *zOut;
244	244	if( zBr==0 ) zBr = "trunk";
245		- if( iBrief ){
	245	+ if( iBrief == 0 ){
246	246	fossil_print("%s ", zDate);
247	247	zOut = mprintf(
248	248	"[%S] %s (user: %s, artifact: [%S], branch: %s)",
249	249	zCiUuid, zCom, zUser, zFileUuid, zBr);
250	250	comment_print(zOut, zCom, 11, iWidth, get_comment_format());
		@@ -620,10 +620,11 @@
620	620	int fnid = db_column_int(&q, 13);
621	621	const char *zFName = db_column_text(&q,14);
622	622	int gidx;
623	623	char zTime[10];
624	624	int nParent = 0;
	625	+ int bIsModified = 0;
625	626	GraphRowId aParent[GR_MAX_RAIL];
626	627
627	628	db_bind_int(&qparent, ":fid", frid);
628	629	db_bind_int(&qparent, ":mid", fmid);
629	630	db_bind_int(&qparent, ":fnid", fnid);
		@@ -661,19 +662,26 @@
661	662	@ </td>
662	663	if( zBgClr && zBgClr[0] ){
663	664	@ <td class="timeline%s(zStyle)Cell" id='mc%d(gidx)'>
664	665	}else{
665	666	@ <td class="timeline%s(zStyle)Cell">
	667	+ }
	668	+ if( zPUuid && zUuid && fossil_strcmp(zPUuid, zUuid)!=0 ){
	669	+ bIsModified = 1;
666	670	}
667	671	if( tmFlags & TIMELINE_COMPACT ){
668	672	@ <span class='timelineCompactComment' data-id='%d(frid)'>
669	673	}else{
670	674	@ <span class='timeline%s(zStyle)Comment'>
671	675	if( pfnid ){
672	676	char *zPrevName = db_text(0,"SELECT name FROM filename WHERE fnid=%d",
673	677	pfnid);
674		- @ <b>Renamed</b> %h(zPrevName) → %h(zFName).
	678	+ if( bIsModified ){
	679	+ @ <b>Renamed and modified</b> %h(zPrevName) → %h(zFName).
	680	+ }else{
	681	+ @ <b>Renamed</b> %h(zPrevName) → %h(zFName).
	682	+ }
675	683	fossil_free(zPrevName);
676	684	}
677	685	if( zUuid && ridTo==0 && nParent==0 ){
678	686	@ <b>Added:</b>
679	687	}
		@@ -756,11 +764,11 @@
756	764	@ %z(href("%R/annotate?filename=%h&checkin=%s",z,zCkin))
757	765	@ [annotate]</a>
758	766	@ %z(href("%R/blame?filename=%h&checkin=%s",z,zCkin))
759	767	@ [blame]</a>
760	768	@ %z(href("%R/timeline?uf=%!S",zUuid))[check-ins using]</a>
761		- if( fpid>0 ){
	769	+ if( fpid>0 && bIsModified!=0 ){
762	770	@ %z(href("%R/fdiff?v1=%!S&v2=%!S",zPUuid,zUuid))[diff]</a>
763	771	}
764	772	if( fileedit_is_editable(zFName) ){
765	773	@ %z(href("%R/fileedit?filename=%T&checkin=%!S",zFName,zCkin))\
766	774	@ [edit]</a>
767	775

	--- src/finfo.c
	+++ src/finfo.c
	@@ -185,11 +185,11 @@
185	zLimit = find_option("limit","n",1);
186	zWidth = find_option("width","W",1);
187	iLimit = zLimit ? atoi(zLimit) : -1;
188	zOffset = find_option("offset",0,1);
189	iOffset = zOffset ? atoi(zOffset) : 0;
190	iBrief = (find_option("brief","b",0) == 0);
191	if( iLimit==0 ){
192	iLimit = -1;
193	}
194	if( zWidth ){
195	iWidth = atoi(zWidth);
	@@ -228,11 +228,11 @@
228	" ORDER BY event.mtime DESC LIMIT %d OFFSET %d",
229	TAG_BRANCH, zFilename, filename_collation(),
230	iLimit, iOffset
231	);
232	blob_zero(&line);
233	if( iBrief ){
234	fossil_print("History for %s\n", blob_str(&fname));
235	}
236	while( db_step(&q)==SQLITE_ROW ){
237	const char *zFileUuid = db_column_text(&q, 0);
238	const char *zCiUuid = db_column_text(&q,1);
	@@ -240,11 +240,11 @@
240	const char *zCom = db_column_text(&q, 3);
241	const char *zUser = db_column_text(&q, 4);
242	const char *zBr = db_column_text(&q, 5);
243	char *zOut;
244	if( zBr==0 ) zBr = "trunk";
245	if( iBrief ){
246	fossil_print("%s ", zDate);
247	zOut = mprintf(
248	"[%S] %s (user: %s, artifact: [%S], branch: %s)",
249	zCiUuid, zCom, zUser, zFileUuid, zBr);
250	comment_print(zOut, zCom, 11, iWidth, get_comment_format());
	@@ -620,10 +620,11 @@
620	int fnid = db_column_int(&q, 13);
621	const char *zFName = db_column_text(&q,14);
622	int gidx;
623	char zTime[10];
624	int nParent = 0;

625	GraphRowId aParent[GR_MAX_RAIL];
626
627	db_bind_int(&qparent, ":fid", frid);
628	db_bind_int(&qparent, ":mid", fmid);
629	db_bind_int(&qparent, ":fnid", fnid);
	@@ -661,19 +662,26 @@
661	@ </td>
662	if( zBgClr && zBgClr[0] ){
663	@ <td class="timeline%s(zStyle)Cell" id='mc%d(gidx)'>
664	}else{
665	@ <td class="timeline%s(zStyle)Cell">



666	}
667	if( tmFlags & TIMELINE_COMPACT ){
668	@ <span class='timelineCompactComment' data-id='%d(frid)'>
669	}else{
670	@ <span class='timeline%s(zStyle)Comment'>
671	if( pfnid ){
672	char *zPrevName = db_text(0,"SELECT name FROM filename WHERE fnid=%d",
673	pfnid);
674	@ <b>Renamed</b> %h(zPrevName) → %h(zFName).




675	fossil_free(zPrevName);
676	}
677	if( zUuid && ridTo==0 && nParent==0 ){
678	@ <b>Added:</b>
679	}
	@@ -756,11 +764,11 @@
756	@ %z(href("%R/annotate?filename=%h&checkin=%s",z,zCkin))
757	@ [annotate]</a>
758	@ %z(href("%R/blame?filename=%h&checkin=%s",z,zCkin))
759	@ [blame]</a>
760	@ %z(href("%R/timeline?uf=%!S",zUuid))[check-ins using]</a>
761	if( fpid>0 ){
762	@ %z(href("%R/fdiff?v1=%!S&v2=%!S",zPUuid,zUuid))[diff]</a>
763	}
764	if( fileedit_is_editable(zFName) ){
765	@ %z(href("%R/fileedit?filename=%T&checkin=%!S",zFName,zCkin))\
766	@ [edit]</a>
767

	--- src/finfo.c
	+++ src/finfo.c
	@@ -185,11 +185,11 @@
185	zLimit = find_option("limit","n",1);
186	zWidth = find_option("width","W",1);
187	iLimit = zLimit ? atoi(zLimit) : -1;
188	zOffset = find_option("offset",0,1);
189	iOffset = zOffset ? atoi(zOffset) : 0;
190	iBrief = find_option("brief","b",0) != 0;
191	if( iLimit==0 ){
192	iLimit = -1;
193	}
194	if( zWidth ){
195	iWidth = atoi(zWidth);
	@@ -228,11 +228,11 @@
228	" ORDER BY event.mtime DESC LIMIT %d OFFSET %d",
229	TAG_BRANCH, zFilename, filename_collation(),
230	iLimit, iOffset
231	);
232	blob_zero(&line);
233	if( iBrief == 0 ){
234	fossil_print("History for %s\n", blob_str(&fname));
235	}
236	while( db_step(&q)==SQLITE_ROW ){
237	const char *zFileUuid = db_column_text(&q, 0);
238	const char *zCiUuid = db_column_text(&q,1);
	@@ -240,11 +240,11 @@
240	const char *zCom = db_column_text(&q, 3);
241	const char *zUser = db_column_text(&q, 4);
242	const char *zBr = db_column_text(&q, 5);
243	char *zOut;
244	if( zBr==0 ) zBr = "trunk";
245	if( iBrief == 0 ){
246	fossil_print("%s ", zDate);
247	zOut = mprintf(
248	"[%S] %s (user: %s, artifact: [%S], branch: %s)",
249	zCiUuid, zCom, zUser, zFileUuid, zBr);
250	comment_print(zOut, zCom, 11, iWidth, get_comment_format());
	@@ -620,10 +620,11 @@
620	int fnid = db_column_int(&q, 13);
621	const char *zFName = db_column_text(&q,14);
622	int gidx;
623	char zTime[10];
624	int nParent = 0;
625	int bIsModified = 0;
626	GraphRowId aParent[GR_MAX_RAIL];
627
628	db_bind_int(&qparent, ":fid", frid);
629	db_bind_int(&qparent, ":mid", fmid);
630	db_bind_int(&qparent, ":fnid", fnid);
	@@ -661,19 +662,26 @@
662	@ </td>
663	if( zBgClr && zBgClr[0] ){
664	@ <td class="timeline%s(zStyle)Cell" id='mc%d(gidx)'>
665	}else{
666	@ <td class="timeline%s(zStyle)Cell">
667	}
668	if( zPUuid && zUuid && fossil_strcmp(zPUuid, zUuid)!=0 ){
669	bIsModified = 1;
670	}
671	if( tmFlags & TIMELINE_COMPACT ){
672	@ <span class='timelineCompactComment' data-id='%d(frid)'>
673	}else{
674	@ <span class='timeline%s(zStyle)Comment'>
675	if( pfnid ){
676	char *zPrevName = db_text(0,"SELECT name FROM filename WHERE fnid=%d",
677	pfnid);
678	if( bIsModified ){
679	@ <b>Renamed and modified</b> %h(zPrevName) → %h(zFName).
680	}else{
681	@ <b>Renamed</b> %h(zPrevName) → %h(zFName).
682	}
683	fossil_free(zPrevName);
684	}
685	if( zUuid && ridTo==0 && nParent==0 ){
686	@ <b>Added:</b>
687	}
	@@ -756,11 +764,11 @@
764	@ %z(href("%R/annotate?filename=%h&checkin=%s",z,zCkin))
765	@ [annotate]</a>
766	@ %z(href("%R/blame?filename=%h&checkin=%s",z,zCkin))
767	@ [blame]</a>
768	@ %z(href("%R/timeline?uf=%!S",zUuid))[check-ins using]</a>
769	if( fpid>0 && bIsModified!=0 ){
770	@ %z(href("%R/fdiff?v1=%!S&v2=%!S",zPUuid,zUuid))[diff]</a>
771	}
772	if( fileedit_is_editable(zFName) ){
773	@ %z(href("%R/fileedit?filename=%T&checkin=%!S",zFName,zCkin))\
774	@ [edit]</a>
775

M src/forum.c

+63 -35

		--- src/forum.c
		+++ src/forum.c
		@@ -1397,12 +1397,13 @@
1397	1397	@ Title: <input type="input" name="title" value="%h(zTitle)" size="50"
1398	1398	@ maxlength="125"><br>
1399	1399	}
1400	1400	@ %z(href("%R/markup_help"))Markup style</a>:
1401	1401	mimetype_option_menu(zMimetype, "mimetype");
1402		- @ <br><textarea aria-label="Content:" name="content" class="wikiedit" \
1403		- @ cols="80" rows="25" wrap="virtual">%h(zContent)</textarea><br>
	1402	+ @ <div class="forum-editor-widget">
	1403	+ @ <textarea aria-label="Content:" name="content" class="wikiedit" \
	1404	+ @ cols="80" rows="25" wrap="virtual">%h(zContent)</textarea></div>
1404	1405	}
1405	1406
1406	1407	/*
1407	1408	** WEBPAGE: forumpost_close hidden
1408	1409	** WEBPAGE: forumpost_reopen hidden
		@@ -1760,20 +1761,35 @@
1760	1761	@ </form>
1761	1762	forum_emit_js();
1762	1763	style_finish_page();
1763	1764	}
1764	1765
	1766	+/*
	1767	+** SETTING: forum-close-policy boolean default=off
	1768	+** If true, forum moderators may close/re-open forum posts, and reply
	1769	+** to closed posts. If false, only administrators may do so. Note that
	1770	+** this only affects the forum web UI, not post-closing tags which
	1771	+** arrive via the command-line or from synchronization with a remote.
	1772	+*/
	1773	+/*
	1774	+** SETTING: forum-title width=20 default=Forum
	1775	+** This is the name or "title" of the Forum for this repository. The
	1776	+** default is just "Forum". But in some setups, admins might want to
	1777	+** change it to "Developer Forum" or "User Forum" or whatever other name
	1778	+** seems more appropriate for the particular usage.
	1779	+*/
	1780	+
1765	1781	/*
1766	1782	** WEBPAGE: setup_forum
1767	1783	**
1768	1784	** Forum configuration and metrics.
1769	1785	*/
1770	1786	void forum_setup(void){
1771	1787	/* boolean config settings specific to the forum. */
1772		- const char * zSettingsBool[] = {
1773		- "forum-close-policy",
1774		- NULL /* sentinel entry */
	1788	+ static const char *azForumSettings[] = {
	1789	+ "forum-close-policy",
	1790	+ "forum-title",
1775	1791	};
1776	1792
1777	1793	login_check_credentials();
1778	1794	if( !g.perm.Setup ){
1779	1795	login_needed(g.anon.Setup);
		@@ -1788,45 +1804,40 @@
1788	1804	@ <p><a href='%R/forum'>Forum posts</a>:
1789	1805	@ <a href='%R/timeline?y=f'>%d(nPosts)</a></p>
1790	1806	}
1791	1807
1792	1808	@ <h2>Supervisors</h2>
1793		- @ <p>Users with capabilities 's', 'a', or '6'.</p>
1794	1809	{
1795	1810	Stmt q = empty_Stmt;
1796		- int nRows = 0;
1797	1811	db_prepare(&q, "SELECT uid, login, cap FROM user "
1798	1812	"WHERE cap GLOB '[as6]' ORDER BY login");
1799	1813	@ <table class='bordered'>
1800	1814	@ <thead><tr><th>User</th><th>Capabilities</th></tr></thead>
1801	1815	@ <tbody>
1802	1816	while( SQLITE_ROW==db_step(&q) ){
1803	1817	const int iUid = db_column_int(&q, 0);
1804	1818	const char *zUser = db_column_text(&q, 1);
1805	1819	const char *zCap = db_column_text(&q, 2);
1806		- ++nRows;
1807	1820	@ <tr>
1808	1821	@ <td><a href='%R/setup_uedit?id=%d(iUid)'>%h(zUser)</a></td>
1809	1822	@ <td>(%h(zCap))</td>
1810	1823	@ </tr>
1811	1824	}
1812	1825	db_finalize(&q);
1813	1826	@</tbody></table>
1814		- if( 0==nRows ){
1815		- @ No supervisors
1816		- }else{
1817		- @ %d(nRows) supervisor(s)
1818		- }
1819	1827	}
1820	1828
1821	1829	@ <h2>Moderators</h2>
1822		- @ <p>Users with capability '5'.</p>
1823		- {
	1830	+ if( db_int(0, "SELECT count(*) FROM user "
	1831	+ " WHERE cap GLOB '5' AND cap NOT GLOB '[as6]'")==0 ){
	1832	+ @ <p>No non-supervisor moderators
	1833	+ }else{
1824	1834	Stmt q = empty_Stmt;
1825	1835	int nRows = 0;
1826	1836	db_prepare(&q, "SELECT uid, login, cap FROM user "
1827		- "WHERE cap GLOB '5' ORDER BY login");
	1837	+ "WHERE cap GLOB '5' AND cap NOT GLOB '[as6]'"
	1838	+ " ORDER BY login");
1828	1839	@ <table class='bordered'>
1829	1840	@ <thead><tr><th>User</th><th>Capabilities</th></tr></thead>
1830	1841	@ <tbody>
1831	1842	while( SQLITE_ROW==db_step(&q) ){
1832	1843	const int iUid = db_column_int(&q, 0);
		@@ -1838,43 +1849,59 @@
1838	1849	@ <td>(%h(zCap))</td>
1839	1850	@ </tr>
1840	1851	}
1841	1852	db_finalize(&q);
1842	1853	@ </tbody></table>
1843		- if( 0==nRows ){
1844		- @ No non-supervisor moderators
1845		- }else{
1846		- @ %d(nRows) moderator(s)
1847		- }
1848	1854	}
1849	1855
1850	1856	@ <h2>Settings</h2>
1851		- @ <p>Configuration settings specific to the forum.</p>
1852	1857	if( P("submit") && cgi_csrf_safe(2) ){
1853	1858	int i = 0;
1854		- const char *zSetting;
1855	1859	db_begin_transaction();
1856		- while( (zSetting = zSettingsBool[i++]) ){
1857		- const char *z = P(zSetting);
1858		- if( !z \|\| !z[0] ) z = "off";
1859		- db_set(zSetting/works-like:"x"/, z, 0);
	1860	+ for(i=0; i<ArraySize(azForumSettings); i++){
	1861	+ char zQP[4];
	1862	+ const char *z;
	1863	+ const Setting *pSetting = setting_find(azForumSettings[i]);
	1864	+ if( pSetting==0 ) continue;
	1865	+ zQP[0] = 'a'+i;
	1866	+ zQP[1] = zQP[0];
	1867	+ zQP[2] = 0;
	1868	+ z = P(zQP);
	1869	+ if( z==0 \|\| z[0]==0 ) continue;
	1870	+ db_set(pSetting->name/works-like:"x"/, z, 0);
1860	1871	}
1861	1872	db_end_transaction(0);
1862	1873	@ <p><em>Settings saved.</em></p>
1863	1874	}
1864	1875	{
1865	1876	int i = 0;
1866		- const char *zSetting;
1867	1877	@ <form action="%R/setup_forum" method="post">
1868	1878	login_insert_csrf_secret();
1869	1879	@ <table class='forum-settings-list'><tbody>
1870		- while( (zSetting = zSettingsBool[i++]) ){
1871		- @ <tr><td>
1872		- onoff_attribute("", zSetting, zSetting/works-like:"x"/, 0, 0);
1873		- @ </td><td>
1874		- @ <a href='%R/help?cmd=%h(zSetting)'>%h(zSetting)</a>
1875		- @ </td></tr>
	1880	+ for(i=0; i<ArraySize(azForumSettings); i++){
	1881	+ char zQP[4];
	1882	+ const Setting *pSetting = setting_find(azForumSettings[i]);
	1883	+ if( pSetting==0 ) continue;
	1884	+ zQP[0] = 'a'+i;
	1885	+ zQP[1] = zQP[0];
	1886	+ zQP[2] = 0;
	1887	+ if( pSetting->width==0 ){
	1888	+ /* Boolean setting */
	1889	+ @ <tr><td align="right">
	1890	+ @ <a href='%R/help?cmd=%h(pSetting->name)'>%h(pSetting->name)</a>:
	1891	+ @ </td><td>
	1892	+ onoff_attribute("", zQP, pSetting->name/works-like:"x"/, 0, 0);
	1893	+ @ </td></tr>
	1894	+ }else{
	1895	+ /* Text value setting */
	1896	+ @ <tr><td align="right">
	1897	+ @ <a href='%R/help?cmd=%h(pSetting->name)'>%h(pSetting->name)</a>:
	1898	+ @ </td><td>
	1899	+ entry_attribute("", 25, pSetting->name, zQP/works-like:""/,
	1900	+ pSetting->def, 0);
	1901	+ @ </td></tr>
	1902	+ }
1876	1903	}
1877	1904	@ </tbody></table>
1878	1905	@ <input type='submit' name='submit' value='Apply changes'>
1879	1906	@ </form>
1880	1907	}
		@@ -1909,11 +1936,12 @@
1909	1936	login_needed(g.anon.RdForum);
1910	1937	return;
1911	1938	}
1912	1939	cgi_check_for_malice();
1913	1940	style_set_current_feature("forum");
1914		- style_header( "%s", isSearch ? "Forum Search Results" : "Forum" );
	1941	+ style_header("%s%s", db_get("forum-title","Forum"),
	1942	+ isSearch ? " Search Results" : "");
1915	1943	style_submenu_element("Timeline", "%R/timeline?ss=v&y=f&vfx");
1916	1944	if( g.perm.WrForum ){
1917	1945	style_submenu_element("New Thread","%R/forumnew");
1918	1946	}else{
1919	1947	/* Can't combine this with previous case using the ternary operator
1920	1948

	--- src/forum.c
	+++ src/forum.c
	@@ -1397,12 +1397,13 @@
1397	@ Title: <input type="input" name="title" value="%h(zTitle)" size="50"
1398	@ maxlength="125"><br>
1399	}
1400	@ %z(href("%R/markup_help"))Markup style</a>:
1401	mimetype_option_menu(zMimetype, "mimetype");
1402	@ <br><textarea aria-label="Content:" name="content" class="wikiedit" \
1403	@ cols="80" rows="25" wrap="virtual">%h(zContent)</textarea><br>

1404	}
1405
1406	/*
1407	** WEBPAGE: forumpost_close hidden
1408	** WEBPAGE: forumpost_reopen hidden
	@@ -1760,20 +1761,35 @@
1760	@ </form>
1761	forum_emit_js();
1762	style_finish_page();
1763	}
1764















1765	/*
1766	** WEBPAGE: setup_forum
1767	**
1768	** Forum configuration and metrics.
1769	*/
1770	void forum_setup(void){
1771	/* boolean config settings specific to the forum. */
1772	const char * zSettingsBool[] = {
1773	"forum-close-policy",
1774	NULL /* sentinel entry */
1775	};
1776
1777	login_check_credentials();
1778	if( !g.perm.Setup ){
1779	login_needed(g.anon.Setup);
	@@ -1788,45 +1804,40 @@
1788	@ <p><a href='%R/forum'>Forum posts</a>:
1789	@ <a href='%R/timeline?y=f'>%d(nPosts)</a></p>
1790	}
1791
1792	@ <h2>Supervisors</h2>
1793	@ <p>Users with capabilities 's', 'a', or '6'.</p>
1794	{
1795	Stmt q = empty_Stmt;
1796	int nRows = 0;
1797	db_prepare(&q, "SELECT uid, login, cap FROM user "
1798	"WHERE cap GLOB '[as6]' ORDER BY login");
1799	@ <table class='bordered'>
1800	@ <thead><tr><th>User</th><th>Capabilities</th></tr></thead>
1801	@ <tbody>
1802	while( SQLITE_ROW==db_step(&q) ){
1803	const int iUid = db_column_int(&q, 0);
1804	const char *zUser = db_column_text(&q, 1);
1805	const char *zCap = db_column_text(&q, 2);
1806	++nRows;
1807	@ <tr>
1808	@ <td><a href='%R/setup_uedit?id=%d(iUid)'>%h(zUser)</a></td>
1809	@ <td>(%h(zCap))</td>
1810	@ </tr>
1811	}
1812	db_finalize(&q);
1813	@</tbody></table>
1814	if( 0==nRows ){
1815	@ No supervisors
1816	}else{
1817	@ %d(nRows) supervisor(s)
1818	}
1819	}
1820
1821	@ <h2>Moderators</h2>
1822	@ <p>Users with capability '5'.</p>
1823	{


1824	Stmt q = empty_Stmt;
1825	int nRows = 0;
1826	db_prepare(&q, "SELECT uid, login, cap FROM user "
1827	"WHERE cap GLOB '5' ORDER BY login");

1828	@ <table class='bordered'>
1829	@ <thead><tr><th>User</th><th>Capabilities</th></tr></thead>
1830	@ <tbody>
1831	while( SQLITE_ROW==db_step(&q) ){
1832	const int iUid = db_column_int(&q, 0);
	@@ -1838,43 +1849,59 @@
1838	@ <td>(%h(zCap))</td>
1839	@ </tr>
1840	}
1841	db_finalize(&q);
1842	@ </tbody></table>
1843	if( 0==nRows ){
1844	@ No non-supervisor moderators
1845	}else{
1846	@ %d(nRows) moderator(s)
1847	}
1848	}
1849
1850	@ <h2>Settings</h2>
1851	@ <p>Configuration settings specific to the forum.</p>
1852	if( P("submit") && cgi_csrf_safe(2) ){
1853	int i = 0;
1854	const char *zSetting;
1855	db_begin_transaction();
1856	while( (zSetting = zSettingsBool[i++]) ){
1857	const char *z = P(zSetting);
1858	if( !z \|\| !z[0] ) z = "off";
1859	db_set(zSetting/works-like:"x"/, z, 0);







1860	}
1861	db_end_transaction(0);
1862	@ <p><em>Settings saved.</em></p>
1863	}
1864	{
1865	int i = 0;
1866	const char *zSetting;
1867	@ <form action="%R/setup_forum" method="post">
1868	login_insert_csrf_secret();
1869	@ <table class='forum-settings-list'><tbody>
1870	while( (zSetting = zSettingsBool[i++]) ){
1871	@ <tr><td>
1872	onoff_attribute("", zSetting, zSetting/works-like:"x"/, 0, 0);
1873	@ </td><td>
1874	@ <a href='%R/help?cmd=%h(zSetting)'>%h(zSetting)</a>
1875	@ </td></tr>

















1876	}
1877	@ </tbody></table>
1878	@ <input type='submit' name='submit' value='Apply changes'>
1879	@ </form>
1880	}
	@@ -1909,11 +1936,12 @@
1909	login_needed(g.anon.RdForum);
1910	return;
1911	}
1912	cgi_check_for_malice();
1913	style_set_current_feature("forum");
1914	style_header( "%s", isSearch ? "Forum Search Results" : "Forum" );

1915	style_submenu_element("Timeline", "%R/timeline?ss=v&y=f&vfx");
1916	if( g.perm.WrForum ){
1917	style_submenu_element("New Thread","%R/forumnew");
1918	}else{
1919	/* Can't combine this with previous case using the ternary operator
1920

	--- src/forum.c
	+++ src/forum.c
	@@ -1397,12 +1397,13 @@
1397	@ Title: <input type="input" name="title" value="%h(zTitle)" size="50"
1398	@ maxlength="125"><br>
1399	}
1400	@ %z(href("%R/markup_help"))Markup style</a>:
1401	mimetype_option_menu(zMimetype, "mimetype");
1402	@ <div class="forum-editor-widget">
1403	@ <textarea aria-label="Content:" name="content" class="wikiedit" \
1404	@ cols="80" rows="25" wrap="virtual">%h(zContent)</textarea></div>
1405	}
1406
1407	/*
1408	** WEBPAGE: forumpost_close hidden
1409	** WEBPAGE: forumpost_reopen hidden
	@@ -1760,20 +1761,35 @@
1761	@ </form>
1762	forum_emit_js();
1763	style_finish_page();
1764	}
1765
1766	/*
1767	** SETTING: forum-close-policy boolean default=off
1768	** If true, forum moderators may close/re-open forum posts, and reply
1769	** to closed posts. If false, only administrators may do so. Note that
1770	** this only affects the forum web UI, not post-closing tags which
1771	** arrive via the command-line or from synchronization with a remote.
1772	*/
1773	/*
1774	** SETTING: forum-title width=20 default=Forum
1775	** This is the name or "title" of the Forum for this repository. The
1776	** default is just "Forum". But in some setups, admins might want to
1777	** change it to "Developer Forum" or "User Forum" or whatever other name
1778	** seems more appropriate for the particular usage.
1779	*/
1780
1781	/*
1782	** WEBPAGE: setup_forum
1783	**
1784	** Forum configuration and metrics.
1785	*/
1786	void forum_setup(void){
1787	/* boolean config settings specific to the forum. */
1788	static const char *azForumSettings[] = {
1789	"forum-close-policy",
1790	"forum-title",
1791	};
1792
1793	login_check_credentials();
1794	if( !g.perm.Setup ){
1795	login_needed(g.anon.Setup);
	@@ -1788,45 +1804,40 @@
1804	@ <p><a href='%R/forum'>Forum posts</a>:
1805	@ <a href='%R/timeline?y=f'>%d(nPosts)</a></p>
1806	}
1807
1808	@ <h2>Supervisors</h2>

1809	{
1810	Stmt q = empty_Stmt;

1811	db_prepare(&q, "SELECT uid, login, cap FROM user "
1812	"WHERE cap GLOB '[as6]' ORDER BY login");
1813	@ <table class='bordered'>
1814	@ <thead><tr><th>User</th><th>Capabilities</th></tr></thead>
1815	@ <tbody>
1816	while( SQLITE_ROW==db_step(&q) ){
1817	const int iUid = db_column_int(&q, 0);
1818	const char *zUser = db_column_text(&q, 1);
1819	const char *zCap = db_column_text(&q, 2);

1820	@ <tr>
1821	@ <td><a href='%R/setup_uedit?id=%d(iUid)'>%h(zUser)</a></td>
1822	@ <td>(%h(zCap))</td>
1823	@ </tr>
1824	}
1825	db_finalize(&q);
1826	@</tbody></table>





1827	}
1828
1829	@ <h2>Moderators</h2>
1830	if( db_int(0, "SELECT count(*) FROM user "
1831	" WHERE cap GLOB '5' AND cap NOT GLOB '[as6]'")==0 ){
1832	@ <p>No non-supervisor moderators
1833	}else{
1834	Stmt q = empty_Stmt;
1835	int nRows = 0;
1836	db_prepare(&q, "SELECT uid, login, cap FROM user "
1837	"WHERE cap GLOB '5' AND cap NOT GLOB '[as6]'"
1838	" ORDER BY login");
1839	@ <table class='bordered'>
1840	@ <thead><tr><th>User</th><th>Capabilities</th></tr></thead>
1841	@ <tbody>
1842	while( SQLITE_ROW==db_step(&q) ){
1843	const int iUid = db_column_int(&q, 0);
	@@ -1838,43 +1849,59 @@
1849	@ <td>(%h(zCap))</td>
1850	@ </tr>
1851	}
1852	db_finalize(&q);
1853	@ </tbody></table>





1854	}
1855
1856	@ <h2>Settings</h2>

1857	if( P("submit") && cgi_csrf_safe(2) ){
1858	int i = 0;

1859	db_begin_transaction();
1860	for(i=0; i<ArraySize(azForumSettings); i++){
1861	char zQP[4];
1862	const char *z;
1863	const Setting *pSetting = setting_find(azForumSettings[i]);
1864	if( pSetting==0 ) continue;
1865	zQP[0] = 'a'+i;
1866	zQP[1] = zQP[0];
1867	zQP[2] = 0;
1868	z = P(zQP);
1869	if( z==0 \|\| z[0]==0 ) continue;
1870	db_set(pSetting->name/works-like:"x"/, z, 0);
1871	}
1872	db_end_transaction(0);
1873	@ <p><em>Settings saved.</em></p>
1874	}
1875	{
1876	int i = 0;

1877	@ <form action="%R/setup_forum" method="post">
1878	login_insert_csrf_secret();
1879	@ <table class='forum-settings-list'><tbody>
1880	for(i=0; i<ArraySize(azForumSettings); i++){
1881	char zQP[4];
1882	const Setting *pSetting = setting_find(azForumSettings[i]);
1883	if( pSetting==0 ) continue;
1884	zQP[0] = 'a'+i;
1885	zQP[1] = zQP[0];
1886	zQP[2] = 0;
1887	if( pSetting->width==0 ){
1888	/* Boolean setting */
1889	@ <tr><td align="right">
1890	@ <a href='%R/help?cmd=%h(pSetting->name)'>%h(pSetting->name)</a>:
1891	@ </td><td>
1892	onoff_attribute("", zQP, pSetting->name/works-like:"x"/, 0, 0);
1893	@ </td></tr>
1894	}else{
1895	/* Text value setting */
1896	@ <tr><td align="right">
1897	@ <a href='%R/help?cmd=%h(pSetting->name)'>%h(pSetting->name)</a>:
1898	@ </td><td>
1899	entry_attribute("", 25, pSetting->name, zQP/works-like:""/,
1900	pSetting->def, 0);
1901	@ </td></tr>
1902	}
1903	}
1904	@ </tbody></table>
1905	@ <input type='submit' name='submit' value='Apply changes'>
1906	@ </form>
1907	}
	@@ -1909,11 +1936,12 @@
1936	login_needed(g.anon.RdForum);
1937	return;
1938	}
1939	cgi_check_for_malice();
1940	style_set_current_feature("forum");
1941	style_header("%s%s", db_get("forum-title","Forum"),
1942	isSearch ? " Search Results" : "");
1943	style_submenu_element("Timeline", "%R/timeline?ss=v&y=f&vfx");
1944	if( g.perm.WrForum ){
1945	style_submenu_element("New Thread","%R/forumnew");
1946	}else{
1947	/* Can't combine this with previous case using the ternary operator
1948

M src/fossil.diff.js

+25 -5

		--- src/fossil.diff.js
		+++ src/fossil.diff.js
		@@ -6,20 +6,39 @@
6	6	/**
7	7	Adds toggle checkboxes to each file entry in the diff views for
8	8	/info and similar pages.
9	9	*/
10	10	const D = window.fossil.dom;
	11	+ const allToggles = [/collection of all diff-toggle checkboxes /];
11	12	const addToggle = function(diffElem){
12	13	const sib = diffElem.previousElementSibling,
13		- btn = sib ? D.addClass(D.checkbox(true), 'diff-toggle') : 0;
	14	+ btnOne = sib ? D.addClass(D.checkbox(true), 'diff-toggle') : 0;
14	15	if(!sib) return;
15		- D.append(sib,btn);
16		- btn.addEventListener('click', function(){
17		- diffElem.classList.toggle('hidden');
	16	+ const lblToggle = D.append(D.label(null, " Toggle "), btnOne);
	17	+ const wrapper = D.append(D.span(), lblToggle);
	18	+ const btnAll = D.button("all");
	19	+ btnAll.$cb = btnOne;
	20	+ allToggles.push(btnOne);
	21	+ D.append(sib, D.append(wrapper, lblToggle, D.text(" "), btnAll));
	22	+ btnOne.addEventListener('change', function(){
	23	+ diffElem.classList[this.checked ? 'remove' : 'add']('hidden');
	24	+ }, false);
	25	+ btnAll.addEventListener('click', function(){
	26	+ /* Toggle all entries to match this line's new state. Note that
	27	+ we use click() instead of cb.checked=... so that the
	28	+ on-change event handler fires. */
	29	+ const checked = !this.$cb.checked;
	30	+ allToggles.forEach( (cb)=>{
	31	+ if(cb.checked!==checked) cb.click();
	32	+ });
18	33	}, false);
19	34	};
20		- document.querySelectorAll('table.diff').forEach(addToggle);
	35	+ if( !document.querySelector('body.fdiff') ){
	36	+ /* Don't show the diff toggle button for /fdiff because it only
	37	+ has a single file to show (and also a different DOM layout). */
	38	+ document.querySelectorAll('table.diff').forEach(addToggle);
	39	+ }
21	40	});
22	41
23	42	window.fossil.onPageLoad(function(){
24	43	const F = window.fossil, D = F.dom;
25	44	const Diff = F.diff = {
		@@ -638,10 +657,11 @@
638	657	const potentialParents = [ /* possible parents for the checkbox */
639	658	/* Put the most likely pages at the end, as array.pop() is more
640	659	efficient than array.shift() (see loop below). */
641	660	/* /filedit */ 'body.cpage-fileedit #fileedit-tab-diff-buttons',
642	661	/* /wikiedit */ 'body.cpage-wikiedit #wikiedit-tab-diff-buttons',
	662	+ /* /fdiff */ 'body.fdiff form div.submenu',
643	663	/* /vdiff */ 'body.vdiff form div.submenu',
644	664	/* /info, /vinfo */ 'body.vinfo div.sectionmenu.info-changes-menu'
645	665	];
646	666	while( potentialParents.length ){
647	667	if( (eToggleParent = document.querySelector(potentialParents.pop())) ){
648	668

	--- src/fossil.diff.js
	+++ src/fossil.diff.js
	@@ -6,20 +6,39 @@
6	/**
7	Adds toggle checkboxes to each file entry in the diff views for
8	/info and similar pages.
9	*/
10	const D = window.fossil.dom;

11	const addToggle = function(diffElem){
12	const sib = diffElem.previousElementSibling,
13	btn = sib ? D.addClass(D.checkbox(true), 'diff-toggle') : 0;
14	if(!sib) return;
15	D.append(sib,btn);
16	btn.addEventListener('click', function(){
17	diffElem.classList.toggle('hidden');














18	}, false);
19	};
20	document.querySelectorAll('table.diff').forEach(addToggle);




21	});
22
23	window.fossil.onPageLoad(function(){
24	const F = window.fossil, D = F.dom;
25	const Diff = F.diff = {
	@@ -638,10 +657,11 @@
638	const potentialParents = [ /* possible parents for the checkbox */
639	/* Put the most likely pages at the end, as array.pop() is more
640	efficient than array.shift() (see loop below). */
641	/* /filedit */ 'body.cpage-fileedit #fileedit-tab-diff-buttons',
642	/* /wikiedit */ 'body.cpage-wikiedit #wikiedit-tab-diff-buttons',

643	/* /vdiff */ 'body.vdiff form div.submenu',
644	/* /info, /vinfo */ 'body.vinfo div.sectionmenu.info-changes-menu'
645	];
646	while( potentialParents.length ){
647	if( (eToggleParent = document.querySelector(potentialParents.pop())) ){
648

	--- src/fossil.diff.js
	+++ src/fossil.diff.js
	@@ -6,20 +6,39 @@
6	/**
7	Adds toggle checkboxes to each file entry in the diff views for
8	/info and similar pages.
9	*/
10	const D = window.fossil.dom;
11	const allToggles = [/collection of all diff-toggle checkboxes /];
12	const addToggle = function(diffElem){
13	const sib = diffElem.previousElementSibling,
14	btnOne = sib ? D.addClass(D.checkbox(true), 'diff-toggle') : 0;
15	if(!sib) return;
16	const lblToggle = D.append(D.label(null, " Toggle "), btnOne);
17	const wrapper = D.append(D.span(), lblToggle);
18	const btnAll = D.button("all");
19	btnAll.$cb = btnOne;
20	allToggles.push(btnOne);
21	D.append(sib, D.append(wrapper, lblToggle, D.text(" "), btnAll));
22	btnOne.addEventListener('change', function(){
23	diffElem.classList[this.checked ? 'remove' : 'add']('hidden');
24	}, false);
25	btnAll.addEventListener('click', function(){
26	/* Toggle all entries to match this line's new state. Note that
27	we use click() instead of cb.checked=... so that the
28	on-change event handler fires. */
29	const checked = !this.$cb.checked;
30	allToggles.forEach( (cb)=>{
31	if(cb.checked!==checked) cb.click();
32	});
33	}, false);
34	};
35	if( !document.querySelector('body.fdiff') ){
36	/* Don't show the diff toggle button for /fdiff because it only
37	has a single file to show (and also a different DOM layout). */
38	document.querySelectorAll('table.diff').forEach(addToggle);
39	}
40	});
41
42	window.fossil.onPageLoad(function(){
43	const F = window.fossil, D = F.dom;
44	const Diff = F.diff = {
	@@ -638,10 +657,11 @@
657	const potentialParents = [ /* possible parents for the checkbox */
658	/* Put the most likely pages at the end, as array.pop() is more
659	efficient than array.shift() (see loop below). */
660	/* /filedit */ 'body.cpage-fileedit #fileedit-tab-diff-buttons',
661	/* /wikiedit */ 'body.cpage-wikiedit #wikiedit-tab-diff-buttons',
662	/* /fdiff */ 'body.fdiff form div.submenu',
663	/* /vdiff */ 'body.vdiff form div.submenu',
664	/* /info, /vinfo */ 'body.vinfo div.sectionmenu.info-changes-menu'
665	];
666	while( potentialParents.length ){
667	if( (eToggleParent = document.querySelector(potentialParents.pop())) ){
668

M src/fossil.dom.js

+3 -1

		--- src/fossil.dom.js
		+++ src/fossil.dom.js
		@@ -87,11 +87,13 @@
87	87	const rc = document.createElement('label');
88	88	if(forElem){
89	89	if(forElem instanceof HTMLElement){
90	90	forElem = this.attr(forElem, 'id');
91	91	}
92		- dom.attr(rc, 'for', forElem);
	92	+ if(forElem){
	93	+ dom.attr(rc, 'for', forElem);
	94	+ }
93	95	}
94	96	if(text) this.append(rc, text);
95	97	return rc;
96	98	};
97	99	/**
98	100

	--- src/fossil.dom.js
	+++ src/fossil.dom.js
	@@ -87,11 +87,13 @@
87	const rc = document.createElement('label');
88	if(forElem){
89	if(forElem instanceof HTMLElement){
90	forElem = this.attr(forElem, 'id');
91	}
92	dom.attr(rc, 'for', forElem);


93	}
94	if(text) this.append(rc, text);
95	return rc;
96	};
97	/**
98

	--- src/fossil.dom.js
	+++ src/fossil.dom.js
	@@ -87,11 +87,13 @@
87	const rc = document.createElement('label');
88	if(forElem){
89	if(forElem instanceof HTMLElement){
90	forElem = this.attr(forElem, 'id');
91	}
92	if(forElem){
93	dom.attr(rc, 'for', forElem);
94	}
95	}
96	if(text) this.append(rc, text);
97	return rc;
98	};
99	/**
100

M src/fossil.page.chat.js

+1 -1

		--- src/fossil.page.chat.js
		+++ src/fossil.page.chat.js
		@@ -2132,11 +2132,11 @@
2132	2132	s.value ? 'add' : 'remove'
2133	2133	]('compact');
2134	2134	Chat.e.inputFields[Chat.e.inputFields.$currentIndex].focus();
2135	2135	});
2136	2136	Chat.settings.addListener('edit-ctrl-send',function(s){
2137		- const label = (s.value ? "Ctrl-" : "")+"Enter submits message";
	2137	+ const label = (s.value ? "Ctrl-" : "")+"Enter submits message.";
2138	2138	Chat.e.inputFields.forEach((e)=>{
2139	2139	const v = e.dataset.placeholder0 + " " +label;
2140	2140	if(e.isContentEditable) e.dataset.placeholder = v;
2141	2141	else D.attr(e,'placeholder',v);
2142	2142	});
2143	2143

	--- src/fossil.page.chat.js
	+++ src/fossil.page.chat.js
	@@ -2132,11 +2132,11 @@
2132	s.value ? 'add' : 'remove'
2133	]('compact');
2134	Chat.e.inputFields[Chat.e.inputFields.$currentIndex].focus();
2135	});
2136	Chat.settings.addListener('edit-ctrl-send',function(s){
2137	const label = (s.value ? "Ctrl-" : "")+"Enter submits message";
2138	Chat.e.inputFields.forEach((e)=>{
2139	const v = e.dataset.placeholder0 + " " +label;
2140	if(e.isContentEditable) e.dataset.placeholder = v;
2141	else D.attr(e,'placeholder',v);
2142	});
2143

	--- src/fossil.page.chat.js
	+++ src/fossil.page.chat.js
	@@ -2132,11 +2132,11 @@
2132	s.value ? 'add' : 'remove'
2133	]('compact');
2134	Chat.e.inputFields[Chat.e.inputFields.$currentIndex].focus();
2135	});
2136	Chat.settings.addListener('edit-ctrl-send',function(s){
2137	const label = (s.value ? "Ctrl-" : "")+"Enter submits message.";
2138	Chat.e.inputFields.forEach((e)=>{
2139	const v = e.dataset.placeholder0 + " " +label;
2140	if(e.isContentEditable) e.dataset.placeholder = v;
2141	else D.attr(e,'placeholder',v);
2142	});
2143

M src/fossil.page.pikchrshowasm.js

+15 -3

		--- src/fossil.page.pikchrshowasm.js
		+++ src/fossil.page.pikchrshowasm.js
		@@ -206,11 +206,11 @@
206	206	const text = getCurrentText();
207	207	if(text) PS.render(text);
208	208	};
209	209	const setCurrentText = function(txt){
210	210	taInput.value = txt;
211		- renderCurrentText();
	211	+ renderCurrentText();
212	212	};
213	213	PS.e.btnRender.addEventListener('click',function(ev){
214	214	ev.preventDefault();
215	215	renderCurrentText();
216	216	},false);
		@@ -427,13 +427,25 @@
427	427	btnToggle.addEventListener('click', function(){
428	428	fs.classList.toggle('collapsed');
429	429	content.forEach((d)=>d.classList.toggle('hidden'));
430	430	}, false);
431	431	});
	432	+
	433	+ if(window.sessionStorage){
	434	+ /* If sessionStorage['pikchr-xfer'] exists and the "fromSession"
	435	+ URL argument was passed to this page, load the pikchr source
	436	+ from the session. This is used by the "open in pikchrshow"
	437	+ link in the forum. */
	438	+ const src = window.sessionStorage.getItem('pikchr-xfer');
	439	+ if( src && (new URL(self.location.href).searchParams).has('fromSession') ){
	440	+ taInput.value = src;
	441	+ window.sessionStorage.removeItem('pikchr-xfer');
	442	+ }
	443	+ }
432	444
433	445	PS.e.btnRender.click();
434		-
	446	+
435	447	/** Debounce handler for auto-rendering while typing. */
436	448	const debounceAutoRender = F.debounce(function f(){
437	449	if(!PS._isDirty) return;
438	450	const text = getCurrentText();
439	451	if(f._ === text){
		@@ -618,11 +630,11 @@
618	630	"Darlene" above aligned
619	631
620	632	# fill in content for the Alice lane
621	633	right
622	634	A1: circle rad 0.1in at end of first line + (0.2,-0.2) \
623		- fill white thickness 1.5px "1"
	635	+ fill white thickness 1.5px "1"
624	636	arrow right 50%
625	637	circle same "2"
626	638	arrow right until even with first box.e - (0.65,0.0)
627	639	ellipse "future" fit fill white height 0.2 width 0.5 thickness 1.5px
628	640	A3: circle same at A1+(0.8,-0.3) "3" fill 0xc0c0c0
629	641

	--- src/fossil.page.pikchrshowasm.js
	+++ src/fossil.page.pikchrshowasm.js
	@@ -206,11 +206,11 @@
206	const text = getCurrentText();
207	if(text) PS.render(text);
208	};
209	const setCurrentText = function(txt){
210	taInput.value = txt;
211	renderCurrentText();
212	};
213	PS.e.btnRender.addEventListener('click',function(ev){
214	ev.preventDefault();
215	renderCurrentText();
216	},false);
	@@ -427,13 +427,25 @@
427	btnToggle.addEventListener('click', function(){
428	fs.classList.toggle('collapsed');
429	content.forEach((d)=>d.classList.toggle('hidden'));
430	}, false);
431	});












432
433	PS.e.btnRender.click();
434
435	/** Debounce handler for auto-rendering while typing. */
436	const debounceAutoRender = F.debounce(function f(){
437	if(!PS._isDirty) return;
438	const text = getCurrentText();
439	if(f._ === text){
	@@ -618,11 +630,11 @@
618	"Darlene" above aligned
619
620	# fill in content for the Alice lane
621	right
622	A1: circle rad 0.1in at end of first line + (0.2,-0.2) \
623	fill white thickness 1.5px "1"
624	arrow right 50%
625	circle same "2"
626	arrow right until even with first box.e - (0.65,0.0)
627	ellipse "future" fit fill white height 0.2 width 0.5 thickness 1.5px
628	A3: circle same at A1+(0.8,-0.3) "3" fill 0xc0c0c0
629

	--- src/fossil.page.pikchrshowasm.js
	+++ src/fossil.page.pikchrshowasm.js
	@@ -206,11 +206,11 @@
206	const text = getCurrentText();
207	if(text) PS.render(text);
208	};
209	const setCurrentText = function(txt){
210	taInput.value = txt;
211	renderCurrentText();
212	};
213	PS.e.btnRender.addEventListener('click',function(ev){
214	ev.preventDefault();
215	renderCurrentText();
216	},false);
	@@ -427,13 +427,25 @@
427	btnToggle.addEventListener('click', function(){
428	fs.classList.toggle('collapsed');
429	content.forEach((d)=>d.classList.toggle('hidden'));
430	}, false);
431	});
432
433	if(window.sessionStorage){
434	/* If sessionStorage['pikchr-xfer'] exists and the "fromSession"
435	URL argument was passed to this page, load the pikchr source
436	from the session. This is used by the "open in pikchrshow"
437	link in the forum. */
438	const src = window.sessionStorage.getItem('pikchr-xfer');
439	if( src && (new URL(self.location.href).searchParams).has('fromSession') ){
440	taInput.value = src;
441	window.sessionStorage.removeItem('pikchr-xfer');
442	}
443	}
444
445	PS.e.btnRender.click();
446
447	/** Debounce handler for auto-rendering while typing. */
448	const debounceAutoRender = F.debounce(function f(){
449	if(!PS._isDirty) return;
450	const text = getCurrentText();
451	if(f._ === text){
	@@ -618,11 +630,11 @@
630	"Darlene" above aligned
631
632	# fill in content for the Alice lane
633	right
634	A1: circle rad 0.1in at end of first line + (0.2,-0.2) \
635	fill white thickness 1.5px "1"
636	arrow right 50%
637	circle same "2"
638	arrow right until even with first box.e - (0.65,0.0)
639	ellipse "future" fit fill white height 0.2 width 0.5 thickness 1.5px
640	A3: circle same at A1+(0.8,-0.3) "3" fill 0xc0c0c0
641

M src/fossil.pikchr.js

+54 -6

		--- src/fossil.pikchr.js
		+++ src/fossil.pikchr.js
		@@ -38,11 +38,14 @@
38	38	This code expects the following structure around the SVGs, and
39	39	will not process any which don't match this:
40	40
41	41	<DIV.pikchr-wrapper>
42	42	<DIV.pikchr-svg><SVG.pikchr></SVG></DIV>
43		- <PRE.pikchr-src></PRE>
	43	+ <DIV.pikchr-src>
	44	+ <PRE>pikchr source code</PRE>
	45	+ <SPAN class='hidden'><A>link to open pikchr in /pikchrshow</A></SPAN>
	46	+ </DIV>
44	47	</DIV>
45	48	*/
46	49	P.addSrcView = function f(svg){
47	50	if(!f.hasOwnProperty('parentClick')){
48	51	f.parentClick = function(ev){
		@@ -58,10 +61,46 @@
58	61	this.classList.toggle('source');
59	62	ev.stopPropagation();
60	63	ev.preventDefault();
61	64	}
62	65	};
	66	+ /**
	67	+ Event handler for the "open in pikchrshow" links: store the
	68	+ source code for the link's pikchr in
	69	+ window.sessionStorage['pikchr-xfer'] then open
	70	+ /pikchrshow?fromSession to trigger loading of that pikchr.
	71	+ */
	72	+ f.clickPikchrShow = function(ev){
	73	+ const pId = this.dataset['pikchrid'] /* ID of the associated pikchr source code element */;
	74	+ if(!pId) return;
	75	+ const ePikchr = this.parentNode.parentNode.querySelector('#'+pId);
	76	+ if(!ePikchr) return;
	77	+ ev.stopPropagation() /* keep pikchr source view from toggling */;
	78	+ window.sessionStorage.setItem('pikchr-xfer', ePikchr.innerText);
	79	+ /*
	80	+ After returning from this function the link element will
	81	+ open [/pikchrshow?fromSession], and pikchrshow will extract
	82	+ the pikchr source code from sessionStorage['pikchr-xfer'].
	83	+
	84	+ Quirks of this ^^^ design:
	85	+
	86	+ We use only a single slot in sessionStorage. We could
	87	+ alternately use a key like pikchr-$pId and pass that key on
	88	+ to /pikchrshow via fromSession=pikchr-$pId, but that would
	89	+ eventually lead to stale session entries if loading of
	90	+ pikchrshow were interrupted at an untimely point. The
	91	+ down-side of _not_ doing that is that some user (or
	92	+ automation) options multiple "open in pikchrshow" links
	93	+ rapidly enough, the will open the same pikchr (the one which
	94	+ was stored in the session's slot most recently). The
	95	+ current approach should be fine for normal human interaction
	96	+ speeds, but if it proves to be a problem we can instead use
	97	+ the above-described approach of storing each pikchr in its
	98	+ own session slot and simply accept that there may be stale
	99	+ entries at some point.
	100	+ */
	101	+ };
63	102	};
64	103	if(!svg) svg = 'svg.pikchr';
65	104	if('string' === typeof svg){
66	105	document.querySelectorAll(svg).forEach((e)=>f.call(this, e));
67	106	return this;
		@@ -71,15 +110,24 @@
71	110	}
72	111	if(svg.dataset.pikchrProcessed){
73	112	return this;
74	113	}
75	114	svg.dataset.pikchrProcessed = 1;
76		- const parent = svg.parentNode.parentNode /* outermost div.pikchr-wrapper */;
77		- const srcView = parent ? svg.parentNode.nextElementSibling : undefined;
78		- if(!srcView \|\| !srcView.classList.contains('pikchr-src')){
	115	+ const parent = svg.parentNode.parentNode /* outermost DIV.pikchr-wrapper */;
	116	+ const srcView = parent ? svg.parentNode.nextElementSibling /* DIV.pikchr-src */ : undefined;
	117	+ if(srcView && srcView.classList.contains('pikchr-src')){
79	118	/* Without this element, there's nothing for us to do here. */
80		- return this;
	119	+ parent.addEventListener('click', f.parentClick, false);
	120	+ const eSpan = window.sessionStorage
	121	+ ? srcView.querySelector('span') /* "open in..." link wrapper */
	122	+ : undefined;
	123	+ if(eSpan){
	124	+ const openLink = eSpan.querySelector('a');
	125	+ if(openLink){
	126	+ openLink.addEventListener('click', f.clickPikchrShow, false);
	127	+ eSpan.classList.remove('hidden');
	128	+ }
	129	+ }
81	130	}
82		- parent.addEventListener('click', f.parentClick, false);
83	131	return this;
84	132	};
85	133	})(window.fossil);
86	134

	--- src/fossil.pikchr.js
	+++ src/fossil.pikchr.js
	@@ -38,11 +38,14 @@
38	This code expects the following structure around the SVGs, and
39	will not process any which don't match this:
40
41	<DIV.pikchr-wrapper>
42	<DIV.pikchr-svg><SVG.pikchr></SVG></DIV>
43	<PRE.pikchr-src></PRE>



44	</DIV>
45	*/
46	P.addSrcView = function f(svg){
47	if(!f.hasOwnProperty('parentClick')){
48	f.parentClick = function(ev){
	@@ -58,10 +61,46 @@
58	this.classList.toggle('source');
59	ev.stopPropagation();
60	ev.preventDefault();
61	}
62	};




































63	};
64	if(!svg) svg = 'svg.pikchr';
65	if('string' === typeof svg){
66	document.querySelectorAll(svg).forEach((e)=>f.call(this, e));
67	return this;
	@@ -71,15 +110,24 @@
71	}
72	if(svg.dataset.pikchrProcessed){
73	return this;
74	}
75	svg.dataset.pikchrProcessed = 1;
76	const parent = svg.parentNode.parentNode /* outermost div.pikchr-wrapper */;
77	const srcView = parent ? svg.parentNode.nextElementSibling : undefined;
78	if(!srcView \|\| !srcView.classList.contains('pikchr-src')){
79	/* Without this element, there's nothing for us to do here. */
80	return this;










81	}
82	parent.addEventListener('click', f.parentClick, false);
83	return this;
84	};
85	})(window.fossil);
86

	--- src/fossil.pikchr.js
	+++ src/fossil.pikchr.js
	@@ -38,11 +38,14 @@
38	This code expects the following structure around the SVGs, and
39	will not process any which don't match this:
40
41	<DIV.pikchr-wrapper>
42	<DIV.pikchr-svg><SVG.pikchr></SVG></DIV>
43	<DIV.pikchr-src>
44	<PRE>pikchr source code</PRE>
45	<SPAN class='hidden'><A>link to open pikchr in /pikchrshow</A></SPAN>
46	</DIV>
47	</DIV>
48	*/
49	P.addSrcView = function f(svg){
50	if(!f.hasOwnProperty('parentClick')){
51	f.parentClick = function(ev){
	@@ -58,10 +61,46 @@
61	this.classList.toggle('source');
62	ev.stopPropagation();
63	ev.preventDefault();
64	}
65	};
66	/**
67	Event handler for the "open in pikchrshow" links: store the
68	source code for the link's pikchr in
69	window.sessionStorage['pikchr-xfer'] then open
70	/pikchrshow?fromSession to trigger loading of that pikchr.
71	*/
72	f.clickPikchrShow = function(ev){
73	const pId = this.dataset['pikchrid'] /* ID of the associated pikchr source code element */;
74	if(!pId) return;
75	const ePikchr = this.parentNode.parentNode.querySelector('#'+pId);
76	if(!ePikchr) return;
77	ev.stopPropagation() /* keep pikchr source view from toggling */;
78	window.sessionStorage.setItem('pikchr-xfer', ePikchr.innerText);
79	/*
80	After returning from this function the link element will
81	open [/pikchrshow?fromSession], and pikchrshow will extract
82	the pikchr source code from sessionStorage['pikchr-xfer'].
83
84	Quirks of this ^^^ design:
85
86	We use only a single slot in sessionStorage. We could
87	alternately use a key like pikchr-$pId and pass that key on
88	to /pikchrshow via fromSession=pikchr-$pId, but that would
89	eventually lead to stale session entries if loading of
90	pikchrshow were interrupted at an untimely point. The
91	down-side of _not_ doing that is that some user (or
92	automation) options multiple "open in pikchrshow" links
93	rapidly enough, the will open the same pikchr (the one which
94	was stored in the session's slot most recently). The
95	current approach should be fine for normal human interaction
96	speeds, but if it proves to be a problem we can instead use
97	the above-described approach of storing each pikchr in its
98	own session slot and simply accept that there may be stale
99	entries at some point.
100	*/
101	};
102	};
103	if(!svg) svg = 'svg.pikchr';
104	if('string' === typeof svg){
105	document.querySelectorAll(svg).forEach((e)=>f.call(this, e));
106	return this;
	@@ -71,15 +110,24 @@
110	}
111	if(svg.dataset.pikchrProcessed){
112	return this;
113	}
114	svg.dataset.pikchrProcessed = 1;
115	const parent = svg.parentNode.parentNode /* outermost DIV.pikchr-wrapper */;
116	const srcView = parent ? svg.parentNode.nextElementSibling /* DIV.pikchr-src */ : undefined;
117	if(srcView && srcView.classList.contains('pikchr-src')){
118	/* Without this element, there's nothing for us to do here. */
119	parent.addEventListener('click', f.parentClick, false);
120	const eSpan = window.sessionStorage
121	? srcView.querySelector('span') /* "open in..." link wrapper */
122	: undefined;
123	if(eSpan){
124	const openLink = eSpan.querySelector('a');
125	if(openLink){
126	openLink.addEventListener('click', f.clickPikchrShow, false);
127	eSpan.classList.remove('hidden');
128	}
129	}
130	}

131	return this;
132	};
133	})(window.fossil);
134

M src/glob.c

+2 -2

		--- src/glob.c
		+++ src/glob.c
		@@ -34,12 +34,12 @@
34	34	** Commas and whitespace are considered to be element delimters. Each
35	35	** element of the GLOB list may optionally be enclosed in either '...' or
36	36	** "...". This allows commas and/or whitespace to be used in the elements
37	37	** themselves.
38	38	**
39		-** This routine makes no effort to free the memory space it uses, which
40		-** currently consists of a blob object and its contents.
	39	+** The returned string is owned by the caller, who must fossil_free()
	40	+** it.
41	41	*/
42	42	char glob_expr(const char zVal, const char *zGlobList){
43	43	Blob expr;
44	44	const char *zSep = "(";
45	45	int nTerm = 0;
46	46

	--- src/glob.c
	+++ src/glob.c
	@@ -34,12 +34,12 @@
34	** Commas and whitespace are considered to be element delimters. Each
35	** element of the GLOB list may optionally be enclosed in either '...' or
36	** "...". This allows commas and/or whitespace to be used in the elements
37	** themselves.
38	**
39	** This routine makes no effort to free the memory space it uses, which
40	** currently consists of a blob object and its contents.
41	*/
42	char glob_expr(const char zVal, const char *zGlobList){
43	Blob expr;
44	const char *zSep = "(";
45	int nTerm = 0;
46

	--- src/glob.c
	+++ src/glob.c
	@@ -34,12 +34,12 @@
34	** Commas and whitespace are considered to be element delimters. Each
35	** element of the GLOB list may optionally be enclosed in either '...' or
36	** "...". This allows commas and/or whitespace to be used in the elements
37	** themselves.
38	**
39	** The returned string is owned by the caller, who must fossil_free()
40	** it.
41	*/
42	char glob_expr(const char zVal, const char *zGlobList){
43	Blob expr;
44	const char *zSep = "(";
45	int nTerm = 0;
46

M src/http.c

		--- src/http.c
		+++ src/http.c
		@@ -768,10 +768,11 @@
768	768	** a GET request where there is no PAYLOAD.
769	769	**
770	770	** Options:
771	771	** --compress Use ZLIB compression on the payload
772	772	** --mimetype TYPE Mimetype of the payload
	773	+** --no-cert-verify Disable TLS cert verification
773	774	** --out FILE Store the reply in FILE
774	775	** -v Verbose output
775	776	** --xfer PAYLOAD in a Fossil xfer protocol message
776	777	*/
777	778	void test_httpmsg_command(void){
		@@ -783,10 +784,15 @@
783	784
784	785	zMimetype = find_option("mimetype",0,1);
785	786	zOutFile = find_option("out","o",1);
786	787	if( find_option("verbose","v",0)!=0 ) mHttpFlags \|= HTTP_VERBOSE;
787	788	if( find_option("compress",0,0)!=0 ) mHttpFlags &= ~HTTP_NOCOMPRESS;
	789	+ if( find_option("no-cert-verify",0,0)!=0 ){
	790	+ #ifdef FOSSIL_ENABLE_SSL
	791	+ ssl_disable_cert_verification();
	792	+ #endif
	793	+ }
788	794	if( find_option("xfer",0,0)!=0 ){
789	795	mHttpFlags \|= HTTP_USE_LOGIN;
790	796	mHttpFlags &= ~HTTP_GENERIC;
791	797	}
792	798	verify_all_options();
793	799

	--- src/http.c
	+++ src/http.c
	@@ -768,10 +768,11 @@
768	** a GET request where there is no PAYLOAD.
769	**
770	** Options:
771	** --compress Use ZLIB compression on the payload
772	** --mimetype TYPE Mimetype of the payload

773	** --out FILE Store the reply in FILE
774	** -v Verbose output
775	** --xfer PAYLOAD in a Fossil xfer protocol message
776	*/
777	void test_httpmsg_command(void){
	@@ -783,10 +784,15 @@
783
784	zMimetype = find_option("mimetype",0,1);
785	zOutFile = find_option("out","o",1);
786	if( find_option("verbose","v",0)!=0 ) mHttpFlags \|= HTTP_VERBOSE;
787	if( find_option("compress",0,0)!=0 ) mHttpFlags &= ~HTTP_NOCOMPRESS;





788	if( find_option("xfer",0,0)!=0 ){
789	mHttpFlags \|= HTTP_USE_LOGIN;
790	mHttpFlags &= ~HTTP_GENERIC;
791	}
792	verify_all_options();
793

	--- src/http.c
	+++ src/http.c
	@@ -768,10 +768,11 @@
768	** a GET request where there is no PAYLOAD.
769	**
770	** Options:
771	** --compress Use ZLIB compression on the payload
772	** --mimetype TYPE Mimetype of the payload
773	** --no-cert-verify Disable TLS cert verification
774	** --out FILE Store the reply in FILE
775	** -v Verbose output
776	** --xfer PAYLOAD in a Fossil xfer protocol message
777	*/
778	void test_httpmsg_command(void){
	@@ -783,10 +784,15 @@
784
785	zMimetype = find_option("mimetype",0,1);
786	zOutFile = find_option("out","o",1);
787	if( find_option("verbose","v",0)!=0 ) mHttpFlags \|= HTTP_VERBOSE;
788	if( find_option("compress",0,0)!=0 ) mHttpFlags &= ~HTTP_NOCOMPRESS;
789	if( find_option("no-cert-verify",0,0)!=0 ){
790	#ifdef FOSSIL_ENABLE_SSL
791	ssl_disable_cert_verification();
792	#endif
793	}
794	if( find_option("xfer",0,0)!=0 ){
795	mHttpFlags \|= HTTP_USE_LOGIN;
796	mHttpFlags &= ~HTTP_GENERIC;
797	}
798	verify_all_options();
799

M src/http_ssl.c

+4 -6

		--- src/http_ssl.c
		+++ src/http_ssl.c
		@@ -403,16 +403,14 @@
403	403	blob_append(&reply, buf, len);
404	404
405	405	bbuf = blob_buffer(&reply);
406	406	len = blob_size(&reply);
407	407	while(end < len) {
408		- if(bbuf[end] == '\r') {
409		- if(len - end < 4) {
410		- /* need more data */
411		- break;
412		- }
413		- if(memcmp(&bbuf[end], "\r\n\r\n", 4) == 0) {
	408	+ if( bbuf[end]=='\n' ) {
	409	+ if( (end+1<len && bbuf[end+1]=='\n')
	410	+ \|\| (end+2<len && bbuf[end+1]=='\r' && bbuf[end+2]=='\n')
	411	+ ){
414	412	done = 1;
415	413	break;
416	414	}
417	415	}
418	416	end++;
419	417

	--- src/http_ssl.c
	+++ src/http_ssl.c
	@@ -403,16 +403,14 @@
403	blob_append(&reply, buf, len);
404
405	bbuf = blob_buffer(&reply);
406	len = blob_size(&reply);
407	while(end < len) {
408	if(bbuf[end] == '\r') {
409	if(len - end < 4) {
410	/* need more data */
411	break;
412	}
413	if(memcmp(&bbuf[end], "\r\n\r\n", 4) == 0) {
414	done = 1;
415	break;
416	}
417	}
418	end++;
419

	--- src/http_ssl.c
	+++ src/http_ssl.c
	@@ -403,16 +403,14 @@
403	blob_append(&reply, buf, len);
404
405	bbuf = blob_buffer(&reply);
406	len = blob_size(&reply);
407	while(end < len) {
408	if( bbuf[end]=='\n' ) {
409	if( (end+1<len && bbuf[end+1]=='\n')
410	\|\| (end+2<len && bbuf[end+1]=='\r' && bbuf[end+2]=='\n')
411	){


412	done = 1;
413	break;
414	}
415	}
416	end++;
417

M src/http_transport.c

+4 -1

		--- src/http_transport.c
		+++ src/http_transport.c
		@@ -103,11 +103,14 @@
103	103	/*
104	104	** Initialize a Blob to the name of the configured SSH command.
105	105	*/
106	106	void transport_ssh_command(Blob *p){
107	107	char zSsh; / The base SSH command */
108		- zSsh = db_get("ssh-command", zDefaultSshCmd);
	108	+ zSsh = g.zSshCmd;
	109	+ if( zSsh==0 \|\| zSsh[0]==0 ){
	110	+ zSsh = db_get("ssh-command", zDefaultSshCmd);
	111	+ }
109	112	blob_init(p, zSsh, -1);
110	113	}
111	114
112	115	/*
113	116	** SSH initialization of the transport layer
114	117

	--- src/http_transport.c
	+++ src/http_transport.c
	@@ -103,11 +103,14 @@
103	/*
104	** Initialize a Blob to the name of the configured SSH command.
105	*/
106	void transport_ssh_command(Blob *p){
107	char zSsh; / The base SSH command */
108	zSsh = db_get("ssh-command", zDefaultSshCmd);



109	blob_init(p, zSsh, -1);
110	}
111
112	/*
113	** SSH initialization of the transport layer
114

	--- src/http_transport.c
	+++ src/http_transport.c
	@@ -103,11 +103,14 @@
103	/*
104	** Initialize a Blob to the name of the configured SSH command.
105	*/
106	void transport_ssh_command(Blob *p){
107	char zSsh; / The base SSH command */
108	zSsh = g.zSshCmd;
109	if( zSsh==0 \|\| zSsh[0]==0 ){
110	zSsh = db_get("ssh-command", zDefaultSshCmd);
111	}
112	blob_init(p, zSsh, -1);
113	}
114
115	/*
116	** SSH initialization of the transport layer
117

M src/info.c

+152 -17

		--- src/info.c
		+++ src/info.c
		@@ -372,11 +372,13 @@
372	372	const char zNew, / blob.uuid after change. NULL for deletes */
373	373	const char zOldName, / Prior name. NULL if no name change. */
374	374	DiffConfig pCfg, / Flags for text_diff() or NULL to omit all */
375	375	int mperm /* executable or symlink permission for zNew */
376	376	){
377		- @ <p>
	377	+ @ <div class='file-change-line'><span>
	378	+ /* Maintenance reminder: the extra level of SPAN is for
	379	+ ** arranging new elements via JS. */
378	380	if( !g.perm.Hyperlink ){
379	381	if( zNew==0 ){
380	382	@ Deleted %h(zName).
381	383	}else if( zOld==0 ){
382	384	@ Added %h(zName).
		@@ -391,10 +393,11 @@
391	393	@ %h(zName) became a regular file.
392	394	}
393	395	}else{
394	396	@ Changes to %h(zName).
395	397	}
	398	+ @ </span></div>
396	399	if( pCfg ){
397	400	append_diff(zOld, zNew, pCfg);
398	401	}
399	402	}else{
400	403	if( zOld && zNew ){
		@@ -436,18 +439,22 @@
436	439	@ %h(zName)</a> version %z(href("%R/artifact/%!S",zOld))[%S(zOld)]</a>.
437	440	}else{
438	441	@ Added %z(href("%R/finfo?name=%T&m=%!S&ci=%!S",zName,zNew,zCkin))\
439	442	@ %h(zName)</a> version %z(href("%R/artifact/%!S",zNew))[%S(zNew)]</a>.
440	443	}
441		- if( pCfg ){
442		- append_diff(zOld, zNew, pCfg);
443		- }else if( zOld && zNew && fossil_strcmp(zOld,zNew)!=0 ){
444		- @
445		- @ %z(href("%R/fdiff?v1=%!S&v2=%!S",zOld,zNew))[diff]</a>
	444	+ if( zOld && zNew && fossil_strcmp(zOld,zNew)!=0 ){
	445	+ if( pCfg ){
	446	+ @ </span></div>
	447	+ append_diff(zOld, zNew, pCfg);
	448	+ }else{
	449	+ @ %z(href("%R/fdiff?v1=%!S&v2=%!S",zOld,zNew))[diff]</a>
	450	+ @ </span></div>
	451	+ }
	452	+ }else{
	453	+ @ </span></div>
446	454	}
447	455	}
448		- @ </p>
449	456	}
450	457
451	458	/*
452	459	** Generate javascript to enhance HTML diffs.
453	460	*/
		@@ -600,10 +607,130 @@
600	607	www_print_timeline(&q, TIMELINE_DISJOINT\|TIMELINE_GRAPH\|TIMELINE_NOSCROLL,
601	608	0, 0, 0, rid, 0, 0);
602	609	db_finalize(&q);
603	610	style_finish_page();
604	611	}
	612	+
	613	+/*
	614	+** WEBPAGE: ckout
	615	+**
	616	+** Show information about the current checkout. This page only functions
	617	+** if the web server is run on a loopback interface (in other words, was
	618	+** started using "fossil ui" or similar) from with on open check-out.
	619	+*/
	620	+void ckout_page(void){
	621	+ int vid;
	622	+ char *zHostname;
	623	+ char *zCwd;
	624	+ int diffType; /* 0: no diff, 1: unified, 2: side-by-side */
	625	+ DiffConfig DCfg,pCfg; / Diff details */
	626	+ const char zHome; / Home directory */
	627	+ Stmt q;
	628	+
	629	+ if( !db_open_local(0) \|\| !cgi_is_loopback(g.zIpAddr) ){
	630	+ cgi_redirect("%R/home");
	631	+ return;
	632	+ }
	633	+ diffType = preferred_diff_type();
	634	+ pCfg = construct_diff_flags(diffType, &DCfg);
	635	+ vid = db_lget_int("checkout", 0);
	636	+ db_unprotect(PROTECT_ALL);
	637	+ vfile_check_signature(vid, CKSIG_ENOTFILE);
	638	+ db_protect_pop();
	639	+ style_set_current_feature("vinfo");
	640	+ zHostname = fossil_hostname();
	641	+ zCwd = file_getcwd(0,0);
	642	+ zHome = fossil_getenv("HOME");
	643	+ if( zHome ){
	644	+ int nHome = (int)strlen(zHome);
	645	+ if( strncmp(zCwd, zHome, nHome)==0 && zCwd[nHome]=='/' ){
	646	+ zCwd = mprintf("~%s", zCwd+nHome);
	647	+ }
	648	+ }
	649	+ if( zHostname ){
	650	+ style_header("Checkout Status: %h on %h", zCwd, zHostname);
	651	+ }else{
	652	+ style_header("Checkout Status: %h", zCwd);
	653	+ }
	654	+ render_checkin_context(vid, 0, 0, 0);
	655	+ if( pCfg==0 ){
	656	+ style_finish_page();
	657	+ return;
	658	+ }
	659	+ db_prepare(&q,
	660	+ /* 0 1 2 3 4 5 6 */
	661	+ "SELECT pathname, deleted, chnged , rid==0, rid, islink, uuid"
	662	+ " FROM vfile LEFT JOIN blob USING(rid)"
	663	+ " WHERE vid=%d"
	664	+ " AND (deleted OR chnged OR rid==0)"
	665	+ " ORDER BY pathname /scan/",
	666	+ vid
	667	+ );
	668	+ if( pCfg->diffFlags & DIFF_SIDEBYSIDE ){
	669	+ pCfg->diffFlags \|= DIFF_HTML \| DIFF_NOTTOOBIG;
	670	+ }else{
	671	+ pCfg->diffFlags \|= DIFF_LINENO \| DIFF_HTML \| DIFF_NOTTOOBIG;
	672	+ }
	673	+ while( db_step(&q)==SQLITE_ROW ){
	674	+ const char *zTreename = db_column_text(&q,0);
	675	+ int isDeleted = db_column_int(&q, 1);
	676	+ int isChnged = db_column_int(&q,2);
	677	+ int isNew = db_column_int(&q,3);
	678	+ int srcid = db_column_int(&q, 4);
	679	+ int isLink = db_column_int(&q, 5);
	680	+ const char *zUuid = db_column_text(&q, 6);
	681	+ int showDiff = 1;
	682	+
	683	+ pCfg->diffFlags &= (~DIFF_FILE_MASK);
	684	+ if( isDeleted ){
	685	+ @ <p>DELETED %h(zTreename)</p>
	686	+ pCfg->diffFlags \|= DIFF_FILE_DELETED;
	687	+ showDiff = 0;
	688	+ }else if( file_access(zTreename, F_OK) ){
	689	+ @ <p>MISSING %h(zTreename)</p>
	690	+ showDiff = 0;
	691	+ }else if( isNew ){
	692	+ @ <p>ADDED %h(zTreename)</p>
	693	+ pCfg->diffFlags \|= DIFF_FILE_ADDED;
	694	+ srcid = 0;
	695	+ showDiff = 0;
	696	+ }else if( isChnged==3 ){
	697	+ @ <p>ADDED_BY_MERGE %h(zTreename)</p>
	698	+ pCfg->diffFlags \|= DIFF_FILE_ADDED;
	699	+ srcid = 0;
	700	+ showDiff = 0;
	701	+ }else if( isChnged==5 ){
	702	+ @ <p>ADDED_BY_INTEGRATE %h(zTreename)</p>
	703	+ pCfg->diffFlags \|= DIFF_FILE_ADDED;
	704	+ srcid = 0;
	705	+ showDiff = 0;
	706	+ }else{
	707	+ @ <p>CHANGED %h(zTreename)</p>
	708	+ }
	709	+ if( showDiff ){
	710	+ Blob old, new;
	711	+ if( !isLink != !file_islink(zTreename) ){
	712	+ @ %s(DIFF_CANNOT_COMPUTE_SYMLINK)
	713	+ continue;
	714	+ }
	715	+ if( srcid>0 ){
	716	+ content_get(srcid, &old);
	717	+ pCfg->zLeftHash = zUuid;
	718	+ }else{
	719	+ blob_zero(&old);
	720	+ pCfg->zLeftHash = 0;
	721	+ }
	722	+ blob_read_from_file(&new, zTreename, ExtFILE);
	723	+ text_diff(&old, &new, cgi_output_blob(), pCfg);
	724	+ blob_reset(&old);
	725	+ blob_reset(&new);
	726	+ }
	727	+ }
	728	+ db_finalize(&q);
	729	+ append_diff_javascript(diffType);
	730	+ style_finish_page();
	731	+}
605	732
606	733	/*
607	734	** WEBPAGE: vinfo
608	735	** WEBPAGE: ci
609	736	** URL: /ci/ARTIFACTID
		@@ -1191,22 +1318,25 @@
1191	1318	const char *zBranch;
1192	1319	const char *zFrom;
1193	1320	const char *zTo;
1194	1321	const char *zRe;
1195	1322	const char *zGlob;
	1323	+ Glob * pGlob = 0;
1196	1324	char *zMergeOrigin = 0;
1197	1325	ReCompiled *pRe = 0;
1198	1326	DiffConfig DCfg, *pCfg = 0;
1199	1327	int graphFlags = 0;
1200		- Blob qp;
	1328	+ Blob qp; /* non-glob= query parameters for generated links */
	1329	+ Blob qpGlob; /* glob= query parameter for generated links */
1201	1330	int bInvert = PB("inv");
1202	1331
1203	1332	login_check_credentials();
1204	1333	if( !g.perm.Read ){ login_needed(g.anon.Read); return; }
1205	1334	login_anonymous_available();
1206	1335	fossil_nice_default();
1207	1336	blob_init(&qp, 0, 0);
	1337	+ blob_init(&qpGlob, 0, 0);
1208	1338	diffType = preferred_diff_type();
1209	1339	zRe = P("regex");
1210	1340	if( zRe ) re_compile(&pRe, zRe, 0);
1211	1341	zBranch = P("branch");
1212	1342	if( zBranch && zBranch[0]==0 ) zBranch = 0;
		@@ -1246,11 +1376,12 @@
1246	1376	}
1247	1377	if( zGlob ){
1248	1378	if( !*zGlob ){
1249	1379	zGlob = NULL;
1250	1380	}else{
1251		- blob_appendf(&qp, "&glob=%T", zGlob);
	1381	+ blob_appendf(&qpGlob, "&glob=%T", zGlob);
	1382	+ pGlob = glob_create(zGlob);
1252	1383	}
1253	1384	}
1254	1385	if( PB("nc") ){
1255	1386	graphFlags \|= TIMELINE_NOCOLOR;
1256	1387	blob_appendf(&qp, "&nc");
		@@ -1263,20 +1394,22 @@
1263	1394	style_set_current_feature("vdiff");
1264	1395	if( zBranch==0 ){
1265	1396	style_submenu_element("Path", "%R/timeline?me=%T&you=%T", zFrom, zTo);
1266	1397	}
1267	1398	if( diffType!=0 ){
1268		- style_submenu_element("Hide Diff", "%R/vdiff?diff=0&%b", &qp);
	1399	+ style_submenu_element("Hide Diff", "%R/vdiff?diff=0&%b%b", &qp, &qpGlob);
1269	1400	}
1270	1401	if( diffType!=2 ){
1271		- style_submenu_element("Side-by-Side Diff", "%R/vdiff?diff=2&%b", &qp);
	1402	+ style_submenu_element("Side-by-Side Diff", "%R/vdiff?diff=2&%b%b", &qp,
	1403	+ &qpGlob);
1272	1404	}
1273	1405	if( diffType!=1 ) {
1274		- style_submenu_element("Unified Diff", "%R/vdiff?diff=1&%b", &qp);
	1406	+ style_submenu_element("Unified Diff", "%R/vdiff?diff=1&%b%b", &qp, &qpGlob);
1275	1407	}
1276	1408	if( zBranch==0 ){
1277		- style_submenu_element("Invert","%R/vdiff?diff=%d&inv&%b", diffType, &qp);
	1409	+ style_submenu_element("Invert","%R/vdiff?diff=%d&inv&%b%b", diffType,
	1410	+ &qp, &qpGlob);
1278	1411	}
1279	1412	if( zGlob ){
1280	1413	style_submenu_element("Clear glob", "%R/vdiff?diff=%d&%b", diffType, &qp);
1281	1414	}else{
1282	1415	style_submenu_element("Patch", "%R/vpatch?from=%T&to=%T%s", zFrom, zTo,
		@@ -1321,10 +1454,11 @@
1321	1454	@ <p><b>Only files matching the glob "%h(zGlob)" are shown.</b></p>
1322	1455	}
1323	1456	@<hr><p>
1324	1457	}
1325	1458	blob_reset(&qp);
	1459	+ blob_reset(&qpGlob);
1326	1460
1327	1461	manifest_file_rewind(pFrom);
1328	1462	pFileFrom = manifest_file_next(pFrom, 0);
1329	1463	manifest_file_rewind(pTo);
1330	1464	pFileTo = manifest_file_next(pTo, 0);
		@@ -1337,37 +1471,38 @@
1337	1471	cmp = -1;
1338	1472	}else{
1339	1473	cmp = fossil_strcmp(pFileFrom->zName, pFileTo->zName);
1340	1474	}
1341	1475	if( cmp<0 ){
1342		- if( !zGlob \|\| sqlite3_strglob(zGlob, pFileFrom->zName)==0 ){
	1476	+ if( !pGlob \|\| glob_match(pGlob, pFileFrom->zName) ){
1343	1477	append_file_change_line(zFrom, pFileFrom->zName,
1344	1478	pFileFrom->zUuid, 0, 0, pCfg, 0);
1345	1479	}
1346	1480	pFileFrom = manifest_file_next(pFrom, 0);
1347	1481	}else if( cmp>0 ){
1348		- if( !zGlob \|\| sqlite3_strglob(zGlob, pFileTo->zName)==0 ){
	1482	+ if( !pGlob \|\| glob_match(pGlob, pFileTo->zName) ){
1349	1483	append_file_change_line(zTo, pFileTo->zName,
1350	1484	0, pFileTo->zUuid, 0, pCfg,
1351	1485	manifest_file_mperm(pFileTo));
1352	1486	}
1353	1487	pFileTo = manifest_file_next(pTo, 0);
1354	1488	}else if( fossil_strcmp(pFileFrom->zUuid, pFileTo->zUuid)==0 ){
1355	1489	pFileFrom = manifest_file_next(pFrom, 0);
1356	1490	pFileTo = manifest_file_next(pTo, 0);
1357	1491	}else{
1358		- if(!zGlob \|\| (sqlite3_strglob(zGlob, pFileFrom->zName)==0
1359		- \|\| sqlite3_strglob(zGlob, pFileTo->zName)==0) ){
	1492	+ if(!pGlob \|\| (glob_match(pGlob, pFileFrom->zName)
	1493	+ \|\| glob_match(pGlob, pFileTo->zName)) ){
1360	1494	append_file_change_line(zFrom, pFileFrom->zName,
1361	1495	pFileFrom->zUuid,
1362	1496	pFileTo->zUuid, 0, pCfg,
1363	1497	manifest_file_mperm(pFileTo));
1364	1498	}
1365	1499	pFileFrom = manifest_file_next(pFrom, 0);
1366	1500	pFileTo = manifest_file_next(pTo, 0);
1367	1501	}
1368	1502	}
	1503	+ glob_free(pGlob);
1369	1504	manifest_destroy(pFrom);
1370	1505	manifest_destroy(pTo);
1371	1506	append_diff_javascript(diffType);
1372	1507	style_finish_page();
1373	1508	}
1374	1509

	--- src/info.c
	+++ src/info.c
	@@ -372,11 +372,13 @@
372	const char zNew, / blob.uuid after change. NULL for deletes */
373	const char zOldName, / Prior name. NULL if no name change. */
374	DiffConfig pCfg, / Flags for text_diff() or NULL to omit all */
375	int mperm /* executable or symlink permission for zNew */
376	){
377	@ <p>


378	if( !g.perm.Hyperlink ){
379	if( zNew==0 ){
380	@ Deleted %h(zName).
381	}else if( zOld==0 ){
382	@ Added %h(zName).
	@@ -391,10 +393,11 @@
391	@ %h(zName) became a regular file.
392	}
393	}else{
394	@ Changes to %h(zName).
395	}

396	if( pCfg ){
397	append_diff(zOld, zNew, pCfg);
398	}
399	}else{
400	if( zOld && zNew ){
	@@ -436,18 +439,22 @@
436	@ %h(zName)</a> version %z(href("%R/artifact/%!S",zOld))[%S(zOld)]</a>.
437	}else{
438	@ Added %z(href("%R/finfo?name=%T&m=%!S&ci=%!S",zName,zNew,zCkin))\
439	@ %h(zName)</a> version %z(href("%R/artifact/%!S",zNew))[%S(zNew)]</a>.
440	}
441	if( pCfg ){
442	append_diff(zOld, zNew, pCfg);
443	}else if( zOld && zNew && fossil_strcmp(zOld,zNew)!=0 ){
444	@
445	@ %z(href("%R/fdiff?v1=%!S&v2=%!S",zOld,zNew))[diff]</a>





446	}
447	}
448	@ </p>
449	}
450
451	/*
452	** Generate javascript to enhance HTML diffs.
453	*/
	@@ -600,10 +607,130 @@
600	www_print_timeline(&q, TIMELINE_DISJOINT\|TIMELINE_GRAPH\|TIMELINE_NOSCROLL,
601	0, 0, 0, rid, 0, 0);
602	db_finalize(&q);
603	style_finish_page();
604	}
























































































































605
606	/*
607	** WEBPAGE: vinfo
608	** WEBPAGE: ci
609	** URL: /ci/ARTIFACTID
	@@ -1191,22 +1318,25 @@
1191	const char *zBranch;
1192	const char *zFrom;
1193	const char *zTo;
1194	const char *zRe;
1195	const char *zGlob;

1196	char *zMergeOrigin = 0;
1197	ReCompiled *pRe = 0;
1198	DiffConfig DCfg, *pCfg = 0;
1199	int graphFlags = 0;
1200	Blob qp;

1201	int bInvert = PB("inv");
1202
1203	login_check_credentials();
1204	if( !g.perm.Read ){ login_needed(g.anon.Read); return; }
1205	login_anonymous_available();
1206	fossil_nice_default();
1207	blob_init(&qp, 0, 0);

1208	diffType = preferred_diff_type();
1209	zRe = P("regex");
1210	if( zRe ) re_compile(&pRe, zRe, 0);
1211	zBranch = P("branch");
1212	if( zBranch && zBranch[0]==0 ) zBranch = 0;
	@@ -1246,11 +1376,12 @@
1246	}
1247	if( zGlob ){
1248	if( !*zGlob ){
1249	zGlob = NULL;
1250	}else{
1251	blob_appendf(&qp, "&glob=%T", zGlob);

1252	}
1253	}
1254	if( PB("nc") ){
1255	graphFlags \|= TIMELINE_NOCOLOR;
1256	blob_appendf(&qp, "&nc");
	@@ -1263,20 +1394,22 @@
1263	style_set_current_feature("vdiff");
1264	if( zBranch==0 ){
1265	style_submenu_element("Path", "%R/timeline?me=%T&you=%T", zFrom, zTo);
1266	}
1267	if( diffType!=0 ){
1268	style_submenu_element("Hide Diff", "%R/vdiff?diff=0&%b", &qp);
1269	}
1270	if( diffType!=2 ){
1271	style_submenu_element("Side-by-Side Diff", "%R/vdiff?diff=2&%b", &qp);

1272	}
1273	if( diffType!=1 ) {
1274	style_submenu_element("Unified Diff", "%R/vdiff?diff=1&%b", &qp);
1275	}
1276	if( zBranch==0 ){
1277	style_submenu_element("Invert","%R/vdiff?diff=%d&inv&%b", diffType, &qp);

1278	}
1279	if( zGlob ){
1280	style_submenu_element("Clear glob", "%R/vdiff?diff=%d&%b", diffType, &qp);
1281	}else{
1282	style_submenu_element("Patch", "%R/vpatch?from=%T&to=%T%s", zFrom, zTo,
	@@ -1321,10 +1454,11 @@
1321	@ <p><b>Only files matching the glob "%h(zGlob)" are shown.</b></p>
1322	}
1323	@<hr><p>
1324	}
1325	blob_reset(&qp);

1326
1327	manifest_file_rewind(pFrom);
1328	pFileFrom = manifest_file_next(pFrom, 0);
1329	manifest_file_rewind(pTo);
1330	pFileTo = manifest_file_next(pTo, 0);
	@@ -1337,37 +1471,38 @@
1337	cmp = -1;
1338	}else{
1339	cmp = fossil_strcmp(pFileFrom->zName, pFileTo->zName);
1340	}
1341	if( cmp<0 ){
1342	if( !zGlob \|\| sqlite3_strglob(zGlob, pFileFrom->zName)==0 ){
1343	append_file_change_line(zFrom, pFileFrom->zName,
1344	pFileFrom->zUuid, 0, 0, pCfg, 0);
1345	}
1346	pFileFrom = manifest_file_next(pFrom, 0);
1347	}else if( cmp>0 ){
1348	if( !zGlob \|\| sqlite3_strglob(zGlob, pFileTo->zName)==0 ){
1349	append_file_change_line(zTo, pFileTo->zName,
1350	0, pFileTo->zUuid, 0, pCfg,
1351	manifest_file_mperm(pFileTo));
1352	}
1353	pFileTo = manifest_file_next(pTo, 0);
1354	}else if( fossil_strcmp(pFileFrom->zUuid, pFileTo->zUuid)==0 ){
1355	pFileFrom = manifest_file_next(pFrom, 0);
1356	pFileTo = manifest_file_next(pTo, 0);
1357	}else{
1358	if(!zGlob \|\| (sqlite3_strglob(zGlob, pFileFrom->zName)==0
1359	\|\| sqlite3_strglob(zGlob, pFileTo->zName)==0) ){
1360	append_file_change_line(zFrom, pFileFrom->zName,
1361	pFileFrom->zUuid,
1362	pFileTo->zUuid, 0, pCfg,
1363	manifest_file_mperm(pFileTo));
1364	}
1365	pFileFrom = manifest_file_next(pFrom, 0);
1366	pFileTo = manifest_file_next(pTo, 0);
1367	}
1368	}

1369	manifest_destroy(pFrom);
1370	manifest_destroy(pTo);
1371	append_diff_javascript(diffType);
1372	style_finish_page();
1373	}
1374

	--- src/info.c
	+++ src/info.c
	@@ -372,11 +372,13 @@
372	const char zNew, / blob.uuid after change. NULL for deletes */
373	const char zOldName, / Prior name. NULL if no name change. */
374	DiffConfig pCfg, / Flags for text_diff() or NULL to omit all */
375	int mperm /* executable or symlink permission for zNew */
376	){
377	@ <div class='file-change-line'><span>
378	/* Maintenance reminder: the extra level of SPAN is for
379	** arranging new elements via JS. */
380	if( !g.perm.Hyperlink ){
381	if( zNew==0 ){
382	@ Deleted %h(zName).
383	}else if( zOld==0 ){
384	@ Added %h(zName).
	@@ -391,10 +393,11 @@
393	@ %h(zName) became a regular file.
394	}
395	}else{
396	@ Changes to %h(zName).
397	}
398	@ </span></div>
399	if( pCfg ){
400	append_diff(zOld, zNew, pCfg);
401	}
402	}else{
403	if( zOld && zNew ){
	@@ -436,18 +439,22 @@
439	@ %h(zName)</a> version %z(href("%R/artifact/%!S",zOld))[%S(zOld)]</a>.
440	}else{
441	@ Added %z(href("%R/finfo?name=%T&m=%!S&ci=%!S",zName,zNew,zCkin))\
442	@ %h(zName)</a> version %z(href("%R/artifact/%!S",zNew))[%S(zNew)]</a>.
443	}
444	if( zOld && zNew && fossil_strcmp(zOld,zNew)!=0 ){
445	if( pCfg ){
446	@ </span></div>
447	append_diff(zOld, zNew, pCfg);
448	}else{
449	@ %z(href("%R/fdiff?v1=%!S&v2=%!S",zOld,zNew))[diff]</a>
450	@ </span></div>
451	}
452	}else{
453	@ </span></div>
454	}
455	}

456	}
457
458	/*
459	** Generate javascript to enhance HTML diffs.
460	*/
	@@ -600,10 +607,130 @@
607	www_print_timeline(&q, TIMELINE_DISJOINT\|TIMELINE_GRAPH\|TIMELINE_NOSCROLL,
608	0, 0, 0, rid, 0, 0);
609	db_finalize(&q);
610	style_finish_page();
611	}
612
613	/*
614	** WEBPAGE: ckout
615	**
616	** Show information about the current checkout. This page only functions
617	** if the web server is run on a loopback interface (in other words, was
618	** started using "fossil ui" or similar) from with on open check-out.
619	*/
620	void ckout_page(void){
621	int vid;
622	char *zHostname;
623	char *zCwd;
624	int diffType; /* 0: no diff, 1: unified, 2: side-by-side */
625	DiffConfig DCfg,pCfg; / Diff details */
626	const char zHome; / Home directory */
627	Stmt q;
628
629	if( !db_open_local(0) \|\| !cgi_is_loopback(g.zIpAddr) ){
630	cgi_redirect("%R/home");
631	return;
632	}
633	diffType = preferred_diff_type();
634	pCfg = construct_diff_flags(diffType, &DCfg);
635	vid = db_lget_int("checkout", 0);
636	db_unprotect(PROTECT_ALL);
637	vfile_check_signature(vid, CKSIG_ENOTFILE);
638	db_protect_pop();
639	style_set_current_feature("vinfo");
640	zHostname = fossil_hostname();
641	zCwd = file_getcwd(0,0);
642	zHome = fossil_getenv("HOME");
643	if( zHome ){
644	int nHome = (int)strlen(zHome);
645	if( strncmp(zCwd, zHome, nHome)==0 && zCwd[nHome]=='/' ){
646	zCwd = mprintf("~%s", zCwd+nHome);
647	}
648	}
649	if( zHostname ){
650	style_header("Checkout Status: %h on %h", zCwd, zHostname);
651	}else{
652	style_header("Checkout Status: %h", zCwd);
653	}
654	render_checkin_context(vid, 0, 0, 0);
655	if( pCfg==0 ){
656	style_finish_page();
657	return;
658	}
659	db_prepare(&q,
660	/* 0 1 2 3 4 5 6 */
661	"SELECT pathname, deleted, chnged , rid==0, rid, islink, uuid"
662	" FROM vfile LEFT JOIN blob USING(rid)"
663	" WHERE vid=%d"
664	" AND (deleted OR chnged OR rid==0)"
665	" ORDER BY pathname /scan/",
666	vid
667	);
668	if( pCfg->diffFlags & DIFF_SIDEBYSIDE ){
669	pCfg->diffFlags \|= DIFF_HTML \| DIFF_NOTTOOBIG;
670	}else{
671	pCfg->diffFlags \|= DIFF_LINENO \| DIFF_HTML \| DIFF_NOTTOOBIG;
672	}
673	while( db_step(&q)==SQLITE_ROW ){
674	const char *zTreename = db_column_text(&q,0);
675	int isDeleted = db_column_int(&q, 1);
676	int isChnged = db_column_int(&q,2);
677	int isNew = db_column_int(&q,3);
678	int srcid = db_column_int(&q, 4);
679	int isLink = db_column_int(&q, 5);
680	const char *zUuid = db_column_text(&q, 6);
681	int showDiff = 1;
682
683	pCfg->diffFlags &= (~DIFF_FILE_MASK);
684	if( isDeleted ){
685	@ <p>DELETED %h(zTreename)</p>
686	pCfg->diffFlags \|= DIFF_FILE_DELETED;
687	showDiff = 0;
688	}else if( file_access(zTreename, F_OK) ){
689	@ <p>MISSING %h(zTreename)</p>
690	showDiff = 0;
691	}else if( isNew ){
692	@ <p>ADDED %h(zTreename)</p>
693	pCfg->diffFlags \|= DIFF_FILE_ADDED;
694	srcid = 0;
695	showDiff = 0;
696	}else if( isChnged==3 ){
697	@ <p>ADDED_BY_MERGE %h(zTreename)</p>
698	pCfg->diffFlags \|= DIFF_FILE_ADDED;
699	srcid = 0;
700	showDiff = 0;
701	}else if( isChnged==5 ){
702	@ <p>ADDED_BY_INTEGRATE %h(zTreename)</p>
703	pCfg->diffFlags \|= DIFF_FILE_ADDED;
704	srcid = 0;
705	showDiff = 0;
706	}else{
707	@ <p>CHANGED %h(zTreename)</p>
708	}
709	if( showDiff ){
710	Blob old, new;
711	if( !isLink != !file_islink(zTreename) ){
712	@ %s(DIFF_CANNOT_COMPUTE_SYMLINK)
713	continue;
714	}
715	if( srcid>0 ){
716	content_get(srcid, &old);
717	pCfg->zLeftHash = zUuid;
718	}else{
719	blob_zero(&old);
720	pCfg->zLeftHash = 0;
721	}
722	blob_read_from_file(&new, zTreename, ExtFILE);
723	text_diff(&old, &new, cgi_output_blob(), pCfg);
724	blob_reset(&old);
725	blob_reset(&new);
726	}
727	}
728	db_finalize(&q);
729	append_diff_javascript(diffType);
730	style_finish_page();
731	}
732
733	/*
734	** WEBPAGE: vinfo
735	** WEBPAGE: ci
736	** URL: /ci/ARTIFACTID
	@@ -1191,22 +1318,25 @@
1318	const char *zBranch;
1319	const char *zFrom;
1320	const char *zTo;
1321	const char *zRe;
1322	const char *zGlob;
1323	Glob * pGlob = 0;
1324	char *zMergeOrigin = 0;
1325	ReCompiled *pRe = 0;
1326	DiffConfig DCfg, *pCfg = 0;
1327	int graphFlags = 0;
1328	Blob qp; /* non-glob= query parameters for generated links */
1329	Blob qpGlob; /* glob= query parameter for generated links */
1330	int bInvert = PB("inv");
1331
1332	login_check_credentials();
1333	if( !g.perm.Read ){ login_needed(g.anon.Read); return; }
1334	login_anonymous_available();
1335	fossil_nice_default();
1336	blob_init(&qp, 0, 0);
1337	blob_init(&qpGlob, 0, 0);
1338	diffType = preferred_diff_type();
1339	zRe = P("regex");
1340	if( zRe ) re_compile(&pRe, zRe, 0);
1341	zBranch = P("branch");
1342	if( zBranch && zBranch[0]==0 ) zBranch = 0;
	@@ -1246,11 +1376,12 @@
1376	}
1377	if( zGlob ){
1378	if( !*zGlob ){
1379	zGlob = NULL;
1380	}else{
1381	blob_appendf(&qpGlob, "&glob=%T", zGlob);
1382	pGlob = glob_create(zGlob);
1383	}
1384	}
1385	if( PB("nc") ){
1386	graphFlags \|= TIMELINE_NOCOLOR;
1387	blob_appendf(&qp, "&nc");
	@@ -1263,20 +1394,22 @@
1394	style_set_current_feature("vdiff");
1395	if( zBranch==0 ){
1396	style_submenu_element("Path", "%R/timeline?me=%T&you=%T", zFrom, zTo);
1397	}
1398	if( diffType!=0 ){
1399	style_submenu_element("Hide Diff", "%R/vdiff?diff=0&%b%b", &qp, &qpGlob);
1400	}
1401	if( diffType!=2 ){
1402	style_submenu_element("Side-by-Side Diff", "%R/vdiff?diff=2&%b%b", &qp,
1403	&qpGlob);
1404	}
1405	if( diffType!=1 ) {
1406	style_submenu_element("Unified Diff", "%R/vdiff?diff=1&%b%b", &qp, &qpGlob);
1407	}
1408	if( zBranch==0 ){
1409	style_submenu_element("Invert","%R/vdiff?diff=%d&inv&%b%b", diffType,
1410	&qp, &qpGlob);
1411	}
1412	if( zGlob ){
1413	style_submenu_element("Clear glob", "%R/vdiff?diff=%d&%b", diffType, &qp);
1414	}else{
1415	style_submenu_element("Patch", "%R/vpatch?from=%T&to=%T%s", zFrom, zTo,
	@@ -1321,10 +1454,11 @@
1454	@ <p><b>Only files matching the glob "%h(zGlob)" are shown.</b></p>
1455	}
1456	@<hr><p>
1457	}
1458	blob_reset(&qp);
1459	blob_reset(&qpGlob);
1460
1461	manifest_file_rewind(pFrom);
1462	pFileFrom = manifest_file_next(pFrom, 0);
1463	manifest_file_rewind(pTo);
1464	pFileTo = manifest_file_next(pTo, 0);
	@@ -1337,37 +1471,38 @@
1471	cmp = -1;
1472	}else{
1473	cmp = fossil_strcmp(pFileFrom->zName, pFileTo->zName);
1474	}
1475	if( cmp<0 ){
1476	if( !pGlob \|\| glob_match(pGlob, pFileFrom->zName) ){
1477	append_file_change_line(zFrom, pFileFrom->zName,
1478	pFileFrom->zUuid, 0, 0, pCfg, 0);
1479	}
1480	pFileFrom = manifest_file_next(pFrom, 0);
1481	}else if( cmp>0 ){
1482	if( !pGlob \|\| glob_match(pGlob, pFileTo->zName) ){
1483	append_file_change_line(zTo, pFileTo->zName,
1484	0, pFileTo->zUuid, 0, pCfg,
1485	manifest_file_mperm(pFileTo));
1486	}
1487	pFileTo = manifest_file_next(pTo, 0);
1488	}else if( fossil_strcmp(pFileFrom->zUuid, pFileTo->zUuid)==0 ){
1489	pFileFrom = manifest_file_next(pFrom, 0);
1490	pFileTo = manifest_file_next(pTo, 0);
1491	}else{
1492	if(!pGlob \|\| (glob_match(pGlob, pFileFrom->zName)
1493	\|\| glob_match(pGlob, pFileTo->zName)) ){
1494	append_file_change_line(zFrom, pFileFrom->zName,
1495	pFileFrom->zUuid,
1496	pFileTo->zUuid, 0, pCfg,
1497	manifest_file_mperm(pFileTo));
1498	}
1499	pFileFrom = manifest_file_next(pFrom, 0);
1500	pFileTo = manifest_file_next(pTo, 0);
1501	}
1502	}
1503	glob_free(pGlob);
1504	manifest_destroy(pFrom);
1505	manifest_destroy(pTo);
1506	append_diff_javascript(diffType);
1507	style_finish_page();
1508	}
1509

M src/interwiki.c

+2 -1

		--- src/interwiki.c
		+++ src/interwiki.c
		@@ -275,13 +275,14 @@
275	275	db_prepare(&q,
276	276	"SELECT substr(name,11), value->>'base'"
277	277	" FROM config WHERE name glob 'interwiki:*' AND json_valid(value)"
278	278	" ORDER BY name;"
279	279	);
	280	+ blob_append(out, "<blockquote>", -1);
280	281	while( db_step(&q)==SQLITE_ROW ){
281	282	if( n==0 ){
282		- blob_appendf(out, "<blockquote><table>\n");
	283	+ blob_appendf(out, "<table>\n");
283	284	}
284	285	blob_appendf(out,"<tr><td>%h</td><td> → </td>",
285	286	db_column_text(&q,0));
286	287	blob_appendf(out,"<td>%h</td></tr>\n",
287	288	db_column_text(&q,1));
288	289

	--- src/interwiki.c
	+++ src/interwiki.c
	@@ -275,13 +275,14 @@
275	db_prepare(&q,
276	"SELECT substr(name,11), value->>'base'"
277	" FROM config WHERE name glob 'interwiki:*' AND json_valid(value)"
278	" ORDER BY name;"
279	);

280	while( db_step(&q)==SQLITE_ROW ){
281	if( n==0 ){
282	blob_appendf(out, "<blockquote><table>\n");
283	}
284	blob_appendf(out,"<tr><td>%h</td><td> → </td>",
285	db_column_text(&q,0));
286	blob_appendf(out,"<td>%h</td></tr>\n",
287	db_column_text(&q,1));
288

	--- src/interwiki.c
	+++ src/interwiki.c
	@@ -275,13 +275,14 @@
275	db_prepare(&q,
276	"SELECT substr(name,11), value->>'base'"
277	" FROM config WHERE name glob 'interwiki:*' AND json_valid(value)"
278	" ORDER BY name;"
279	);
280	blob_append(out, "<blockquote>", -1);
281	while( db_step(&q)==SQLITE_ROW ){
282	if( n==0 ){
283	blob_appendf(out, "<table>\n");
284	}
285	blob_appendf(out,"<tr><td>%h</td><td> → </td>",
286	db_column_text(&q,0));
287	blob_appendf(out,"<td>%h</td></tr>\n",
288	db_column_text(&q,1));
289

M src/login.c

+16 -4

		--- src/login.c
		+++ src/login.c
		@@ -751,11 +751,11 @@
751	751	if( anonFlag ){
752	752	@ <input type="hidden" name="anon" value="1">
753	753	}
754	754	if( g.zLogin ){
755	755	@ <p>Currently logged in as <b>%h(g.zLogin)</b>.
756		- @ <input type="submit" name="out" value="Logout"></p>
	756	+ @ <input type="submit" name="out" value="Logout" autofocus></p>
757	757	@ </form>
758	758	}else{
759	759	unsigned int uSeed = captcha_seed();
760	760	if( g.zLogin==0 && (anonFlag \|\| zGoto==0) ){
761	761	zAnonPw = db_text(0, "SELECT pw FROM user"
		@@ -777,11 +777,11 @@
777	777	@ </span></td></tr>
778	778	}
779	779	@ <tr>
780	780	@ <td class="form_label" id="userlabel1">User ID:</td>
781	781	@ <td><input type="text" id="u" aria-labelledby="userlabel1" name="u" \
782		- @ size="30" value="%s(anonFlag?"anonymous":"")"></td>
	782	+ @ size="30" value="%s(anonFlag?"anonymous":"")" autofocus></td>
783	783	@ </tr>
784	784	@ <tr>
785	785	@ <td class="form_label" id="pswdlabel">Password:</td>
786	786	@ <td><input aria-labelledby="pswdlabel" type="password" id="p" \
787	787	@ name="p" value="" size="30">\
		@@ -1302,20 +1302,32 @@
1302	1302	*/
1303	1303	void login_restrict_robot_access(void){
1304	1304	const char *zReferer;
1305	1305	const char *zGlob;
1306	1306	int isMatch = 1;
	1307	+ int nQP; /* Number of query parameters other than name= */
1307	1308	if( g.zLogin!=0 ) return;
1308	1309	zGlob = db_get("robot-restrict",0);
1309	1310	if( zGlob==0 \|\| zGlob[0]==0 ) return;
1310	1311	if( g.isHuman ){
1311	1312	zReferer = P("HTTP_REFERER");
1312	1313	if( zReferer && zReferer[0]!=0 ) return;
1313	1314	}
1314		- if( cgi_qp_count()<1 ) return;
	1315	+ nQP = cgi_qp_count();
	1316	+ if( nQP<1 ) return;
1315	1317	isMatch = glob_multi_match(zGlob, g.zPath);
1316	1318	if( !isMatch ) return;
	1319	+
	1320	+ /* Check for exceptions to the restriction on the number of query
	1321	+ ** parameters. */
	1322	+ zGlob = db_get("robot-restrict-qp",0);
	1323	+ if( zGlob && zGlob[0] ){
	1324	+ char *zPath = mprintf("%s/%d", g.zPath, nQP);
	1325	+ isMatch = glob_multi_match(zGlob, zPath);
	1326	+ fossil_free(zPath);
	1327	+ if( isMatch ) return;
	1328	+ }
1317	1329
1318	1330	/* If we reach this point, it means we have a situation where we
1319	1331	** want to restrict the activity of a robot.
1320	1332	*/
1321	1333	g.isHuman = 0;
		@@ -2243,11 +2255,11 @@
2243	2255	@ <p><input type="hidden" name="captchaseed" value="%u(uSeed)">
2244	2256	@ <table class="login_out">
2245	2257	@ <tr>
2246	2258	@ <td class="form_label" align="right" id="uid">User ID:</td>
2247	2259	@ <td><input aria-labelledby="uid" type="text" name="u" \
2248		- @ value="%h(zUserID)" size="30"></td>
	2260	+ @ value="%h(zUserID)" size="30" autofocus></td>
2249	2261	@
2250	2262	if( iErrLine==1 ){
2251	2263	@ <tr><td><td><span class='loginError'>↑ %h(zErr)</span></td></tr>
2252	2264	}
2253	2265	@ <tr>
2254	2266

	--- src/login.c
	+++ src/login.c
	@@ -751,11 +751,11 @@
751	if( anonFlag ){
752	@ <input type="hidden" name="anon" value="1">
753	}
754	if( g.zLogin ){
755	@ <p>Currently logged in as <b>%h(g.zLogin)</b>.
756	@ <input type="submit" name="out" value="Logout"></p>
757	@ </form>
758	}else{
759	unsigned int uSeed = captcha_seed();
760	if( g.zLogin==0 && (anonFlag \|\| zGoto==0) ){
761	zAnonPw = db_text(0, "SELECT pw FROM user"
	@@ -777,11 +777,11 @@
777	@ </span></td></tr>
778	}
779	@ <tr>
780	@ <td class="form_label" id="userlabel1">User ID:</td>
781	@ <td><input type="text" id="u" aria-labelledby="userlabel1" name="u" \
782	@ size="30" value="%s(anonFlag?"anonymous":"")"></td>
783	@ </tr>
784	@ <tr>
785	@ <td class="form_label" id="pswdlabel">Password:</td>
786	@ <td><input aria-labelledby="pswdlabel" type="password" id="p" \
787	@ name="p" value="" size="30">\
	@@ -1302,20 +1302,32 @@
1302	*/
1303	void login_restrict_robot_access(void){
1304	const char *zReferer;
1305	const char *zGlob;
1306	int isMatch = 1;

1307	if( g.zLogin!=0 ) return;
1308	zGlob = db_get("robot-restrict",0);
1309	if( zGlob==0 \|\| zGlob[0]==0 ) return;
1310	if( g.isHuman ){
1311	zReferer = P("HTTP_REFERER");
1312	if( zReferer && zReferer[0]!=0 ) return;
1313	}
1314	if( cgi_qp_count()<1 ) return;

1315	isMatch = glob_multi_match(zGlob, g.zPath);
1316	if( !isMatch ) return;










1317
1318	/* If we reach this point, it means we have a situation where we
1319	** want to restrict the activity of a robot.
1320	*/
1321	g.isHuman = 0;
	@@ -2243,11 +2255,11 @@
2243	@ <p><input type="hidden" name="captchaseed" value="%u(uSeed)">
2244	@ <table class="login_out">
2245	@ <tr>
2246	@ <td class="form_label" align="right" id="uid">User ID:</td>
2247	@ <td><input aria-labelledby="uid" type="text" name="u" \
2248	@ value="%h(zUserID)" size="30"></td>
2249	@
2250	if( iErrLine==1 ){
2251	@ <tr><td><td><span class='loginError'>↑ %h(zErr)</span></td></tr>
2252	}
2253	@ <tr>
2254

	--- src/login.c
	+++ src/login.c
	@@ -751,11 +751,11 @@
751	if( anonFlag ){
752	@ <input type="hidden" name="anon" value="1">
753	}
754	if( g.zLogin ){
755	@ <p>Currently logged in as <b>%h(g.zLogin)</b>.
756	@ <input type="submit" name="out" value="Logout" autofocus></p>
757	@ </form>
758	}else{
759	unsigned int uSeed = captcha_seed();
760	if( g.zLogin==0 && (anonFlag \|\| zGoto==0) ){
761	zAnonPw = db_text(0, "SELECT pw FROM user"
	@@ -777,11 +777,11 @@
777	@ </span></td></tr>
778	}
779	@ <tr>
780	@ <td class="form_label" id="userlabel1">User ID:</td>
781	@ <td><input type="text" id="u" aria-labelledby="userlabel1" name="u" \
782	@ size="30" value="%s(anonFlag?"anonymous":"")" autofocus></td>
783	@ </tr>
784	@ <tr>
785	@ <td class="form_label" id="pswdlabel">Password:</td>
786	@ <td><input aria-labelledby="pswdlabel" type="password" id="p" \
787	@ name="p" value="" size="30">\
	@@ -1302,20 +1302,32 @@
1302	*/
1303	void login_restrict_robot_access(void){
1304	const char *zReferer;
1305	const char *zGlob;
1306	int isMatch = 1;
1307	int nQP; /* Number of query parameters other than name= */
1308	if( g.zLogin!=0 ) return;
1309	zGlob = db_get("robot-restrict",0);
1310	if( zGlob==0 \|\| zGlob[0]==0 ) return;
1311	if( g.isHuman ){
1312	zReferer = P("HTTP_REFERER");
1313	if( zReferer && zReferer[0]!=0 ) return;
1314	}
1315	nQP = cgi_qp_count();
1316	if( nQP<1 ) return;
1317	isMatch = glob_multi_match(zGlob, g.zPath);
1318	if( !isMatch ) return;
1319
1320	/* Check for exceptions to the restriction on the number of query
1321	** parameters. */
1322	zGlob = db_get("robot-restrict-qp",0);
1323	if( zGlob && zGlob[0] ){
1324	char *zPath = mprintf("%s/%d", g.zPath, nQP);
1325	isMatch = glob_multi_match(zGlob, zPath);
1326	fossil_free(zPath);
1327	if( isMatch ) return;
1328	}
1329
1330	/* If we reach this point, it means we have a situation where we
1331	** want to restrict the activity of a robot.
1332	*/
1333	g.isHuman = 0;
	@@ -2243,11 +2255,11 @@
2255	@ <p><input type="hidden" name="captchaseed" value="%u(uSeed)">
2256	@ <table class="login_out">
2257	@ <tr>
2258	@ <td class="form_label" align="right" id="uid">User ID:</td>
2259	@ <td><input aria-labelledby="uid" type="text" name="u" \
2260	@ value="%h(zUserID)" size="30" autofocus></td>
2261	@
2262	if( iErrLine==1 ){
2263	@ <tr><td><td><span class='loginError'>↑ %h(zErr)</span></td></tr>
2264	}
2265	@ <tr>
2266

M src/main.c

+13 -10

		--- src/main.c
		+++ src/main.c
		@@ -2040,20 +2040,27 @@
2040	2040	*/
2041	2041	set_base_url(0);
2042	2042	if( fossil_redirect_to_https_if_needed(2) ) return;
2043	2043	if( zPathInfo==0 \|\| zPathInfo[0]==0
2044	2044	\|\| (zPathInfo[0]=='/' && zPathInfo[1]==0) ){
2045		- /* Second special case: If the PATH_INFO is blank, issue a redirect to
2046		- ** the home page identified by the "index-page" setting in the repository
2047		- ** CONFIG table, to "/index" if there no "index-page" setting. */
	2045	+ /* Second special case: If the PATH_INFO is blank, issue a redirect:
	2046	+ ** (1) to "/ckout" if g.useLocalauth and g.localOpen are both set.
	2047	+ ** (2) to the home page identified by the "index-page" setting
	2048	+ ** in the repository CONFIG table
	2049	+ ** (3) to "/index" if there no "index-page" setting in CONFIG
	2050	+ */
2048	2051	#ifdef FOSSIL_ENABLE_JSON
2049	2052	if(g.json.isJsonMode){
2050	2053	json_err(FSL_JSON_E_RESOURCE_NOT_FOUND,NULL,1);
2051	2054	fossil_exit(0);
2052	2055	}
2053	2056	#endif
2054		- fossil_redirect_home() /does not return/;
	2057	+ if( g.useLocalauth && g.localOpen ){
	2058	+ cgi_redirectf("%R/ckout");
	2059	+ }else{
	2060	+ fossil_redirect_home() /does not return/;
	2061	+ }
2055	2062	}else{
2056	2063	zPath = mprintf("%s", zPathInfo);
2057	2064	}
2058	2065
2059	2066	/* Make g.zPath point to the first element of the path. Make
		@@ -3356,11 +3363,11 @@
3356	3363	const char * zDir = g.argv[2];
3357	3364	if(dir_has_ckout_db(zDir)){
3358	3365	if(0!=file_chdir(zDir, 0)){
3359	3366	fossil_fatal("Cannot chdir to %s", zDir);
3360	3367	}
3361		- findServerArg = 99;
	3368	+ findServerArg = g.argc;
3362	3369	fCreate = 0;
3363	3370	g.argv[2] = 0;
3364	3371	--g.argc;
3365	3372	}
3366	3373	}
		@@ -3384,15 +3391,11 @@
3384	3391	}
3385	3392	if( !zRemote ){
3386	3393	find_server_repository(findServerArg, fCreate);
3387	3394	}
3388	3395	if( zInitPage==0 ){
3389		- if( isUiCmd && g.localOpen ){
3390		- zInitPage = "timeline?c=current";
3391		- }else{
3392		- zInitPage = "";
3393		- }
	3396	+ zInitPage = "";
3394	3397	}
3395	3398	if( zPort ){
3396	3399	if( strchr(zPort,':') ){
3397	3400	int i;
3398	3401	for(i=strlen(zPort)-1; i>=0 && zPort[i]!=':'; i--){}
3399	3402

	--- src/main.c
	+++ src/main.c
	@@ -2040,20 +2040,27 @@
2040	*/
2041	set_base_url(0);
2042	if( fossil_redirect_to_https_if_needed(2) ) return;
2043	if( zPathInfo==0 \|\| zPathInfo[0]==0
2044	\|\| (zPathInfo[0]=='/' && zPathInfo[1]==0) ){
2045	/* Second special case: If the PATH_INFO is blank, issue a redirect to
2046	** the home page identified by the "index-page" setting in the repository
2047	** CONFIG table, to "/index" if there no "index-page" setting. */



2048	#ifdef FOSSIL_ENABLE_JSON
2049	if(g.json.isJsonMode){
2050	json_err(FSL_JSON_E_RESOURCE_NOT_FOUND,NULL,1);
2051	fossil_exit(0);
2052	}
2053	#endif
2054	fossil_redirect_home() /does not return/;




2055	}else{
2056	zPath = mprintf("%s", zPathInfo);
2057	}
2058
2059	/* Make g.zPath point to the first element of the path. Make
	@@ -3356,11 +3363,11 @@
3356	const char * zDir = g.argv[2];
3357	if(dir_has_ckout_db(zDir)){
3358	if(0!=file_chdir(zDir, 0)){
3359	fossil_fatal("Cannot chdir to %s", zDir);
3360	}
3361	findServerArg = 99;
3362	fCreate = 0;
3363	g.argv[2] = 0;
3364	--g.argc;
3365	}
3366	}
	@@ -3384,15 +3391,11 @@
3384	}
3385	if( !zRemote ){
3386	find_server_repository(findServerArg, fCreate);
3387	}
3388	if( zInitPage==0 ){
3389	if( isUiCmd && g.localOpen ){
3390	zInitPage = "timeline?c=current";
3391	}else{
3392	zInitPage = "";
3393	}
3394	}
3395	if( zPort ){
3396	if( strchr(zPort,':') ){
3397	int i;
3398	for(i=strlen(zPort)-1; i>=0 && zPort[i]!=':'; i--){}
3399

	--- src/main.c
	+++ src/main.c
	@@ -2040,20 +2040,27 @@
2040	*/
2041	set_base_url(0);
2042	if( fossil_redirect_to_https_if_needed(2) ) return;
2043	if( zPathInfo==0 \|\| zPathInfo[0]==0
2044	\|\| (zPathInfo[0]=='/' && zPathInfo[1]==0) ){
2045	/* Second special case: If the PATH_INFO is blank, issue a redirect:
2046	** (1) to "/ckout" if g.useLocalauth and g.localOpen are both set.
2047	** (2) to the home page identified by the "index-page" setting
2048	** in the repository CONFIG table
2049	** (3) to "/index" if there no "index-page" setting in CONFIG
2050	*/
2051	#ifdef FOSSIL_ENABLE_JSON
2052	if(g.json.isJsonMode){
2053	json_err(FSL_JSON_E_RESOURCE_NOT_FOUND,NULL,1);
2054	fossil_exit(0);
2055	}
2056	#endif
2057	if( g.useLocalauth && g.localOpen ){
2058	cgi_redirectf("%R/ckout");
2059	}else{
2060	fossil_redirect_home() /does not return/;
2061	}
2062	}else{
2063	zPath = mprintf("%s", zPathInfo);
2064	}
2065
2066	/* Make g.zPath point to the first element of the path. Make
	@@ -3356,11 +3363,11 @@
3363	const char * zDir = g.argv[2];
3364	if(dir_has_ckout_db(zDir)){
3365	if(0!=file_chdir(zDir, 0)){
3366	fossil_fatal("Cannot chdir to %s", zDir);
3367	}
3368	findServerArg = g.argc;
3369	fCreate = 0;
3370	g.argv[2] = 0;
3371	--g.argc;
3372	}
3373	}
	@@ -3384,15 +3391,11 @@
3391	}
3392	if( !zRemote ){
3393	find_server_repository(findServerArg, fCreate);
3394	}
3395	if( zInitPage==0 ){
3396	zInitPage = "";




3397	}
3398	if( zPort ){
3399	if( strchr(zPort,':') ){
3400	int i;
3401	for(i=strlen(zPort)-1; i>=0 && zPort[i]!=':'; i--){}
3402

M src/main.mk

+2 -1

		--- src/main.mk
		+++ src/main.mk
		@@ -248,10 +248,11 @@
248	248	$(SRCDIR)/hbmenu.js \
249	249	$(SRCDIR)/href.js \
250	250	$(SRCDIR)/login.js \
251	251	$(SRCDIR)/markdown.md \
252	252	$(SRCDIR)/menu.js \
	253	+ $(SRCDIR)/merge.tcl \
253	254	$(SRCDIR)/scroll.js \
254	255	$(SRCDIR)/skin.js \
255	256	$(SRCDIR)/sorttable.js \
256	257	$(SRCDIR)/sounds/0.wav \
257	258	$(SRCDIR)/sounds/1.wav \
		@@ -704,11 +705,11 @@
704	705
705	706	# The USE_LINENOISE variable may be undefined, set to 0, or set
706	707	# to 1. If it is set to 0, then there is no need to build or link
707	708	# the linenoise.o object.
708	709	LINENOISE_DEF.0 =
709		-LINENOISE_DEF.1 = -DHAVE_LINENOISE
	710	+LINENOISE_DEF.1 = -DHAVE_LINENOISE=2
710	711	LINENOISE_DEF. = $(LINENOISE_DEF.0)
711	712	LINENOISE_OBJ.0 =
712	713	LINENOISE_OBJ.1 = $(OBJDIR)/linenoise.o
713	714	LINENOISE_OBJ. = $(LINENOISE_OBJ.0)
714	715
715	716

	--- src/main.mk
	+++ src/main.mk
	@@ -248,10 +248,11 @@
248	$(SRCDIR)/hbmenu.js \
249	$(SRCDIR)/href.js \
250	$(SRCDIR)/login.js \
251	$(SRCDIR)/markdown.md \
252	$(SRCDIR)/menu.js \

253	$(SRCDIR)/scroll.js \
254	$(SRCDIR)/skin.js \
255	$(SRCDIR)/sorttable.js \
256	$(SRCDIR)/sounds/0.wav \
257	$(SRCDIR)/sounds/1.wav \
	@@ -704,11 +705,11 @@
704
705	# The USE_LINENOISE variable may be undefined, set to 0, or set
706	# to 1. If it is set to 0, then there is no need to build or link
707	# the linenoise.o object.
708	LINENOISE_DEF.0 =
709	LINENOISE_DEF.1 = -DHAVE_LINENOISE
710	LINENOISE_DEF. = $(LINENOISE_DEF.0)
711	LINENOISE_OBJ.0 =
712	LINENOISE_OBJ.1 = $(OBJDIR)/linenoise.o
713	LINENOISE_OBJ. = $(LINENOISE_OBJ.0)
714
715

	--- src/main.mk
	+++ src/main.mk
	@@ -248,10 +248,11 @@
248	$(SRCDIR)/hbmenu.js \
249	$(SRCDIR)/href.js \
250	$(SRCDIR)/login.js \
251	$(SRCDIR)/markdown.md \
252	$(SRCDIR)/menu.js \
253	$(SRCDIR)/merge.tcl \
254	$(SRCDIR)/scroll.js \
255	$(SRCDIR)/skin.js \
256	$(SRCDIR)/sorttable.js \
257	$(SRCDIR)/sounds/0.wav \
258	$(SRCDIR)/sounds/1.wav \
	@@ -704,11 +705,11 @@
705
706	# The USE_LINENOISE variable may be undefined, set to 0, or set
707	# to 1. If it is set to 0, then there is no need to build or link
708	# the linenoise.o object.
709	LINENOISE_DEF.0 =
710	LINENOISE_DEF.1 = -DHAVE_LINENOISE=2
711	LINENOISE_DEF. = $(LINENOISE_DEF.0)
712	LINENOISE_OBJ.0 =
713	LINENOISE_OBJ.1 = $(OBJDIR)/linenoise.o
714	LINENOISE_OBJ. = $(LINENOISE_OBJ.0)
715
716

M src/manifest.c

+7 -3

		--- src/manifest.c
		+++ src/manifest.c
		@@ -2334,11 +2334,11 @@
2334	2334	int i, rc = TH_OK;
2335	2335	Manifest *p;
2336	2336	int parentid = 0;
2337	2337	int permitHooks = (flags & MC_PERMIT_HOOKS);
2338	2338	const char *zScript = 0;
2339		- const char *zUuid = 0;
	2339	+ char *zUuid = 0;
2340	2340
2341	2341	if( g.fSqlTrace ){
2342	2342	fossil_trace("-- manifest_crosslink(%d)\n", rid);
2343	2343	}
2344	2344	manifest_create_event_triggers();
		@@ -2370,11 +2370,11 @@
2370	2370	}
2371	2371	db_begin_transaction();
2372	2372	if( p->type==CFTYPE_MANIFEST ){
2373	2373	if( permitHooks ){
2374	2374	zScript = xfer_commit_code();
2375		- zUuid = db_text(0, "SELECT uuid FROM blob WHERE rid=%d", rid);
	2375	+ zUuid = rid_to_uuid(rid);
2376	2376	}
2377	2377	if( p->nCherrypick && db_table_exists("repository","cherrypick") ){
2378	2378	int i;
2379	2379	for(i=0; i<p->nCherrypick; i++){
2380	2380	db_multi_exec(
		@@ -2722,11 +2722,12 @@
2722	2722	branchMove = 0;
2723	2723	if( permitHooks && db_exists("SELECT 1 FROM event, blob"
2724	2724	" WHERE event.type='ci' AND event.objid=blob.rid"
2725	2725	" AND blob.uuid=%Q", zTagUuid) ){
2726	2726	zScript = xfer_commit_code();
2727		- zUuid = zTagUuid;
	2727	+ fossil_free(zUuid);
	2728	+ zUuid = fossil_strdup(zTagUuid);
2728	2729	}
2729	2730	}
2730	2731	zName = p->aTag[i].zName;
2731	2732	zValue = p->aTag[i].zValue;
2732	2733	if( strcmp(zName, "*branch")==0 ){
		@@ -2805,10 +2806,12 @@
2805	2806	}
2806	2807	if( p->type==CFTYPE_FORUM ){
2807	2808	int froot, fprev, firt;
2808	2809	char *zFType;
2809	2810	char *zTitle;
	2811	+
	2812	+ assert( 0==zUuid );
2810	2813	schema_forum();
2811	2814	search_doc_touch('f', rid, 0);
2812	2815	froot = p->zThreadRoot ? uuid_to_rid(p->zThreadRoot, 1) : rid;
2813	2816	fprev = p->nParent ? uuid_to_rid(p->azParent[0],1) : 0;
2814	2817	firt = p->zInReplyTo ? uuid_to_rid(p->zInReplyTo,1) : 0;
		@@ -2877,10 +2880,11 @@
2877	2880	rc = xfer_run_common_script();
2878	2881	if( rc==TH_OK ){
2879	2882	rc = xfer_run_script(zScript, zUuid, 0);
2880	2883	}
2881	2884	}
	2885	+ fossil_free(zUuid);
2882	2886	if( p->type==CFTYPE_MANIFEST ){
2883	2887	manifest_cache_insert(p);
2884	2888	}else{
2885	2889	manifest_destroy(p);
2886	2890	}
2887	2891

	--- src/manifest.c
	+++ src/manifest.c
	@@ -2334,11 +2334,11 @@
2334	int i, rc = TH_OK;
2335	Manifest *p;
2336	int parentid = 0;
2337	int permitHooks = (flags & MC_PERMIT_HOOKS);
2338	const char *zScript = 0;
2339	const char *zUuid = 0;
2340
2341	if( g.fSqlTrace ){
2342	fossil_trace("-- manifest_crosslink(%d)\n", rid);
2343	}
2344	manifest_create_event_triggers();
	@@ -2370,11 +2370,11 @@
2370	}
2371	db_begin_transaction();
2372	if( p->type==CFTYPE_MANIFEST ){
2373	if( permitHooks ){
2374	zScript = xfer_commit_code();
2375	zUuid = db_text(0, "SELECT uuid FROM blob WHERE rid=%d", rid);
2376	}
2377	if( p->nCherrypick && db_table_exists("repository","cherrypick") ){
2378	int i;
2379	for(i=0; i<p->nCherrypick; i++){
2380	db_multi_exec(
	@@ -2722,11 +2722,12 @@
2722	branchMove = 0;
2723	if( permitHooks && db_exists("SELECT 1 FROM event, blob"
2724	" WHERE event.type='ci' AND event.objid=blob.rid"
2725	" AND blob.uuid=%Q", zTagUuid) ){
2726	zScript = xfer_commit_code();
2727	zUuid = zTagUuid;

2728	}
2729	}
2730	zName = p->aTag[i].zName;
2731	zValue = p->aTag[i].zValue;
2732	if( strcmp(zName, "*branch")==0 ){
	@@ -2805,10 +2806,12 @@
2805	}
2806	if( p->type==CFTYPE_FORUM ){
2807	int froot, fprev, firt;
2808	char *zFType;
2809	char *zTitle;


2810	schema_forum();
2811	search_doc_touch('f', rid, 0);
2812	froot = p->zThreadRoot ? uuid_to_rid(p->zThreadRoot, 1) : rid;
2813	fprev = p->nParent ? uuid_to_rid(p->azParent[0],1) : 0;
2814	firt = p->zInReplyTo ? uuid_to_rid(p->zInReplyTo,1) : 0;
	@@ -2877,10 +2880,11 @@
2877	rc = xfer_run_common_script();
2878	if( rc==TH_OK ){
2879	rc = xfer_run_script(zScript, zUuid, 0);
2880	}
2881	}

2882	if( p->type==CFTYPE_MANIFEST ){
2883	manifest_cache_insert(p);
2884	}else{
2885	manifest_destroy(p);
2886	}
2887

	--- src/manifest.c
	+++ src/manifest.c
	@@ -2334,11 +2334,11 @@
2334	int i, rc = TH_OK;
2335	Manifest *p;
2336	int parentid = 0;
2337	int permitHooks = (flags & MC_PERMIT_HOOKS);
2338	const char *zScript = 0;
2339	char *zUuid = 0;
2340
2341	if( g.fSqlTrace ){
2342	fossil_trace("-- manifest_crosslink(%d)\n", rid);
2343	}
2344	manifest_create_event_triggers();
	@@ -2370,11 +2370,11 @@
2370	}
2371	db_begin_transaction();
2372	if( p->type==CFTYPE_MANIFEST ){
2373	if( permitHooks ){
2374	zScript = xfer_commit_code();
2375	zUuid = rid_to_uuid(rid);
2376	}
2377	if( p->nCherrypick && db_table_exists("repository","cherrypick") ){
2378	int i;
2379	for(i=0; i<p->nCherrypick; i++){
2380	db_multi_exec(
	@@ -2722,11 +2722,12 @@
2722	branchMove = 0;
2723	if( permitHooks && db_exists("SELECT 1 FROM event, blob"
2724	" WHERE event.type='ci' AND event.objid=blob.rid"
2725	" AND blob.uuid=%Q", zTagUuid) ){
2726	zScript = xfer_commit_code();
2727	fossil_free(zUuid);
2728	zUuid = fossil_strdup(zTagUuid);
2729	}
2730	}
2731	zName = p->aTag[i].zName;
2732	zValue = p->aTag[i].zValue;
2733	if( strcmp(zName, "*branch")==0 ){
	@@ -2805,10 +2806,12 @@
2806	}
2807	if( p->type==CFTYPE_FORUM ){
2808	int froot, fprev, firt;
2809	char *zFType;
2810	char *zTitle;
2811
2812	assert( 0==zUuid );
2813	schema_forum();
2814	search_doc_touch('f', rid, 0);
2815	froot = p->zThreadRoot ? uuid_to_rid(p->zThreadRoot, 1) : rid;
2816	fprev = p->nParent ? uuid_to_rid(p->azParent[0],1) : 0;
2817	firt = p->zInReplyTo ? uuid_to_rid(p->zInReplyTo,1) : 0;
	@@ -2877,10 +2880,11 @@
2880	rc = xfer_run_common_script();
2881	if( rc==TH_OK ){
2882	rc = xfer_run_script(zScript, zUuid, 0);
2883	}
2884	}
2885	fossil_free(zUuid);
2886	if( p->type==CFTYPE_MANIFEST ){
2887	manifest_cache_insert(p);
2888	}else{
2889	manifest_destroy(p);
2890	}
2891

M src/merge.c

+465 -6

		--- src/merge.c
		+++ src/merge.c
		@@ -20,10 +20,372 @@
20	20	*/
21	21	#include "config.h"
22	22	#include "merge.h"
23	23	#include <assert.h>
24	24
	25	+
	26	+/*
	27	+** Bring up a Tcl/Tk GUI to show details of the most recent merge.
	28	+*/
	29	+static void merge_info_tk(int bDark, int bAll, int nContext){
	30	+ int i;
	31	+ Blob script;
	32	+ const char *zTempFile = 0;
	33	+ char *zCmd;
	34	+ const char *zTclsh;
	35	+ zTclsh = find_option("tclsh",0,1);
	36	+ if( zTclsh==0 ){
	37	+ zTclsh = db_get("tclsh",0);
	38	+ }
	39	+ /* The undocumented --script FILENAME option causes the Tk script to
	40	+ ** be written into the FILENAME instead of being run. This is used
	41	+ ** for testing and debugging. */
	42	+ zTempFile = find_option("script",0,1);
	43	+ verify_all_options();
	44	+
	45	+ blob_zero(&script);
	46	+ blob_appendf(&script, "set ncontext %d\n", nContext);
	47	+ blob_appendf(&script, "set fossilcmd {\| \"%/\" merge-info}\n",
	48	+ g.nameOfExe);
	49	+ blob_appendf(&script, "set filelist [list");
	50	+ if( g.argc==2 ){
	51	+ /* No files named on the command-line. Use every file mentioned
	52	+ ** in the MERGESTAT table to generate the file list. */
	53	+ Stmt q;
	54	+ int cnt = 0;
	55	+ db_prepare(&q,
	56	+ "WITH priority(op,pri) AS (VALUES('CONFLICT',0),('ERROR',0),"
	57	+ "('MERGE',1),('ADDED',2),('UPDATE',2))"
	58	+ "SELECT coalesce(fnr,fn), op FROM mergestat JOIN priority USING(op)"
	59	+ " %s ORDER BY pri, 1",
	60	+ bAll ? "" : "WHERE op IN ('MERGE','CONFLICT')" /safe-for-%s/
	61	+ );
	62	+ while( db_step(&q)==SQLITE_ROW ){
	63	+ blob_appendf(&script," %s ", db_column_text(&q,1));
	64	+ blob_append_tcl_literal(&script, db_column_text(&q,0),
	65	+ db_column_bytes(&q,0));
	66	+ cnt++;
	67	+ }
	68	+ db_finalize(&q);
	69	+ if( cnt==0 ){
	70	+ fossil_print(
	71	+ "No interesting changes in this merge. Use --all to see everything\n"
	72	+ );
	73	+ return;
	74	+ }
	75	+ }else{
	76	+ /* Use only files named on the command-line in the file list.
	77	+ ** But verify each file named is actually found in the MERGESTAT
	78	+ ** table first. */
	79	+ for(i=2; i<g.argc; i++){
	80	+ char zFile; / Input filename */
	81	+ char zTreename; / Name of the file in the tree */
	82	+ Blob fname; /* Filename relative to root */
	83	+ char zOp; / Operation on this file */
	84	+ zFile = mprintf("%/", g.argv[i]);
	85	+ file_tree_name(zFile, &fname, 0, 1);
	86	+ fossil_free(zFile);
	87	+ zTreename = blob_str(&fname);
	88	+ zOp = db_text(0, "SELECT op FROM mergestat WHERE fn=%Q or fnr=%Q",
	89	+ zTreename, zTreename);
	90	+ blob_appendf(&script, " %s ", zOp);
	91	+ fossil_free(zOp);
	92	+ blob_append_tcl_literal(&script, zTreename, (int)strlen(zTreename));
	93	+ blob_reset(&fname);
	94	+ }
	95	+ }
	96	+ blob_appendf(&script, "]\n");
	97	+ blob_appendf(&script, "set darkmode %d\n", bDark!=0);
	98	+ blob_appendf(&script, "%s", builtin_file("merge.tcl", 0));
	99	+ if( zTempFile ){
	100	+ blob_write_to_file(&script, zTempFile);
	101	+ fossil_print("To see the merge, run: %s \"%s\"\n", zTclsh, zTempFile);
	102	+ }else{
	103	+#if defined(FOSSIL_ENABLE_TCL)
	104	+ Th_FossilInit(TH_INIT_DEFAULT);
	105	+ if( evaluateTclWithEvents(g.interp, &g.tcl, blob_str(&script),
	106	+ blob_size(&script), 1, 1, 0)==TCL_OK ){
	107	+ blob_reset(&script);
	108	+ return;
	109	+ }
	110	+ /*
	111	+ * If evaluation of the Tcl script fails, the reason may be that Tk
	112	+ * could not be found by the loaded Tcl, or that Tcl cannot be loaded
	113	+ * dynamically (e.g. x64 Tcl with x86 Fossil). Therefore, fallback
	114	+ * to using the external "tclsh", if available.
	115	+ */
	116	+#endif
	117	+ zTempFile = write_blob_to_temp_file(&script);
	118	+ zCmd = mprintf("%$ %$", zTclsh, zTempFile);
	119	+ fossil_system(zCmd);
	120	+ file_delete(zTempFile);
	121	+ fossil_free(zCmd);
	122	+ }
	123	+ blob_reset(&script);
	124	+}
	125	+
	126	+/*
	127	+** Generate a TCL list on standard output that can be fed into the
	128	+** merge.tcl script to show the details of the most recent merge
	129	+** command associated with file "zFName". zFName must be the filename
	130	+** relative to the root of the check-in - in other words a "tree name".
	131	+**
	132	+** When this routine is called, we know that the mergestat table
	133	+** exists, but we do not know if zFName is mentioned in that table.
	134	+*/
	135	+static void merge_info_tcl(const char *zFName, int nContext){
	136	+ const char zTreename;/ Name of the file in the tree */
	137	+ Stmt q; /* To query the MERGESTAT table */
	138	+ MergeBuilder mb; /* The merge builder object */
	139	+ Blob pivot,v1,v2,out; /* Blobs for holding content */
	140	+ const char zFN; / A filename */
	141	+ int rid; /* RID value */
	142	+ int sz; /* File size value */
	143	+
	144	+ zTreename = zFName;
	145	+ db_prepare(&q,
	146	+ /* 0 1 2 3 4 5 6 7 */
	147	+ "SELECT fnp, ridp, fn, ridv, sz, fnm, ridm, fnr"
	148	+ " FROM mergestat"
	149	+ " WHERE fnp=%Q OR fnr=%Q",
	150	+ zTreename, zTreename
	151	+ );
	152	+ if( db_step(&q)!=SQLITE_ROW ){
	153	+ db_finalize(&q);
	154	+ fossil_print("ERROR {don't know anything about file: %s}\n", zTreename);
	155	+ return;
	156	+ }
	157	+ mergebuilder_init_tcl(&mb);
	158	+ mb.nContext = nContext;
	159	+
	160	+ /* Set up the pivot */
	161	+ zFN = db_column_text(&q, 0);
	162	+ if( zFN==0 ){
	163	+ /* No pivot because the file was added */
	164	+ mb.zPivot = "(no baseline)";
	165	+ blob_zero(&pivot);
	166	+ }else{
	167	+ mb.zPivot = mprintf("%s (baseline)", file_tail(zFN));
	168	+ rid = db_column_int(&q, 1);
	169	+ content_get(rid, &pivot);
	170	+ }
	171	+ mb.pPivot = &pivot;
	172	+
	173	+ /* Set up the merge-in as V2 */
	174	+ zFN = db_column_text(&q, 5);
	175	+ if( zFN==0 ){
	176	+ /* File deleted in the merged-in branch */
	177	+ mb.zV2 = "(deleted file)";
	178	+ blob_zero(&v2);
	179	+ }else{
	180	+ mb.zV2 = mprintf("%s (merge-in)", file_tail(zFN));
	181	+ rid = db_column_int(&q, 6);
	182	+ content_get(rid, &v2);
	183	+ }
	184	+ mb.pV2 = &v2;
	185	+
	186	+ /* Set up the local content as V1 */
	187	+ zFN = db_column_text(&q, 2);
	188	+ if( zFN==0 ){
	189	+ /* File added by merge */
	190	+ mb.zV1 = "(no original)";
	191	+ blob_zero(&v1);
	192	+ }else{
	193	+ mb.zV1 = mprintf("%s (local)", file_tail(zFN));
	194	+ rid = db_column_int(&q, 3);
	195	+ sz = db_column_int(&q, 4);
	196	+ if( rid==0 && sz>0 ){
	197	+ /* The origin file had been edited so we'll have to pull its
	198	+ ** original content out of the undo buffer */
	199	+ Stmt q2;
	200	+ db_prepare(&q2,
	201	+ "SELECT content FROM undo"
	202	+ " WHERE pathname=%Q AND octet_length(content)=%d",
	203	+ zFN, sz
	204	+ );
	205	+ blob_zero(&v1);
	206	+ if( db_step(&q2)==SQLITE_ROW ){
	207	+ db_column_blob(&q2, 0, &v1);
	208	+ }else{
	209	+ mb.zV1 = "(local content missing)";
	210	+ }
	211	+ db_finalize(&q2);
	212	+ }else{
	213	+ /* The origin file was unchanged when the merge first occurred */
	214	+ content_get(rid, &v1);
	215	+ }
	216	+ }
	217	+ mb.pV1 = &v1;
	218	+
	219	+ /* Set up the output */
	220	+ zFN = db_column_text(&q, 7);
	221	+ if( zFN==0 ){
	222	+ mb.zOut = "(Merge Result)";
	223	+ }else{
	224	+ mb.zOut = mprintf("%s (after merge)", file_tail(zFN));
	225	+ }
	226	+ blob_zero(&out);
	227	+ mb.pOut = &out;
	228	+
	229	+ merge_three_blobs(&mb);
	230	+ blob_write_to_file(&out, "-");
	231	+
	232	+ mb.xDestroy(&mb);
	233	+ blob_reset(&pivot);
	234	+ blob_reset(&v1);
	235	+ blob_reset(&v2);
	236	+ blob_reset(&out);
	237	+ db_finalize(&q);
	238	+}
	239	+
	240	+/*
	241	+** COMMAND: merge-info
	242	+**
	243	+** Usage: %fossil merge-info [OPTIONS]
	244	+**
	245	+** Display information about the most recent merge operation.
	246	+**
	247	+** Options:
	248	+** -a\|--all Show all file changes that happened because of
	249	+** the merge. Normally only MERGE, CONFLICT, and ERROR
	250	+** lines are shown
	251	+** -c\|--context N Show N lines of context around each change,
	252	+** with negative N meaning show all content. Only
	253	+** meaningful in combination with --tcl or --tk.
	254	+** --dark Use dark mode for the Tcl/Tk-based GUI
	255	+** --tcl FILE Generate (to stdout) a TCL list containing
	256	+** information needed to display the changes to
	257	+** FILE caused by the most recent merge. FILE must
	258	+** be a pathname relative to the root of the check-out.
	259	+** --tk Bring up a Tcl/Tk GUI that shows the changes
	260	+** associated with the most recent merge.
	261	+**
	262	+*/
	263	+void merge_info_cmd(void){
	264	+ const char *zCnt;
	265	+ const char *zTcl;
	266	+ int bTk;
	267	+ int bDark;
	268	+ int bAll;
	269	+ int nContext;
	270	+ Stmt q;
	271	+ const char *zWhere;
	272	+ int cnt = 0;
	273	+
	274	+ db_must_be_within_tree();
	275	+ zTcl = find_option("tcl", 0, 1);
	276	+ bTk = find_option("tk", 0, 0)!=0;
	277	+ zCnt = find_option("context", "c", 1);
	278	+ bDark = find_option("dark", 0, 0)!=0;
	279	+ bAll = find_option("all", "a", 0)!=0;
	280	+ if( bTk==0 ){
	281	+ verify_all_options();
	282	+ if( g.argc>2 ){
	283	+ usage("[OPTIONS]");
	284	+ }
	285	+ }
	286	+ if( zCnt ){
	287	+ nContext = atoi(zCnt);
	288	+ if( nContext<0 ) nContext = 0xfffffff;
	289	+ }else{
	290	+ nContext = 6;
	291	+ }
	292	+ if( !db_table_exists("localdb","mergestat") ){
	293	+ if( zTcl ){
	294	+ fossil_print("ERROR {no merge data available}\n");
	295	+ }else{
	296	+ fossil_print("No merge data is available\n");
	297	+ }
	298	+ return;
	299	+ }
	300	+ if( bTk ){
	301	+ merge_info_tk(bDark, bAll, nContext);
	302	+ return;
	303	+ }
	304	+ if( zTcl ){
	305	+ merge_info_tcl(zTcl, nContext);
	306	+ return;
	307	+ }
	308	+ if( bAll ){
	309	+ zWhere = "";
	310	+ }else{
	311	+ zWhere = "WHERE op IN ('MERGE','CONFLICT','ERROR')";
	312	+ }
	313	+ db_prepare(&q,
	314	+ "WITH priority(op,pri) AS (VALUES('CONFLICT',0),('ERROR',0),"
	315	+ "('MERGE',1),('ADDED',2),('UPDATE',2))"
	316	+
	317	+ /* 0 1 2 */
	318	+ "SELECT op, coalesce(fnr,fn), msg"
	319	+ " FROM mergestat JOIN priority USING(op)"
	320	+ " %s"
	321	+ " ORDER BY pri, coalesce(fnr,fn)",
	322	+ zWhere /safe-for-%s/
	323	+ );
	324	+ while( db_step(&q)==SQLITE_ROW ){
	325	+ const char *zOp = db_column_text(&q, 0);
	326	+ const char *zName = db_column_text(&q, 1);
	327	+ const char *zErr = db_column_text(&q, 2);
	328	+ if( zErr && fossil_strcmp(zOp,"CONFLICT")!=0 ){
	329	+ fossil_print("%-9s %s (%s)\n", zOp, zName, zErr);
	330	+ }else{
	331	+ fossil_print("%-9s %s\n", zOp, zName);
	332	+ }
	333	+ cnt++;
	334	+ }
	335	+ db_finalize(&q);
	336	+ if( !bAll && cnt==0 ){
	337	+ fossil_print(
	338	+ "No interesting changes in this merge. Use --all to see everything.\n"
	339	+ );
	340	+ }
	341	+}
	342	+
	343	+/*
	344	+** Erase all information about prior merges. Do this, for example, after
	345	+** a commit.
	346	+*/
	347	+void merge_info_forget(void){
	348	+ db_multi_exec(
	349	+ "DROP TABLE IF EXISTS localdb.mergestat;"
	350	+ "DELETE FROM localdb.vvar WHERE name glob 'mergestat-*';"
	351	+ );
	352	+}
	353	+
	354	+
	355	+/*
	356	+** Initialize the MERGESTAT table.
	357	+**
	358	+** Notes about mergestat:
	359	+**
	360	+** * ridv is a positive integer and sz is NULL if the V file contained
	361	+** no local edits prior to the merge. If the V file was modified prior
	362	+** to the merge then ridv is NULL and sz is the size of the file prior
	363	+** to merge.
	364	+**
	365	+** * fnp, ridp, fn, ridv, and sz are all NULL for a file that was
	366	+** added by merge.
	367	+*/
	368	+void merge_info_init(void){
	369	+ merge_info_forget();
	370	+ db_multi_exec(
	371	+ "CREATE TABLE localdb.mergestat(\n"
	372	+ " op TEXT, -- 'UPDATE', 'ADDED', 'MERGE', etc...\n"
	373	+ " fnp TEXT, -- Name of the pivot file (P)\n"
	374	+ " ridp INT, -- RID for the pivot file\n"
	375	+ " fn TEXT, -- Name of origin file (V)\n"
	376	+ " ridv INT, -- RID for origin file, or NULL if previously edited\n"
	377	+ " sz INT, -- Size of origin file in bytes, NULL if unedited\n"
	378	+ " fnm TEXT, -- Name of the file being merged in (M)\n"
	379	+ " ridm INT, -- RID for the merge-in file\n"
	380	+ " fnr TEXT, -- Name of the final output file, after all renaming\n"
	381	+ " nc INT DEFAULT 0, -- Number of conflicts\n"
	382	+ " msg TEXT -- Error message\n"
	383	+ ");"
	384	+ );
	385	+}
	386	+
25	387	/*
26	388	** Print information about a particular check-in.
27	389	*/
28	390	void print_checkin_description(int rid, int indent, const char *zLabel){
29	391	Stmt q;
		@@ -295,10 +657,13 @@
295	657	** Files which are renamed in the merged-in branch will be renamed in
296	658	** the current check-out.
297	659	**
298	660	** If the VERSION argument is omitted, then Fossil attempts to find
299	661	** a recent fork on the current branch to merge.
	662	+**
	663	+** Note that this command does not commit the merge, as that is a
	664	+** separate step.
300	665	**
301	666	** If there are multiple VERSION arguments, then each VERSION is merged
302	667	** (or cherrypicked) in the order that they appear on the command-line.
303	668	**
304	669	** Options:
		@@ -320,11 +685,11 @@
320	685	** --force-missing Force the merge even if there is missing content
321	686	** --integrate Merged branch will be closed when committing
322	687	** -K\|--keep-merge-files On merge conflict, retain the temporary files
323	688	** used for merging, named -baseline, -original,
324	689	** and *-merge.
325		-** -n\|--dry-run If given, display instead of run actions
	690	+** -n\|--dry-run Do not actually change files on disk
326	691	** --nosync Do not auto-sync prior to merging
327	692	** -v\|--verbose Show additional details of the merge
328	693	*/
329	694	void merge_cmd(void){
330	695	int vid; /* Current version "V" */
		@@ -797,15 +1162,21 @@
797	1162
798	1163	/************************************************************************
799	1164	** All of the information needed to do the merge is now contained in the
800	1165	** FV table. Starting here, we begin to actually carry out the merge.
801	1166	**
802		- ** First, find files that have changed from P->M but not P->V.
	1167	+ ** Begin by constructing the localdb.mergestat table.
	1168	+ */
	1169	+ merge_info_init();
	1170	+
	1171	+ /*
	1172	+ ** Find files that have changed from P->M but not P->V.
803	1173	** Copy the M content over into V.
804	1174	*/
805	1175	db_prepare(&q,
806		- "SELECT idv, ridm, fn, islinkm FROM fv"
	1176	+ /* 0 1 2 3 4 5 6 7 */
	1177	+ "SELECT idv, ridm, fn, islinkm, fnp, ridp, ridv, fnm FROM fv"
807	1178	" WHERE idp>0 AND idv>0 AND idm>0"
808	1179	" AND ridm!=ridp AND ridv=ridp AND NOT chnged"
809	1180	);
810	1181	while( db_step(&q)==SQLITE_ROW ){
811	1182	int idv = db_column_int(&q, 0);
		@@ -822,10 +1193,21 @@
822	1193	" THEN (SELECT uuid FROM blob WHERE blob.rid=%d) END"
823	1194	" WHERE id=%d", ridm, integrateFlag?4:2, islinkm, ridm, ridm, idv
824	1195	);
825	1196	vfile_to_disk(0, idv, 0, 0);
826	1197	}
	1198	+ db_multi_exec(
	1199	+ "INSERT INTO mergestat(op,fnp,ridp,fn,ridv,fnm,ridm,fnr)"
	1200	+ "VALUES('UPDATE',%Q,%d,%Q,%d,%Q,%d,%Q)",
	1201	+ /* fnp */ db_column_text(&q, 4),
	1202	+ /* ridp */ db_column_int(&q,5),
	1203	+ /* fn */ zName,
	1204	+ /* ridv */ db_column_int(&q,6),
	1205	+ /* fnm */ db_column_text(&q, 7),
	1206	+ /* ridm */ ridm,
	1207	+ /* fnr */ zName
	1208	+ );
827	1209	}
828	1210	db_finalize(&q);
829	1211
830	1212	/*
831	1213	** Do a three-way merge on files that have changes on both P->M and P->V.
		@@ -833,11 +1215,15 @@
833	1215	** Proceed even if the file doesn't exist on P, just like the common ancestor
834	1216	** of M and V is an empty file. In this case, merge conflict marks will be
835	1217	** added to the file and user will be forced to take a decision.
836	1218	*/
837	1219	db_prepare(&q,
838		- "SELECT ridm, idv, ridp, ridv, %s, fn, isexe, islinkv, islinkm FROM fv"
	1220	+ /* 0 1 2 3 4 5 6 7 8 */
	1221	+ "SELECT ridm, idv, ridp, ridv, %z, fn, isexe, islinkv, islinkm,"
	1222	+ /* 9 10 11 */
	1223	+ " fnp, fnm, chnged"
	1224	+ " FROM fv"
839	1225	" WHERE idv>0 AND idm>0"
840	1226	" AND ridm!=ridp AND (ridv!=ridp OR chnged)",
841	1227	glob_expr("fv.fn", zBinGlob)
842	1228	);
843	1229	while( db_step(&q)==SQLITE_ROW ){
		@@ -848,12 +1234,14 @@
848	1234	int isBinary = db_column_int(&q, 4);
849	1235	const char *zName = db_column_text(&q, 5);
850	1236	int isExe = db_column_int(&q, 6);
851	1237	int islinkv = db_column_int(&q, 7);
852	1238	int islinkm = db_column_int(&q, 8);
	1239	+ int chnged = db_column_int(&q, 11);
853	1240	int rc;
854	1241	char *zFullPath;
	1242	+ const char *zType = "MERGE";
855	1243	Blob m, p, r;
856	1244	/* Do a 3-way merge of idp->idm into idp->idv. The results go into idv. */
857	1245	if( verboseFlag ){
858	1246	fossil_print("MERGE %s (pivot=%d v1=%d v2=%d)\n",
859	1247	zName, ridp, ridm, ridv);
		@@ -861,13 +1249,29 @@
861	1249	fossil_print("MERGE %s\n", zName);
862	1250	}
863	1251	if( islinkv \|\| islinkm ){
864	1252	fossil_print("***** Cannot merge symlink %s\n", zName);
865	1253	nConflict++;
	1254	+ db_multi_exec(
	1255	+ "INSERT INTO mergestat(op,fnp,ridp,fn,ridv,fnm,ridm,fnr,nc,msg)"
	1256	+ "VALUES('ERROR',%Q,%d,%Q,%d,%Q,%d,%Q,1,'cannot merge symlink')",
	1257	+ /* fnp */ db_column_text(&q, 9),
	1258	+ /* ridp */ ridp,
	1259	+ /* fn */ zName,
	1260	+ /* ridv */ ridv,
	1261	+ /* fnm */ db_column_text(&q, 10),
	1262	+ /* ridm */ ridm,
	1263	+ /* fnr */ zName
	1264	+ );
866	1265	}else{
	1266	+ i64 sz;
	1267	+ const char *zErrMsg = 0;
	1268	+ int nc = 0;
	1269	+
867	1270	if( !dryRunFlag ) undo_save(zName);
868	1271	zFullPath = mprintf("%s/%s", g.zLocalRoot, zName);
	1272	+ sz = file_size(zFullPath, ExtFILE);
869	1273	content_get(ridp, &p);
870	1274	content_get(ridm, &m);
871	1275	if( isBinary ){
872	1276	rc = -1;
873	1277	blob_zero(&r);
		@@ -884,15 +1288,38 @@
884	1288	db_multi_exec("UPDATE vfile SET mtime=0 WHERE id=%d", idv);
885	1289	if( rc>0 ){
886	1290	fossil_print("***** %d merge conflict%s in %s\n",
887	1291	rc, rc>1 ? "s" : "", zName);
888	1292	nConflict++;
	1293	+ nc = rc;
	1294	+ zErrMsg = "merge conflicts";
	1295	+ zType = "CONFLICT";
889	1296	}
890	1297	}else{
891	1298	fossil_print("***** Cannot merge binary file %s\n", zName);
892	1299	nConflict++;
	1300	+ nc = 1;
	1301	+ zErrMsg = "cannot merge binary file";
	1302	+ zType = "ERROR";
893	1303	}
	1304	+ db_multi_exec(
	1305	+ "INSERT INTO mergestat(op,fnp,ridp,fn,ridv,sz,fnm,ridm,fnr,nc,msg)"
	1306	+ "VALUES(%Q,%Q,%d,%Q,iif(%d,%d,NULL),iif(%d,%lld,NULL),%Q,%d,"
	1307	+ "%Q,%d,%Q)",
	1308	+ /* op */ zType,
	1309	+ /* fnp */ db_column_text(&q, 9),
	1310	+ /* ridp */ ridp,
	1311	+ /* fn */ zName,
	1312	+ /* ridv */ chnged==0, ridv,
	1313	+ /* sz */ chnged!=0, sz,
	1314	+ /* fnm */ db_column_text(&q, 10),
	1315	+ /* ridm */ ridm,
	1316	+ /* fnr */ zName,
	1317	+ /* nc */ nc,
	1318	+ /* msg */ zErrMsg
	1319	+ );
	1320	+ fossil_free(zFullPath);
894	1321	blob_reset(&p);
895	1322	blob_reset(&m);
896	1323	blob_reset(&r);
897	1324	}
898	1325	vmerge_insert(idv, ridm);
		@@ -901,22 +1328,33 @@
901	1328
902	1329	/*
903	1330	** Drop files that are in P and V but not in M
904	1331	*/
905	1332	db_prepare(&q,
906		- "SELECT idv, fn, chnged FROM fv"
	1333	+ "SELECT idv, fn, chnged, ridv FROM fv"
907	1334	" WHERE idp>0 AND idv>0 AND idm=0"
908	1335	);
909	1336	while( db_step(&q)==SQLITE_ROW ){
910	1337	int idv = db_column_int(&q, 0);
911	1338	const char *zName = db_column_text(&q, 1);
912	1339	int chnged = db_column_int(&q, 2);
	1340	+ int ridv = db_column_int(&q, 3);
	1341	+ int sz = -1;
	1342	+ const char *zErrMsg = 0;
	1343	+ int nc = 0;
913	1344	/* Delete the file idv */
914	1345	fossil_print("DELETE %s\n", zName);
915	1346	if( chnged ){
	1347	+ char *zFullPath;
916	1348	fossil_warning("WARNING: local edits lost for %s", zName);
917	1349	nConflict++;
	1350	+ ridv = 0;
	1351	+ nc = 1;
	1352	+ zErrMsg = "local edits lost";
	1353	+ zFullPath = mprintf("%s/%s", g.zLocalRoot, zName);
	1354	+ sz = file_size(zFullPath, ExtFILE);
	1355	+ fossil_free(zFullPath);
918	1356	}
919	1357	if( !dryRunFlag ) undo_save(zName);
920	1358	db_multi_exec(
921	1359	"UPDATE vfile SET deleted=1 WHERE id=%d", idv
922	1360	);
		@@ -923,10 +1361,20 @@
923	1361	if( !dryRunFlag ){
924	1362	char *zFullPath = mprintf("%s%s", g.zLocalRoot, zName);
925	1363	file_delete(zFullPath);
926	1364	free(zFullPath);
927	1365	}
	1366	+ db_multi_exec(
	1367	+ "INSERT INTO localdb.mergestat(op,fnp,ridp,fn,ridv,sz,fnm,ridm,nc,msg)"
	1368	+ "VALUES('DELETE',NULL,NULL,%Q,iif(%d,%d,NULL),iif(%d,%d,NULL),"
	1369	+ "NULL,NULL,%d,%Q)",
	1370	+ /* fn */ zName,
	1371	+ /* ridv */ chnged==0, ridv,
	1372	+ /* sz */ chnged!=0, sz,
	1373	+ /* nc */ nc,
	1374	+ /* msg */ zErrMsg
	1375	+ );
928	1376	}
929	1377	db_finalize(&q);
930	1378
931	1379	/* For certain sets of renames (e.g. A -> B and B -> A), a file that is
932	1380	** being renamed must first be moved to a temporary location to avoid
		@@ -953,10 +1401,14 @@
953	1401	const char *zNewName = db_column_text(&q, 2);
954	1402	int isExe = db_column_int(&q, 3);
955	1403	fossil_print("RENAME %s -> %s\n", zOldName, zNewName);
956	1404	if( !dryRunFlag ) undo_save(zOldName);
957	1405	if( !dryRunFlag ) undo_save(zNewName);
	1406	+ db_multi_exec(
	1407	+ "UPDATE mergestat SET fnr=fnm WHERE fnp=%Q",
	1408	+ zOldName
	1409	+ );
958	1410	db_multi_exec(
959	1411	"UPDATE vfile SET pathname=NULL, origname=pathname"
960	1412	" WHERE vid=%d AND pathname=%Q;"
961	1413	"UPDATE vfile SET pathname=%Q, origname=coalesce(origname,pathname)"
962	1414	" WHERE id=%d;",
		@@ -1006,11 +1458,11 @@
1006	1458
1007	1459	/*
1008	1460	** Insert into V any files that are not in V or P but are in M.
1009	1461	*/
1010	1462	db_prepare(&q,
1011		- "SELECT idm, fnm FROM fv"
	1463	+ "SELECT idm, fnm, ridm FROM fv"
1012	1464	" WHERE idp=0 AND idv=0 AND idm>0"
1013	1465	);
1014	1466	while( db_step(&q)==SQLITE_ROW ){
1015	1467	int idm = db_column_int(&q, 0);
1016	1468	const char *zName;
		@@ -1039,10 +1491,17 @@
1039	1491	nOverwrite++;
1040	1492	}else{
1041	1493	fossil_print("ADDED %s\n", zName);
1042	1494	}
1043	1495	fossil_free(zFullName);
	1496	+ db_multi_exec(
	1497	+ "INSERT INTO mergestat(op,fnm,ridm,fnr)"
	1498	+ "VALUES('ADDED',%Q,%d,%Q)",
	1499	+ /* fnm */ zName,
	1500	+ /* ridm */ db_column_int(&q,2),
	1501	+ /* fnr */ zName
	1502	+ );
1044	1503	if( !dryRunFlag ){
1045	1504	undo_save(zName);
1046	1505	vfile_to_disk(0, idm, 0, 0);
1047	1506	}
1048	1507	}
1049	1508
1050	1509	ADDED src/merge.tcl

	--- src/merge.c
	+++ src/merge.c
	@@ -20,10 +20,372 @@
20	*/
21	#include "config.h"
22	#include "merge.h"
23	#include <assert.h>
24










































































































































































































































































































































































25	/*
26	** Print information about a particular check-in.
27	*/
28	void print_checkin_description(int rid, int indent, const char *zLabel){
29	Stmt q;
	@@ -295,10 +657,13 @@
295	** Files which are renamed in the merged-in branch will be renamed in
296	** the current check-out.
297	**
298	** If the VERSION argument is omitted, then Fossil attempts to find
299	** a recent fork on the current branch to merge.



300	**
301	** If there are multiple VERSION arguments, then each VERSION is merged
302	** (or cherrypicked) in the order that they appear on the command-line.
303	**
304	** Options:
	@@ -320,11 +685,11 @@
320	** --force-missing Force the merge even if there is missing content
321	** --integrate Merged branch will be closed when committing
322	** -K\|--keep-merge-files On merge conflict, retain the temporary files
323	** used for merging, named -baseline, -original,
324	** and *-merge.
325	** -n\|--dry-run If given, display instead of run actions
326	** --nosync Do not auto-sync prior to merging
327	** -v\|--verbose Show additional details of the merge
328	*/
329	void merge_cmd(void){
330	int vid; /* Current version "V" */
	@@ -797,15 +1162,21 @@
797
798	/************************************************************************
799	** All of the information needed to do the merge is now contained in the
800	** FV table. Starting here, we begin to actually carry out the merge.
801	**
802	** First, find files that have changed from P->M but not P->V.





803	** Copy the M content over into V.
804	*/
805	db_prepare(&q,
806	"SELECT idv, ridm, fn, islinkm FROM fv"

807	" WHERE idp>0 AND idv>0 AND idm>0"
808	" AND ridm!=ridp AND ridv=ridp AND NOT chnged"
809	);
810	while( db_step(&q)==SQLITE_ROW ){
811	int idv = db_column_int(&q, 0);
	@@ -822,10 +1193,21 @@
822	" THEN (SELECT uuid FROM blob WHERE blob.rid=%d) END"
823	" WHERE id=%d", ridm, integrateFlag?4:2, islinkm, ridm, ridm, idv
824	);
825	vfile_to_disk(0, idv, 0, 0);
826	}











827	}
828	db_finalize(&q);
829
830	/*
831	** Do a three-way merge on files that have changes on both P->M and P->V.
	@@ -833,11 +1215,15 @@
833	** Proceed even if the file doesn't exist on P, just like the common ancestor
834	** of M and V is an empty file. In this case, merge conflict marks will be
835	** added to the file and user will be forced to take a decision.
836	*/
837	db_prepare(&q,
838	"SELECT ridm, idv, ridp, ridv, %s, fn, isexe, islinkv, islinkm FROM fv"




839	" WHERE idv>0 AND idm>0"
840	" AND ridm!=ridp AND (ridv!=ridp OR chnged)",
841	glob_expr("fv.fn", zBinGlob)
842	);
843	while( db_step(&q)==SQLITE_ROW ){
	@@ -848,12 +1234,14 @@
848	int isBinary = db_column_int(&q, 4);
849	const char *zName = db_column_text(&q, 5);
850	int isExe = db_column_int(&q, 6);
851	int islinkv = db_column_int(&q, 7);
852	int islinkm = db_column_int(&q, 8);

853	int rc;
854	char *zFullPath;

855	Blob m, p, r;
856	/* Do a 3-way merge of idp->idm into idp->idv. The results go into idv. */
857	if( verboseFlag ){
858	fossil_print("MERGE %s (pivot=%d v1=%d v2=%d)\n",
859	zName, ridp, ridm, ridv);
	@@ -861,13 +1249,29 @@
861	fossil_print("MERGE %s\n", zName);
862	}
863	if( islinkv \|\| islinkm ){
864	fossil_print("***** Cannot merge symlink %s\n", zName);
865	nConflict++;











866	}else{




867	if( !dryRunFlag ) undo_save(zName);
868	zFullPath = mprintf("%s/%s", g.zLocalRoot, zName);

869	content_get(ridp, &p);
870	content_get(ridm, &m);
871	if( isBinary ){
872	rc = -1;
873	blob_zero(&r);
	@@ -884,15 +1288,38 @@
884	db_multi_exec("UPDATE vfile SET mtime=0 WHERE id=%d", idv);
885	if( rc>0 ){
886	fossil_print("***** %d merge conflict%s in %s\n",
887	rc, rc>1 ? "s" : "", zName);
888	nConflict++;



889	}
890	}else{
891	fossil_print("***** Cannot merge binary file %s\n", zName);
892	nConflict++;



893	}

















894	blob_reset(&p);
895	blob_reset(&m);
896	blob_reset(&r);
897	}
898	vmerge_insert(idv, ridm);
	@@ -901,22 +1328,33 @@
901
902	/*
903	** Drop files that are in P and V but not in M
904	*/
905	db_prepare(&q,
906	"SELECT idv, fn, chnged FROM fv"
907	" WHERE idp>0 AND idv>0 AND idm=0"
908	);
909	while( db_step(&q)==SQLITE_ROW ){
910	int idv = db_column_int(&q, 0);
911	const char *zName = db_column_text(&q, 1);
912	int chnged = db_column_int(&q, 2);




913	/* Delete the file idv */
914	fossil_print("DELETE %s\n", zName);
915	if( chnged ){

916	fossil_warning("WARNING: local edits lost for %s", zName);
917	nConflict++;






918	}
919	if( !dryRunFlag ) undo_save(zName);
920	db_multi_exec(
921	"UPDATE vfile SET deleted=1 WHERE id=%d", idv
922	);
	@@ -923,10 +1361,20 @@
923	if( !dryRunFlag ){
924	char *zFullPath = mprintf("%s%s", g.zLocalRoot, zName);
925	file_delete(zFullPath);
926	free(zFullPath);
927	}










928	}
929	db_finalize(&q);
930
931	/* For certain sets of renames (e.g. A -> B and B -> A), a file that is
932	** being renamed must first be moved to a temporary location to avoid
	@@ -953,10 +1401,14 @@
953	const char *zNewName = db_column_text(&q, 2);
954	int isExe = db_column_int(&q, 3);
955	fossil_print("RENAME %s -> %s\n", zOldName, zNewName);
956	if( !dryRunFlag ) undo_save(zOldName);
957	if( !dryRunFlag ) undo_save(zNewName);




958	db_multi_exec(
959	"UPDATE vfile SET pathname=NULL, origname=pathname"
960	" WHERE vid=%d AND pathname=%Q;"
961	"UPDATE vfile SET pathname=%Q, origname=coalesce(origname,pathname)"
962	" WHERE id=%d;",
	@@ -1006,11 +1458,11 @@
1006
1007	/*
1008	** Insert into V any files that are not in V or P but are in M.
1009	*/
1010	db_prepare(&q,
1011	"SELECT idm, fnm FROM fv"
1012	" WHERE idp=0 AND idv=0 AND idm>0"
1013	);
1014	while( db_step(&q)==SQLITE_ROW ){
1015	int idm = db_column_int(&q, 0);
1016	const char *zName;
	@@ -1039,10 +1491,17 @@
1039	nOverwrite++;
1040	}else{
1041	fossil_print("ADDED %s\n", zName);
1042	}
1043	fossil_free(zFullName);







1044	if( !dryRunFlag ){
1045	undo_save(zName);
1046	vfile_to_disk(0, idm, 0, 0);
1047	}
1048	}
1049
1050	DDED src/merge.tcl

	--- src/merge.c
	+++ src/merge.c
	@@ -20,10 +20,372 @@
20	*/
21	#include "config.h"
22	#include "merge.h"
23	#include <assert.h>
24
25
26	/*
27	** Bring up a Tcl/Tk GUI to show details of the most recent merge.
28	*/
29	static void merge_info_tk(int bDark, int bAll, int nContext){
30	int i;
31	Blob script;
32	const char *zTempFile = 0;
33	char *zCmd;
34	const char *zTclsh;
35	zTclsh = find_option("tclsh",0,1);
36	if( zTclsh==0 ){
37	zTclsh = db_get("tclsh",0);
38	}
39	/* The undocumented --script FILENAME option causes the Tk script to
40	** be written into the FILENAME instead of being run. This is used
41	** for testing and debugging. */
42	zTempFile = find_option("script",0,1);
43	verify_all_options();
44
45	blob_zero(&script);
46	blob_appendf(&script, "set ncontext %d\n", nContext);
47	blob_appendf(&script, "set fossilcmd {\| \"%/\" merge-info}\n",
48	g.nameOfExe);
49	blob_appendf(&script, "set filelist [list");
50	if( g.argc==2 ){
51	/* No files named on the command-line. Use every file mentioned
52	** in the MERGESTAT table to generate the file list. */
53	Stmt q;
54	int cnt = 0;
55	db_prepare(&q,
56	"WITH priority(op,pri) AS (VALUES('CONFLICT',0),('ERROR',0),"
57	"('MERGE',1),('ADDED',2),('UPDATE',2))"
58	"SELECT coalesce(fnr,fn), op FROM mergestat JOIN priority USING(op)"
59	" %s ORDER BY pri, 1",
60	bAll ? "" : "WHERE op IN ('MERGE','CONFLICT')" /safe-for-%s/
61	);
62	while( db_step(&q)==SQLITE_ROW ){
63	blob_appendf(&script," %s ", db_column_text(&q,1));
64	blob_append_tcl_literal(&script, db_column_text(&q,0),
65	db_column_bytes(&q,0));
66	cnt++;
67	}
68	db_finalize(&q);
69	if( cnt==0 ){
70	fossil_print(
71	"No interesting changes in this merge. Use --all to see everything\n"
72	);
73	return;
74	}
75	}else{
76	/* Use only files named on the command-line in the file list.
77	** But verify each file named is actually found in the MERGESTAT
78	** table first. */
79	for(i=2; i<g.argc; i++){
80	char zFile; / Input filename */
81	char zTreename; / Name of the file in the tree */
82	Blob fname; /* Filename relative to root */
83	char zOp; / Operation on this file */
84	zFile = mprintf("%/", g.argv[i]);
85	file_tree_name(zFile, &fname, 0, 1);
86	fossil_free(zFile);
87	zTreename = blob_str(&fname);
88	zOp = db_text(0, "SELECT op FROM mergestat WHERE fn=%Q or fnr=%Q",
89	zTreename, zTreename);
90	blob_appendf(&script, " %s ", zOp);
91	fossil_free(zOp);
92	blob_append_tcl_literal(&script, zTreename, (int)strlen(zTreename));
93	blob_reset(&fname);
94	}
95	}
96	blob_appendf(&script, "]\n");
97	blob_appendf(&script, "set darkmode %d\n", bDark!=0);
98	blob_appendf(&script, "%s", builtin_file("merge.tcl", 0));
99	if( zTempFile ){
100	blob_write_to_file(&script, zTempFile);
101	fossil_print("To see the merge, run: %s \"%s\"\n", zTclsh, zTempFile);
102	}else{
103	#if defined(FOSSIL_ENABLE_TCL)
104	Th_FossilInit(TH_INIT_DEFAULT);
105	if( evaluateTclWithEvents(g.interp, &g.tcl, blob_str(&script),
106	blob_size(&script), 1, 1, 0)==TCL_OK ){
107	blob_reset(&script);
108	return;
109	}
110	/*
111	* If evaluation of the Tcl script fails, the reason may be that Tk
112	* could not be found by the loaded Tcl, or that Tcl cannot be loaded
113	* dynamically (e.g. x64 Tcl with x86 Fossil). Therefore, fallback
114	* to using the external "tclsh", if available.
115	*/
116	#endif
117	zTempFile = write_blob_to_temp_file(&script);
118	zCmd = mprintf("%$ %$", zTclsh, zTempFile);
119	fossil_system(zCmd);
120	file_delete(zTempFile);
121	fossil_free(zCmd);
122	}
123	blob_reset(&script);
124	}
125
126	/*
127	** Generate a TCL list on standard output that can be fed into the
128	** merge.tcl script to show the details of the most recent merge
129	** command associated with file "zFName". zFName must be the filename
130	** relative to the root of the check-in - in other words a "tree name".
131	**
132	** When this routine is called, we know that the mergestat table
133	** exists, but we do not know if zFName is mentioned in that table.
134	*/
135	static void merge_info_tcl(const char *zFName, int nContext){
136	const char zTreename;/ Name of the file in the tree */
137	Stmt q; /* To query the MERGESTAT table */
138	MergeBuilder mb; /* The merge builder object */
139	Blob pivot,v1,v2,out; /* Blobs for holding content */
140	const char zFN; / A filename */
141	int rid; /* RID value */
142	int sz; /* File size value */
143
144	zTreename = zFName;
145	db_prepare(&q,
146	/* 0 1 2 3 4 5 6 7 */
147	"SELECT fnp, ridp, fn, ridv, sz, fnm, ridm, fnr"
148	" FROM mergestat"
149	" WHERE fnp=%Q OR fnr=%Q",
150	zTreename, zTreename
151	);
152	if( db_step(&q)!=SQLITE_ROW ){
153	db_finalize(&q);
154	fossil_print("ERROR {don't know anything about file: %s}\n", zTreename);
155	return;
156	}
157	mergebuilder_init_tcl(&mb);
158	mb.nContext = nContext;
159
160	/* Set up the pivot */
161	zFN = db_column_text(&q, 0);
162	if( zFN==0 ){
163	/* No pivot because the file was added */
164	mb.zPivot = "(no baseline)";
165	blob_zero(&pivot);
166	}else{
167	mb.zPivot = mprintf("%s (baseline)", file_tail(zFN));
168	rid = db_column_int(&q, 1);
169	content_get(rid, &pivot);
170	}
171	mb.pPivot = &pivot;
172
173	/* Set up the merge-in as V2 */
174	zFN = db_column_text(&q, 5);
175	if( zFN==0 ){
176	/* File deleted in the merged-in branch */
177	mb.zV2 = "(deleted file)";
178	blob_zero(&v2);
179	}else{
180	mb.zV2 = mprintf("%s (merge-in)", file_tail(zFN));
181	rid = db_column_int(&q, 6);
182	content_get(rid, &v2);
183	}
184	mb.pV2 = &v2;
185
186	/* Set up the local content as V1 */
187	zFN = db_column_text(&q, 2);
188	if( zFN==0 ){
189	/* File added by merge */
190	mb.zV1 = "(no original)";
191	blob_zero(&v1);
192	}else{
193	mb.zV1 = mprintf("%s (local)", file_tail(zFN));
194	rid = db_column_int(&q, 3);
195	sz = db_column_int(&q, 4);
196	if( rid==0 && sz>0 ){
197	/* The origin file had been edited so we'll have to pull its
198	** original content out of the undo buffer */
199	Stmt q2;
200	db_prepare(&q2,
201	"SELECT content FROM undo"
202	" WHERE pathname=%Q AND octet_length(content)=%d",
203	zFN, sz
204	);
205	blob_zero(&v1);
206	if( db_step(&q2)==SQLITE_ROW ){
207	db_column_blob(&q2, 0, &v1);
208	}else{
209	mb.zV1 = "(local content missing)";
210	}
211	db_finalize(&q2);
212	}else{
213	/* The origin file was unchanged when the merge first occurred */
214	content_get(rid, &v1);
215	}
216	}
217	mb.pV1 = &v1;
218
219	/* Set up the output */
220	zFN = db_column_text(&q, 7);
221	if( zFN==0 ){
222	mb.zOut = "(Merge Result)";
223	}else{
224	mb.zOut = mprintf("%s (after merge)", file_tail(zFN));
225	}
226	blob_zero(&out);
227	mb.pOut = &out;
228
229	merge_three_blobs(&mb);
230	blob_write_to_file(&out, "-");
231
232	mb.xDestroy(&mb);
233	blob_reset(&pivot);
234	blob_reset(&v1);
235	blob_reset(&v2);
236	blob_reset(&out);
237	db_finalize(&q);
238	}
239
240	/*
241	** COMMAND: merge-info
242	**
243	** Usage: %fossil merge-info [OPTIONS]
244	**
245	** Display information about the most recent merge operation.
246	**
247	** Options:
248	** -a\|--all Show all file changes that happened because of
249	** the merge. Normally only MERGE, CONFLICT, and ERROR
250	** lines are shown
251	** -c\|--context N Show N lines of context around each change,
252	** with negative N meaning show all content. Only
253	** meaningful in combination with --tcl or --tk.
254	** --dark Use dark mode for the Tcl/Tk-based GUI
255	** --tcl FILE Generate (to stdout) a TCL list containing
256	** information needed to display the changes to
257	** FILE caused by the most recent merge. FILE must
258	** be a pathname relative to the root of the check-out.
259	** --tk Bring up a Tcl/Tk GUI that shows the changes
260	** associated with the most recent merge.
261	**
262	*/
263	void merge_info_cmd(void){
264	const char *zCnt;
265	const char *zTcl;
266	int bTk;
267	int bDark;
268	int bAll;
269	int nContext;
270	Stmt q;
271	const char *zWhere;
272	int cnt = 0;
273
274	db_must_be_within_tree();
275	zTcl = find_option("tcl", 0, 1);
276	bTk = find_option("tk", 0, 0)!=0;
277	zCnt = find_option("context", "c", 1);
278	bDark = find_option("dark", 0, 0)!=0;
279	bAll = find_option("all", "a", 0)!=0;
280	if( bTk==0 ){
281	verify_all_options();
282	if( g.argc>2 ){
283	usage("[OPTIONS]");
284	}
285	}
286	if( zCnt ){
287	nContext = atoi(zCnt);
288	if( nContext<0 ) nContext = 0xfffffff;
289	}else{
290	nContext = 6;
291	}
292	if( !db_table_exists("localdb","mergestat") ){
293	if( zTcl ){
294	fossil_print("ERROR {no merge data available}\n");
295	}else{
296	fossil_print("No merge data is available\n");
297	}
298	return;
299	}
300	if( bTk ){
301	merge_info_tk(bDark, bAll, nContext);
302	return;
303	}
304	if( zTcl ){
305	merge_info_tcl(zTcl, nContext);
306	return;
307	}
308	if( bAll ){
309	zWhere = "";
310	}else{
311	zWhere = "WHERE op IN ('MERGE','CONFLICT','ERROR')";
312	}
313	db_prepare(&q,
314	"WITH priority(op,pri) AS (VALUES('CONFLICT',0),('ERROR',0),"
315	"('MERGE',1),('ADDED',2),('UPDATE',2))"
316
317	/* 0 1 2 */
318	"SELECT op, coalesce(fnr,fn), msg"
319	" FROM mergestat JOIN priority USING(op)"
320	" %s"
321	" ORDER BY pri, coalesce(fnr,fn)",
322	zWhere /safe-for-%s/
323	);
324	while( db_step(&q)==SQLITE_ROW ){
325	const char *zOp = db_column_text(&q, 0);
326	const char *zName = db_column_text(&q, 1);
327	const char *zErr = db_column_text(&q, 2);
328	if( zErr && fossil_strcmp(zOp,"CONFLICT")!=0 ){
329	fossil_print("%-9s %s (%s)\n", zOp, zName, zErr);
330	}else{
331	fossil_print("%-9s %s\n", zOp, zName);
332	}
333	cnt++;
334	}
335	db_finalize(&q);
336	if( !bAll && cnt==0 ){
337	fossil_print(
338	"No interesting changes in this merge. Use --all to see everything.\n"
339	);
340	}
341	}
342
343	/*
344	** Erase all information about prior merges. Do this, for example, after
345	** a commit.
346	*/
347	void merge_info_forget(void){
348	db_multi_exec(
349	"DROP TABLE IF EXISTS localdb.mergestat;"
350	"DELETE FROM localdb.vvar WHERE name glob 'mergestat-*';"
351	);
352	}
353
354
355	/*
356	** Initialize the MERGESTAT table.
357	**
358	** Notes about mergestat:
359	**
360	** * ridv is a positive integer and sz is NULL if the V file contained
361	** no local edits prior to the merge. If the V file was modified prior
362	** to the merge then ridv is NULL and sz is the size of the file prior
363	** to merge.
364	**
365	** * fnp, ridp, fn, ridv, and sz are all NULL for a file that was
366	** added by merge.
367	*/
368	void merge_info_init(void){
369	merge_info_forget();
370	db_multi_exec(
371	"CREATE TABLE localdb.mergestat(\n"
372	" op TEXT, -- 'UPDATE', 'ADDED', 'MERGE', etc...\n"
373	" fnp TEXT, -- Name of the pivot file (P)\n"
374	" ridp INT, -- RID for the pivot file\n"
375	" fn TEXT, -- Name of origin file (V)\n"
376	" ridv INT, -- RID for origin file, or NULL if previously edited\n"
377	" sz INT, -- Size of origin file in bytes, NULL if unedited\n"
378	" fnm TEXT, -- Name of the file being merged in (M)\n"
379	" ridm INT, -- RID for the merge-in file\n"
380	" fnr TEXT, -- Name of the final output file, after all renaming\n"
381	" nc INT DEFAULT 0, -- Number of conflicts\n"
382	" msg TEXT -- Error message\n"
383	");"
384	);
385	}
386
387	/*
388	** Print information about a particular check-in.
389	*/
390	void print_checkin_description(int rid, int indent, const char *zLabel){
391	Stmt q;
	@@ -295,10 +657,13 @@
657	** Files which are renamed in the merged-in branch will be renamed in
658	** the current check-out.
659	**
660	** If the VERSION argument is omitted, then Fossil attempts to find
661	** a recent fork on the current branch to merge.
662	**
663	** Note that this command does not commit the merge, as that is a
664	** separate step.
665	**
666	** If there are multiple VERSION arguments, then each VERSION is merged
667	** (or cherrypicked) in the order that they appear on the command-line.
668	**
669	** Options:
	@@ -320,11 +685,11 @@
685	** --force-missing Force the merge even if there is missing content
686	** --integrate Merged branch will be closed when committing
687	** -K\|--keep-merge-files On merge conflict, retain the temporary files
688	** used for merging, named -baseline, -original,
689	** and *-merge.
690	** -n\|--dry-run Do not actually change files on disk
691	** --nosync Do not auto-sync prior to merging
692	** -v\|--verbose Show additional details of the merge
693	*/
694	void merge_cmd(void){
695	int vid; /* Current version "V" */
	@@ -797,15 +1162,21 @@
1162
1163	/************************************************************************
1164	** All of the information needed to do the merge is now contained in the
1165	** FV table. Starting here, we begin to actually carry out the merge.
1166	**
1167	** Begin by constructing the localdb.mergestat table.
1168	*/
1169	merge_info_init();
1170
1171	/*
1172	** Find files that have changed from P->M but not P->V.
1173	** Copy the M content over into V.
1174	*/
1175	db_prepare(&q,
1176	/* 0 1 2 3 4 5 6 7 */
1177	"SELECT idv, ridm, fn, islinkm, fnp, ridp, ridv, fnm FROM fv"
1178	" WHERE idp>0 AND idv>0 AND idm>0"
1179	" AND ridm!=ridp AND ridv=ridp AND NOT chnged"
1180	);
1181	while( db_step(&q)==SQLITE_ROW ){
1182	int idv = db_column_int(&q, 0);
	@@ -822,10 +1193,21 @@
1193	" THEN (SELECT uuid FROM blob WHERE blob.rid=%d) END"
1194	" WHERE id=%d", ridm, integrateFlag?4:2, islinkm, ridm, ridm, idv
1195	);
1196	vfile_to_disk(0, idv, 0, 0);
1197	}
1198	db_multi_exec(
1199	"INSERT INTO mergestat(op,fnp,ridp,fn,ridv,fnm,ridm,fnr)"
1200	"VALUES('UPDATE',%Q,%d,%Q,%d,%Q,%d,%Q)",
1201	/* fnp */ db_column_text(&q, 4),
1202	/* ridp */ db_column_int(&q,5),
1203	/* fn */ zName,
1204	/* ridv */ db_column_int(&q,6),
1205	/* fnm */ db_column_text(&q, 7),
1206	/* ridm */ ridm,
1207	/* fnr */ zName
1208	);
1209	}
1210	db_finalize(&q);
1211
1212	/*
1213	** Do a three-way merge on files that have changes on both P->M and P->V.
	@@ -833,11 +1215,15 @@
1215	** Proceed even if the file doesn't exist on P, just like the common ancestor
1216	** of M and V is an empty file. In this case, merge conflict marks will be
1217	** added to the file and user will be forced to take a decision.
1218	*/
1219	db_prepare(&q,
1220	/* 0 1 2 3 4 5 6 7 8 */
1221	"SELECT ridm, idv, ridp, ridv, %z, fn, isexe, islinkv, islinkm,"
1222	/* 9 10 11 */
1223	" fnp, fnm, chnged"
1224	" FROM fv"
1225	" WHERE idv>0 AND idm>0"
1226	" AND ridm!=ridp AND (ridv!=ridp OR chnged)",
1227	glob_expr("fv.fn", zBinGlob)
1228	);
1229	while( db_step(&q)==SQLITE_ROW ){
	@@ -848,12 +1234,14 @@
1234	int isBinary = db_column_int(&q, 4);
1235	const char *zName = db_column_text(&q, 5);
1236	int isExe = db_column_int(&q, 6);
1237	int islinkv = db_column_int(&q, 7);
1238	int islinkm = db_column_int(&q, 8);
1239	int chnged = db_column_int(&q, 11);
1240	int rc;
1241	char *zFullPath;
1242	const char *zType = "MERGE";
1243	Blob m, p, r;
1244	/* Do a 3-way merge of idp->idm into idp->idv. The results go into idv. */
1245	if( verboseFlag ){
1246	fossil_print("MERGE %s (pivot=%d v1=%d v2=%d)\n",
1247	zName, ridp, ridm, ridv);
	@@ -861,13 +1249,29 @@
1249	fossil_print("MERGE %s\n", zName);
1250	}
1251	if( islinkv \|\| islinkm ){
1252	fossil_print("***** Cannot merge symlink %s\n", zName);
1253	nConflict++;
1254	db_multi_exec(
1255	"INSERT INTO mergestat(op,fnp,ridp,fn,ridv,fnm,ridm,fnr,nc,msg)"
1256	"VALUES('ERROR',%Q,%d,%Q,%d,%Q,%d,%Q,1,'cannot merge symlink')",
1257	/* fnp */ db_column_text(&q, 9),
1258	/* ridp */ ridp,
1259	/* fn */ zName,
1260	/* ridv */ ridv,
1261	/* fnm */ db_column_text(&q, 10),
1262	/* ridm */ ridm,
1263	/* fnr */ zName
1264	);
1265	}else{
1266	i64 sz;
1267	const char *zErrMsg = 0;
1268	int nc = 0;
1269
1270	if( !dryRunFlag ) undo_save(zName);
1271	zFullPath = mprintf("%s/%s", g.zLocalRoot, zName);
1272	sz = file_size(zFullPath, ExtFILE);
1273	content_get(ridp, &p);
1274	content_get(ridm, &m);
1275	if( isBinary ){
1276	rc = -1;
1277	blob_zero(&r);
	@@ -884,15 +1288,38 @@
1288	db_multi_exec("UPDATE vfile SET mtime=0 WHERE id=%d", idv);
1289	if( rc>0 ){
1290	fossil_print("***** %d merge conflict%s in %s\n",
1291	rc, rc>1 ? "s" : "", zName);
1292	nConflict++;
1293	nc = rc;
1294	zErrMsg = "merge conflicts";
1295	zType = "CONFLICT";
1296	}
1297	}else{
1298	fossil_print("***** Cannot merge binary file %s\n", zName);
1299	nConflict++;
1300	nc = 1;
1301	zErrMsg = "cannot merge binary file";
1302	zType = "ERROR";
1303	}
1304	db_multi_exec(
1305	"INSERT INTO mergestat(op,fnp,ridp,fn,ridv,sz,fnm,ridm,fnr,nc,msg)"
1306	"VALUES(%Q,%Q,%d,%Q,iif(%d,%d,NULL),iif(%d,%lld,NULL),%Q,%d,"
1307	"%Q,%d,%Q)",
1308	/* op */ zType,
1309	/* fnp */ db_column_text(&q, 9),
1310	/* ridp */ ridp,
1311	/* fn */ zName,
1312	/* ridv */ chnged==0, ridv,
1313	/* sz */ chnged!=0, sz,
1314	/* fnm */ db_column_text(&q, 10),
1315	/* ridm */ ridm,
1316	/* fnr */ zName,
1317	/* nc */ nc,
1318	/* msg */ zErrMsg
1319	);
1320	fossil_free(zFullPath);
1321	blob_reset(&p);
1322	blob_reset(&m);
1323	blob_reset(&r);
1324	}
1325	vmerge_insert(idv, ridm);
	@@ -901,22 +1328,33 @@
1328
1329	/*
1330	** Drop files that are in P and V but not in M
1331	*/
1332	db_prepare(&q,
1333	"SELECT idv, fn, chnged, ridv FROM fv"
1334	" WHERE idp>0 AND idv>0 AND idm=0"
1335	);
1336	while( db_step(&q)==SQLITE_ROW ){
1337	int idv = db_column_int(&q, 0);
1338	const char *zName = db_column_text(&q, 1);
1339	int chnged = db_column_int(&q, 2);
1340	int ridv = db_column_int(&q, 3);
1341	int sz = -1;
1342	const char *zErrMsg = 0;
1343	int nc = 0;
1344	/* Delete the file idv */
1345	fossil_print("DELETE %s\n", zName);
1346	if( chnged ){
1347	char *zFullPath;
1348	fossil_warning("WARNING: local edits lost for %s", zName);
1349	nConflict++;
1350	ridv = 0;
1351	nc = 1;
1352	zErrMsg = "local edits lost";
1353	zFullPath = mprintf("%s/%s", g.zLocalRoot, zName);
1354	sz = file_size(zFullPath, ExtFILE);
1355	fossil_free(zFullPath);
1356	}
1357	if( !dryRunFlag ) undo_save(zName);
1358	db_multi_exec(
1359	"UPDATE vfile SET deleted=1 WHERE id=%d", idv
1360	);
	@@ -923,10 +1361,20 @@
1361	if( !dryRunFlag ){
1362	char *zFullPath = mprintf("%s%s", g.zLocalRoot, zName);
1363	file_delete(zFullPath);
1364	free(zFullPath);
1365	}
1366	db_multi_exec(
1367	"INSERT INTO localdb.mergestat(op,fnp,ridp,fn,ridv,sz,fnm,ridm,nc,msg)"
1368	"VALUES('DELETE',NULL,NULL,%Q,iif(%d,%d,NULL),iif(%d,%d,NULL),"
1369	"NULL,NULL,%d,%Q)",
1370	/* fn */ zName,
1371	/* ridv */ chnged==0, ridv,
1372	/* sz */ chnged!=0, sz,
1373	/* nc */ nc,
1374	/* msg */ zErrMsg
1375	);
1376	}
1377	db_finalize(&q);
1378
1379	/* For certain sets of renames (e.g. A -> B and B -> A), a file that is
1380	** being renamed must first be moved to a temporary location to avoid
	@@ -953,10 +1401,14 @@
1401	const char *zNewName = db_column_text(&q, 2);
1402	int isExe = db_column_int(&q, 3);
1403	fossil_print("RENAME %s -> %s\n", zOldName, zNewName);
1404	if( !dryRunFlag ) undo_save(zOldName);
1405	if( !dryRunFlag ) undo_save(zNewName);
1406	db_multi_exec(
1407	"UPDATE mergestat SET fnr=fnm WHERE fnp=%Q",
1408	zOldName
1409	);
1410	db_multi_exec(
1411	"UPDATE vfile SET pathname=NULL, origname=pathname"
1412	" WHERE vid=%d AND pathname=%Q;"
1413	"UPDATE vfile SET pathname=%Q, origname=coalesce(origname,pathname)"
1414	" WHERE id=%d;",
	@@ -1006,11 +1458,11 @@
1458
1459	/*
1460	** Insert into V any files that are not in V or P but are in M.
1461	*/
1462	db_prepare(&q,
1463	"SELECT idm, fnm, ridm FROM fv"
1464	" WHERE idp=0 AND idv=0 AND idm>0"
1465	);
1466	while( db_step(&q)==SQLITE_ROW ){
1467	int idm = db_column_int(&q, 0);
1468	const char *zName;
	@@ -1039,10 +1491,17 @@
1491	nOverwrite++;
1492	}else{
1493	fossil_print("ADDED %s\n", zName);
1494	}
1495	fossil_free(zFullName);
1496	db_multi_exec(
1497	"INSERT INTO mergestat(op,fnm,ridm,fnr)"
1498	"VALUES('ADDED',%Q,%d,%Q)",
1499	/* fnm */ zName,
1500	/* ridm */ db_column_int(&q,2),
1501	/* fnr */ zName
1502	);
1503	if( !dryRunFlag ){
1504	undo_save(zName);
1505	vfile_to_disk(0, idm, 0, 0);
1506	}
1507	}
1508
1509	DDED src/merge.tcl

M src/merge.tcl

+581

		--- a/src/merge.tcl
		+++ b/src/merge.tcl
		@@ -0,0 +1,581 @@
	1	+# Show details of a 3-way merge operation. The left-most column is the
	2	+# common ancestor. The next two columns are edits of that common ancestor.
	3	+# The right-most column is the result of the merge.
	4	+#
	5	+# There is always a "fossilcmd" variable which tells the script how to
	6	+# invoke Fossil to get the information it needs. This script will
	7	+# automatically append "-c N" to tell Fossil how much context it wants. True for debugging output
	8	+#
	9	+# If the "filelist" global variable is defined, then it is a list of
	10	+# alternating "merge-type names" (ex: UPDATE, MERGE, CONFLICT, ERROR) and
	11	+# filenames. In that case, the initial display shows the changes for
	12	+# the first pair on the list and there is a optionmenu that allows the
	13	+# useelect otheere should also be a global variable named "ncontext" which is the
	14	+# number of lines of context to display. The value of this variable
	15	+# controls the "-c N" argument that is appended to fossilcmdht {
	16	+ TITLE {Fossil Merge}
	17	+ LN_COL_BG #dddddd
	18	+ LN_COL_FG #444444
	19	+ TXT_COL_BG #ffffff
	20	+ TXT_COL_FG #000000
	21	+ MKR_COL_BG #444444
	22	+ MKR_COL_FG #dddddd
	23	+ CHNG_BG #d0d070
	24	+ ADD_BG #c0ffc0
	25	+ RM_BG #ffc0c0
	26	+ HR_FG #444444
	27	+ HR_PAD_TOP 4
	28	+ HR_PAD_BTM 8
	29	+ FN_BG #444444
	30	+ FN_FG #ffffff
	31	+ FN_PAD 5
	32	+ ERR_FG #ee0000
	33	+ PADX 5
	34	+ WIDTH 80
	35	+ HEIGHT 45
	36	+ LB_HEIGHT 25
	37	+}
	38	+
	39	+array set CFG_dark {
	40	+ TITLE {Fossil Merge}
	41	+ LN_COL_BG #dddddd
	42	+ LN_COL_FG #444444
	43	+ TXT_COL_BG #3f3f3f
	44	+ TXT_COL_FG #dcdccc
	45	+ MKR_COL_BG #444444
	46	+ MKR_COL_FG #dddddd
	47	+ CHNG_BG #6a6a00
	48	+ ADD_BG #57934c
	49	+ RM_BG #ef6767
	50	+ HR_FG #444444
	51	+ HR_PAD_TOP 4
	52	+ HR_PAD_BTM 8
	53	+ FN_BG #5e5e5e
	54	+ FN_FG #ffffff
	55	+ FN_PAD 5
	56	+ ERR_FG #ee0000
	57	+ PADX 5
	58	+ WIDTH 80
	59	+ HEIGHT 45
	60	+ LB_HEIGHT 25
	61	+}
	62	+
	63	+array set CFG_arr {
	64	+ 0 CFG_light
	65	+ 1 CFG_dark
	66	+}
	67	+
	68	+array set CFG [array get $CFG_arr($darkmode)]
	69	+
	70	+if {![namespace exists ttk]} {
	71	+ interp alias {} ::ttk::scrollbar {} ::scrollbar
	72	+ interp alias {} ::ttk::menubutton {} ::menubutton
	73	+}
	74	+
	75	+proc dehtml {x} {
	76	+ set x [regsub -all {<[^>]*>} $x {}]
	77	+ return [string map {& & < < > > ' ' " \"} $x]
	78	+}
	79	+
	80	+proc cols {} {
	81	+ return [list .lnA .txtA .lnB .txtB .lnC .txtC .lnD .txtD]
	82	+}
	83	+
	84	+proc colType {c} {
	85	+ regexp {[a-z]+} $c type
	86	+ return $type
	87	+}
	88	+
	89	+proc readMercmdes currentails of a 3-way merge operation# Show det for
	90	+# the first pair on the list and there is a optionmenu that allows the
	91	+# user to select other fiels on the list.
	92	+#
	93	+# This header comment is stripped off by the "mkbuiltin.c" program.
	94	+#
	95	+package require Tk
	96	+
	97	+array set CFG_light {
	98	+ TITLE {Fossil Merge}
	99	+ LN_COL_BG #dddddd
	100	+ LN_COL_FG #444444
	101	+ TXT_COL_BG #ffffff
	102	+ TXT_COL_FG #000000
	103	+ MKR_COL_BG #444444
	104	+ MKR_COL_FG #dddddd
	105	+ CHNG_BG #d0d070
	106	+ ADD_BG #c0ffc0
	107	+ RM_BG #ffc0c0
	108	+ HR_FG #444444
	109	+ HR_PAD_TOP 4
	110	+ HR_PAD_BTM 8
	111	+ FN_BG #444444
	112	+ FN_FG #ffffff
	113	+ FN_PAD 5
	114	+ ERR_FG #ee0000
	115	+ PADX 5
	116	+ WIDTH 80
	117	+ HEIGHT 45
	118	+ LB_HEIGHT 25
	119	+}
	120	+
	121	+array set CFG_dark {
	122	+ TITLE {Fossil Merge}
	123	+ LN_COL_BG #dddddd
	124	+ LN_COL_FG #444444
	125	+ TXT_COL_BG #3f3f3f
	126	+ TXT_COL_FG #dcdccc
	127	+ MKR_COL_BG #444444
	128	+ MKR_COL_FG #dddddd
	129	+ CHNG_BG #6a6a00
	130	+ ADD_BG #57934c
	131	+ RM_BG #ef6767
	132	+ HR_FG #444444
	133	+ HR_PAD_TOP 4
	134	+ HR_PAD_BTM 8
	135	+ FN_BG #5e5e5e
	136	+ FN_FG #ffffff
	137	+ FN_PAD 5
	138	+ ERR_FG #ee0000
	139	+ PADX 5
	140	+ WIDTH 80
	141	+ HEIGHT 45
	142	+ LB_HEIGHT 25
	143	+}
	144	+
	145	+array set CFG_arr {
	146	+ 0 CFG_light
	147	+ 1 CFG_dark
	148	+}
	149	+
	150	+array set CFG [array get $CFG_arr($darkmode)]
	151	+
	152	+if {![namespace exists ttk]} {
	153	+ interp alias {} ::ttk::scrollbar {} ::scrollbar
	154	+ interp alias {} ::ttk::menubutton {} ::menubutton
	155	+}
	156	+
	157	+proc dehtml {x} {
	158	+ set x [regsub -all {<[^>]*>} $x {}]
	159	+ return [string map {& & < < > > ' ' " \"} $x]
	160	+}
	161	+
	162	+proc cols {} {
	163	+ return [list .lnA .txtA .lnB .txtB .lnC .txtC .lnD .txtD]
	164	+}
	165	+
	166	+proc colType {c} {
	167	+ regexp {[a-z]+} $c type
	168	+ return $type
	169	+}
	170	+
	171	+proc readMerge {args} {
	172	+ global fossilexe ncontext current_file debug
	173	+ if {$ncontext=="All"} {
	174	+ set cmd "\| $fossilexe merge-info -c -1"
	175	+ } else {
	176	+ set cmd "\| $fossilexe merge-info -c $ncontext"
	177	+ }
	178	+ if {[info exists current_file]} {
	179	+ regsub {^[A-Z]+ } $current_file {} fn
	180	+ lappend cmd -tcl $fn
	181	+ }
	182	+ if {$debug} {
	183	+ regsub {^\\| +} $cmd {} cmd2
	184	+ puts $cmd2
	185	+ flush stdout
	186	+ }
	187	+ if {[catch {
	188	+ set in [open $cmd r]
	189	+ fconfigure $in -encoding utf-8
	190	+ set mergetxt [read $in]
	191	+ close $in
	192	+ } msg]} {
	193	+ tk_messageBox -message "Unable to run command: \"$cmd\""
	194	+ set mergetxt {}
	195	+ }
	196	+ foreach c [cols] {
	197	+ $c config -state normal
	198	+ $c delete 1.0 end
	199	+ }
	200	+ set lnA 1
	201	+ set lnB 1
	202	+ set lnC 1
	203	+ set lnD 1
	204	+ foreach {A B C D} $mergetxt {
	205	+ set key1 [string index $A 0]
	206	+ if {$key1=="S"} {
	207	+ scan [string range $A 1 end] "%d %d %d %d" nA nB nC nD
	208	+ foreach x {A B C D} {
	209	+ set N [set n$x]
	210	+ incr ln$x $N
	211	+ if {$N>0} {
	212	+ .ln$x insert end ...\n hrln
	213	+ .txt$x insert end [string repeat . 30]\n hrtxt
	214	+ } else {
	215	+ .ln$x insert end \n hrln
	216	+ .txt$x insert end \n hrtxt
	217	+ }
	218	+ }
	219	+ continue
	220	+ }
	221	+ set key2 [string index $B 0]
	222	+ set key3 [string index $C 0]
	223	+ set key4 [string index $D 0]
	224	+ if {$key1=="."} {
	225	+ .lnA insert end \n -
	226	+ .txtA insert end \n -
	227	+ } elseif {$key1=="N"} {
	228	+ .nameA config -text [string range $A 1 end]
	229	+ } else {
	230	+ .lnA insert end $lnA\n -
	231	+ incr lnA
	232	+ if {$key1=="X"} {
	233	+ .txtA insert end [string range $A 1 end]\n rm
	234	+ } else {
	235	+ .txtA insert end [string range $A 1 end]\n -
	236	+ }
	237	+ }
	238	+ if {$key2=="."} {
	239	+ .lnB insert end \n -
	240	+ .txtB insert end \n -
	241	+ } elseif {$key2=="N"} {
	242	+ .nameB config -text [string range $B 1 end]
	243	+ } else {
	244	+ .lnB insert end $lnB\n -
	245	+ incr lnB
	246	+ if {$key4=="2"} {set tag chng} {set tag -}
	247	+ if {$key2=="1"} {
	248	+ .txtB insert end [string range $A 1 end]\n $tag
	249	+ } elseif {$key2=="X"} {
	250	+ .txtB insert end [string range $B 1 end]\n rm
	251	+ } else {
	252	+ .txtB insert end [string range $B 1 end]\n $tag
	253	+ }
	254	+ }
	255	+ if {$key3=="."} {
	256	+ .lnC insert end \n -
	257	+ .txtC insert end \n -
	258	+ } elseif {$key3=="N"} {
	259	+ .nameC config -text [string range $C 1 end]
	260	+ } else {
	261	+ .lnC insert end $lnC\n -
	262	+ incr lnC
	263	+ if {$key4=="3"} {set tag add} {set tag -}
	264	+ if {$key3=="1"} {
	265	+ .txtC insert end [string range $A 1 end]\n $tag
	266	+ } elseif {$key3=="2"} {
	267	+ .txtC insert end [string range $B 1 end]\n chng
	268	+ } elseif {$key3=="X"} {
	269	+ .txtC insert end [string range $C 1 end]\n rm
	270	+ } else {
	271	+ .txtC insert end [string range $C 1 end]\n $tag
	272	+ }
	273	+ }
	274	+ if {$key4=="."} {
	275	+ .lnD insert end \n -
	276	+ .txtD insert end \n -
	277	+ } elseif {$key4=="N"} {
	278	+ .nameD config -text [string range $D 1 end]
	279	+ } else {
	280	+ .lnD insert end $lnD\n -
	281	+ incr lnD
	282	+ if {$key4=="1"} {
	283	+ .txtD insert end [string range $A 1 end]\n -
	284	+ } elseif {$key4=="2"} {
	285	+ .txtD insert end [string range $B 1 end]\n chng
	286	+ } elseif {$key4=="3"} {
	287	+ .txtD insert end [string range $C 1 end]\n add
	288	+ } elseif {$key4=="X"} {
	289	+ .txtD insert end [string range $D 1 end]\n rm
	290	+ } else {
	291	+ .txtD insert end [string range $D 1 end]\n -
	292	+ }
	293	+ }
	294	+ }
	295	+ foreach c [cols] {
	296	+ set type [colType $c]
	297	+ if {$type ne "txt"} {
	298	+ $c config -width 6; # $widths($type)
	299	+ }
	300	+ $c config -state disabled
	301	+ }
	302	+ set mx $lnA
	303	+ if {$lnB>$mx} {set mx $lnB}
	304	+ if {$lnC>$mx} {set mx $lnC}
	305	+ if {$lnD>$mx} {set mx $lnD}
	306	+ global lnWidth
	307	+ set lnWidth [string length [format +%d $mx]]
	308	+ .lnA config -width $lnWidth
	309	+ .lnB config -width $lnWidth
	310	+ .lnC config -width $lnWidth
	311	+ .lnD config -width $lnWidth
	312	+ grid columnconfig . {0 2 4 6} -minsize $lnWidth
	313	+}
	314	+
	315	+proc viewDiff {idx} {
	316	+ .txtA yview $idx
	317	+ .txtA xview moveto 0
	318	+}
	319	+
	320	+proc cycleDiffs {{reverse 0}} {
	321	+ if {$reverse} {
	322	+ set range [.txtA tag prevrange fn @0,0 1.0]
	323	+ if {$range eq ""} {
	324	+ viewDiff {fn.last -1c}
	325	+ } else {
	326	+ viewDiff [lindex $range 0]
	327	+ }
	328	+ } else {
	329	+ set range [.txtA tag nextrange fn {@0,0 +1c} end]
	330	+ if {$range eq "" \|\| [lindex [.txtA yview] 1] == 1} {
	331	+ viewDiff fn.first
	332	+ } else {
	333	+ viewDiff [lindex $range 0]
	334	+ }
	335	+ }
	336	+}
	337	+
	338	+proc xvis {col} {
	339	+ set view [$col xview]
	340	+ return [expr {[lindex $view 1]-[lindex $view 0]}]
	341	+}
	342	+
	343	+proc scroll-x {args} {
	344	+ set c .txt[expr {[xvis .txtA] < [xvis .txtB] ? "A" : "B"}]
	345	+ eval $c xview $args
	346	+}
	347	+
	348	+interp alias {} scroll-y {} .txtA yview
	349	+
	350	+proc noop {args} {}
	351	+
	352	+proc enableSync {axis} {
	353	+ update idletasks
	354	+ interp alias {} sync-$axis {}
	355	+ rename _sync-$axis sync-$axis
	356	+}
	357	+
	358	+proc disableSync {axis} {
	359	+ rename sync-$axis _sync-$axis
	360	+ interp alias {} sync-$axis {} noop
	361	+}
	362	+
	363	+proc sync-y {first last} {
	364	+ disableSync y
	365	+ foreach c [cols] {
	366	+ $c yview moveto $first
	367	+ }
	368	+ if {$first > 0 \|\| $last < 1} {
	369	+ grid .sby
	370	+ .sby set $first $last
	371	+ } else {
	372	+ grid remove .sby
	373	+ }
	374	+ enableSync y
	375	+}
	376	+
	377	+wm withdraw .
	378	+wm title . $CFG(TITLE)
	379	+wm iconname . $CFG(TITLE)
	380	+# Keystroke bindings for on the top-level window for navigation and
	381	+# control also fire when those same keystrokes are pressed in the
	382	+# Search entry box. Disable them, to prevent the diff screen from
	383	+# disappearing abruptly and unexpectedly when searching for "q".
	384	+#
	385	+bind . <Control-q> exit
	386	+bind . <Control-p> {catch searchPrev; break}
	387	+bind . <Control-n> {catch searchNext; break}
	388	+bind . <Escape><Escape> exit
	389	+bind . <Destroy> {after 0 exit}
	390	+bind . <Tab> {cycleDiffs; break}
	391	+bind . <<PrevWindow>> {cycleDiffs 1; break}
	392	+bind . <Control-f> {searchOnOff; break}
	393	+bind . <Control-g> {catch searchNext; break}
	394	+bind . <Return> {
	395	+ event generate .bb.files <1>
	396	+ event generate .bb.files <ButtonRelease-1>
	397	+ break
	398	+}
	399	+foreach {key axis args} {
	400	+ Up y {scroll -5 units}
	401	+ k y {scroll -5 units}
	402	+ Down y {scroll 5 units}
	403	+ j y {scroll 5 units}
	404	+ Left x {scroll -5 units}
	405	+ h x {scroll -5 units}
	406	+ Right x {scroll 5 units}
	407	+ l x {scroll 5 units}
	408	+ Prior y {scroll -1 page}
	409	+ b y {scroll -1 page}
	410	+ Next y {scroll 1 page}
	411	+ space y {scroll 1 page}
	412	+ Home y {moveto 0}
	413	+ g y {moveto 0}
	414	+ End y {moveto 1}
	415	+} {
	416	+ bind . <$key> "scroll-$axis $args; break"
	417	+ bind . <Shift-$key> continue
	418	+}
	419	+
	420	+frame .bb
	421	+::ttk::menubutton .bb.diff2 -text {2-way diffs
	422	+.bb.diff2.m add command -label {baseline vs. local} -command {two-way 12}
	423	+.bb.diff2.m add command -label {baseline vs. merge-in} -command {two-way 13}
	424	+.bb.diff2.m add command -label {local vs. merge-in} -command {two-way 23}
	425	+
	426	+# Bring up a separate two-way diff between a pair of columns
	427	+# the argument is one of:
	428	+# 12 Baseline versus Local
	429	+# 13 Baseline versus Merge-in
	430	+# 23 Local versus Merge-in
	431	+#
	432	+proc two-way {mode} {
	433	+ global current_file fossilexe debug darkmode ncontext
	434	+ regsub {^[A-Z]+ } $current_file {} fn
	435	+ set cmd $fossilexe
	436	+ lappend cmd merge-info --diff$mode $fn -c $ncontext
	437	+ if {$darkmode} {
	438	+ lappend cmd --dark
	439	+ }
	440	+ if {$debug} {
	441	+ lappend cc {*}$cmd &
	442	+}
	443	+
	444	+set useOptionMenu 1
	445	+if {[info exists filelist]} {
	446	+ set current_file "[lindex $filelist 0] [lindex $filelist 1]"
	447	+ if {[llength $filelist]>2} {
	448	+ trace add variable current_file write readMerge
	449	+
	450	+ if {$tcl_platform(os)=="Darwin" \|\| [llength $filelist]<30} {
	451	+ set fnlist {}
	452	+ foreach {op fn} $filelist {lappend fnlist "$op $fn"}
	453	+ tk_optionMenu .bb.files current_file {*}$fnlist
	454	+ } else {
	455	+ set useOptionMenu 0
	456	+ ::ttk::menubutton .bb.files -text $current_file
	457	+ if {[tk windowingsystem] eq "win32"} {
	458	+ ::ttk::style theme use winnative
	459	+ .bb.files configure -padding {20 1 10 2}
	460	+ }
	461	+ toplevel .wfiles
	462	+ wm withdraw .wfiles
	463	+ update idletasks
	464	+ wm transient .wfiles .
	465	+ wm overrideredirect .wfiles 1
	466	+ set ht [expr {[llength $filelist]/2}]
	467	+ if {$ht>$CFG(LB_HEIGHT)} {set ht $CFG(LB_HEIGHT)}
	468	+ listbox .wfiles.lb -width 0 -height $ht -activestyle none \
	469	+ -yscroll {.wfiles.sb set}
	470	+ set mx 1
	471	+ foreach {op fn} $filelist {
	472	+ set n [string length $fn]
	473	+ if {$n>$mx} {set mx $n}
	474	+ .wfiles.lb insert end "$op $fn"
	475	+ }
	476	+ .bb.files config -width $mx
	477	+ ::ttk::scrollbar .wfiles.sb -command {.wfiles.lb yview}
	478	+ grid .wfiles.lb .wfiles.sb -sticky ns
	479	+ bind .bb.files <1> {
	480	+ set x [winfo rootx %W]
	481	+ set y [expr {[winfo rooty %W]+[winfo height %W]}]
	482	+ wm geometry .wfiles +$x+$y
	483	+ wm deiconify .wfiles
	484	+ focus .wfiles.lb
	485	+ }
	486	+ bind .wfiles <FocusOut> {wm withdraw .wfiles}
	487	+ bind .wfiles <Escape> {focus .}
	488	+ foreach evt {1 Return} {
	489	+ bind .wfiles.lb <$evt> {
	490	+ set ii [%W curselection]
	491	+ set ::current_file [%W get $ii]
	492	+ .bb.files config -text $::current_file
	493	+ focus .
	494	+ break
	495	+ }
	496	+ }
	497	+ bind .wfiles.lb <Motion> {
	498	+ %W selection clear 0 end
	499	+ %W selection set @%x,%y
	500	+ }
	501	+ }
	502	+ }
	503	+}
	504	+
	505	+label .bb.ctxtag -text "Context:"
	506	+set context_choices {3 6 12 25 50 100 All}
	507	+if {$ncontext<0} {set ncontext All}
	508	+trace add variable ncontext write readMerge
	509	+if {$tcl_platform(os)=="Darwin" \|\| $useOptionMenu} {
	510	+ tk_optionMenu .bb.ctx ncontext {*}$context_choices
	511	+} else {
	512	+ ::ttk::menubutton .bb.ctx -text $ncontext
	513	+ if {[tk windowingsystem] eq "win32"} {
	514	+ ::ttk::style theme use winnative
	515	+ .bb.ctx configure -padding {20 1 10 2}
	516	+ }
	517	+ toplevel .wctx
	518	+ wm withdraw .wctx
	519	+ update idletasks
	520	+ wm transient .wctx .
	521	+ wm overrideredirect .wctx 1
	522	+ listbox .wctx.lb -width 0 -height 7 -activestyle none
	523	+ .wctx.lb insert end {*}$context_choices
	524	+ pack .wctx.lb
	525	+ bind .bb.ctx <1> {
	526	+ set x [winfo rootx %W]
	527	+ set y [expr {[winfo rooty %W]+[winfo height %W]}]
	528	+ wm geometry .wctx +$x+$y
	529	+ wm deiconify .wctx
	530	+ focus .wctx.lb
	531	+ }
	532	+ bind .wctx <FocusOut> {wm withdraw .wctx}
	533	+ bind .wctx <Escape> {focus .}
	534	+ foreach evt {1 Return} {
	535	+ bind .wctx.lb <$evt> {
	536	+ set ::ncontext [lindex $::context_choices [%W curselection]]
	537	+ .bb.ctx config -text $::ncontext
	538	+ focus .
	539	+ break
	540	+ }
	541	+ }
	542	+ bind .wctx.lb <Motion> {
	543	+ %W selection clear 0 end
	544	+ %W selection set @%x,%y
	545	+ }
	546	+}
	547	+
	548	+foreach {side syncCol} {A .txtA B .txtB C .txtC D .txtD} {
	549	+ set ln .ln$side
	550	+ text $ln -width 6
	551	+ $ln tag config - -justify right
	552	+
	553	+ set txt .txt$side
	554	+ text $txt -width $CFG(WIDTH) -height $CFG(HEIGHT) -wrap none \
	555	+ -xscroll ".sbx$side set"
	556	+ catch {$txt config -tabstyle wordprocessor} ;# Required for Tk>=8.5
	557	+ foreach tag {add rm chng} {
	558	+ $txt tag config $tag -background $CFG([string toupper $tag]_BG)
	559	+ $txt tag lower $tag
	560	+ }
	561	+ $txt tag config fn -background $CFG(FN_BG) -foreground $CFG(FN_FG) \
	562	+ -justify center
	563	+ $txt tag config err -foreground $CFG(ERR_FG)
	564	+}
	565	+text .mkr
	566	+
	567	+set mxwidth [lindex [wm maxsize .] 0]
	568	+while {$CFG(WIDTH)>=40} {
	569	+ set wanted [expr {([winfo reqwidth .lnA]+[winfo reqwidth .txtA])*4+30}]
	570	+ if {$wanted<=$mxwidth} break
	571	+ incr CFG(WIDTH) -10
	572	+ .txtA config -width $CFG(WIDTH)
	573	+ .txtB config -width $CFG(WIDTH)
	574	+ .txtC config -width $CFG(WIDTH)
	575	+ .txtD config -width $CFG(WIDTH)
	576	+}
	577	+
	578	+foreach c [cols] {
	579	+ set keyPrefix [string toupper [colType $c]]_COL_
	580	+ if {[tk windowingsystem] eq "win32"} {$c config -font {courier 9}}
	581	+ $c co

	--- a/src/merge.tcl
	+++ b/src/merge.tcl
	@@ -0,0 +1,581 @@

	--- a/src/merge.tcl
	+++ b/src/merge.tcl
	@@ -0,0 +1,581 @@
1	# Show details of a 3-way merge operation. The left-most column is the
2	# common ancestor. The next two columns are edits of that common ancestor.
3	# The right-most column is the result of the merge.
4	#
5	# There is always a "fossilcmd" variable which tells the script how to
6	# invoke Fossil to get the information it needs. This script will
7	# automatically append "-c N" to tell Fossil how much context it wants. True for debugging output
8	#
9	# If the "filelist" global variable is defined, then it is a list of
10	# alternating "merge-type names" (ex: UPDATE, MERGE, CONFLICT, ERROR) and
11	# filenames. In that case, the initial display shows the changes for
12	# the first pair on the list and there is a optionmenu that allows the
13	# useelect otheere should also be a global variable named "ncontext" which is the
14	# number of lines of context to display. The value of this variable
15	# controls the "-c N" argument that is appended to fossilcmdht {
16	TITLE {Fossil Merge}
17	LN_COL_BG #dddddd
18	LN_COL_FG #444444
19	TXT_COL_BG #ffffff
20	TXT_COL_FG #000000
21	MKR_COL_BG #444444
22	MKR_COL_FG #dddddd
23	CHNG_BG #d0d070
24	ADD_BG #c0ffc0
25	RM_BG #ffc0c0
26	HR_FG #444444
27	HR_PAD_TOP 4
28	HR_PAD_BTM 8
29	FN_BG #444444
30	FN_FG #ffffff
31	FN_PAD 5
32	ERR_FG #ee0000
33	PADX 5
34	WIDTH 80
35	HEIGHT 45
36	LB_HEIGHT 25
37	}
38
39	array set CFG_dark {
40	TITLE {Fossil Merge}
41	LN_COL_BG #dddddd
42	LN_COL_FG #444444
43	TXT_COL_BG #3f3f3f
44	TXT_COL_FG #dcdccc
45	MKR_COL_BG #444444
46	MKR_COL_FG #dddddd
47	CHNG_BG #6a6a00
48	ADD_BG #57934c
49	RM_BG #ef6767
50	HR_FG #444444
51	HR_PAD_TOP 4
52	HR_PAD_BTM 8
53	FN_BG #5e5e5e
54	FN_FG #ffffff
55	FN_PAD 5
56	ERR_FG #ee0000
57	PADX 5
58	WIDTH 80
59	HEIGHT 45
60	LB_HEIGHT 25
61	}
62
63	array set CFG_arr {
64	0 CFG_light
65	1 CFG_dark
66	}
67
68	array set CFG [array get $CFG_arr($darkmode)]
69
70	if {![namespace exists ttk]} {
71	interp alias {} ::ttk::scrollbar {} ::scrollbar
72	interp alias {} ::ttk::menubutton {} ::menubutton
73	}
74
75	proc dehtml {x} {
76	set x [regsub -all {<[^>]*>} $x {}]
77	return [string map {& & < < > > ' ' " \"} $x]
78	}
79
80	proc cols {} {
81	return [list .lnA .txtA .lnB .txtB .lnC .txtC .lnD .txtD]
82	}
83
84	proc colType {c} {
85	regexp {[a-z]+} $c type
86	return $type
87	}
88
89	proc readMercmdes currentails of a 3-way merge operation# Show det for
90	# the first pair on the list and there is a optionmenu that allows the
91	# user to select other fiels on the list.
92	#
93	# This header comment is stripped off by the "mkbuiltin.c" program.
94	#
95	package require Tk
96
97	array set CFG_light {
98	TITLE {Fossil Merge}
99	LN_COL_BG #dddddd
100	LN_COL_FG #444444
101	TXT_COL_BG #ffffff
102	TXT_COL_FG #000000
103	MKR_COL_BG #444444
104	MKR_COL_FG #dddddd
105	CHNG_BG #d0d070
106	ADD_BG #c0ffc0
107	RM_BG #ffc0c0
108	HR_FG #444444
109	HR_PAD_TOP 4
110	HR_PAD_BTM 8
111	FN_BG #444444
112	FN_FG #ffffff
113	FN_PAD 5
114	ERR_FG #ee0000
115	PADX 5
116	WIDTH 80
117	HEIGHT 45
118	LB_HEIGHT 25
119	}
120
121	array set CFG_dark {
122	TITLE {Fossil Merge}
123	LN_COL_BG #dddddd
124	LN_COL_FG #444444
125	TXT_COL_BG #3f3f3f
126	TXT_COL_FG #dcdccc
127	MKR_COL_BG #444444
128	MKR_COL_FG #dddddd
129	CHNG_BG #6a6a00
130	ADD_BG #57934c
131	RM_BG #ef6767
132	HR_FG #444444
133	HR_PAD_TOP 4
134	HR_PAD_BTM 8
135	FN_BG #5e5e5e
136	FN_FG #ffffff
137	FN_PAD 5
138	ERR_FG #ee0000
139	PADX 5
140	WIDTH 80
141	HEIGHT 45
142	LB_HEIGHT 25
143	}
144
145	array set CFG_arr {
146	0 CFG_light
147	1 CFG_dark
148	}
149
150	array set CFG [array get $CFG_arr($darkmode)]
151
152	if {![namespace exists ttk]} {
153	interp alias {} ::ttk::scrollbar {} ::scrollbar
154	interp alias {} ::ttk::menubutton {} ::menubutton
155	}
156
157	proc dehtml {x} {
158	set x [regsub -all {<[^>]*>} $x {}]
159	return [string map {& & < < > > ' ' " \"} $x]
160	}
161
162	proc cols {} {
163	return [list .lnA .txtA .lnB .txtB .lnC .txtC .lnD .txtD]
164	}
165
166	proc colType {c} {
167	regexp {[a-z]+} $c type
168	return $type
169	}
170
171	proc readMerge {args} {
172	global fossilexe ncontext current_file debug
173	if {$ncontext=="All"} {
174	set cmd "\| $fossilexe merge-info -c -1"
175	} else {
176	set cmd "\| $fossilexe merge-info -c $ncontext"
177	}
178	if {[info exists current_file]} {
179	regsub {^[A-Z]+ } $current_file {} fn
180	lappend cmd -tcl $fn
181	}
182	if {$debug} {
183	regsub {^\\| +} $cmd {} cmd2
184	puts $cmd2
185	flush stdout
186	}
187	if {[catch {
188	set in [open $cmd r]
189	fconfigure $in -encoding utf-8
190	set mergetxt [read $in]
191	close $in
192	} msg]} {
193	tk_messageBox -message "Unable to run command: \"$cmd\""
194	set mergetxt {}
195	}
196	foreach c [cols] {
197	$c config -state normal
198	$c delete 1.0 end
199	}
200	set lnA 1
201	set lnB 1
202	set lnC 1
203	set lnD 1
204	foreach {A B C D} $mergetxt {
205	set key1 [string index $A 0]
206	if {$key1=="S"} {
207	scan [string range $A 1 end] "%d %d %d %d" nA nB nC nD
208	foreach x {A B C D} {
209	set N [set n$x]
210	incr ln$x $N
211	if {$N>0} {
212	.ln$x insert end ...\n hrln
213	.txt$x insert end [string repeat . 30]\n hrtxt
214	} else {
215	.ln$x insert end \n hrln
216	.txt$x insert end \n hrtxt
217	}
218	}
219	continue
220	}
221	set key2 [string index $B 0]
222	set key3 [string index $C 0]
223	set key4 [string index $D 0]
224	if {$key1=="."} {
225	.lnA insert end \n -
226	.txtA insert end \n -
227	} elseif {$key1=="N"} {
228	.nameA config -text [string range $A 1 end]
229	} else {
230	.lnA insert end $lnA\n -
231	incr lnA
232	if {$key1=="X"} {
233	.txtA insert end [string range $A 1 end]\n rm
234	} else {
235	.txtA insert end [string range $A 1 end]\n -
236	}
237	}
238	if {$key2=="."} {
239	.lnB insert end \n -
240	.txtB insert end \n -
241	} elseif {$key2=="N"} {
242	.nameB config -text [string range $B 1 end]
243	} else {
244	.lnB insert end $lnB\n -
245	incr lnB
246	if {$key4=="2"} {set tag chng} {set tag -}
247	if {$key2=="1"} {
248	.txtB insert end [string range $A 1 end]\n $tag
249	} elseif {$key2=="X"} {
250	.txtB insert end [string range $B 1 end]\n rm
251	} else {
252	.txtB insert end [string range $B 1 end]\n $tag
253	}
254	}
255	if {$key3=="."} {
256	.lnC insert end \n -
257	.txtC insert end \n -
258	} elseif {$key3=="N"} {
259	.nameC config -text [string range $C 1 end]
260	} else {
261	.lnC insert end $lnC\n -
262	incr lnC
263	if {$key4=="3"} {set tag add} {set tag -}
264	if {$key3=="1"} {
265	.txtC insert end [string range $A 1 end]\n $tag
266	} elseif {$key3=="2"} {
267	.txtC insert end [string range $B 1 end]\n chng
268	} elseif {$key3=="X"} {
269	.txtC insert end [string range $C 1 end]\n rm
270	} else {
271	.txtC insert end [string range $C 1 end]\n $tag
272	}
273	}
274	if {$key4=="."} {
275	.lnD insert end \n -
276	.txtD insert end \n -
277	} elseif {$key4=="N"} {
278	.nameD config -text [string range $D 1 end]
279	} else {
280	.lnD insert end $lnD\n -
281	incr lnD
282	if {$key4=="1"} {
283	.txtD insert end [string range $A 1 end]\n -
284	} elseif {$key4=="2"} {
285	.txtD insert end [string range $B 1 end]\n chng
286	} elseif {$key4=="3"} {
287	.txtD insert end [string range $C 1 end]\n add
288	} elseif {$key4=="X"} {
289	.txtD insert end [string range $D 1 end]\n rm
290	} else {
291	.txtD insert end [string range $D 1 end]\n -
292	}
293	}
294	}
295	foreach c [cols] {
296	set type [colType $c]
297	if {$type ne "txt"} {
298	$c config -width 6; # $widths($type)
299	}
300	$c config -state disabled
301	}
302	set mx $lnA
303	if {$lnB>$mx} {set mx $lnB}
304	if {$lnC>$mx} {set mx $lnC}
305	if {$lnD>$mx} {set mx $lnD}
306	global lnWidth
307	set lnWidth [string length [format +%d $mx]]
308	.lnA config -width $lnWidth
309	.lnB config -width $lnWidth
310	.lnC config -width $lnWidth
311	.lnD config -width $lnWidth
312	grid columnconfig . {0 2 4 6} -minsize $lnWidth
313	}
314
315	proc viewDiff {idx} {
316	.txtA yview $idx
317	.txtA xview moveto 0
318	}
319
320	proc cycleDiffs {{reverse 0}} {
321	if {$reverse} {
322	set range [.txtA tag prevrange fn @0,0 1.0]
323	if {$range eq ""} {
324	viewDiff {fn.last -1c}
325	} else {
326	viewDiff [lindex $range 0]
327	}
328	} else {
329	set range [.txtA tag nextrange fn {@0,0 +1c} end]
330	if {$range eq "" \|\| [lindex [.txtA yview] 1] == 1} {
331	viewDiff fn.first
332	} else {
333	viewDiff [lindex $range 0]
334	}
335	}
336	}
337
338	proc xvis {col} {
339	set view [$col xview]
340	return [expr {[lindex $view 1]-[lindex $view 0]}]
341	}
342
343	proc scroll-x {args} {
344	set c .txt[expr {[xvis .txtA] < [xvis .txtB] ? "A" : "B"}]
345	eval $c xview $args
346	}
347
348	interp alias {} scroll-y {} .txtA yview
349
350	proc noop {args} {}
351
352	proc enableSync {axis} {
353	update idletasks
354	interp alias {} sync-$axis {}
355	rename _sync-$axis sync-$axis
356	}
357
358	proc disableSync {axis} {
359	rename sync-$axis _sync-$axis
360	interp alias {} sync-$axis {} noop
361	}
362
363	proc sync-y {first last} {
364	disableSync y
365	foreach c [cols] {
366	$c yview moveto $first
367	}
368	if {$first > 0 \|\| $last < 1} {
369	grid .sby
370	.sby set $first $last
371	} else {
372	grid remove .sby
373	}
374	enableSync y
375	}
376
377	wm withdraw .
378	wm title . $CFG(TITLE)
379	wm iconname . $CFG(TITLE)
380	# Keystroke bindings for on the top-level window for navigation and
381	# control also fire when those same keystrokes are pressed in the
382	# Search entry box. Disable them, to prevent the diff screen from
383	# disappearing abruptly and unexpectedly when searching for "q".
384	#
385	bind . <Control-q> exit
386	bind . <Control-p> {catch searchPrev; break}
387	bind . <Control-n> {catch searchNext; break}
388	bind . <Escape><Escape> exit
389	bind . <Destroy> {after 0 exit}
390	bind . <Tab> {cycleDiffs; break}
391	bind . <<PrevWindow>> {cycleDiffs 1; break}
392	bind . <Control-f> {searchOnOff; break}
393	bind . <Control-g> {catch searchNext; break}
394	bind . <Return> {
395	event generate .bb.files <1>
396	event generate .bb.files <ButtonRelease-1>
397	break
398	}
399	foreach {key axis args} {
400	Up y {scroll -5 units}
401	k y {scroll -5 units}
402	Down y {scroll 5 units}
403	j y {scroll 5 units}
404	Left x {scroll -5 units}
405	h x {scroll -5 units}
406	Right x {scroll 5 units}
407	l x {scroll 5 units}
408	Prior y {scroll -1 page}
409	b y {scroll -1 page}
410	Next y {scroll 1 page}
411	space y {scroll 1 page}
412	Home y {moveto 0}
413	g y {moveto 0}
414	End y {moveto 1}
415	} {
416	bind . <$key> "scroll-$axis $args; break"
417	bind . <Shift-$key> continue
418	}
419
420	frame .bb
421	::ttk::menubutton .bb.diff2 -text {2-way diffs
422	.bb.diff2.m add command -label {baseline vs. local} -command {two-way 12}
423	.bb.diff2.m add command -label {baseline vs. merge-in} -command {two-way 13}
424	.bb.diff2.m add command -label {local vs. merge-in} -command {two-way 23}
425
426	# Bring up a separate two-way diff between a pair of columns
427	# the argument is one of:
428	# 12 Baseline versus Local
429	# 13 Baseline versus Merge-in
430	# 23 Local versus Merge-in
431	#
432	proc two-way {mode} {
433	global current_file fossilexe debug darkmode ncontext
434	regsub {^[A-Z]+ } $current_file {} fn
435	set cmd $fossilexe
436	lappend cmd merge-info --diff$mode $fn -c $ncontext
437	if {$darkmode} {
438	lappend cmd --dark
439	}
440	if {$debug} {
441	lappend cc {*}$cmd &
442	}
443
444	set useOptionMenu 1
445	if {[info exists filelist]} {
446	set current_file "[lindex $filelist 0] [lindex $filelist 1]"
447	if {[llength $filelist]>2} {
448	trace add variable current_file write readMerge
449
450	if {$tcl_platform(os)=="Darwin" \|\| [llength $filelist]<30} {
451	set fnlist {}
452	foreach {op fn} $filelist {lappend fnlist "$op $fn"}
453	tk_optionMenu .bb.files current_file {*}$fnlist
454	} else {
455	set useOptionMenu 0
456	::ttk::menubutton .bb.files -text $current_file
457	if {[tk windowingsystem] eq "win32"} {
458	::ttk::style theme use winnative
459	.bb.files configure -padding {20 1 10 2}
460	}
461	toplevel .wfiles
462	wm withdraw .wfiles
463	update idletasks
464	wm transient .wfiles .
465	wm overrideredirect .wfiles 1
466	set ht [expr {[llength $filelist]/2}]
467	if {$ht>$CFG(LB_HEIGHT)} {set ht $CFG(LB_HEIGHT)}
468	listbox .wfiles.lb -width 0 -height $ht -activestyle none \
469	-yscroll {.wfiles.sb set}
470	set mx 1
471	foreach {op fn} $filelist {
472	set n [string length $fn]
473	if {$n>$mx} {set mx $n}
474	.wfiles.lb insert end "$op $fn"
475	}
476	.bb.files config -width $mx
477	::ttk::scrollbar .wfiles.sb -command {.wfiles.lb yview}
478	grid .wfiles.lb .wfiles.sb -sticky ns
479	bind .bb.files <1> {
480	set x [winfo rootx %W]
481	set y [expr {[winfo rooty %W]+[winfo height %W]}]
482	wm geometry .wfiles +$x+$y
483	wm deiconify .wfiles
484	focus .wfiles.lb
485	}
486	bind .wfiles <FocusOut> {wm withdraw .wfiles}
487	bind .wfiles <Escape> {focus .}
488	foreach evt {1 Return} {
489	bind .wfiles.lb <$evt> {
490	set ii [%W curselection]
491	set ::current_file [%W get $ii]
492	.bb.files config -text $::current_file
493	focus .
494	break
495	}
496	}
497	bind .wfiles.lb <Motion> {
498	%W selection clear 0 end
499	%W selection set @%x,%y
500	}
501	}
502	}
503	}
504
505	label .bb.ctxtag -text "Context:"
506	set context_choices {3 6 12 25 50 100 All}
507	if {$ncontext<0} {set ncontext All}
508	trace add variable ncontext write readMerge
509	if {$tcl_platform(os)=="Darwin" \|\| $useOptionMenu} {
510	tk_optionMenu .bb.ctx ncontext {*}$context_choices
511	} else {
512	::ttk::menubutton .bb.ctx -text $ncontext
513	if {[tk windowingsystem] eq "win32"} {
514	::ttk::style theme use winnative
515	.bb.ctx configure -padding {20 1 10 2}
516	}
517	toplevel .wctx
518	wm withdraw .wctx
519	update idletasks
520	wm transient .wctx .
521	wm overrideredirect .wctx 1
522	listbox .wctx.lb -width 0 -height 7 -activestyle none
523	.wctx.lb insert end {*}$context_choices
524	pack .wctx.lb
525	bind .bb.ctx <1> {
526	set x [winfo rootx %W]
527	set y [expr {[winfo rooty %W]+[winfo height %W]}]
528	wm geometry .wctx +$x+$y
529	wm deiconify .wctx
530	focus .wctx.lb
531	}
532	bind .wctx <FocusOut> {wm withdraw .wctx}
533	bind .wctx <Escape> {focus .}
534	foreach evt {1 Return} {
535	bind .wctx.lb <$evt> {
536	set ::ncontext [lindex $::context_choices [%W curselection]]
537	.bb.ctx config -text $::ncontext
538	focus .
539	break
540	}
541	}
542	bind .wctx.lb <Motion> {
543	%W selection clear 0 end
544	%W selection set @%x,%y
545	}
546	}
547
548	foreach {side syncCol} {A .txtA B .txtB C .txtC D .txtD} {
549	set ln .ln$side
550	text $ln -width 6
551	$ln tag config - -justify right
552
553	set txt .txt$side
554	text $txt -width $CFG(WIDTH) -height $CFG(HEIGHT) -wrap none \
555	-xscroll ".sbx$side set"
556	catch {$txt config -tabstyle wordprocessor} ;# Required for Tk>=8.5
557	foreach tag {add rm chng} {
558	$txt tag config $tag -background $CFG([string toupper $tag]_BG)
559	$txt tag lower $tag
560	}
561	$txt tag config fn -background $CFG(FN_BG) -foreground $CFG(FN_FG) \
562	-justify center
563	$txt tag config err -foreground $CFG(ERR_FG)
564	}
565	text .mkr
566
567	set mxwidth [lindex [wm maxsize .] 0]
568	while {$CFG(WIDTH)>=40} {
569	set wanted [expr {([winfo reqwidth .lnA]+[winfo reqwidth .txtA])*4+30}]
570	if {$wanted<=$mxwidth} break
571	incr CFG(WIDTH) -10
572	.txtA config -width $CFG(WIDTH)
573	.txtB config -width $CFG(WIDTH)
574	.txtC config -width $CFG(WIDTH)
575	.txtD config -width $CFG(WIDTH)
576	}
577
578	foreach c [cols] {
579	set keyPrefix [string toupper [colType $c]]_COL_
580	if {[tk windowingsystem] eq "win32"} {$c config -font {courier 9}}
581	$c co

M src/merge3.c

+804 -168

		--- src/merge3.c
		+++ src/merge3.c
		@@ -75,75 +75,17 @@
75	75	if( aC1[2]!=aC2[2] ) return 0;
76	76	if( sameLines(pV1, pV2, aC1[2]) ) return 1;
77	77	return 0;
78	78	}
79	79
80		-/*
81		-** The aC[] array contains triples of integers. Within each triple, the
82		-** elements are:
83		-**
84		-** (0) The number of lines to copy
85		-** (1) The number of lines to delete
86		-** (2) The number of liens to insert
87		-**
88		-** Suppose we want to advance over sz lines of the original file. This routine
89		-** returns true if that advance would land us on a copy operation. It
90		-** returns false if the advance would end on a delete.
91		-*/
92		-static int ends_at_CPY(int *aC, int sz){
93		- while( sz>0 && (aC[0]>0 \|\| aC[1]>0 \|\| aC[2]>0) ){
94		- if( aC[0]>=sz ) return 1;
95		- sz -= aC[0];
96		- if( aC[1]>sz ) return 0;
97		- sz -= aC[1];
98		- aC += 3;
99		- }
100		- return 1;
101		-}
102		-
103		-/*
104		-** pSrc contains an edited file where aC[] describes the edit. Part of
105		-** pSrc has already been output. This routine outputs additional lines
106		-** of pSrc - lines that correspond to the next sz lines of the original
107		-** unedited file.
108		-**
109		-** Note that sz counts the number of lines of text in the original file.
110		-** But text is output from the edited file. So the number of lines transfer
111		-** to pOut might be different from sz. Fewer lines appear in pOut if there
112		-** are deletes. More lines appear if there are inserts.
113		-**
114		-** The aC[] array is updated and the new index into aC[] is returned.
115		-*/
116		-static int output_one_side(
117		- Blob pOut, / Write to this blob */
118		- Blob pSrc, / The edited file that is to be copied to pOut */
119		- int aC, / Array of integer triples describing the edit */
120		- int i, /* Index in aC[] of current location in pSrc */
121		- int sz, /* Number of lines in unedited source to output */
122		- int pLn / Line number counter */
123		-){
124		- while( sz>0 ){
125		- if( aC[i]==0 && aC[i+1]==0 && aC[i+2]==0 ) break;
126		- if( aC[i]>=sz ){
127		- blob_copy_lines(pOut, pSrc, sz); *pLn += sz;
128		- aC[i] -= sz;
129		- break;
130		- }
131		- blob_copy_lines(pOut, pSrc, aC[i]); *pLn += aC[i];
132		- blob_copy_lines(pOut, pSrc, aC[i+2]); *pLn += aC[i+2];
133		- sz -= aC[i] + aC[i+1];
134		- i += 3;
135		- }
136		- return i;
137		-}
138		-
139	80	/*
140	81	** Text of boundary markers for merge conflicts.
141	82	*/
142	83	static const char *const mergeMarker[] = {
143	84	/123456789 123456789 123456789 123456789 123456789 123456789 123456789/
144	85	"<<<<<<< BEGIN MERGE CONFLICT: local copy shown first <<<<<<<<<<<<",
	86	+ "####### SUGGESTED CONFLICT RESOLUTION follows ###################",
145	87	"\|\|\|\|\|\|\| COMMON ANCESTOR content follows \|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|",
146	88	"======= MERGED IN content follows ===============================",
147	89	">>>>>>> END MERGE CONFLICT >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>"
148	90	};
149	91
		@@ -186,10 +128,615 @@
186	128	ensure_line_end(pOut, useCrLf);
187	129	blob_append(pOut, mergeMarker[iMark], -1);
188	130	if( ln>0 ) blob_appendf(pOut, " (line %d)", ln);
189	131	ensure_line_end(pOut, useCrLf);
190	132	}
	133	+
	134	+#if INTERFACE
	135	+/*
	136	+** This is an abstract class for constructing a merge.
	137	+** Subclasses of this object format the merge output in different ways.
	138	+**
	139	+** To subclass, create an instance of the MergeBuilder object and fill
	140	+** in appropriate method implementations.
	141	+*/
	142	+struct MergeBuilder {
	143	+ void (xStart)(MergeBuilder);
	144	+ void (xSame)(MergeBuilder, unsigned int);
	145	+ void (xChngV1)(MergeBuilder, unsigned int, unsigned int);
	146	+ void (xChngV2)(MergeBuilder, unsigned int, unsigned int);
	147	+ void (xChngBoth)(MergeBuilder, unsigned int, unsigned int);
	148	+ void (xConflict)(MergeBuilder, unsigned int, unsigned int, unsigned int);
	149	+ void (xEnd)(MergeBuilder);
	150	+ void (xDestroy)(MergeBuilder);
	151	+ const char zPivot; / Label or name for the pivot */
	152	+ const char zV1; / Label or name for the V1 file */
	153	+ const char zV2; / Label or name for the V2 file */
	154	+ const char zOut; / Label or name for the output */
	155	+ Blob pPivot; / The common ancestor */
	156	+ Blob pV1; / First variant (local copy) */
	157	+ Blob pV2; / Second variant (merged in) */
	158	+ Blob pOut; / Write merge results here */
	159	+ int useCrLf; /* Use CRLF line endings */
	160	+ int nContext; /* Size of unchanged line boundaries */
	161	+ unsigned int mxPivot; /* Number of lines in the pivot */
	162	+ unsigned int mxV1; /* Number of lines in V1 */
	163	+ unsigned int mxV2; /* Number of lines in V2 */
	164	+ unsigned int lnPivot; /* Lines read from pivot */
	165	+ unsigned int lnV1; /* Lines read from v1 */
	166	+ unsigned int lnV2; /* Lines read from v2 */
	167	+ unsigned int lnOut; /* Lines written to out */
	168	+ unsigned int nConflict; /* Number of conflicts seen */
	169	+ u64 diffFlags; /* Flags for difference engine */
	170	+};
	171	+#endif /* INTERFACE */
	172	+
	173	+
	174	+/*********************** Generic MergeBuilder ****************************/
	175	+/* These are generic methods for MergeBuilder. They just output debugging
	176	+** information. But some of them are useful as base methods for other useful
	177	+** implementations of MergeBuilder.
	178	+*/
	179	+
	180	+/* xStart() and xEnd() are called to generate header and fotter information
	181	+** in the output. This is a no-op in the generic implementation.
	182	+*/
	183	+static void dbgStartEnd(MergeBuilder *p){ (void)p; }
	184	+
	185	+/* The next N lines of PIVOT are unchanged in both V1 and V2
	186	+*/
	187	+static void dbgSame(MergeBuilder *p, unsigned int N){
	188	+ blob_appendf(p->pOut,
	189	+ "COPY %u from BASELINE(%u..%u) or V1(%u..%u) or V2(%u..%u)\n",
	190	+ N, p->lnPivot+1, p->lnPivot+N, p->lnV1+1, p->lnV1+N,
	191	+ p->lnV2+1, p->lnV2+N);
	192	+ p->lnPivot += N;
	193	+ p->lnV1 += N;
	194	+ p->lnV2 += N;
	195	+}
	196	+
	197	+/* The next nPivot lines of the PIVOT are changed into nV1 lines by V1
	198	+*/
	199	+static void dbgChngV1(MergeBuilder *p, unsigned int nPivot, unsigned int nV1){
	200	+ blob_appendf(p->pOut, "COPY %u from V1(%u..%u)\n",
	201	+ nV1, p->lnV1+1, p->lnV1+nV1);
	202	+ p->lnPivot += nPivot;
	203	+ p->lnV2 += nPivot;
	204	+ p->lnV1 += nV1;
	205	+}
	206	+
	207	+/* The next nPivot lines of the PIVOT are changed into nV2 lines by V2
	208	+*/
	209	+static void dbgChngV2(MergeBuilder *p, unsigned int nPivot, unsigned int nV2){
	210	+ blob_appendf(p->pOut, "COPY %u lines FROM V2(%u..%u)\n",
	211	+ nV2, p->lnV2+1, p->lnV2+nV2);
	212	+ p->lnPivot += nPivot;
	213	+ p->lnV1 += nPivot;
	214	+ p->lnV2 += nV2;
	215	+}
	216	+
	217	+/* The next nPivot lines of the PIVOT are changed into nV lines from V1 and
	218	+** V2, which should be the same. In other words, the same change is found
	219	+** in both V1 and V2.
	220	+*/
	221	+static void dbgChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned int nV){
	222	+ blob_appendf(p->pOut, "COPY %u lines from V1(%u..%u) or V2(%u..%u)\n",
	223	+ nV, p->lnV1+1, p->lnV1+nV, p->lnV2+1, p->lnV2+nV);
	224	+ p->lnPivot += nPivot;
	225	+ p->lnV1 += nV;
	226	+ p->lnV2 += nV;
	227	+}
	228	+
	229	+/* V1 and V2 have different and overlapping changes. The next nPivot lines
	230	+** of the PIVOT are converted into nV1 lines of V1 and nV2 lines of V2.
	231	+*/
	232	+static void dbgConflict(
	233	+ MergeBuilder *p,
	234	+ unsigned int nPivot,
	235	+ unsigned int nV1,
	236	+ unsigned int nV2
	237	+){
	238	+ blob_appendf(p->pOut,
	239	+ "CONFLICT %u,%u,%u BASELINE(%u..%u) versus V1(%u..%u) versus V2(%u..%u)\n",
	240	+ nPivot, nV1, nV2,
	241	+ p->lnPivot+1, p->lnPivot+nPivot,
	242	+ p->lnV1+1, p->lnV1+nV1,
	243	+ p->lnV2+1, p->lnV2+nV2);
	244	+ p->lnV1 += nV1;
	245	+ p->lnPivot += nPivot;
	246	+ p->lnV2 += nV2;
	247	+}
	248	+
	249	+/* Generic destructor for the MergeBuilder object
	250	+*/
	251	+static void dbgDestroy(MergeBuilder *p){
	252	+ memset(p, 0, sizeof(*p));
	253	+}
	254	+
	255	+/* Generic initializer for a MergeBuilder object
	256	+*/
	257	+static void mergebuilder_init(MergeBuilder *p){
	258	+ memset(p, 0, sizeof(*p));
	259	+ p->xStart = dbgStartEnd;
	260	+ p->xSame = dbgSame;
	261	+ p->xChngV1 = dbgChngV1;
	262	+ p->xChngV2 = dbgChngV2;
	263	+ p->xChngBoth = dbgChngBoth;
	264	+ p->xConflict = dbgConflict;
	265	+ p->xEnd = dbgStartEnd;
	266	+ p->xDestroy = dbgDestroy;
	267	+}
	268	+
	269	+/*********************** MergeBuilderToken ******************************/
	270	+/* This version of MergeBuilder actually performs a merge on file that
	271	+** are broken up into tokens instead of lines, and puts the result in pOut.
	272	+*/
	273	+static void tokenSame(MergeBuilder *p, unsigned int N){
	274	+ blob_append(p->pOut, p->pPivot->aData+p->pPivot->iCursor, N);
	275	+ p->pPivot->iCursor += N;
	276	+ p->pV1->iCursor += N;
	277	+ p->pV2->iCursor += N;
	278	+}
	279	+static void tokenChngV1(MergeBuilder *p, unsigned int nPivot, unsigned nV1){
	280	+ blob_append(p->pOut, p->pV1->aData+p->pV1->iCursor, nV1);
	281	+ p->pPivot->iCursor += nPivot;
	282	+ p->pV1->iCursor += nV1;
	283	+ p->pV2->iCursor += nPivot;
	284	+}
	285	+static void tokenChngV2(MergeBuilder *p, unsigned int nPivot, unsigned nV2){
	286	+ blob_append(p->pOut, p->pV2->aData+p->pV2->iCursor, nV2);
	287	+ p->pPivot->iCursor += nPivot;
	288	+ p->pV1->iCursor += nPivot;
	289	+ p->pV2->iCursor += nV2;
	290	+}
	291	+static void tokenChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned nV){
	292	+ blob_append(p->pOut, p->pV2->aData+p->pV2->iCursor, nV);
	293	+ p->pPivot->iCursor += nPivot;
	294	+ p->pV1->iCursor += nV;
	295	+ p->pV2->iCursor += nV;
	296	+}
	297	+static void tokenConflict(
	298	+ MergeBuilder *p,
	299	+ unsigned int nPivot,
	300	+ unsigned int nV1,
	301	+ unsigned int nV2
	302	+){
	303	+ /* For a token-merge conflict, use the text from the merge-in */
	304	+ blob_append(p->pOut, p->pV2->aData+p->pV2->iCursor, nV2);
	305	+ p->pPivot->iCursor += nPivot;
	306	+ p->pV1->iCursor += nV1;
	307	+ p->pV2->iCursor += nV2;
	308	+}
	309	+static void mergebuilder_init_token(MergeBuilder *p){
	310	+ mergebuilder_init(p);
	311	+ p->xSame = tokenSame;
	312	+ p->xChngV1 = tokenChngV1;
	313	+ p->xChngV2 = tokenChngV2;
	314	+ p->xChngBoth = tokenChngBoth;
	315	+ p->xConflict = tokenConflict;
	316	+ p->diffFlags = DIFF_BY_TOKEN;
	317	+}
	318	+
	319	+/*
	320	+** Attempt to do a low-level merge on a conflict. The conflict is
	321	+** described by the first four parameters, which are the same as the
	322	+** arguments to the xConflict method of the MergeBuilder object.
	323	+** This routine attempts to resolve the conflict by looking at
	324	+** elements of the conflict region that are finer grain than complete
	325	+** lines of text.
	326	+**
	327	+** The result is written into Blob pOut. pOut is initialized by this
	328	+** routine.
	329	+*/
	330	+int merge_try_to_resolve_conflict(
	331	+ MergeBuilder pMB, / MergeBuilder that encounter conflict */
	332	+ unsigned int nPivot, /* Lines of conflict in the pivot */
	333	+ unsigned int nV1, /* Lines of conflict in V1 */
	334	+ unsigned int nV2, /* Lines of conflict in V2 */
	335	+ Blob pOut / Write resolution text here */
	336	+){
	337	+ int nConflict;
	338	+ MergeBuilder mb;
	339	+ Blob pv, v1, v2;
	340	+ mergebuilder_init_token(&mb);
	341	+ blob_extract_lines(pMB->pPivot, nPivot, &pv);
	342	+ blob_extract_lines(pMB->pV1, nV1, &v1);
	343	+ blob_extract_lines(pMB->pV2, nV2, &v2);
	344	+ blob_zero(pOut);
	345	+ blob_materialize(&pv);
	346	+ blob_materialize(&v1);
	347	+ blob_materialize(&v2);
	348	+ mb.pPivot = &pv;
	349	+ mb.pV1 = &v1;
	350	+ mb.pV2 = &v2;
	351	+ mb.pOut = pOut;
	352	+ nConflict = merge_three_blobs(&mb);
	353	+ blob_reset(&pv);
	354	+ blob_reset(&v1);
	355	+ blob_reset(&v2);
	356	+ /* mb has not allocated any resources, so we do not need to invoke
	357	+ ** the xDestroy method. */
	358	+ blob_add_final_newline(pOut);
	359	+ return nConflict;
	360	+}
	361	+
	362	+
	363	+/*********************** MergeBuilderText ********************************/
	364	+/* This version of MergeBuilder actually performs a merge on file and puts
	365	+** the result in pOut
	366	+*/
	367	+static void txtStart(MergeBuilder *p){
	368	+ /* If both pV1 and pV2 start with a UTF-8 byte-order-mark (BOM),
	369	+ ** keep it in the output. This should be secure enough not to cause
	370	+ ** unintended changes to the merged file and consistent with what
	371	+ ** users are using in their source files.
	372	+ */
	373	+ if( starts_with_utf8_bom(p->pV1, 0) && starts_with_utf8_bom(p->pV2, 0) ){
	374	+ blob_append(p->pOut, (char*)get_utf8_bom(0), -1);
	375	+ }
	376	+ if( contains_crlf(p->pV1) && contains_crlf(p->pV2) ){
	377	+ p->useCrLf = 1;
	378	+ }
	379	+}
	380	+static void txtSame(MergeBuilder *p, unsigned int N){
	381	+ blob_copy_lines(p->pOut, p->pPivot, N); p->lnPivot += N;
	382	+ blob_copy_lines(0, p->pV1, N); p->lnV1 += N;
	383	+ blob_copy_lines(0, p->pV2, N); p->lnV2 += N;
	384	+}
	385	+static void txtChngV1(MergeBuilder *p, unsigned int nPivot, unsigned int nV1){
	386	+ blob_copy_lines(0, p->pPivot, nPivot); p->lnPivot += nPivot;
	387	+ blob_copy_lines(0, p->pV2, nPivot); p->lnV2 += nPivot;
	388	+ blob_copy_lines(p->pOut, p->pV1, nV1); p->lnV1 += nV1;
	389	+}
	390	+static void txtChngV2(MergeBuilder *p, unsigned int nPivot, unsigned int nV2){
	391	+ blob_copy_lines(0, p->pPivot, nPivot); p->lnPivot += nPivot;
	392	+ blob_copy_lines(0, p->pV1, nPivot); p->lnV1 += nPivot;
	393	+ blob_copy_lines(p->pOut, p->pV2, nV2); p->lnV2 += nV2;
	394	+}
	395	+static void txtChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned int nV){
	396	+ blob_copy_lines(0, p->pPivot, nPivot); p->lnPivot += nPivot;
	397	+ blob_copy_lines(0, p->pV1, nV); p->lnV1 += nV;
	398	+ blob_copy_lines(p->pOut, p->pV2, nV); p->lnV2 += nV;
	399	+}
	400	+static void txtConflict(
	401	+ MergeBuilder *p,
	402	+ unsigned int nPivot,
	403	+ unsigned int nV1,
	404	+ unsigned int nV2
	405	+){
	406	+ int nRes; /* Lines in the computed conflict resolution */
	407	+ Blob res; /* Text of the conflict resolution */
	408	+
	409	+ merge_try_to_resolve_conflict(p, nPivot, nV1, nV2, &res);
	410	+ nRes = blob_linecount(&res);
	411	+
	412	+ append_merge_mark(p->pOut, 0, p->lnV1+1, p->useCrLf);
	413	+ blob_copy_lines(p->pOut, p->pV1, nV1); p->lnV1 += nV1;
	414	+
	415	+ if( nRes>0 ){
	416	+ append_merge_mark(p->pOut, 1, 0, p->useCrLf);
	417	+ blob_copy_lines(p->pOut, &res, nRes);
	418	+ }
	419	+ blob_reset(&res);
	420	+
	421	+ append_merge_mark(p->pOut, 2, p->lnPivot+1, p->useCrLf);
	422	+ blob_copy_lines(p->pOut, p->pPivot, nPivot); p->lnPivot += nPivot;
	423	+
	424	+ append_merge_mark(p->pOut, 3, p->lnV2+1, p->useCrLf);
	425	+ blob_copy_lines(p->pOut, p->pV2, nV2); p->lnV2 += nV2;
	426	+
	427	+ append_merge_mark(p->pOut, 4, -1, p->useCrLf);
	428	+}
	429	+static void mergebuilder_init_text(MergeBuilder *p){
	430	+ mergebuilder_init(p);
	431	+ p->xStart = txtStart;
	432	+ p->xSame = txtSame;
	433	+ p->xChngV1 = txtChngV1;
	434	+ p->xChngV2 = txtChngV2;
	435	+ p->xChngBoth = txtChngBoth;
	436	+ p->xConflict = txtConflict;
	437	+}
	438	+
	439	+/*********************** MergeBuilderTcl ********************************/
	440	+/* Generate merge output formatted for reading by a TCL script.
	441	+**
	442	+** The output consists of lines of text, each with 4 tokens. The tokens
	443	+** represent the content for one line from baseline, v1, v2, and output
	444	+** respectively. The first character of each token provides auxiliary
	445	+** information:
	446	+**
	447	+** . This line is omitted.
	448	+** N Name of the file.
	449	+** T Literal text follows that should have a \n terminator.
	450	+** R Literal text follows that needs a \r\n terminator.
	451	+** X Merge conflict.
	452	+** Z Literal text without a line terminator.
	453	+** S Skipped lines. Followed by number of lines to skip.
	454	+** 1 Text is a copy of token 1
	455	+** 2 Use data from data-token 2
	456	+** 3 Use data from data-token 3
	457	+*/
	458	+
	459	+/* Write text that goes into the interior of a double-quoted string in TCL */
	460	+static void tclWriteQuotedText(Blob pOut, const char zIn, int nIn){
	461	+ int j;
	462	+ for(j=0; j<nIn; j++){
	463	+ char c = zIn[j];
	464	+ if( c=='\\' ){
	465	+ blob_append(pOut, "\\\\", 2);
	466	+ }else if( c=='"' ){
	467	+ blob_append(pOut, "\\\"", 2);
	468	+ }else if( c<' ' \|\| c>0x7e ){
	469	+ char z[5];
	470	+ z[0] = '\\';
	471	+ z[1] = "01234567"[(c>>6)&0x3];
	472	+ z[2] = "01234567"[(c>>3)&0x7];
	473	+ z[3] = "01234567"[c&0x7];
	474	+ z[4] = 0;
	475	+ blob_append(pOut, z, 4);
	476	+ }else{
	477	+ blob_append_char(pOut, c);
	478	+ }
	479	+ }
	480	+}
	481	+
	482	+/* Copy one line of text from pIn and append to pOut, encoded as TCL */
	483	+static void tclLineOfText(Blob pOut, Blob pIn, char cType){
	484	+ int i, k;
	485	+ for(i=pIn->iCursor; i<pIn->nUsed && pIn->aData[i]!='\n'; i++){}
	486	+ if( i==pIn->nUsed ){
	487	+ k = i;
	488	+ }else if( i>pIn->iCursor && pIn->aData[i-1]=='\r' ){
	489	+ k = i-1;
	490	+ i++;
	491	+ }else{
	492	+ k = i;
	493	+ i++;
	494	+ }
	495	+ blob_append_char(pOut, '"');
	496	+ blob_append_char(pOut, cType);
	497	+ tclWriteQuotedText(pOut, pIn->aData+pIn->iCursor, k-pIn->iCursor);
	498	+ pIn->iCursor = i;
	499	+ blob_append_char(pOut, '"');
	500	+}
	501	+static void tclStart(MergeBuilder *p){
	502	+ Blob *pOut = p->pOut;
	503	+ blob_append(pOut, "\"N", 2);
	504	+ tclWriteQuotedText(pOut, p->zPivot, (int)strlen(p->zPivot));
	505	+ blob_append(pOut, "\" \"N", 4);
	506	+ tclWriteQuotedText(pOut, p->zV1, (int)strlen(p->zV1));
	507	+ blob_append(pOut, "\" \"N", 4);
	508	+ tclWriteQuotedText(pOut, p->zV2, (int)strlen(p->zV2));
	509	+ blob_append(pOut, "\" \"N", 4);
	510	+ if( p->zOut ){
	511	+ tclWriteQuotedText(pOut, p->zOut, (int)strlen(p->zOut));
	512	+ }else{
	513	+ blob_append(pOut, "(Merge Result)", -1);
	514	+ }
	515	+ blob_append(pOut, "\"\n", 2);
	516	+}
	517	+static void tclSame(MergeBuilder *p, unsigned int N){
	518	+ int i = 0;
	519	+ int nSkip;
	520	+
	521	+ if( p->lnPivot>=2 \|\| p->lnV1>2 \|\| p->lnV2>2 ){
	522	+ while( i<N && i<p->nContext ){
	523	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	524	+ blob_append(p->pOut, " 1 1 1\n", 7);
	525	+ i++;
	526	+ }
	527	+ nSkip = N - p->nContext*2;
	528	+ }else{
	529	+ nSkip = N - p->nContext;
	530	+ }
	531	+ if( nSkip>0 ){
	532	+ blob_appendf(p->pOut, "\"S%d %d %d %d\" . . .\n",
	533	+ nSkip, nSkip, nSkip, nSkip);
	534	+ blob_copy_lines(0, p->pPivot, nSkip);
	535	+ i += nSkip;
	536	+ }
	537	+
	538	+ p->lnPivot += N;
	539	+ p->lnV1 += N;
	540	+ p->lnV2 += N;
	541	+
	542	+ if( p->lnPivot<p->mxPivot \|\| p->lnV1<p->mxV1 \|\| p->lnV2<p->mxV2 ){
	543	+ while( i<N ){
	544	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	545	+ blob_append(p->pOut, " 1 1 1\n", 7);
	546	+ i++;
	547	+ }
	548	+ }
	549	+
	550	+ blob_copy_lines(0, p->pV1, N);
	551	+ blob_copy_lines(0, p->pV2, N);
	552	+}
	553	+static void tclChngV1(MergeBuilder *p, unsigned int nPivot, unsigned int nV1){
	554	+ int i;
	555	+ for(i=0; i<nPivot && i<nV1; i++){
	556	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	557	+ blob_append_char(p->pOut, ' ');
	558	+ tclLineOfText(p->pOut, p->pV1, 'T');
	559	+ blob_append(p->pOut, " 1 2\n", 5);
	560	+ }
	561	+ while( i<nPivot ){
	562	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	563	+ blob_append(p->pOut, " . 1 .\n", 7);
	564	+ i++;
	565	+ }
	566	+ while( i<nV1 ){
	567	+ blob_append(p->pOut, ". ", 2);
	568	+ tclLineOfText(p->pOut, p->pV1, 'T');
	569	+ blob_append(p->pOut, " . 2\n", 5);
	570	+ i++;
	571	+ }
	572	+ p->lnPivot += nPivot;
	573	+ p->lnV1 += nV1;
	574	+ p->lnV2 += nPivot;
	575	+ blob_copy_lines(0, p->pV2, nPivot);
	576	+}
	577	+static void tclChngV2(MergeBuilder *p, unsigned int nPivot, unsigned int nV2){
	578	+ int i;
	579	+ for(i=0; i<nPivot && i<nV2; i++){
	580	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	581	+ blob_append(p->pOut, " 1 ", 3);
	582	+ tclLineOfText(p->pOut, p->pV2, 'T');
	583	+ blob_append(p->pOut, " 3\n", 3);
	584	+ }
	585	+ while( i<nPivot ){
	586	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	587	+ blob_append(p->pOut, " 1 . .\n", 7);
	588	+ i++;
	589	+ }
	590	+ while( i<nV2 ){
	591	+ blob_append(p->pOut, ". . ", 4);
	592	+ tclLineOfText(p->pOut, p->pV2, 'T');
	593	+ blob_append(p->pOut, " 3\n", 3);
	594	+ i++;
	595	+ }
	596	+ p->lnPivot += nPivot;
	597	+ p->lnV1 += nPivot;
	598	+ p->lnV2 += nV2;
	599	+ blob_copy_lines(0, p->pV1, nPivot);
	600	+}
	601	+static void tclChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned int nV){
	602	+ int i;
	603	+ for(i=0; i<nPivot && i<nV; i++){
	604	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	605	+ blob_append_char(p->pOut, ' ');
	606	+ tclLineOfText(p->pOut, p->pV1, 'T');
	607	+ blob_append(p->pOut, " 2 2\n", 5);
	608	+ }
	609	+ while( i<nPivot ){
	610	+ tclLineOfText(p->pOut, p->pPivot, 'T');
	611	+ blob_append(p->pOut, " . . .\n", 7);
	612	+ i++;
	613	+ }
	614	+ while( i<nV ){
	615	+ blob_append(p->pOut, ". ", 2);
	616	+ tclLineOfText(p->pOut, p->pV1, 'T');
	617	+ blob_append(p->pOut, " 2 2\n", 5);
	618	+ i++;
	619	+ }
	620	+ p->lnPivot += nPivot;
	621	+ p->lnV1 += nV;
	622	+ p->lnV2 += nV;
	623	+ blob_copy_lines(0, p->pV2, nV);
	624	+}
	625	+static void tclConflict(
	626	+ MergeBuilder *p,
	627	+ unsigned int nPivot,
	628	+ unsigned int nV1,
	629	+ unsigned int nV2
	630	+){
	631	+ int mx = nPivot;
	632	+ int i;
	633	+ int nRes;
	634	+ Blob res;
	635	+
	636	+ merge_try_to_resolve_conflict(p, nPivot, nV1, nV2, &res);
	637	+ nRes = blob_linecount(&res);
	638	+ if( nV1>mx ) mx = nV1;
	639	+ if( nV2>mx ) mx = nV2;
	640	+ if( nRes>mx ) mx = nRes;
	641	+ if( nRes>0 ){
	642	+ blob_appendf(p->pOut, "\"S0 0 0 %d\" . . .\n", nV2+2);
	643	+ }
	644	+ for(i=0; i<mx; i++){
	645	+ if( i<nPivot ){
	646	+ tclLineOfText(p->pOut, p->pPivot, 'X');
	647	+ }else{
	648	+ blob_append_char(p->pOut, '.');
	649	+ }
	650	+ blob_append_char(p->pOut, ' ');
	651	+ if( i<nV1 ){
	652	+ tclLineOfText(p->pOut, p->pV1, 'X');
	653	+ }else{
	654	+ blob_append_char(p->pOut, '.');
	655	+ }
	656	+ blob_append_char(p->pOut, ' ');
	657	+ if( i<nV2 ){
	658	+ tclLineOfText(p->pOut, p->pV2, 'X');
	659	+ }else{
	660	+ blob_append_char(p->pOut, '.');
	661	+ }
	662	+ if( i<nRes ){
	663	+ blob_append_char(p->pOut, ' ');
	664	+ tclLineOfText(p->pOut, &res, 'X');
	665	+ blob_append_char(p->pOut, '\n');
	666	+ }else{
	667	+ blob_append(p->pOut, " .\n", 3);
	668	+ }
	669	+ if( i==mx-1 ){
	670	+ blob_appendf(p->pOut, "\"S0 0 0 %d\" . . .\n", nPivot+nV1+3);
	671	+ }
	672	+ }
	673	+ blob_reset(&res);
	674	+ p->lnPivot += nPivot;
	675	+ p->lnV1 += nV1;
	676	+ p->lnV2 += nV2;
	677	+}
	678	+void mergebuilder_init_tcl(MergeBuilder *p){
	679	+ mergebuilder_init(p);
	680	+ p->xStart = tclStart;
	681	+ p->xSame = tclSame;
	682	+ p->xChngV1 = tclChngV1;
	683	+ p->xChngV2 = tclChngV2;
	684	+ p->xChngBoth = tclChngBoth;
	685	+ p->xConflict = tclConflict;
	686	+}
	687	+/*****************************************************************************/
	688	+
	689	+/*
	690	+** The aC[] array contains triples of integers. Within each triple, the
	691	+** elements are:
	692	+**
	693	+** (0) The number of lines to copy
	694	+** (1) The number of lines to delete
	695	+** (2) The number of liens to insert
	696	+**
	697	+** Suppose we want to advance over sz lines of the original file. This routine
	698	+** returns true if that advance would land us on a copy operation. It
	699	+** returns false if the advance would end on a delete.
	700	+*/
	701	+static int ends_with_copy(int *aC, int sz){
	702	+ while( sz>0 && (aC[0]>0 \|\| aC[1]>0 \|\| aC[2]>0) ){
	703	+ if( aC[0]>=sz ) return 1;
	704	+ sz -= aC[0];
	705	+ if( aC[1]>sz ) return 0;
	706	+ sz -= aC[1];
	707	+ aC += 3;
	708	+ }
	709	+ return 1;
	710	+}
	711	+
	712	+/*
	713	+** aC[] is an "edit triple" for changes from A to B. Advance through
	714	+** this triple to determine the number of lines to bypass on B in order
	715	+** to match an advance of sz lines on A.
	716	+*/
	717	+static int skip_conflict(
	718	+ int aC, / Array of integer triples describing the edit */
	719	+ int i, /* Index in aC[] of current location */
	720	+ int sz, /* Lines of A that have been skipped */
	721	+ unsigned int pLn / OUT: Lines of B to skip to keep aligment with A */
	722	+){
	723	+ *pLn = 0;
	724	+ while( sz>0 ){
	725	+ if( aC[i]==0 && aC[i+1]==0 && aC[i+2]==0 ) break;
	726	+ if( aC[i]>=sz ){
	727	+ aC[i] -= sz;
	728	+ *pLn += sz;
	729	+ break;
	730	+ }
	731	+ *pLn += aC[i];
	732	+ *pLn += aC[i+2];
	733	+ sz -= aC[i] + aC[i+1];
	734	+ i += 3;
	735	+ }
	736	+ return i;
	737	+}
191	738
192	739	/*
193	740	** Do a three-way merge. Initialize pOut to contain the result.
194	741	**
195	742	** The merge is an edit against pV2. Both pV1 and pV2 have a
		@@ -199,156 +746,113 @@
199	746	** The return is 0 upon complete success. If any input file is binary,
200	747	** -1 is returned and pOut is unmodified. If there are merge
201	748	** conflicts, the merge proceeds as best as it can and the number
202	749	** of conflicts is returns
203	750	*/
204		-static int blob_merge(Blob pPivot, Blob pV1, Blob pV2, Blob pOut){
	751	+int merge_three_blobs(MergeBuilder *p){
205	752	int aC1; / Changes from pPivot to pV1 */
206	753	int aC2; / Changes from pPivot to pV2 */
207	754	int i1, i2; /* Index into aC1[] and aC2[] */
208	755	int nCpy, nDel, nIns; /* Number of lines to copy, delete, or insert */
209	756	int limit1, limit2; /* Sizes of aC1[] and aC2[] */
210	757	int nConflict = 0; /* Number of merge conflicts seen so far */
211		- int useCrLf = 0;
212		- int ln1, ln2, lnPivot; /* Line numbers for all files */
213	758	DiffConfig DCfg;
214	759
215		- blob_zero(pOut); /* Merge results stored in pOut */
216		-
217		- /* If both pV1 and pV2 start with a UTF-8 byte-order-mark (BOM),
218		- ** keep it in the output. This should be secure enough not to cause
219		- ** unintended changes to the merged file and consistent with what
220		- ** users are using in their source files.
221		- */
222		- if( starts_with_utf8_bom(pV1, 0) && starts_with_utf8_bom(pV2, 0) ){
223		- blob_append(pOut, (char*)get_utf8_bom(0), -1);
224		- }
225		-
226		- /* Check once to see if both pV1 and pV2 contains CR/LF endings.
227		- ** If true, CR/LF pair will be used later to append the
228		- ** boundary markers for merge conflicts.
229		- */
230		- if( contains_crlf(pV1) && contains_crlf(pV2) ){
231		- useCrLf = 1;
232		- }
233		-
234	760	/* Compute the edits that occur from pPivot => pV1 (into aC1)
235	761	** and pPivot => pV2 (into aC2). Each of the aC1 and aC2 arrays is
236	762	** an array of integer triples. Within each triple, the first integer
237	763	** is the number of lines of text to copy directly from the pivot,
238	764	** the second integer is the number of lines of text to omit from the
239	765	** pivot, and the third integer is the number of lines of text that are
240	766	** inserted. The edit array ends with a triple of 0,0,0.
241	767	*/
242	768	diff_config_init(&DCfg, 0);
243		- aC1 = text_diff(pPivot, pV1, 0, &DCfg);
244		- aC2 = text_diff(pPivot, pV2, 0, &DCfg);
	769	+ DCfg.diffFlags = p->diffFlags;
	770	+ aC1 = text_diff(p->pPivot, p->pV1, 0, &DCfg);
	771	+ aC2 = text_diff(p->pPivot, p->pV2, 0, &DCfg);
245	772	if( aC1==0 \|\| aC2==0 ){
246	773	free(aC1);
247	774	free(aC2);
248	775	return -1;
249	776	}
250	777
251		- blob_rewind(pV1); /* Rewind inputs: Needed to reconstruct output */
252		- blob_rewind(pV2);
253		- blob_rewind(pPivot);
	778	+ blob_rewind(p->pV1); /* Rewind inputs: Needed to reconstruct output */
	779	+ blob_rewind(p->pV2);
	780	+ blob_rewind(p->pPivot);
254	781
255	782	/* Determine the length of the aC1[] and aC2[] change vectors */
256		- for(i1=0; aC1[i1] \|\| aC1[i1+1] \|\| aC1[i1+2]; i1+=3){}
	783	+ p->mxPivot = 0;
	784	+ p->mxV1 = 0;
	785	+ for(i1=0; aC1[i1] \|\| aC1[i1+1] \|\| aC1[i1+2]; i1+=3){
	786	+ p->mxPivot += aC1[i1] + aC1[i1+1];
	787	+ p->mxV1 += aC1[i1] + aC1[i1+2];
	788	+ }
257	789	limit1 = i1;
258		- for(i2=0; aC2[i2] \|\| aC2[i2+1] \|\| aC2[i2+2]; i2+=3){}
	790	+ p->mxV2 = 0;
	791	+ for(i2=0; aC2[i2] \|\| aC2[i2+1] \|\| aC2[i2+2]; i2+=3){
	792	+ p->mxV2 += aC2[i2] + aC2[i2+2];
	793	+ }
259	794	limit2 = i2;
260	795
261		- DEBUG(
262		- for(i1=0; i1<limit1; i1+=3){
263		- printf("c1: %4d %4d %4d\n", aC1[i1], aC1[i1+1], aC1[i1+2]);
264		- }
265		- for(i2=0; i2<limit2; i2+=3){
266		- printf("c2: %4d %4d %4d\n", aC2[i2], aC2[i2+1], aC2[i2+2]);
267		- }
268		- )
	796	+ /* Output header text and do any other required initialization */
	797	+ p->xStart(p);
269	798
270	799	/* Loop over the two edit vectors and use them to compute merged text
271	800	** which is written into pOut. i1 and i2 are multiples of 3 which are
272	801	** indices into aC1[] and aC2[] to the edit triple currently being
273	802	** processed
274	803	*/
275	804	i1 = i2 = 0;
276		- ln1 = ln2 = lnPivot = 1;
277	805	while( i1<limit1 && i2<limit2 ){
278		- DEBUG( printf("%d: %2d %2d %2d %d: %2d %2d %2d\n",
279		- i1/3, aC1[i1], aC1[i1+1], aC1[i1+2],
280		- i2/3, aC2[i2], aC2[i2+1], aC2[i2+2]); )
281		-
282	806	if( aC1[i1]>0 && aC2[i2]>0 ){
283	807	/* Output text that is unchanged in both V1 and V2 */
284	808	nCpy = min(aC1[i1], aC2[i2]);
285		- DEBUG( printf("COPY %d\n", nCpy); )
286		- blob_copy_lines(pOut, pPivot, nCpy); lnPivot += nCpy;
287		- blob_copy_lines(0, pV1, nCpy); ln1 += nCpy;
288		- blob_copy_lines(0, pV2, nCpy); ln2 += nCpy;
	809	+ p->xSame(p, nCpy);
289	810	aC1[i1] -= nCpy;
290	811	aC2[i2] -= nCpy;
291	812	}else
292	813	if( aC1[i1] >= aC2[i2+1] && aC1[i1]>0 && aC2[i2+1]+aC2[i2+2]>0 ){
293	814	/* Output edits to V2 that occurs within unchanged regions of V1 */
294	815	nDel = aC2[i2+1];
295	816	nIns = aC2[i2+2];
296		- DEBUG( printf("EDIT -%d+%d left\n", nDel, nIns); )
297		- blob_copy_lines(0, pPivot, nDel); lnPivot += nDel;
298		- blob_copy_lines(0, pV1, nDel); ln1 += nDel;
299		- blob_copy_lines(pOut, pV2, nIns); ln2 += nIns;
	817	+ p->xChngV2(p, nDel, nIns);
300	818	aC1[i1] -= nDel;
301	819	i2 += 3;
302	820	}else
303	821	if( aC2[i2] >= aC1[i1+1] && aC2[i2]>0 && aC1[i1+1]+aC1[i1+2]>0 ){
304	822	/* Output edits to V1 that occur within unchanged regions of V2 */
305	823	nDel = aC1[i1+1];
306	824	nIns = aC1[i1+2];
307		- DEBUG( printf("EDIT -%d+%d right\n", nDel, nIns); )
308		- blob_copy_lines(0, pPivot, nDel); lnPivot += nDel;
309		- blob_copy_lines(0, pV2, nDel); ln2 += nDel;
310		- blob_copy_lines(pOut, pV1, nIns); ln1 += nIns;
	825	+ p->xChngV1(p, nDel, nIns);
311	826	aC2[i2] -= nDel;
312	827	i1 += 3;
313	828	}else
314		- if( sameEdit(&aC1[i1], &aC2[i2], pV1, pV2) ){
	829	+ if( sameEdit(&aC1[i1], &aC2[i2], p->pV1, p->pV2) ){
315	830	/* Output edits that are identical in both V1 and V2. */
316	831	assert( aC1[i1]==0 );
317	832	nDel = aC1[i1+1];
318	833	nIns = aC1[i1+2];
319		- DEBUG( printf("EDIT -%d+%d both\n", nDel, nIns); )
320		- blob_copy_lines(0, pPivot, nDel); lnPivot += nDel;
321		- blob_copy_lines(pOut, pV1, nIns); ln1 += nIns;
322		- blob_copy_lines(0, pV2, nIns); ln2 += nIns;
	834	+ p->xChngBoth(p, nDel, nIns);
323	835	i1 += 3;
324	836	i2 += 3;
325	837	}else
326	838	{
327	839	/* We have found a region where different edits to V1 and V2 overlap.
328	840	** This is a merge conflict. Find the size of the conflict, then
329	841	** output both possible edits separated by distinctive marks.
330	842	*/
331		- int sz = 1; /* Size of the conflict in lines */
	843	+ unsigned int sz = 1; /* Size of the conflict in the pivot, in lines */
	844	+ unsigned int nV1, nV2; /* Size of conflict in V1 and V2, in lines */
332	845	nConflict++;
333		- while( !ends_at_CPY(&aC1[i1], sz) \|\| !ends_at_CPY(&aC2[i2], sz) ){
	846	+ while( !ends_with_copy(&aC1[i1], sz) \|\| !ends_with_copy(&aC2[i2], sz) ){
334	847	sz++;
335	848	}
336		- DEBUG( printf("CONFLICT %d\n", sz); )
337		-
338		- append_merge_mark(pOut, 0, ln1, useCrLf);
339		- i1 = output_one_side(pOut, pV1, aC1, i1, sz, &ln1);
340		-
341		- append_merge_mark(pOut, 1, lnPivot, useCrLf);
342		- blob_copy_lines(pOut, pPivot, sz); lnPivot += sz;
343		-
344		- append_merge_mark(pOut, 2, ln2, useCrLf);
345		- i2 = output_one_side(pOut, pV2, aC2, i2, sz, &ln2);
346		-
347		- append_merge_mark(pOut, 3, -1, useCrLf);
348		- }
349		-
	849	+ i1 = skip_conflict(aC1, i1, sz, &nV1);
	850	+ i2 = skip_conflict(aC2, i2, sz, &nV2);
	851	+ p->xConflict(p, sz, nV1, nV2);
	852	+ }
	853	+
350	854	/* If we are finished with an edit triple, advance to the next
351	855	** triple.
352	856	*/
353	857	if( i1<limit1 && aC1[i1]==0 && aC1[i1+1]==0 && aC1[i1+2]==0 ) i1+=3;
354	858	if( i2<limit2 && aC2[i2]==0 && aC2[i2+1]==0 && aC2[i2+2]==0 ) i2+=3;
		@@ -356,20 +860,18 @@
356	860
357	861	/* When one of the two edit vectors reaches its end, there might still
358	862	** be an insert in the other edit vector. Output this remaining
359	863	** insert.
360	864	*/
361		- DEBUG( printf("%d: %2d %2d %2d %d: %2d %2d %2d\n",
362		- i1/3, aC1[i1], aC1[i1+1], aC1[i1+2],
363		- i2/3, aC2[i2], aC2[i2+1], aC2[i2+2]); )
364	865	if( i1<limit1 && aC1[i1+2]>0 ){
365		- DEBUG( printf("INSERT +%d left\n", aC1[i1+2]); )
366		- blob_copy_lines(pOut, pV1, aC1[i1+2]);
	866	+ p->xChngV1(p, 0, aC1[i1+2]);
367	867	}else if( i2<limit2 && aC2[i2+2]>0 ){
368		- DEBUG( printf("INSERT +%d right\n", aC2[i2+2]); )
369		- blob_copy_lines(pOut, pV2, aC2[i2+2]);
	868	+ p->xChngV2(p, 0, aC2[i2+2]);
370	869	}
	870	+
	871	+ /* Output footer text */
	872	+ p->xEnd(p);
371	873
372	874	free(aC1);
373	875	free(aC2);
374	876	return nConflict;
375	877	}
		@@ -384,11 +886,12 @@
384	886	const char *z = blob_buffer(p);
385	887	int n = blob_size(p) - len + 1;
386	888	assert( len==(int)strlen(mergeMarker[1]) );
387	889	assert( len==(int)strlen(mergeMarker[2]) );
388	890	assert( len==(int)strlen(mergeMarker[3]) );
389		- assert( count(mergeMarker)==4 );
	891	+ assert( len==(int)strlen(mergeMarker[4]) );
	892	+ assert( count(mergeMarker)==5 );
390	893	for(i=0; i<n; ){
391	894	for(j=0; j<4; j++){
392	895	if( (memcmp(&z[i], mergeMarker[j], len)==0) ){
393	896	return 1;
394	897	}
		@@ -408,18 +911,106 @@
408	911	blob_read_from_file(&file, zFullpath, ExtFILE);
409	912	rc = contains_merge_marker(&file);
410	913	blob_reset(&file);
411	914	return rc;
412	915	}
	916	+
	917	+/*
	918	+** Show merge output in a Tcl/Tk window, in response to the --tk option
	919	+** to the "merge" or "3-way-merge" command.
	920	+**
	921	+** If fossil has direct access to a Tcl interpreter (either loaded
	922	+** dynamically through stubs or linked in statically), we can use it
	923	+** directly. Otherwise:
	924	+** (1) Write the Tcl/Tk script used for rendering into a temp file.
	925	+** (2) Invoke "tclsh" on the temp file using fossil_system().
	926	+** (3) Delete the temp file.
	927	+*/
	928	+void merge_tk(const char *zSubCmd, int firstArg){
	929	+ int i;
	930	+ Blob script;
	931	+ const char *zTempFile = 0;
	932	+ char *zCmd;
	933	+ const char *zTclsh;
	934	+ const char *zCnt;
	935	+ int bDarkMode = find_option("dark",0,0)!=0;
	936	+ int nContext;
	937	+ zCnt = find_option("context", "c", 1);
	938	+ if( zCnt==0 ){
	939	+ nContext = 6;
	940	+ }else{
	941	+ nContext = atoi(zCnt);
	942	+ if( nContext<0 ) nContext = 0xfffffff;
	943	+ }
	944	+ blob_zero(&script);
	945	+ blob_appendf(&script, "set ncontext %d\n", nContext);
	946	+ blob_appendf(&script, "set fossilcmd {\| \"%/\" %s -tcl",
	947	+ g.nameOfExe, zSubCmd);
	948	+ find_option("tcl",0,0);
	949	+ find_option("debug",0,0);
	950	+ zTclsh = find_option("tclsh",0,1);
	951	+ if( zTclsh==0 ){
	952	+ zTclsh = db_get("tclsh",0);
	953	+ }
	954	+ /* The undocumented --script FILENAME option causes the Tk script to
	955	+ ** be written into the FILENAME instead of being run. This is used
	956	+ ** for testing and debugging. */
	957	+ zTempFile = find_option("script",0,1);
	958	+ verify_all_options();
	959	+
	960	+ if( (g.argc - firstArg)!=3 ){
	961	+ fossil_fatal("Requires 3 filename arguments");
	962	+ }
	963	+
	964	+ for(i=firstArg; i<g.argc; i++){
	965	+ const char *z = g.argv[i];
	966	+ if( sqlite3_strglob("}",z) ){
	967	+ blob_appendf(&script, " {%/}", z);
	968	+ }else{
	969	+ int j;
	970	+ blob_append(&script, " ", 1);
	971	+ for(j=0; z[j]; j++) blob_appendf(&script, "\\%03o", (unsigned char)z[j]);
	972	+ }
	973	+ }
	974	+ blob_appendf(&script, "}\nset darkmode %d\n", bDarkMode);
	975	+ blob_appendf(&script, "%s", builtin_file("merge.tcl", 0));
	976	+ if( zTempFile ){
	977	+ blob_write_to_file(&script, zTempFile);
	978	+ fossil_print("To see the merge, run: %s \"%s\"\n", zTclsh, zTempFile);
	979	+ }else{
	980	+#if defined(FOSSIL_ENABLE_TCL)
	981	+ Th_FossilInit(TH_INIT_DEFAULT);
	982	+ if( evaluateTclWithEvents(g.interp, &g.tcl, blob_str(&script),
	983	+ blob_size(&script), 1, 1, 0)==TCL_OK ){
	984	+ blob_reset(&script);
	985	+ return;
	986	+ }
	987	+ /*
	988	+ * If evaluation of the Tcl script fails, the reason may be that Tk
	989	+ * could not be found by the loaded Tcl, or that Tcl cannot be loaded
	990	+ * dynamically (e.g. x64 Tcl with x86 Fossil). Therefore, fallback
	991	+ * to using the external "tclsh", if available.
	992	+ */
	993	+#endif
	994	+ zTempFile = write_blob_to_temp_file(&script);
	995	+ zCmd = mprintf("%$ %$", zTclsh, zTempFile);
	996	+ fossil_system(zCmd);
	997	+ file_delete(zTempFile);
	998	+ fossil_free(zCmd);
	999	+ }
	1000	+ blob_reset(&script);
	1001	+}
	1002	+
413	1003
414	1004	/*
415	1005	** COMMAND: 3-way-merge*
416	1006	**
417		-** Usage: %fossil 3-way-merge BASELINE V1 V2 MERGED
	1007	+** Usage: %fossil 3-way-merge BASELINE V1 V2 [MERGED]
418	1008	**
419	1009	** Inputs are files BASELINE, V1, and V2. The file MERGED is generated
420		-** as output.
	1010	+** as output. If no MERGED file is specified, output is sent to
	1011	+** stdout.
421	1012	**
422	1013	** BASELINE is a common ancestor of two files V1 and V2 that have diverging
423	1014	** edits. The generated output file MERGED is the combination of all
424	1015	** changes in both V1 and V2.
425	1016	**
		@@ -436,38 +1027,75 @@
436	1027	** cp Xup.c Xbase.c
437	1028	** # Verify that everything still works
438	1029	** fossil commit
439	1030	**
440	1031	*/
441		-void delta_3waymerge_cmd(void){
442		- Blob pivot, v1, v2, merged;
	1032	+void merge_3way_cmd(void){
	1033	+ MergeBuilder s;
443	1034	int nConflict;
	1035	+ Blob pivot, v1, v2, out;
	1036	+ int noWarn = 0;
	1037	+ const char *zCnt;
	1038	+
	1039	+ if( find_option("tk", 0, 0)!=0 ){
	1040	+ merge_tk("3-way-merge", 2);
	1041	+ return;
	1042	+ }
	1043	+ mergebuilder_init_text(&s);
	1044	+ if( find_option("debug", 0, 0) ){
	1045	+ mergebuilder_init(&s);
	1046	+ }
	1047	+ if( find_option("tcl", 0, 0) ){
	1048	+ mergebuilder_init_tcl(&s);
	1049	+ noWarn = 1;
	1050	+ }
	1051	+ zCnt = find_option("context", "c", 1);
	1052	+ if( zCnt ){
	1053	+ s.nContext = atoi(zCnt);
	1054	+ if( s.nContext<0 ) s.nContext = 0xfffffff;
	1055	+ }else{
	1056	+ s.nContext = 6;
	1057	+ }
	1058	+ blob_zero(&pivot); s.pPivot = &pivot;
	1059	+ blob_zero(&v1); s.pV1 = &v1;
	1060	+ blob_zero(&v2); s.pV2 = &v2;
	1061	+ blob_zero(&out); s.pOut = &out;
444	1062
445	1063	/* We should be done with options.. */
446	1064	verify_all_options();
447	1065
448		- if( g.argc!=6 ){
449		- usage("PIVOT V1 V2 MERGED");
	1066	+ if( g.argc!=6 && g.argc!=5 ){
	1067	+ usage("[OPTIONS] PIVOT V1 V2 [MERGED]");
450	1068	}
451		- if( blob_read_from_file(&pivot, g.argv[2], ExtFILE)<0 ){
	1069	+ s.zPivot = file_tail(g.argv[2]);
	1070	+ s.zV1 = file_tail(g.argv[3]);
	1071	+ s.zV2 = file_tail(g.argv[4]);
	1072	+ if( blob_read_from_file(s.pPivot, g.argv[2], ExtFILE)<0 ){
452	1073	fossil_fatal("cannot read %s", g.argv[2]);
453	1074	}
454		- if( blob_read_from_file(&v1, g.argv[3], ExtFILE)<0 ){
	1075	+ if( blob_read_from_file(s.pV1, g.argv[3], ExtFILE)<0 ){
455	1076	fossil_fatal("cannot read %s", g.argv[3]);
456	1077	}
457		- if( blob_read_from_file(&v2, g.argv[4], ExtFILE)<0 ){
	1078	+ if( blob_read_from_file(s.pV2, g.argv[4], ExtFILE)<0 ){
458	1079	fossil_fatal("cannot read %s", g.argv[4]);
459	1080	}
460		- nConflict = blob_merge(&pivot, &v1, &v2, &merged);
461		- if( blob_write_to_file(&merged, g.argv[5])<(int)blob_size(&merged) ){
462		- fossil_fatal("cannot write %s", g.argv[4]);
	1081	+ nConflict = merge_three_blobs(&s);
	1082	+ if( g.argc==6 ){
	1083	+ s.zOut = file_tail(g.argv[5]);
	1084	+ blob_write_to_file(s.pOut, g.argv[5]);
	1085	+ }else{
	1086	+ s.zOut = "(Merge Result)";
	1087	+ blob_write_to_file(s.pOut, "-");
463	1088	}
	1089	+ s.xDestroy(&s);
464	1090	blob_reset(&pivot);
465	1091	blob_reset(&v1);
466	1092	blob_reset(&v2);
467		- blob_reset(&merged);
468		- if( nConflict>0 ) fossil_warning("WARNING: %d merge conflicts", nConflict);
	1093	+ blob_reset(&out);
	1094	+ if( nConflict>0 && !noWarn ){
	1095	+ fossil_warning("WARNING: %d merge conflicts", nConflict);
	1096	+ }
469	1097	}
470	1098
471	1099	/*
472	1100	** aSubst is an array of string pairs. The first element of each pair is
473	1101	** a string that begins with %. The second element is a replacement for that
		@@ -505,32 +1133,32 @@
505	1133
506	1134	#if INTERFACE
507	1135	/*
508	1136	** Flags to the 3-way merger
509	1137	*/
510		-#define MERGE_DRYRUN 0x0001
	1138	+#define MERGE_DRYRUN 0x0001
511	1139	/*
512	1140	** The MERGE_KEEP_FILES flag specifies that merge_3way() should retain
513	1141	** its temporary files on error. By default they are removed after the
514	1142	** merge, regardless of success or failure.
515	1143	*/
516		-#define MERGE_KEEP_FILES 0x0002
	1144	+#define MERGE_KEEP_FILES 0x0002
517	1145	#endif
518	1146
519	1147
520	1148	/*
521		-** This routine is a wrapper around blob_merge() with the following
	1149	+** This routine is a wrapper around merge_three_blobs() with the following
522	1150	** enhancements:
523	1151	**
524	1152	** (1) If the merge-command is defined, then use the external merging
525	1153	** program specified instead of the built-in blob-merge to do the
526	1154	** merging. Panic if the external merger fails.
527	1155	Not currently implemented **
528	1156	**
529	1157	** (2) If gmerge-command is defined and there are merge conflicts in
530		-** blob_merge() then invoke the external graphical merger to resolve
531		-** the conflicts.
	1158	+** merge_three_blobs() then invoke the external graphical merger
	1159	+** to resolve the conflicts.
532	1160	**
533	1161	** (3) If a merge conflict occurs and gmerge-command is not defined,
534	1162	** then write the pivot, original, and merge-in files to the
535	1163	** filesystem.
536	1164	*/
		@@ -539,30 +1167,37 @@
539	1167	const char zV1, / Name of file for version merging into (mine) */
540	1168	Blob pV2, / Version merging from (yours) */
541	1169	Blob pOut, / Output written here */
542	1170	unsigned mergeFlags /* Flags that control operation */
543	1171	){
544		- Blob v1; /* Content of zV1 */
545		- int rc; /* Return code of subroutines and this routine */
	1172	+ Blob v1; /* Content of zV1 */
	1173	+ int rc; /* Return code of subroutines and this routine */
546	1174	const char zGMerge; / Name of the gmerge command */
	1175	+ MergeBuilder s; /* The merge state */
547	1176
548		- blob_read_from_file(&v1, zV1, ExtFILE);
549		- rc = blob_merge(pPivot, &v1, pV2, pOut);
	1177	+ mergebuilder_init_text(&s);
	1178	+ s.pPivot = pPivot;
	1179	+ s.pV1 = &v1;
	1180	+ s.pV2 = pV2;
	1181	+ blob_zero(pOut);
	1182	+ s.pOut = pOut;
	1183	+ blob_read_from_file(s.pV1, zV1, ExtFILE);
	1184	+ rc = merge_three_blobs(&s);
550	1185	zGMerge = rc<=0 ? 0 : db_get("gmerge-command", 0);
551	1186	if( (mergeFlags & MERGE_DRYRUN)==0
552	1187	&& ((zGMerge!=0 && zGMerge[0]!=0)
553	1188	\|\| (rc!=0 && (mergeFlags & MERGE_KEEP_FILES)!=0)) ){
554	1189	char zPivot; / Name of the pivot file */
555	1190	char zOrig; / Name of the original content file */
556	1191	char zOther; / Name of the merge file */
557	1192
558	1193	zPivot = file_newname(zV1, "baseline", 1);
559		- blob_write_to_file(pPivot, zPivot);
	1194	+ blob_write_to_file(s.pPivot, zPivot);
560	1195	zOrig = file_newname(zV1, "original", 1);
561		- blob_write_to_file(&v1, zOrig);
	1196	+ blob_write_to_file(s.pV1, zOrig);
562	1197	zOther = file_newname(zV1, "merge", 1);
563		- blob_write_to_file(pV2, zOther);
	1198	+ blob_write_to_file(s.pV2, zOther);
564	1199	if( rc>0 ){
565	1200	if( zGMerge && zGMerge[0] ){
566	1201	char zOut; / Temporary output file */
567	1202	char zCmd; / Command to invoke */
568	1203	const char azSubst[8]; / Strings to be substituted */
		@@ -590,7 +1225,8 @@
590	1225	fossil_free(zPivot);
591	1226	fossil_free(zOrig);
592	1227	fossil_free(zOther);
593	1228	}
594	1229	blob_reset(&v1);
	1230	+ s.xDestroy(&s);
595	1231	return rc;
596	1232	}
597	1233

	--- src/merge3.c
	+++ src/merge3.c
	@@ -75,75 +75,17 @@
75	if( aC1[2]!=aC2[2] ) return 0;
76	if( sameLines(pV1, pV2, aC1[2]) ) return 1;
77	return 0;
78	}
79
80	/*
81	** The aC[] array contains triples of integers. Within each triple, the
82	** elements are:
83	**
84	** (0) The number of lines to copy
85	** (1) The number of lines to delete
86	** (2) The number of liens to insert
87	**
88	** Suppose we want to advance over sz lines of the original file. This routine
89	** returns true if that advance would land us on a copy operation. It
90	** returns false if the advance would end on a delete.
91	*/
92	static int ends_at_CPY(int *aC, int sz){
93	while( sz>0 && (aC[0]>0 \|\| aC[1]>0 \|\| aC[2]>0) ){
94	if( aC[0]>=sz ) return 1;
95	sz -= aC[0];
96	if( aC[1]>sz ) return 0;
97	sz -= aC[1];
98	aC += 3;
99	}
100	return 1;
101	}
102
103	/*
104	** pSrc contains an edited file where aC[] describes the edit. Part of
105	** pSrc has already been output. This routine outputs additional lines
106	** of pSrc - lines that correspond to the next sz lines of the original
107	** unedited file.
108	**
109	** Note that sz counts the number of lines of text in the original file.
110	** But text is output from the edited file. So the number of lines transfer
111	** to pOut might be different from sz. Fewer lines appear in pOut if there
112	** are deletes. More lines appear if there are inserts.
113	**
114	** The aC[] array is updated and the new index into aC[] is returned.
115	*/
116	static int output_one_side(
117	Blob pOut, / Write to this blob */
118	Blob pSrc, / The edited file that is to be copied to pOut */
119	int aC, / Array of integer triples describing the edit */
120	int i, /* Index in aC[] of current location in pSrc */
121	int sz, /* Number of lines in unedited source to output */
122	int pLn / Line number counter */
123	){
124	while( sz>0 ){
125	if( aC[i]==0 && aC[i+1]==0 && aC[i+2]==0 ) break;
126	if( aC[i]>=sz ){
127	blob_copy_lines(pOut, pSrc, sz); *pLn += sz;
128	aC[i] -= sz;
129	break;
130	}
131	blob_copy_lines(pOut, pSrc, aC[i]); *pLn += aC[i];
132	blob_copy_lines(pOut, pSrc, aC[i+2]); *pLn += aC[i+2];
133	sz -= aC[i] + aC[i+1];
134	i += 3;
135	}
136	return i;
137	}
138
139	/*
140	** Text of boundary markers for merge conflicts.
141	*/
142	static const char *const mergeMarker[] = {
143	/123456789 123456789 123456789 123456789 123456789 123456789 123456789/
144	"<<<<<<< BEGIN MERGE CONFLICT: local copy shown first <<<<<<<<<<<<",

145	"\|\|\|\|\|\|\| COMMON ANCESTOR content follows \|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|",
146	"======= MERGED IN content follows ===============================",
147	">>>>>>> END MERGE CONFLICT >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>"
148	};
149
	@@ -186,10 +128,615 @@
186	ensure_line_end(pOut, useCrLf);
187	blob_append(pOut, mergeMarker[iMark], -1);
188	if( ln>0 ) blob_appendf(pOut, " (line %d)", ln);
189	ensure_line_end(pOut, useCrLf);
190	}





























































































































































































































































































































































































































































































































































































































191
192	/*
193	** Do a three-way merge. Initialize pOut to contain the result.
194	**
195	** The merge is an edit against pV2. Both pV1 and pV2 have a
	@@ -199,156 +746,113 @@
199	** The return is 0 upon complete success. If any input file is binary,
200	** -1 is returned and pOut is unmodified. If there are merge
201	** conflicts, the merge proceeds as best as it can and the number
202	** of conflicts is returns
203	*/
204	static int blob_merge(Blob pPivot, Blob pV1, Blob pV2, Blob pOut){
205	int aC1; / Changes from pPivot to pV1 */
206	int aC2; / Changes from pPivot to pV2 */
207	int i1, i2; /* Index into aC1[] and aC2[] */
208	int nCpy, nDel, nIns; /* Number of lines to copy, delete, or insert */
209	int limit1, limit2; /* Sizes of aC1[] and aC2[] */
210	int nConflict = 0; /* Number of merge conflicts seen so far */
211	int useCrLf = 0;
212	int ln1, ln2, lnPivot; /* Line numbers for all files */
213	DiffConfig DCfg;
214
215	blob_zero(pOut); /* Merge results stored in pOut */
216
217	/* If both pV1 and pV2 start with a UTF-8 byte-order-mark (BOM),
218	** keep it in the output. This should be secure enough not to cause
219	** unintended changes to the merged file and consistent with what
220	** users are using in their source files.
221	*/
222	if( starts_with_utf8_bom(pV1, 0) && starts_with_utf8_bom(pV2, 0) ){
223	blob_append(pOut, (char*)get_utf8_bom(0), -1);
224	}
225
226	/* Check once to see if both pV1 and pV2 contains CR/LF endings.
227	** If true, CR/LF pair will be used later to append the
228	** boundary markers for merge conflicts.
229	*/
230	if( contains_crlf(pV1) && contains_crlf(pV2) ){
231	useCrLf = 1;
232	}
233
234	/* Compute the edits that occur from pPivot => pV1 (into aC1)
235	** and pPivot => pV2 (into aC2). Each of the aC1 and aC2 arrays is
236	** an array of integer triples. Within each triple, the first integer
237	** is the number of lines of text to copy directly from the pivot,
238	** the second integer is the number of lines of text to omit from the
239	** pivot, and the third integer is the number of lines of text that are
240	** inserted. The edit array ends with a triple of 0,0,0.
241	*/
242	diff_config_init(&DCfg, 0);
243	aC1 = text_diff(pPivot, pV1, 0, &DCfg);
244	aC2 = text_diff(pPivot, pV2, 0, &DCfg);

245	if( aC1==0 \|\| aC2==0 ){
246	free(aC1);
247	free(aC2);
248	return -1;
249	}
250
251	blob_rewind(pV1); /* Rewind inputs: Needed to reconstruct output */
252	blob_rewind(pV2);
253	blob_rewind(pPivot);
254
255	/* Determine the length of the aC1[] and aC2[] change vectors */
256	for(i1=0; aC1[i1] \|\| aC1[i1+1] \|\| aC1[i1+2]; i1+=3){}





257	limit1 = i1;
258	for(i2=0; aC2[i2] \|\| aC2[i2+1] \|\| aC2[i2+2]; i2+=3){}



259	limit2 = i2;
260
261	DEBUG(
262	for(i1=0; i1<limit1; i1+=3){
263	printf("c1: %4d %4d %4d\n", aC1[i1], aC1[i1+1], aC1[i1+2]);
264	}
265	for(i2=0; i2<limit2; i2+=3){
266	printf("c2: %4d %4d %4d\n", aC2[i2], aC2[i2+1], aC2[i2+2]);
267	}
268	)
269
270	/* Loop over the two edit vectors and use them to compute merged text
271	** which is written into pOut. i1 and i2 are multiples of 3 which are
272	** indices into aC1[] and aC2[] to the edit triple currently being
273	** processed
274	*/
275	i1 = i2 = 0;
276	ln1 = ln2 = lnPivot = 1;
277	while( i1<limit1 && i2<limit2 ){
278	DEBUG( printf("%d: %2d %2d %2d %d: %2d %2d %2d\n",
279	i1/3, aC1[i1], aC1[i1+1], aC1[i1+2],
280	i2/3, aC2[i2], aC2[i2+1], aC2[i2+2]); )
281
282	if( aC1[i1]>0 && aC2[i2]>0 ){
283	/* Output text that is unchanged in both V1 and V2 */
284	nCpy = min(aC1[i1], aC2[i2]);
285	DEBUG( printf("COPY %d\n", nCpy); )
286	blob_copy_lines(pOut, pPivot, nCpy); lnPivot += nCpy;
287	blob_copy_lines(0, pV1, nCpy); ln1 += nCpy;
288	blob_copy_lines(0, pV2, nCpy); ln2 += nCpy;
289	aC1[i1] -= nCpy;
290	aC2[i2] -= nCpy;
291	}else
292	if( aC1[i1] >= aC2[i2+1] && aC1[i1]>0 && aC2[i2+1]+aC2[i2+2]>0 ){
293	/* Output edits to V2 that occurs within unchanged regions of V1 */
294	nDel = aC2[i2+1];
295	nIns = aC2[i2+2];
296	DEBUG( printf("EDIT -%d+%d left\n", nDel, nIns); )
297	blob_copy_lines(0, pPivot, nDel); lnPivot += nDel;
298	blob_copy_lines(0, pV1, nDel); ln1 += nDel;
299	blob_copy_lines(pOut, pV2, nIns); ln2 += nIns;
300	aC1[i1] -= nDel;
301	i2 += 3;
302	}else
303	if( aC2[i2] >= aC1[i1+1] && aC2[i2]>0 && aC1[i1+1]+aC1[i1+2]>0 ){
304	/* Output edits to V1 that occur within unchanged regions of V2 */
305	nDel = aC1[i1+1];
306	nIns = aC1[i1+2];
307	DEBUG( printf("EDIT -%d+%d right\n", nDel, nIns); )
308	blob_copy_lines(0, pPivot, nDel); lnPivot += nDel;
309	blob_copy_lines(0, pV2, nDel); ln2 += nDel;
310	blob_copy_lines(pOut, pV1, nIns); ln1 += nIns;
311	aC2[i2] -= nDel;
312	i1 += 3;
313	}else
314	if( sameEdit(&aC1[i1], &aC2[i2], pV1, pV2) ){
315	/* Output edits that are identical in both V1 and V2. */
316	assert( aC1[i1]==0 );
317	nDel = aC1[i1+1];
318	nIns = aC1[i1+2];
319	DEBUG( printf("EDIT -%d+%d both\n", nDel, nIns); )
320	blob_copy_lines(0, pPivot, nDel); lnPivot += nDel;
321	blob_copy_lines(pOut, pV1, nIns); ln1 += nIns;
322	blob_copy_lines(0, pV2, nIns); ln2 += nIns;
323	i1 += 3;
324	i2 += 3;
325	}else
326	{
327	/* We have found a region where different edits to V1 and V2 overlap.
328	** This is a merge conflict. Find the size of the conflict, then
329	** output both possible edits separated by distinctive marks.
330	*/
331	int sz = 1; /* Size of the conflict in lines */

332	nConflict++;
333	while( !ends_at_CPY(&aC1[i1], sz) \|\| !ends_at_CPY(&aC2[i2], sz) ){
334	sz++;
335	}
336	DEBUG( printf("CONFLICT %d\n", sz); )
337
338	append_merge_mark(pOut, 0, ln1, useCrLf);
339	i1 = output_one_side(pOut, pV1, aC1, i1, sz, &ln1);
340
341	append_merge_mark(pOut, 1, lnPivot, useCrLf);
342	blob_copy_lines(pOut, pPivot, sz); lnPivot += sz;
343
344	append_merge_mark(pOut, 2, ln2, useCrLf);
345	i2 = output_one_side(pOut, pV2, aC2, i2, sz, &ln2);
346
347	append_merge_mark(pOut, 3, -1, useCrLf);
348	}
349
350	/* If we are finished with an edit triple, advance to the next
351	** triple.
352	*/
353	if( i1<limit1 && aC1[i1]==0 && aC1[i1+1]==0 && aC1[i1+2]==0 ) i1+=3;
354	if( i2<limit2 && aC2[i2]==0 && aC2[i2+1]==0 && aC2[i2+2]==0 ) i2+=3;
	@@ -356,20 +860,18 @@
356
357	/* When one of the two edit vectors reaches its end, there might still
358	** be an insert in the other edit vector. Output this remaining
359	** insert.
360	*/
361	DEBUG( printf("%d: %2d %2d %2d %d: %2d %2d %2d\n",
362	i1/3, aC1[i1], aC1[i1+1], aC1[i1+2],
363	i2/3, aC2[i2], aC2[i2+1], aC2[i2+2]); )
364	if( i1<limit1 && aC1[i1+2]>0 ){
365	DEBUG( printf("INSERT +%d left\n", aC1[i1+2]); )
366	blob_copy_lines(pOut, pV1, aC1[i1+2]);
367	}else if( i2<limit2 && aC2[i2+2]>0 ){
368	DEBUG( printf("INSERT +%d right\n", aC2[i2+2]); )
369	blob_copy_lines(pOut, pV2, aC2[i2+2]);
370	}



371
372	free(aC1);
373	free(aC2);
374	return nConflict;
375	}
	@@ -384,11 +886,12 @@
384	const char *z = blob_buffer(p);
385	int n = blob_size(p) - len + 1;
386	assert( len==(int)strlen(mergeMarker[1]) );
387	assert( len==(int)strlen(mergeMarker[2]) );
388	assert( len==(int)strlen(mergeMarker[3]) );
389	assert( count(mergeMarker)==4 );

390	for(i=0; i<n; ){
391	for(j=0; j<4; j++){
392	if( (memcmp(&z[i], mergeMarker[j], len)==0) ){
393	return 1;
394	}
	@@ -408,18 +911,106 @@
408	blob_read_from_file(&file, zFullpath, ExtFILE);
409	rc = contains_merge_marker(&file);
410	blob_reset(&file);
411	return rc;
412	}























































































413
414	/*
415	** COMMAND: 3-way-merge*
416	**
417	** Usage: %fossil 3-way-merge BASELINE V1 V2 MERGED
418	**
419	** Inputs are files BASELINE, V1, and V2. The file MERGED is generated
420	** as output.

421	**
422	** BASELINE is a common ancestor of two files V1 and V2 that have diverging
423	** edits. The generated output file MERGED is the combination of all
424	** changes in both V1 and V2.
425	**
	@@ -436,38 +1027,75 @@
436	** cp Xup.c Xbase.c
437	** # Verify that everything still works
438	** fossil commit
439	**
440	*/
441	void delta_3waymerge_cmd(void){
442	Blob pivot, v1, v2, merged;
443	int nConflict;



























444
445	/* We should be done with options.. */
446	verify_all_options();
447
448	if( g.argc!=6 ){
449	usage("PIVOT V1 V2 MERGED");
450	}
451	if( blob_read_from_file(&pivot, g.argv[2], ExtFILE)<0 ){



452	fossil_fatal("cannot read %s", g.argv[2]);
453	}
454	if( blob_read_from_file(&v1, g.argv[3], ExtFILE)<0 ){
455	fossil_fatal("cannot read %s", g.argv[3]);
456	}
457	if( blob_read_from_file(&v2, g.argv[4], ExtFILE)<0 ){
458	fossil_fatal("cannot read %s", g.argv[4]);
459	}
460	nConflict = blob_merge(&pivot, &v1, &v2, &merged);
461	if( blob_write_to_file(&merged, g.argv[5])<(int)blob_size(&merged) ){
462	fossil_fatal("cannot write %s", g.argv[4]);




463	}

464	blob_reset(&pivot);
465	blob_reset(&v1);
466	blob_reset(&v2);
467	blob_reset(&merged);
468	if( nConflict>0 ) fossil_warning("WARNING: %d merge conflicts", nConflict);


469	}
470
471	/*
472	** aSubst is an array of string pairs. The first element of each pair is
473	** a string that begins with %. The second element is a replacement for that
	@@ -505,32 +1133,32 @@
505
506	#if INTERFACE
507	/*
508	** Flags to the 3-way merger
509	*/
510	#define MERGE_DRYRUN 0x0001
511	/*
512	** The MERGE_KEEP_FILES flag specifies that merge_3way() should retain
513	** its temporary files on error. By default they are removed after the
514	** merge, regardless of success or failure.
515	*/
516	#define MERGE_KEEP_FILES 0x0002
517	#endif
518
519
520	/*
521	** This routine is a wrapper around blob_merge() with the following
522	** enhancements:
523	**
524	** (1) If the merge-command is defined, then use the external merging
525	** program specified instead of the built-in blob-merge to do the
526	** merging. Panic if the external merger fails.
527	Not currently implemented **
528	**
529	** (2) If gmerge-command is defined and there are merge conflicts in
530	** blob_merge() then invoke the external graphical merger to resolve
531	** the conflicts.
532	**
533	** (3) If a merge conflict occurs and gmerge-command is not defined,
534	** then write the pivot, original, and merge-in files to the
535	** filesystem.
536	*/
	@@ -539,30 +1167,37 @@
539	const char zV1, / Name of file for version merging into (mine) */
540	Blob pV2, / Version merging from (yours) */
541	Blob pOut, / Output written here */
542	unsigned mergeFlags /* Flags that control operation */
543	){
544	Blob v1; /* Content of zV1 */
545	int rc; /* Return code of subroutines and this routine */
546	const char zGMerge; / Name of the gmerge command */

547
548	blob_read_from_file(&v1, zV1, ExtFILE);
549	rc = blob_merge(pPivot, &v1, pV2, pOut);






550	zGMerge = rc<=0 ? 0 : db_get("gmerge-command", 0);
551	if( (mergeFlags & MERGE_DRYRUN)==0
552	&& ((zGMerge!=0 && zGMerge[0]!=0)
553	\|\| (rc!=0 && (mergeFlags & MERGE_KEEP_FILES)!=0)) ){
554	char zPivot; / Name of the pivot file */
555	char zOrig; / Name of the original content file */
556	char zOther; / Name of the merge file */
557
558	zPivot = file_newname(zV1, "baseline", 1);
559	blob_write_to_file(pPivot, zPivot);
560	zOrig = file_newname(zV1, "original", 1);
561	blob_write_to_file(&v1, zOrig);
562	zOther = file_newname(zV1, "merge", 1);
563	blob_write_to_file(pV2, zOther);
564	if( rc>0 ){
565	if( zGMerge && zGMerge[0] ){
566	char zOut; / Temporary output file */
567	char zCmd; / Command to invoke */
568	const char azSubst[8]; / Strings to be substituted */
	@@ -590,7 +1225,8 @@
590	fossil_free(zPivot);
591	fossil_free(zOrig);
592	fossil_free(zOther);
593	}
594	blob_reset(&v1);

595	return rc;
596	}
597

	--- src/merge3.c
	+++ src/merge3.c
	@@ -75,75 +75,17 @@
75	if( aC1[2]!=aC2[2] ) return 0;
76	if( sameLines(pV1, pV2, aC1[2]) ) return 1;
77	return 0;
78	}
79



























































80	/*
81	** Text of boundary markers for merge conflicts.
82	*/
83	static const char *const mergeMarker[] = {
84	/123456789 123456789 123456789 123456789 123456789 123456789 123456789/
85	"<<<<<<< BEGIN MERGE CONFLICT: local copy shown first <<<<<<<<<<<<",
86	"####### SUGGESTED CONFLICT RESOLUTION follows ###################",
87	"\|\|\|\|\|\|\| COMMON ANCESTOR content follows \|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|\|",
88	"======= MERGED IN content follows ===============================",
89	">>>>>>> END MERGE CONFLICT >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>"
90	};
91
	@@ -186,10 +128,615 @@
128	ensure_line_end(pOut, useCrLf);
129	blob_append(pOut, mergeMarker[iMark], -1);
130	if( ln>0 ) blob_appendf(pOut, " (line %d)", ln);
131	ensure_line_end(pOut, useCrLf);
132	}
133
134	#if INTERFACE
135	/*
136	** This is an abstract class for constructing a merge.
137	** Subclasses of this object format the merge output in different ways.
138	**
139	** To subclass, create an instance of the MergeBuilder object and fill
140	** in appropriate method implementations.
141	*/
142	struct MergeBuilder {
143	void (xStart)(MergeBuilder);
144	void (xSame)(MergeBuilder, unsigned int);
145	void (xChngV1)(MergeBuilder, unsigned int, unsigned int);
146	void (xChngV2)(MergeBuilder, unsigned int, unsigned int);
147	void (xChngBoth)(MergeBuilder, unsigned int, unsigned int);
148	void (xConflict)(MergeBuilder, unsigned int, unsigned int, unsigned int);
149	void (xEnd)(MergeBuilder);
150	void (xDestroy)(MergeBuilder);
151	const char zPivot; / Label or name for the pivot */
152	const char zV1; / Label or name for the V1 file */
153	const char zV2; / Label or name for the V2 file */
154	const char zOut; / Label or name for the output */
155	Blob pPivot; / The common ancestor */
156	Blob pV1; / First variant (local copy) */
157	Blob pV2; / Second variant (merged in) */
158	Blob pOut; / Write merge results here */
159	int useCrLf; /* Use CRLF line endings */
160	int nContext; /* Size of unchanged line boundaries */
161	unsigned int mxPivot; /* Number of lines in the pivot */
162	unsigned int mxV1; /* Number of lines in V1 */
163	unsigned int mxV2; /* Number of lines in V2 */
164	unsigned int lnPivot; /* Lines read from pivot */
165	unsigned int lnV1; /* Lines read from v1 */
166	unsigned int lnV2; /* Lines read from v2 */
167	unsigned int lnOut; /* Lines written to out */
168	unsigned int nConflict; /* Number of conflicts seen */
169	u64 diffFlags; /* Flags for difference engine */
170	};
171	#endif /* INTERFACE */
172
173
174	/*********************** Generic MergeBuilder ****************************/
175	/* These are generic methods for MergeBuilder. They just output debugging
176	** information. But some of them are useful as base methods for other useful
177	** implementations of MergeBuilder.
178	*/
179
180	/* xStart() and xEnd() are called to generate header and fotter information
181	** in the output. This is a no-op in the generic implementation.
182	*/
183	static void dbgStartEnd(MergeBuilder *p){ (void)p; }
184
185	/* The next N lines of PIVOT are unchanged in both V1 and V2
186	*/
187	static void dbgSame(MergeBuilder *p, unsigned int N){
188	blob_appendf(p->pOut,
189	"COPY %u from BASELINE(%u..%u) or V1(%u..%u) or V2(%u..%u)\n",
190	N, p->lnPivot+1, p->lnPivot+N, p->lnV1+1, p->lnV1+N,
191	p->lnV2+1, p->lnV2+N);
192	p->lnPivot += N;
193	p->lnV1 += N;
194	p->lnV2 += N;
195	}
196
197	/* The next nPivot lines of the PIVOT are changed into nV1 lines by V1
198	*/
199	static void dbgChngV1(MergeBuilder *p, unsigned int nPivot, unsigned int nV1){
200	blob_appendf(p->pOut, "COPY %u from V1(%u..%u)\n",
201	nV1, p->lnV1+1, p->lnV1+nV1);
202	p->lnPivot += nPivot;
203	p->lnV2 += nPivot;
204	p->lnV1 += nV1;
205	}
206
207	/* The next nPivot lines of the PIVOT are changed into nV2 lines by V2
208	*/
209	static void dbgChngV2(MergeBuilder *p, unsigned int nPivot, unsigned int nV2){
210	blob_appendf(p->pOut, "COPY %u lines FROM V2(%u..%u)\n",
211	nV2, p->lnV2+1, p->lnV2+nV2);
212	p->lnPivot += nPivot;
213	p->lnV1 += nPivot;
214	p->lnV2 += nV2;
215	}
216
217	/* The next nPivot lines of the PIVOT are changed into nV lines from V1 and
218	** V2, which should be the same. In other words, the same change is found
219	** in both V1 and V2.
220	*/
221	static void dbgChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned int nV){
222	blob_appendf(p->pOut, "COPY %u lines from V1(%u..%u) or V2(%u..%u)\n",
223	nV, p->lnV1+1, p->lnV1+nV, p->lnV2+1, p->lnV2+nV);
224	p->lnPivot += nPivot;
225	p->lnV1 += nV;
226	p->lnV2 += nV;
227	}
228
229	/* V1 and V2 have different and overlapping changes. The next nPivot lines
230	** of the PIVOT are converted into nV1 lines of V1 and nV2 lines of V2.
231	*/
232	static void dbgConflict(
233	MergeBuilder *p,
234	unsigned int nPivot,
235	unsigned int nV1,
236	unsigned int nV2
237	){
238	blob_appendf(p->pOut,
239	"CONFLICT %u,%u,%u BASELINE(%u..%u) versus V1(%u..%u) versus V2(%u..%u)\n",
240	nPivot, nV1, nV2,
241	p->lnPivot+1, p->lnPivot+nPivot,
242	p->lnV1+1, p->lnV1+nV1,
243	p->lnV2+1, p->lnV2+nV2);
244	p->lnV1 += nV1;
245	p->lnPivot += nPivot;
246	p->lnV2 += nV2;
247	}
248
249	/* Generic destructor for the MergeBuilder object
250	*/
251	static void dbgDestroy(MergeBuilder *p){
252	memset(p, 0, sizeof(*p));
253	}
254
255	/* Generic initializer for a MergeBuilder object
256	*/
257	static void mergebuilder_init(MergeBuilder *p){
258	memset(p, 0, sizeof(*p));
259	p->xStart = dbgStartEnd;
260	p->xSame = dbgSame;
261	p->xChngV1 = dbgChngV1;
262	p->xChngV2 = dbgChngV2;
263	p->xChngBoth = dbgChngBoth;
264	p->xConflict = dbgConflict;
265	p->xEnd = dbgStartEnd;
266	p->xDestroy = dbgDestroy;
267	}
268
269	/*********************** MergeBuilderToken ******************************/
270	/* This version of MergeBuilder actually performs a merge on file that
271	** are broken up into tokens instead of lines, and puts the result in pOut.
272	*/
273	static void tokenSame(MergeBuilder *p, unsigned int N){
274	blob_append(p->pOut, p->pPivot->aData+p->pPivot->iCursor, N);
275	p->pPivot->iCursor += N;
276	p->pV1->iCursor += N;
277	p->pV2->iCursor += N;
278	}
279	static void tokenChngV1(MergeBuilder *p, unsigned int nPivot, unsigned nV1){
280	blob_append(p->pOut, p->pV1->aData+p->pV1->iCursor, nV1);
281	p->pPivot->iCursor += nPivot;
282	p->pV1->iCursor += nV1;
283	p->pV2->iCursor += nPivot;
284	}
285	static void tokenChngV2(MergeBuilder *p, unsigned int nPivot, unsigned nV2){
286	blob_append(p->pOut, p->pV2->aData+p->pV2->iCursor, nV2);
287	p->pPivot->iCursor += nPivot;
288	p->pV1->iCursor += nPivot;
289	p->pV2->iCursor += nV2;
290	}
291	static void tokenChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned nV){
292	blob_append(p->pOut, p->pV2->aData+p->pV2->iCursor, nV);
293	p->pPivot->iCursor += nPivot;
294	p->pV1->iCursor += nV;
295	p->pV2->iCursor += nV;
296	}
297	static void tokenConflict(
298	MergeBuilder *p,
299	unsigned int nPivot,
300	unsigned int nV1,
301	unsigned int nV2
302	){
303	/* For a token-merge conflict, use the text from the merge-in */
304	blob_append(p->pOut, p->pV2->aData+p->pV2->iCursor, nV2);
305	p->pPivot->iCursor += nPivot;
306	p->pV1->iCursor += nV1;
307	p->pV2->iCursor += nV2;
308	}
309	static void mergebuilder_init_token(MergeBuilder *p){
310	mergebuilder_init(p);
311	p->xSame = tokenSame;
312	p->xChngV1 = tokenChngV1;
313	p->xChngV2 = tokenChngV2;
314	p->xChngBoth = tokenChngBoth;
315	p->xConflict = tokenConflict;
316	p->diffFlags = DIFF_BY_TOKEN;
317	}
318
319	/*
320	** Attempt to do a low-level merge on a conflict. The conflict is
321	** described by the first four parameters, which are the same as the
322	** arguments to the xConflict method of the MergeBuilder object.
323	** This routine attempts to resolve the conflict by looking at
324	** elements of the conflict region that are finer grain than complete
325	** lines of text.
326	**
327	** The result is written into Blob pOut. pOut is initialized by this
328	** routine.
329	*/
330	int merge_try_to_resolve_conflict(
331	MergeBuilder pMB, / MergeBuilder that encounter conflict */
332	unsigned int nPivot, /* Lines of conflict in the pivot */
333	unsigned int nV1, /* Lines of conflict in V1 */
334	unsigned int nV2, /* Lines of conflict in V2 */
335	Blob pOut / Write resolution text here */
336	){
337	int nConflict;
338	MergeBuilder mb;
339	Blob pv, v1, v2;
340	mergebuilder_init_token(&mb);
341	blob_extract_lines(pMB->pPivot, nPivot, &pv);
342	blob_extract_lines(pMB->pV1, nV1, &v1);
343	blob_extract_lines(pMB->pV2, nV2, &v2);
344	blob_zero(pOut);
345	blob_materialize(&pv);
346	blob_materialize(&v1);
347	blob_materialize(&v2);
348	mb.pPivot = &pv;
349	mb.pV1 = &v1;
350	mb.pV2 = &v2;
351	mb.pOut = pOut;
352	nConflict = merge_three_blobs(&mb);
353	blob_reset(&pv);
354	blob_reset(&v1);
355	blob_reset(&v2);
356	/* mb has not allocated any resources, so we do not need to invoke
357	** the xDestroy method. */
358	blob_add_final_newline(pOut);
359	return nConflict;
360	}
361
362
363	/*********************** MergeBuilderText ********************************/
364	/* This version of MergeBuilder actually performs a merge on file and puts
365	** the result in pOut
366	*/
367	static void txtStart(MergeBuilder *p){
368	/* If both pV1 and pV2 start with a UTF-8 byte-order-mark (BOM),
369	** keep it in the output. This should be secure enough not to cause
370	** unintended changes to the merged file and consistent with what
371	** users are using in their source files.
372	*/
373	if( starts_with_utf8_bom(p->pV1, 0) && starts_with_utf8_bom(p->pV2, 0) ){
374	blob_append(p->pOut, (char*)get_utf8_bom(0), -1);
375	}
376	if( contains_crlf(p->pV1) && contains_crlf(p->pV2) ){
377	p->useCrLf = 1;
378	}
379	}
380	static void txtSame(MergeBuilder *p, unsigned int N){
381	blob_copy_lines(p->pOut, p->pPivot, N); p->lnPivot += N;
382	blob_copy_lines(0, p->pV1, N); p->lnV1 += N;
383	blob_copy_lines(0, p->pV2, N); p->lnV2 += N;
384	}
385	static void txtChngV1(MergeBuilder *p, unsigned int nPivot, unsigned int nV1){
386	blob_copy_lines(0, p->pPivot, nPivot); p->lnPivot += nPivot;
387	blob_copy_lines(0, p->pV2, nPivot); p->lnV2 += nPivot;
388	blob_copy_lines(p->pOut, p->pV1, nV1); p->lnV1 += nV1;
389	}
390	static void txtChngV2(MergeBuilder *p, unsigned int nPivot, unsigned int nV2){
391	blob_copy_lines(0, p->pPivot, nPivot); p->lnPivot += nPivot;
392	blob_copy_lines(0, p->pV1, nPivot); p->lnV1 += nPivot;
393	blob_copy_lines(p->pOut, p->pV2, nV2); p->lnV2 += nV2;
394	}
395	static void txtChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned int nV){
396	blob_copy_lines(0, p->pPivot, nPivot); p->lnPivot += nPivot;
397	blob_copy_lines(0, p->pV1, nV); p->lnV1 += nV;
398	blob_copy_lines(p->pOut, p->pV2, nV); p->lnV2 += nV;
399	}
400	static void txtConflict(
401	MergeBuilder *p,
402	unsigned int nPivot,
403	unsigned int nV1,
404	unsigned int nV2
405	){
406	int nRes; /* Lines in the computed conflict resolution */
407	Blob res; /* Text of the conflict resolution */
408
409	merge_try_to_resolve_conflict(p, nPivot, nV1, nV2, &res);
410	nRes = blob_linecount(&res);
411
412	append_merge_mark(p->pOut, 0, p->lnV1+1, p->useCrLf);
413	blob_copy_lines(p->pOut, p->pV1, nV1); p->lnV1 += nV1;
414
415	if( nRes>0 ){
416	append_merge_mark(p->pOut, 1, 0, p->useCrLf);
417	blob_copy_lines(p->pOut, &res, nRes);
418	}
419	blob_reset(&res);
420
421	append_merge_mark(p->pOut, 2, p->lnPivot+1, p->useCrLf);
422	blob_copy_lines(p->pOut, p->pPivot, nPivot); p->lnPivot += nPivot;
423
424	append_merge_mark(p->pOut, 3, p->lnV2+1, p->useCrLf);
425	blob_copy_lines(p->pOut, p->pV2, nV2); p->lnV2 += nV2;
426
427	append_merge_mark(p->pOut, 4, -1, p->useCrLf);
428	}
429	static void mergebuilder_init_text(MergeBuilder *p){
430	mergebuilder_init(p);
431	p->xStart = txtStart;
432	p->xSame = txtSame;
433	p->xChngV1 = txtChngV1;
434	p->xChngV2 = txtChngV2;
435	p->xChngBoth = txtChngBoth;
436	p->xConflict = txtConflict;
437	}
438
439	/*********************** MergeBuilderTcl ********************************/
440	/* Generate merge output formatted for reading by a TCL script.
441	**
442	** The output consists of lines of text, each with 4 tokens. The tokens
443	** represent the content for one line from baseline, v1, v2, and output
444	** respectively. The first character of each token provides auxiliary
445	** information:
446	**
447	** . This line is omitted.
448	** N Name of the file.
449	** T Literal text follows that should have a \n terminator.
450	** R Literal text follows that needs a \r\n terminator.
451	** X Merge conflict.
452	** Z Literal text without a line terminator.
453	** S Skipped lines. Followed by number of lines to skip.
454	** 1 Text is a copy of token 1
455	** 2 Use data from data-token 2
456	** 3 Use data from data-token 3
457	*/
458
459	/* Write text that goes into the interior of a double-quoted string in TCL */
460	static void tclWriteQuotedText(Blob pOut, const char zIn, int nIn){
461	int j;
462	for(j=0; j<nIn; j++){
463	char c = zIn[j];
464	if( c=='\\' ){
465	blob_append(pOut, "\\\\", 2);
466	}else if( c=='"' ){
467	blob_append(pOut, "\\\"", 2);
468	}else if( c<' ' \|\| c>0x7e ){
469	char z[5];
470	z[0] = '\\';
471	z[1] = "01234567"[(c>>6)&0x3];
472	z[2] = "01234567"[(c>>3)&0x7];
473	z[3] = "01234567"[c&0x7];
474	z[4] = 0;
475	blob_append(pOut, z, 4);
476	}else{
477	blob_append_char(pOut, c);
478	}
479	}
480	}
481
482	/* Copy one line of text from pIn and append to pOut, encoded as TCL */
483	static void tclLineOfText(Blob pOut, Blob pIn, char cType){
484	int i, k;
485	for(i=pIn->iCursor; i<pIn->nUsed && pIn->aData[i]!='\n'; i++){}
486	if( i==pIn->nUsed ){
487	k = i;
488	}else if( i>pIn->iCursor && pIn->aData[i-1]=='\r' ){
489	k = i-1;
490	i++;
491	}else{
492	k = i;
493	i++;
494	}
495	blob_append_char(pOut, '"');
496	blob_append_char(pOut, cType);
497	tclWriteQuotedText(pOut, pIn->aData+pIn->iCursor, k-pIn->iCursor);
498	pIn->iCursor = i;
499	blob_append_char(pOut, '"');
500	}
501	static void tclStart(MergeBuilder *p){
502	Blob *pOut = p->pOut;
503	blob_append(pOut, "\"N", 2);
504	tclWriteQuotedText(pOut, p->zPivot, (int)strlen(p->zPivot));
505	blob_append(pOut, "\" \"N", 4);
506	tclWriteQuotedText(pOut, p->zV1, (int)strlen(p->zV1));
507	blob_append(pOut, "\" \"N", 4);
508	tclWriteQuotedText(pOut, p->zV2, (int)strlen(p->zV2));
509	blob_append(pOut, "\" \"N", 4);
510	if( p->zOut ){
511	tclWriteQuotedText(pOut, p->zOut, (int)strlen(p->zOut));
512	}else{
513	blob_append(pOut, "(Merge Result)", -1);
514	}
515	blob_append(pOut, "\"\n", 2);
516	}
517	static void tclSame(MergeBuilder *p, unsigned int N){
518	int i = 0;
519	int nSkip;
520
521	if( p->lnPivot>=2 \|\| p->lnV1>2 \|\| p->lnV2>2 ){
522	while( i<N && i<p->nContext ){
523	tclLineOfText(p->pOut, p->pPivot, 'T');
524	blob_append(p->pOut, " 1 1 1\n", 7);
525	i++;
526	}
527	nSkip = N - p->nContext*2;
528	}else{
529	nSkip = N - p->nContext;
530	}
531	if( nSkip>0 ){
532	blob_appendf(p->pOut, "\"S%d %d %d %d\" . . .\n",
533	nSkip, nSkip, nSkip, nSkip);
534	blob_copy_lines(0, p->pPivot, nSkip);
535	i += nSkip;
536	}
537
538	p->lnPivot += N;
539	p->lnV1 += N;
540	p->lnV2 += N;
541
542	if( p->lnPivot<p->mxPivot \|\| p->lnV1<p->mxV1 \|\| p->lnV2<p->mxV2 ){
543	while( i<N ){
544	tclLineOfText(p->pOut, p->pPivot, 'T');
545	blob_append(p->pOut, " 1 1 1\n", 7);
546	i++;
547	}
548	}
549
550	blob_copy_lines(0, p->pV1, N);
551	blob_copy_lines(0, p->pV2, N);
552	}
553	static void tclChngV1(MergeBuilder *p, unsigned int nPivot, unsigned int nV1){
554	int i;
555	for(i=0; i<nPivot && i<nV1; i++){
556	tclLineOfText(p->pOut, p->pPivot, 'T');
557	blob_append_char(p->pOut, ' ');
558	tclLineOfText(p->pOut, p->pV1, 'T');
559	blob_append(p->pOut, " 1 2\n", 5);
560	}
561	while( i<nPivot ){
562	tclLineOfText(p->pOut, p->pPivot, 'T');
563	blob_append(p->pOut, " . 1 .\n", 7);
564	i++;
565	}
566	while( i<nV1 ){
567	blob_append(p->pOut, ". ", 2);
568	tclLineOfText(p->pOut, p->pV1, 'T');
569	blob_append(p->pOut, " . 2\n", 5);
570	i++;
571	}
572	p->lnPivot += nPivot;
573	p->lnV1 += nV1;
574	p->lnV2 += nPivot;
575	blob_copy_lines(0, p->pV2, nPivot);
576	}
577	static void tclChngV2(MergeBuilder *p, unsigned int nPivot, unsigned int nV2){
578	int i;
579	for(i=0; i<nPivot && i<nV2; i++){
580	tclLineOfText(p->pOut, p->pPivot, 'T');
581	blob_append(p->pOut, " 1 ", 3);
582	tclLineOfText(p->pOut, p->pV2, 'T');
583	blob_append(p->pOut, " 3\n", 3);
584	}
585	while( i<nPivot ){
586	tclLineOfText(p->pOut, p->pPivot, 'T');
587	blob_append(p->pOut, " 1 . .\n", 7);
588	i++;
589	}
590	while( i<nV2 ){
591	blob_append(p->pOut, ". . ", 4);
592	tclLineOfText(p->pOut, p->pV2, 'T');
593	blob_append(p->pOut, " 3\n", 3);
594	i++;
595	}
596	p->lnPivot += nPivot;
597	p->lnV1 += nPivot;
598	p->lnV2 += nV2;
599	blob_copy_lines(0, p->pV1, nPivot);
600	}
601	static void tclChngBoth(MergeBuilder *p, unsigned int nPivot, unsigned int nV){
602	int i;
603	for(i=0; i<nPivot && i<nV; i++){
604	tclLineOfText(p->pOut, p->pPivot, 'T');
605	blob_append_char(p->pOut, ' ');
606	tclLineOfText(p->pOut, p->pV1, 'T');
607	blob_append(p->pOut, " 2 2\n", 5);
608	}
609	while( i<nPivot ){
610	tclLineOfText(p->pOut, p->pPivot, 'T');
611	blob_append(p->pOut, " . . .\n", 7);
612	i++;
613	}
614	while( i<nV ){
615	blob_append(p->pOut, ". ", 2);
616	tclLineOfText(p->pOut, p->pV1, 'T');
617	blob_append(p->pOut, " 2 2\n", 5);
618	i++;
619	}
620	p->lnPivot += nPivot;
621	p->lnV1 += nV;
622	p->lnV2 += nV;
623	blob_copy_lines(0, p->pV2, nV);
624	}
625	static void tclConflict(
626	MergeBuilder *p,
627	unsigned int nPivot,
628	unsigned int nV1,
629	unsigned int nV2
630	){
631	int mx = nPivot;
632	int i;
633	int nRes;
634	Blob res;
635
636	merge_try_to_resolve_conflict(p, nPivot, nV1, nV2, &res);
637	nRes = blob_linecount(&res);
638	if( nV1>mx ) mx = nV1;
639	if( nV2>mx ) mx = nV2;
640	if( nRes>mx ) mx = nRes;
641	if( nRes>0 ){
642	blob_appendf(p->pOut, "\"S0 0 0 %d\" . . .\n", nV2+2);
643	}
644	for(i=0; i<mx; i++){
645	if( i<nPivot ){
646	tclLineOfText(p->pOut, p->pPivot, 'X');
647	}else{
648	blob_append_char(p->pOut, '.');
649	}
650	blob_append_char(p->pOut, ' ');
651	if( i<nV1 ){
652	tclLineOfText(p->pOut, p->pV1, 'X');
653	}else{
654	blob_append_char(p->pOut, '.');
655	}
656	blob_append_char(p->pOut, ' ');
657	if( i<nV2 ){
658	tclLineOfText(p->pOut, p->pV2, 'X');
659	}else{
660	blob_append_char(p->pOut, '.');
661	}
662	if( i<nRes ){
663	blob_append_char(p->pOut, ' ');
664	tclLineOfText(p->pOut, &res, 'X');
665	blob_append_char(p->pOut, '\n');
666	}else{
667	blob_append(p->pOut, " .\n", 3);
668	}
669	if( i==mx-1 ){
670	blob_appendf(p->pOut, "\"S0 0 0 %d\" . . .\n", nPivot+nV1+3);
671	}
672	}
673	blob_reset(&res);
674	p->lnPivot += nPivot;
675	p->lnV1 += nV1;
676	p->lnV2 += nV2;
677	}
678	void mergebuilder_init_tcl(MergeBuilder *p){
679	mergebuilder_init(p);
680	p->xStart = tclStart;
681	p->xSame = tclSame;
682	p->xChngV1 = tclChngV1;
683	p->xChngV2 = tclChngV2;
684	p->xChngBoth = tclChngBoth;
685	p->xConflict = tclConflict;
686	}
687	/*****************************************************************************/
688
689	/*
690	** The aC[] array contains triples of integers. Within each triple, the
691	** elements are:
692	**
693	** (0) The number of lines to copy
694	** (1) The number of lines to delete
695	** (2) The number of liens to insert
696	**
697	** Suppose we want to advance over sz lines of the original file. This routine
698	** returns true if that advance would land us on a copy operation. It
699	** returns false if the advance would end on a delete.
700	*/
701	static int ends_with_copy(int *aC, int sz){
702	while( sz>0 && (aC[0]>0 \|\| aC[1]>0 \|\| aC[2]>0) ){
703	if( aC[0]>=sz ) return 1;
704	sz -= aC[0];
705	if( aC[1]>sz ) return 0;
706	sz -= aC[1];
707	aC += 3;
708	}
709	return 1;
710	}
711
712	/*
713	** aC[] is an "edit triple" for changes from A to B. Advance through
714	** this triple to determine the number of lines to bypass on B in order
715	** to match an advance of sz lines on A.
716	*/
717	static int skip_conflict(
718	int aC, / Array of integer triples describing the edit */
719	int i, /* Index in aC[] of current location */
720	int sz, /* Lines of A that have been skipped */
721	unsigned int pLn / OUT: Lines of B to skip to keep aligment with A */
722	){
723	*pLn = 0;
724	while( sz>0 ){
725	if( aC[i]==0 && aC[i+1]==0 && aC[i+2]==0 ) break;
726	if( aC[i]>=sz ){
727	aC[i] -= sz;
728	*pLn += sz;
729	break;
730	}
731	*pLn += aC[i];
732	*pLn += aC[i+2];
733	sz -= aC[i] + aC[i+1];
734	i += 3;
735	}
736	return i;
737	}
738
739	/*
740	** Do a three-way merge. Initialize pOut to contain the result.
741	**
742	** The merge is an edit against pV2. Both pV1 and pV2 have a
	@@ -199,156 +746,113 @@
746	** The return is 0 upon complete success. If any input file is binary,
747	** -1 is returned and pOut is unmodified. If there are merge
748	** conflicts, the merge proceeds as best as it can and the number
749	** of conflicts is returns
750	*/
751	int merge_three_blobs(MergeBuilder *p){
752	int aC1; / Changes from pPivot to pV1 */
753	int aC2; / Changes from pPivot to pV2 */
754	int i1, i2; /* Index into aC1[] and aC2[] */
755	int nCpy, nDel, nIns; /* Number of lines to copy, delete, or insert */
756	int limit1, limit2; /* Sizes of aC1[] and aC2[] */
757	int nConflict = 0; /* Number of merge conflicts seen so far */


758	DiffConfig DCfg;
759



















760	/* Compute the edits that occur from pPivot => pV1 (into aC1)
761	** and pPivot => pV2 (into aC2). Each of the aC1 and aC2 arrays is
762	** an array of integer triples. Within each triple, the first integer
763	** is the number of lines of text to copy directly from the pivot,
764	** the second integer is the number of lines of text to omit from the
765	** pivot, and the third integer is the number of lines of text that are
766	** inserted. The edit array ends with a triple of 0,0,0.
767	*/
768	diff_config_init(&DCfg, 0);
769	DCfg.diffFlags = p->diffFlags;
770	aC1 = text_diff(p->pPivot, p->pV1, 0, &DCfg);
771	aC2 = text_diff(p->pPivot, p->pV2, 0, &DCfg);
772	if( aC1==0 \|\| aC2==0 ){
773	free(aC1);
774	free(aC2);
775	return -1;
776	}
777
778	blob_rewind(p->pV1); /* Rewind inputs: Needed to reconstruct output */
779	blob_rewind(p->pV2);
780	blob_rewind(p->pPivot);
781
782	/* Determine the length of the aC1[] and aC2[] change vectors */
783	p->mxPivot = 0;
784	p->mxV1 = 0;
785	for(i1=0; aC1[i1] \|\| aC1[i1+1] \|\| aC1[i1+2]; i1+=3){
786	p->mxPivot += aC1[i1] + aC1[i1+1];
787	p->mxV1 += aC1[i1] + aC1[i1+2];
788	}
789	limit1 = i1;
790	p->mxV2 = 0;
791	for(i2=0; aC2[i2] \|\| aC2[i2+1] \|\| aC2[i2+2]; i2+=3){
792	p->mxV2 += aC2[i2] + aC2[i2+2];
793	}
794	limit2 = i2;
795
796	/* Output header text and do any other required initialization */
797	p->xStart(p);






798
799	/* Loop over the two edit vectors and use them to compute merged text
800	** which is written into pOut. i1 and i2 are multiples of 3 which are
801	** indices into aC1[] and aC2[] to the edit triple currently being
802	** processed
803	*/
804	i1 = i2 = 0;

805	while( i1<limit1 && i2<limit2 ){




806	if( aC1[i1]>0 && aC2[i2]>0 ){
807	/* Output text that is unchanged in both V1 and V2 */
808	nCpy = min(aC1[i1], aC2[i2]);
809	p->xSame(p, nCpy);



810	aC1[i1] -= nCpy;
811	aC2[i2] -= nCpy;
812	}else
813	if( aC1[i1] >= aC2[i2+1] && aC1[i1]>0 && aC2[i2+1]+aC2[i2+2]>0 ){
814	/* Output edits to V2 that occurs within unchanged regions of V1 */
815	nDel = aC2[i2+1];
816	nIns = aC2[i2+2];
817	p->xChngV2(p, nDel, nIns);



818	aC1[i1] -= nDel;
819	i2 += 3;
820	}else
821	if( aC2[i2] >= aC1[i1+1] && aC2[i2]>0 && aC1[i1+1]+aC1[i1+2]>0 ){
822	/* Output edits to V1 that occur within unchanged regions of V2 */
823	nDel = aC1[i1+1];
824	nIns = aC1[i1+2];
825	p->xChngV1(p, nDel, nIns);



826	aC2[i2] -= nDel;
827	i1 += 3;
828	}else
829	if( sameEdit(&aC1[i1], &aC2[i2], p->pV1, p->pV2) ){
830	/* Output edits that are identical in both V1 and V2. */
831	assert( aC1[i1]==0 );
832	nDel = aC1[i1+1];
833	nIns = aC1[i1+2];
834	p->xChngBoth(p, nDel, nIns);



835	i1 += 3;
836	i2 += 3;
837	}else
838	{
839	/* We have found a region where different edits to V1 and V2 overlap.
840	** This is a merge conflict. Find the size of the conflict, then
841	** output both possible edits separated by distinctive marks.
842	*/
843	unsigned int sz = 1; /* Size of the conflict in the pivot, in lines */
844	unsigned int nV1, nV2; /* Size of conflict in V1 and V2, in lines */
845	nConflict++;
846	while( !ends_with_copy(&aC1[i1], sz) \|\| !ends_with_copy(&aC2[i2], sz) ){
847	sz++;
848	}
849	i1 = skip_conflict(aC1, i1, sz, &nV1);
850	i2 = skip_conflict(aC2, i2, sz, &nV2);
851	p->xConflict(p, sz, nV1, nV2);
852	}
853









854	/* If we are finished with an edit triple, advance to the next
855	** triple.
856	*/
857	if( i1<limit1 && aC1[i1]==0 && aC1[i1+1]==0 && aC1[i1+2]==0 ) i1+=3;
858	if( i2<limit2 && aC2[i2]==0 && aC2[i2+1]==0 && aC2[i2+2]==0 ) i2+=3;
	@@ -356,20 +860,18 @@
860
861	/* When one of the two edit vectors reaches its end, there might still
862	** be an insert in the other edit vector. Output this remaining
863	** insert.
864	*/



865	if( i1<limit1 && aC1[i1+2]>0 ){
866	p->xChngV1(p, 0, aC1[i1+2]);

867	}else if( i2<limit2 && aC2[i2+2]>0 ){
868	p->xChngV2(p, 0, aC2[i2+2]);

869	}
870
871	/* Output footer text */
872	p->xEnd(p);
873
874	free(aC1);
875	free(aC2);
876	return nConflict;
877	}
	@@ -384,11 +886,12 @@
886	const char *z = blob_buffer(p);
887	int n = blob_size(p) - len + 1;
888	assert( len==(int)strlen(mergeMarker[1]) );
889	assert( len==(int)strlen(mergeMarker[2]) );
890	assert( len==(int)strlen(mergeMarker[3]) );
891	assert( len==(int)strlen(mergeMarker[4]) );
892	assert( count(mergeMarker)==5 );
893	for(i=0; i<n; ){
894	for(j=0; j<4; j++){
895	if( (memcmp(&z[i], mergeMarker[j], len)==0) ){
896	return 1;
897	}
	@@ -408,18 +911,106 @@
911	blob_read_from_file(&file, zFullpath, ExtFILE);
912	rc = contains_merge_marker(&file);
913	blob_reset(&file);
914	return rc;
915	}
916
917	/*
918	** Show merge output in a Tcl/Tk window, in response to the --tk option
919	** to the "merge" or "3-way-merge" command.
920	**
921	** If fossil has direct access to a Tcl interpreter (either loaded
922	** dynamically through stubs or linked in statically), we can use it
923	** directly. Otherwise:
924	** (1) Write the Tcl/Tk script used for rendering into a temp file.
925	** (2) Invoke "tclsh" on the temp file using fossil_system().
926	** (3) Delete the temp file.
927	*/
928	void merge_tk(const char *zSubCmd, int firstArg){
929	int i;
930	Blob script;
931	const char *zTempFile = 0;
932	char *zCmd;
933	const char *zTclsh;
934	const char *zCnt;
935	int bDarkMode = find_option("dark",0,0)!=0;
936	int nContext;
937	zCnt = find_option("context", "c", 1);
938	if( zCnt==0 ){
939	nContext = 6;
940	}else{
941	nContext = atoi(zCnt);
942	if( nContext<0 ) nContext = 0xfffffff;
943	}
944	blob_zero(&script);
945	blob_appendf(&script, "set ncontext %d\n", nContext);
946	blob_appendf(&script, "set fossilcmd {\| \"%/\" %s -tcl",
947	g.nameOfExe, zSubCmd);
948	find_option("tcl",0,0);
949	find_option("debug",0,0);
950	zTclsh = find_option("tclsh",0,1);
951	if( zTclsh==0 ){
952	zTclsh = db_get("tclsh",0);
953	}
954	/* The undocumented --script FILENAME option causes the Tk script to
955	** be written into the FILENAME instead of being run. This is used
956	** for testing and debugging. */
957	zTempFile = find_option("script",0,1);
958	verify_all_options();
959
960	if( (g.argc - firstArg)!=3 ){
961	fossil_fatal("Requires 3 filename arguments");
962	}
963
964	for(i=firstArg; i<g.argc; i++){
965	const char *z = g.argv[i];
966	if( sqlite3_strglob("}",z) ){
967	blob_appendf(&script, " {%/}", z);
968	}else{
969	int j;
970	blob_append(&script, " ", 1);
971	for(j=0; z[j]; j++) blob_appendf(&script, "\\%03o", (unsigned char)z[j]);
972	}
973	}
974	blob_appendf(&script, "}\nset darkmode %d\n", bDarkMode);
975	blob_appendf(&script, "%s", builtin_file("merge.tcl", 0));
976	if( zTempFile ){
977	blob_write_to_file(&script, zTempFile);
978	fossil_print("To see the merge, run: %s \"%s\"\n", zTclsh, zTempFile);
979	}else{
980	#if defined(FOSSIL_ENABLE_TCL)
981	Th_FossilInit(TH_INIT_DEFAULT);
982	if( evaluateTclWithEvents(g.interp, &g.tcl, blob_str(&script),
983	blob_size(&script), 1, 1, 0)==TCL_OK ){
984	blob_reset(&script);
985	return;
986	}
987	/*
988	* If evaluation of the Tcl script fails, the reason may be that Tk
989	* could not be found by the loaded Tcl, or that Tcl cannot be loaded
990	* dynamically (e.g. x64 Tcl with x86 Fossil). Therefore, fallback
991	* to using the external "tclsh", if available.
992	*/
993	#endif
994	zTempFile = write_blob_to_temp_file(&script);
995	zCmd = mprintf("%$ %$", zTclsh, zTempFile);
996	fossil_system(zCmd);
997	file_delete(zTempFile);
998	fossil_free(zCmd);
999	}
1000	blob_reset(&script);
1001	}
1002
1003
1004	/*
1005	** COMMAND: 3-way-merge*
1006	**
1007	** Usage: %fossil 3-way-merge BASELINE V1 V2 [MERGED]
1008	**
1009	** Inputs are files BASELINE, V1, and V2. The file MERGED is generated
1010	** as output. If no MERGED file is specified, output is sent to
1011	** stdout.
1012	**
1013	** BASELINE is a common ancestor of two files V1 and V2 that have diverging
1014	** edits. The generated output file MERGED is the combination of all
1015	** changes in both V1 and V2.
1016	**
	@@ -436,38 +1027,75 @@
1027	** cp Xup.c Xbase.c
1028	** # Verify that everything still works
1029	** fossil commit
1030	**
1031	*/
1032	void merge_3way_cmd(void){
1033	MergeBuilder s;
1034	int nConflict;
1035	Blob pivot, v1, v2, out;
1036	int noWarn = 0;
1037	const char *zCnt;
1038
1039	if( find_option("tk", 0, 0)!=0 ){
1040	merge_tk("3-way-merge", 2);
1041	return;
1042	}
1043	mergebuilder_init_text(&s);
1044	if( find_option("debug", 0, 0) ){
1045	mergebuilder_init(&s);
1046	}
1047	if( find_option("tcl", 0, 0) ){
1048	mergebuilder_init_tcl(&s);
1049	noWarn = 1;
1050	}
1051	zCnt = find_option("context", "c", 1);
1052	if( zCnt ){
1053	s.nContext = atoi(zCnt);
1054	if( s.nContext<0 ) s.nContext = 0xfffffff;
1055	}else{
1056	s.nContext = 6;
1057	}
1058	blob_zero(&pivot); s.pPivot = &pivot;
1059	blob_zero(&v1); s.pV1 = &v1;
1060	blob_zero(&v2); s.pV2 = &v2;
1061	blob_zero(&out); s.pOut = &out;
1062
1063	/* We should be done with options.. */
1064	verify_all_options();
1065
1066	if( g.argc!=6 && g.argc!=5 ){
1067	usage("[OPTIONS] PIVOT V1 V2 [MERGED]");
1068	}
1069	s.zPivot = file_tail(g.argv[2]);
1070	s.zV1 = file_tail(g.argv[3]);
1071	s.zV2 = file_tail(g.argv[4]);
1072	if( blob_read_from_file(s.pPivot, g.argv[2], ExtFILE)<0 ){
1073	fossil_fatal("cannot read %s", g.argv[2]);
1074	}
1075	if( blob_read_from_file(s.pV1, g.argv[3], ExtFILE)<0 ){
1076	fossil_fatal("cannot read %s", g.argv[3]);
1077	}
1078	if( blob_read_from_file(s.pV2, g.argv[4], ExtFILE)<0 ){
1079	fossil_fatal("cannot read %s", g.argv[4]);
1080	}
1081	nConflict = merge_three_blobs(&s);
1082	if( g.argc==6 ){
1083	s.zOut = file_tail(g.argv[5]);
1084	blob_write_to_file(s.pOut, g.argv[5]);
1085	}else{
1086	s.zOut = "(Merge Result)";
1087	blob_write_to_file(s.pOut, "-");
1088	}
1089	s.xDestroy(&s);
1090	blob_reset(&pivot);
1091	blob_reset(&v1);
1092	blob_reset(&v2);
1093	blob_reset(&out);
1094	if( nConflict>0 && !noWarn ){
1095	fossil_warning("WARNING: %d merge conflicts", nConflict);
1096	}
1097	}
1098
1099	/*
1100	** aSubst is an array of string pairs. The first element of each pair is
1101	** a string that begins with %. The second element is a replacement for that
	@@ -505,32 +1133,32 @@
1133
1134	#if INTERFACE
1135	/*
1136	** Flags to the 3-way merger
1137	*/
1138	#define MERGE_DRYRUN 0x0001
1139	/*
1140	** The MERGE_KEEP_FILES flag specifies that merge_3way() should retain
1141	** its temporary files on error. By default they are removed after the
1142	** merge, regardless of success or failure.
1143	*/
1144	#define MERGE_KEEP_FILES 0x0002
1145	#endif
1146
1147
1148	/*
1149	** This routine is a wrapper around merge_three_blobs() with the following
1150	** enhancements:
1151	**
1152	** (1) If the merge-command is defined, then use the external merging
1153	** program specified instead of the built-in blob-merge to do the
1154	** merging. Panic if the external merger fails.
1155	Not currently implemented **
1156	**
1157	** (2) If gmerge-command is defined and there are merge conflicts in
1158	** merge_three_blobs() then invoke the external graphical merger
1159	** to resolve the conflicts.
1160	**
1161	** (3) If a merge conflict occurs and gmerge-command is not defined,
1162	** then write the pivot, original, and merge-in files to the
1163	** filesystem.
1164	*/
	@@ -539,30 +1167,37 @@
1167	const char zV1, / Name of file for version merging into (mine) */
1168	Blob pV2, / Version merging from (yours) */
1169	Blob pOut, / Output written here */
1170	unsigned mergeFlags /* Flags that control operation */
1171	){
1172	Blob v1; /* Content of zV1 */
1173	int rc; /* Return code of subroutines and this routine */
1174	const char zGMerge; / Name of the gmerge command */
1175	MergeBuilder s; /* The merge state */
1176
1177	mergebuilder_init_text(&s);
1178	s.pPivot = pPivot;
1179	s.pV1 = &v1;
1180	s.pV2 = pV2;
1181	blob_zero(pOut);
1182	s.pOut = pOut;
1183	blob_read_from_file(s.pV1, zV1, ExtFILE);
1184	rc = merge_three_blobs(&s);
1185	zGMerge = rc<=0 ? 0 : db_get("gmerge-command", 0);
1186	if( (mergeFlags & MERGE_DRYRUN)==0
1187	&& ((zGMerge!=0 && zGMerge[0]!=0)
1188	\|\| (rc!=0 && (mergeFlags & MERGE_KEEP_FILES)!=0)) ){
1189	char zPivot; / Name of the pivot file */
1190	char zOrig; / Name of the original content file */
1191	char zOther; / Name of the merge file */
1192
1193	zPivot = file_newname(zV1, "baseline", 1);
1194	blob_write_to_file(s.pPivot, zPivot);
1195	zOrig = file_newname(zV1, "original", 1);
1196	blob_write_to_file(s.pV1, zOrig);
1197	zOther = file_newname(zV1, "merge", 1);
1198	blob_write_to_file(s.pV2, zOther);
1199	if( rc>0 ){
1200	if( zGMerge && zGMerge[0] ){
1201	char zOut; / Temporary output file */
1202	char zCmd; / Command to invoke */
1203	const char azSubst[8]; / Strings to be substituted */
	@@ -590,7 +1225,8 @@
1225	fossil_free(zPivot);
1226	fossil_free(zOrig);
1227	fossil_free(zOther);
1228	}
1229	blob_reset(&v1);
1230	s.xDestroy(&s);
1231	return rc;
1232	}
1233

M src/patch.c

+6 -2

		--- src/patch.c
		+++ src/patch.c
		@@ -464,18 +464,22 @@
464	464
465	465	/* Deletions */
466	466	db_prepare(&q, "SELECT pathname FROM patch.chng"
467	467	" WHERE origname IS NULL AND delta IS NULL");
468	468	while( db_step(&q)==SQLITE_ROW ){
469		- blob_append_escaped_arg(&cmd, g.nameOfExe, 1);
470		- blob_appendf(&cmd, " rm --hard %$\n", db_column_text(&q,0));
	469	+ if( blob_size(&cmd)==0 ){
	470	+ blob_append_escaped_arg(&cmd, g.nameOfExe, 1);
	471	+ blob_appendf(&cmd, " rm --hard");
	472	+ }
	473	+ blob_appendf(&cmd, " %$", db_column_text(&q,0));
471	474	if( mFlags & PATCH_VERBOSE ){
472	475	fossil_print("%-10s %s\n", "DELETE", db_column_text(&q,0));
473	476	}
474	477	}
475	478	db_finalize(&q);
476	479	if( blob_size(&cmd)>0 ){
	480	+ blob_appendf(&cmd, "\n");
477	481	if( mFlags & PATCH_DRYRUN ){
478	482	fossil_print("%s", blob_str(&cmd));
479	483	}else{
480	484	int rc = fossil_unsafe_system(blob_str(&cmd));
481	485	if( rc ){
482	486

	--- src/patch.c
	+++ src/patch.c
	@@ -464,18 +464,22 @@
464
465	/* Deletions */
466	db_prepare(&q, "SELECT pathname FROM patch.chng"
467	" WHERE origname IS NULL AND delta IS NULL");
468	while( db_step(&q)==SQLITE_ROW ){
469	blob_append_escaped_arg(&cmd, g.nameOfExe, 1);
470	blob_appendf(&cmd, " rm --hard %$\n", db_column_text(&q,0));



471	if( mFlags & PATCH_VERBOSE ){
472	fossil_print("%-10s %s\n", "DELETE", db_column_text(&q,0));
473	}
474	}
475	db_finalize(&q);
476	if( blob_size(&cmd)>0 ){

477	if( mFlags & PATCH_DRYRUN ){
478	fossil_print("%s", blob_str(&cmd));
479	}else{
480	int rc = fossil_unsafe_system(blob_str(&cmd));
481	if( rc ){
482

	--- src/patch.c
	+++ src/patch.c
	@@ -464,18 +464,22 @@
464
465	/* Deletions */
466	db_prepare(&q, "SELECT pathname FROM patch.chng"
467	" WHERE origname IS NULL AND delta IS NULL");
468	while( db_step(&q)==SQLITE_ROW ){
469	if( blob_size(&cmd)==0 ){
470	blob_append_escaped_arg(&cmd, g.nameOfExe, 1);
471	blob_appendf(&cmd, " rm --hard");
472	}
473	blob_appendf(&cmd, " %$", db_column_text(&q,0));
474	if( mFlags & PATCH_VERBOSE ){
475	fossil_print("%-10s %s\n", "DELETE", db_column_text(&q,0));
476	}
477	}
478	db_finalize(&q);
479	if( blob_size(&cmd)>0 ){
480	blob_appendf(&cmd, "\n");
481	if( mFlags & PATCH_DRYRUN ){
482	fossil_print("%s", blob_str(&cmd));
483	}else{
484	int rc = fossil_unsafe_system(blob_str(&cmd));
485	if( rc ){
486

M src/path.c

		--- src/path.c
		+++ src/path.c
		@@ -542,10 +542,11 @@
542	542	pChng = pAll;
543	543	pAll = pAll->pNext;
544	544	fossil_free(pChng);
545	545	}
546	546	}
	547	+ path_reset();
547	548	}
548	549
549	550	/*
550	551	** COMMAND: test-name-changes
551	552	**
552	553

	--- src/path.c
	+++ src/path.c
	@@ -542,10 +542,11 @@
542	pChng = pAll;
543	pAll = pAll->pNext;
544	fossil_free(pChng);
545	}
546	}

547	}
548
549	/*
550	** COMMAND: test-name-changes
551	**
552

	--- src/path.c
	+++ src/path.c
	@@ -542,10 +542,11 @@
542	pChng = pAll;
543	pAll = pAll->pNext;
544	fossil_free(pChng);
545	}
546	}
547	path_reset();
548	}
549
550	/*
551	** COMMAND: test-name-changes
552	**
553

M src/pikchrshow.c

+12 -2

		--- src/pikchrshow.c
		+++ src/pikchrshow.c
		@@ -207,12 +207,22 @@
207	207	blob_append(pOut, zOut, -1);
208	208	if(PIKCHR_PROCESS_DIV & pikFlags){
209	209	blob_append(pOut, "</div>\n", 7);
210	210	}
211	211	if(PIKCHR_PROCESS_SRC & pikFlags){
212		- blob_appendf(pOut, "<pre class='pikchr-src'>%h</pre>\n",
213		- blob_str(&bIn));
	212	+ static int counter = 0;
	213	+ ++counter;
	214	+ blob_appendf(pOut, "<div class='pikchr-src'>"
	215	+ "<pre id='pikchr-src-%d'>%h</pre>"
	216	+ "<span class='hidden'>"
	217	+ "<a href='%R/pikchrshow?fromSession' "
	218	+ "class='pikchr-src-pikchrshow' target='_new-%d' "
	219	+ "data-pikchrid='pikchr-src-%d' "
	220	+ "title='Open this pikchr in /pikchrshow'"
	221	+ ">→ /pikchrshow</a></span>"
	222	+ "</div>\n",
	223	+ counter, blob_str(&bIn), counter, counter);
214	224	}
215	225	if(PIKCHR_PROCESS_DIV & pikFlags){
216	226	blob_append(pOut, "</div>\n", 7);
217	227	}
218	228	}else{
219	229

	--- src/pikchrshow.c
	+++ src/pikchrshow.c
	@@ -207,12 +207,22 @@
207	blob_append(pOut, zOut, -1);
208	if(PIKCHR_PROCESS_DIV & pikFlags){
209	blob_append(pOut, "</div>\n", 7);
210	}
211	if(PIKCHR_PROCESS_SRC & pikFlags){
212	blob_appendf(pOut, "<pre class='pikchr-src'>%h</pre>\n",
213	blob_str(&bIn));










214	}
215	if(PIKCHR_PROCESS_DIV & pikFlags){
216	blob_append(pOut, "</div>\n", 7);
217	}
218	}else{
219

	--- src/pikchrshow.c
	+++ src/pikchrshow.c
	@@ -207,12 +207,22 @@
207	blob_append(pOut, zOut, -1);
208	if(PIKCHR_PROCESS_DIV & pikFlags){
209	blob_append(pOut, "</div>\n", 7);
210	}
211	if(PIKCHR_PROCESS_SRC & pikFlags){
212	static int counter = 0;
213	++counter;
214	blob_appendf(pOut, "<div class='pikchr-src'>"
215	"<pre id='pikchr-src-%d'>%h</pre>"
216	"<span class='hidden'>"
217	"<a href='%R/pikchrshow?fromSession' "
218	"class='pikchr-src-pikchrshow' target='_new-%d' "
219	"data-pikchrid='pikchr-src-%d' "
220	"title='Open this pikchr in /pikchrshow'"
221	">→ /pikchrshow</a></span>"
222	"</div>\n",
223	counter, blob_str(&bIn), counter, counter);
224	}
225	if(PIKCHR_PROCESS_DIV & pikFlags){
226	blob_append(pOut, "</div>\n", 7);
227	}
228	}else{
229

M src/rebuild.c

+9 -3

		--- src/rebuild.c
		+++ src/rebuild.c
		@@ -390,25 +390,31 @@
390	390	}
391	391	manifest_disable_event_triggers();
392	392	rebuild_update_schema();
393	393	blob_init(&sql, 0, 0);
394	394	db_unprotect(PROTECT_ALL);
	395	+#ifndef SQLITE_PREPARE_DONT_LOG
	396	+ g.dbIgnoreErrors++;
	397	+#endif
395	398	db_prepare(&q,
396		- "SELECT name FROM sqlite_schema /scan/"
397		- " WHERE type='table'"
	399	+ "SELECT name FROM pragma_table_list /scan/"
	400	+ " WHERE schema='repository' AND type IN ('table','virtual')"
398	401	" AND name NOT IN ('admin_log', 'blob','delta','rcvfrom','user','alias',"
399	402	"'config','shun','private','reportfmt',"
400	403	"'concealed','accesslog','modreq',"
401	404	"'purgeevent','purgeitem','unversioned',"
402	405	"'subscriber','pending_alert','chat')"
403	406	" AND name NOT GLOB 'sqlite_*'"
404		- " AND name NOT GLOB 'fx_*'"
	407	+ " AND name NOT GLOB 'fx_*';"
405	408	);
406	409	while( db_step(&q)==SQLITE_ROW ){
407	410	blob_appendf(&sql, "DROP TABLE IF EXISTS \"%w\";\n", db_column_text(&q,0));
408	411	}
409	412	db_finalize(&q);
	413	+#ifndef SQLITE_PREPARE_DONT_LOG
	414	+ g.dbIgnoreErrors--;
	415	+#endif
410	416	db_multi_exec("%s", blob_str(&sql)/safe-for-%s/);
411	417	blob_reset(&sql);
412	418	db_multi_exec("%s", zRepositorySchema2/safe-for-%s/);
413	419	ticket_create_table(0);
414	420	shun_artifacts();
415	421

	--- src/rebuild.c
	+++ src/rebuild.c
	@@ -390,25 +390,31 @@
390	}
391	manifest_disable_event_triggers();
392	rebuild_update_schema();
393	blob_init(&sql, 0, 0);
394	db_unprotect(PROTECT_ALL);



395	db_prepare(&q,
396	"SELECT name FROM sqlite_schema /scan/"
397	" WHERE type='table'"
398	" AND name NOT IN ('admin_log', 'blob','delta','rcvfrom','user','alias',"
399	"'config','shun','private','reportfmt',"
400	"'concealed','accesslog','modreq',"
401	"'purgeevent','purgeitem','unversioned',"
402	"'subscriber','pending_alert','chat')"
403	" AND name NOT GLOB 'sqlite_*'"
404	" AND name NOT GLOB 'fx_*'"
405	);
406	while( db_step(&q)==SQLITE_ROW ){
407	blob_appendf(&sql, "DROP TABLE IF EXISTS \"%w\";\n", db_column_text(&q,0));
408	}
409	db_finalize(&q);



410	db_multi_exec("%s", blob_str(&sql)/safe-for-%s/);
411	blob_reset(&sql);
412	db_multi_exec("%s", zRepositorySchema2/safe-for-%s/);
413	ticket_create_table(0);
414	shun_artifacts();
415

	--- src/rebuild.c
	+++ src/rebuild.c
	@@ -390,25 +390,31 @@
390	}
391	manifest_disable_event_triggers();
392	rebuild_update_schema();
393	blob_init(&sql, 0, 0);
394	db_unprotect(PROTECT_ALL);
395	#ifndef SQLITE_PREPARE_DONT_LOG
396	g.dbIgnoreErrors++;
397	#endif
398	db_prepare(&q,
399	"SELECT name FROM pragma_table_list /scan/"
400	" WHERE schema='repository' AND type IN ('table','virtual')"
401	" AND name NOT IN ('admin_log', 'blob','delta','rcvfrom','user','alias',"
402	"'config','shun','private','reportfmt',"
403	"'concealed','accesslog','modreq',"
404	"'purgeevent','purgeitem','unversioned',"
405	"'subscriber','pending_alert','chat')"
406	" AND name NOT GLOB 'sqlite_*'"
407	" AND name NOT GLOB 'fx_*';"
408	);
409	while( db_step(&q)==SQLITE_ROW ){
410	blob_appendf(&sql, "DROP TABLE IF EXISTS \"%w\";\n", db_column_text(&q,0));
411	}
412	db_finalize(&q);
413	#ifndef SQLITE_PREPARE_DONT_LOG
414	g.dbIgnoreErrors--;
415	#endif
416	db_multi_exec("%s", blob_str(&sql)/safe-for-%s/);
417	blob_reset(&sql);
418	db_multi_exec("%s", zRepositorySchema2/safe-for-%s/);
419	ticket_create_table(0);
420	shun_artifacts();
421

M src/rss.c

+4 -5

		--- src/rss.c
		+++ src/rss.c
		@@ -39,11 +39,10 @@
39	39	**
40	40	** In addition, name=FILENAME filters for a specific file. This may be
41	41	** combined with one of the other filters (useful for looking at a specific
42	42	** branch).
43	43	*/
44		-
45	44	void page_timeline_rss(void){
46	45	Stmt q;
47	46	int nLine=0;
48	47	char zPubDate, zProjectName, zProjectDescr, zFreeProjectName=0;
49	48	Blob bSQL;
		@@ -79,13 +78,13 @@
79	78	blob_zero(&bSQL);
80	79	blob_append_sql( &bSQL, "%s", zSQL1/safe-for-%s/ );
81	80
82	81	if( zType[0]!='a' ){
83	82	if( zType[0]=='c' && !g.perm.Read ) zType = "x";
84		- if( zType[0]=='w' && !g.perm.RdWiki ) zType = "x";
85		- if( zType[0]=='t' && !g.perm.RdTkt ) zType = "x";
86		- if( zType[0]=='f' && !g.perm.RdForum ) zType = "x";
	83	+ else if( (zType[0]=='w' \|\| zType[0]=='e') && !g.perm.RdWiki ) zType = "x";
	84	+ else if( zType[0]=='t' && !g.perm.RdTkt ) zType = "x";
	85	+ else if( zType[0]=='f' && !g.perm.RdForum ) zType = "x";
87	86	blob_append_sql(&bSQL, " AND event.type=%Q", zType);
88	87	}else{
89	88	blob_append_sql(&bSQL, " AND event.type in (");
90	89	if( g.perm.Read ){
91	90	blob_append_sql(&bSQL, "'ci',");
		@@ -92,11 +91,11 @@
92	91	}
93	92	if( g.perm.RdTkt ){
94	93	blob_append_sql(&bSQL, "'t',");
95	94	}
96	95	if( g.perm.RdWiki ){
97		- blob_append_sql(&bSQL, "'w',");
	96	+ blob_append_sql(&bSQL, "'w','e',");
98	97	}
99	98	if( g.perm.RdForum ){
100	99	blob_append_sql(&bSQL, "'f',");
101	100	}
102	101	blob_append_sql(&bSQL, "'x')");
103	102

	--- src/rss.c
	+++ src/rss.c
	@@ -39,11 +39,10 @@
39	**
40	** In addition, name=FILENAME filters for a specific file. This may be
41	** combined with one of the other filters (useful for looking at a specific
42	** branch).
43	*/
44
45	void page_timeline_rss(void){
46	Stmt q;
47	int nLine=0;
48	char zPubDate, zProjectName, zProjectDescr, zFreeProjectName=0;
49	Blob bSQL;
	@@ -79,13 +78,13 @@
79	blob_zero(&bSQL);
80	blob_append_sql( &bSQL, "%s", zSQL1/safe-for-%s/ );
81
82	if( zType[0]!='a' ){
83	if( zType[0]=='c' && !g.perm.Read ) zType = "x";
84	if( zType[0]=='w' && !g.perm.RdWiki ) zType = "x";
85	if( zType[0]=='t' && !g.perm.RdTkt ) zType = "x";
86	if( zType[0]=='f' && !g.perm.RdForum ) zType = "x";
87	blob_append_sql(&bSQL, " AND event.type=%Q", zType);
88	}else{
89	blob_append_sql(&bSQL, " AND event.type in (");
90	if( g.perm.Read ){
91	blob_append_sql(&bSQL, "'ci',");
	@@ -92,11 +91,11 @@
92	}
93	if( g.perm.RdTkt ){
94	blob_append_sql(&bSQL, "'t',");
95	}
96	if( g.perm.RdWiki ){
97	blob_append_sql(&bSQL, "'w',");
98	}
99	if( g.perm.RdForum ){
100	blob_append_sql(&bSQL, "'f',");
101	}
102	blob_append_sql(&bSQL, "'x')");
103

	--- src/rss.c
	+++ src/rss.c
	@@ -39,11 +39,10 @@
39	**
40	** In addition, name=FILENAME filters for a specific file. This may be
41	** combined with one of the other filters (useful for looking at a specific
42	** branch).
43	*/

44	void page_timeline_rss(void){
45	Stmt q;
46	int nLine=0;
47	char zPubDate, zProjectName, zProjectDescr, zFreeProjectName=0;
48	Blob bSQL;
	@@ -79,13 +78,13 @@
78	blob_zero(&bSQL);
79	blob_append_sql( &bSQL, "%s", zSQL1/safe-for-%s/ );
80
81	if( zType[0]!='a' ){
82	if( zType[0]=='c' && !g.perm.Read ) zType = "x";
83	else if( (zType[0]=='w' \|\| zType[0]=='e') && !g.perm.RdWiki ) zType = "x";
84	else if( zType[0]=='t' && !g.perm.RdTkt ) zType = "x";
85	else if( zType[0]=='f' && !g.perm.RdForum ) zType = "x";
86	blob_append_sql(&bSQL, " AND event.type=%Q", zType);
87	}else{
88	blob_append_sql(&bSQL, " AND event.type in (");
89	if( g.perm.Read ){
90	blob_append_sql(&bSQL, "'ci',");
	@@ -92,11 +91,11 @@
91	}
92	if( g.perm.RdTkt ){
93	blob_append_sql(&bSQL, "'t',");
94	}
95	if( g.perm.RdWiki ){
96	blob_append_sql(&bSQL, "'w','e',");
97	}
98	if( g.perm.RdForum ){
99	blob_append_sql(&bSQL, "'f',");
100	}
101	blob_append_sql(&bSQL, "'x')");
102

M src/schema.c

+7 -7

		--- src/schema.c
		+++ src/schema.c
		@@ -28,11 +28,11 @@
28	28	@ -- ~/.fossil file and that stores information about the users setup.
29	29	@ --
30	30	@ CREATE TABLE global_config(
31	31	@ name TEXT PRIMARY KEY,
32	32	@ value TEXT
33		-@ );
	33	+@ ) WITHOUT ROWID;
34	34	@
35	35	@ -- Identifier for this file type.
36	36	@ -- The integer is the same as 'FSLG'.
37	37	@ PRAGMA application_id=252006675;
38	38	;
		@@ -138,11 +138,11 @@
138	138	@ CREATE TABLE config(
139	139	@ name TEXT PRIMARY KEY NOT NULL, -- Primary name of the entry
140	140	@ value CLOB, -- Content of the named parameter
141	141	@ mtime DATE, -- last modified. seconds since 1970
142	142	@ CHECK( typeof(name)='text' AND length(name)>=1 )
143		-@ );
	143	+@ ) WITHOUT ROWID;
144	144	@
145	145	@ -- Artifacts that should not be processed are identified in the
146	146	@ -- "shun" table. Artifacts that are control-file forgeries or
147	147	@ -- spam or artifacts whose contents violate administrative policy
148	148	@ -- can be shunned in order to prevent them from contaminating
		@@ -151,14 +151,14 @@
151	151	@ -- Shunned artifacts do not exist in the blob table. Hence they
152	152	@ -- have not artifact ID (rid) and we thus must store their full
153	153	@ -- UUID.
154	154	@ --
155	155	@ CREATE TABLE shun(
156		-@ uuid UNIQUE, -- UUID of artifact to be shunned. Canonical form
	156	+@ uuid TEXT PRIMARY KEY,-- UUID of artifact to be shunned. Canonical form
157	157	@ mtime DATE, -- When added. seconds since 1970
158	158	@ scom TEXT -- Optional text explaining why the shun occurred
159		-@ );
	159	+@ ) WITHOUT ROWID;
160	160	@
161	161	@ -- Artifacts that should not be pushed are stored in the "private"
162	162	@ -- table. Private artifacts are omitted from the "unclustered" and
163	163	@ -- "unsent" tables.
164	164	@ --
		@@ -193,11 +193,11 @@
193	193	@ --
194	194	@ CREATE TABLE concealed(
195	195	@ hash TEXT PRIMARY KEY, -- The SHA1 hash of content
196	196	@ mtime DATE, -- Time created. Seconds since 1970
197	197	@ content TEXT -- Content intended to be concealed
198		-@ );
	198	+@ ) WITHOUT ROWID;
199	199	@
200	200	@ -- The application ID helps the unix "file" command to identify the
201	201	@ -- database as a fossil repository.
202	202	@ PRAGMA application_id=252006673;
203	203	;
		@@ -528,11 +528,11 @@
528	528	** The schema for the local FOSSIL database file found at the root
529	529	** of every check-out. This database contains the complete state of
530	530	** the check-out. See also the addendum in zLocalSchemaVmerge[].
531	531	*/
532	532	const char zLocalSchema[] =
533		-@ -- The VVAR table holds miscellanous information about the local database
	533	+@ -- The VVAR table holds miscellanous information about the local checkout
534	534	@ -- in the form of name-value pairs. This is similar to the VAR table
535	535	@ -- table in the repository except that this table holds information that
536	536	@ -- is specific to the local check-out.
537	537	@ --
538	538	@ -- Important Variables:
		@@ -542,11 +542,11 @@
542	542	@ --
543	543	@ CREATE TABLE vvar(
544	544	@ name TEXT PRIMARY KEY NOT NULL, -- Primary name of the entry
545	545	@ value CLOB, -- Content of the named parameter
546	546	@ CHECK( typeof(name)='text' AND length(name)>=1 )
547		-@ );
	547	+@ ) WITHOUT ROWID;
548	548	@
549	549	@ -- Each entry in the vfile table represents a single file in the
550	550	@ -- current check-out.
551	551	@ --
552	552	@ -- The file.rid field is 0 for files or folders that have been
553	553

	--- src/schema.c
	+++ src/schema.c
	@@ -28,11 +28,11 @@
28	@ -- ~/.fossil file and that stores information about the users setup.
29	@ --
30	@ CREATE TABLE global_config(
31	@ name TEXT PRIMARY KEY,
32	@ value TEXT
33	@ );
34	@
35	@ -- Identifier for this file type.
36	@ -- The integer is the same as 'FSLG'.
37	@ PRAGMA application_id=252006675;
38	;
	@@ -138,11 +138,11 @@
138	@ CREATE TABLE config(
139	@ name TEXT PRIMARY KEY NOT NULL, -- Primary name of the entry
140	@ value CLOB, -- Content of the named parameter
141	@ mtime DATE, -- last modified. seconds since 1970
142	@ CHECK( typeof(name)='text' AND length(name)>=1 )
143	@ );
144	@
145	@ -- Artifacts that should not be processed are identified in the
146	@ -- "shun" table. Artifacts that are control-file forgeries or
147	@ -- spam or artifacts whose contents violate administrative policy
148	@ -- can be shunned in order to prevent them from contaminating
	@@ -151,14 +151,14 @@
151	@ -- Shunned artifacts do not exist in the blob table. Hence they
152	@ -- have not artifact ID (rid) and we thus must store their full
153	@ -- UUID.
154	@ --
155	@ CREATE TABLE shun(
156	@ uuid UNIQUE, -- UUID of artifact to be shunned. Canonical form
157	@ mtime DATE, -- When added. seconds since 1970
158	@ scom TEXT -- Optional text explaining why the shun occurred
159	@ );
160	@
161	@ -- Artifacts that should not be pushed are stored in the "private"
162	@ -- table. Private artifacts are omitted from the "unclustered" and
163	@ -- "unsent" tables.
164	@ --
	@@ -193,11 +193,11 @@
193	@ --
194	@ CREATE TABLE concealed(
195	@ hash TEXT PRIMARY KEY, -- The SHA1 hash of content
196	@ mtime DATE, -- Time created. Seconds since 1970
197	@ content TEXT -- Content intended to be concealed
198	@ );
199	@
200	@ -- The application ID helps the unix "file" command to identify the
201	@ -- database as a fossil repository.
202	@ PRAGMA application_id=252006673;
203	;
	@@ -528,11 +528,11 @@
528	** The schema for the local FOSSIL database file found at the root
529	** of every check-out. This database contains the complete state of
530	** the check-out. See also the addendum in zLocalSchemaVmerge[].
531	*/
532	const char zLocalSchema[] =
533	@ -- The VVAR table holds miscellanous information about the local database
534	@ -- in the form of name-value pairs. This is similar to the VAR table
535	@ -- table in the repository except that this table holds information that
536	@ -- is specific to the local check-out.
537	@ --
538	@ -- Important Variables:
	@@ -542,11 +542,11 @@
542	@ --
543	@ CREATE TABLE vvar(
544	@ name TEXT PRIMARY KEY NOT NULL, -- Primary name of the entry
545	@ value CLOB, -- Content of the named parameter
546	@ CHECK( typeof(name)='text' AND length(name)>=1 )
547	@ );
548	@
549	@ -- Each entry in the vfile table represents a single file in the
550	@ -- current check-out.
551	@ --
552	@ -- The file.rid field is 0 for files or folders that have been
553

	--- src/schema.c
	+++ src/schema.c
	@@ -28,11 +28,11 @@
28	@ -- ~/.fossil file and that stores information about the users setup.
29	@ --
30	@ CREATE TABLE global_config(
31	@ name TEXT PRIMARY KEY,
32	@ value TEXT
33	@ ) WITHOUT ROWID;
34	@
35	@ -- Identifier for this file type.
36	@ -- The integer is the same as 'FSLG'.
37	@ PRAGMA application_id=252006675;
38	;
	@@ -138,11 +138,11 @@
138	@ CREATE TABLE config(
139	@ name TEXT PRIMARY KEY NOT NULL, -- Primary name of the entry
140	@ value CLOB, -- Content of the named parameter
141	@ mtime DATE, -- last modified. seconds since 1970
142	@ CHECK( typeof(name)='text' AND length(name)>=1 )
143	@ ) WITHOUT ROWID;
144	@
145	@ -- Artifacts that should not be processed are identified in the
146	@ -- "shun" table. Artifacts that are control-file forgeries or
147	@ -- spam or artifacts whose contents violate administrative policy
148	@ -- can be shunned in order to prevent them from contaminating
	@@ -151,14 +151,14 @@
151	@ -- Shunned artifacts do not exist in the blob table. Hence they
152	@ -- have not artifact ID (rid) and we thus must store their full
153	@ -- UUID.
154	@ --
155	@ CREATE TABLE shun(
156	@ uuid TEXT PRIMARY KEY,-- UUID of artifact to be shunned. Canonical form
157	@ mtime DATE, -- When added. seconds since 1970
158	@ scom TEXT -- Optional text explaining why the shun occurred
159	@ ) WITHOUT ROWID;
160	@
161	@ -- Artifacts that should not be pushed are stored in the "private"
162	@ -- table. Private artifacts are omitted from the "unclustered" and
163	@ -- "unsent" tables.
164	@ --
	@@ -193,11 +193,11 @@
193	@ --
194	@ CREATE TABLE concealed(
195	@ hash TEXT PRIMARY KEY, -- The SHA1 hash of content
196	@ mtime DATE, -- Time created. Seconds since 1970
197	@ content TEXT -- Content intended to be concealed
198	@ ) WITHOUT ROWID;
199	@
200	@ -- The application ID helps the unix "file" command to identify the
201	@ -- database as a fossil repository.
202	@ PRAGMA application_id=252006673;
203	;
	@@ -528,11 +528,11 @@
528	** The schema for the local FOSSIL database file found at the root
529	** of every check-out. This database contains the complete state of
530	** the check-out. See also the addendum in zLocalSchemaVmerge[].
531	*/
532	const char zLocalSchema[] =
533	@ -- The VVAR table holds miscellanous information about the local checkout
534	@ -- in the form of name-value pairs. This is similar to the VAR table
535	@ -- table in the repository except that this table holds information that
536	@ -- is specific to the local check-out.
537	@ --
538	@ -- Important Variables:
	@@ -542,11 +542,11 @@
542	@ --
543	@ CREATE TABLE vvar(
544	@ name TEXT PRIMARY KEY NOT NULL, -- Primary name of the entry
545	@ value CLOB, -- Content of the named parameter
546	@ CHECK( typeof(name)='text' AND length(name)>=1 )
547	@ ) WITHOUT ROWID;
548	@
549	@ -- Each entry in the vfile table represents a single file in the
550	@ -- current check-out.
551	@ --
552	@ -- The file.rid field is 0 for files or folders that have been
553

M src/search.c

+83 -64

		--- src/search.c
		+++ src/search.c
		@@ -794,31 +794,41 @@
794	794	SRCHFLG_STATIC\|SRCHFLG_HTML);
795	795	if( (srchFlags & SRCH_DOC)!=0 ){
796	796	char *zDocGlob = db_get("doc-glob","");
797	797	char *zDocBr = db_get("doc-branch","trunk");
798	798	if( zDocGlob && zDocGlob[0] && zDocBr && zDocBr[0] ){
	799	+ Glob * pGlob = glob_create(zDocBr)
	800	+ /* We're misusing a Glob as a list of comma-/space-delimited
	801	+ ** tokens. We're not actually doing glob matches here. */;
	802	+ int i;
799	803	db_multi_exec(
800	804	"CREATE VIRTUAL TABLE IF NOT EXISTS temp.foci USING files_of_checkin;"
801	805	);
802		- db_multi_exec(
803		- "INSERT INTO x(label,url,score,id,date,snip)"
804		- " SELECT printf('Document: %%s',title('d',blob.rid,foci.filename)),"
805		- " printf('/doc/%T/%%s',foci.filename),"
806		- " search_score(),"
807		- " 'd'\|\|blob.rid,"
808		- " (SELECT datetime(event.mtime) FROM event"
809		- " WHERE objid=symbolic_name_to_rid('trunk')),"
810		- " search_snippet()"
811		- " FROM foci CROSS JOIN blob"
812		- " WHERE checkinID=symbolic_name_to_rid('trunk')"
813		- " AND blob.uuid=foci.uuid"
814		- " AND search_match(title('d',blob.rid,foci.filename),"
815		- " body('d',blob.rid,foci.filename))"
816		- " AND %z",
817		- zDocBr, glob_expr("foci.filename", zDocGlob)
818		- );
819		- }
	806	+ for( i = 0; i < pGlob->nPattern; ++i ){
	807	+ const char * zBranch = pGlob->azPattern[i];
	808	+ db_multi_exec(
	809	+ "INSERT INTO x(label,url,score,id,date,snip)"
	810	+ " SELECT printf('Document: %%s',title('d',blob.rid,foci.filename)),"
	811	+ " printf('/doc/%T/%%s',foci.filename),"
	812	+ " search_score(),"
	813	+ " 'd'\|\|blob.rid,"
	814	+ " (SELECT datetime(event.mtime) FROM event"
	815	+ " WHERE objid=symbolic_name_to_rid(%Q)),"
	816	+ " search_snippet()"
	817	+ " FROM foci CROSS JOIN blob"
	818	+ " WHERE checkinID=symbolic_name_to_rid(%Q)"
	819	+ " AND blob.uuid=foci.uuid"
	820	+ " AND search_match(title('d',blob.rid,foci.filename),"
	821	+ " body('d',blob.rid,foci.filename))"
	822	+ " AND %z",
	823	+ zBranch, zBranch, zBranch, glob_expr("foci.filename", zDocGlob)
	824	+ );
	825	+ }
	826	+ glob_free(pGlob);
	827	+ }
	828	+ fossil_free(zDocGlob);
	829	+ fossil_free(zDocBr);
820	830	}
821	831	if( (srchFlags & SRCH_WIKI)!=0 ){
822	832	db_multi_exec(
823	833	"WITH wiki(name,rid,mtime) AS ("
824	834	" SELECT substr(tagname,6), tagxref.rid, max(tagxref.mtime)"
		@@ -1878,59 +1888,68 @@
1878	1888	** and if the latest check-in on doc-br is in the unindexed set of
1879	1889	** check-ins, then update all 'd' entries in FTSDOCS that have
1880	1890	** changed.
1881	1891	*/
1882	1892	static void search_update_doc_index(void){
1883		- const char *zDocBr = db_get("doc-branch","trunk");
1884		- int ckid = zDocBr ? symbolic_name_to_rid(zDocBr,"ci") : 0;
1885		- double rTime;
1886		- if( ckid==0 ) return;
1887		- if( !db_exists("SELECT 1 FROM ftsdocs WHERE type='c' AND rid=%d"
1888		- " AND NOT idxed", ckid) ) return;
1889		-
1890		- /* If we get this far, it means that changes to 'd' entries are
1891		- ** required. */
1892		- rTime = db_double(0.0, "SELECT mtime FROM event WHERE objid=%d", ckid);
	1893	+ const char *zDocBranches = db_get("doc-branch","trunk");
	1894	+ int i;
	1895	+ Glob * pGlob = glob_create(zDocBranches)
	1896	+ /* We're misusing a Glob as a list of comma-/space-delimited
	1897	+ ** tokens. We're not actually doing glob matches here. */;
	1898	+ if( !pGlob ) return;
1893	1899	db_multi_exec(
1894	1900	"CREATE TEMP TABLE current_docs(rid INTEGER PRIMARY KEY, name);"
1895	1901	"CREATE VIRTUAL TABLE IF NOT EXISTS temp.foci USING files_of_checkin;"
1896		- "INSERT OR IGNORE INTO current_docs(rid, name)"
1897		- " SELECT blob.rid, foci.filename FROM foci, blob"
1898		- " WHERE foci.checkinID=%d AND blob.uuid=foci.uuid"
1899		- " AND %z",
1900		- ckid, glob_expr("foci.filename", db_get("doc-glob",""))
1901		- );
1902		- db_multi_exec(
1903		- "DELETE FROM ftsidx WHERE rowid IN"
1904		- " (SELECT rowid FROM ftsdocs WHERE type='d'"
1905		- " AND rid NOT IN (SELECT rid FROM current_docs))"
1906		- );
1907		- db_multi_exec(
1908		- "DELETE FROM ftsdocs WHERE type='d'"
1909		- " AND rid NOT IN (SELECT rid FROM current_docs)"
1910		- );
1911		- db_multi_exec(
1912		- "INSERT OR IGNORE INTO ftsdocs(type,rid,name,idxed,label,bx,url,mtime)"
1913		- " SELECT 'd', rid, name, 0,"
1914		- " title('d',rid,name),"
1915		- " body('d',rid,name),"
1916		- " printf('/doc/%T/%%s',urlencode(name)),"
1917		- " %.17g"
1918		- " FROM current_docs",
1919		- zDocBr, rTime
1920		- );
1921		- db_multi_exec(
1922		- "INSERT INTO ftsidx(rowid,title,body)"
1923		- " SELECT rowid, label, bx FROM ftsdocs WHERE type='d' AND NOT idxed"
1924		- );
1925		- db_multi_exec(
1926		- "UPDATE ftsdocs SET"
1927		- " idxed=1,"
1928		- " bx=NULL,"
1929		- " label='Document: '\|\|label"
1930		- " WHERE type='d' AND NOT idxed"
1931		- );
	1902	+ );
	1903	+ for( i = 0; i < pGlob->nPattern; ++i ){
	1904	+ const char *zDocBr = pGlob->azPattern[i];
	1905	+ int ckid = symbolic_name_to_rid(zDocBr,"ci");
	1906	+ double rTime;
	1907	+ if( !db_exists("SELECT 1 FROM ftsdocs WHERE type='c' AND rid=%d"
	1908	+ " AND NOT idxed", ckid) ) continue;
	1909	+ /* If we get this far, it means that changes to 'd' entries are
	1910	+ ** required. */
	1911	+ rTime = db_double(0.0, "SELECT mtime FROM event WHERE objid=%d", ckid);
	1912	+ db_multi_exec(
	1913	+ "INSERT OR IGNORE INTO current_docs(rid, name)"
	1914	+ " SELECT blob.rid, foci.filename FROM foci, blob"
	1915	+ " WHERE foci.checkinID=%d AND blob.uuid=foci.uuid"
	1916	+ " AND %z",
	1917	+ ckid, glob_expr("foci.filename", db_get("doc-glob",""))
	1918	+ );
	1919	+ db_multi_exec(
	1920	+ "DELETE FROM ftsidx WHERE rowid IN"
	1921	+ " (SELECT rowid FROM ftsdocs WHERE type='d'"
	1922	+ " AND rid NOT IN (SELECT rid FROM current_docs))"
	1923	+ );
	1924	+ db_multi_exec(
	1925	+ "DELETE FROM ftsdocs WHERE type='d'"
	1926	+ " AND rid NOT IN (SELECT rid FROM current_docs)"
	1927	+ );
	1928	+ db_multi_exec(
	1929	+ "INSERT OR IGNORE INTO ftsdocs(type,rid,name,idxed,label,bx,url,mtime)"
	1930	+ " SELECT 'd', rid, name, 0,"
	1931	+ " title('d',rid,name),"
	1932	+ " body('d',rid,name),"
	1933	+ " printf('/doc/%T/%%s',urlencode(name)),"
	1934	+ " %.17g"
	1935	+ " FROM current_docs",
	1936	+ zDocBr, rTime
	1937	+ );
	1938	+ db_multi_exec(
	1939	+ "INSERT INTO ftsidx(rowid,title,body)"
	1940	+ " SELECT rowid, label, bx FROM ftsdocs WHERE type='d' AND NOT idxed"
	1941	+ );
	1942	+ db_multi_exec(
	1943	+ "UPDATE ftsdocs SET"
	1944	+ " idxed=1,"
	1945	+ " bx=NULL,"
	1946	+ " label='Document: '\|\|label"
	1947	+ " WHERE type='d' AND NOT idxed"
	1948	+ );
	1949	+ }
	1950	+ glob_free(pGlob);
1932	1951	}
1933	1952
1934	1953	/*
1935	1954	** Deal with all of the unindexed 'c' terms in FTSDOCS
1936	1955	*/
1937	1956

	--- src/search.c
	+++ src/search.c
	@@ -794,31 +794,41 @@
794	SRCHFLG_STATIC\|SRCHFLG_HTML);
795	if( (srchFlags & SRCH_DOC)!=0 ){
796	char *zDocGlob = db_get("doc-glob","");
797	char *zDocBr = db_get("doc-branch","trunk");
798	if( zDocGlob && zDocGlob[0] && zDocBr && zDocBr[0] ){




799	db_multi_exec(
800	"CREATE VIRTUAL TABLE IF NOT EXISTS temp.foci USING files_of_checkin;"
801	);
802	db_multi_exec(
803	"INSERT INTO x(label,url,score,id,date,snip)"
804	" SELECT printf('Document: %%s',title('d',blob.rid,foci.filename)),"
805	" printf('/doc/%T/%%s',foci.filename),"
806	" search_score(),"
807	" 'd'\|\|blob.rid,"
808	" (SELECT datetime(event.mtime) FROM event"
809	" WHERE objid=symbolic_name_to_rid('trunk')),"
810	" search_snippet()"
811	" FROM foci CROSS JOIN blob"
812	" WHERE checkinID=symbolic_name_to_rid('trunk')"
813	" AND blob.uuid=foci.uuid"
814	" AND search_match(title('d',blob.rid,foci.filename),"
815	" body('d',blob.rid,foci.filename))"
816	" AND %z",
817	zDocBr, glob_expr("foci.filename", zDocGlob)
818	);
819	}






820	}
821	if( (srchFlags & SRCH_WIKI)!=0 ){
822	db_multi_exec(
823	"WITH wiki(name,rid,mtime) AS ("
824	" SELECT substr(tagname,6), tagxref.rid, max(tagxref.mtime)"
	@@ -1878,59 +1888,68 @@
1878	** and if the latest check-in on doc-br is in the unindexed set of
1879	** check-ins, then update all 'd' entries in FTSDOCS that have
1880	** changed.
1881	*/
1882	static void search_update_doc_index(void){
1883	const char *zDocBr = db_get("doc-branch","trunk");
1884	int ckid = zDocBr ? symbolic_name_to_rid(zDocBr,"ci") : 0;
1885	double rTime;
1886	if( ckid==0 ) return;
1887	if( !db_exists("SELECT 1 FROM ftsdocs WHERE type='c' AND rid=%d"
1888	" AND NOT idxed", ckid) ) return;
1889
1890	/* If we get this far, it means that changes to 'd' entries are
1891	** required. */
1892	rTime = db_double(0.0, "SELECT mtime FROM event WHERE objid=%d", ckid);
1893	db_multi_exec(
1894	"CREATE TEMP TABLE current_docs(rid INTEGER PRIMARY KEY, name);"
1895	"CREATE VIRTUAL TABLE IF NOT EXISTS temp.foci USING files_of_checkin;"
1896	"INSERT OR IGNORE INTO current_docs(rid, name)"
1897	" SELECT blob.rid, foci.filename FROM foci, blob"
1898	" WHERE foci.checkinID=%d AND blob.uuid=foci.uuid"
1899	" AND %z",
1900	ckid, glob_expr("foci.filename", db_get("doc-glob",""))
1901	);
1902	db_multi_exec(
1903	"DELETE FROM ftsidx WHERE rowid IN"
1904	" (SELECT rowid FROM ftsdocs WHERE type='d'"
1905	" AND rid NOT IN (SELECT rid FROM current_docs))"
1906	);
1907	db_multi_exec(
1908	"DELETE FROM ftsdocs WHERE type='d'"
1909	" AND rid NOT IN (SELECT rid FROM current_docs)"
1910	);
1911	db_multi_exec(
1912	"INSERT OR IGNORE INTO ftsdocs(type,rid,name,idxed,label,bx,url,mtime)"
1913	" SELECT 'd', rid, name, 0,"
1914	" title('d',rid,name),"
1915	" body('d',rid,name),"
1916	" printf('/doc/%T/%%s',urlencode(name)),"
1917	" %.17g"
1918	" FROM current_docs",
1919	zDocBr, rTime
1920	);
1921	db_multi_exec(
1922	"INSERT INTO ftsidx(rowid,title,body)"
1923	" SELECT rowid, label, bx FROM ftsdocs WHERE type='d' AND NOT idxed"
1924	);
1925	db_multi_exec(
1926	"UPDATE ftsdocs SET"
1927	" idxed=1,"
1928	" bx=NULL,"
1929	" label='Document: '\|\|label"
1930	" WHERE type='d' AND NOT idxed"
1931	);













1932	}
1933
1934	/*
1935	** Deal with all of the unindexed 'c' terms in FTSDOCS
1936	*/
1937

	--- src/search.c
	+++ src/search.c
	@@ -794,31 +794,41 @@
794	SRCHFLG_STATIC\|SRCHFLG_HTML);
795	if( (srchFlags & SRCH_DOC)!=0 ){
796	char *zDocGlob = db_get("doc-glob","");
797	char *zDocBr = db_get("doc-branch","trunk");
798	if( zDocGlob && zDocGlob[0] && zDocBr && zDocBr[0] ){
799	Glob * pGlob = glob_create(zDocBr)
800	/* We're misusing a Glob as a list of comma-/space-delimited
801	** tokens. We're not actually doing glob matches here. */;
802	int i;
803	db_multi_exec(
804	"CREATE VIRTUAL TABLE IF NOT EXISTS temp.foci USING files_of_checkin;"
805	);
806	for( i = 0; i < pGlob->nPattern; ++i ){
807	const char * zBranch = pGlob->azPattern[i];
808	db_multi_exec(
809	"INSERT INTO x(label,url,score,id,date,snip)"
810	" SELECT printf('Document: %%s',title('d',blob.rid,foci.filename)),"
811	" printf('/doc/%T/%%s',foci.filename),"
812	" search_score(),"
813	" 'd'\|\|blob.rid,"
814	" (SELECT datetime(event.mtime) FROM event"
815	" WHERE objid=symbolic_name_to_rid(%Q)),"
816	" search_snippet()"
817	" FROM foci CROSS JOIN blob"
818	" WHERE checkinID=symbolic_name_to_rid(%Q)"
819	" AND blob.uuid=foci.uuid"
820	" AND search_match(title('d',blob.rid,foci.filename),"
821	" body('d',blob.rid,foci.filename))"
822	" AND %z",
823	zBranch, zBranch, zBranch, glob_expr("foci.filename", zDocGlob)
824	);
825	}
826	glob_free(pGlob);
827	}
828	fossil_free(zDocGlob);
829	fossil_free(zDocBr);
830	}
831	if( (srchFlags & SRCH_WIKI)!=0 ){
832	db_multi_exec(
833	"WITH wiki(name,rid,mtime) AS ("
834	" SELECT substr(tagname,6), tagxref.rid, max(tagxref.mtime)"
	@@ -1878,59 +1888,68 @@
1888	** and if the latest check-in on doc-br is in the unindexed set of
1889	** check-ins, then update all 'd' entries in FTSDOCS that have
1890	** changed.
1891	*/
1892	static void search_update_doc_index(void){
1893	const char *zDocBranches = db_get("doc-branch","trunk");
1894	int i;
1895	Glob * pGlob = glob_create(zDocBranches)
1896	/* We're misusing a Glob as a list of comma-/space-delimited
1897	** tokens. We're not actually doing glob matches here. */;
1898	if( !pGlob ) return;




1899	db_multi_exec(
1900	"CREATE TEMP TABLE current_docs(rid INTEGER PRIMARY KEY, name);"
1901	"CREATE VIRTUAL TABLE IF NOT EXISTS temp.foci USING files_of_checkin;"
1902	);
1903	for( i = 0; i < pGlob->nPattern; ++i ){
1904	const char *zDocBr = pGlob->azPattern[i];
1905	int ckid = symbolic_name_to_rid(zDocBr,"ci");
1906	double rTime;
1907	if( !db_exists("SELECT 1 FROM ftsdocs WHERE type='c' AND rid=%d"
1908	" AND NOT idxed", ckid) ) continue;
1909	/* If we get this far, it means that changes to 'd' entries are
1910	** required. */
1911	rTime = db_double(0.0, "SELECT mtime FROM event WHERE objid=%d", ckid);
1912	db_multi_exec(
1913	"INSERT OR IGNORE INTO current_docs(rid, name)"
1914	" SELECT blob.rid, foci.filename FROM foci, blob"
1915	" WHERE foci.checkinID=%d AND blob.uuid=foci.uuid"
1916	" AND %z",
1917	ckid, glob_expr("foci.filename", db_get("doc-glob",""))
1918	);
1919	db_multi_exec(
1920	"DELETE FROM ftsidx WHERE rowid IN"
1921	" (SELECT rowid FROM ftsdocs WHERE type='d'"
1922	" AND rid NOT IN (SELECT rid FROM current_docs))"
1923	);
1924	db_multi_exec(
1925	"DELETE FROM ftsdocs WHERE type='d'"
1926	" AND rid NOT IN (SELECT rid FROM current_docs)"
1927	);
1928	db_multi_exec(
1929	"INSERT OR IGNORE INTO ftsdocs(type,rid,name,idxed,label,bx,url,mtime)"
1930	" SELECT 'd', rid, name, 0,"
1931	" title('d',rid,name),"
1932	" body('d',rid,name),"
1933	" printf('/doc/%T/%%s',urlencode(name)),"
1934	" %.17g"
1935	" FROM current_docs",
1936	zDocBr, rTime
1937	);
1938	db_multi_exec(
1939	"INSERT INTO ftsidx(rowid,title,body)"
1940	" SELECT rowid, label, bx FROM ftsdocs WHERE type='d' AND NOT idxed"
1941	);
1942	db_multi_exec(
1943	"UPDATE ftsdocs SET"
1944	" idxed=1,"
1945	" bx=NULL,"
1946	" label='Document: '\|\|label"
1947	" WHERE type='d' AND NOT idxed"
1948	);
1949	}
1950	glob_free(pGlob);
1951	}
1952
1953	/*
1954	** Deal with all of the unindexed 'c' terms in FTSDOCS
1955	*/
1956

M src/setup.c

+16 -4

		--- src/setup.c
		+++ src/setup.c
		@@ -501,13 +501,24 @@
501	501	@ behavior or to find an SQL injection opportunity or similar. This can
502	502	@ waste hours of CPU time and gigabytes of bandwidth on the server. A
503	503	@ suggested value for this setting is:
504	504	@ "<tt>timeline,*diff,vpatch,annotate,blame,praise,dir,tree</tt>".
505	505	@ (Property: robot-restrict)
506		- @ <p>
	506	+ @ <br>
507	507	textarea_attribute("", 2, 80,
508	508	"robot-restrict", "rbrestrict", "", 0);
	509	+ @ <br> The following comma-separated GLOB pattern allows for exceptions
	510	+ @ in the maximum number of query parameters before a request is considered
	511	+ @ complex. If this GLOB pattern exists and is non-empty and if it
	512	+ @ matches against the pagename followed by "/" and the number of query
	513	+ @ parameters, then the request is allowed through. For example, the
	514	+ @ suggested pattern of "timeline/[012]" allows the /timeline page to
	515	+ @ pass with up to 2 query parameters besides "name".
	516	+ @ (Property: robot-restrict-qp)
	517	+ @ <br>
	518	+ textarea_attribute("", 2, 80,
	519	+ "robot-restrict-qp", "rbrestrictqp", "", 0);
509	520
510	521	@ <hr>
511	522	@ <p><input type="submit" name="submit" value="Apply Changes"></p>
512	523	@ </div></form>
513	524	db_end_transaction(0);
		@@ -2187,14 +2198,15 @@
2187	2198	@ <tr><td>*<td><td>Search all checked-in files</tr>
2188	2199	@ <tr><td><i>(blank)</i><td>
2189	2200	@ <td>Search nothing. (Disables document search).</tr>
2190	2201	@ </table>
2191	2202	@ <hr>
2192		- entry_attribute("Document Branch", 20, "doc-branch", "db", "trunk", 0);
	2203	+ entry_attribute("Document Branches", 20, "doc-branch", "db", "trunk", 0);
2193	2204	@ <p>When searching documents, use the versions of the files found at the
2194		- @ type of the "Document Branch" branch. Recommended value: "trunk".
2195		- @ Document search is disabled if blank.
	2205	+ @ type of the "Document Branches" branch. Recommended value: "trunk".
	2206	+ @ Document search is disabled if blank. It may be a list of branch names
	2207	+ @ separated by spaces and/or commas.
2196	2208	@ <hr>
2197	2209	onoff_attribute("Search Check-in Comments", "search-ci", "sc", 0, 0);
2198	2210	@ <br>
2199	2211	onoff_attribute("Search Documents", "search-doc", "sd", 0, 0);
2200	2212	@ <br>
2201	2213

	--- src/setup.c
	+++ src/setup.c
	@@ -501,13 +501,24 @@
501	@ behavior or to find an SQL injection opportunity or similar. This can
502	@ waste hours of CPU time and gigabytes of bandwidth on the server. A
503	@ suggested value for this setting is:
504	@ "<tt>timeline,*diff,vpatch,annotate,blame,praise,dir,tree</tt>".
505	@ (Property: robot-restrict)
506	@ <p>
507	textarea_attribute("", 2, 80,
508	"robot-restrict", "rbrestrict", "", 0);











509
510	@ <hr>
511	@ <p><input type="submit" name="submit" value="Apply Changes"></p>
512	@ </div></form>
513	db_end_transaction(0);
	@@ -2187,14 +2198,15 @@
2187	@ <tr><td>*<td><td>Search all checked-in files</tr>
2188	@ <tr><td><i>(blank)</i><td>
2189	@ <td>Search nothing. (Disables document search).</tr>
2190	@ </table>
2191	@ <hr>
2192	entry_attribute("Document Branch", 20, "doc-branch", "db", "trunk", 0);
2193	@ <p>When searching documents, use the versions of the files found at the
2194	@ type of the "Document Branch" branch. Recommended value: "trunk".
2195	@ Document search is disabled if blank.

2196	@ <hr>
2197	onoff_attribute("Search Check-in Comments", "search-ci", "sc", 0, 0);
2198	@ <br>
2199	onoff_attribute("Search Documents", "search-doc", "sd", 0, 0);
2200	@ <br>
2201

	--- src/setup.c
	+++ src/setup.c
	@@ -501,13 +501,24 @@
501	@ behavior or to find an SQL injection opportunity or similar. This can
502	@ waste hours of CPU time and gigabytes of bandwidth on the server. A
503	@ suggested value for this setting is:
504	@ "<tt>timeline,*diff,vpatch,annotate,blame,praise,dir,tree</tt>".
505	@ (Property: robot-restrict)
506	@ <br>
507	textarea_attribute("", 2, 80,
508	"robot-restrict", "rbrestrict", "", 0);
509	@ <br> The following comma-separated GLOB pattern allows for exceptions
510	@ in the maximum number of query parameters before a request is considered
511	@ complex. If this GLOB pattern exists and is non-empty and if it
512	@ matches against the pagename followed by "/" and the number of query
513	@ parameters, then the request is allowed through. For example, the
514	@ suggested pattern of "timeline/[012]" allows the /timeline page to
515	@ pass with up to 2 query parameters besides "name".
516	@ (Property: robot-restrict-qp)
517	@ <br>
518	textarea_attribute("", 2, 80,
519	"robot-restrict-qp", "rbrestrictqp", "", 0);
520
521	@ <hr>
522	@ <p><input type="submit" name="submit" value="Apply Changes"></p>
523	@ </div></form>
524	db_end_transaction(0);
	@@ -2187,14 +2198,15 @@
2198	@ <tr><td>*<td><td>Search all checked-in files</tr>
2199	@ <tr><td><i>(blank)</i><td>
2200	@ <td>Search nothing. (Disables document search).</tr>
2201	@ </table>
2202	@ <hr>
2203	entry_attribute("Document Branches", 20, "doc-branch", "db", "trunk", 0);
2204	@ <p>When searching documents, use the versions of the files found at the
2205	@ type of the "Document Branches" branch. Recommended value: "trunk".
2206	@ Document search is disabled if blank. It may be a list of branch names
2207	@ separated by spaces and/or commas.
2208	@ <hr>
2209	onoff_attribute("Search Check-in Comments", "search-ci", "sc", 0, 0);
2210	@ <br>
2211	onoff_attribute("Search Documents", "search-doc", "sd", 0, 0);
2212	@ <br>
2213

M src/setupuser.c

+7 -5

		--- src/setupuser.c
		+++ src/setupuser.c
		@@ -115,11 +115,11 @@
115	115	}
116	116	if( !bUnusedOnly ){
117	117	style_submenu_element("Unused", "setup_ulist?unused");
118	118	}
119	119	@ <table border=1 cellpadding=2 cellspacing=0 class='userTable sortable' \
120		- @ data-column-types='ktxTTKt' data-init-sort='2'>
	120	+ @ data-column-types='ktxKTKt' data-init-sort='4'>
121	121	@ <thead><tr>
122	122	@ <th>Login Name<th>Caps<th>Info<th>Date<th>Expire<th>Last Login\
123	123	@ <th>Alerts</tr></thead>
124	124	@ <tbody>
125	125	db_multi_exec(
		@@ -161,15 +161,16 @@
161	161	" lower(login) AS sortkey, "
162	162	" CASE WHEN info LIKE '%%expires 20%%'"
163	163	" THEN substr(info,instr(lower(info),'expires')+8,10)"
164	164	" END AS exp,"
165	165	"atime,"
166		- " subscriber.ssub, subscriber.subscriberId"
	166	+ " subscriber.ssub, subscriber.subscriberId,"
	167	+ " user.mtime AS sorttime"
167	168	" FROM user LEFT JOIN lastAccess ON login=uname"
168	169	" LEFT JOIN subscriber ON login=suname"
169	170	" WHERE login NOT IN ('anonymous','nobody','developer','reader') %s"
170		- " ORDER BY sortkey", zWith/safe-for-%s/
	171	+ " ORDER BY sorttime DESC", zWith/safe-for-%s/
171	172	);
172	173	rNow = db_double(0.0, "SELECT julianday('now');");
173	174	while( db_step(&s)==SQLITE_ROW ){
174	175	int uid = db_column_int(&s, 0);
175	176	const char *zLogin = db_column_text(&s, 1);
		@@ -180,10 +181,11 @@
180	181	const char *zExp = db_column_text(&s,6);
181	182	double rATime = db_column_double(&s,7);
182	183	char *zAge = 0;
183	184	const char *zSub;
184	185	int sid = db_column_int(&s,9);
	186	+ sqlite3_int64 sorttime = db_column_int64(&s, 10);
185	187	if( rATime>0.0 ){
186	188	zAge = human_readable_age(rNow - rATime);
187	189	}
188	190	if( bUbg ){
189	191	@ <tr style='background-color: %h(user_color(zLogin));'>
		@@ -192,11 +194,11 @@
192	194	}
193	195	@ <td data-sortkey='%h(zSortKey)'>\
194	196	@ <a href='setup_uedit?id=%d(uid)'>%h(zLogin)</a>
195	197	@ <td>%h(zCap)
196	198	@ <td>%h(zInfo)
197		- @ <td>%h(zDate?zDate:"")
	199	+ @ <td data-sortkey='%09llx(sorttime)'>%h(zDate?zDate:"")
198	200	@ <td>%h(zExp?zExp:"")
199	201	@ <td data-sortkey='%f(rATime)' style='white-space:nowrap'>%s(zAge?zAge:"")
200	202	if( db_column_type(&s,8)==SQLITE_NULL ){
201	203	@ <td>
202	204	}else if( (zSub = db_column_text(&s,8))==0 \|\| zSub[0]==0 ){
		@@ -1016,11 +1018,11 @@
1016	1018	style_header("User %h", db_column_text(&q,1));
1017	1019	@ <table class="label-value">
1018	1020	@ <tr><th>uid:</th><td>%d(db_column_int(&q,0))
1019	1021	@ (<a href="%R/setup_uedit?id=%d(db_column_int(&q,0))">edit</a>)</td></tr>
1020	1022	@ <tr><th>login:</th><td>%h(db_column_text(&q,1))</td></tr>
1021		- @ <tr><th>capabilities:</th><td>%h(db_column_text(&q,2))</th></tr>
	1023	+ @ <tr><th>capabilities:</th><td>%h(db_column_text(&q,2))</td></tr>
1022	1024	@ <tr><th valign="top">info:</th>
1023	1025	@ <td valign="top"><span style='white-space:pre-line;'>\
1024	1026	@ %h(db_column_text(&q,5))</span></td></tr>
1025	1027	@ <tr><th>user.mtime:</th><td>%h(db_column_text(&q,6))</td></tr>
1026	1028	if( db_column_type(&q,7)!=SQLITE_NULL ){
1027	1029

	--- src/setupuser.c
	+++ src/setupuser.c
	@@ -115,11 +115,11 @@
115	}
116	if( !bUnusedOnly ){
117	style_submenu_element("Unused", "setup_ulist?unused");
118	}
119	@ <table border=1 cellpadding=2 cellspacing=0 class='userTable sortable' \
120	@ data-column-types='ktxTTKt' data-init-sort='2'>
121	@ <thead><tr>
122	@ <th>Login Name<th>Caps<th>Info<th>Date<th>Expire<th>Last Login\
123	@ <th>Alerts</tr></thead>
124	@ <tbody>
125	db_multi_exec(
	@@ -161,15 +161,16 @@
161	" lower(login) AS sortkey, "
162	" CASE WHEN info LIKE '%%expires 20%%'"
163	" THEN substr(info,instr(lower(info),'expires')+8,10)"
164	" END AS exp,"
165	"atime,"
166	" subscriber.ssub, subscriber.subscriberId"

167	" FROM user LEFT JOIN lastAccess ON login=uname"
168	" LEFT JOIN subscriber ON login=suname"
169	" WHERE login NOT IN ('anonymous','nobody','developer','reader') %s"
170	" ORDER BY sortkey", zWith/safe-for-%s/
171	);
172	rNow = db_double(0.0, "SELECT julianday('now');");
173	while( db_step(&s)==SQLITE_ROW ){
174	int uid = db_column_int(&s, 0);
175	const char *zLogin = db_column_text(&s, 1);
	@@ -180,10 +181,11 @@
180	const char *zExp = db_column_text(&s,6);
181	double rATime = db_column_double(&s,7);
182	char *zAge = 0;
183	const char *zSub;
184	int sid = db_column_int(&s,9);

185	if( rATime>0.0 ){
186	zAge = human_readable_age(rNow - rATime);
187	}
188	if( bUbg ){
189	@ <tr style='background-color: %h(user_color(zLogin));'>
	@@ -192,11 +194,11 @@
192	}
193	@ <td data-sortkey='%h(zSortKey)'>\
194	@ <a href='setup_uedit?id=%d(uid)'>%h(zLogin)</a>
195	@ <td>%h(zCap)
196	@ <td>%h(zInfo)
197	@ <td>%h(zDate?zDate:"")
198	@ <td>%h(zExp?zExp:"")
199	@ <td data-sortkey='%f(rATime)' style='white-space:nowrap'>%s(zAge?zAge:"")
200	if( db_column_type(&s,8)==SQLITE_NULL ){
201	@ <td>
202	}else if( (zSub = db_column_text(&s,8))==0 \|\| zSub[0]==0 ){
	@@ -1016,11 +1018,11 @@
1016	style_header("User %h", db_column_text(&q,1));
1017	@ <table class="label-value">
1018	@ <tr><th>uid:</th><td>%d(db_column_int(&q,0))
1019	@ (<a href="%R/setup_uedit?id=%d(db_column_int(&q,0))">edit</a>)</td></tr>
1020	@ <tr><th>login:</th><td>%h(db_column_text(&q,1))</td></tr>
1021	@ <tr><th>capabilities:</th><td>%h(db_column_text(&q,2))</th></tr>
1022	@ <tr><th valign="top">info:</th>
1023	@ <td valign="top"><span style='white-space:pre-line;'>\
1024	@ %h(db_column_text(&q,5))</span></td></tr>
1025	@ <tr><th>user.mtime:</th><td>%h(db_column_text(&q,6))</td></tr>
1026	if( db_column_type(&q,7)!=SQLITE_NULL ){
1027

	--- src/setupuser.c
	+++ src/setupuser.c
	@@ -115,11 +115,11 @@
115	}
116	if( !bUnusedOnly ){
117	style_submenu_element("Unused", "setup_ulist?unused");
118	}
119	@ <table border=1 cellpadding=2 cellspacing=0 class='userTable sortable' \
120	@ data-column-types='ktxKTKt' data-init-sort='4'>
121	@ <thead><tr>
122	@ <th>Login Name<th>Caps<th>Info<th>Date<th>Expire<th>Last Login\
123	@ <th>Alerts</tr></thead>
124	@ <tbody>
125	db_multi_exec(
	@@ -161,15 +161,16 @@
161	" lower(login) AS sortkey, "
162	" CASE WHEN info LIKE '%%expires 20%%'"
163	" THEN substr(info,instr(lower(info),'expires')+8,10)"
164	" END AS exp,"
165	"atime,"
166	" subscriber.ssub, subscriber.subscriberId,"
167	" user.mtime AS sorttime"
168	" FROM user LEFT JOIN lastAccess ON login=uname"
169	" LEFT JOIN subscriber ON login=suname"
170	" WHERE login NOT IN ('anonymous','nobody','developer','reader') %s"
171	" ORDER BY sorttime DESC", zWith/safe-for-%s/
172	);
173	rNow = db_double(0.0, "SELECT julianday('now');");
174	while( db_step(&s)==SQLITE_ROW ){
175	int uid = db_column_int(&s, 0);
176	const char *zLogin = db_column_text(&s, 1);
	@@ -180,10 +181,11 @@
181	const char *zExp = db_column_text(&s,6);
182	double rATime = db_column_double(&s,7);
183	char *zAge = 0;
184	const char *zSub;
185	int sid = db_column_int(&s,9);
186	sqlite3_int64 sorttime = db_column_int64(&s, 10);
187	if( rATime>0.0 ){
188	zAge = human_readable_age(rNow - rATime);
189	}
190	if( bUbg ){
191	@ <tr style='background-color: %h(user_color(zLogin));'>
	@@ -192,11 +194,11 @@
194	}
195	@ <td data-sortkey='%h(zSortKey)'>\
196	@ <a href='setup_uedit?id=%d(uid)'>%h(zLogin)</a>
197	@ <td>%h(zCap)
198	@ <td>%h(zInfo)
199	@ <td data-sortkey='%09llx(sorttime)'>%h(zDate?zDate:"")
200	@ <td>%h(zExp?zExp:"")
201	@ <td data-sortkey='%f(rATime)' style='white-space:nowrap'>%s(zAge?zAge:"")
202	if( db_column_type(&s,8)==SQLITE_NULL ){
203	@ <td>
204	}else if( (zSub = db_column_text(&s,8))==0 \|\| zSub[0]==0 ){
	@@ -1016,11 +1018,11 @@
1018	style_header("User %h", db_column_text(&q,1));
1019	@ <table class="label-value">
1020	@ <tr><th>uid:</th><td>%d(db_column_int(&q,0))
1021	@ (<a href="%R/setup_uedit?id=%d(db_column_int(&q,0))">edit</a>)</td></tr>
1022	@ <tr><th>login:</th><td>%h(db_column_text(&q,1))</td></tr>
1023	@ <tr><th>capabilities:</th><td>%h(db_column_text(&q,2))</td></tr>
1024	@ <tr><th valign="top">info:</th>
1025	@ <td valign="top"><span style='white-space:pre-line;'>\
1026	@ %h(db_column_text(&q,5))</span></td></tr>
1027	@ <tr><th>user.mtime:</th><td>%h(db_column_text(&q,6))</td></tr>
1028	if( db_column_type(&q,7)!=SQLITE_NULL ){
1029

M src/sitemap.c

+5 -40

		--- src/sitemap.c
		+++ src/sitemap.c
		@@ -56,22 +56,10 @@
56	56	int i;
57	57	int isPopup = 0; /* This is an XMLHttpRequest() for /sitemap */
58	58	int e = atoi(PD("e","0"));
59	59	const char *zExtra;
60	60
61		-#if 0 /* Removed 2021-01-26 */
62		- const struct {
63		- const char *zTitle;
64		- const char *zProperty;
65		- } aExtra[] = {
66		- { "Documentation", "sitemap-docidx" },
67		- { "Download", "sitemap-download" },
68		- { "License", "sitemap-license" },
69		- { "Contact", "sitemap-contact" },
70		- };
71		-#endif
72		-
73	61	login_check_credentials();
74	62	if( P("popup")!=0 ){
75	63	/* The "popup" query parameter
76	64	** then disable anti-robot defenses */
77	65	isPopup = 1;
		@@ -87,26 +75,10 @@
87	75	@ <ul id="sitemap" class="columns" style="column-width:20em">
88	76	if( (e&1)==0 ){
89	77	@ <li>%z(href("%R/home"))Home Page</a>
90	78	}
91	79
92		-#if 0 /* Removed 2021-01-26 */
93		- for(i=0; i<sizeof(aExtra)/sizeof(aExtra[0]); i++){
94		- char *z = db_get(aExtra[i].zProperty,0);
95		- if( z==0 \|\| z[0]==0 ) continue;
96		- if( !inSublist ){
97		- @ <ul>
98		- inSublist = 1;
99		- }
100		- if( z[0]=='/' ){
101		- @ <li>%z(href("%R%s",z))%s(aExtra[i].zTitle)</a></li>
102		- }else{
103		- @ <li>%z(href("%s",z))%s(aExtra[i].zTitle)</a></li>
104		- }
105		- }
106		-#endif
107		-
108	80	zExtra = db_get("sitemap-extra",0);
109	81	if( zExtra && (e&2)==0 ){
110	82	int rc;
111	83	char **azExtra = 0;
112	84	int *anExtra;
		@@ -139,26 +111,18 @@
139	111	}
140	112	Th_Free(g.interp, azExtra);
141	113	}
142	114	if( (e&1)!=0 ) goto end_of_sitemap;
143	115
144		-#if 0 /* Removed on 2021-02-11. Make a sitemap-extra entry if you */
145		- /* really want this */
146		- if( srchFlags & SRCH_DOC ){
147		- if( !inSublist ){
148		- @ <ul>
149		- inSublist = 1;
150		- }
151		- @ <li>%z(href("%R/docsrch"))Documentation Search</a></li>
152		- }
153		-#endif
154		-
155	116	if( inSublist ){
156	117	@ </ul>
157	118	inSublist = 0;
158	119	}
159	120	@ </li>
	121	+ if( db_open_local(0) && cgi_is_loopback(g.zIpAddr) ){
	122	+ @ <li>%z(href("%R/ckout"))Checkout Status</a></li>
	123	+ }
160	124	if( g.perm.Read ){
161	125	const char *zEditGlob = db_get("fileedit-glob","");
162	126	@ <li>%z(href("%R/tree"))File Browser</a>
163	127	@ <ul>
164	128	@ <li>%z(href("%R/tree?type=tree&ci=trunk"))Tree-view,
		@@ -192,11 +156,12 @@
192	156	}
193	157	if( g.perm.Chat ){
194	158	@ <li>%z(href("%R/chat"))Chat</a></li>
195	159	}
196	160	if( g.perm.RdForum ){
197		- @ <li>%z(href("%R/forum"))Forum</a>
	161	+ const char *zTitle = db_get("forum-title","Forum");
	162	+ @ <li>%z(href("%R/forum"))%h(zTitle)</a>
198	163	@ <ul>
199	164	@ <li>%z(href("%R/timeline?y=f"))Recent activity</a></li>
200	165	@ </ul>
201	166	@ </li>
202	167	}
203	168

	--- src/sitemap.c
	+++ src/sitemap.c
	@@ -56,22 +56,10 @@
56	int i;
57	int isPopup = 0; /* This is an XMLHttpRequest() for /sitemap */
58	int e = atoi(PD("e","0"));
59	const char *zExtra;
60
61	#if 0 /* Removed 2021-01-26 */
62	const struct {
63	const char *zTitle;
64	const char *zProperty;
65	} aExtra[] = {
66	{ "Documentation", "sitemap-docidx" },
67	{ "Download", "sitemap-download" },
68	{ "License", "sitemap-license" },
69	{ "Contact", "sitemap-contact" },
70	};
71	#endif
72
73	login_check_credentials();
74	if( P("popup")!=0 ){
75	/* The "popup" query parameter
76	** then disable anti-robot defenses */
77	isPopup = 1;
	@@ -87,26 +75,10 @@
87	@ <ul id="sitemap" class="columns" style="column-width:20em">
88	if( (e&1)==0 ){
89	@ <li>%z(href("%R/home"))Home Page</a>
90	}
91
92	#if 0 /* Removed 2021-01-26 */
93	for(i=0; i<sizeof(aExtra)/sizeof(aExtra[0]); i++){
94	char *z = db_get(aExtra[i].zProperty,0);
95	if( z==0 \|\| z[0]==0 ) continue;
96	if( !inSublist ){
97	@ <ul>
98	inSublist = 1;
99	}
100	if( z[0]=='/' ){
101	@ <li>%z(href("%R%s",z))%s(aExtra[i].zTitle)</a></li>
102	}else{
103	@ <li>%z(href("%s",z))%s(aExtra[i].zTitle)</a></li>
104	}
105	}
106	#endif
107
108	zExtra = db_get("sitemap-extra",0);
109	if( zExtra && (e&2)==0 ){
110	int rc;
111	char **azExtra = 0;
112	int *anExtra;
	@@ -139,26 +111,18 @@
139	}
140	Th_Free(g.interp, azExtra);
141	}
142	if( (e&1)!=0 ) goto end_of_sitemap;
143
144	#if 0 /* Removed on 2021-02-11. Make a sitemap-extra entry if you */
145	/* really want this */
146	if( srchFlags & SRCH_DOC ){
147	if( !inSublist ){
148	@ <ul>
149	inSublist = 1;
150	}
151	@ <li>%z(href("%R/docsrch"))Documentation Search</a></li>
152	}
153	#endif
154
155	if( inSublist ){
156	@ </ul>
157	inSublist = 0;
158	}
159	@ </li>



160	if( g.perm.Read ){
161	const char *zEditGlob = db_get("fileedit-glob","");
162	@ <li>%z(href("%R/tree"))File Browser</a>
163	@ <ul>
164	@ <li>%z(href("%R/tree?type=tree&ci=trunk"))Tree-view,
	@@ -192,11 +156,12 @@
192	}
193	if( g.perm.Chat ){
194	@ <li>%z(href("%R/chat"))Chat</a></li>
195	}
196	if( g.perm.RdForum ){
197	@ <li>%z(href("%R/forum"))Forum</a>

198	@ <ul>
199	@ <li>%z(href("%R/timeline?y=f"))Recent activity</a></li>
200	@ </ul>
201	@ </li>
202	}
203

	--- src/sitemap.c
	+++ src/sitemap.c
	@@ -56,22 +56,10 @@
56	int i;
57	int isPopup = 0; /* This is an XMLHttpRequest() for /sitemap */
58	int e = atoi(PD("e","0"));
59	const char *zExtra;
60












61	login_check_credentials();
62	if( P("popup")!=0 ){
63	/* The "popup" query parameter
64	** then disable anti-robot defenses */
65	isPopup = 1;
	@@ -87,26 +75,10 @@
75	@ <ul id="sitemap" class="columns" style="column-width:20em">
76	if( (e&1)==0 ){
77	@ <li>%z(href("%R/home"))Home Page</a>
78	}
79
















80	zExtra = db_get("sitemap-extra",0);
81	if( zExtra && (e&2)==0 ){
82	int rc;
83	char **azExtra = 0;
84	int *anExtra;
	@@ -139,26 +111,18 @@
111	}
112	Th_Free(g.interp, azExtra);
113	}
114	if( (e&1)!=0 ) goto end_of_sitemap;
115











116	if( inSublist ){
117	@ </ul>
118	inSublist = 0;
119	}
120	@ </li>
121	if( db_open_local(0) && cgi_is_loopback(g.zIpAddr) ){
122	@ <li>%z(href("%R/ckout"))Checkout Status</a></li>
123	}
124	if( g.perm.Read ){
125	const char *zEditGlob = db_get("fileedit-glob","");
126	@ <li>%z(href("%R/tree"))File Browser</a>
127	@ <ul>
128	@ <li>%z(href("%R/tree?type=tree&ci=trunk"))Tree-view,
	@@ -192,11 +156,12 @@
156	}
157	if( g.perm.Chat ){
158	@ <li>%z(href("%R/chat"))Chat</a></li>
159	}
160	if( g.perm.RdForum ){
161	const char *zTitle = db_get("forum-title","Forum");
162	@ <li>%z(href("%R/forum"))%h(zTitle)</a>
163	@ <ul>
164	@ <li>%z(href("%R/timeline?y=f"))Recent activity</a></li>
165	@ </ul>
166	@ </li>
167	}
168

M src/th_tcl.c

		--- src/th_tcl.c
		+++ src/th_tcl.c
		@@ -731,10 +731,19 @@
731	731	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDIN));
732	732	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDOUT));
733	733	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDERR));
734	734	}
735	735	Tcl_Preserve((ClientData)tclInterp);
	736	+#if ((TCL_MAJOR_VERSION==8 && TCL_MINOR_VERSION>6) \
	737	+ \|\| (TCL_MAJOR_VERSION>8))
	738	+ /* TCL 8.7+ removes Tcl_MakeSafe():
	739	+ ** https://core.tcl-lang.org/tcl/tktview?name=655300
	740	+ ** https://core.tcl-lang.org/tips/doc/trunk/tip/624.md
	741	+ ** 8.7 has it in the headers but not in the libs.
	742	+ */
	743	+# define Tcl_MakeSafe(X) TCL_OK
	744	+#endif
736	745	if( Tcl_MakeSafe(tclInterp)!=TCL_OK ){
737	746	int nResult;
738	747	const char *zResult = getTclResult(tclInterp, &nResult);
739	748	Th_ErrorMessage(interp,
740	749	"could not make Tcl interpreter 'safe':", zResult, nResult);
741	750

	--- src/th_tcl.c
	+++ src/th_tcl.c
	@@ -731,10 +731,19 @@
731	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDIN));
732	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDOUT));
733	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDERR));
734	}
735	Tcl_Preserve((ClientData)tclInterp);









736	if( Tcl_MakeSafe(tclInterp)!=TCL_OK ){
737	int nResult;
738	const char *zResult = getTclResult(tclInterp, &nResult);
739	Th_ErrorMessage(interp,
740	"could not make Tcl interpreter 'safe':", zResult, nResult);
741

	--- src/th_tcl.c
	+++ src/th_tcl.c
	@@ -731,10 +731,19 @@
731	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDIN));
732	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDOUT));
733	Tcl_RegisterChannel(NULL, Tcl_GetStdChannel(TCL_STDERR));
734	}
735	Tcl_Preserve((ClientData)tclInterp);
736	#if ((TCL_MAJOR_VERSION==8 && TCL_MINOR_VERSION>6) \
737	\|\| (TCL_MAJOR_VERSION>8))
738	/* TCL 8.7+ removes Tcl_MakeSafe():
739	** https://core.tcl-lang.org/tcl/tktview?name=655300
740	** https://core.tcl-lang.org/tips/doc/trunk/tip/624.md
741	** 8.7 has it in the headers but not in the libs.
742	*/
743	# define Tcl_MakeSafe(X) TCL_OK
744	#endif
745	if( Tcl_MakeSafe(tclInterp)!=TCL_OK ){
746	int nResult;
747	const char *zResult = getTclResult(tclInterp, &nResult);
748	Th_ErrorMessage(interp,
749	"could not make Tcl interpreter 'safe':", zResult, nResult);
750

M src/timeline.c

+2 -1

		--- src/timeline.c
		+++ src/timeline.c
		@@ -1885,10 +1885,11 @@
1885	1885	char zPlural; / Ending for plural forms */
1886	1886	int showCherrypicks = 1; /* True to show cherrypick merges */
1887	1887	int haveParameterN; /* True if n= query parameter present */
1888	1888	int from_to_mode = 0; /* 0: from,to. 1: from,ft 2: from,bt */
1889	1889
	1890	+ login_check_credentials();
1890	1891	url_initialize(&url, "timeline");
1891	1892	cgi_query_parameters_to_url(&url);
1892	1893
1893	1894	(void)P_NoBot("ss")
1894	1895	/* "ss" is processed via the udc but at least one spider likes to
		@@ -1965,11 +1966,10 @@
1965	1966	*/
1966	1967	pd_rid = name_choice("dp","dp2",&zDPName);
1967	1968	if( pd_rid ){
1968	1969	p_rid = d_rid = pd_rid;
1969	1970	}
1970		- login_check_credentials();
1971	1971	if( (!g.perm.Read && !g.perm.RdTkt && !g.perm.RdWiki && !g.perm.RdForum)
1972	1972	\|\| (bisectLocal && !g.perm.Setup)
1973	1973	){
1974	1974	login_needed(g.anon.Read && g.anon.RdTkt && g.anon.RdWiki);
1975	1975	return;
		@@ -2012,10 +2012,11 @@
2012	2012	zTagName = z;
2013	2013	zMatchStyle = "brlist";
2014	2014	}
2015	2015	if( (z = P("rl"))!=0 ){
2016	2016	zBrName = z;
	2017	+ related = 1;
2017	2018	zMatchStyle = "brlist";
2018	2019	}
2019	2020	}
2020	2021
2021	2022	/* Convert r=TAG to t=TAG&rel in order to populate the UI style widgets. */
2022	2023

	--- src/timeline.c
	+++ src/timeline.c
	@@ -1885,10 +1885,11 @@
1885	char zPlural; / Ending for plural forms */
1886	int showCherrypicks = 1; /* True to show cherrypick merges */
1887	int haveParameterN; /* True if n= query parameter present */
1888	int from_to_mode = 0; /* 0: from,to. 1: from,ft 2: from,bt */
1889

1890	url_initialize(&url, "timeline");
1891	cgi_query_parameters_to_url(&url);
1892
1893	(void)P_NoBot("ss")
1894	/* "ss" is processed via the udc but at least one spider likes to
	@@ -1965,11 +1966,10 @@
1965	*/
1966	pd_rid = name_choice("dp","dp2",&zDPName);
1967	if( pd_rid ){
1968	p_rid = d_rid = pd_rid;
1969	}
1970	login_check_credentials();
1971	if( (!g.perm.Read && !g.perm.RdTkt && !g.perm.RdWiki && !g.perm.RdForum)
1972	\|\| (bisectLocal && !g.perm.Setup)
1973	){
1974	login_needed(g.anon.Read && g.anon.RdTkt && g.anon.RdWiki);
1975	return;
	@@ -2012,10 +2012,11 @@
2012	zTagName = z;
2013	zMatchStyle = "brlist";
2014	}
2015	if( (z = P("rl"))!=0 ){
2016	zBrName = z;

2017	zMatchStyle = "brlist";
2018	}
2019	}
2020
2021	/* Convert r=TAG to t=TAG&rel in order to populate the UI style widgets. */
2022

	--- src/timeline.c
	+++ src/timeline.c
	@@ -1885,10 +1885,11 @@
1885	char zPlural; / Ending for plural forms */
1886	int showCherrypicks = 1; /* True to show cherrypick merges */
1887	int haveParameterN; /* True if n= query parameter present */
1888	int from_to_mode = 0; /* 0: from,to. 1: from,ft 2: from,bt */
1889
1890	login_check_credentials();
1891	url_initialize(&url, "timeline");
1892	cgi_query_parameters_to_url(&url);
1893
1894	(void)P_NoBot("ss")
1895	/* "ss" is processed via the udc but at least one spider likes to
	@@ -1965,11 +1966,10 @@
1966	*/
1967	pd_rid = name_choice("dp","dp2",&zDPName);
1968	if( pd_rid ){
1969	p_rid = d_rid = pd_rid;
1970	}

1971	if( (!g.perm.Read && !g.perm.RdTkt && !g.perm.RdWiki && !g.perm.RdForum)
1972	\|\| (bisectLocal && !g.perm.Setup)
1973	){
1974	login_needed(g.anon.Read && g.anon.RdTkt && g.anon.RdWiki);
1975	return;
	@@ -2012,10 +2012,11 @@
2012	zTagName = z;
2013	zMatchStyle = "brlist";
2014	}
2015	if( (z = P("rl"))!=0 ){
2016	zBrName = z;
2017	related = 1;
2018	zMatchStyle = "brlist";
2019	}
2020	}
2021
2022	/* Convert r=TAG to t=TAG&rel in order to populate the UI style widgets. */
2023

M src/timeline.c

+2 -1

		--- src/timeline.c
		+++ src/timeline.c
		@@ -1885,10 +1885,11 @@
1885	1885	char zPlural; / Ending for plural forms */
1886	1886	int showCherrypicks = 1; /* True to show cherrypick merges */
1887	1887	int haveParameterN; /* True if n= query parameter present */
1888	1888	int from_to_mode = 0; /* 0: from,to. 1: from,ft 2: from,bt */
1889	1889
	1890	+ login_check_credentials();
1890	1891	url_initialize(&url, "timeline");
1891	1892	cgi_query_parameters_to_url(&url);
1892	1893
1893	1894	(void)P_NoBot("ss")
1894	1895	/* "ss" is processed via the udc but at least one spider likes to
		@@ -1965,11 +1966,10 @@
1965	1966	*/
1966	1967	pd_rid = name_choice("dp","dp2",&zDPName);
1967	1968	if( pd_rid ){
1968	1969	p_rid = d_rid = pd_rid;
1969	1970	}
1970		- login_check_credentials();
1971	1971	if( (!g.perm.Read && !g.perm.RdTkt && !g.perm.RdWiki && !g.perm.RdForum)
1972	1972	\|\| (bisectLocal && !g.perm.Setup)
1973	1973	){
1974	1974	login_needed(g.anon.Read && g.anon.RdTkt && g.anon.RdWiki);
1975	1975	return;
		@@ -2012,10 +2012,11 @@
2012	2012	zTagName = z;
2013	2013	zMatchStyle = "brlist";
2014	2014	}
2015	2015	if( (z = P("rl"))!=0 ){
2016	2016	zBrName = z;
	2017	+ related = 1;
2017	2018	zMatchStyle = "brlist";
2018	2019	}
2019	2020	}
2020	2021
2021	2022	/* Convert r=TAG to t=TAG&rel in order to populate the UI style widgets. */
2022	2023

	--- src/timeline.c
	+++ src/timeline.c
	@@ -1885,10 +1885,11 @@
1885	char zPlural; / Ending for plural forms */
1886	int showCherrypicks = 1; /* True to show cherrypick merges */
1887	int haveParameterN; /* True if n= query parameter present */
1888	int from_to_mode = 0; /* 0: from,to. 1: from,ft 2: from,bt */
1889

1890	url_initialize(&url, "timeline");
1891	cgi_query_parameters_to_url(&url);
1892
1893	(void)P_NoBot("ss")
1894	/* "ss" is processed via the udc but at least one spider likes to
	@@ -1965,11 +1966,10 @@
1965	*/
1966	pd_rid = name_choice("dp","dp2",&zDPName);
1967	if( pd_rid ){
1968	p_rid = d_rid = pd_rid;
1969	}
1970	login_check_credentials();
1971	if( (!g.perm.Read && !g.perm.RdTkt && !g.perm.RdWiki && !g.perm.RdForum)
1972	\|\| (bisectLocal && !g.perm.Setup)
1973	){
1974	login_needed(g.anon.Read && g.anon.RdTkt && g.anon.RdWiki);
1975	return;
	@@ -2012,10 +2012,11 @@
2012	zTagName = z;
2013	zMatchStyle = "brlist";
2014	}
2015	if( (z = P("rl"))!=0 ){
2016	zBrName = z;

2017	zMatchStyle = "brlist";
2018	}
2019	}
2020
2021	/* Convert r=TAG to t=TAG&rel in order to populate the UI style widgets. */
2022

	--- src/timeline.c
	+++ src/timeline.c
	@@ -1885,10 +1885,11 @@
1885	char zPlural; / Ending for plural forms */
1886	int showCherrypicks = 1; /* True to show cherrypick merges */
1887	int haveParameterN; /* True if n= query parameter present */
1888	int from_to_mode = 0; /* 0: from,to. 1: from,ft 2: from,bt */
1889
1890	login_check_credentials();
1891	url_initialize(&url, "timeline");
1892	cgi_query_parameters_to_url(&url);
1893
1894	(void)P_NoBot("ss")
1895	/* "ss" is processed via the udc but at least one spider likes to
	@@ -1965,11 +1966,10 @@
1966	*/
1967	pd_rid = name_choice("dp","dp2",&zDPName);
1968	if( pd_rid ){
1969	p_rid = d_rid = pd_rid;
1970	}

1971	if( (!g.perm.Read && !g.perm.RdTkt && !g.perm.RdWiki && !g.perm.RdForum)
1972	\|\| (bisectLocal && !g.perm.Setup)
1973	){
1974	login_needed(g.anon.Read && g.anon.RdTkt && g.anon.RdWiki);
1975	return;
	@@ -2012,10 +2012,11 @@
2012	zTagName = z;
2013	zMatchStyle = "brlist";
2014	}
2015	if( (z = P("rl"))!=0 ){
2016	zBrName = z;
2017	related = 1;
2018	zMatchStyle = "brlist";
2019	}
2020	}
2021
2022	/* Convert r=TAG to t=TAG&rel in order to populate the UI style widgets. */
2023

M src/tkt.c

+1 -1

		--- src/tkt.c
		+++ src/tkt.c
		@@ -781,11 +781,11 @@
781	781
782	782	zFullName = db_text(0,
783	783	"SELECT tkt_uuid FROM ticket"
784	784	" WHERE tkt_uuid GLOB '%q*'", zUuid);
785	785	if( zFullName ){
786		- attachment_list(zFullName, "<hr><h2>Attachments:</h2><ul>");
	786	+ attachment_list(zFullName, "<h2>Attachments:</h2>", 1);
787	787	}
788	788
789	789	style_finish_page();
790	790	}
791	791
792	792

	--- src/tkt.c
	+++ src/tkt.c
	@@ -781,11 +781,11 @@
781
782	zFullName = db_text(0,
783	"SELECT tkt_uuid FROM ticket"
784	" WHERE tkt_uuid GLOB '%q*'", zUuid);
785	if( zFullName ){
786	attachment_list(zFullName, "<hr><h2>Attachments:</h2><ul>");
787	}
788
789	style_finish_page();
790	}
791
792

	--- src/tkt.c
	+++ src/tkt.c
	@@ -781,11 +781,11 @@
781
782	zFullName = db_text(0,
783	"SELECT tkt_uuid FROM ticket"
784	" WHERE tkt_uuid GLOB '%q*'", zUuid);
785	if( zFullName ){
786	attachment_list(zFullName, "<h2>Attachments:</h2>", 1);
787	}
788
789	style_finish_page();
790	}
791
792

M src/update.c

+55 -3

		--- src/update.c
		+++ src/update.c
		@@ -132,10 +132,13 @@
132	132	int nUpdate = 0; /* Number of changes of any kind */
133	133	int bNosync = 0; /* --nosync. Omit the auto-sync */
134	134	int width; /* Width of printed comment lines */
135	135	Stmt mtimeXfer; /* Statement to transfer mtimes */
136	136	const char zWidth; / Width option string value */
	137	+ const char zCurBrName; / Current branch name */
	138	+ const char zNewBrName; / New branch name */
	139	+ const char zBrChgMsg = ""; / Message to display if branch changes */
137	140
138	141	if( !internalUpdate ){
139	142	undo_capture_command_line();
140	143	url_proxy_options();
141	144	}
		@@ -163,10 +166,11 @@
163	166	/* We should be done with options.. */
164	167	verify_all_options();
165	168
166	169	db_must_be_within_tree();
167	170	vid = db_lget_int("checkout", 0);
	171	+ zCurBrName = branch_of_rid(vid);
168	172	user_select();
169	173	if( !dryRunFlag && !internalUpdate && !bNosync ){
170	174	if( autosync_loop(SYNC_PULL + SYNC_VERBOSE*verboseFlag, 1, "update") ){
171	175	fossil_fatal("update abandoned due to sync failure");
172	176	}
		@@ -403,10 +407,11 @@
403	407	" WHERE id=:idt"
404	408	);
405	409	assert( g.zLocalRoot!=0 );
406	410	assert( strlen(g.zLocalRoot)>0 );
407	411	assert( g.zLocalRoot[strlen(g.zLocalRoot)-1]=='/' );
	412	+ merge_info_init();
408	413	while( db_step(&q)==SQLITE_ROW ){
409	414	const char zName = db_column_text(&q, 0); / The filename from root */
410	415	int idv = db_column_int(&q, 1); /* VFILE entry for current */
411	416	int ridv = db_column_int(&q, 2); /* RecordID for current */
412	417	int idt = db_column_int(&q, 3); /* VFILE entry for target */
		@@ -418,13 +423,18 @@
418	423	int islinkt = db_column_int(&q, 9); /* Is target file is a link */
419	424	int deleted = db_column_int(&q, 10); /* Marked for deletion */
420	425	char zFullPath; / Full pathname of the file */
421	426	char zFullNewPath; / Full pathname of dest */
422	427	char nameChng; /* True if the name changed */
	428	+ const char zOp = 0; / Type of change. */
	429	+ i64 sz = 0; /* Size of the file */
	430	+ int nc = 0; /* Number of conflicts */
	431	+ const char zErrMsg = 0; / Error message */
423	432
424	433	zFullPath = mprintf("%s%s", g.zLocalRoot, zName);
425	434	zFullNewPath = mprintf("%s%s", g.zLocalRoot, zNewName);
	435	+ sz = file_size(zFullNewPath, ExtFILE);
426	436	nameChng = fossil_strcmp(zName, zNewName);
427	437	nUpdate++;
428	438	if( deleted ){
429	439	db_multi_exec("UPDATE vfile SET deleted=1 WHERE id=%d", idt);
430	440	}
		@@ -432,10 +442,13 @@
432	442	/* Conflict. This file has been added to the current check-out
433	443	** but also exists in the target check-out. Use the current version.
434	444	*/
435	445	fossil_print("CONFLICT %s\n", zName);
436	446	nConflict++;
	447	+ zOp = "CONFLICT";
	448	+ nc = 1;
	449	+ zErrMsg = "duplicate file";
437	450	}else if( idt>0 && idv==0 ){
438	451	/* File added in the target. */
439	452	if( file_isfile_or_link(zFullPath) ){
440	453	/* Name of backup file with Original content */
441	454	char *zOrig = file_newname(zFullPath, "original", 1);
		@@ -444,10 +457,13 @@
444	457	fossil_free(zOrig);
445	458	fossil_print("ADD %s - overwrites an unmanaged file", zName);
446	459	if( !dryRunFlag ) fossil_print(", original copy backed up locally");
447	460	fossil_print("\n");
448	461	nOverwrite++;
	462	+ nc = 1;
	463	+ zOp = "CONFLICT";
	464	+ zErrMsg = "new file overwrites unmanaged file";
449	465	}else{
450	466	fossil_print("ADD %s\n", zName);
451	467	}
452	468	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
453	469	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
		@@ -458,16 +474,18 @@
458	474	}else{
459	475	fossil_print("UPDATE %s\n", zName);
460	476	}
461	477	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
462	478	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
	479	+ zOp = "UPDATE";
463	480	}else if( idt>0 && idv>0 && !deleted && file_size(zFullPath, RepoFILE)<0 ){
464	481	/* The file missing from the local check-out. Restore it to the
465	482	** version that appears in the target. */
466	483	fossil_print("UPDATE %s\n", zName);
467	484	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
468	485	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
	486	+ zOp = "UPDATE";
469	487	}else if( idt==0 && idv>0 ){
470	488	if( ridv==0 ){
471	489	/* Added in current check-out. Continue to hold the file as
472	490	** as an addition */
473	491	db_multi_exec("UPDATE vfile SET vid=%d WHERE id=%d", tid, idv);
		@@ -474,10 +492,13 @@
474	492	}else if( chnged ){
475	493	/* Edited locally but deleted from the target. Do not track the
476	494	** file but keep the edited version around. */
477	495	fossil_print("CONFLICT %s - edited locally but deleted by update\n",
478	496	zName);
	497	+ zOp = "CONFLICT";
	498	+ zErrMsg = "edited locally but deleted by update";
	499	+ nc = 1;
479	500	nConflict++;
480	501	}else{
481	502	fossil_print("REMOVE %s\n", zName);
482	503	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
483	504	if( !dryRunFlag ){
		@@ -496,17 +517,19 @@
496	517	}
497	518	}else if( idt>0 && idv>0 && ridt!=ridv && chnged ){
498	519	/* Merge the changes in the current tree into the target version */
499	520	Blob r, t, v;
500	521	int rc;
	522	+
501	523	if( nameChng ){
502	524	fossil_print("MERGE %s -> %s\n", zName, zNewName);
503	525	}else{
504	526	fossil_print("MERGE %s\n", zName);
505	527	}
506	528	if( islinkv \|\| islinkt ){
507	529	fossil_print("***** Cannot merge symlink %s\n", zNewName);
	530	+ zOp = "CONFLICT";
508	531	nConflict++;
509	532	}else{
510	533	unsigned mergeFlags = dryRunFlag ? MERGE_DRYRUN : 0;
511	534	if(keepMergeFlag!=0) mergeFlags \|= MERGE_KEEP_FILES;
512	535	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
		@@ -517,12 +540,17 @@
517	540	if( !dryRunFlag ){
518	541	blob_write_to_file(&r, zFullNewPath);
519	542	file_setexe(zFullNewPath, isexe);
520	543	}
521	544	if( rc>0 ){
	545	+ nc = rc;
	546	+ zOp = "CONFLICT";
	547	+ zErrMsg = "merge conflicts";
522	548	fossil_print("***** %d merge conflicts in %s\n", rc, zNewName);
523	549	nConflict++;
	550	+ }else{
	551	+ zOp = "MERGE";
524	552	}
525	553	}else{
526	554	if( !dryRunFlag ){
527	555	if( !keepMergeFlag ){
528	556	/* Name of backup file with Original content */
		@@ -539,10 +567,13 @@
539	567	if( !dryRunFlag ){
540	568	fossil_print(", original copy backed up locally");
541	569	}
542	570	fossil_print("\n");
543	571	nConflict++;
	572	+ zOp = "ERROR";
	573	+ zErrMsg = "cannot merge binary file";
	574	+ nc = 1;
544	575	}
545	576	}
546	577	if( nameChng && !dryRunFlag ) file_delete(zFullPath);
547	578	blob_reset(&v);
548	579	blob_reset(&t);
		@@ -556,27 +587,48 @@
556	587	db_bind_int(&mtimeXfer, ":idt", idt);
557	588	db_step(&mtimeXfer);
558	589	db_reset(&mtimeXfer);
559	590	if( verboseFlag ) fossil_print("UNCHANGED %s\n", zName);
560	591	}
	592	+ }
	593	+ if( zOp!=0 ){
	594	+ db_multi_exec(
	595	+ "INSERT INTO mergestat(op,fnp,ridp,fn,ridv,sz,fnm,ridm,fnr,nc,msg)"
	596	+ "VALUES(%Q,%Q,%d,%Q,NULL,%lld,%Q,%d,%Q,%d,%Q)",
	597	+ /* op */ zOp,
	598	+ /* fnp */ zName,
	599	+ /* ridp */ ridv,
	600	+ /* fn */ zNewName,
	601	+ /* sz */ sz,
	602	+ /* fnm */ zName,
	603	+ /* ridm */ ridt,
	604	+ /* fnr */ zNewName,
	605	+ /* nc */ nc,
	606	+ /* msg */ zErrMsg
	607	+ );
561	608	}
562	609	free(zFullPath);
563	610	free(zFullNewPath);
564	611	}
565	612	db_finalize(&q);
566	613	db_finalize(&mtimeXfer);
567	614	fossil_print("%.79c\n",'-');
	615	+ zNewBrName = branch_of_rid(tid);
	616	+ if( g.argc<3 && fossil_strcmp(zCurBrName, zNewBrName)!=0 ){
	617	+ zBrChgMsg = mprintf(" Branch changed from %s to %s.",
	618	+ zCurBrName, zNewBrName);
	619	+ }
568	620	if( nUpdate==0 ){
569	621	show_common_info(tid, "checkout:", 1, 0);
570		- fossil_print("%-13s None. Already up-to-date\n", "changes:");
	622	+ fossil_print("%-13s None. Already up-to-date.%s\n", "changes:", zBrChgMsg);
571	623	}else{
572	624	fossil_print("%-13s %.40s %s\n", "updated-from:", rid_to_uuid(vid),
573	625	db_text("", "SELECT datetime(mtime) \|\| ' UTC' FROM event "
574	626	" WHERE objid=%d", vid));
575	627	show_common_info(tid, "updated-to:", 1, 0);
576		- fossil_print("%-13s %d file%s modified.\n", "changes:",
577		- nUpdate, nUpdate>1 ? "s" : "");
	628	+ fossil_print("%-13s %d file%s modified.%s\n", "changes:",
	629	+ nUpdate, nUpdate>1 ? "s" : "", zBrChgMsg);
578	630	}
579	631
580	632	/* Report on conflicts
581	633	*/
582	634	if( !dryRunFlag ){
583	635

	--- src/update.c
	+++ src/update.c
	@@ -132,10 +132,13 @@
132	int nUpdate = 0; /* Number of changes of any kind */
133	int bNosync = 0; /* --nosync. Omit the auto-sync */
134	int width; /* Width of printed comment lines */
135	Stmt mtimeXfer; /* Statement to transfer mtimes */
136	const char zWidth; / Width option string value */



137
138	if( !internalUpdate ){
139	undo_capture_command_line();
140	url_proxy_options();
141	}
	@@ -163,10 +166,11 @@
163	/* We should be done with options.. */
164	verify_all_options();
165
166	db_must_be_within_tree();
167	vid = db_lget_int("checkout", 0);

168	user_select();
169	if( !dryRunFlag && !internalUpdate && !bNosync ){
170	if( autosync_loop(SYNC_PULL + SYNC_VERBOSE*verboseFlag, 1, "update") ){
171	fossil_fatal("update abandoned due to sync failure");
172	}
	@@ -403,10 +407,11 @@
403	" WHERE id=:idt"
404	);
405	assert( g.zLocalRoot!=0 );
406	assert( strlen(g.zLocalRoot)>0 );
407	assert( g.zLocalRoot[strlen(g.zLocalRoot)-1]=='/' );

408	while( db_step(&q)==SQLITE_ROW ){
409	const char zName = db_column_text(&q, 0); / The filename from root */
410	int idv = db_column_int(&q, 1); /* VFILE entry for current */
411	int ridv = db_column_int(&q, 2); /* RecordID for current */
412	int idt = db_column_int(&q, 3); /* VFILE entry for target */
	@@ -418,13 +423,18 @@
418	int islinkt = db_column_int(&q, 9); /* Is target file is a link */
419	int deleted = db_column_int(&q, 10); /* Marked for deletion */
420	char zFullPath; / Full pathname of the file */
421	char zFullNewPath; / Full pathname of dest */
422	char nameChng; /* True if the name changed */




423
424	zFullPath = mprintf("%s%s", g.zLocalRoot, zName);
425	zFullNewPath = mprintf("%s%s", g.zLocalRoot, zNewName);

426	nameChng = fossil_strcmp(zName, zNewName);
427	nUpdate++;
428	if( deleted ){
429	db_multi_exec("UPDATE vfile SET deleted=1 WHERE id=%d", idt);
430	}
	@@ -432,10 +442,13 @@
432	/* Conflict. This file has been added to the current check-out
433	** but also exists in the target check-out. Use the current version.
434	*/
435	fossil_print("CONFLICT %s\n", zName);
436	nConflict++;



437	}else if( idt>0 && idv==0 ){
438	/* File added in the target. */
439	if( file_isfile_or_link(zFullPath) ){
440	/* Name of backup file with Original content */
441	char *zOrig = file_newname(zFullPath, "original", 1);
	@@ -444,10 +457,13 @@
444	fossil_free(zOrig);
445	fossil_print("ADD %s - overwrites an unmanaged file", zName);
446	if( !dryRunFlag ) fossil_print(", original copy backed up locally");
447	fossil_print("\n");
448	nOverwrite++;



449	}else{
450	fossil_print("ADD %s\n", zName);
451	}
452	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
453	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
	@@ -458,16 +474,18 @@
458	}else{
459	fossil_print("UPDATE %s\n", zName);
460	}
461	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
462	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);

463	}else if( idt>0 && idv>0 && !deleted && file_size(zFullPath, RepoFILE)<0 ){
464	/* The file missing from the local check-out. Restore it to the
465	** version that appears in the target. */
466	fossil_print("UPDATE %s\n", zName);
467	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
468	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);

469	}else if( idt==0 && idv>0 ){
470	if( ridv==0 ){
471	/* Added in current check-out. Continue to hold the file as
472	** as an addition */
473	db_multi_exec("UPDATE vfile SET vid=%d WHERE id=%d", tid, idv);
	@@ -474,10 +492,13 @@
474	}else if( chnged ){
475	/* Edited locally but deleted from the target. Do not track the
476	** file but keep the edited version around. */
477	fossil_print("CONFLICT %s - edited locally but deleted by update\n",
478	zName);



479	nConflict++;
480	}else{
481	fossil_print("REMOVE %s\n", zName);
482	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
483	if( !dryRunFlag ){
	@@ -496,17 +517,19 @@
496	}
497	}else if( idt>0 && idv>0 && ridt!=ridv && chnged ){
498	/* Merge the changes in the current tree into the target version */
499	Blob r, t, v;
500	int rc;

501	if( nameChng ){
502	fossil_print("MERGE %s -> %s\n", zName, zNewName);
503	}else{
504	fossil_print("MERGE %s\n", zName);
505	}
506	if( islinkv \|\| islinkt ){
507	fossil_print("***** Cannot merge symlink %s\n", zNewName);

508	nConflict++;
509	}else{
510	unsigned mergeFlags = dryRunFlag ? MERGE_DRYRUN : 0;
511	if(keepMergeFlag!=0) mergeFlags \|= MERGE_KEEP_FILES;
512	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
	@@ -517,12 +540,17 @@
517	if( !dryRunFlag ){
518	blob_write_to_file(&r, zFullNewPath);
519	file_setexe(zFullNewPath, isexe);
520	}
521	if( rc>0 ){



522	fossil_print("***** %d merge conflicts in %s\n", rc, zNewName);
523	nConflict++;


524	}
525	}else{
526	if( !dryRunFlag ){
527	if( !keepMergeFlag ){
528	/* Name of backup file with Original content */
	@@ -539,10 +567,13 @@
539	if( !dryRunFlag ){
540	fossil_print(", original copy backed up locally");
541	}
542	fossil_print("\n");
543	nConflict++;



544	}
545	}
546	if( nameChng && !dryRunFlag ) file_delete(zFullPath);
547	blob_reset(&v);
548	blob_reset(&t);
	@@ -556,27 +587,48 @@
556	db_bind_int(&mtimeXfer, ":idt", idt);
557	db_step(&mtimeXfer);
558	db_reset(&mtimeXfer);
559	if( verboseFlag ) fossil_print("UNCHANGED %s\n", zName);
560	}
















561	}
562	free(zFullPath);
563	free(zFullNewPath);
564	}
565	db_finalize(&q);
566	db_finalize(&mtimeXfer);
567	fossil_print("%.79c\n",'-');





568	if( nUpdate==0 ){
569	show_common_info(tid, "checkout:", 1, 0);
570	fossil_print("%-13s None. Already up-to-date\n", "changes:");
571	}else{
572	fossil_print("%-13s %.40s %s\n", "updated-from:", rid_to_uuid(vid),
573	db_text("", "SELECT datetime(mtime) \|\| ' UTC' FROM event "
574	" WHERE objid=%d", vid));
575	show_common_info(tid, "updated-to:", 1, 0);
576	fossil_print("%-13s %d file%s modified.\n", "changes:",
577	nUpdate, nUpdate>1 ? "s" : "");
578	}
579
580	/* Report on conflicts
581	*/
582	if( !dryRunFlag ){
583

	--- src/update.c
	+++ src/update.c
	@@ -132,10 +132,13 @@
132	int nUpdate = 0; /* Number of changes of any kind */
133	int bNosync = 0; /* --nosync. Omit the auto-sync */
134	int width; /* Width of printed comment lines */
135	Stmt mtimeXfer; /* Statement to transfer mtimes */
136	const char zWidth; / Width option string value */
137	const char zCurBrName; / Current branch name */
138	const char zNewBrName; / New branch name */
139	const char zBrChgMsg = ""; / Message to display if branch changes */
140
141	if( !internalUpdate ){
142	undo_capture_command_line();
143	url_proxy_options();
144	}
	@@ -163,10 +166,11 @@
166	/* We should be done with options.. */
167	verify_all_options();
168
169	db_must_be_within_tree();
170	vid = db_lget_int("checkout", 0);
171	zCurBrName = branch_of_rid(vid);
172	user_select();
173	if( !dryRunFlag && !internalUpdate && !bNosync ){
174	if( autosync_loop(SYNC_PULL + SYNC_VERBOSE*verboseFlag, 1, "update") ){
175	fossil_fatal("update abandoned due to sync failure");
176	}
	@@ -403,10 +407,11 @@
407	" WHERE id=:idt"
408	);
409	assert( g.zLocalRoot!=0 );
410	assert( strlen(g.zLocalRoot)>0 );
411	assert( g.zLocalRoot[strlen(g.zLocalRoot)-1]=='/' );
412	merge_info_init();
413	while( db_step(&q)==SQLITE_ROW ){
414	const char zName = db_column_text(&q, 0); / The filename from root */
415	int idv = db_column_int(&q, 1); /* VFILE entry for current */
416	int ridv = db_column_int(&q, 2); /* RecordID for current */
417	int idt = db_column_int(&q, 3); /* VFILE entry for target */
	@@ -418,13 +423,18 @@
423	int islinkt = db_column_int(&q, 9); /* Is target file is a link */
424	int deleted = db_column_int(&q, 10); /* Marked for deletion */
425	char zFullPath; / Full pathname of the file */
426	char zFullNewPath; / Full pathname of dest */
427	char nameChng; /* True if the name changed */
428	const char zOp = 0; / Type of change. */
429	i64 sz = 0; /* Size of the file */
430	int nc = 0; /* Number of conflicts */
431	const char zErrMsg = 0; / Error message */
432
433	zFullPath = mprintf("%s%s", g.zLocalRoot, zName);
434	zFullNewPath = mprintf("%s%s", g.zLocalRoot, zNewName);
435	sz = file_size(zFullNewPath, ExtFILE);
436	nameChng = fossil_strcmp(zName, zNewName);
437	nUpdate++;
438	if( deleted ){
439	db_multi_exec("UPDATE vfile SET deleted=1 WHERE id=%d", idt);
440	}
	@@ -432,10 +442,13 @@
442	/* Conflict. This file has been added to the current check-out
443	** but also exists in the target check-out. Use the current version.
444	*/
445	fossil_print("CONFLICT %s\n", zName);
446	nConflict++;
447	zOp = "CONFLICT";
448	nc = 1;
449	zErrMsg = "duplicate file";
450	}else if( idt>0 && idv==0 ){
451	/* File added in the target. */
452	if( file_isfile_or_link(zFullPath) ){
453	/* Name of backup file with Original content */
454	char *zOrig = file_newname(zFullPath, "original", 1);
	@@ -444,10 +457,13 @@
457	fossil_free(zOrig);
458	fossil_print("ADD %s - overwrites an unmanaged file", zName);
459	if( !dryRunFlag ) fossil_print(", original copy backed up locally");
460	fossil_print("\n");
461	nOverwrite++;
462	nc = 1;
463	zOp = "CONFLICT";
464	zErrMsg = "new file overwrites unmanaged file";
465	}else{
466	fossil_print("ADD %s\n", zName);
467	}
468	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
469	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
	@@ -458,16 +474,18 @@
474	}else{
475	fossil_print("UPDATE %s\n", zName);
476	}
477	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
478	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
479	zOp = "UPDATE";
480	}else if( idt>0 && idv>0 && !deleted && file_size(zFullPath, RepoFILE)<0 ){
481	/* The file missing from the local check-out. Restore it to the
482	** version that appears in the target. */
483	fossil_print("UPDATE %s\n", zName);
484	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
485	if( !dryRunFlag ) vfile_to_disk(0, idt, 0, 0);
486	zOp = "UPDATE";
487	}else if( idt==0 && idv>0 ){
488	if( ridv==0 ){
489	/* Added in current check-out. Continue to hold the file as
490	** as an addition */
491	db_multi_exec("UPDATE vfile SET vid=%d WHERE id=%d", tid, idv);
	@@ -474,10 +492,13 @@
492	}else if( chnged ){
493	/* Edited locally but deleted from the target. Do not track the
494	** file but keep the edited version around. */
495	fossil_print("CONFLICT %s - edited locally but deleted by update\n",
496	zName);
497	zOp = "CONFLICT";
498	zErrMsg = "edited locally but deleted by update";
499	nc = 1;
500	nConflict++;
501	}else{
502	fossil_print("REMOVE %s\n", zName);
503	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
504	if( !dryRunFlag ){
	@@ -496,17 +517,19 @@
517	}
518	}else if( idt>0 && idv>0 && ridt!=ridv && chnged ){
519	/* Merge the changes in the current tree into the target version */
520	Blob r, t, v;
521	int rc;
522
523	if( nameChng ){
524	fossil_print("MERGE %s -> %s\n", zName, zNewName);
525	}else{
526	fossil_print("MERGE %s\n", zName);
527	}
528	if( islinkv \|\| islinkt ){
529	fossil_print("***** Cannot merge symlink %s\n", zNewName);
530	zOp = "CONFLICT";
531	nConflict++;
532	}else{
533	unsigned mergeFlags = dryRunFlag ? MERGE_DRYRUN : 0;
534	if(keepMergeFlag!=0) mergeFlags \|= MERGE_KEEP_FILES;
535	if( !dryRunFlag && !internalUpdate ) undo_save(zName);
	@@ -517,12 +540,17 @@
540	if( !dryRunFlag ){
541	blob_write_to_file(&r, zFullNewPath);
542	file_setexe(zFullNewPath, isexe);
543	}
544	if( rc>0 ){
545	nc = rc;
546	zOp = "CONFLICT";
547	zErrMsg = "merge conflicts";
548	fossil_print("***** %d merge conflicts in %s\n", rc, zNewName);
549	nConflict++;
550	}else{
551	zOp = "MERGE";
552	}
553	}else{
554	if( !dryRunFlag ){
555	if( !keepMergeFlag ){
556	/* Name of backup file with Original content */
	@@ -539,10 +567,13 @@
567	if( !dryRunFlag ){
568	fossil_print(", original copy backed up locally");
569	}
570	fossil_print("\n");
571	nConflict++;
572	zOp = "ERROR";
573	zErrMsg = "cannot merge binary file";
574	nc = 1;
575	}
576	}
577	if( nameChng && !dryRunFlag ) file_delete(zFullPath);
578	blob_reset(&v);
579	blob_reset(&t);
	@@ -556,27 +587,48 @@
587	db_bind_int(&mtimeXfer, ":idt", idt);
588	db_step(&mtimeXfer);
589	db_reset(&mtimeXfer);
590	if( verboseFlag ) fossil_print("UNCHANGED %s\n", zName);
591	}
592	}
593	if( zOp!=0 ){
594	db_multi_exec(
595	"INSERT INTO mergestat(op,fnp,ridp,fn,ridv,sz,fnm,ridm,fnr,nc,msg)"
596	"VALUES(%Q,%Q,%d,%Q,NULL,%lld,%Q,%d,%Q,%d,%Q)",
597	/* op */ zOp,
598	/* fnp */ zName,
599	/* ridp */ ridv,
600	/* fn */ zNewName,
601	/* sz */ sz,
602	/* fnm */ zName,
603	/* ridm */ ridt,
604	/* fnr */ zNewName,
605	/* nc */ nc,
606	/* msg */ zErrMsg
607	);
608	}
609	free(zFullPath);
610	free(zFullNewPath);
611	}
612	db_finalize(&q);
613	db_finalize(&mtimeXfer);
614	fossil_print("%.79c\n",'-');
615	zNewBrName = branch_of_rid(tid);
616	if( g.argc<3 && fossil_strcmp(zCurBrName, zNewBrName)!=0 ){
617	zBrChgMsg = mprintf(" Branch changed from %s to %s.",
618	zCurBrName, zNewBrName);
619	}
620	if( nUpdate==0 ){
621	show_common_info(tid, "checkout:", 1, 0);
622	fossil_print("%-13s None. Already up-to-date.%s\n", "changes:", zBrChgMsg);
623	}else{
624	fossil_print("%-13s %.40s %s\n", "updated-from:", rid_to_uuid(vid),
625	db_text("", "SELECT datetime(mtime) \|\| ' UTC' FROM event "
626	" WHERE objid=%d", vid));
627	show_common_info(tid, "updated-to:", 1, 0);
628	fossil_print("%-13s %d file%s modified.%s\n", "changes:",
629	nUpdate, nUpdate>1 ? "s" : "", zBrChgMsg);
630	}
631
632	/* Report on conflicts
633	*/
634	if( !dryRunFlag ){
635

M src/wiki.c

+4 -4

		--- src/wiki.c
		+++ src/wiki.c
		@@ -622,14 +622,14 @@
622	622	wiki_render_by_mimetype(&wiki, zMimetype);
623	623	blob_reset(&wiki);
624	624	}
625	625	manifest_destroy(pWiki);
626	626	if( !isPopup ){
627		- char * zLabel = mprintf("<hr><h2><a href='%R/attachlist?name=%T'>"
628		- "Attachments</a>:</h2><ul>",
	627	+ char * zLabel = mprintf("<h2><a href='%R/attachlist?page=%T'>"
	628	+ "Attachments</a>:</h2>",
629	629	zPageName);
630		- attachment_list(zPageName, zLabel);
	630	+ attachment_list(zPageName, zLabel, 1);
631	631	fossil_free(zLabel);
632	632	document_emit_js(/for optional pikchr support/);
633	633	style_finish_page();
634	634	}
635	635	}
		@@ -1331,11 +1331,11 @@
1331	1331	CX("<div id='fossil-status-bar' "
1332	1332	"title='Status message area. Double-click to clear them.'>"
1333	1333	"Status messages will go here.</div>\n"
1334	1334	/* will be moved into the tab container via JS */);
1335	1335
1336		- CX("<div id='wikiedit-edit-status''>"
	1336	+ CX("<div id='wikiedit-edit-status'>"
1337	1337	"<span class='name'></span>"
1338	1338	"<span class='links'></span>"
1339	1339	"</div>");
1340	1340
1341	1341	/* Main tab container... */
1342	1342

	--- src/wiki.c
	+++ src/wiki.c
	@@ -622,14 +622,14 @@
622	wiki_render_by_mimetype(&wiki, zMimetype);
623	blob_reset(&wiki);
624	}
625	manifest_destroy(pWiki);
626	if( !isPopup ){
627	char * zLabel = mprintf("<hr><h2><a href='%R/attachlist?name=%T'>"
628	"Attachments</a>:</h2><ul>",
629	zPageName);
630	attachment_list(zPageName, zLabel);
631	fossil_free(zLabel);
632	document_emit_js(/for optional pikchr support/);
633	style_finish_page();
634	}
635	}
	@@ -1331,11 +1331,11 @@
1331	CX("<div id='fossil-status-bar' "
1332	"title='Status message area. Double-click to clear them.'>"
1333	"Status messages will go here.</div>\n"
1334	/* will be moved into the tab container via JS */);
1335
1336	CX("<div id='wikiedit-edit-status''>"
1337	"<span class='name'></span>"
1338	"<span class='links'></span>"
1339	"</div>");
1340
1341	/* Main tab container... */
1342

	--- src/wiki.c
	+++ src/wiki.c
	@@ -622,14 +622,14 @@
622	wiki_render_by_mimetype(&wiki, zMimetype);
623	blob_reset(&wiki);
624	}
625	manifest_destroy(pWiki);
626	if( !isPopup ){
627	char * zLabel = mprintf("<h2><a href='%R/attachlist?page=%T'>"
628	"Attachments</a>:</h2>",
629	zPageName);
630	attachment_list(zPageName, zLabel, 1);
631	fossil_free(zLabel);
632	document_emit_js(/for optional pikchr support/);
633	style_finish_page();
634	}
635	}
	@@ -1331,11 +1331,11 @@
1331	CX("<div id='fossil-status-bar' "
1332	"title='Status message area. Double-click to clear them.'>"
1333	"Status messages will go here.</div>\n"
1334	/* will be moved into the tab container via JS */);
1335
1336	CX("<div id='wikiedit-edit-status'>"
1337	"<span class='name'></span>"
1338	"<span class='links'></span>"
1339	"</div>");
1340
1341	/* Main tab container... */
1342

M src/winfile.c

+229

		--- src/winfile.c
		+++ src/winfile.c
		@@ -290,6 +290,235 @@
290	290	fossil_free(zWide);
291	291	for(i=0; zUtf8[i]; i++) if( zUtf8[i]=='\\' ) zUtf8[i] = '/';
292	292	strncpy(zBuf, zUtf8, nBuf);
293	293	fossil_path_free(zUtf8);
294	294	}
	295	+
	296	+/* Perform case-insensitive comparison of two UTF-16 file names. Try to load the
	297	+** CompareStringOrdinal() function on Windows Vista and newer, and resort to the
	298	+** RtlEqualUnicodeString() function on Windows XP.
	299	+** The dance to invoke RtlEqualUnicodeString() is necessary because lstrcmpiW()
	300	+** performs linguistic comparison, while the former performs binary comparison.
	301	+** As an example, matching "ß" (U+00DF Latin Small Letter Sharp S) with "ss" is
	302	+** undesirable in file name comparison, so lstrcmpiW() is only invoked in cases
	303	+** that are technically impossible and contradicting all known laws of physics.
	304	+*/
	305	+int win32_filenames_equal_nocase(
	306	+ const wchar_t *fn1,
	307	+ const wchar_t *fn2
	308	+){
	309	+ static FARPROC fnCompareStringOrdinal;
	310	+ static FARPROC fnRtlInitUnicodeString;
	311	+ static FARPROC fnRtlEqualUnicodeString;
	312	+ static int loaded_CompareStringOrdinal;
	313	+ static int loaded_RtlUnicodeStringAPIs;
	314	+ if( !loaded_CompareStringOrdinal ){
	315	+ fnCompareStringOrdinal =
	316	+ GetProcAddress(GetModuleHandleA("kernel32"),"CompareStringOrdinal");
	317	+ loaded_CompareStringOrdinal = 1;
	318	+ }
	319	+ if( fnCompareStringOrdinal ){
	320	+ return fnCompareStringOrdinal(fn1,-1,fn2,-1,1)-2==0;
	321	+ }
	322	+ if( !loaded_RtlUnicodeStringAPIs ){
	323	+ fnRtlInitUnicodeString =
	324	+ GetProcAddress(GetModuleHandleA("ntdll"),"RtlInitUnicodeString");
	325	+ fnRtlEqualUnicodeString =
	326	+ GetProcAddress(GetModuleHandleA("ntdll"),"RtlEqualUnicodeString");
	327	+ loaded_RtlUnicodeStringAPIs = 1;
	328	+ }
	329	+ if( fnRtlInitUnicodeString && fnRtlEqualUnicodeString ){
	330	+ struct { /* UNICODE_STRING from <ntdef.h> */
	331	+ unsigned short Length;
	332	+ unsigned short MaximumLength;
	333	+ wchar_t *Buffer;
	334	+ } u1, u2;
	335	+ fnRtlInitUnicodeString(&u1,fn1);
	336	+ fnRtlInitUnicodeString(&u2,fn2);
	337	+ return (unsigned char)fnRtlEqualUnicodeString(&u1,&u2,1);
	338	+ }
	339	+ /* In what kind of strange parallel universe are we? */
	340	+ return lstrcmpiW(fn1,fn2)==0;
	341	+}
	342	+
	343	+/* Helper macros to deal with directory separators. */
	344	+#define IS_DIRSEP(s,i) ( s[i]=='/' \|\| s[i]=='\\' )
	345	+#define NEXT_DIRSEP(s,i) while( s[i] && s[i]!='/' && s[i]!='\\' ){i++;}
	346	+
	347	+/* The Win32 version of file_case_preferred_name() from file.c, which is able to
	348	+** find case-preserved file names containing non-ASCII characters. The result is
	349	+** allocated by fossil_malloc() and should be free'd by the caller. While this
	350	+** function usually gets canonicalized paths, it is able to handle any input and
	351	+** figure out more cases than the original:
	352	+**
	353	+** fossil test-case-filename C:/ .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
	354	+** → Original: .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
	355	+** → Modified: .//..\Windows\/.//.\System32\.\notepad.exe
	356	+**
	357	+** md ÄÖÜ
	358	+** fossil test-case-filename ./\ .\äöü\/[empty]\\/
	359	+** → Original: ./äöü\/[empty]\\/
	360	+** → Modified: .\ÄÖÜ\/[empty]\\/
	361	+**
	362	+** The function preserves slashes and backslashes: only single file or directory
	363	+** components without directory separators ("basenames") are converted to UTF-16
	364	+** using fossil_utf8_to_path(), so bypassing its slash ↔ backslash translations.
	365	+** Note that the original function doesn't preserve all slashes and backslashes,
	366	+** for example in the second example above.
	367	+**
	368	+** NOTE: As of Windows 10, version 1803, case sensitivity may be enabled on a
	369	+** per-directory basis, as returned by NtQueryInformationFile() with the file
	370	+** information class FILE_CASE_SENSITIVE_INFORMATION. So this function may be
	371	+** changed to act like fossil_strdup() for files located in such directories.
	372	+*/
	373	+char *win32_file_case_preferred_name(
	374	+ const char *zBase,
	375	+ const char *zPath
	376	+){
	377	+ int cchBase;
	378	+ int cchPath;
	379	+ int cchBuf;
	380	+ int cchRes;
	381	+ char *zBuf;
	382	+ char *zRes;
	383	+ int ncUsed;
	384	+ int i, j;
	385	+ if( filenames_are_case_sensitive() ){
	386	+ return fossil_strdup(zPath);
	387	+ }
	388	+ cchBase = strlen(zBase);
	389	+ cchPath = strlen(zPath);
	390	+ cchBuf = cchBase + cchPath + 2; /* + NULL + optional directory slash */
	391	+ cchRes = cchPath + 1; /* + NULL */
	392	+ zBuf = fossil_malloc(cchBuf);
	393	+ zRes = fossil_malloc(cchRes);
	394	+ ncUsed = 0;
	395	+ memcpy(zBuf,zBase,cchBase);
	396	+ if( !IS_DIRSEP(zBuf,cchBase-1) ){
	397	+ zBuf[cchBase++]=L'/';
	398	+ }
	399	+ memcpy(zBuf+cchBase,zPath,cchPath+1);
	400	+ i = j = cchBase;
	401	+ while( 1 ){
	402	+ WIN32_FIND_DATAW fd;
	403	+ HANDLE hFind;
	404	+ wchar_t *wzBuf;
	405	+ char *zCompBuf = 0;
	406	+ char *zComp = &zBuf[i];
	407	+ int cchComp;
	408	+ char chSep;
	409	+ int fDone;
	410	+ if( IS_DIRSEP(zBuf,i) ){
	411	+ if( ncUsed+2>cchRes ){ /* Directory slash + NULL*/
	412	+ cchRes += 32; /* Overprovisioning. */
	413	+ zRes = fossil_realloc(zRes,cchRes);
	414	+ }
	415	+ zRes[ncUsed++] = zBuf[i];
	416	+ i = j = i+1;
	417	+ continue;
	418	+ }
	419	+ NEXT_DIRSEP(zBuf,j);
	420	+ fDone = zBuf[j]==0;
	421	+ chSep = zBuf[j];
	422	+ zBuf[j] = 0; /* Truncate working buffer. */
	423	+ wzBuf = fossil_utf8_to_path(zBuf,0);
	424	+ hFind = FindFirstFileW(wzBuf,&fd);
	425	+ if( hFind!=INVALID_HANDLE_VALUE ){
	426	+ wchar_t *wzComp = fossil_utf8_to_path(zComp,0);
	427	+ FindClose(hFind);
	428	+ /* Test fd.cFileName, not fd.cAlternateFileName (classic 8.3 format). */
	429	+ if( win32_filenames_equal_nocase(wzComp,fd.cFileName) ){
	430	+ zCompBuf = fossil_path_to_utf8(fd.cFileName);
	431	+ zComp = zCompBuf;
	432	+ }
	433	+ fossil_path_free(wzComp);
	434	+ }
	435	+ fossil_path_free(wzBuf);
	436	+ cchComp = strlen(zComp);
	437	+ if( ncUsed+cchComp+1>cchRes ){ /* Current component + NULL */
	438	+ cchRes += cchComp + 32; /* Overprovisioning. */
	439	+ zRes = fossil_realloc(zRes,cchRes);
	440	+ }
	441	+ memcpy(zRes+ncUsed,zComp,cchComp);
	442	+ ncUsed += cchComp;
	443	+ if( zCompBuf ){
	444	+ fossil_path_free(zCompBuf);
	445	+ }
	446	+ if( fDone ){
	447	+ zRes[ncUsed] = 0;
	448	+ break;
	449	+ }
	450	+ zBuf[j] = chSep; /* Undo working buffer truncation. */
	451	+ i = j;
	452	+ }
	453	+ fossil_free(zBuf);
	454	+ return zRes;
	455	+}
	456	+
	457	+/* Return the unique identifier (UID) for a file, made up of the file identifier
	458	+** (equal to "inode" for Unix-style file systems) plus the volume serial number.
	459	+** Call the GetFileInformationByHandleEx() function on Windows Vista, and resort
	460	+** to the GetFileInformationByHandle() function on Windows XP. The result string
	461	+** is allocated by mprintf(), or NULL on failure.
	462	+*/
	463	+char *win32_file_id(
	464	+ const char *zFileName
	465	+){
	466	+ static FARPROC fnGetFileInformationByHandleEx;
	467	+ static int loaded_fnGetFileInformationByHandleEx;
	468	+ wchar_t *wzFileName = fossil_utf8_to_path(zFileName,0);
	469	+ HANDLE hFile;
	470	+ char *zFileId = 0;
	471	+ hFile = CreateFileW(
	472	+ wzFileName,
	473	+ 0,
	474	+ FILE_SHARE_READ \| FILE_SHARE_WRITE \| FILE_SHARE_DELETE,
	475	+ NULL,
	476	+ OPEN_EXISTING,
	477	+ FILE_FLAG_BACKUP_SEMANTICS,
	478	+ NULL);
	479	+ if( hFile!=INVALID_HANDLE_VALUE ){
	480	+ BY_HANDLE_FILE_INFORMATION fi;
	481	+ struct { /* FILE_ID_INFO from <winbase.h> */
	482	+ u64 VolumeSerialNumber;
	483	+ unsigned char FileId[16];
	484	+ } fi2;
	485	+ if( !loaded_fnGetFileInformationByHandleEx ){
	486	+ fnGetFileInformationByHandleEx = GetProcAddress(
	487	+ GetModuleHandleA("kernel32"),"GetFileInformationByHandleEx");
	488	+ loaded_fnGetFileInformationByHandleEx = 1;
	489	+ }
	490	+ if( fnGetFileInformationByHandleEx ){
	491	+ if( fnGetFileInformationByHandleEx(
	492	+ hFile,/FileIdInfo/0x12,&fi2,sizeof(fi2)) ){
	493	+ zFileId = mprintf(
	494	+ "%016llx/"
	495	+ "%02x%02x%02x%02x%02x%02x%02x%02x"
	496	+ "%02x%02x%02x%02x%02x%02x%02x%02x",
	497	+ fi2.VolumeSerialNumber,
	498	+ fi2.FileId[15], fi2.FileId[14],
	499	+ fi2.FileId[13], fi2.FileId[12],
	500	+ fi2.FileId[11], fi2.FileId[10],
	501	+ fi2.FileId[9], fi2.FileId[8],
	502	+ fi2.FileId[7], fi2.FileId[6],
	503	+ fi2.FileId[5], fi2.FileId[4],
	504	+ fi2.FileId[3], fi2.FileId[2],
	505	+ fi2.FileId[1], fi2.FileId[0]);
	506	+ }
	507	+ }
	508	+ if( zFileId==0 ){
	509	+ if( GetFileInformationByHandle(hFile,&fi) ){
	510	+ ULARGE_INTEGER FileId = {
	511	+ /.LowPart = / fi.nFileIndexLow,
	512	+ /.HighPart = / fi.nFileIndexHigh
	513	+ };
	514	+ zFileId = mprintf(
	515	+ "%08x/%016llx",
	516	+ fi.dwVolumeSerialNumber,(u64)FileId.QuadPart);
	517	+ }
	518	+ }
	519	+ CloseHandle(hFile);
	520	+ }
	521	+ fossil_path_free(wzFileName);
	522	+ return zFileId;
	523	+}
295	524	#endif /* _WIN32 -- This code is for win32 only */
296	525

	--- src/winfile.c
	+++ src/winfile.c
	@@ -290,6 +290,235 @@
290	fossil_free(zWide);
291	for(i=0; zUtf8[i]; i++) if( zUtf8[i]=='\\' ) zUtf8[i] = '/';
292	strncpy(zBuf, zUtf8, nBuf);
293	fossil_path_free(zUtf8);
294	}





































































































































































































































295	#endif /* _WIN32 -- This code is for win32 only */
296

	--- src/winfile.c
	+++ src/winfile.c
	@@ -290,6 +290,235 @@
290	fossil_free(zWide);
291	for(i=0; zUtf8[i]; i++) if( zUtf8[i]=='\\' ) zUtf8[i] = '/';
292	strncpy(zBuf, zUtf8, nBuf);
293	fossil_path_free(zUtf8);
294	}
295
296	/* Perform case-insensitive comparison of two UTF-16 file names. Try to load the
297	** CompareStringOrdinal() function on Windows Vista and newer, and resort to the
298	** RtlEqualUnicodeString() function on Windows XP.
299	** The dance to invoke RtlEqualUnicodeString() is necessary because lstrcmpiW()
300	** performs linguistic comparison, while the former performs binary comparison.
301	** As an example, matching "ß" (U+00DF Latin Small Letter Sharp S) with "ss" is
302	** undesirable in file name comparison, so lstrcmpiW() is only invoked in cases
303	** that are technically impossible and contradicting all known laws of physics.
304	*/
305	int win32_filenames_equal_nocase(
306	const wchar_t *fn1,
307	const wchar_t *fn2
308	){
309	static FARPROC fnCompareStringOrdinal;
310	static FARPROC fnRtlInitUnicodeString;
311	static FARPROC fnRtlEqualUnicodeString;
312	static int loaded_CompareStringOrdinal;
313	static int loaded_RtlUnicodeStringAPIs;
314	if( !loaded_CompareStringOrdinal ){
315	fnCompareStringOrdinal =
316	GetProcAddress(GetModuleHandleA("kernel32"),"CompareStringOrdinal");
317	loaded_CompareStringOrdinal = 1;
318	}
319	if( fnCompareStringOrdinal ){
320	return fnCompareStringOrdinal(fn1,-1,fn2,-1,1)-2==0;
321	}
322	if( !loaded_RtlUnicodeStringAPIs ){
323	fnRtlInitUnicodeString =
324	GetProcAddress(GetModuleHandleA("ntdll"),"RtlInitUnicodeString");
325	fnRtlEqualUnicodeString =
326	GetProcAddress(GetModuleHandleA("ntdll"),"RtlEqualUnicodeString");
327	loaded_RtlUnicodeStringAPIs = 1;
328	}
329	if( fnRtlInitUnicodeString && fnRtlEqualUnicodeString ){
330	struct { /* UNICODE_STRING from <ntdef.h> */
331	unsigned short Length;
332	unsigned short MaximumLength;
333	wchar_t *Buffer;
334	} u1, u2;
335	fnRtlInitUnicodeString(&u1,fn1);
336	fnRtlInitUnicodeString(&u2,fn2);
337	return (unsigned char)fnRtlEqualUnicodeString(&u1,&u2,1);
338	}
339	/* In what kind of strange parallel universe are we? */
340	return lstrcmpiW(fn1,fn2)==0;
341	}
342
343	/* Helper macros to deal with directory separators. */
344	#define IS_DIRSEP(s,i) ( s[i]=='/' \|\| s[i]=='\\' )
345	#define NEXT_DIRSEP(s,i) while( s[i] && s[i]!='/' && s[i]!='\\' ){i++;}
346
347	/* The Win32 version of file_case_preferred_name() from file.c, which is able to
348	** find case-preserved file names containing non-ASCII characters. The result is
349	** allocated by fossil_malloc() and should be free'd by the caller. While this
350	** function usually gets canonicalized paths, it is able to handle any input and
351	** figure out more cases than the original:
352	**
353	** fossil test-case-filename C:/ .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
354	** → Original: .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
355	** → Modified: .//..\Windows\/.//.\System32\.\notepad.exe
356	**
357	** md ÄÖÜ
358	** fossil test-case-filename ./\ .\äöü\/[empty]\\/
359	** → Original: ./äöü\/[empty]\\/
360	** → Modified: .\ÄÖÜ\/[empty]\\/
361	**
362	** The function preserves slashes and backslashes: only single file or directory
363	** components without directory separators ("basenames") are converted to UTF-16
364	** using fossil_utf8_to_path(), so bypassing its slash ↔ backslash translations.
365	** Note that the original function doesn't preserve all slashes and backslashes,
366	** for example in the second example above.
367	**
368	** NOTE: As of Windows 10, version 1803, case sensitivity may be enabled on a
369	** per-directory basis, as returned by NtQueryInformationFile() with the file
370	** information class FILE_CASE_SENSITIVE_INFORMATION. So this function may be
371	** changed to act like fossil_strdup() for files located in such directories.
372	*/
373	char *win32_file_case_preferred_name(
374	const char *zBase,
375	const char *zPath
376	){
377	int cchBase;
378	int cchPath;
379	int cchBuf;
380	int cchRes;
381	char *zBuf;
382	char *zRes;
383	int ncUsed;
384	int i, j;
385	if( filenames_are_case_sensitive() ){
386	return fossil_strdup(zPath);
387	}
388	cchBase = strlen(zBase);
389	cchPath = strlen(zPath);
390	cchBuf = cchBase + cchPath + 2; /* + NULL + optional directory slash */
391	cchRes = cchPath + 1; /* + NULL */
392	zBuf = fossil_malloc(cchBuf);
393	zRes = fossil_malloc(cchRes);
394	ncUsed = 0;
395	memcpy(zBuf,zBase,cchBase);
396	if( !IS_DIRSEP(zBuf,cchBase-1) ){
397	zBuf[cchBase++]=L'/';
398	}
399	memcpy(zBuf+cchBase,zPath,cchPath+1);
400	i = j = cchBase;
401	while( 1 ){
402	WIN32_FIND_DATAW fd;
403	HANDLE hFind;
404	wchar_t *wzBuf;
405	char *zCompBuf = 0;
406	char *zComp = &zBuf[i];
407	int cchComp;
408	char chSep;
409	int fDone;
410	if( IS_DIRSEP(zBuf,i) ){
411	if( ncUsed+2>cchRes ){ /* Directory slash + NULL*/
412	cchRes += 32; /* Overprovisioning. */
413	zRes = fossil_realloc(zRes,cchRes);
414	}
415	zRes[ncUsed++] = zBuf[i];
416	i = j = i+1;
417	continue;
418	}
419	NEXT_DIRSEP(zBuf,j);
420	fDone = zBuf[j]==0;
421	chSep = zBuf[j];
422	zBuf[j] = 0; /* Truncate working buffer. */
423	wzBuf = fossil_utf8_to_path(zBuf,0);
424	hFind = FindFirstFileW(wzBuf,&fd);
425	if( hFind!=INVALID_HANDLE_VALUE ){
426	wchar_t *wzComp = fossil_utf8_to_path(zComp,0);
427	FindClose(hFind);
428	/* Test fd.cFileName, not fd.cAlternateFileName (classic 8.3 format). */
429	if( win32_filenames_equal_nocase(wzComp,fd.cFileName) ){
430	zCompBuf = fossil_path_to_utf8(fd.cFileName);
431	zComp = zCompBuf;
432	}
433	fossil_path_free(wzComp);
434	}
435	fossil_path_free(wzBuf);
436	cchComp = strlen(zComp);
437	if( ncUsed+cchComp+1>cchRes ){ /* Current component + NULL */
438	cchRes += cchComp + 32; /* Overprovisioning. */
439	zRes = fossil_realloc(zRes,cchRes);
440	}
441	memcpy(zRes+ncUsed,zComp,cchComp);
442	ncUsed += cchComp;
443	if( zCompBuf ){
444	fossil_path_free(zCompBuf);
445	}
446	if( fDone ){
447	zRes[ncUsed] = 0;
448	break;
449	}
450	zBuf[j] = chSep; /* Undo working buffer truncation. */
451	i = j;
452	}
453	fossil_free(zBuf);
454	return zRes;
455	}
456
457	/* Return the unique identifier (UID) for a file, made up of the file identifier
458	** (equal to "inode" for Unix-style file systems) plus the volume serial number.
459	** Call the GetFileInformationByHandleEx() function on Windows Vista, and resort
460	** to the GetFileInformationByHandle() function on Windows XP. The result string
461	** is allocated by mprintf(), or NULL on failure.
462	*/
463	char *win32_file_id(
464	const char *zFileName
465	){
466	static FARPROC fnGetFileInformationByHandleEx;
467	static int loaded_fnGetFileInformationByHandleEx;
468	wchar_t *wzFileName = fossil_utf8_to_path(zFileName,0);
469	HANDLE hFile;
470	char *zFileId = 0;
471	hFile = CreateFileW(
472	wzFileName,
473	0,
474	FILE_SHARE_READ \| FILE_SHARE_WRITE \| FILE_SHARE_DELETE,
475	NULL,
476	OPEN_EXISTING,
477	FILE_FLAG_BACKUP_SEMANTICS,
478	NULL);
479	if( hFile!=INVALID_HANDLE_VALUE ){
480	BY_HANDLE_FILE_INFORMATION fi;
481	struct { /* FILE_ID_INFO from <winbase.h> */
482	u64 VolumeSerialNumber;
483	unsigned char FileId[16];
484	} fi2;
485	if( !loaded_fnGetFileInformationByHandleEx ){
486	fnGetFileInformationByHandleEx = GetProcAddress(
487	GetModuleHandleA("kernel32"),"GetFileInformationByHandleEx");
488	loaded_fnGetFileInformationByHandleEx = 1;
489	}
490	if( fnGetFileInformationByHandleEx ){
491	if( fnGetFileInformationByHandleEx(
492	hFile,/FileIdInfo/0x12,&fi2,sizeof(fi2)) ){
493	zFileId = mprintf(
494	"%016llx/"
495	"%02x%02x%02x%02x%02x%02x%02x%02x"
496	"%02x%02x%02x%02x%02x%02x%02x%02x",
497	fi2.VolumeSerialNumber,
498	fi2.FileId[15], fi2.FileId[14],
499	fi2.FileId[13], fi2.FileId[12],
500	fi2.FileId[11], fi2.FileId[10],
501	fi2.FileId[9], fi2.FileId[8],
502	fi2.FileId[7], fi2.FileId[6],
503	fi2.FileId[5], fi2.FileId[4],
504	fi2.FileId[3], fi2.FileId[2],
505	fi2.FileId[1], fi2.FileId[0]);
506	}
507	}
508	if( zFileId==0 ){
509	if( GetFileInformationByHandle(hFile,&fi) ){
510	ULARGE_INTEGER FileId = {
511	/.LowPart = / fi.nFileIndexLow,
512	/.HighPart = / fi.nFileIndexHigh
513	};
514	zFileId = mprintf(
515	"%08x/%016llx",
516	fi.dwVolumeSerialNumber,(u64)FileId.QuadPart);
517	}
518	}
519	CloseHandle(hFile);
520	}
521	fossil_path_free(wzFileName);
522	return zFileId;
523	}
524	#endif /* _WIN32 -- This code is for win32 only */
525

M src/xfer.c

+40 -1

		--- src/xfer.c
		+++ src/xfer.c
		@@ -336,10 +336,11 @@
336	336	\|\| !blob_is_filename(&pXfer->aToken[1])
337	337	\|\| !blob_is_int64(&pXfer->aToken[2], &mtime)
338	338	\|\| (!blob_eq(pHash,"-") && !blob_is_hname(pHash))
339	339	\|\| !blob_is_int(&pXfer->aToken[4], &sz)
340	340	\|\| !blob_is_int(&pXfer->aToken[5], &flags)
	341	+ \|\| (mtime<0 \|\| sz<0 \|\| flags<0)
341	342	){
342	343	blob_appendf(&pXfer->err, "malformed uvfile line");
343	344	return;
344	345	}
345	346	blob_init(&content, 0, 0);
		@@ -1119,10 +1120,24 @@
1119	1120	** Return the TH1 code to evaluate when a ticket change is processed.
1120	1121	*/
1121	1122	const char *xfer_ticket_code(void){
1122	1123	return db_get("xfer-ticket-script", 0);
1123	1124	}
	1125	+
	1126	+/*
	1127	+** Reset the CGI content, roll back any pending db transaction, and
	1128	+** emit an "error" xfer message. The message text gets fossil-encoded
	1129	+** by this function. This is only intended for use with
	1130	+** fail-fast/fatal errors, not ones which can be skipped over.
	1131	+*/
	1132	+static void xfer_fatal_error(const char *zMsg){
	1133	+ cgi_reset_content();
	1134	+ if( db_transaction_nesting_depth()>0 ){
	1135	+ db_rollback_transaction();
	1136	+ }
	1137	+ @ error %F(zMsg)
	1138	+}
1124	1139
1125	1140	/*
1126	1141	** Run the specified TH1 script, if any, and returns 1 on error.
1127	1142	*/
1128	1143	int xfer_run_script(
		@@ -1461,10 +1476,14 @@
1461	1476	int seqno, max;
1462	1477	if( iVers>=3 ){
1463	1478	cgi_set_content_type("application/x-fossil-uncompressed");
1464	1479	}
1465	1480	blob_is_int(&xfer.aToken[2], &seqno);
	1481	+ if( seqno<=0 ){
	1482	+ xfer_fatal_error("invalid clone sequence number");
	1483	+ return;
	1484	+ }
1466	1485	max = db_int(0, "SELECT max(rid) FROM blob");
1467	1486	while( xfer.mxSend>(int)blob_size(xfer.pOut) && seqno<=max){
1468	1487	if( time(NULL) >= xfer.maxTime ) break;
1469	1488	if( iVers>=3 ){
1470	1489	send_compressed_file(&xfer, seqno);
		@@ -1532,10 +1551,14 @@
1532	1551	*/
1533	1552	if( blob_eq(&xfer.aToken[0],"config") && xfer.nToken==3
1534	1553	&& blob_is_int(&xfer.aToken[2], &size) ){
1535	1554	const char *zName = blob_str(&xfer.aToken[1]);
1536	1555	Blob content;
	1556	+ if( size<0 ){
	1557	+ xfer_fatal_error("invalid config record");
	1558	+ return;
	1559	+ }
1537	1560	blob_zero(&content);
1538	1561	blob_extract(xfer.pIn, size, &content);
1539	1562	if( !g.perm.Admin ){
1540	1563	cgi_reset_content();
1541	1564	@ error not\sauthorized\sto\spush\sconfiguration
		@@ -1844,11 +1867,11 @@
1844	1867	/* Send the server timestamp last, in case prior processing happened
1845	1868	** to use up a significant fraction of our time window.
1846	1869	*/
1847	1870	zNow = db_text(0, "SELECT strftime('%%Y-%%m-%%dT%%H:%%M:%%S', 'now')");
1848	1871	@ # timestamp %s(zNow) errors %d(nErr)
1849		- free(zNow);
	1872	+ fossil_free(zNow);
1850	1873
1851	1874	db_commit_transaction();
1852	1875	configure_rebuild();
1853	1876	}
1854	1877
		@@ -2487,10 +2510,14 @@
2487	2510	&& (blob_eq(&xfer.aToken[3],"-") \|\| blob_is_hname(&xfer.aToken[3]))
2488	2511	){
2489	2512	const char *zName = blob_str(&xfer.aToken[1]);
2490	2513	const char *zHash = blob_str(&xfer.aToken[3]);
2491	2514	int iStatus;
	2515	+ if( mtime<0 \|\| size<0 ){
	2516	+ xfer_fatal_error("invalid uvigot");
	2517	+ return ++nErr;
	2518	+ }
2492	2519	iStatus = unversioned_status(zName, mtime, zHash);
2493	2520	if( (syncFlags & SYNC_UV_REVERT)!=0 ){
2494	2521	if( iStatus==4 ) iStatus = 2;
2495	2522	if( iStatus==5 ) iStatus = 1;
2496	2523	}
		@@ -2573,10 +2600,14 @@
2573	2600	*/
2574	2601	if( blob_eq(&xfer.aToken[0],"config") && xfer.nToken==3
2575	2602	&& blob_is_int(&xfer.aToken[2], &size) ){
2576	2603	const char *zName = blob_str(&xfer.aToken[1]);
2577	2604	Blob content;
	2605	+ if( size<0 ){
	2606	+ xfer_fatal_error("invalid config record");
	2607	+ return ++nErr;
	2608	+ }
2578	2609	blob_zero(&content);
2579	2610	blob_extract(xfer.pIn, size, &content);
2580	2611	g.perm.Admin = g.perm.RdAddr = 1;
2581	2612	configure_receive(zName, &content, origConfigRcvMask);
2582	2613	nCardRcvd++;
		@@ -2617,10 +2648,14 @@
2617	2648	** blob that needs to be sent. If N<=0 that indicates that all blobs
2618	2649	** have been sent.
2619	2650	*/
2620	2651	if( blob_eq(&xfer.aToken[0], "clone_seqno") && xfer.nToken==2 ){
2621	2652	blob_is_int(&xfer.aToken[1], &cloneSeqno);
	2653	+ if( cloneSeqno<0 ){
	2654	+ xfer_fatal_error("invalid clone_seqno");
	2655	+ return ++nErr;
	2656	+ }
2622	2657	}else
2623	2658
2624	2659	/* message MESSAGE
2625	2660	**
2626	2661	** A message is received from the server. Print it.
		@@ -2694,10 +2729,14 @@
2694	2729	char *zUser = blob_terminate(&xfer.aToken[2]);
2695	2730	sqlite3_int64 mtime, iNow;
2696	2731	defossilize(zUser);
2697	2732	iNow = time(NULL);
2698	2733	if( blob_is_int64(&xfer.aToken[3], &mtime) && iNow>mtime ){
	2734	+ if( mtime<0 ){
	2735	+ xfer_fatal_error("invalid ci-lock-fail time");
	2736	+ return ++nErr;
	2737	+ }
2699	2738	iNow = time(NULL);
2700	2739	fossil_print("\nParent check-in locked by %s %s ago\n",
2701	2740	zUser, human_readable_age((iNow+1-mtime)/86400.0));
2702	2741	}else{
2703	2742	fossil_print("\nParent check-in locked by %s\n", zUser);
2704	2743

	--- src/xfer.c
	+++ src/xfer.c
	@@ -336,10 +336,11 @@
336	\|\| !blob_is_filename(&pXfer->aToken[1])
337	\|\| !blob_is_int64(&pXfer->aToken[2], &mtime)
338	\|\| (!blob_eq(pHash,"-") && !blob_is_hname(pHash))
339	\|\| !blob_is_int(&pXfer->aToken[4], &sz)
340	\|\| !blob_is_int(&pXfer->aToken[5], &flags)

341	){
342	blob_appendf(&pXfer->err, "malformed uvfile line");
343	return;
344	}
345	blob_init(&content, 0, 0);
	@@ -1119,10 +1120,24 @@
1119	** Return the TH1 code to evaluate when a ticket change is processed.
1120	*/
1121	const char *xfer_ticket_code(void){
1122	return db_get("xfer-ticket-script", 0);
1123	}














1124
1125	/*
1126	** Run the specified TH1 script, if any, and returns 1 on error.
1127	*/
1128	int xfer_run_script(
	@@ -1461,10 +1476,14 @@
1461	int seqno, max;
1462	if( iVers>=3 ){
1463	cgi_set_content_type("application/x-fossil-uncompressed");
1464	}
1465	blob_is_int(&xfer.aToken[2], &seqno);




1466	max = db_int(0, "SELECT max(rid) FROM blob");
1467	while( xfer.mxSend>(int)blob_size(xfer.pOut) && seqno<=max){
1468	if( time(NULL) >= xfer.maxTime ) break;
1469	if( iVers>=3 ){
1470	send_compressed_file(&xfer, seqno);
	@@ -1532,10 +1551,14 @@
1532	*/
1533	if( blob_eq(&xfer.aToken[0],"config") && xfer.nToken==3
1534	&& blob_is_int(&xfer.aToken[2], &size) ){
1535	const char *zName = blob_str(&xfer.aToken[1]);
1536	Blob content;




1537	blob_zero(&content);
1538	blob_extract(xfer.pIn, size, &content);
1539	if( !g.perm.Admin ){
1540	cgi_reset_content();
1541	@ error not\sauthorized\sto\spush\sconfiguration
	@@ -1844,11 +1867,11 @@
1844	/* Send the server timestamp last, in case prior processing happened
1845	** to use up a significant fraction of our time window.
1846	*/
1847	zNow = db_text(0, "SELECT strftime('%%Y-%%m-%%dT%%H:%%M:%%S', 'now')");
1848	@ # timestamp %s(zNow) errors %d(nErr)
1849	free(zNow);
1850
1851	db_commit_transaction();
1852	configure_rebuild();
1853	}
1854
	@@ -2487,10 +2510,14 @@
2487	&& (blob_eq(&xfer.aToken[3],"-") \|\| blob_is_hname(&xfer.aToken[3]))
2488	){
2489	const char *zName = blob_str(&xfer.aToken[1]);
2490	const char *zHash = blob_str(&xfer.aToken[3]);
2491	int iStatus;




2492	iStatus = unversioned_status(zName, mtime, zHash);
2493	if( (syncFlags & SYNC_UV_REVERT)!=0 ){
2494	if( iStatus==4 ) iStatus = 2;
2495	if( iStatus==5 ) iStatus = 1;
2496	}
	@@ -2573,10 +2600,14 @@
2573	*/
2574	if( blob_eq(&xfer.aToken[0],"config") && xfer.nToken==3
2575	&& blob_is_int(&xfer.aToken[2], &size) ){
2576	const char *zName = blob_str(&xfer.aToken[1]);
2577	Blob content;




2578	blob_zero(&content);
2579	blob_extract(xfer.pIn, size, &content);
2580	g.perm.Admin = g.perm.RdAddr = 1;
2581	configure_receive(zName, &content, origConfigRcvMask);
2582	nCardRcvd++;
	@@ -2617,10 +2648,14 @@
2617	** blob that needs to be sent. If N<=0 that indicates that all blobs
2618	** have been sent.
2619	*/
2620	if( blob_eq(&xfer.aToken[0], "clone_seqno") && xfer.nToken==2 ){
2621	blob_is_int(&xfer.aToken[1], &cloneSeqno);




2622	}else
2623
2624	/* message MESSAGE
2625	**
2626	** A message is received from the server. Print it.
	@@ -2694,10 +2729,14 @@
2694	char *zUser = blob_terminate(&xfer.aToken[2]);
2695	sqlite3_int64 mtime, iNow;
2696	defossilize(zUser);
2697	iNow = time(NULL);
2698	if( blob_is_int64(&xfer.aToken[3], &mtime) && iNow>mtime ){




2699	iNow = time(NULL);
2700	fossil_print("\nParent check-in locked by %s %s ago\n",
2701	zUser, human_readable_age((iNow+1-mtime)/86400.0));
2702	}else{
2703	fossil_print("\nParent check-in locked by %s\n", zUser);
2704

	--- src/xfer.c
	+++ src/xfer.c
	@@ -336,10 +336,11 @@
336	\|\| !blob_is_filename(&pXfer->aToken[1])
337	\|\| !blob_is_int64(&pXfer->aToken[2], &mtime)
338	\|\| (!blob_eq(pHash,"-") && !blob_is_hname(pHash))
339	\|\| !blob_is_int(&pXfer->aToken[4], &sz)
340	\|\| !blob_is_int(&pXfer->aToken[5], &flags)
341	\|\| (mtime<0 \|\| sz<0 \|\| flags<0)
342	){
343	blob_appendf(&pXfer->err, "malformed uvfile line");
344	return;
345	}
346	blob_init(&content, 0, 0);
	@@ -1119,10 +1120,24 @@
1120	** Return the TH1 code to evaluate when a ticket change is processed.
1121	*/
1122	const char *xfer_ticket_code(void){
1123	return db_get("xfer-ticket-script", 0);
1124	}
1125
1126	/*
1127	** Reset the CGI content, roll back any pending db transaction, and
1128	** emit an "error" xfer message. The message text gets fossil-encoded
1129	** by this function. This is only intended for use with
1130	** fail-fast/fatal errors, not ones which can be skipped over.
1131	*/
1132	static void xfer_fatal_error(const char *zMsg){
1133	cgi_reset_content();
1134	if( db_transaction_nesting_depth()>0 ){
1135	db_rollback_transaction();
1136	}
1137	@ error %F(zMsg)
1138	}
1139
1140	/*
1141	** Run the specified TH1 script, if any, and returns 1 on error.
1142	*/
1143	int xfer_run_script(
	@@ -1461,10 +1476,14 @@
1476	int seqno, max;
1477	if( iVers>=3 ){
1478	cgi_set_content_type("application/x-fossil-uncompressed");
1479	}
1480	blob_is_int(&xfer.aToken[2], &seqno);
1481	if( seqno<=0 ){
1482	xfer_fatal_error("invalid clone sequence number");
1483	return;
1484	}
1485	max = db_int(0, "SELECT max(rid) FROM blob");
1486	while( xfer.mxSend>(int)blob_size(xfer.pOut) && seqno<=max){
1487	if( time(NULL) >= xfer.maxTime ) break;
1488	if( iVers>=3 ){
1489	send_compressed_file(&xfer, seqno);
	@@ -1532,10 +1551,14 @@
1551	*/
1552	if( blob_eq(&xfer.aToken[0],"config") && xfer.nToken==3
1553	&& blob_is_int(&xfer.aToken[2], &size) ){
1554	const char *zName = blob_str(&xfer.aToken[1]);
1555	Blob content;
1556	if( size<0 ){
1557	xfer_fatal_error("invalid config record");
1558	return;
1559	}
1560	blob_zero(&content);
1561	blob_extract(xfer.pIn, size, &content);
1562	if( !g.perm.Admin ){
1563	cgi_reset_content();
1564	@ error not\sauthorized\sto\spush\sconfiguration
	@@ -1844,11 +1867,11 @@
1867	/* Send the server timestamp last, in case prior processing happened
1868	** to use up a significant fraction of our time window.
1869	*/
1870	zNow = db_text(0, "SELECT strftime('%%Y-%%m-%%dT%%H:%%M:%%S', 'now')");
1871	@ # timestamp %s(zNow) errors %d(nErr)
1872	fossil_free(zNow);
1873
1874	db_commit_transaction();
1875	configure_rebuild();
1876	}
1877
	@@ -2487,10 +2510,14 @@
2510	&& (blob_eq(&xfer.aToken[3],"-") \|\| blob_is_hname(&xfer.aToken[3]))
2511	){
2512	const char *zName = blob_str(&xfer.aToken[1]);
2513	const char *zHash = blob_str(&xfer.aToken[3]);
2514	int iStatus;
2515	if( mtime<0 \|\| size<0 ){
2516	xfer_fatal_error("invalid uvigot");
2517	return ++nErr;
2518	}
2519	iStatus = unversioned_status(zName, mtime, zHash);
2520	if( (syncFlags & SYNC_UV_REVERT)!=0 ){
2521	if( iStatus==4 ) iStatus = 2;
2522	if( iStatus==5 ) iStatus = 1;
2523	}
	@@ -2573,10 +2600,14 @@
2600	*/
2601	if( blob_eq(&xfer.aToken[0],"config") && xfer.nToken==3
2602	&& blob_is_int(&xfer.aToken[2], &size) ){
2603	const char *zName = blob_str(&xfer.aToken[1]);
2604	Blob content;
2605	if( size<0 ){
2606	xfer_fatal_error("invalid config record");
2607	return ++nErr;
2608	}
2609	blob_zero(&content);
2610	blob_extract(xfer.pIn, size, &content);
2611	g.perm.Admin = g.perm.RdAddr = 1;
2612	configure_receive(zName, &content, origConfigRcvMask);
2613	nCardRcvd++;
	@@ -2617,10 +2648,14 @@
2648	** blob that needs to be sent. If N<=0 that indicates that all blobs
2649	** have been sent.
2650	*/
2651	if( blob_eq(&xfer.aToken[0], "clone_seqno") && xfer.nToken==2 ){
2652	blob_is_int(&xfer.aToken[1], &cloneSeqno);
2653	if( cloneSeqno<0 ){
2654	xfer_fatal_error("invalid clone_seqno");
2655	return ++nErr;
2656	}
2657	}else
2658
2659	/* message MESSAGE
2660	**
2661	** A message is received from the server. Print it.
	@@ -2694,10 +2729,14 @@
2729	char *zUser = blob_terminate(&xfer.aToken[2]);
2730	sqlite3_int64 mtime, iNow;
2731	defossilize(zUser);
2732	iNow = time(NULL);
2733	if( blob_is_int64(&xfer.aToken[3], &mtime) && iNow>mtime ){
2734	if( mtime<0 ){
2735	xfer_fatal_error("invalid ci-lock-fail time");
2736	return ++nErr;
2737	}
2738	iNow = time(NULL);
2739	fossil_print("\nParent check-in locked by %s %s ago\n",
2740	zUser, human_readable_age((iNow+1-mtime)/86400.0));
2741	}else{
2742	fossil_print("\nParent check-in locked by %s\n", zUser);
2743

M test/comment.test

+20

		--- test/comment.test
		+++ test/comment.test
		@@ -319,8 +319,28 @@
319	319	###############################################################################
320	320
321	321	fossil test-comment-format --width 81 --indent 9 --decode --trimcrlf --origbreak "00:00:00 " "\[0000000000\] CURRENT $orig" $orig
322	322	test comment-60 {$RESULT eq "00:00:00 \[0000000000\] CURRENT \n xxxx xx xxxxxxx xxxx xxxxxx xxxxxxx, xxxxxxx, x xxxx xxxxxx xx xxxx xxxx\n xxxxxxx xxxxx xxxx xxxx xx xxxxxxx xxxxxxx (xxxxxx xxxxxxxxx x xxxxx).\n xxx'x xxx xxx xx xxxxx xxxx xxx xxx --xxxxxxxxxxx xxxxxx xx xx xxxx. x\n xxxxx x xxxxxx xxxx xxxx xxxx xxxx xxxx x xxxxx xx xxx x xxxxxxxx\n xxxxxxx.\n(6 lines output)"}
323	323
	324	+###############################################################################
	325	+
	326	+fossil test-comment-format --width 72 --file "" [file join $testdir "utf8-comment.txt"]
	327	+test comment-61 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\] an\nd symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
	328	+
	329	+###############################################################################
	330	+
	331	+fossil test-comment-format --width 72 --wordbreak --file "" [file join $testdir "utf8-comment.txt"]
	332	+test comment-62 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\]\nand symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
	333	+
	334	+###############################################################################
	335	+
	336	+fossil test-comment-format --width 72 --legacy --file "" [file join $testdir "utf8-comment.txt"]
	337	+test comment-63 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\]\nand symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
	338	+
	339	+###############################################################################
	340	+
	341	+fossil test-comment-format --width 72 --legacy --wordbreak --file "" [file join $testdir "utf8-comment.txt"]
	342	+test comment-64 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\]\nand symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
	343	+
324	344	###############################################################################
325	345
326	346	test_cleanup
327	347

	--- test/comment.test
	+++ test/comment.test
	@@ -319,8 +319,28 @@
319	###############################################################################
320
321	fossil test-comment-format --width 81 --indent 9 --decode --trimcrlf --origbreak "00:00:00 " "\[0000000000\] CURRENT $orig" $orig
322	test comment-60 {$RESULT eq "00:00:00 \[0000000000\] CURRENT \n xxxx xx xxxxxxx xxxx xxxxxx xxxxxxx, xxxxxxx, x xxxx xxxxxx xx xxxx xxxx\n xxxxxxx xxxxx xxxx xxxx xx xxxxxxx xxxxxxx (xxxxxx xxxxxxxxx x xxxxx).\n xxx'x xxx xxx xx xxxxx xxxx xxx xxx --xxxxxxxxxxx xxxxxx xx xx xxxx. x\n xxxxx x xxxxxx xxxx xxxx xxxx xxxx xxxx x xxxxx xx xxx x xxxxxxxx\n xxxxxxx.\n(6 lines output)"}
323




















324	###############################################################################
325
326	test_cleanup
327

	--- test/comment.test
	+++ test/comment.test
	@@ -319,8 +319,28 @@
319	###############################################################################
320
321	fossil test-comment-format --width 81 --indent 9 --decode --trimcrlf --origbreak "00:00:00 " "\[0000000000\] CURRENT $orig" $orig
322	test comment-60 {$RESULT eq "00:00:00 \[0000000000\] CURRENT \n xxxx xx xxxxxxx xxxx xxxxxx xxxxxxx, xxxxxxx, x xxxx xxxxxx xx xxxx xxxx\n xxxxxxx xxxxx xxxx xxxx xx xxxxxxx xxxxxxx (xxxxxx xxxxxxxxx x xxxxx).\n xxx'x xxx xxx xx xxxxx xxxx xxx xxx --xxxxxxxxxxx xxxxxx xx xx xxxx. x\n xxxxx x xxxxxx xxxx xxxx xxxx xxxx xxxx x xxxxx xx xxx x xxxxxxxx\n xxxxxxx.\n(6 lines output)"}
323
324	###############################################################################
325
326	fossil test-comment-format --width 72 --file "" [file join $testdir "utf8-comment.txt"]
327	test comment-61 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\] an\nd symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
328
329	###############################################################################
330
331	fossil test-comment-format --width 72 --wordbreak --file "" [file join $testdir "utf8-comment.txt"]
332	test comment-62 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\]\nand symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
333
334	###############################################################################
335
336	fossil test-comment-format --width 72 --legacy --file "" [file join $testdir "utf8-comment.txt"]
337	test comment-63 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\]\nand symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
338
339	###############################################################################
340
341	fossil test-comment-format --width 72 --legacy --wordbreak --file "" [file join $testdir "utf8-comment.txt"]
342	test comment-64 {$RESULT eq "The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte \[äöü\]\nand symbols \[☃\] and emoji \[💾\] characters!\n(2 lines output)"}
343
344	###############################################################################
345
346	test_cleanup
347

M test/glob.test

+20 -20

		--- test/glob.test
		+++ test/glob.test
		@@ -30,40 +30,40 @@
30	30	}
31	31
32	32	glob-parse 100 test test [string map [list \r\n \n] \
33	33	{SQL expression: (x GLOB 'test')
34	34	pattern[0] = [test]
35		-1 test}]
	35	+1 1 test}]
36	36
37	37	glob-parse 101 "one two" one [string map [list \r\n \n] \
38	38	{SQL expression: (x GLOB 'one' OR x GLOB 'two')
39	39	pattern[0] = [one]
40	40	pattern[1] = [two]
41		-1 one}]
	41	+1 1 one}]
42	42
43	43	glob-parse 102 t* test [string map [list \r\n \n] \
44	44	{SQL expression: (x GLOB 't*')
45	45	pattern[0] = [t*]
46		-1 test}]
	46	+1 1 test}]
47	47
48	48	glob-parse 103 "o* two" one [string map [list \r\n \n] \
49	49	{SQL expression: (x GLOB 'o*' OR x GLOB 'two')
50	50	pattern[0] = [o*]
51	51	pattern[1] = [two]
52		-1 one}]
	52	+1 1 one}]
53	53
54	54	glob-parse 104 {"o* two" "three four"} "one two" [string map [list \r\n \n] \
55	55	{SQL expression: (x GLOB 'o* two' OR x GLOB 'three four')
56	56	pattern[0] = [o* two]
57	57	pattern[1] = [three four]
58		-1 one two}]
	58	+1 1 one two}]
59	59
60	60	glob-parse 105 {"o* two" "three four"} "two one" [string map [list \r\n \n] \
61	61	{SQL expression: (x GLOB 'o* two' OR x GLOB 'three four')
62	62	pattern[0] = [o* two]
63	63	pattern[1] = [three four]
64		-0 two one}]
	64	+0 0 two one}]
65	65
66	66	glob-parse 106 "\"o*\ntwo\" \"three\nfour\"" "one\ntwo" \
67	67	[string map [list \r\n \n] \
68	68	{SQL expression: (x GLOB 'o*
69	69	two' OR x GLOB 'three
		@@ -70,11 +70,11 @@
70	70	four')
71	71	pattern[0] = [o*
72	72	two]
73	73	pattern[1] = [three
74	74	four]
75		-1 one
	75	+1 1 one
76	76	two}]
77	77
78	78	glob-parse 107 "\"o*\ntwo\" \"three\nfour\"" "two\none" \
79	79	[string map [list \r\n \n] \
80	80	{SQL expression: (x GLOB 'o*
		@@ -82,11 +82,11 @@
82	82	four')
83	83	pattern[0] = [o*
84	84	two]
85	85	pattern[1] = [three
86	86	four]
87		-0 two
	87	+0 0 two
88	88	one}]
89	89
90	90	glob-parse 108 "\"o*\rtwo\" \"three\rfour\"" "one\rtwo" \
91	91	[string map [list \r\n \n] \
92	92	{SQL expression: (x GLOB 'o*
		@@ -94,11 +94,11 @@
94	94	four')
95	95	pattern[0] = [o*
96	96	two]
97	97	pattern[1] = [three
98	98	four]
99		-1 one
	99	+1 1 one
100	100	two}]
101	101
102	102	glob-parse 109 "\"o*\rtwo\" \"three\rfour\"" "two\rone" \
103	103	[string map [list \r\n \n] \
104	104	{SQL expression: (x GLOB 'o*
		@@ -106,11 +106,11 @@
106	106	four')
107	107	pattern[0] = [o*
108	108	two]
109	109	pattern[1] = [three
110	110	four]
111		-0 two
	111	+0 0 two
112	112	one}]
113	113
114	114	glob-parse 110 "'o*\ntwo' 'three\nfour'" "one\ntwo" \
115	115	[string map [list \r\n \n] \
116	116	{SQL expression: (x GLOB 'o*
		@@ -118,11 +118,11 @@
118	118	four')
119	119	pattern[0] = [o*
120	120	two]
121	121	pattern[1] = [three
122	122	four]
123		-1 one
	123	+1 1 one
124	124	two}]
125	125
126	126	glob-parse 111 "'o*\ntwo' 'three\nfour'" "two\none" \
127	127	[string map [list \r\n \n] \
128	128	{SQL expression: (x GLOB 'o*
		@@ -130,61 +130,61 @@
130	130	four')
131	131	pattern[0] = [o*
132	132	two]
133	133	pattern[1] = [three
134	134	four]
135		-0 two
	135	+0 0 two
136	136	one}]
137	137
138	138	glob-parse 112 "\"'o*' 'two'\" \"'three' 'four'\"" "'one' 'two'" \
139	139	[string map [list \r\n \n] \
140	140	{SQL expression: (x GLOB '''o*'' ''two''' OR x GLOB '''three'' ''four''')
141	141	pattern[0] = ['o*' 'two']
142	142	pattern[1] = ['three' 'four']
143		-1 'one' 'two'}]
	143	+1 1 'one' 'two'}]
144	144
145	145	glob-parse 113 "\"'o*' 'two'\" \"'three' 'four'\"" "two one" \
146	146	[string map [list \r\n \n] \
147	147	{SQL expression: (x GLOB '''o*'' ''two''' OR x GLOB '''three'' ''four''')
148	148	pattern[0] = ['o*' 'two']
149	149	pattern[1] = ['three' 'four']
150		-0 two one}]
	150	+0 0 two one}]
151	151
152	152	glob-parse 114 o*,two one [string map [list \r\n \n] \
153	153	{SQL expression: (x GLOB 'o*' OR x GLOB 'two')
154	154	pattern[0] = [o*]
155	155	pattern[1] = [two]
156		-1 one}]
	156	+1 1 one}]
157	157
158	158	glob-parse 115 "o*,two three,four" "one two" [string map [list \r\n \n] \
159	159	{SQL expression: (x GLOB 'o*' OR x GLOB 'two' OR x GLOB 'three' OR x GLOB 'four')
160	160	pattern[0] = [o*]
161	161	pattern[1] = [two]
162	162	pattern[2] = [three]
163	163	pattern[3] = [four]
164		-1 one two}]
	164	+1 1 one two}]
165	165
166	166	glob-parse 116 'o*,two' one [string map [list \r\n \n] \
167	167	{SQL expression: (x GLOB 'o*,two')
168	168	pattern[0] = [o*,two]
169		-0 one}]
	169	+0 0 one}]
170	170
171	171	glob-parse 117 'o*,two' one,two [string map [list \r\n \n] \
172	172	{SQL expression: (x GLOB 'o*,two')
173	173	pattern[0] = [o*,two]
174		-1 one,two}]
	174	+1 1 one,two}]
175	175
176	176	glob-parse 118 "'o*,two three,four'" "one two three,four" \
177	177	[string map [list \r\n \n] \
178	178	{SQL expression: (x GLOB 'o*,two three,four')
179	179	pattern[0] = [o*,two three,four]
180		-0 one two three,four}]
	180	+0 0 one two three,four}]
181	181
182	182	glob-parse 119 "'o*,two three,four'" "one,two three,four" \
183	183	[string map [list \r\n \n] \
184	184	{SQL expression: (x GLOB 'o*,two three,four')
185	185	pattern[0] = [o*,two three,four]
186		-1 one,two three,four}]
	186	+1 1 one,two three,four}]
187	187
188	188	###############################################################################
189	189
190	190	test_cleanup
191	191

	--- test/glob.test
	+++ test/glob.test
	@@ -30,40 +30,40 @@
30	}
31
32	glob-parse 100 test test [string map [list \r\n \n] \
33	{SQL expression: (x GLOB 'test')
34	pattern[0] = [test]
35	1 test}]
36
37	glob-parse 101 "one two" one [string map [list \r\n \n] \
38	{SQL expression: (x GLOB 'one' OR x GLOB 'two')
39	pattern[0] = [one]
40	pattern[1] = [two]
41	1 one}]
42
43	glob-parse 102 t* test [string map [list \r\n \n] \
44	{SQL expression: (x GLOB 't*')
45	pattern[0] = [t*]
46	1 test}]
47
48	glob-parse 103 "o* two" one [string map [list \r\n \n] \
49	{SQL expression: (x GLOB 'o*' OR x GLOB 'two')
50	pattern[0] = [o*]
51	pattern[1] = [two]
52	1 one}]
53
54	glob-parse 104 {"o* two" "three four"} "one two" [string map [list \r\n \n] \
55	{SQL expression: (x GLOB 'o* two' OR x GLOB 'three four')
56	pattern[0] = [o* two]
57	pattern[1] = [three four]
58	1 one two}]
59
60	glob-parse 105 {"o* two" "three four"} "two one" [string map [list \r\n \n] \
61	{SQL expression: (x GLOB 'o* two' OR x GLOB 'three four')
62	pattern[0] = [o* two]
63	pattern[1] = [three four]
64	0 two one}]
65
66	glob-parse 106 "\"o*\ntwo\" \"three\nfour\"" "one\ntwo" \
67	[string map [list \r\n \n] \
68	{SQL expression: (x GLOB 'o*
69	two' OR x GLOB 'three
	@@ -70,11 +70,11 @@
70	four')
71	pattern[0] = [o*
72	two]
73	pattern[1] = [three
74	four]
75	1 one
76	two}]
77
78	glob-parse 107 "\"o*\ntwo\" \"three\nfour\"" "two\none" \
79	[string map [list \r\n \n] \
80	{SQL expression: (x GLOB 'o*
	@@ -82,11 +82,11 @@
82	four')
83	pattern[0] = [o*
84	two]
85	pattern[1] = [three
86	four]
87	0 two
88	one}]
89
90	glob-parse 108 "\"o*\rtwo\" \"three\rfour\"" "one\rtwo" \
91	[string map [list \r\n \n] \
92	{SQL expression: (x GLOB 'o*
	@@ -94,11 +94,11 @@
94	four')
95	pattern[0] = [o*
96	two]
97	pattern[1] = [three
98	four]
99	1 one
100	two}]
101
102	glob-parse 109 "\"o*\rtwo\" \"three\rfour\"" "two\rone" \
103	[string map [list \r\n \n] \
104	{SQL expression: (x GLOB 'o*
	@@ -106,11 +106,11 @@
106	four')
107	pattern[0] = [o*
108	two]
109	pattern[1] = [three
110	four]
111	0 two
112	one}]
113
114	glob-parse 110 "'o*\ntwo' 'three\nfour'" "one\ntwo" \
115	[string map [list \r\n \n] \
116	{SQL expression: (x GLOB 'o*
	@@ -118,11 +118,11 @@
118	four')
119	pattern[0] = [o*
120	two]
121	pattern[1] = [three
122	four]
123	1 one
124	two}]
125
126	glob-parse 111 "'o*\ntwo' 'three\nfour'" "two\none" \
127	[string map [list \r\n \n] \
128	{SQL expression: (x GLOB 'o*
	@@ -130,61 +130,61 @@
130	four')
131	pattern[0] = [o*
132	two]
133	pattern[1] = [three
134	four]
135	0 two
136	one}]
137
138	glob-parse 112 "\"'o*' 'two'\" \"'three' 'four'\"" "'one' 'two'" \
139	[string map [list \r\n \n] \
140	{SQL expression: (x GLOB '''o*'' ''two''' OR x GLOB '''three'' ''four''')
141	pattern[0] = ['o*' 'two']
142	pattern[1] = ['three' 'four']
143	1 'one' 'two'}]
144
145	glob-parse 113 "\"'o*' 'two'\" \"'three' 'four'\"" "two one" \
146	[string map [list \r\n \n] \
147	{SQL expression: (x GLOB '''o*'' ''two''' OR x GLOB '''three'' ''four''')
148	pattern[0] = ['o*' 'two']
149	pattern[1] = ['three' 'four']
150	0 two one}]
151
152	glob-parse 114 o*,two one [string map [list \r\n \n] \
153	{SQL expression: (x GLOB 'o*' OR x GLOB 'two')
154	pattern[0] = [o*]
155	pattern[1] = [two]
156	1 one}]
157
158	glob-parse 115 "o*,two three,four" "one two" [string map [list \r\n \n] \
159	{SQL expression: (x GLOB 'o*' OR x GLOB 'two' OR x GLOB 'three' OR x GLOB 'four')
160	pattern[0] = [o*]
161	pattern[1] = [two]
162	pattern[2] = [three]
163	pattern[3] = [four]
164	1 one two}]
165
166	glob-parse 116 'o*,two' one [string map [list \r\n \n] \
167	{SQL expression: (x GLOB 'o*,two')
168	pattern[0] = [o*,two]
169	0 one}]
170
171	glob-parse 117 'o*,two' one,two [string map [list \r\n \n] \
172	{SQL expression: (x GLOB 'o*,two')
173	pattern[0] = [o*,two]
174	1 one,two}]
175
176	glob-parse 118 "'o*,two three,four'" "one two three,four" \
177	[string map [list \r\n \n] \
178	{SQL expression: (x GLOB 'o*,two three,four')
179	pattern[0] = [o*,two three,four]
180	0 one two three,four}]
181
182	glob-parse 119 "'o*,two three,four'" "one,two three,four" \
183	[string map [list \r\n \n] \
184	{SQL expression: (x GLOB 'o*,two three,four')
185	pattern[0] = [o*,two three,four]
186	1 one,two three,four}]
187
188	###############################################################################
189
190	test_cleanup
191

	--- test/glob.test
	+++ test/glob.test
	@@ -30,40 +30,40 @@
30	}
31
32	glob-parse 100 test test [string map [list \r\n \n] \
33	{SQL expression: (x GLOB 'test')
34	pattern[0] = [test]
35	1 1 test}]
36
37	glob-parse 101 "one two" one [string map [list \r\n \n] \
38	{SQL expression: (x GLOB 'one' OR x GLOB 'two')
39	pattern[0] = [one]
40	pattern[1] = [two]
41	1 1 one}]
42
43	glob-parse 102 t* test [string map [list \r\n \n] \
44	{SQL expression: (x GLOB 't*')
45	pattern[0] = [t*]
46	1 1 test}]
47
48	glob-parse 103 "o* two" one [string map [list \r\n \n] \
49	{SQL expression: (x GLOB 'o*' OR x GLOB 'two')
50	pattern[0] = [o*]
51	pattern[1] = [two]
52	1 1 one}]
53
54	glob-parse 104 {"o* two" "three four"} "one two" [string map [list \r\n \n] \
55	{SQL expression: (x GLOB 'o* two' OR x GLOB 'three four')
56	pattern[0] = [o* two]
57	pattern[1] = [three four]
58	1 1 one two}]
59
60	glob-parse 105 {"o* two" "three four"} "two one" [string map [list \r\n \n] \
61	{SQL expression: (x GLOB 'o* two' OR x GLOB 'three four')
62	pattern[0] = [o* two]
63	pattern[1] = [three four]
64	0 0 two one}]
65
66	glob-parse 106 "\"o*\ntwo\" \"three\nfour\"" "one\ntwo" \
67	[string map [list \r\n \n] \
68	{SQL expression: (x GLOB 'o*
69	two' OR x GLOB 'three
	@@ -70,11 +70,11 @@
70	four')
71	pattern[0] = [o*
72	two]
73	pattern[1] = [three
74	four]
75	1 1 one
76	two}]
77
78	glob-parse 107 "\"o*\ntwo\" \"three\nfour\"" "two\none" \
79	[string map [list \r\n \n] \
80	{SQL expression: (x GLOB 'o*
	@@ -82,11 +82,11 @@
82	four')
83	pattern[0] = [o*
84	two]
85	pattern[1] = [three
86	four]
87	0 0 two
88	one}]
89
90	glob-parse 108 "\"o*\rtwo\" \"three\rfour\"" "one\rtwo" \
91	[string map [list \r\n \n] \
92	{SQL expression: (x GLOB 'o*
	@@ -94,11 +94,11 @@
94	four')
95	pattern[0] = [o*
96	two]
97	pattern[1] = [three
98	four]
99	1 1 one
100	two}]
101
102	glob-parse 109 "\"o*\rtwo\" \"three\rfour\"" "two\rone" \
103	[string map [list \r\n \n] \
104	{SQL expression: (x GLOB 'o*
	@@ -106,11 +106,11 @@
106	four')
107	pattern[0] = [o*
108	two]
109	pattern[1] = [three
110	four]
111	0 0 two
112	one}]
113
114	glob-parse 110 "'o*\ntwo' 'three\nfour'" "one\ntwo" \
115	[string map [list \r\n \n] \
116	{SQL expression: (x GLOB 'o*
	@@ -118,11 +118,11 @@
118	four')
119	pattern[0] = [o*
120	two]
121	pattern[1] = [three
122	four]
123	1 1 one
124	two}]
125
126	glob-parse 111 "'o*\ntwo' 'three\nfour'" "two\none" \
127	[string map [list \r\n \n] \
128	{SQL expression: (x GLOB 'o*
	@@ -130,61 +130,61 @@
130	four')
131	pattern[0] = [o*
132	two]
133	pattern[1] = [three
134	four]
135	0 0 two
136	one}]
137
138	glob-parse 112 "\"'o*' 'two'\" \"'three' 'four'\"" "'one' 'two'" \
139	[string map [list \r\n \n] \
140	{SQL expression: (x GLOB '''o*'' ''two''' OR x GLOB '''three'' ''four''')
141	pattern[0] = ['o*' 'two']
142	pattern[1] = ['three' 'four']
143	1 1 'one' 'two'}]
144
145	glob-parse 113 "\"'o*' 'two'\" \"'three' 'four'\"" "two one" \
146	[string map [list \r\n \n] \
147	{SQL expression: (x GLOB '''o*'' ''two''' OR x GLOB '''three'' ''four''')
148	pattern[0] = ['o*' 'two']
149	pattern[1] = ['three' 'four']
150	0 0 two one}]
151
152	glob-parse 114 o*,two one [string map [list \r\n \n] \
153	{SQL expression: (x GLOB 'o*' OR x GLOB 'two')
154	pattern[0] = [o*]
155	pattern[1] = [two]
156	1 1 one}]
157
158	glob-parse 115 "o*,two three,four" "one two" [string map [list \r\n \n] \
159	{SQL expression: (x GLOB 'o*' OR x GLOB 'two' OR x GLOB 'three' OR x GLOB 'four')
160	pattern[0] = [o*]
161	pattern[1] = [two]
162	pattern[2] = [three]
163	pattern[3] = [four]
164	1 1 one two}]
165
166	glob-parse 116 'o*,two' one [string map [list \r\n \n] \
167	{SQL expression: (x GLOB 'o*,two')
168	pattern[0] = [o*,two]
169	0 0 one}]
170
171	glob-parse 117 'o*,two' one,two [string map [list \r\n \n] \
172	{SQL expression: (x GLOB 'o*,two')
173	pattern[0] = [o*,two]
174	1 1 one,two}]
175
176	glob-parse 118 "'o*,two three,four'" "one two three,four" \
177	[string map [list \r\n \n] \
178	{SQL expression: (x GLOB 'o*,two three,four')
179	pattern[0] = [o*,two three,four]
180	0 0 one two three,four}]
181
182	glob-parse 119 "'o*,two three,four'" "one,two three,four" \
183	[string map [list \r\n \n] \
184	{SQL expression: (x GLOB 'o*,two three,four')
185	pattern[0] = [o*,two three,four]
186	1 1 one,two three,four}]
187
188	###############################################################################
189
190	test_cleanup
191

M test/settings.test

		--- test/settings.test
		+++ test/settings.test
		@@ -26,10 +26,16 @@
26	26	#
27	27	# fossil settings ?PROPERTY? ?VALUE? ?OPTIONS?
28	28	# fossil unset PROPERTY ?OPTIONS?
29	29	#
30	30	# Where the only supported options are "--global" and "--exact".
	31	+#
	32	+###############################################################################
	33	+#
	34	+# NOTE: The [get_all_settings] procedure from test/tester.tcl returns the list
	35	+# of settings to test and needs to be manually updated when new settings
	36	+# are added.
31	37	#
32	38	###############################################################################
33	39	#
34	40	# NOTE: The [extract_setting_names] procedure extracts the list of setting
35	41	# names from the line-ending normalized output of the "fossil settings"
36	42

	--- test/settings.test
	+++ test/settings.test
	@@ -26,10 +26,16 @@
26	#
27	# fossil settings ?PROPERTY? ?VALUE? ?OPTIONS?
28	# fossil unset PROPERTY ?OPTIONS?
29	#
30	# Where the only supported options are "--global" and "--exact".






31	#
32	###############################################################################
33	#
34	# NOTE: The [extract_setting_names] procedure extracts the list of setting
35	# names from the line-ending normalized output of the "fossil settings"
36

	--- test/settings.test
	+++ test/settings.test
	@@ -26,10 +26,16 @@
26	#
27	# fossil settings ?PROPERTY? ?VALUE? ?OPTIONS?
28	# fossil unset PROPERTY ?OPTIONS?
29	#
30	# Where the only supported options are "--global" and "--exact".
31	#
32	###############################################################################
33	#
34	# NOTE: The [get_all_settings] procedure from test/tester.tcl returns the list
35	# of settings to test and needs to be manually updated when new settings
36	# are added.
37	#
38	###############################################################################
39	#
40	# NOTE: The [extract_setting_names] procedure extracts the list of setting
41	# names from the line-ending normalized output of the "fossil settings"
42

M test/tester.tcl

		--- test/tester.tcl
		+++ test/tester.tcl
		@@ -359,10 +359,12 @@
359	359	proxy \
360	360	redirect-to-https \
361	361	relative-paths \
362	362	repo-cksum \
363	363	repolist-skin \
	364	+ robot-restrict \
	365	+ robots-txt \
364	366	safe-html \
365	367	self-pw-reset \
366	368	self-register \
367	369	sitemap-extra \
368	370	ssh-command \
369	371
370	372	ADDED test/utf8-comment.txt

	--- test/tester.tcl
	+++ test/tester.tcl
	@@ -359,10 +359,12 @@
359	proxy \
360	redirect-to-https \
361	relative-paths \
362	repo-cksum \
363	repolist-skin \


364	safe-html \
365	self-pw-reset \
366	self-register \
367	sitemap-extra \
368	ssh-command \
369
370	DDED test/utf8-comment.txt

	--- test/tester.tcl
	+++ test/tester.tcl
	@@ -359,10 +359,12 @@
359	proxy \
360	redirect-to-https \
361	relative-paths \
362	repo-cksum \
363	repolist-skin \
364	robot-restrict \
365	robots-txt \
366	safe-html \
367	self-pw-reset \
368	self-register \
369	sitemap-extra \
370	ssh-command \
371
372	DDED test/utf8-comment.txt

M test/utf8-comment.txt

		--- a/test/utf8-comment.txt
		+++ b/test/utf8-comment.txt
		@@ -0,0 +1 @@
	1	+The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte [äöü] and symbols [☃] and emoji [💾] characters!

	--- a/test/utf8-comment.txt
	+++ b/test/utf8-comment.txt
	@@ -0,0 +1 @@

	--- a/test/utf8-comment.txt
	+++ b/test/utf8-comment.txt
	@@ -0,0 +1 @@
1	The comment formatter handles ｆｕｌｌｗｉｄｔｈ and multi-byte [äöü] and symbols [☃] and emoji [💾] characters!

M tools/makemake.tcl

+2 -1

		--- tools/makemake.tcl
		+++ tools/makemake.tcl
		@@ -209,10 +209,11 @@
209	209	# Additional resource files that get built into the executable.
210	210	# These paths are all resolved from the src/ directory, so must
211	211	# be relative to that.
212	212	set extra_files {
213	213	diff.tcl
	214	+ merge.tcl
214	215	markdown.md
215	216	wiki.wiki
216	217	*.js
217	218	default.css
218	219	style.*.css
		@@ -445,11 +446,11 @@
445	446
446	447	# The USE_LINENOISE variable may be undefined, set to 0, or set
447	448	# to 1. If it is set to 0, then there is no need to build or link
448	449	# the linenoise.o object.
449	450	LINENOISE_DEF.0 =
450		-LINENOISE_DEF.1 = -DHAVE_LINENOISE
	451	+LINENOISE_DEF.1 = -DHAVE_LINENOISE=2
451	452	LINENOISE_DEF. = $(LINENOISE_DEF.0)
452	453	LINENOISE_OBJ.0 =
453	454	LINENOISE_OBJ.1 = $(OBJDIR)/linenoise.o
454	455	LINENOISE_OBJ. = $(LINENOISE_OBJ.0)
455	456
456	457

	--- tools/makemake.tcl
	+++ tools/makemake.tcl
	@@ -209,10 +209,11 @@
209	# Additional resource files that get built into the executable.
210	# These paths are all resolved from the src/ directory, so must
211	# be relative to that.
212	set extra_files {
213	diff.tcl

214	markdown.md
215	wiki.wiki
216	*.js
217	default.css
218	style.*.css
	@@ -445,11 +446,11 @@
445
446	# The USE_LINENOISE variable may be undefined, set to 0, or set
447	# to 1. If it is set to 0, then there is no need to build or link
448	# the linenoise.o object.
449	LINENOISE_DEF.0 =
450	LINENOISE_DEF.1 = -DHAVE_LINENOISE
451	LINENOISE_DEF. = $(LINENOISE_DEF.0)
452	LINENOISE_OBJ.0 =
453	LINENOISE_OBJ.1 = $(OBJDIR)/linenoise.o
454	LINENOISE_OBJ. = $(LINENOISE_OBJ.0)
455
456

	--- tools/makemake.tcl
	+++ tools/makemake.tcl
	@@ -209,10 +209,11 @@
209	# Additional resource files that get built into the executable.
210	# These paths are all resolved from the src/ directory, so must
211	# be relative to that.
212	set extra_files {
213	diff.tcl
214	merge.tcl
215	markdown.md
216	wiki.wiki
217	*.js
218	default.css
219	style.*.css
	@@ -445,11 +446,11 @@
446
447	# The USE_LINENOISE variable may be undefined, set to 0, or set
448	# to 1. If it is set to 0, then there is no need to build or link
449	# the linenoise.o object.
450	LINENOISE_DEF.0 =
451	LINENOISE_DEF.1 = -DHAVE_LINENOISE=2
452	LINENOISE_DEF. = $(LINENOISE_DEF.0)
453	LINENOISE_OBJ.0 =
454	LINENOISE_OBJ.1 = $(OBJDIR)/linenoise.o
455	LINENOISE_OBJ. = $(LINENOISE_OBJ.0)
456
457

M tools/man_page_command_list.tcl

+1 -1

		--- tools/man_page_command_list.tcl
		+++ tools/man_page_command_list.tcl
		@@ -10,11 +10,11 @@
10	10	set file [lindex $argv 0]
11	11	}
12	12
13	13	# Get list of common commands.
14	14	set commands [exec fossil help]
15		-regsub -nocase {.?\ncommon commands:.\n} $commands {} commands
	15	+regsub -nocase {.?\nfrequently used commands:.\n} $commands {} commands
16	16	regsub -nocase {\nthis is fossil version.*} $commands {} commands
17	17	regsub -all {\s+} $commands " " commands
18	18	set commands [lsort $commands]
19	19
20	20	# Compute number of rows.
21	21

	--- tools/man_page_command_list.tcl
	+++ tools/man_page_command_list.tcl
	@@ -10,11 +10,11 @@
10	set file [lindex $argv 0]
11	}
12
13	# Get list of common commands.
14	set commands [exec fossil help]
15	regsub -nocase {.?\ncommon commands:.\n} $commands {} commands
16	regsub -nocase {\nthis is fossil version.*} $commands {} commands
17	regsub -all {\s+} $commands " " commands
18	set commands [lsort $commands]
19
20	# Compute number of rows.
21

	--- tools/man_page_command_list.tcl
	+++ tools/man_page_command_list.tcl
	@@ -10,11 +10,11 @@
10	set file [lindex $argv 0]
11	}
12
13	# Get list of common commands.
14	set commands [exec fossil help]
15	regsub -nocase {.?\nfrequently used commands:.\n} $commands {} commands
16	regsub -nocase {\nthis is fossil version.*} $commands {} commands
17	regsub -all {\s+} $commands " " commands
18	set commands [lsort $commands]
19
20	# Compute number of rows.
21

M win/Makefile.mingw

		--- win/Makefile.mingw
		+++ win/Makefile.mingw
		@@ -634,10 +634,11 @@
634	634	$(SRCDIR)/hbmenu.js \
635	635	$(SRCDIR)/href.js \
636	636	$(SRCDIR)/login.js \
637	637	$(SRCDIR)/markdown.md \
638	638	$(SRCDIR)/menu.js \
	639	+ $(SRCDIR)/merge.tcl \
639	640	$(SRCDIR)/scroll.js \
640	641	$(SRCDIR)/skin.js \
641	642	$(SRCDIR)/sorttable.js \
642	643	$(SRCDIR)/sounds/0.wav \
643	644	$(SRCDIR)/sounds/1.wav \
644	645

	--- win/Makefile.mingw
	+++ win/Makefile.mingw
	@@ -634,10 +634,11 @@
634	$(SRCDIR)/hbmenu.js \
635	$(SRCDIR)/href.js \
636	$(SRCDIR)/login.js \
637	$(SRCDIR)/markdown.md \
638	$(SRCDIR)/menu.js \

639	$(SRCDIR)/scroll.js \
640	$(SRCDIR)/skin.js \
641	$(SRCDIR)/sorttable.js \
642	$(SRCDIR)/sounds/0.wav \
643	$(SRCDIR)/sounds/1.wav \
644

	--- win/Makefile.mingw
	+++ win/Makefile.mingw
	@@ -634,10 +634,11 @@
634	$(SRCDIR)/hbmenu.js \
635	$(SRCDIR)/href.js \
636	$(SRCDIR)/login.js \
637	$(SRCDIR)/markdown.md \
638	$(SRCDIR)/menu.js \
639	$(SRCDIR)/merge.tcl \
640	$(SRCDIR)/scroll.js \
641	$(SRCDIR)/skin.js \
642	$(SRCDIR)/sorttable.js \
643	$(SRCDIR)/sounds/0.wav \
644	$(SRCDIR)/sounds/1.wav \
645

M win/Makefile.msc

		--- win/Makefile.msc
		+++ win/Makefile.msc
		@@ -592,10 +592,11 @@
592	592	"$(SRCDIR)\hbmenu.js" \
593	593	"$(SRCDIR)\href.js" \
594	594	"$(SRCDIR)\login.js" \
595	595	"$(SRCDIR)\markdown.md" \
596	596	"$(SRCDIR)\menu.js" \
	597	+ "$(SRCDIR)\merge.tcl" \
597	598	"$(SRCDIR)\scroll.js" \
598	599	"$(SRCDIR)\skin.js" \
599	600	"$(SRCDIR)\sorttable.js" \
600	601	"$(SRCDIR)\sounds\0.wav" \
601	602	"$(SRCDIR)\sounds\1.wav" \
		@@ -1222,10 +1223,11 @@
1222	1223	echo "$(SRCDIR)\hbmenu.js" >> $@
1223	1224	echo "$(SRCDIR)\href.js" >> $@
1224	1225	echo "$(SRCDIR)\login.js" >> $@
1225	1226	echo "$(SRCDIR)\markdown.md" >> $@
1226	1227	echo "$(SRCDIR)\menu.js" >> $@
	1228	+ echo "$(SRCDIR)\merge.tcl" >> $@
1227	1229	echo "$(SRCDIR)\scroll.js" >> $@
1228	1230	echo "$(SRCDIR)\skin.js" >> $@
1229	1231	echo "$(SRCDIR)\sorttable.js" >> $@
1230	1232	echo "$(SRCDIR)\sounds/0.wav" >> $@
1231	1233	echo "$(SRCDIR)\sounds/1.wav" >> $@
1232	1234

	--- win/Makefile.msc
	+++ win/Makefile.msc
	@@ -592,10 +592,11 @@
592	"$(SRCDIR)\hbmenu.js" \
593	"$(SRCDIR)\href.js" \
594	"$(SRCDIR)\login.js" \
595	"$(SRCDIR)\markdown.md" \
596	"$(SRCDIR)\menu.js" \

597	"$(SRCDIR)\scroll.js" \
598	"$(SRCDIR)\skin.js" \
599	"$(SRCDIR)\sorttable.js" \
600	"$(SRCDIR)\sounds\0.wav" \
601	"$(SRCDIR)\sounds\1.wav" \
	@@ -1222,10 +1223,11 @@
1222	echo "$(SRCDIR)\hbmenu.js" >> $@
1223	echo "$(SRCDIR)\href.js" >> $@
1224	echo "$(SRCDIR)\login.js" >> $@
1225	echo "$(SRCDIR)\markdown.md" >> $@
1226	echo "$(SRCDIR)\menu.js" >> $@

1227	echo "$(SRCDIR)\scroll.js" >> $@
1228	echo "$(SRCDIR)\skin.js" >> $@
1229	echo "$(SRCDIR)\sorttable.js" >> $@
1230	echo "$(SRCDIR)\sounds/0.wav" >> $@
1231	echo "$(SRCDIR)\sounds/1.wav" >> $@
1232

	--- win/Makefile.msc
	+++ win/Makefile.msc
	@@ -592,10 +592,11 @@
592	"$(SRCDIR)\hbmenu.js" \
593	"$(SRCDIR)\href.js" \
594	"$(SRCDIR)\login.js" \
595	"$(SRCDIR)\markdown.md" \
596	"$(SRCDIR)\menu.js" \
597	"$(SRCDIR)\merge.tcl" \
598	"$(SRCDIR)\scroll.js" \
599	"$(SRCDIR)\skin.js" \
600	"$(SRCDIR)\sorttable.js" \
601	"$(SRCDIR)\sounds\0.wav" \
602	"$(SRCDIR)\sounds\1.wav" \
	@@ -1222,10 +1223,11 @@
1223	echo "$(SRCDIR)\hbmenu.js" >> $@
1224	echo "$(SRCDIR)\href.js" >> $@
1225	echo "$(SRCDIR)\login.js" >> $@
1226	echo "$(SRCDIR)\markdown.md" >> $@
1227	echo "$(SRCDIR)\menu.js" >> $@
1228	echo "$(SRCDIR)\merge.tcl" >> $@
1229	echo "$(SRCDIR)\scroll.js" >> $@
1230	echo "$(SRCDIR)\skin.js" >> $@
1231	echo "$(SRCDIR)\sorttable.js" >> $@
1232	echo "$(SRCDIR)\sounds/0.wav" >> $@
1233	echo "$(SRCDIR)\sounds/1.wav" >> $@
1234

M www/aboutcgi.wiki

+4 -4

		--- www/aboutcgi.wiki
		+++ www/aboutcgi.wiki
		@@ -8,30 +8,30 @@
8	8
9	9	This is a "how it works" guide. This document provides background
10	10	information on the CGI protocol so that you can better understand what
11	11	is going on behind the scenes. If you just want to set up Fossil
12	12	as a CGI server, see the [./server/ \| Fossil Server Setup] page. Or
13		-if you want to development CGI-based extensions to Fossil, see
	13	+if you want to develop CGI-based extensions to Fossil, see
14	14	the [./serverext.wiki\|CGI Server Extensions] page.
15	15
16	16	<h2>A Quick Review Of CGI</h2>
17	17
18	18	An HTTP request is a block of text that is sent by a client application
19	19	(usually a web browser) and arrives at the web server over a network
20	20	connection. The HTTP request contains a URL that describes the information
21	21	being requested. The URL in the HTTP request is typically the same URL
22	22	that appears in the URL bar at the top of the web browser that is making
23		-the request. The URL might contain a "?" character followed
	23	+the request. The URL might contain a "?" character followed by
24	24	query parameters. The HTTP will usually also contain other information
25	25	such as the name of the application that made the request, whether or
26	26	not the requesting application can accept a compressed reply, POST
27	27	parameters from forms, and so forth.
28	28
29	29	The job of the web server is to interpret the HTTP request and formulate
30	30	an appropriate reply.
31		-The web server is free to interpret the HTTP request in any way it wants.
32		-But most web servers follow a similar pattern, described below.
	31	+The web server is free to interpret the HTTP request in any way it wants,
	32	+but most web servers follow a similar pattern, described below.
33	33	(Note: details may vary from one web server to another.)
34	34
35	35	Suppose the filename component of the URL in the HTTP request looks like this:
36	36
37	37	<pre>/one/two/timeline/four</pre>
38	38

	--- www/aboutcgi.wiki
	+++ www/aboutcgi.wiki
	@@ -8,30 +8,30 @@
8
9	This is a "how it works" guide. This document provides background
10	information on the CGI protocol so that you can better understand what
11	is going on behind the scenes. If you just want to set up Fossil
12	as a CGI server, see the [./server/ \| Fossil Server Setup] page. Or
13	if you want to development CGI-based extensions to Fossil, see
14	the [./serverext.wiki\|CGI Server Extensions] page.
15
16	<h2>A Quick Review Of CGI</h2>
17
18	An HTTP request is a block of text that is sent by a client application
19	(usually a web browser) and arrives at the web server over a network
20	connection. The HTTP request contains a URL that describes the information
21	being requested. The URL in the HTTP request is typically the same URL
22	that appears in the URL bar at the top of the web browser that is making
23	the request. The URL might contain a "?" character followed
24	query parameters. The HTTP will usually also contain other information
25	such as the name of the application that made the request, whether or
26	not the requesting application can accept a compressed reply, POST
27	parameters from forms, and so forth.
28
29	The job of the web server is to interpret the HTTP request and formulate
30	an appropriate reply.
31	The web server is free to interpret the HTTP request in any way it wants.
32	But most web servers follow a similar pattern, described below.
33	(Note: details may vary from one web server to another.)
34
35	Suppose the filename component of the URL in the HTTP request looks like this:
36
37	<pre>/one/two/timeline/four</pre>
38

	--- www/aboutcgi.wiki
	+++ www/aboutcgi.wiki
	@@ -8,30 +8,30 @@
8
9	This is a "how it works" guide. This document provides background
10	information on the CGI protocol so that you can better understand what
11	is going on behind the scenes. If you just want to set up Fossil
12	as a CGI server, see the [./server/ \| Fossil Server Setup] page. Or
13	if you want to develop CGI-based extensions to Fossil, see
14	the [./serverext.wiki\|CGI Server Extensions] page.
15
16	<h2>A Quick Review Of CGI</h2>
17
18	An HTTP request is a block of text that is sent by a client application
19	(usually a web browser) and arrives at the web server over a network
20	connection. The HTTP request contains a URL that describes the information
21	being requested. The URL in the HTTP request is typically the same URL
22	that appears in the URL bar at the top of the web browser that is making
23	the request. The URL might contain a "?" character followed by
24	query parameters. The HTTP will usually also contain other information
25	such as the name of the application that made the request, whether or
26	not the requesting application can accept a compressed reply, POST
27	parameters from forms, and so forth.
28
29	The job of the web server is to interpret the HTTP request and formulate
30	an appropriate reply.
31	The web server is free to interpret the HTTP request in any way it wants,
32	but most web servers follow a similar pattern, described below.
33	(Note: details may vary from one web server to another.)
34
35	Suppose the filename component of the URL in the HTTP request looks like this:
36
37	<pre>/one/two/timeline/four</pre>
38

M www/adding_code.wiki

+3 -3

		--- www/adding_code.wiki
		+++ www/adding_code.wiki
		@@ -37,11 +37,11 @@
37	37	The [./makefile.wiki\|Makefile] for Fossil takes care of running these
38	38	preprocessors with all the right arguments and in the right order. So it is
39	39	not necessary to understand the details of how these preprocessors work.
40	40	(Though, the sources for all three preprocessors are included in the source
41	41	tree and are well commented, if you want to dig deeper.) It is only necessary
42		-to know that these preprocessors exist and hence will effect the way you
	42	+to know that these preprocessors exist and hence will affect the way you
43	43	write code.
44	44
45	45	<h2>3.0 Adding New Source Code Files</h2>
46	46
47	47	New source code files are added in the "src/" subdirectory of the Fossil
		@@ -98,11 +98,11 @@
98	98	files. See [../tools/makeheaders.html \| makeheaders.html] for additional
99	99	information.
100	100
101	101	After creating a template file such as shown above, and after updating
102	102	the makefiles, you should be able to recompile Fossil and have it include
103		-your new source file, even before you source file contains any code.
	103	+your new source file, even before your source file contains any code.
104	104	It is recommended that you try this.
105	105
106	106	Be sure to [/help/add\|fossil add] your new source file to the self-hosting
107	107	Fossil repository and then [/help/commit\|commit] your changes!
108	108
		@@ -212,11 +212,11 @@
212	212	non-HTML content. In the unlikely event that you need to generate
213	213	non-HTML content, look at existing webpage implementations
214	214	(ex: "logo" or "style.css") to see how that is done.
215	215
216	216	There are lots of other things that a real web-page implementation will
217		-need to do, such verifying user credentials, parsing query parameters,
	217	+need to do, such as verifying user credentials, parsing query parameters,
218	218	and interacting with the repository. But now that you have the general
219	219	idea of how webpages are implemented, you can look at the many other
220	220	webpage implementations already built into Fossil to see how all that
221	221	works.
222	222
223	223

	--- www/adding_code.wiki
	+++ www/adding_code.wiki
	@@ -37,11 +37,11 @@
37	The [./makefile.wiki\|Makefile] for Fossil takes care of running these
38	preprocessors with all the right arguments and in the right order. So it is
39	not necessary to understand the details of how these preprocessors work.
40	(Though, the sources for all three preprocessors are included in the source
41	tree and are well commented, if you want to dig deeper.) It is only necessary
42	to know that these preprocessors exist and hence will effect the way you
43	write code.
44
45	<h2>3.0 Adding New Source Code Files</h2>
46
47	New source code files are added in the "src/" subdirectory of the Fossil
	@@ -98,11 +98,11 @@
98	files. See [../tools/makeheaders.html \| makeheaders.html] for additional
99	information.
100
101	After creating a template file such as shown above, and after updating
102	the makefiles, you should be able to recompile Fossil and have it include
103	your new source file, even before you source file contains any code.
104	It is recommended that you try this.
105
106	Be sure to [/help/add\|fossil add] your new source file to the self-hosting
107	Fossil repository and then [/help/commit\|commit] your changes!
108
	@@ -212,11 +212,11 @@
212	non-HTML content. In the unlikely event that you need to generate
213	non-HTML content, look at existing webpage implementations
214	(ex: "logo" or "style.css") to see how that is done.
215
216	There are lots of other things that a real web-page implementation will
217	need to do, such verifying user credentials, parsing query parameters,
218	and interacting with the repository. But now that you have the general
219	idea of how webpages are implemented, you can look at the many other
220	webpage implementations already built into Fossil to see how all that
221	works.
222
223

	--- www/adding_code.wiki
	+++ www/adding_code.wiki
	@@ -37,11 +37,11 @@
37	The [./makefile.wiki\|Makefile] for Fossil takes care of running these
38	preprocessors with all the right arguments and in the right order. So it is
39	not necessary to understand the details of how these preprocessors work.
40	(Though, the sources for all three preprocessors are included in the source
41	tree and are well commented, if you want to dig deeper.) It is only necessary
42	to know that these preprocessors exist and hence will affect the way you
43	write code.
44
45	<h2>3.0 Adding New Source Code Files</h2>
46
47	New source code files are added in the "src/" subdirectory of the Fossil
	@@ -98,11 +98,11 @@
98	files. See [../tools/makeheaders.html \| makeheaders.html] for additional
99	information.
100
101	After creating a template file such as shown above, and after updating
102	the makefiles, you should be able to recompile Fossil and have it include
103	your new source file, even before your source file contains any code.
104	It is recommended that you try this.
105
106	Be sure to [/help/add\|fossil add] your new source file to the self-hosting
107	Fossil repository and then [/help/commit\|commit] your changes!
108
	@@ -212,11 +212,11 @@
212	non-HTML content. In the unlikely event that you need to generate
213	non-HTML content, look at existing webpage implementations
214	(ex: "logo" or "style.css") to see how that is done.
215
216	There are lots of other things that a real web-page implementation will
217	need to do, such as verifying user credentials, parsing query parameters,
218	and interacting with the repository. But now that you have the general
219	idea of how webpages are implemented, you can look at the many other
220	webpage implementations already built into Fossil to see how all that
221	works.
222
223

M www/alerts.md

+2 -2

		--- www/alerts.md
		+++ www/alerts.md
		@@ -311,11 +311,11 @@
311	311	## Advanced Email Setups
312	312
313	313	Fossil offers several methods of sending email:
314	314
315	315	1. Pipe the email message text into a command.
316		- 2. Store email messages as entries in a SQLite database.
	316	+ 2. Store email messages as entries in an SQLite database.
317	317	3. Store email messages as individual files in a directory.
318	318	4. Send emails to an SMTP relay.
319	319	5. Send emails directly to the recipients via SMTP.
320	320
321	321	This wide range of options allows Fossil to talk to pretty much any
		@@ -390,11 +390,11 @@
390	390	currently uses this method rather than [the pipe method](#pipe) because
391	391	it is running inside of a restrictive [chroot jail][cj] which is unable
392	392	to hand off messages to the local MTA directly.
393	393
394	394	When you configure a Fossil server this way, it adds outgoing email
395		-messages to a SQLite database file. A separate daemon process can then
	395	+messages to an SQLite database file. A separate daemon process can then
396	396	extract those messages for further disposition.
397	397
398	398	Fossil includes a copy of [the daemon](/file/tools/email-sender.tcl)
399	399	used on `fossil-scm.org`: it is just a short Tcl script that
400	400	continuously monitors this database for new messages and hands any that
401	401

	--- www/alerts.md
	+++ www/alerts.md
	@@ -311,11 +311,11 @@
311	## Advanced Email Setups
312
313	Fossil offers several methods of sending email:
314
315	1. Pipe the email message text into a command.
316	2. Store email messages as entries in a SQLite database.
317	3. Store email messages as individual files in a directory.
318	4. Send emails to an SMTP relay.
319	5. Send emails directly to the recipients via SMTP.
320
321	This wide range of options allows Fossil to talk to pretty much any
	@@ -390,11 +390,11 @@
390	currently uses this method rather than [the pipe method](#pipe) because
391	it is running inside of a restrictive [chroot jail][cj] which is unable
392	to hand off messages to the local MTA directly.
393
394	When you configure a Fossil server this way, it adds outgoing email
395	messages to a SQLite database file. A separate daemon process can then
396	extract those messages for further disposition.
397
398	Fossil includes a copy of [the daemon](/file/tools/email-sender.tcl)
399	used on `fossil-scm.org`: it is just a short Tcl script that
400	continuously monitors this database for new messages and hands any that
401

	--- www/alerts.md
	+++ www/alerts.md
	@@ -311,11 +311,11 @@
311	## Advanced Email Setups
312
313	Fossil offers several methods of sending email:
314
315	1. Pipe the email message text into a command.
316	2. Store email messages as entries in an SQLite database.
317	3. Store email messages as individual files in a directory.
318	4. Send emails to an SMTP relay.
319	5. Send emails directly to the recipients via SMTP.
320
321	This wide range of options allows Fossil to talk to pretty much any
	@@ -390,11 +390,11 @@
390	currently uses this method rather than [the pipe method](#pipe) because
391	it is running inside of a restrictive [chroot jail][cj] which is unable
392	to hand off messages to the local MTA directly.
393
394	When you configure a Fossil server this way, it adds outgoing email
395	messages to an SQLite database file. A separate daemon process can then
396	extract those messages for further disposition.
397
398	Fossil includes a copy of [the daemon](/file/tools/email-sender.tcl)
399	used on `fossil-scm.org`: it is just a short Tcl script that
400	continuously monitors this database for new messages and hands any that
401

M www/antibot.wiki

+2 -2

		--- www/antibot.wiki
		+++ www/antibot.wiki
		@@ -73,17 +73,17 @@
73	73	In Fossil, under the Admin/Robot-Defense menu, there is a setting entitled
74	74	"<b>Enable hyperlinks based on User-Agent and/or Javascript</b>".
75	75	If this setting is set to "UserAgent only" or "UserAgent and Javascript",
76	76	and if the UserAgent string looks like a human and not a robot, then
77	77	Fossil will enable hyperlinks even if the <b>Hyperlink</b> capability
78		-is omitted from the user permissions. This settingn gives humans easy
	78	+is omitted from the user permissions. This setting gives humans easy
79	79	access to the hyperlinks while preventing robots
80	80	from walking the billions of pages on a typical Fossil site.
81	81
82	82	If the setting is "UserAgent only", then the hyperlinks are simply
83	83	enabled and that is all. But if the setting is "UserAgent and Javascript",
84		-then the hyperlinks are not enabled directly .
	84	+then the hyperlinks are not enabled directly.
85	85	Instead, the HTML code that is generated contains anchor tags ("<a>")
86	86	with "href=" attributes that point to [/honeypot] rather than the correct
87	87	link. JavaScript code is added to the end of the page that goes back and
88	88	fills in the correct "href=" attributes of
89	89	the anchor tags with the true hyperlink targets, thus enabling the hyperlinks.
90	90

	--- www/antibot.wiki
	+++ www/antibot.wiki
	@@ -73,17 +73,17 @@
73	In Fossil, under the Admin/Robot-Defense menu, there is a setting entitled
74	"<b>Enable hyperlinks based on User-Agent and/or Javascript</b>".
75	If this setting is set to "UserAgent only" or "UserAgent and Javascript",
76	and if the UserAgent string looks like a human and not a robot, then
77	Fossil will enable hyperlinks even if the <b>Hyperlink</b> capability
78	is omitted from the user permissions. This settingn gives humans easy
79	access to the hyperlinks while preventing robots
80	from walking the billions of pages on a typical Fossil site.
81
82	If the setting is "UserAgent only", then the hyperlinks are simply
83	enabled and that is all. But if the setting is "UserAgent and Javascript",
84	then the hyperlinks are not enabled directly .
85	Instead, the HTML code that is generated contains anchor tags ("<a>")
86	with "href=" attributes that point to [/honeypot] rather than the correct
87	link. JavaScript code is added to the end of the page that goes back and
88	fills in the correct "href=" attributes of
89	the anchor tags with the true hyperlink targets, thus enabling the hyperlinks.
90

	--- www/antibot.wiki
	+++ www/antibot.wiki
	@@ -73,17 +73,17 @@
73	In Fossil, under the Admin/Robot-Defense menu, there is a setting entitled
74	"<b>Enable hyperlinks based on User-Agent and/or Javascript</b>".
75	If this setting is set to "UserAgent only" or "UserAgent and Javascript",
76	and if the UserAgent string looks like a human and not a robot, then
77	Fossil will enable hyperlinks even if the <b>Hyperlink</b> capability
78	is omitted from the user permissions. This setting gives humans easy
79	access to the hyperlinks while preventing robots
80	from walking the billions of pages on a typical Fossil site.
81
82	If the setting is "UserAgent only", then the hyperlinks are simply
83	enabled and that is all. But if the setting is "UserAgent and Javascript",
84	then the hyperlinks are not enabled directly.
85	Instead, the HTML code that is generated contains anchor tags ("<a>")
86	with "href=" attributes that point to [/honeypot] rather than the correct
87	link. JavaScript code is added to the end of the page that goes back and
88	fills in the correct "href=" attributes of
89	the anchor tags with the true hyperlink targets, thus enabling the hyperlinks.
90

M www/backoffice.md

+3 -3

		--- www/backoffice.md
		+++ www/backoffice.md
		@@ -115,14 +115,14 @@
115	115	launching the backoffice for all of them. There are additional useful
116	116	command-line options. See the "[fossil backoffice](/help?cmd=backoffice)"
117	117	documentation for details.
118	118
119	119	The backoffice processes run manually using the "fossil backoffice"
120		-command do not normally use a lease. That means that you run the
	120	+command do not normally use a lease. That means that if you run the
121	121	"fossil backoffice" command with --poll and you forget to disable
122	122	automatic backoffice by setting the "backoffice-disable" flag, then
123		-you might have one backoffice running due command and another due
	123	+you might have one backoffice running due to a command and another due
124	124	to a webpage access, both at the same time. This is harmless. The
125	125	only downside is that it uses extra CPU time.
126	126
127	127	How Backoffice Is Implemented
128	128	-----------------------------
		@@ -210,11 +210,11 @@
210	210	backoffice processes to exit either immediately or after doing whatever
211	211	backoffice works needs to be done. If something is going wrong and
212	212	backoffice leases are causing delays in webpage processing, then setting
213	213	"backoffice-nodelay" to true can work around the problem until the bug
214	214	can be fixed. The "backoffice-logfile" setting is the name of a log
215		-file onto which is appended a short message everything a backoffice
	215	+file onto which is appended a short message every time a backoffice
216	216	process actually starts to do the backoffice work. This log file can
217	217	be used to verify that backoffice really is running, if there is any
218	218	doubt. The "backoffice-disable" setting prevents automatic backoffice
219	219	processing, if true. Use this to completely disable backoffice processing
220	220	that occurs automatically after each HTTP request. The "backoffice-disable"
221	221

	--- www/backoffice.md
	+++ www/backoffice.md
	@@ -115,14 +115,14 @@
115	launching the backoffice for all of them. There are additional useful
116	command-line options. See the "[fossil backoffice](/help?cmd=backoffice)"
117	documentation for details.
118
119	The backoffice processes run manually using the "fossil backoffice"
120	command do not normally use a lease. That means that you run the
121	"fossil backoffice" command with --poll and you forget to disable
122	automatic backoffice by setting the "backoffice-disable" flag, then
123	you might have one backoffice running due command and another due
124	to a webpage access, both at the same time. This is harmless. The
125	only downside is that it uses extra CPU time.
126
127	How Backoffice Is Implemented
128	-----------------------------
	@@ -210,11 +210,11 @@
210	backoffice processes to exit either immediately or after doing whatever
211	backoffice works needs to be done. If something is going wrong and
212	backoffice leases are causing delays in webpage processing, then setting
213	"backoffice-nodelay" to true can work around the problem until the bug
214	can be fixed. The "backoffice-logfile" setting is the name of a log
215	file onto which is appended a short message everything a backoffice
216	process actually starts to do the backoffice work. This log file can
217	be used to verify that backoffice really is running, if there is any
218	doubt. The "backoffice-disable" setting prevents automatic backoffice
219	processing, if true. Use this to completely disable backoffice processing
220	that occurs automatically after each HTTP request. The "backoffice-disable"
221

	--- www/backoffice.md
	+++ www/backoffice.md
	@@ -115,14 +115,14 @@
115	launching the backoffice for all of them. There are additional useful
116	command-line options. See the "[fossil backoffice](/help?cmd=backoffice)"
117	documentation for details.
118
119	The backoffice processes run manually using the "fossil backoffice"
120	command do not normally use a lease. That means that if you run the
121	"fossil backoffice" command with --poll and you forget to disable
122	automatic backoffice by setting the "backoffice-disable" flag, then
123	you might have one backoffice running due to a command and another due
124	to a webpage access, both at the same time. This is harmless. The
125	only downside is that it uses extra CPU time.
126
127	How Backoffice Is Implemented
128	-----------------------------
	@@ -210,11 +210,11 @@
210	backoffice processes to exit either immediately or after doing whatever
211	backoffice works needs to be done. If something is going wrong and
212	backoffice leases are causing delays in webpage processing, then setting
213	"backoffice-nodelay" to true can work around the problem until the bug
214	can be fixed. The "backoffice-logfile" setting is the name of a log
215	file onto which is appended a short message every time a backoffice
216	process actually starts to do the backoffice work. This log file can
217	be used to verify that backoffice really is running, if there is any
218	doubt. The "backoffice-disable" setting prevents automatic backoffice
219	processing, if true. Use this to completely disable backoffice processing
220	that occurs automatically after each HTTP request. The "backoffice-disable"
221

M www/blame.wiki

+2 -2

		--- www/blame.wiki
		+++ www/blame.wiki
		@@ -19,11 +19,11 @@
19	19	annotated. Call this check-in C0.
20	20	<li>Find all direct ancestors of C0. A direct ancestor is the closure
21	21	of the primary parent of C0. Merged in branches are not part of
22	22	the direct ancestors of C0.
23	23	<li>Prune the list of ancestors of C0 so that it contains only
24		- check-in in which the file to be annotated was modified.
	24	+ check-ins in which the file to be annotated was modified.
25	25	<li>Load the complete text of the file to be annotated from check-in C0.
26	26	Call this version of the file F0.
27	27	<li>Parse F0 into lines. Mark each line as "unchanged".
28	28	<li>For each ancestor of C0 on the pruned list (call the ancestor CX),
29	29	beginning with the most
		@@ -45,11 +45,11 @@
45	45
46	46	The time-consuming part of this algorithm is step 6b - computing the
47	47	diff from all historical versions of the file to the version of the file
48	48	under analysis. For a large file that has many historical changes, this
49	49	can take several seconds. For this reason, the default
50		-[/help?cmd=/annotate\|/annotate] webpage only shows those lines that where
	50	+[/help?cmd=/annotate\|/annotate] webpage only shows those lines that were
51	51	changed by the 20 most recent modifications to the file. This allows
52	52	the loop on step 6 to terminate after only 19 diffs instead of the hundreds
53	53	or thousands of diffs that might be required for a frequently modified file.
54	54
55	55	As currently implemented (as of 2015-12-12) the annotate algorithm does not
56	56

	--- www/blame.wiki
	+++ www/blame.wiki
	@@ -19,11 +19,11 @@
19	annotated. Call this check-in C0.
20	<li>Find all direct ancestors of C0. A direct ancestor is the closure
21	of the primary parent of C0. Merged in branches are not part of
22	the direct ancestors of C0.
23	<li>Prune the list of ancestors of C0 so that it contains only
24	check-in in which the file to be annotated was modified.
25	<li>Load the complete text of the file to be annotated from check-in C0.
26	Call this version of the file F0.
27	<li>Parse F0 into lines. Mark each line as "unchanged".
28	<li>For each ancestor of C0 on the pruned list (call the ancestor CX),
29	beginning with the most
	@@ -45,11 +45,11 @@
45
46	The time-consuming part of this algorithm is step 6b - computing the
47	diff from all historical versions of the file to the version of the file
48	under analysis. For a large file that has many historical changes, this
49	can take several seconds. For this reason, the default
50	[/help?cmd=/annotate\|/annotate] webpage only shows those lines that where
51	changed by the 20 most recent modifications to the file. This allows
52	the loop on step 6 to terminate after only 19 diffs instead of the hundreds
53	or thousands of diffs that might be required for a frequently modified file.
54
55	As currently implemented (as of 2015-12-12) the annotate algorithm does not
56

	--- www/blame.wiki
	+++ www/blame.wiki
	@@ -19,11 +19,11 @@
19	annotated. Call this check-in C0.
20	<li>Find all direct ancestors of C0. A direct ancestor is the closure
21	of the primary parent of C0. Merged in branches are not part of
22	the direct ancestors of C0.
23	<li>Prune the list of ancestors of C0 so that it contains only
24	check-ins in which the file to be annotated was modified.
25	<li>Load the complete text of the file to be annotated from check-in C0.
26	Call this version of the file F0.
27	<li>Parse F0 into lines. Mark each line as "unchanged".
28	<li>For each ancestor of C0 on the pruned list (call the ancestor CX),
29	beginning with the most
	@@ -45,11 +45,11 @@
45
46	The time-consuming part of this algorithm is step 6b - computing the
47	diff from all historical versions of the file to the version of the file
48	under analysis. For a large file that has many historical changes, this
49	can take several seconds. For this reason, the default
50	[/help?cmd=/annotate\|/annotate] webpage only shows those lines that were
51	changed by the 20 most recent modifications to the file. This allows
52	the loop on step 6 to terminate after only 19 diffs instead of the hundreds
53	or thousands of diffs that might be required for a frequently modified file.
54
55	As currently implemented (as of 2015-12-12) the annotate algorithm does not
56

M www/bugtheory.wiki

+6 -6

		--- www/bugtheory.wiki
		+++ www/bugtheory.wiki
		@@ -112,13 +112,13 @@
112	112	and is not shared with other repositories during a sync, push, or pull.
113	113
114	114	Each repository also defines scripts used to generate web pages for
115	115	creating new tickets, viewing existing tickets, and modifying an
116	116	existing ticket. These scripts consist of HTML with an embedded
117		-scripts written a Tcl-like language called "[./th1.md\|TH1]". Every new fossil
118		-repository is created with default scripts. Administrators wishing to
119		-customize their ticket entry, viewing, and editing screens should
120		-modify the default scripts to suit their needs. These screen generator
121		-scripts are part of the local state of a repository and are not shared
122		-with other repositories during a sync, push, or pull.
	117	+scripts written in a Tcl-like language called "[./th1.md\|TH1]". Every
	118	+new fossil repository is created with default scripts. Administrators
	119	+wishing to customize their ticket entry, viewing, and editing screens
	120	+should modify the default scripts to suit their needs. These screen
	121	+generator scripts are part of the local state of a repository and are
	122	+not shared with other repositories during a sync, push, or pull.
123	123
124	124	<i>To be continued...</i>
125	125

	--- www/bugtheory.wiki
	+++ www/bugtheory.wiki
	@@ -112,13 +112,13 @@
112	and is not shared with other repositories during a sync, push, or pull.
113
114	Each repository also defines scripts used to generate web pages for
115	creating new tickets, viewing existing tickets, and modifying an
116	existing ticket. These scripts consist of HTML with an embedded
117	scripts written a Tcl-like language called "[./th1.md\|TH1]". Every new fossil
118	repository is created with default scripts. Administrators wishing to
119	customize their ticket entry, viewing, and editing screens should
120	modify the default scripts to suit their needs. These screen generator
121	scripts are part of the local state of a repository and are not shared
122	with other repositories during a sync, push, or pull.
123
124	<i>To be continued...</i>
125

	--- www/bugtheory.wiki
	+++ www/bugtheory.wiki
	@@ -112,13 +112,13 @@
112	and is not shared with other repositories during a sync, push, or pull.
113
114	Each repository also defines scripts used to generate web pages for
115	creating new tickets, viewing existing tickets, and modifying an
116	existing ticket. These scripts consist of HTML with an embedded
117	scripts written in a Tcl-like language called "[./th1.md\|TH1]". Every
118	new fossil repository is created with default scripts. Administrators
119	wishing to customize their ticket entry, viewing, and editing screens
120	should modify the default scripts to suit their needs. These screen
121	generator scripts are part of the local state of a repository and are
122	not shared with other repositories during a sync, push, or pull.
123
124	<i>To be continued...</i>
125

M www/build.wiki

+2 -2

		--- www/build.wiki
		+++ www/build.wiki
		@@ -55,11 +55,11 @@
55	55	<h2>Aside: Is it really safe to use an unreleased development version of
56	56	the Fossil source code?</h2>
57	57
58	58	Yes! Any check-in on the
59	59	[/timeline?t=trunk \| trunk branch] of the Fossil
60		-[https://fossil-scm.org/ \| Fossil self-hosting repository]
	60	+[https://fossil-scm.org/ \| self-hosting repository]
61	61	will work fine. (Dodgy code is always on a branch.) In the unlikely
62	62	event that you pick a version with a serious bug, it still won't
63	63	clobber your files. Fossil uses several
64	64	[./selfcheck.wiki \| self-checks] prior to committing any
65	65	repository change that prevent loss-of-work due to bugs.
		@@ -475,11 +475,11 @@
475	475	environment variables which are required by <tt>emcc</tt> and will
476	476	fail if it cannot do so. Once it's set up the environment, it passes
477	477	on all of its arguments to <tt>emcc</tt>.
478	478
479	479	The WASM-related build parts are set up such that none of them should
480		-ever trigger implicity (e.g. via dependencies resolution) in a normal
	480	+ever trigger implicitly (e.g. via dependencies resolution) in a normal
481	481	build cycle. They are instead explicitly built as described below.
482	482
483	483	From the top of the source tree, all WASM-related components can be
484	484	built with:
485	485
486	486

	--- www/build.wiki
	+++ www/build.wiki
	@@ -55,11 +55,11 @@
55	<h2>Aside: Is it really safe to use an unreleased development version of
56	the Fossil source code?</h2>
57
58	Yes! Any check-in on the
59	[/timeline?t=trunk \| trunk branch] of the Fossil
60	[https://fossil-scm.org/ \| Fossil self-hosting repository]
61	will work fine. (Dodgy code is always on a branch.) In the unlikely
62	event that you pick a version with a serious bug, it still won't
63	clobber your files. Fossil uses several
64	[./selfcheck.wiki \| self-checks] prior to committing any
65	repository change that prevent loss-of-work due to bugs.
	@@ -475,11 +475,11 @@
475	environment variables which are required by <tt>emcc</tt> and will
476	fail if it cannot do so. Once it's set up the environment, it passes
477	on all of its arguments to <tt>emcc</tt>.
478
479	The WASM-related build parts are set up such that none of them should
480	ever trigger implicity (e.g. via dependencies resolution) in a normal
481	build cycle. They are instead explicitly built as described below.
482
483	From the top of the source tree, all WASM-related components can be
484	built with:
485
486

	--- www/build.wiki
	+++ www/build.wiki
	@@ -55,11 +55,11 @@
55	<h2>Aside: Is it really safe to use an unreleased development version of
56	the Fossil source code?</h2>
57
58	Yes! Any check-in on the
59	[/timeline?t=trunk \| trunk branch] of the Fossil
60	[https://fossil-scm.org/ \| self-hosting repository]
61	will work fine. (Dodgy code is always on a branch.) In the unlikely
62	event that you pick a version with a serious bug, it still won't
63	clobber your files. Fossil uses several
64	[./selfcheck.wiki \| self-checks] prior to committing any
65	repository change that prevent loss-of-work due to bugs.
	@@ -475,11 +475,11 @@
475	environment variables which are required by <tt>emcc</tt> and will
476	fail if it cannot do so. Once it's set up the environment, it passes
477	on all of its arguments to <tt>emcc</tt>.
478
479	The WASM-related build parts are set up such that none of them should
480	ever trigger implicitly (e.g. via dependencies resolution) in a normal
481	build cycle. They are instead explicitly built as described below.
482
483	From the top of the source tree, all WASM-related components can be
484	built with:
485
486

M www/caps/index.md

+1 -1

		--- www/caps/index.md
		+++ www/caps/index.md
		@@ -294,11 +294,11 @@
294	294	Fossil requires write access to a repo DB while cloning from it, so you
295	295	can’t clone from a read-only repo DB file over a local file path.
296	296
297	297	Even more surprising to you may be the fact that user caps do not affect
298	298	cloning and syncing over SSH! (Not unless you go [out of your way][sshfc]
299		-patch around it, at any rate.) When you make a change to such a
	299	+to patch around it, at any rate.) When you make a change to such a
300	300	repository, the stock Fossil behavior is that the change first goes to the
301	301	local repo clone where file system
302	302	permissions are all that matter, but then upon sync, the situation is
303	303	effectively the same as when the parent repo is on the local file
304	304	system. The reason behind this is that if you can log into the remote
305	305

	--- www/caps/index.md
	+++ www/caps/index.md
	@@ -294,11 +294,11 @@
294	Fossil requires write access to a repo DB while cloning from it, so you
295	can’t clone from a read-only repo DB file over a local file path.
296
297	Even more surprising to you may be the fact that user caps do not affect
298	cloning and syncing over SSH! (Not unless you go [out of your way][sshfc]
299	patch around it, at any rate.) When you make a change to such a
300	repository, the stock Fossil behavior is that the change first goes to the
301	local repo clone where file system
302	permissions are all that matter, but then upon sync, the situation is
303	effectively the same as when the parent repo is on the local file
304	system. The reason behind this is that if you can log into the remote
305

	--- www/caps/index.md
	+++ www/caps/index.md
	@@ -294,11 +294,11 @@
294	Fossil requires write access to a repo DB while cloning from it, so you
295	can’t clone from a read-only repo DB file over a local file path.
296
297	Even more surprising to you may be the fact that user caps do not affect
298	cloning and syncing over SSH! (Not unless you go [out of your way][sshfc]
299	to patch around it, at any rate.) When you make a change to such a
300	repository, the stock Fossil behavior is that the change first goes to the
301	local repo clone where file system
302	permissions are all that matter, but then upon sync, the situation is
303	effectively the same as when the parent repo is on the local file
304	system. The reason behind this is that if you can log into the remote
305

M www/caps/ref.html

+1 -1

		--- www/caps/ref.html
		+++ www/caps/ref.html
		@@ -252,11 +252,11 @@
252	252	Create new ticket report formats. Note that although this allows
253	253	the user to provide SQL code to be run in the server’s context,
254	254	and this capability is given to the untrusted “anonymous” user
255	255	category by default, this is a safe capability to give to users
256	256	because it is internally restricted to read-only queries on the
257		- tickets table only. (This restriction is done with a SQLite
	257	+ tickets table only. (This restriction is done with an SQLite
258	258	authorization hook, not by any method so weak as SQL text
259	259	filtering.) Mnemonic: new <b>t</b>icket report.
260	260	</td>
261	261	</tr>
262	262
263	263

	--- www/caps/ref.html
	+++ www/caps/ref.html
	@@ -252,11 +252,11 @@
252	Create new ticket report formats. Note that although this allows
253	the user to provide SQL code to be run in the server’s context,
254	and this capability is given to the untrusted “anonymous” user
255	category by default, this is a safe capability to give to users
256	because it is internally restricted to read-only queries on the
257	tickets table only. (This restriction is done with a SQLite
258	authorization hook, not by any method so weak as SQL text
259	filtering.) Mnemonic: new <b>t</b>icket report.
260	</td>
261	</tr>
262
263

	--- www/caps/ref.html
	+++ www/caps/ref.html
	@@ -252,11 +252,11 @@
252	Create new ticket report formats. Note that although this allows
253	the user to provide SQL code to be run in the server’s context,
254	and this capability is given to the untrusted “anonymous” user
255	category by default, this is a safe capability to give to users
256	because it is internally restricted to read-only queries on the
257	tickets table only. (This restriction is done with an SQLite
258	authorization hook, not by any method so weak as SQL text
259	filtering.) Mnemonic: new <b>t</b>icket report.
260	</td>
261	</tr>
262
263

M www/cgi.wiki

+2 -2

		--- www/cgi.wiki
		+++ www/cgi.wiki
		@@ -13,11 +13,11 @@
13	13	setting up a Fossil server, one of which is [./server/any/cgi.md \| as a CGI
14	14	script]. CGI is the technique that the three
15	15	[./selfhost.wiki\|self-hosting Fossil repositories] all use.
16	16
17	17	Setting up a Fossil server using CGI is mostly about writing a short
18		-script (usually just 2 lines line) in the cgi-bin folder of an ordinary
	18	+script (usually just 2 lines) in the cgi-bin folder of an ordinary
19	19	web-server. But there are a lot of extra options that can be added
20	20	to this script, to customize the configuration. This article describes
21	21	those options.
22	22
23	23	<h1>CGI Script Options</h1>
		@@ -191,11 +191,11 @@
191	191
192	192
193	193	<h2 id="mainmenu">mainmenu: <i>FILE</i></h2>
194	194
195	195	This parameter causes the contents of the given file to override the
196		-site's <tt>mainmenu</tt> configuration setting, much in the same way
	196	+site's <tt>mainmenu</tt> configuration setting, in much the same way
197	197	that the <tt>skin</tt> setting overrides the skin. This can be used to
198	198	apply a common main menu to a number of sites, and centrally maintain
199	199	it, without having to copy its contents into each site. Note, however,
200	200	that the contents of this setting are not stored in the repository and
201	201	will not be cloned along with the repository.
202	202

	--- www/cgi.wiki
	+++ www/cgi.wiki
	@@ -13,11 +13,11 @@
13	setting up a Fossil server, one of which is [./server/any/cgi.md \| as a CGI
14	script]. CGI is the technique that the three
15	[./selfhost.wiki\|self-hosting Fossil repositories] all use.
16
17	Setting up a Fossil server using CGI is mostly about writing a short
18	script (usually just 2 lines line) in the cgi-bin folder of an ordinary
19	web-server. But there are a lot of extra options that can be added
20	to this script, to customize the configuration. This article describes
21	those options.
22
23	<h1>CGI Script Options</h1>
	@@ -191,11 +191,11 @@
191
192
193	<h2 id="mainmenu">mainmenu: <i>FILE</i></h2>
194
195	This parameter causes the contents of the given file to override the
196	site's <tt>mainmenu</tt> configuration setting, much in the same way
197	that the <tt>skin</tt> setting overrides the skin. This can be used to
198	apply a common main menu to a number of sites, and centrally maintain
199	it, without having to copy its contents into each site. Note, however,
200	that the contents of this setting are not stored in the repository and
201	will not be cloned along with the repository.
202

	--- www/cgi.wiki
	+++ www/cgi.wiki
	@@ -13,11 +13,11 @@
13	setting up a Fossil server, one of which is [./server/any/cgi.md \| as a CGI
14	script]. CGI is the technique that the three
15	[./selfhost.wiki\|self-hosting Fossil repositories] all use.
16
17	Setting up a Fossil server using CGI is mostly about writing a short
18	script (usually just 2 lines) in the cgi-bin folder of an ordinary
19	web-server. But there are a lot of extra options that can be added
20	to this script, to customize the configuration. This article describes
21	those options.
22
23	<h1>CGI Script Options</h1>
	@@ -191,11 +191,11 @@
191
192
193	<h2 id="mainmenu">mainmenu: <i>FILE</i></h2>
194
195	This parameter causes the contents of the given file to override the
196	site's <tt>mainmenu</tt> configuration setting, in much the same way
197	that the <tt>skin</tt> setting overrides the skin. This can be used to
198	apply a common main menu to a number of sites, and centrally maintain
199	it, without having to copy its contents into each site. Note, however,
200	that the contents of this setting are not stored in the repository and
201	will not be cloned along with the repository.
202

M www/changes.wiki

+26 -4

		--- www/changes.wiki
		+++ www/changes.wiki
		@@ -1,8 +1,22 @@
1	1	<title>Change Log</title>
2	2
3		-<h2 id='v2_25'>Changes for version 2.25 (pending)</h2>
	3	+<h2 id='v2_26'>Changes for version 2.26 (pending)</h2>
	4	+
	5	+ * Added the [/help?cmd=/ckout\|/ckout web page] to provide information
	6	+ about pending changes in a working check-out
	7	+ * The [/help?cmd=ui\|fossil ui] command defaults to using the /ckout
	8	+ page as its start page.
	9	+ * Added the [/help?cmd=merge-info\|fossil merge-info] command and especially
	10	+ the --tk option to that command, to provide analysis of the most recent
	11	+ merge or update operation.
	12	+ * Issue a warning if a user tries to commit on a check-in where the
	13	+ branch has been changed.
	14	+ * When a merge conflict occurs, a new section is added to the conflict
	15	+ text that shows Fossil's suggested resolution to the conflict.
	16	+
	17	+<h2 id='v2_25'>Changes for version 2.25 (2024-11-06)</h2>
4	18
5	19	* The "[/help?cmd=ui\|fossil ui /]" command now works even for repositories
6	20	that have non-ASCII filenames
7	21	* Add the [/help?cmd=tree\|fossil tree] command.
8	22	* On case-insensitive filesystems, store files using the filesystem's
		@@ -11,12 +25,21 @@
11	25	which is more familiar to Git users. Retain the legacy name for
12	26	compatibility.
13	27	* Add new query parameters to the [/help?cmd=/timeline\|/timeline page]:
14	28	d2=, p2=, and dp2=.
15	29	* Add options to the [/help?cmd=tag\|fossil tag] command that will list tag values.
	30	+ * Add the -b\|--brief option to the [/help?cmd=status\|fossil status] command.
16	31	* Add ability to upload unversioned files via the [/help?cmd=/uvlist\|/uvlist page].
17	32	* Add history search to the [/help?cmd=/chat\|/chat page].
	33	+ * Add Unix socket support to the [/help?cmd=server\|server command].
	34	+ * On Windows, use the root certificates managed by the operating system
	35	+ (requires OpenSSL 3.2.0 or greater).
	36	+ * Take into account zero-width and double-width unicode characters when
	37	+ formatting the command-line timeline.
	38	+ * Update the built-in SQLite to version 3.47.0. Precompiled binaries are
	39	+ linked against OpenSSL 3.4.0.
	40	+ * Numerous minor fixes and additions.
18	41
19	42
20	43	<h2 id='v2_24'>Changes for version 2.24 (2024-04-23)</h2>
21	44
22	45	* Apache change work-around → As part of a security fix, the Apache webserver
		@@ -85,11 +108,11 @@
85	108	* Add ability to "close" forum threads, such that unprivileged users
86	109	may no longer respond to them. Only administrators can close
87	110	threads or respond to them by default, and the
88	111	[/help?cmd=forum-close-policy\|forum-close-policy setting] can be
89	112	used to add that capability to moderators.
90		- * Add the [/help?cmd=all\|fossil all whatis] command.
	113	+ * Add the [/help?cmd=all\|fossil all whatis] command.
91	114	* The [/help?cmd=status\|fossil status] command and relevant UI pages now
92	115	correctly report files which were both renamed <b>and</b> edited as such.
93	116	* Show default value of settings that have a default in
94	117	[/help?cmd=help\|fossil help SETTING] output.
95	118	* On timeline graphs, show closed check-ins using an X in the middle of the
		@@ -127,11 +150,10 @@
127	150	they work when the "remote" machine is a Mac and the "fossil"
128	151	executable is in the $HOME/bin directory.
129	152	</ul>
130	153	* Update built-in libraries SQLite, ZLib, Pikchr to their latest versions.
131	154	* Documentation enhancements and typo fixes.
132		-
133	155
134	156
135	157	<h2 id='v2_22'>Changes for version 2.22 (2023-05-31)</h2>
136	158	* Enhancements to the [/help?cmd=/timeline\|/timeline webpage]: <ol type="a">
137	159	<li> Add the ft=TAG query parameter which in combination with d=Y
		@@ -155,11 +177,11 @@
155	177	frees us from needing to build and install the container as root,
156	178	since it no longer has to create a private <tt>/dev</tt> tree
157	179	inside the jail for Fossil's use.
158	180	* Add support for the trigram tokenizer for FTS5 search to enable
159	181	searching in Chinese.
160		- * Comment lines (starting with a '#') are now supported inside
	182	+ * Comment lines (starting with a '#') are now supported inside
161	183	[./settings.wiki#versionable\|versioned settings].
162	184	* Default permissions for anonymous users in new repositories are
163	185	changed to "hz".
164	186	* The [/help?cmd=status\|fossil status] command now detects when a
165	187	file used to be a symlink and has been replaced by a regular file.
166	188

	--- www/changes.wiki
	+++ www/changes.wiki
	@@ -1,8 +1,22 @@
1	<title>Change Log</title>
2
3	<h2 id='v2_25'>Changes for version 2.25 (pending)</h2>














4
5	* The "[/help?cmd=ui\|fossil ui /]" command now works even for repositories
6	that have non-ASCII filenames
7	* Add the [/help?cmd=tree\|fossil tree] command.
8	* On case-insensitive filesystems, store files using the filesystem's
	@@ -11,12 +25,21 @@
11	which is more familiar to Git users. Retain the legacy name for
12	compatibility.
13	* Add new query parameters to the [/help?cmd=/timeline\|/timeline page]:
14	d2=, p2=, and dp2=.
15	* Add options to the [/help?cmd=tag\|fossil tag] command that will list tag values.

16	* Add ability to upload unversioned files via the [/help?cmd=/uvlist\|/uvlist page].
17	* Add history search to the [/help?cmd=/chat\|/chat page].








18
19
20	<h2 id='v2_24'>Changes for version 2.24 (2024-04-23)</h2>
21
22	* Apache change work-around → As part of a security fix, the Apache webserver
	@@ -85,11 +108,11 @@
85	* Add ability to "close" forum threads, such that unprivileged users
86	may no longer respond to them. Only administrators can close
87	threads or respond to them by default, and the
88	[/help?cmd=forum-close-policy\|forum-close-policy setting] can be
89	used to add that capability to moderators.
90	* Add the [/help?cmd=all\|fossil all whatis] command.
91	* The [/help?cmd=status\|fossil status] command and relevant UI pages now
92	correctly report files which were both renamed <b>and</b> edited as such.
93	* Show default value of settings that have a default in
94	[/help?cmd=help\|fossil help SETTING] output.
95	* On timeline graphs, show closed check-ins using an X in the middle of the
	@@ -127,11 +150,10 @@
127	they work when the "remote" machine is a Mac and the "fossil"
128	executable is in the $HOME/bin directory.
129	</ul>
130	* Update built-in libraries SQLite, ZLib, Pikchr to their latest versions.
131	* Documentation enhancements and typo fixes.
132
133
134
135	<h2 id='v2_22'>Changes for version 2.22 (2023-05-31)</h2>
136	* Enhancements to the [/help?cmd=/timeline\|/timeline webpage]: <ol type="a">
137	<li> Add the ft=TAG query parameter which in combination with d=Y
	@@ -155,11 +177,11 @@
155	frees us from needing to build and install the container as root,
156	since it no longer has to create a private <tt>/dev</tt> tree
157	inside the jail for Fossil's use.
158	* Add support for the trigram tokenizer for FTS5 search to enable
159	searching in Chinese.
160	* Comment lines (starting with a '#') are now supported inside
161	[./settings.wiki#versionable\|versioned settings].
162	* Default permissions for anonymous users in new repositories are
163	changed to "hz".
164	* The [/help?cmd=status\|fossil status] command now detects when a
165	file used to be a symlink and has been replaced by a regular file.
166

	--- www/changes.wiki
	+++ www/changes.wiki
	@@ -1,8 +1,22 @@
1	<title>Change Log</title>
2
3	<h2 id='v2_26'>Changes for version 2.26 (pending)</h2>
4
5	* Added the [/help?cmd=/ckout\|/ckout web page] to provide information
6	about pending changes in a working check-out
7	* The [/help?cmd=ui\|fossil ui] command defaults to using the /ckout
8	page as its start page.
9	* Added the [/help?cmd=merge-info\|fossil merge-info] command and especially
10	the --tk option to that command, to provide analysis of the most recent
11	merge or update operation.
12	* Issue a warning if a user tries to commit on a check-in where the
13	branch has been changed.
14	* When a merge conflict occurs, a new section is added to the conflict
15	text that shows Fossil's suggested resolution to the conflict.
16
17	<h2 id='v2_25'>Changes for version 2.25 (2024-11-06)</h2>
18
19	* The "[/help?cmd=ui\|fossil ui /]" command now works even for repositories
20	that have non-ASCII filenames
21	* Add the [/help?cmd=tree\|fossil tree] command.
22	* On case-insensitive filesystems, store files using the filesystem's
	@@ -11,12 +25,21 @@
25	which is more familiar to Git users. Retain the legacy name for
26	compatibility.
27	* Add new query parameters to the [/help?cmd=/timeline\|/timeline page]:
28	d2=, p2=, and dp2=.
29	* Add options to the [/help?cmd=tag\|fossil tag] command that will list tag values.
30	* Add the -b\|--brief option to the [/help?cmd=status\|fossil status] command.
31	* Add ability to upload unversioned files via the [/help?cmd=/uvlist\|/uvlist page].
32	* Add history search to the [/help?cmd=/chat\|/chat page].
33	* Add Unix socket support to the [/help?cmd=server\|server command].
34	* On Windows, use the root certificates managed by the operating system
35	(requires OpenSSL 3.2.0 or greater).
36	* Take into account zero-width and double-width unicode characters when
37	formatting the command-line timeline.
38	* Update the built-in SQLite to version 3.47.0. Precompiled binaries are
39	linked against OpenSSL 3.4.0.
40	* Numerous minor fixes and additions.
41
42
43	<h2 id='v2_24'>Changes for version 2.24 (2024-04-23)</h2>
44
45	* Apache change work-around → As part of a security fix, the Apache webserver
	@@ -85,11 +108,11 @@
108	* Add ability to "close" forum threads, such that unprivileged users
109	may no longer respond to them. Only administrators can close
110	threads or respond to them by default, and the
111	[/help?cmd=forum-close-policy\|forum-close-policy setting] can be
112	used to add that capability to moderators.
113	* Add the [/help?cmd=all\|fossil all whatis] command.
114	* The [/help?cmd=status\|fossil status] command and relevant UI pages now
115	correctly report files which were both renamed <b>and</b> edited as such.
116	* Show default value of settings that have a default in
117	[/help?cmd=help\|fossil help SETTING] output.
118	* On timeline graphs, show closed check-ins using an X in the middle of the
	@@ -127,11 +150,10 @@
150	they work when the "remote" machine is a Mac and the "fossil"
151	executable is in the $HOME/bin directory.
152	</ul>
153	* Update built-in libraries SQLite, ZLib, Pikchr to their latest versions.
154	* Documentation enhancements and typo fixes.

155
156
157	<h2 id='v2_22'>Changes for version 2.22 (2023-05-31)</h2>
158	* Enhancements to the [/help?cmd=/timeline\|/timeline webpage]: <ol type="a">
159	<li> Add the ft=TAG query parameter which in combination with d=Y
	@@ -155,11 +177,11 @@
177	frees us from needing to build and install the container as root,
178	since it no longer has to create a private <tt>/dev</tt> tree
179	inside the jail for Fossil's use.
180	* Add support for the trigram tokenizer for FTS5 search to enable
181	searching in Chinese.
182	* Comment lines (starting with a '#') are now supported inside
183	[./settings.wiki#versionable\|versioned settings].
184	* Default permissions for anonymous users in new repositories are
185	changed to "hz".
186	* The [/help?cmd=status\|fossil status] command now detects when a
187	file used to be a symlink and has been replaced by a regular file.
188

M www/chat.md

+2 -2

		--- www/chat.md
		+++ www/chat.md
		@@ -155,11 +155,11 @@
155	155	Substitute the appropriate project URL, robot account
156	156	name and password, message text and file attachment, of course.
157	157
158	158	### <a id="chat-robot"></a> Chat Messages For Timeline Events
159	159
160		-If the [chat-timeline-user setting](/help?cmd=chat-timeline-user) is not a
	160	+If the [chat-timeline-user setting](/help?cmd=chat-timeline-user) is not an
161	161	empty string, then any change to the repository that would normally result
162	162	in a new timeline entry is announced in the chatroom. The announcement
163	163	appears to come from a user whose name is given by the chat-timeline-user
164	164	setting.
165	165
		@@ -193,11 +193,11 @@
193	193	* [/chat-delete](/help?name=/chat-delete) →
194	194	Deletes a chat message.
195	195
196	196	Fossil chat uses the venerable "hanging GET" or
197	197	"[long polling](wikipedia:/wiki/Push_technology#Long_polling)"
198		-technique to recieve asynchronous notification of new messages.
	198	+technique to receive asynchronous notification of new messages.
199	199	This is done because long polling works well with CGI and SCGI,
200	200	which are the usual mechanisms for setting up a Fossil server.
201	201	More advanced notification techniques such as
202	202	[Server-sent events](wikipedia:/wiki/Server-sent_events) and especially
203	203	[WebSockets](wikipedia:/wiki/WebSocket) might seem more appropriate for
204	204

	--- www/chat.md
	+++ www/chat.md
	@@ -155,11 +155,11 @@
155	Substitute the appropriate project URL, robot account
156	name and password, message text and file attachment, of course.
157
158	### <a id="chat-robot"></a> Chat Messages For Timeline Events
159
160	If the [chat-timeline-user setting](/help?cmd=chat-timeline-user) is not a
161	empty string, then any change to the repository that would normally result
162	in a new timeline entry is announced in the chatroom. The announcement
163	appears to come from a user whose name is given by the chat-timeline-user
164	setting.
165
	@@ -193,11 +193,11 @@
193	* [/chat-delete](/help?name=/chat-delete) →
194	Deletes a chat message.
195
196	Fossil chat uses the venerable "hanging GET" or
197	"[long polling](wikipedia:/wiki/Push_technology#Long_polling)"
198	technique to recieve asynchronous notification of new messages.
199	This is done because long polling works well with CGI and SCGI,
200	which are the usual mechanisms for setting up a Fossil server.
201	More advanced notification techniques such as
202	[Server-sent events](wikipedia:/wiki/Server-sent_events) and especially
203	[WebSockets](wikipedia:/wiki/WebSocket) might seem more appropriate for
204

	--- www/chat.md
	+++ www/chat.md
	@@ -155,11 +155,11 @@
155	Substitute the appropriate project URL, robot account
156	name and password, message text and file attachment, of course.
157
158	### <a id="chat-robot"></a> Chat Messages For Timeline Events
159
160	If the [chat-timeline-user setting](/help?cmd=chat-timeline-user) is not an
161	empty string, then any change to the repository that would normally result
162	in a new timeline entry is announced in the chatroom. The announcement
163	appears to come from a user whose name is given by the chat-timeline-user
164	setting.
165
	@@ -193,11 +193,11 @@
193	* [/chat-delete](/help?name=/chat-delete) →
194	Deletes a chat message.
195
196	Fossil chat uses the venerable "hanging GET" or
197	"[long polling](wikipedia:/wiki/Push_technology#Long_polling)"
198	technique to receive asynchronous notification of new messages.
199	This is done because long polling works well with CGI and SCGI,
200	which are the usual mechanisms for setting up a Fossil server.
201	More advanced notification techniques such as
202	[Server-sent events](wikipedia:/wiki/Server-sent_events) and especially
203	[WebSockets](wikipedia:/wiki/WebSocket) might seem more appropriate for
204

M www/checkin_names.wiki

+1 -1

		--- www/checkin_names.wiki
		+++ www/checkin_names.wiki
		@@ -246,11 +246,11 @@
246	246
247	247	<pre>
248	248	fossil info merge-in:xyzzy:2022-03-01
249	249	</pre>
250	250
251		-to get informations about the most recent merge-in point on the branch
	251	+to get information about the most recent merge-in point on the branch
252	252	"xyzzy" that happened on or before March 1, 2022.
253	253
254	254	<h2 id="special">Special Tags</h2>
255	255
256	256	The tag "tip" means the most recent check-in. The "tip" tag is practically
257	257

	--- www/checkin_names.wiki
	+++ www/checkin_names.wiki
	@@ -246,11 +246,11 @@
246
247	<pre>
248	fossil info merge-in:xyzzy:2022-03-01
249	</pre>
250
251	to get informations about the most recent merge-in point on the branch
252	"xyzzy" that happened on or before March 1, 2022.
253
254	<h2 id="special">Special Tags</h2>
255
256	The tag "tip" means the most recent check-in. The "tip" tag is practically
257

	--- www/checkin_names.wiki
	+++ www/checkin_names.wiki
	@@ -246,11 +246,11 @@
246
247	<pre>
248	fossil info merge-in:xyzzy:2022-03-01
249	</pre>
250
251	to get information about the most recent merge-in point on the branch
252	"xyzzy" that happened on or before March 1, 2022.
253
254	<h2 id="special">Special Tags</h2>
255
256	The tag "tip" means the most recent check-in. The "tip" tag is practically
257

M www/concepts.wiki

+4 -4

		--- www/concepts.wiki
		+++ www/concepts.wiki
		@@ -163,11 +163,11 @@
163	163	The manifest file is not normally a real file on disk. Instead,
164	164	the manifest is computed in memory by Fossil whenever it needs it.
165	165	However, the "fossil setting manifest on" command will cause the
166	166	manifest file to be materialized to disk, if desired. Both Fossil
167	167	itself, and SQLite cause the manifest file to be materialized to disk
168		-so that the makefiles for these project can read the manifest and
	168	+so that the makefiles for these projects can read the manifest and
169	169	embed version information in generated binaries.
170	170
171	171	Fossil automatically generates a manifest whenever you "commit"
172	172	a new check-in. So this is not something that you, the developer,
173	173	need to worry with. The format of a manifest is intentionally
		@@ -178,11 +178,11 @@
178	178
179	179	In addition to identifying all files in the check-in, a
180	180	manifest also contains a check-in comment, the date and time
181	181	when the check-in was established, who created the check-in,
182	182	and links to other check-ins from which the current check-in
183		-is derived. There is also a couple of checksums used to verify
	183	+is derived. There are also a couple of checksums used to verify
184	184	the integrity of the check-in. And the whole manifest might
185	185	be PGP clearsigned.
186	186
187	187	<h3 id="keyconc">2.3 Key concepts</h3>
188	188
		@@ -217,11 +217,11 @@
217	217	SQLite, patch, or any similar software on your system in order to use
218	218	Fossil effectively. You will want to have some kind of text editor
219	219	for entering check-in comments. Fossil will use whatever text editor
220	220	is identified by your VISUAL environment variable. Fossil will also
221	221	use GPG to clearsign your manifests if you happen to have it installed,
222		-but Fossil will skip that step if GPG missing from your system.
	222	+but Fossil will skip that step if GPG is missing from your system.
223	223	You can optionally set up Fossil to use external "diff" programs,
224	224	though Fossil has an excellent built-in "diff" algorithm that works
225	225	fine for most people. If you happen to have Tcl/Tk installed on your
226	226	system, Fossil will use it to generate a graphical "diff" display when
227	227	you use the --tk option to the "diff" command, but this too is entirely
		@@ -286,11 +286,11 @@
286	286	fossil setting autosync off
287	287	fossil settings
288	288	</pre>
289	289
290	290	By default, Fossil runs with autosync mode turned on. The
291		-authors finds that projects run more smoothly in autosync mode since
	291	+authors find that projects run more smoothly in autosync mode since
292	292	autosync helps to prevent pointless forking and merging and helps keeps
293	293	all collaborators working on exactly the same code rather than on their
294	294	own personal forks of the code. In the author's view, manual-merge mode
295	295	should be reserved for disconnected operation.
296	296
297	297

	--- www/concepts.wiki
	+++ www/concepts.wiki
	@@ -163,11 +163,11 @@
163	The manifest file is not normally a real file on disk. Instead,
164	the manifest is computed in memory by Fossil whenever it needs it.
165	However, the "fossil setting manifest on" command will cause the
166	manifest file to be materialized to disk, if desired. Both Fossil
167	itself, and SQLite cause the manifest file to be materialized to disk
168	so that the makefiles for these project can read the manifest and
169	embed version information in generated binaries.
170
171	Fossil automatically generates a manifest whenever you "commit"
172	a new check-in. So this is not something that you, the developer,
173	need to worry with. The format of a manifest is intentionally
	@@ -178,11 +178,11 @@
178
179	In addition to identifying all files in the check-in, a
180	manifest also contains a check-in comment, the date and time
181	when the check-in was established, who created the check-in,
182	and links to other check-ins from which the current check-in
183	is derived. There is also a couple of checksums used to verify
184	the integrity of the check-in. And the whole manifest might
185	be PGP clearsigned.
186
187	<h3 id="keyconc">2.3 Key concepts</h3>
188
	@@ -217,11 +217,11 @@
217	SQLite, patch, or any similar software on your system in order to use
218	Fossil effectively. You will want to have some kind of text editor
219	for entering check-in comments. Fossil will use whatever text editor
220	is identified by your VISUAL environment variable. Fossil will also
221	use GPG to clearsign your manifests if you happen to have it installed,
222	but Fossil will skip that step if GPG missing from your system.
223	You can optionally set up Fossil to use external "diff" programs,
224	though Fossil has an excellent built-in "diff" algorithm that works
225	fine for most people. If you happen to have Tcl/Tk installed on your
226	system, Fossil will use it to generate a graphical "diff" display when
227	you use the --tk option to the "diff" command, but this too is entirely
	@@ -286,11 +286,11 @@
286	fossil setting autosync off
287	fossil settings
288	</pre>
289
290	By default, Fossil runs with autosync mode turned on. The
291	authors finds that projects run more smoothly in autosync mode since
292	autosync helps to prevent pointless forking and merging and helps keeps
293	all collaborators working on exactly the same code rather than on their
294	own personal forks of the code. In the author's view, manual-merge mode
295	should be reserved for disconnected operation.
296
297

	--- www/concepts.wiki
	+++ www/concepts.wiki
	@@ -163,11 +163,11 @@
163	The manifest file is not normally a real file on disk. Instead,
164	the manifest is computed in memory by Fossil whenever it needs it.
165	However, the "fossil setting manifest on" command will cause the
166	manifest file to be materialized to disk, if desired. Both Fossil
167	itself, and SQLite cause the manifest file to be materialized to disk
168	so that the makefiles for these projects can read the manifest and
169	embed version information in generated binaries.
170
171	Fossil automatically generates a manifest whenever you "commit"
172	a new check-in. So this is not something that you, the developer,
173	need to worry with. The format of a manifest is intentionally
	@@ -178,11 +178,11 @@
178
179	In addition to identifying all files in the check-in, a
180	manifest also contains a check-in comment, the date and time
181	when the check-in was established, who created the check-in,
182	and links to other check-ins from which the current check-in
183	is derived. There are also a couple of checksums used to verify
184	the integrity of the check-in. And the whole manifest might
185	be PGP clearsigned.
186
187	<h3 id="keyconc">2.3 Key concepts</h3>
188
	@@ -217,11 +217,11 @@
217	SQLite, patch, or any similar software on your system in order to use
218	Fossil effectively. You will want to have some kind of text editor
219	for entering check-in comments. Fossil will use whatever text editor
220	is identified by your VISUAL environment variable. Fossil will also
221	use GPG to clearsign your manifests if you happen to have it installed,
222	but Fossil will skip that step if GPG is missing from your system.
223	You can optionally set up Fossil to use external "diff" programs,
224	though Fossil has an excellent built-in "diff" algorithm that works
225	fine for most people. If you happen to have Tcl/Tk installed on your
226	system, Fossil will use it to generate a graphical "diff" display when
227	you use the --tk option to the "diff" command, but this too is entirely
	@@ -286,11 +286,11 @@
286	fossil setting autosync off
287	fossil settings
288	</pre>
289
290	By default, Fossil runs with autosync mode turned on. The
291	authors find that projects run more smoothly in autosync mode since
292	autosync helps to prevent pointless forking and merging and helps keeps
293	all collaborators working on exactly the same code rather than on their
294	own personal forks of the code. In the author's view, manual-merge mode
295	should be reserved for disconnected operation.
296
297

M www/contribute.wiki

+1 -1

		--- www/contribute.wiki
		+++ www/contribute.wiki
		@@ -1,8 +1,8 @@
1	1	<title>Contributing To Fossil</title>
2	2
3		-Fossil users are encouraged to contributed enhancements back to the
	3	+Fossil users are encouraged to contribute enhancements back to the
4	4	project. This note outlines some of the procedures for making
5	5	useful contributions.
6	6
7	7	<h2>1.0 Contributor Agreement</h2>
8	8
9	9

	--- www/contribute.wiki
	+++ www/contribute.wiki
	@@ -1,8 +1,8 @@
1	<title>Contributing To Fossil</title>
2
3	Fossil users are encouraged to contributed enhancements back to the
4	project. This note outlines some of the procedures for making
5	useful contributions.
6
7	<h2>1.0 Contributor Agreement</h2>
8
9

	--- www/contribute.wiki
	+++ www/contribute.wiki
	@@ -1,8 +1,8 @@
1	<title>Contributing To Fossil</title>
2
3	Fossil users are encouraged to contribute enhancements back to the
4	project. This note outlines some of the procedures for making
5	useful contributions.
6
7	<h2>1.0 Contributor Agreement</h2>
8
9

M www/custom_ticket.wiki

+10 -8

		--- www/custom_ticket.wiki
		+++ www/custom_ticket.wiki
		@@ -82,16 +82,18 @@
82	82
83	83	Look for the text "Contact:" (about halfway through). Then insert these lines
84	84	after the closing tr tag and before the "enable_output" line:
85	85
86	86	<verbatim>
87		-<td align="right">Assigned to:</td><td bgcolor="#d0d0d0">
88		- $<assigned_to>
89		-</td>
90		-<td align="right">Opened by:</td><td bgcolor="#d0d0d0">
91		- $<opened_by>
92		-</td>
	87	+<tr>
	88	+ <td align="right">Assigned to:</td><td bgcolor="#d0d0d0">
	89	+ $<assigned_to>
	90	+ </td>
	91	+ <td align="right">Opened by:</td><td bgcolor="#d0d0d0">
	92	+ $<opened_by>
	93	+ </td>
	94	+</tr>
93	95	</verbatim>
94	96
95	97	This will add a row which displays these two fields, in the event the user has
96	98	<a href="./caps/ref.html#w">ticket "edit" capability</a>.
97	99
		@@ -109,12 +111,12 @@
109	111	</verbatim>
110	112
111	113	That will give you a drop-down list of assignees. The first argument to the TH1
112	114	command 'combobox' is the database field which the combobox is associated to.
113	115	The next argument is the list of choices you want to show in the combobox (and
114		-that you specified in the second step above. The last argument should be 1 for a
115		-true combobox (see the <a href="th1.md#combobox">TH1 documentation</a> for
	116	+that you specified in the second step above.) The last argument should be 1 for
	117	+a true combobox (see the <a href="th1.md#combobox">TH1 documentation</a> for
116	118	details).
117	119
118	120	Now, similar to the previous
119	121	section, look for "Contact:" and add this:
120	122
121	123

	--- www/custom_ticket.wiki
	+++ www/custom_ticket.wiki
	@@ -82,16 +82,18 @@
82
83	Look for the text "Contact:" (about halfway through). Then insert these lines
84	after the closing tr tag and before the "enable_output" line:
85
86	<verbatim>
87	<td align="right">Assigned to:</td><td bgcolor="#d0d0d0">
88	$<assigned_to>
89	</td>
90	<td align="right">Opened by:</td><td bgcolor="#d0d0d0">
91	$<opened_by>
92	</td>


93	</verbatim>
94
95	This will add a row which displays these two fields, in the event the user has
96	<a href="./caps/ref.html#w">ticket "edit" capability</a>.
97
	@@ -109,12 +111,12 @@
109	</verbatim>
110
111	That will give you a drop-down list of assignees. The first argument to the TH1
112	command 'combobox' is the database field which the combobox is associated to.
113	The next argument is the list of choices you want to show in the combobox (and
114	that you specified in the second step above. The last argument should be 1 for a
115	true combobox (see the <a href="th1.md#combobox">TH1 documentation</a> for
116	details).
117
118	Now, similar to the previous
119	section, look for "Contact:" and add this:
120
121

	--- www/custom_ticket.wiki
	+++ www/custom_ticket.wiki
	@@ -82,16 +82,18 @@
82
83	Look for the text "Contact:" (about halfway through). Then insert these lines
84	after the closing tr tag and before the "enable_output" line:
85
86	<verbatim>
87	<tr>
88	<td align="right">Assigned to:</td><td bgcolor="#d0d0d0">
89	$<assigned_to>
90	</td>
91	<td align="right">Opened by:</td><td bgcolor="#d0d0d0">
92	$<opened_by>
93	</td>
94	</tr>
95	</verbatim>
96
97	This will add a row which displays these two fields, in the event the user has
98	<a href="./caps/ref.html#w">ticket "edit" capability</a>.
99
	@@ -109,12 +111,12 @@
111	</verbatim>
112
113	That will give you a drop-down list of assignees. The first argument to the TH1
114	command 'combobox' is the database field which the combobox is associated to.
115	The next argument is the list of choices you want to show in the combobox (and
116	that you specified in the second step above.) The last argument should be 1 for
117	a true combobox (see the <a href="th1.md#combobox">TH1 documentation</a> for
118	details).
119
120	Now, similar to the previous
121	section, look for "Contact:" and add this:
122
123

M www/embeddeddoc.wiki

+13 -5

		--- www/embeddeddoc.wiki
		+++ www/embeddeddoc.wiki
		@@ -178,17 +178,25 @@
178	178	As with "$ROOT", this substitution only works for Markdown and HTML
179	179	documents. For Wiki documents, you would need to use a relative URL.
180	180
181	181	<h2 id="th1">2.3 TH1 Documents</h2>
182	182
183		-Fossil will substitute the value of [./th1.md \| TH1 expressions] within
184		-<tt>{</tt> curly braces <tt>}</tt> into the output HTML if you have
185		-configured it with the <tt>--with-th1-docs</tt> option, which is
186		-disabled by default.
	183	+Enabling TH1 document support requires the following:
	184	+
	185	+ * Configure the build with the <code>--with-th1-docs</code> flag.
	186	+ * Enable the <code>th1-docs</code> setting, which is only available
	187	+ after building with <code>--with-th1-docs</code>.
	188	+ * Affected files must have a <code>.th1</code> file extension.
	189	+ * The code to run must be embedded in blocks of
	190	+ <code><th1>...</th1></code>.
	191	+
	192	+Fossil will substitute the value of [./th1.md \| TH1 expressions]
	193	+within the <code><th1>...</th1></code> blocks into
	194	+the output HTML.
187	195
188	196	Since TH1 is a full scripting language, this feature essential grants
189		-the ability to execute code on the server to anyone with check-in
	197	+the ability to execute code on the server to anyone with check-in
190	198	privilege for the project.
191	199	This is a security risk that needs to be carefully managed.
192	200	The feature is off by default.
193	201	Administrators should understand and carefully assess the risks
194	202	before enabling the use of TH1 within embedded documentation.
195	203

	--- www/embeddeddoc.wiki
	+++ www/embeddeddoc.wiki
	@@ -178,17 +178,25 @@
178	As with "$ROOT", this substitution only works for Markdown and HTML
179	documents. For Wiki documents, you would need to use a relative URL.
180
181	<h2 id="th1">2.3 TH1 Documents</h2>
182
183	Fossil will substitute the value of [./th1.md \| TH1 expressions] within
184	<tt>{</tt> curly braces <tt>}</tt> into the output HTML if you have
185	configured it with the <tt>--with-th1-docs</tt> option, which is
186	disabled by default.








187
188	Since TH1 is a full scripting language, this feature essential grants
189	the ability to execute code on the server to anyone with check-in
190	privilege for the project.
191	This is a security risk that needs to be carefully managed.
192	The feature is off by default.
193	Administrators should understand and carefully assess the risks
194	before enabling the use of TH1 within embedded documentation.
195

	--- www/embeddeddoc.wiki
	+++ www/embeddeddoc.wiki
	@@ -178,17 +178,25 @@
178	As with "$ROOT", this substitution only works for Markdown and HTML
179	documents. For Wiki documents, you would need to use a relative URL.
180
181	<h2 id="th1">2.3 TH1 Documents</h2>
182
183	Enabling TH1 document support requires the following:
184
185	* Configure the build with the <code>--with-th1-docs</code> flag.
186	* Enable the <code>th1-docs</code> setting, which is only available
187	after building with <code>--with-th1-docs</code>.
188	* Affected files must have a <code>.th1</code> file extension.
189	* The code to run must be embedded in blocks of
190	<code><th1>...</th1></code>.
191
192	Fossil will substitute the value of [./th1.md \| TH1 expressions]
193	within the <code><th1>...</th1></code> blocks into
194	the output HTML.
195
196	Since TH1 is a full scripting language, this feature essential grants
197	the ability to execute code on the server to anyone with check-in
198	privilege for the project.
199	This is a security risk that needs to be carefully managed.
200	The feature is off by default.
201	Administrators should understand and carefully assess the risks
202	before enabling the use of TH1 within embedded documentation.
203

M www/env-opts.md

+3 -3

		--- www/env-opts.md
		+++ www/env-opts.md
		@@ -468,21 +468,21 @@
468	468	obviously as part of the `fossil ui` command. In that specific case,
469	469	the browser is launched pointing at the web server started by `fossil
470	470	ui` listening on a private TCP port.
471	471
472	472	On all platforms, if the local or global settings `web-browser` is
473		-set, that is the command used to open an URL.
	473	+set, that is the command used to open a URL.
474	474
475	475	Otherwise, the specific actions vary by platform.
476	476
477	477	On Unix-like platforms other than Apple's, it looks for the first
478	478	program from the list `xdg-open`, `gnome-open`, `firefox`, and
479	479	`google-chrome` that it can find on the `PATH`.
480	480
481		-On Apple platforms, it assumes that `open` is the command to open an
	481	+On Apple platforms, it assumes that `open` is the command to open a
482	482	URL in the user's configured default browser.
483	483
484	484	On Windows platforms, it assumes that `start` is the command to open
485		-an URL in the user's configured default browser.
	485	+a URL in the user's configured default browser.
486	486
487	487	[configdb]: ./tech_overview.wiki#configdb
488	488	[configloc]: ./tech_overview.wiki#configloc
489	489

	--- www/env-opts.md
	+++ www/env-opts.md
	@@ -468,21 +468,21 @@
468	obviously as part of the `fossil ui` command. In that specific case,
469	the browser is launched pointing at the web server started by `fossil
470	ui` listening on a private TCP port.
471
472	On all platforms, if the local or global settings `web-browser` is
473	set, that is the command used to open an URL.
474
475	Otherwise, the specific actions vary by platform.
476
477	On Unix-like platforms other than Apple's, it looks for the first
478	program from the list `xdg-open`, `gnome-open`, `firefox`, and
479	`google-chrome` that it can find on the `PATH`.
480
481	On Apple platforms, it assumes that `open` is the command to open an
482	URL in the user's configured default browser.
483
484	On Windows platforms, it assumes that `start` is the command to open
485	an URL in the user's configured default browser.
486
487	[configdb]: ./tech_overview.wiki#configdb
488	[configloc]: ./tech_overview.wiki#configloc
489

	--- www/env-opts.md
	+++ www/env-opts.md
	@@ -468,21 +468,21 @@
468	obviously as part of the `fossil ui` command. In that specific case,
469	the browser is launched pointing at the web server started by `fossil
470	ui` listening on a private TCP port.
471
472	On all platforms, if the local or global settings `web-browser` is
473	set, that is the command used to open a URL.
474
475	Otherwise, the specific actions vary by platform.
476
477	On Unix-like platforms other than Apple's, it looks for the first
478	program from the list `xdg-open`, `gnome-open`, `firefox`, and
479	`google-chrome` that it can find on the `PATH`.
480
481	On Apple platforms, it assumes that `open` is the command to open a
482	URL in the user's configured default browser.
483
484	On Windows platforms, it assumes that `start` is the command to open
485	a URL in the user's configured default browser.
486
487	[configdb]: ./tech_overview.wiki#configdb
488	[configloc]: ./tech_overview.wiki#configloc
489

M www/event.wiki

+1 -1

		--- www/event.wiki
		+++ www/event.wiki
		@@ -115,6 +115,6 @@
115	115	create or edit technotes. In addition, users must have create-wiki
116	116	privilege (permission "f") to create new technotes and edit-wiki
117	117	privilege (permission "k") in order to edit existing technotes.
118	118
119	119	Technote content may be formatted as [/wiki_rules \| Fossil wiki],
120		-[/md_rules \| Markdown], or a plain text.
	120	+[/md_rules \| Markdown], or plain text.
121	121

	--- www/event.wiki
	+++ www/event.wiki
	@@ -115,6 +115,6 @@
115	create or edit technotes. In addition, users must have create-wiki
116	privilege (permission "f") to create new technotes and edit-wiki
117	privilege (permission "k") in order to edit existing technotes.
118
119	Technote content may be formatted as [/wiki_rules \| Fossil wiki],
120	[/md_rules \| Markdown], or a plain text.
121

	--- www/event.wiki
	+++ www/event.wiki
	@@ -115,6 +115,6 @@
115	create or edit technotes. In addition, users must have create-wiki
116	privilege (permission "f") to create new technotes and edit-wiki
117	privilege (permission "k") in order to edit existing technotes.
118
119	Technote content may be formatted as [/wiki_rules \| Fossil wiki],
120	[/md_rules \| Markdown], or plain text.
121

M www/fileedit-page.md

+3 -3

		--- www/fileedit-page.md
		+++ www/fileedit-page.md
		@@ -1,11 +1,11 @@
1	1	# The fileedit Page
2	2
3	3	This document describes the limitations of, caveats for, and
4	4	disclaimers for the [](/fileedit) page, which provides users with
5		-[checkin privileges](./caps/index.md) basic editing features for files
6		-via the web interface.
	5	+ basic editing features for files via the web interface when they
	6	+ have [checkin privileges](./caps/index.md).
7	7
8	8	# Important Caveats and Disclaimers
9	9
10	10	Predictably, the ability to edit files in a repository from a web
11	11	browser halfway around the world comes with several obligatory caveats
		@@ -219,11 +219,11 @@
219	219	- `element`: the DOM element in which the preview is rendered.
220	220	- `mimetype`: the mimetype of the being-previewed content, as determined
221	221	by Fossil (by its file extension).
222	222
223	223	The event listener callback shown above doesn't use the `mimetype`,
224		-but makes used of the other two. It fishes all `code` blocks out of
	224	+but makes use of the other two. It fishes all `code` blocks out of
225	225	the preview which explicitly have a CSS class named
226	226	`language-`something, and then asks highlightjs to highlight them.
227	227
228	228	## <a id="editor"></a> Integrating a Custom Editor Widget
229	229
230	230

	--- www/fileedit-page.md
	+++ www/fileedit-page.md
	@@ -1,11 +1,11 @@
1	# The fileedit Page
2
3	This document describes the limitations of, caveats for, and
4	disclaimers for the [](/fileedit) page, which provides users with
5	[checkin privileges](./caps/index.md) basic editing features for files
6	via the web interface.
7
8	# Important Caveats and Disclaimers
9
10	Predictably, the ability to edit files in a repository from a web
11	browser halfway around the world comes with several obligatory caveats
	@@ -219,11 +219,11 @@
219	- `element`: the DOM element in which the preview is rendered.
220	- `mimetype`: the mimetype of the being-previewed content, as determined
221	by Fossil (by its file extension).
222
223	The event listener callback shown above doesn't use the `mimetype`,
224	but makes used of the other two. It fishes all `code` blocks out of
225	the preview which explicitly have a CSS class named
226	`language-`something, and then asks highlightjs to highlight them.
227
228	## <a id="editor"></a> Integrating a Custom Editor Widget
229
230

	--- www/fileedit-page.md
	+++ www/fileedit-page.md
	@@ -1,11 +1,11 @@
1	# The fileedit Page
2
3	This document describes the limitations of, caveats for, and
4	disclaimers for the [](/fileedit) page, which provides users with
5	basic editing features for files via the web interface when they
6	have [checkin privileges](./caps/index.md).
7
8	# Important Caveats and Disclaimers
9
10	Predictably, the ability to edit files in a repository from a web
11	browser halfway around the world comes with several obligatory caveats
	@@ -219,11 +219,11 @@
219	- `element`: the DOM element in which the preview is rendered.
220	- `mimetype`: the mimetype of the being-previewed content, as determined
221	by Fossil (by its file extension).
222
223	The event listener callback shown above doesn't use the `mimetype`,
224	but makes use of the other two. It fishes all `code` blocks out of
225	the preview which explicitly have a CSS class named
226	`language-`something, and then asks highlightjs to highlight them.
227
228	## <a id="editor"></a> Integrating a Custom Editor Widget
229
230

M www/fileformat.wiki

+6 -6

		--- www/fileformat.wiki
		+++ www/fileformat.wiki
		@@ -360,11 +360,11 @@
360	360	of text in the wiki page. That text follows the newline character
361	361	that terminates the <b>W</b> card. The wiki text is always followed by one
362	362	extra newline.
363	363
364	364	The <b>C</b> card on a wiki page is optional. The argument is a comment
365		-that explains why the changes was made. The ability to have a <b>C</b>
	365	+that explains why the changes were made. The ability to have a <b>C</b>
366	366	card on a wiki page artifact was added on 2019-12-02 at the suggestion
367	367	of user George Krivov and is not currently used or generated by the
368	368	implementation. Older versions of Fossil will reject a wiki-page
369	369	artifact that includes a <b>C</b> card.
370	370
		@@ -396,14 +396,14 @@
396	396	ticket into existence.
397	397
398	398	<b>J</b> cards specify changes to the "value" of "fields" in the ticket.
399	399	If the <i>value</i> parameter of the <b>J</b> card is omitted, then the
400	400	field is set to an empty string.
401		-Each fossil server has a ticket configuration which specifies the fields its
	401	+Each fossil server has a ticket configuration which specifies the fields it
402	402	understands. The ticket configuration is part of the local state for
403	403	the repository and thus can vary from one repository to another.
404		-Hence a <b>J</b> card might specify a <i>field</i> that do not exist in the
	404	+Hence a <b>J</b> card might specify a <i>field</i> that does not exist in the
405	405	local ticket configuration. If a <b>J</b> card specifies a <i>field</i> that
406	406	is not in the local configuration, then that <b>J</b> card
407	407	is simply ignored.
408	408
409	409	The first argument of the <b>J</b> card is the field name. The second
		@@ -462,11 +462,11 @@
462	462	A technical note or "technote" artifact (formerly known as an "event" artifact)
463	463	associates a timeline comment and a page of text
464	464	(similar to a wiki page) with a point in time. Technotes can be used
465	465	to record project milestones, release notes, blog entries, process
466	466	checkpoints, or news articles.
467		-The following cards are allowed on an technote artifact:
	467	+The following cards are allowed on a technote artifact:
468	468
469	469	<div class="indent">
470	470	<b>C</b> <i>comment</i><br>
471	471	<b>D</b> <i>time-and-date-stamp</i><br />
472	472	<b>E</b> <i>technote-time</i> <i>technote-id</i><br />
		@@ -510,15 +510,15 @@
510	510	The <b>*</b> in place of the artifact ID indicates that
511	511	the tag or property applies to the current artifact. It is not
512	512	possible to encode the current artifact ID as part of an artifact,
513	513	since the act of inserting the artifact ID would change the artifact ID,
514	514	hence a <b>*</b> is used to represent "self". The "<b>+</b>" on the
515		-name means that tags can only be add and they can only be non-propagating
	515	+name means that tags can only be "add" and they can only be non-propagating
516	516	tags. In a technote, <b>T</b> cards are normally used to set the background
517	517	display color for timelines.
518	518
519		-The optional <b>U</b> card gives name of the user who entered the technote.
	519	+The optional <b>U</b> card gives the name of the user who entered the technote.
520	520
521	521	A single <b>W</b> card provides wiki text for the document associated with the
522	522	technote. The format of the <b>W</b> card is exactly the same as for a
523	523	[#wikichng \| wiki artifact].
524	524
525	525

	--- www/fileformat.wiki
	+++ www/fileformat.wiki
	@@ -360,11 +360,11 @@
360	of text in the wiki page. That text follows the newline character
361	that terminates the <b>W</b> card. The wiki text is always followed by one
362	extra newline.
363
364	The <b>C</b> card on a wiki page is optional. The argument is a comment
365	that explains why the changes was made. The ability to have a <b>C</b>
366	card on a wiki page artifact was added on 2019-12-02 at the suggestion
367	of user George Krivov and is not currently used or generated by the
368	implementation. Older versions of Fossil will reject a wiki-page
369	artifact that includes a <b>C</b> card.
370
	@@ -396,14 +396,14 @@
396	ticket into existence.
397
398	<b>J</b> cards specify changes to the "value" of "fields" in the ticket.
399	If the <i>value</i> parameter of the <b>J</b> card is omitted, then the
400	field is set to an empty string.
401	Each fossil server has a ticket configuration which specifies the fields its
402	understands. The ticket configuration is part of the local state for
403	the repository and thus can vary from one repository to another.
404	Hence a <b>J</b> card might specify a <i>field</i> that do not exist in the
405	local ticket configuration. If a <b>J</b> card specifies a <i>field</i> that
406	is not in the local configuration, then that <b>J</b> card
407	is simply ignored.
408
409	The first argument of the <b>J</b> card is the field name. The second
	@@ -462,11 +462,11 @@
462	A technical note or "technote" artifact (formerly known as an "event" artifact)
463	associates a timeline comment and a page of text
464	(similar to a wiki page) with a point in time. Technotes can be used
465	to record project milestones, release notes, blog entries, process
466	checkpoints, or news articles.
467	The following cards are allowed on an technote artifact:
468
469	<div class="indent">
470	<b>C</b> <i>comment</i><br>
471	<b>D</b> <i>time-and-date-stamp</i><br />
472	<b>E</b> <i>technote-time</i> <i>technote-id</i><br />
	@@ -510,15 +510,15 @@
510	The <b>*</b> in place of the artifact ID indicates that
511	the tag or property applies to the current artifact. It is not
512	possible to encode the current artifact ID as part of an artifact,
513	since the act of inserting the artifact ID would change the artifact ID,
514	hence a <b>*</b> is used to represent "self". The "<b>+</b>" on the
515	name means that tags can only be add and they can only be non-propagating
516	tags. In a technote, <b>T</b> cards are normally used to set the background
517	display color for timelines.
518
519	The optional <b>U</b> card gives name of the user who entered the technote.
520
521	A single <b>W</b> card provides wiki text for the document associated with the
522	technote. The format of the <b>W</b> card is exactly the same as for a
523	[#wikichng \| wiki artifact].
524
525

	--- www/fileformat.wiki
	+++ www/fileformat.wiki
	@@ -360,11 +360,11 @@
360	of text in the wiki page. That text follows the newline character
361	that terminates the <b>W</b> card. The wiki text is always followed by one
362	extra newline.
363
364	The <b>C</b> card on a wiki page is optional. The argument is a comment
365	that explains why the changes were made. The ability to have a <b>C</b>
366	card on a wiki page artifact was added on 2019-12-02 at the suggestion
367	of user George Krivov and is not currently used or generated by the
368	implementation. Older versions of Fossil will reject a wiki-page
369	artifact that includes a <b>C</b> card.
370
	@@ -396,14 +396,14 @@
396	ticket into existence.
397
398	<b>J</b> cards specify changes to the "value" of "fields" in the ticket.
399	If the <i>value</i> parameter of the <b>J</b> card is omitted, then the
400	field is set to an empty string.
401	Each fossil server has a ticket configuration which specifies the fields it
402	understands. The ticket configuration is part of the local state for
403	the repository and thus can vary from one repository to another.
404	Hence a <b>J</b> card might specify a <i>field</i> that does not exist in the
405	local ticket configuration. If a <b>J</b> card specifies a <i>field</i> that
406	is not in the local configuration, then that <b>J</b> card
407	is simply ignored.
408
409	The first argument of the <b>J</b> card is the field name. The second
	@@ -462,11 +462,11 @@
462	A technical note or "technote" artifact (formerly known as an "event" artifact)
463	associates a timeline comment and a page of text
464	(similar to a wiki page) with a point in time. Technotes can be used
465	to record project milestones, release notes, blog entries, process
466	checkpoints, or news articles.
467	The following cards are allowed on a technote artifact:
468
469	<div class="indent">
470	<b>C</b> <i>comment</i><br>
471	<b>D</b> <i>time-and-date-stamp</i><br />
472	<b>E</b> <i>technote-time</i> <i>technote-id</i><br />
	@@ -510,15 +510,15 @@
510	The <b>*</b> in place of the artifact ID indicates that
511	the tag or property applies to the current artifact. It is not
512	possible to encode the current artifact ID as part of an artifact,
513	since the act of inserting the artifact ID would change the artifact ID,
514	hence a <b>*</b> is used to represent "self". The "<b>+</b>" on the
515	name means that tags can only be "add" and they can only be non-propagating
516	tags. In a technote, <b>T</b> cards are normally used to set the background
517	display color for timelines.
518
519	The optional <b>U</b> card gives the name of the user who entered the technote.
520
521	A single <b>W</b> card provides wiki text for the document associated with the
522	technote. The format of the <b>W</b> card is exactly the same as for a
523	[#wikichng \| wiki artifact].
524
525

M www/forum.wiki

+1 -1

		--- www/forum.wiki
		+++ www/forum.wiki
		@@ -405,7 +405,7 @@
405	405
406	406	Though forum users are permitted to delete their own posts, they are
407	407	not permitted, without appropriate permissions, to close their own
408	408	posts. This is intentional, as closing one's own post can be used to
409	409	antagonize other forum users. For example, by posting something
410		-trollish or highly contraversial in nature and closing the post to
	410	+trollish or highly controversial in nature and closing the post to
411	411	further responses.
412	412

	--- www/forum.wiki
	+++ www/forum.wiki
	@@ -405,7 +405,7 @@
405
406	Though forum users are permitted to delete their own posts, they are
407	not permitted, without appropriate permissions, to close their own
408	posts. This is intentional, as closing one's own post can be used to
409	antagonize other forum users. For example, by posting something
410	trollish or highly contraversial in nature and closing the post to
411	further responses.
412

	--- www/forum.wiki
	+++ www/forum.wiki
	@@ -405,7 +405,7 @@
405
406	Though forum users are permitted to delete their own posts, they are
407	not permitted, without appropriate permissions, to close their own
408	posts. This is intentional, as closing one's own post can be used to
409	antagonize other forum users. For example, by posting something
410	trollish or highly controversial in nature and closing the post to
411	further responses.
412

M www/fossil-is-not-relational.md

+1 -1

		--- www/fossil-is-not-relational.md
		+++ www/fossil-is-not-relational.md
		@@ -131,11 +131,11 @@
131	131	metadata.
132	132
133	133	- Raw file content of versioned files. These data are external to
134	134	artifacts, which refer to them by their hashes. How they are stored
135	135	is not the concern of the data model, but (spoiler alert!) Fossil
136		- stores in them an sqlite database, one record per distinct hash, in
	136	+ stores them in an SQLite database, one record per distinct hash, in
137	137	its `blob` table (which we will cover more very soon).
138	138
139	139	Non-SCM-relevant state includes:
140	140
141	141	- Fossil's list of users and their metadata (permissions, email
142	142

	--- www/fossil-is-not-relational.md
	+++ www/fossil-is-not-relational.md
	@@ -131,11 +131,11 @@
131	metadata.
132
133	- Raw file content of versioned files. These data are external to
134	artifacts, which refer to them by their hashes. How they are stored
135	is not the concern of the data model, but (spoiler alert!) Fossil
136	stores in them an sqlite database, one record per distinct hash, in
137	its `blob` table (which we will cover more very soon).
138
139	Non-SCM-relevant state includes:
140
141	- Fossil's list of users and their metadata (permissions, email
142

	--- www/fossil-is-not-relational.md
	+++ www/fossil-is-not-relational.md
	@@ -131,11 +131,11 @@
131	metadata.
132
133	- Raw file content of versioned files. These data are external to
134	artifacts, which refer to them by their hashes. How they are stored
135	is not the concern of the data model, but (spoiler alert!) Fossil
136	stores them in an SQLite database, one record per distinct hash, in
137	its `blob` table (which we will cover more very soon).
138
139	Non-SCM-relevant state includes:
140
141	- Fossil's list of users and their metadata (permissions, email
142

M www/fossil-v-git.wiki

+9 -10

		--- www/fossil-v-git.wiki
		+++ www/fossil-v-git.wiki
		@@ -70,18 +70,13 @@
70	70	<td>Designed for Linux kernel development</td>
71	71	<td>Designed for SQLite development</td>
72	72	<td><a href="#scale">2.5.2 ↓</a></td>
73	73	</tr>
74	74	<tr>
75		- <td>Many contributors</td>
76		- <td>Select contributors</td>
77		- <td><a href="#contrib">2.5.3 ↓</a></td>
78		-</tr>
79		-<tr>
80	75	<td>Focus on individual branches</td>
81	76	<td>Focus on the entire tree of changes</td>
82		- <td><a href="#branches">2.5.4 ↓</a></td>
	77	+ <td><a href="#branches">2.5.3 ↓</a></td>
83	78	</tr>
84	79	<tr>
85	80	<td>One check-out per repository</td>
86	81	<td>Many check-outs per repository</td>
87	82	<td><a href="#checkouts">2.6 ↓</a></td>
		@@ -591,18 +586,21 @@
591	586	and its handful of active committers. Seeing all
592	587	changes on all branches all at once helps keep the whole team
593	588	up-to-date with what everybody else is doing, resulting in a more
594	589	tightly focused and cohesive implementation.
595	590
	591	+Parts of this section are [https://fossil-scm.org/forum/forumpost/5961e969fa\|disputed]
	592	+by [https://github.com/olorin37\|Jakub A. G.].
	593	+
596	594
597	595	<h3 id="checkouts">2.6 One vs. Many Check-outs per Repository</h3>
598	596
599	597	Because Git commingles the repository data with the initial checkout of
600	598	that repository, the default mode of operation in Git is to stick to that
601	599	single work/repo tree, even when that's a shortsighted way of working.
602	600
603		-Fossil doesn't work that way. A Fossil repository is a SQLite database
	601	+Fossil doesn't work that way. A Fossil repository is an SQLite database
604	602	file which is normally stored outside the working checkout directory. You can
605	603	[/help?cmd=open \| open] a Fossil repository any number of times into
606	604	any number of working directories. A common usage pattern is to have one
607	605	working directory per active working branch, so that switching branches
608	606	is done with a <tt>cd</tt> command rather than by checking out the
		@@ -660,10 +658,12 @@
660	658
661	659	Plus,
662	660	<tt>cd</tt> is faster to type than <tt>git checkout</tt> or <tt>fossil
663	661	update</tt>.
664	662
	663	+Parts of this section are [https://fossil-scm.org/forum/forumpost/5961e969fa\|disputed]
	664	+by [https://github.com/olorin37\|Jakub A. G.].
665	665
666	666	<h3 id="history">2.7 What you should have done vs. What you actually did</h3>
667	667
668	668	Git puts a lot of emphasis on maintaining
669	669	a "clean" check-in history. Extraneous and experimental branches by
		@@ -861,13 +861,12 @@
861	861	as this author is aware, but there is now
862	862	[https://lwn.net/ml/git/[email protected]/
863	863	\| a competing SHA-256 based plan] which requires complete repository
864	864	conversion from SHA-1 to SHA-256, breaking all public hashes in the
865	865	repo. One way to characterize such a massive upheaval in Git terms is a
866		-whole-project rebase, which violates
867		-[https://blog.axosoft.com/golden-rule-of-rebasing-in-git/ \| Git's own
868		-Golden Rule of Rebasing].
	866	+whole-project rebase, which violates the
	867	+[https://www.atlassian.com/git/tutorials/merging-vs-rebasing#the-golden-rule-of-rebasing\|Golden Rule of Rebasing].
869	868
870	869	Regardless of the eventual implementation details, we fully expect Git
871	870	to move off SHA-1 eventually and for the changes to take years more to
872	871	percolate through the community.
873	872
874	873

	--- www/fossil-v-git.wiki
	+++ www/fossil-v-git.wiki
	@@ -70,18 +70,13 @@
70	<td>Designed for Linux kernel development</td>
71	<td>Designed for SQLite development</td>
72	<td><a href="#scale">2.5.2 ↓</a></td>
73	</tr>
74	<tr>
75	<td>Many contributors</td>
76	<td>Select contributors</td>
77	<td><a href="#contrib">2.5.3 ↓</a></td>
78	</tr>
79	<tr>
80	<td>Focus on individual branches</td>
81	<td>Focus on the entire tree of changes</td>
82	<td><a href="#branches">2.5.4 ↓</a></td>
83	</tr>
84	<tr>
85	<td>One check-out per repository</td>
86	<td>Many check-outs per repository</td>
87	<td><a href="#checkouts">2.6 ↓</a></td>
	@@ -591,18 +586,21 @@
591	and its handful of active committers. Seeing all
592	changes on all branches all at once helps keep the whole team
593	up-to-date with what everybody else is doing, resulting in a more
594	tightly focused and cohesive implementation.
595



596
597	<h3 id="checkouts">2.6 One vs. Many Check-outs per Repository</h3>
598
599	Because Git commingles the repository data with the initial checkout of
600	that repository, the default mode of operation in Git is to stick to that
601	single work/repo tree, even when that's a shortsighted way of working.
602
603	Fossil doesn't work that way. A Fossil repository is a SQLite database
604	file which is normally stored outside the working checkout directory. You can
605	[/help?cmd=open \| open] a Fossil repository any number of times into
606	any number of working directories. A common usage pattern is to have one
607	working directory per active working branch, so that switching branches
608	is done with a <tt>cd</tt> command rather than by checking out the
	@@ -660,10 +658,12 @@
660
661	Plus,
662	<tt>cd</tt> is faster to type than <tt>git checkout</tt> or <tt>fossil
663	update</tt>.
664


665
666	<h3 id="history">2.7 What you should have done vs. What you actually did</h3>
667
668	Git puts a lot of emphasis on maintaining
669	a "clean" check-in history. Extraneous and experimental branches by
	@@ -861,13 +861,12 @@
861	as this author is aware, but there is now
862	[https://lwn.net/ml/git/[email protected]/
863	\| a competing SHA-256 based plan] which requires complete repository
864	conversion from SHA-1 to SHA-256, breaking all public hashes in the
865	repo. One way to characterize such a massive upheaval in Git terms is a
866	whole-project rebase, which violates
867	[https://blog.axosoft.com/golden-rule-of-rebasing-in-git/ \| Git's own
868	Golden Rule of Rebasing].
869
870	Regardless of the eventual implementation details, we fully expect Git
871	to move off SHA-1 eventually and for the changes to take years more to
872	percolate through the community.
873
874

	--- www/fossil-v-git.wiki
	+++ www/fossil-v-git.wiki
	@@ -70,18 +70,13 @@
70	<td>Designed for Linux kernel development</td>
71	<td>Designed for SQLite development</td>
72	<td><a href="#scale">2.5.2 ↓</a></td>
73	</tr>
74	<tr>





75	<td>Focus on individual branches</td>
76	<td>Focus on the entire tree of changes</td>
77	<td><a href="#branches">2.5.3 ↓</a></td>
78	</tr>
79	<tr>
80	<td>One check-out per repository</td>
81	<td>Many check-outs per repository</td>
82	<td><a href="#checkouts">2.6 ↓</a></td>
	@@ -591,18 +586,21 @@
586	and its handful of active committers. Seeing all
587	changes on all branches all at once helps keep the whole team
588	up-to-date with what everybody else is doing, resulting in a more
589	tightly focused and cohesive implementation.
590
591	Parts of this section are [https://fossil-scm.org/forum/forumpost/5961e969fa\|disputed]
592	by [https://github.com/olorin37\|Jakub A. G.].
593
594
595	<h3 id="checkouts">2.6 One vs. Many Check-outs per Repository</h3>
596
597	Because Git commingles the repository data with the initial checkout of
598	that repository, the default mode of operation in Git is to stick to that
599	single work/repo tree, even when that's a shortsighted way of working.
600
601	Fossil doesn't work that way. A Fossil repository is an SQLite database
602	file which is normally stored outside the working checkout directory. You can
603	[/help?cmd=open \| open] a Fossil repository any number of times into
604	any number of working directories. A common usage pattern is to have one
605	working directory per active working branch, so that switching branches
606	is done with a <tt>cd</tt> command rather than by checking out the
	@@ -660,10 +658,12 @@
658
659	Plus,
660	<tt>cd</tt> is faster to type than <tt>git checkout</tt> or <tt>fossil
661	update</tt>.
662
663	Parts of this section are [https://fossil-scm.org/forum/forumpost/5961e969fa\|disputed]
664	by [https://github.com/olorin37\|Jakub A. G.].
665
666	<h3 id="history">2.7 What you should have done vs. What you actually did</h3>
667
668	Git puts a lot of emphasis on maintaining
669	a "clean" check-in history. Extraneous and experimental branches by
	@@ -861,13 +861,12 @@
861	as this author is aware, but there is now
862	[https://lwn.net/ml/git/[email protected]/
863	\| a competing SHA-256 based plan] which requires complete repository
864	conversion from SHA-1 to SHA-256, breaking all public hashes in the
865	repo. One way to characterize such a massive upheaval in Git terms is a
866	whole-project rebase, which violates the
867	[https://www.atlassian.com/git/tutorials/merging-vs-rebasing#the-golden-rule-of-rebasing\|Golden Rule of Rebasing].

868
869	Regardless of the eventual implementation details, we fully expect Git
870	to move off SHA-1 eventually and for the changes to take years more to
871	percolate through the community.
872
873

M www/fossil_prompt.wiki

+1 -1

		--- www/fossil_prompt.wiki
		+++ www/fossil_prompt.wiki
		@@ -17,6 +17,6 @@
17	17
18	18	For a permanent installation, you can graft the code into your
19	19	<tt>.bashrc</tt> file in your home directory.
20	20
21	21	The code is very simple (only 32 non-comment lines, as of this writing)
22		-and hence easy to customized.
	22	+and hence easy to customize.
23	23

	--- www/fossil_prompt.wiki
	+++ www/fossil_prompt.wiki
	@@ -17,6 +17,6 @@
17
18	For a permanent installation, you can graft the code into your
19	<tt>.bashrc</tt> file in your home directory.
20
21	The code is very simple (only 32 non-comment lines, as of this writing)
22	and hence easy to customized.
23

	--- www/fossil_prompt.wiki
	+++ www/fossil_prompt.wiki
	@@ -17,6 +17,6 @@
17
18	For a permanent installation, you can graft the code into your
19	<tt>.bashrc</tt> file in your home directory.
20
21	The code is very simple (only 32 non-comment lines, as of this writing)
22	and hence easy to customize.
23

M www/gitusers.md

+3 -3

		--- www/gitusers.md
		+++ www/gitusers.md
		@@ -56,11 +56,11 @@
56	56
57	57
58	58
59	59	#### <a id="cwork" name="scw"></a> Checkout Workflows
60	60
61		-A Fossil repository is a SQLite database storing the entire history of a
	61	+A Fossil repository is an SQLite database storing the entire history of a
62	62	project. It is not normally stored inside the working tree.
63	63	A Fossil working tree — [also called a check-out](./glossary.md#check-out) — is a directory
64	64	that contains a snapshot of your project that you are currently working
65	65	on, extracted for you from the repository database file by the `fossil`
66	66	program.
		@@ -148,11 +148,11 @@
148	148	option, it won’t let you close the check-out with uncommitted changes to
149	149	those managed files.
150	150
151	151	The `close` command also refuses to run without `--force` when you have
152	152	certain other precious per-checkout data that Fossil stores in the
153		-`.fslckout` file at the root of a check-out directory. This is a SQLite
	153	+`.fslckout` file at the root of a check-out directory. This is an SQLite
154	154	database that keeps track of local state such as what version you have
155	155	checked out, the contents of the [stash] for that working directory, the
156	156	[undo] buffers, per-checkout [settings][set], and so forth. The stash
157	157	and undo buffers are considered precious uncommitted changes,
158	158	so you have to force Fossil to discard these as part of closing the
		@@ -773,11 +773,11 @@
773	773	hashes.)
774	774
775	775	In this scheme, Alice then needs to say “`fossil update trunk`” in order
776	776	to return her check-out’s parent commit to the previous version lest her
777	777	next attempted commit land atop this mistake branch. The fact that Bob
778		-marked the branch as closed will prevent that from going thru, cluing
	778	+marked the branch as closed will prevent that from going through, cluing
779	779	Alice into what she needs to do to remedy the situation, but that merely
780	780	shows why it’s a better workflow if Alice makes the amendment herself:
781	781
782	782	```
783	783	fossil amend --branch MISTAKE --hide --close \
784	784

	--- www/gitusers.md
	+++ www/gitusers.md
	@@ -56,11 +56,11 @@
56
57
58
59	#### <a id="cwork" name="scw"></a> Checkout Workflows
60
61	A Fossil repository is a SQLite database storing the entire history of a
62	project. It is not normally stored inside the working tree.
63	A Fossil working tree — [also called a check-out](./glossary.md#check-out) — is a directory
64	that contains a snapshot of your project that you are currently working
65	on, extracted for you from the repository database file by the `fossil`
66	program.
	@@ -148,11 +148,11 @@
148	option, it won’t let you close the check-out with uncommitted changes to
149	those managed files.
150
151	The `close` command also refuses to run without `--force` when you have
152	certain other precious per-checkout data that Fossil stores in the
153	`.fslckout` file at the root of a check-out directory. This is a SQLite
154	database that keeps track of local state such as what version you have
155	checked out, the contents of the [stash] for that working directory, the
156	[undo] buffers, per-checkout [settings][set], and so forth. The stash
157	and undo buffers are considered precious uncommitted changes,
158	so you have to force Fossil to discard these as part of closing the
	@@ -773,11 +773,11 @@
773	hashes.)
774
775	In this scheme, Alice then needs to say “`fossil update trunk`” in order
776	to return her check-out’s parent commit to the previous version lest her
777	next attempted commit land atop this mistake branch. The fact that Bob
778	marked the branch as closed will prevent that from going thru, cluing
779	Alice into what she needs to do to remedy the situation, but that merely
780	shows why it’s a better workflow if Alice makes the amendment herself:
781
782	```
783	fossil amend --branch MISTAKE --hide --close \
784

	--- www/gitusers.md
	+++ www/gitusers.md
	@@ -56,11 +56,11 @@
56
57
58
59	#### <a id="cwork" name="scw"></a> Checkout Workflows
60
61	A Fossil repository is an SQLite database storing the entire history of a
62	project. It is not normally stored inside the working tree.
63	A Fossil working tree — [also called a check-out](./glossary.md#check-out) — is a directory
64	that contains a snapshot of your project that you are currently working
65	on, extracted for you from the repository database file by the `fossil`
66	program.
	@@ -148,11 +148,11 @@
148	option, it won’t let you close the check-out with uncommitted changes to
149	those managed files.
150
151	The `close` command also refuses to run without `--force` when you have
152	certain other precious per-checkout data that Fossil stores in the
153	`.fslckout` file at the root of a check-out directory. This is an SQLite
154	database that keeps track of local state such as what version you have
155	checked out, the contents of the [stash] for that working directory, the
156	[undo] buffers, per-checkout [settings][set], and so forth. The stash
157	and undo buffers are considered precious uncommitted changes,
158	so you have to force Fossil to discard these as part of closing the
	@@ -773,11 +773,11 @@
773	hashes.)
774
775	In this scheme, Alice then needs to say “`fossil update trunk`” in order
776	to return her check-out’s parent commit to the previous version lest her
777	next attempted commit land atop this mistake branch. The fact that Bob
778	marked the branch as closed will prevent that from going through, cluing
779	Alice into what she needs to do to remedy the situation, but that merely
780	shows why it’s a better workflow if Alice makes the amendment herself:
781
782	```
783	fossil amend --branch MISTAKE --hide --close \
784

M www/glossary.md

+2 -2

		--- www/glossary.md
		+++ www/glossary.md
		@@ -73,11 +73,11 @@
73	73	because you’ll have all of your OS’s other files intermixed.
74	74	Worse, Fossil doesn’t track OS permissions, so even if you were to
75	75	try to use Fossil as a system deployment tool by archiving versions
76	76	of the OS configuration files and then unpacking them on a new
77	77	system, the extracted project files would have read/write access by
78		- the user who did the extraction, which probably isn’t want you were
	78	+ the user who did the extraction, which probably isn’t what you were
79	79	wanting.
80	80
81	81	Even with these problems aside, do you really want a `.fslckout`
82	82	SQLite database at the root of your filesystem? Are you prepared for
83	83	the consequences of saying `fossil clean --verily` on such a system?
		@@ -348,11 +348,11 @@
348	348
349	349	* In the same way that one cannot extract files from a zip archive
350	350	without having a copy of that zip file, one cannot make check-outs
351	351	without access to the repository file or a clone thereof.
352	352
353		-* Because a Fossil repository is a SQLite database file, the same
	353	+* Because a Fossil repository is an SQLite database file, the same
354	354	rules for avoiding data corruption apply to it. In particular, it is
355	355	[nearly a hard requirement][h2cflp] that the repository clone be on
356	356	the same machine as the one where you make check-outs and the
357	357	subsequent check-ins.
358	358
359	359

	--- www/glossary.md
	+++ www/glossary.md
	@@ -73,11 +73,11 @@
73	because you’ll have all of your OS’s other files intermixed.
74	Worse, Fossil doesn’t track OS permissions, so even if you were to
75	try to use Fossil as a system deployment tool by archiving versions
76	of the OS configuration files and then unpacking them on a new
77	system, the extracted project files would have read/write access by
78	the user who did the extraction, which probably isn’t want you were
79	wanting.
80
81	Even with these problems aside, do you really want a `.fslckout`
82	SQLite database at the root of your filesystem? Are you prepared for
83	the consequences of saying `fossil clean --verily` on such a system?
	@@ -348,11 +348,11 @@
348
349	* In the same way that one cannot extract files from a zip archive
350	without having a copy of that zip file, one cannot make check-outs
351	without access to the repository file or a clone thereof.
352
353	* Because a Fossil repository is a SQLite database file, the same
354	rules for avoiding data corruption apply to it. In particular, it is
355	[nearly a hard requirement][h2cflp] that the repository clone be on
356	the same machine as the one where you make check-outs and the
357	subsequent check-ins.
358
359

	--- www/glossary.md
	+++ www/glossary.md
	@@ -73,11 +73,11 @@
73	because you’ll have all of your OS’s other files intermixed.
74	Worse, Fossil doesn’t track OS permissions, so even if you were to
75	try to use Fossil as a system deployment tool by archiving versions
76	of the OS configuration files and then unpacking them on a new
77	system, the extracted project files would have read/write access by
78	the user who did the extraction, which probably isn’t what you were
79	wanting.
80
81	Even with these problems aside, do you really want a `.fslckout`
82	SQLite database at the root of your filesystem? Are you prepared for
83	the consequences of saying `fossil clean --verily` on such a system?
	@@ -348,11 +348,11 @@
348
349	* In the same way that one cannot extract files from a zip archive
350	without having a copy of that zip file, one cannot make check-outs
351	without access to the repository file or a clone thereof.
352
353	* Because a Fossil repository is an SQLite database file, the same
354	rules for avoiding data corruption apply to it. In particular, it is
355	[nearly a hard requirement][h2cflp] that the repository clone be on
356	the same machine as the one where you make check-outs and the
357	subsequent check-ins.
358
359

M www/grep.md

+1 -1

		--- www/grep.md
		+++ www/grep.md
		@@ -49,11 +49,11 @@
49	49
50	50	$ fossil grep COMMAND: $(fossil ls src)
51	51
52	52	If you run that in a check-out of the [Fossil self-hosting source
53	53	repository][fshsr], that returns the first line of the built-in
54		-documentation for each Fossil command, across all historical verisons.
	54	+documentation for each Fossil command, across all historical versions.
55	55
56	56	Fossil `grep` has extensions relative to these other `grep` standards,
57	57	such as `--verbose` to print each checkin ID considered, regardless of
58	58	whether it matches. This one is noteworthy here because the behavior
59	59	used to be under `-v` before we reassigned it to give POSIX-like `grep
60	60

	--- www/grep.md
	+++ www/grep.md
	@@ -49,11 +49,11 @@
49
50	$ fossil grep COMMAND: $(fossil ls src)
51
52	If you run that in a check-out of the [Fossil self-hosting source
53	repository][fshsr], that returns the first line of the built-in
54	documentation for each Fossil command, across all historical verisons.
55
56	Fossil `grep` has extensions relative to these other `grep` standards,
57	such as `--verbose` to print each checkin ID considered, regardless of
58	whether it matches. This one is noteworthy here because the behavior
59	used to be under `-v` before we reassigned it to give POSIX-like `grep
60

	--- www/grep.md
	+++ www/grep.md
	@@ -49,11 +49,11 @@
49
50	$ fossil grep COMMAND: $(fossil ls src)
51
52	If you run that in a check-out of the [Fossil self-hosting source
53	repository][fshsr], that returns the first line of the built-in
54	documentation for each Fossil command, across all historical versions.
55
56	Fossil `grep` has extensions relative to these other `grep` standards,
57	such as `--verbose` to print each checkin ID considered, regardless of
58	whether it matches. This one is noteworthy here because the behavior
59	used to be under `-v` before we reassigned it to give POSIX-like `grep
60

M www/hashpolicy.wiki

+1 -1

		--- www/hashpolicy.wiki
		+++ www/hashpolicy.wiki
		@@ -78,11 +78,11 @@
78	78	Version 2.0 extended the [./fileformat.wiki\|Fossil file format]
79	79	to allow artifacts to be named by either SHA1 or SHA3-256 hashes.
80	80	(SHA3-256 is the only variant of SHA3 that
81	81	Fossil uses for artifact naming, so for the remainder of this article
82	82	it will be called simply "SHA3". Similarly, "Hardened SHA1" will
83		-shortened to "SHA1" in the remaining text.)
	83	+be shortened to "SHA1" in the remaining text.)
84	84
85	85	To be clear: Fossil (version 2.0 and later)
86	86	allows the SHA1 and SHA3 hashes to be mixed within
87	87	the same repository. Older check-ins, created years ago,
88	88	continue to be named using their legacy SHA1 hashes while
89	89

	--- www/hashpolicy.wiki
	+++ www/hashpolicy.wiki
	@@ -78,11 +78,11 @@
78	Version 2.0 extended the [./fileformat.wiki\|Fossil file format]
79	to allow artifacts to be named by either SHA1 or SHA3-256 hashes.
80	(SHA3-256 is the only variant of SHA3 that
81	Fossil uses for artifact naming, so for the remainder of this article
82	it will be called simply "SHA3". Similarly, "Hardened SHA1" will
83	shortened to "SHA1" in the remaining text.)
84
85	To be clear: Fossil (version 2.0 and later)
86	allows the SHA1 and SHA3 hashes to be mixed within
87	the same repository. Older check-ins, created years ago,
88	continue to be named using their legacy SHA1 hashes while
89

	--- www/hashpolicy.wiki
	+++ www/hashpolicy.wiki
	@@ -78,11 +78,11 @@
78	Version 2.0 extended the [./fileformat.wiki\|Fossil file format]
79	to allow artifacts to be named by either SHA1 or SHA3-256 hashes.
80	(SHA3-256 is the only variant of SHA3 that
81	Fossil uses for artifact naming, so for the remainder of this article
82	it will be called simply "SHA3". Similarly, "Hardened SHA1" will
83	be shortened to "SHA1" in the remaining text.)
84
85	To be clear: Fossil (version 2.0 and later)
86	allows the SHA1 and SHA3 hashes to be mixed within
87	the same repository. Older check-ins, created years ago,
88	continue to be named using their legacy SHA1 hashes while
89

M www/hints.wiki

+1 -1

		--- www/hints.wiki
		+++ www/hints.wiki
		@@ -8,11 +8,11 @@
8	8	2. Add the "--tk" option to "[/help?cmd=diff \| fossil diff]" commands
9	9	to get a pop-up
10	10	window containing a complete side-by-side diff. (NB: The pop-up
11	11	window is run as a separate Tcl/Tk process, so you will need to
12	12	have Tcl/Tk installed on your machine for this to work. Visit
13		- [http://www.activestate.com/activetcl] to for a quick download of
	13	+ [http://www.activestate.com/activetcl] for a quick download of
14	14	Tcl/Tk if you do not already have it on your system.)
15	15
16	16	3. The "[/help/clean \| fossil clean -x]" command is a great
17	17	alternative to "make clean". You can use "[/help/clean \| fossil clean -f]"
18	18	as a slightly safer alternative if the "ignore-glob" setting is
19	19

	--- www/hints.wiki
	+++ www/hints.wiki
	@@ -8,11 +8,11 @@
8	2. Add the "--tk" option to "[/help?cmd=diff \| fossil diff]" commands
9	to get a pop-up
10	window containing a complete side-by-side diff. (NB: The pop-up
11	window is run as a separate Tcl/Tk process, so you will need to
12	have Tcl/Tk installed on your machine for this to work. Visit
13	[http://www.activestate.com/activetcl] to for a quick download of
14	Tcl/Tk if you do not already have it on your system.)
15
16	3. The "[/help/clean \| fossil clean -x]" command is a great
17	alternative to "make clean". You can use "[/help/clean \| fossil clean -f]"
18	as a slightly safer alternative if the "ignore-glob" setting is
19

	--- www/hints.wiki
	+++ www/hints.wiki
	@@ -8,11 +8,11 @@
8	2. Add the "--tk" option to "[/help?cmd=diff \| fossil diff]" commands
9	to get a pop-up
10	window containing a complete side-by-side diff. (NB: The pop-up
11	window is run as a separate Tcl/Tk process, so you will need to
12	have Tcl/Tk installed on your machine for this to work. Visit
13	[http://www.activestate.com/activetcl] for a quick download of
14	Tcl/Tk if you do not already have it on your system.)
15
16	3. The "[/help/clean \| fossil clean -x]" command is a great
17	alternative to "make clean". You can use "[/help/clean \| fossil clean -f]"
18	as a slightly safer alternative if the "ignore-glob" setting is
19

M www/history.md

+1 -1

		--- www/history.md
		+++ www/history.md
		@@ -16,11 +16,11 @@
16	16
17	17	[120]: /reports?type=ci&view=byuser
18	18
19	19	## History
20	20
21		-The SQLite project start out using [CVS][300], as CVS was the most
	21	+The SQLite project started out using [CVS][300], as CVS was the most
22	22	commonly used version control system in that era (circa 2000). CVS
23	23	was an amazing version control system for its day in that it allowed
24	24	multiple developers to be editing the same file at the same time.
25	25
26	26	[300]: https://en.wikipedia.org/wiki/Concurrent_Versions_System
27	27

	--- www/history.md
	+++ www/history.md
	@@ -16,11 +16,11 @@
16
17	[120]: /reports?type=ci&view=byuser
18
19	## History
20
21	The SQLite project start out using [CVS][300], as CVS was the most
22	commonly used version control system in that era (circa 2000). CVS
23	was an amazing version control system for its day in that it allowed
24	multiple developers to be editing the same file at the same time.
25
26	[300]: https://en.wikipedia.org/wiki/Concurrent_Versions_System
27

	--- www/history.md
	+++ www/history.md
	@@ -16,11 +16,11 @@
16
17	[120]: /reports?type=ci&view=byuser
18
19	## History
20
21	The SQLite project started out using [CVS][300], as CVS was the most
22	commonly used version control system in that era (circa 2000). CVS
23	was an amazing version control system for its day in that it allowed
24	multiple developers to be editing the same file at the same time.
25
26	[300]: https://en.wikipedia.org/wiki/Concurrent_Versions_System
27

M www/hooks.md

+4 -4

		--- www/hooks.md
		+++ www/hooks.md
		@@ -5,11 +5,11 @@
5	5	a continuous integration (CI) system.
6	6
7	7	## Interim Documentation.
8	8
9	9	* This is a work-in-progress. The interface is in flux.
10		- For the time being, the documentation as a list of
	10	+ For the time being, the documentation is a list of
11	11	bullet points. We hope to transform this into a proper document
12	12	later, after things settle down.
13	13
14	14	* Contributions and suggestions to the hook system and/or the
15	15	documentation are welcomed.
		@@ -35,14 +35,14 @@
35	35	that the original script can return.
36	36
37	37	* The "%F" sequence inside the script is translated into the
38	38	name of the fossil executable.
39	39
40		- * The "%R" sequence in the script is translated in to the name of
	40	+ * The "%R" sequence in the script is translated into the name of
41	41	the repository.
42	42
43		- * The "%A" sequence becomes the name of an auxiliary input files,
	43	+ * The "%A" sequence becomes the name of an auxiliary input file,
44	44	the meaning of which depends on the hook type. The auxiliary filename
45	45	might be an empty string. Take care to use appropriate quoting!
46	46
47	47	## Disabled Hooks
48	48
		@@ -144,11 +144,11 @@
144	144	## Commit-Msg Hooks
145	145
146	146	* Commit-msg hooks are not yet implemented.
147	147
148	148	* The commit-msg hooks run during "fossil commit" after the check-in
149		- messages has been entered by the user. The "%A" argument to the
	149	+ message has been entered by the user. The "%A" argument to the
150	150	commit-msg hook is the text of the commit message. The intent
151	151	of the commit-msg hook is to validate the text of the commit
152	152	message to (for example) check for typos or ensure that it
153	153	conforms to standards.
154	154
155	155

	--- www/hooks.md
	+++ www/hooks.md
	@@ -5,11 +5,11 @@
5	a continuous integration (CI) system.
6
7	## Interim Documentation.
8
9	* This is a work-in-progress. The interface is in flux.
10	For the time being, the documentation as a list of
11	bullet points. We hope to transform this into a proper document
12	later, after things settle down.
13
14	* Contributions and suggestions to the hook system and/or the
15	documentation are welcomed.
	@@ -35,14 +35,14 @@
35	that the original script can return.
36
37	* The "%F" sequence inside the script is translated into the
38	name of the fossil executable.
39
40	* The "%R" sequence in the script is translated in to the name of
41	the repository.
42
43	* The "%A" sequence becomes the name of an auxiliary input files,
44	the meaning of which depends on the hook type. The auxiliary filename
45	might be an empty string. Take care to use appropriate quoting!
46
47	## Disabled Hooks
48
	@@ -144,11 +144,11 @@
144	## Commit-Msg Hooks
145
146	* Commit-msg hooks are not yet implemented.
147
148	* The commit-msg hooks run during "fossil commit" after the check-in
149	messages has been entered by the user. The "%A" argument to the
150	commit-msg hook is the text of the commit message. The intent
151	of the commit-msg hook is to validate the text of the commit
152	message to (for example) check for typos or ensure that it
153	conforms to standards.
154
155

	--- www/hooks.md
	+++ www/hooks.md
	@@ -5,11 +5,11 @@
5	a continuous integration (CI) system.
6
7	## Interim Documentation.
8
9	* This is a work-in-progress. The interface is in flux.
10	For the time being, the documentation is a list of
11	bullet points. We hope to transform this into a proper document
12	later, after things settle down.
13
14	* Contributions and suggestions to the hook system and/or the
15	documentation are welcomed.
	@@ -35,14 +35,14 @@
35	that the original script can return.
36
37	* The "%F" sequence inside the script is translated into the
38	name of the fossil executable.
39
40	* The "%R" sequence in the script is translated into the name of
41	the repository.
42
43	* The "%A" sequence becomes the name of an auxiliary input file,
44	the meaning of which depends on the hook type. The auxiliary filename
45	might be an empty string. Take care to use appropriate quoting!
46
47	## Disabled Hooks
48
	@@ -144,11 +144,11 @@
144	## Commit-Msg Hooks
145
146	* Commit-msg hooks are not yet implemented.
147
148	* The commit-msg hooks run during "fossil commit" after the check-in
149	message has been entered by the user. The "%A" argument to the
150	commit-msg hook is the text of the commit message. The intent
151	of the commit-msg hook is to validate the text of the commit
152	message to (for example) check for typos or ensure that it
153	conforms to standards.
154
155

M www/index.wiki

+4 -4

		--- www/index.wiki
		+++ www/index.wiki
		@@ -84,16 +84,16 @@
84	84	the repository are consistent prior to each commit.
85	85
86	86	8. <b>Free and Open-Source</b> — [../COPYRIGHT-BSD2.txt\|2-clause BSD license].
87	87
88	88	<hr>
89		-<h3>Latest Release: 2.24 ([/timeline?c=version-2.24\|2024-04-23])</h3>
	89	+<h3>Latest Release: 2.25 ([/timeline?c=version-2.25\|2024-11-06])</h3>
90	90
91	91	* [/uv/download.html\|Download]
92		- * [./changes.wiki#v2_24\|Change Summary]
93		- * [/timeline?p=version-2.24&bt=version-2.23&y=ci\|Check-ins in version 2.24]
94		- * [/timeline?df=version-2.24&y=ci\|Check-ins derived from the 2.24 release]
	92	+ * [./changes.wiki#v2_25\|Change Summary]
	93	+ * [/timeline?p=version-2.25&bt=version-2.24&y=ci\|Check-ins in version 2.25]
	94	+ * [/timeline?df=version-2.25&y=ci\|Check-ins derived from the 2.25 release]
95	95	* [/timeline?t=release\|Timeline of all past releases]
96	96
97	97	<hr>
98	98	<h3>Quick Start</h3>
99	99
100	100

	--- www/index.wiki
	+++ www/index.wiki
	@@ -84,16 +84,16 @@
84	the repository are consistent prior to each commit.
85
86	8. <b>Free and Open-Source</b> — [../COPYRIGHT-BSD2.txt\|2-clause BSD license].
87
88	<hr>
89	<h3>Latest Release: 2.24 ([/timeline?c=version-2.24\|2024-04-23])</h3>
90
91	* [/uv/download.html\|Download]
92	* [./changes.wiki#v2_24\|Change Summary]
93	* [/timeline?p=version-2.24&bt=version-2.23&y=ci\|Check-ins in version 2.24]
94	* [/timeline?df=version-2.24&y=ci\|Check-ins derived from the 2.24 release]
95	* [/timeline?t=release\|Timeline of all past releases]
96
97	<hr>
98	<h3>Quick Start</h3>
99
100

	--- www/index.wiki
	+++ www/index.wiki
	@@ -84,16 +84,16 @@
84	the repository are consistent prior to each commit.
85
86	8. <b>Free and Open-Source</b> — [../COPYRIGHT-BSD2.txt\|2-clause BSD license].
87
88	<hr>
89	<h3>Latest Release: 2.25 ([/timeline?c=version-2.25\|2024-11-06])</h3>
90
91	* [/uv/download.html\|Download]
92	* [./changes.wiki#v2_25\|Change Summary]
93	* [/timeline?p=version-2.25&bt=version-2.24&y=ci\|Check-ins in version 2.25]
94	* [/timeline?df=version-2.25&y=ci\|Check-ins derived from the 2.25 release]
95	* [/timeline?t=release\|Timeline of all past releases]
96
97	<hr>
98	<h3>Quick Start</h3>
99
100

M www/inout.wiki

+2 -2

		--- www/inout.wiki
		+++ www/inout.wiki
		@@ -48,11 +48,11 @@
48	48	emitted by "<tt>git fast-export</tt>" before sending it along to Fossil,
49	49	but we've seen the problem recur on multiple machines.
50	50
51	51	While one workaround is to fall back to <tt>cmd.exe</tt> — which doesn't
52	52	seem to be affected by this problem — we instead recommend using
53		-Mirosoft's own [https://learn.microsoft.com/en-us/windows/wsl/ \| Windows
	53	+Microsoft's own [https://learn.microsoft.com/en-us/windows/wsl/ \| Windows
54	54	Subsystem for Linux] or either of the two popular "Git for Windows"
55	55	distributions based on MSYS2. They handle pipes the POSIX way, avoiding
56	56	any dependency on the amount of data involved.
57	57
58	58	<h2>Fossil → Git</h2>
		@@ -76,11 +76,11 @@
76	76	As with the "import" command, the --git option is not required
77	77	since the git-fast-export file format is currently the only VCS interchange
78	78	format that Fossil will generate. However,
79	79	future versions of Fossil might add the ability to generate other
80	80	VCS interchange formats, and so for compatibility, the use of the --git
81		-option recommended.
	81	+option is recommended.
82	82
83	83	<h2>Mirror A Fossil Repository In Git</h2>
84	84
85	85	Fossil version 2.9 and later supports a simple mechanism for
86	86	doing a Git or
87	87

	--- www/inout.wiki
	+++ www/inout.wiki
	@@ -48,11 +48,11 @@
48	emitted by "<tt>git fast-export</tt>" before sending it along to Fossil,
49	but we've seen the problem recur on multiple machines.
50
51	While one workaround is to fall back to <tt>cmd.exe</tt> — which doesn't
52	seem to be affected by this problem — we instead recommend using
53	Mirosoft's own [https://learn.microsoft.com/en-us/windows/wsl/ \| Windows
54	Subsystem for Linux] or either of the two popular "Git for Windows"
55	distributions based on MSYS2. They handle pipes the POSIX way, avoiding
56	any dependency on the amount of data involved.
57
58	<h2>Fossil → Git</h2>
	@@ -76,11 +76,11 @@
76	As with the "import" command, the --git option is not required
77	since the git-fast-export file format is currently the only VCS interchange
78	format that Fossil will generate. However,
79	future versions of Fossil might add the ability to generate other
80	VCS interchange formats, and so for compatibility, the use of the --git
81	option recommended.
82
83	<h2>Mirror A Fossil Repository In Git</h2>
84
85	Fossil version 2.9 and later supports a simple mechanism for
86	doing a Git or
87

	--- www/inout.wiki
	+++ www/inout.wiki
	@@ -48,11 +48,11 @@
48	emitted by "<tt>git fast-export</tt>" before sending it along to Fossil,
49	but we've seen the problem recur on multiple machines.
50
51	While one workaround is to fall back to <tt>cmd.exe</tt> — which doesn't
52	seem to be affected by this problem — we instead recommend using
53	Microsoft's own [https://learn.microsoft.com/en-us/windows/wsl/ \| Windows
54	Subsystem for Linux] or either of the two popular "Git for Windows"
55	distributions based on MSYS2. They handle pipes the POSIX way, avoiding
56	any dependency on the amount of data involved.
57
58	<h2>Fossil → Git</h2>
	@@ -76,11 +76,11 @@
76	As with the "import" command, the --git option is not required
77	since the git-fast-export file format is currently the only VCS interchange
78	format that Fossil will generate. However,
79	future versions of Fossil might add the ability to generate other
80	VCS interchange formats, and so for compatibility, the use of the --git
81	option is recommended.
82
83	<h2>Mirror A Fossil Repository In Git</h2>
84
85	Fossil version 2.9 and later supports a simple mechanism for
86	doing a Git or
87

M www/interwiki.md

+5 -5

		--- www/interwiki.md
		+++ www/interwiki.md
		@@ -50,11 +50,11 @@
50	50	or is an empty string.
51	51
52	52	2. <b>Hash Links</b> → the PageName is a hexadecimal number with
53	53	at least four digits.
54	54
55		- 3. <b>Wiki Links</b> → An PageName that is not a Path or Hash.
	55	+ 3. <b>Wiki Links</b> → A PageName that is not a Path or Hash.
56	56
57	57	The Intermap defines a base URL for each Tag. Path links are appended
58	58	directly to the URL contained in the Intermap. The Intermap can define
59	59	additional text to put in between the base URL and the PageName for
60	60	Hash and Wiki links, respectively.
		@@ -62,11 +62,11 @@
62	62	<a id="intermap"></a>
63	63	## Intermap
64	64
65	65	The intermap defines a mapping from interwiki Tags to full URLs. The
66	66	Intermap can be viewed and managed using the [fossil interwiki][iwiki]
67		-command or the [/intermap][imap] webpages.
	67	+command or the [/intermap][imap] webpage.
68	68
69	69	[iwiki]: /help?cmd=interwiki
70	70	[imap]: /intermap
71	71
72	72	The current intermap for a server is seen on the [/intermap][imap] page
		@@ -77,19 +77,19 @@
77	77	Each intermap entry stores, at a minimum, the base URL for the remote
78	78	wiki. The intermap entry might also store additional path text that
79	79	is used for Hash and Wiki links. If only the base URL is provided,
80	80	then the intermap will only allow Path style interwiki links. The
81	81	Hash and Wiki style interwiki links are only allowed if the necessary
82		-extensions for provided in the intermap.
	82	+extensions are provided in the intermap.
83	83
84	84
85	85	## Disadvantages and Limitations
86	86
87	87	* Configuration is required. The intermap must be set up correctly
88	88	before interwiki links will work. This contrasts with ordinary
89	89	links that just work without any configuration. Cloning a repository
90		- copies the intermap, but normal syncs to not keep the intermap in
	90	+ copies the intermap, but normal syncs do not keep the intermap in
91	91	sync. Use the "[fossil config pull interwiki][fcfg]" command to
92	92	synchronize the intermap.
93	93
94	94	* The is no backlink tracking. For ordinary intrawiki links, Fossil keeps
95	95	track of both the source and target, and when displaying targets it
		@@ -96,11 +96,11 @@
96	96	commonly shows links to that target. For example, if you mention a
97	97	check-in as part of a comment of another check-in, that new check-in
98	98	shows up in the "References" section of the target check-in.
99	99	([example](31af805348690958). In other words, Fossil tracks not just
100	100	"_source→target_", but it also tracks "_target→source_".
101		- But backtracking do not work for interwiki links, since the Fossil
	101	+ But backtracking does not work for interwiki links, since the Fossil
102	102	running on the target has no way of scanning the source text and
103	103	hence has no way of knowing that it is a target of a link from the source.
104	104
105	105	[fcfg]: /help?cmd=config
106	106
107	107

	--- www/interwiki.md
	+++ www/interwiki.md
	@@ -50,11 +50,11 @@
50	or is an empty string.
51
52	2. <b>Hash Links</b> → the PageName is a hexadecimal number with
53	at least four digits.
54
55	3. <b>Wiki Links</b> → An PageName that is not a Path or Hash.
56
57	The Intermap defines a base URL for each Tag. Path links are appended
58	directly to the URL contained in the Intermap. The Intermap can define
59	additional text to put in between the base URL and the PageName for
60	Hash and Wiki links, respectively.
	@@ -62,11 +62,11 @@
62	<a id="intermap"></a>
63	## Intermap
64
65	The intermap defines a mapping from interwiki Tags to full URLs. The
66	Intermap can be viewed and managed using the [fossil interwiki][iwiki]
67	command or the [/intermap][imap] webpages.
68
69	[iwiki]: /help?cmd=interwiki
70	[imap]: /intermap
71
72	The current intermap for a server is seen on the [/intermap][imap] page
	@@ -77,19 +77,19 @@
77	Each intermap entry stores, at a minimum, the base URL for the remote
78	wiki. The intermap entry might also store additional path text that
79	is used for Hash and Wiki links. If only the base URL is provided,
80	then the intermap will only allow Path style interwiki links. The
81	Hash and Wiki style interwiki links are only allowed if the necessary
82	extensions for provided in the intermap.
83
84
85	## Disadvantages and Limitations
86
87	* Configuration is required. The intermap must be set up correctly
88	before interwiki links will work. This contrasts with ordinary
89	links that just work without any configuration. Cloning a repository
90	copies the intermap, but normal syncs to not keep the intermap in
91	sync. Use the "[fossil config pull interwiki][fcfg]" command to
92	synchronize the intermap.
93
94	* The is no backlink tracking. For ordinary intrawiki links, Fossil keeps
95	track of both the source and target, and when displaying targets it
	@@ -96,11 +96,11 @@
96	commonly shows links to that target. For example, if you mention a
97	check-in as part of a comment of another check-in, that new check-in
98	shows up in the "References" section of the target check-in.
99	([example](31af805348690958). In other words, Fossil tracks not just
100	"_source→target_", but it also tracks "_target→source_".
101	But backtracking do not work for interwiki links, since the Fossil
102	running on the target has no way of scanning the source text and
103	hence has no way of knowing that it is a target of a link from the source.
104
105	[fcfg]: /help?cmd=config
106
107

	--- www/interwiki.md
	+++ www/interwiki.md
	@@ -50,11 +50,11 @@
50	or is an empty string.
51
52	2. <b>Hash Links</b> → the PageName is a hexadecimal number with
53	at least four digits.
54
55	3. <b>Wiki Links</b> → A PageName that is not a Path or Hash.
56
57	The Intermap defines a base URL for each Tag. Path links are appended
58	directly to the URL contained in the Intermap. The Intermap can define
59	additional text to put in between the base URL and the PageName for
60	Hash and Wiki links, respectively.
	@@ -62,11 +62,11 @@
62	<a id="intermap"></a>
63	## Intermap
64
65	The intermap defines a mapping from interwiki Tags to full URLs. The
66	Intermap can be viewed and managed using the [fossil interwiki][iwiki]
67	command or the [/intermap][imap] webpage.
68
69	[iwiki]: /help?cmd=interwiki
70	[imap]: /intermap
71
72	The current intermap for a server is seen on the [/intermap][imap] page
	@@ -77,19 +77,19 @@
77	Each intermap entry stores, at a minimum, the base URL for the remote
78	wiki. The intermap entry might also store additional path text that
79	is used for Hash and Wiki links. If only the base URL is provided,
80	then the intermap will only allow Path style interwiki links. The
81	Hash and Wiki style interwiki links are only allowed if the necessary
82	extensions are provided in the intermap.
83
84
85	## Disadvantages and Limitations
86
87	* Configuration is required. The intermap must be set up correctly
88	before interwiki links will work. This contrasts with ordinary
89	links that just work without any configuration. Cloning a repository
90	copies the intermap, but normal syncs do not keep the intermap in
91	sync. Use the "[fossil config pull interwiki][fcfg]" command to
92	synchronize the intermap.
93
94	* The is no backlink tracking. For ordinary intrawiki links, Fossil keeps
95	track of both the source and target, and when displaying targets it
	@@ -96,11 +96,11 @@
96	commonly shows links to that target. For example, if you mention a
97	check-in as part of a comment of another check-in, that new check-in
98	shows up in the "References" section of the target check-in.
99	([example](31af805348690958). In other words, Fossil tracks not just
100	"_source→target_", but it also tracks "_target→source_".
101	But backtracking does not work for interwiki links, since the Fossil
102	running on the target has no way of scanning the source text and
103	hence has no way of knowing that it is a target of a link from the source.
104
105	[fcfg]: /help?cmd=config
106
107

M www/json-api/api-artifact.md

+1 -1

		--- www/json-api/api-artifact.md
		+++ www/json-api/api-artifact.md
		@@ -71,11 +71,11 @@
71	71	# File Artifacts
72	72
73	73	Fetches information about file artifacts.
74	74
75	75	FIXME: the content type guessing is currently very primitive, and
76		-may (but i haven't seen this) mis-diagnose some non-binary files as
	76	+may (but I haven't seen this) mis-diagnose some non-binary files as
77	77	binary. Fossil doesn't yet have a mechanism for mime-type mappings.
78	78
79	79	Status: implemented 20111020
80	80
81	81	Required permissions: "o"
82	82

	--- www/json-api/api-artifact.md
	+++ www/json-api/api-artifact.md
	@@ -71,11 +71,11 @@
71	# File Artifacts
72
73	Fetches information about file artifacts.
74
75	FIXME: the content type guessing is currently very primitive, and
76	may (but i haven't seen this) mis-diagnose some non-binary files as
77	binary. Fossil doesn't yet have a mechanism for mime-type mappings.
78
79	Status: implemented 20111020
80
81	Required permissions: "o"
82

	--- www/json-api/api-artifact.md
	+++ www/json-api/api-artifact.md
	@@ -71,11 +71,11 @@
71	# File Artifacts
72
73	Fetches information about file artifacts.
74
75	FIXME: the content type guessing is currently very primitive, and
76	may (but I haven't seen this) mis-diagnose some non-binary files as
77	binary. Fossil doesn't yet have a mechanism for mime-type mappings.
78
79	Status: implemented 20111020
80
81	Required permissions: "o"
82

M www/json-api/api-auth.md

+3 -3

		--- www/json-api/api-auth.md
		+++ www/json-api/api-auth.md
		@@ -33,11 +33,11 @@
33	33	purposes, the "auth token" and the "login cookie" are the same thing (or
34	34	serve the same purpose), and the auth token is in fact just the value
35	35	part of the login cookie (which has a project-specific key).
36	36
37	37	Note that fossil has two conventional user names which can show up in
38		-various response but do not refer to specific people: nobody and
	38	+various responses but do not refer to specific people: nobody and
39	39	anonymous. The nobody user is anyone who is not logged in. The anonymous
40	40	user is logged in but has no persistent user data (no associated user
41	41	name, email address, or similar). Normally the guest (nobody) user has
42	42	more access restrictions. The distinction between the two is largely
43	43	historical - it is a mechanism to keep bots from following the
		@@ -158,11 +158,11 @@
158	158	tickets is disabled for the guest user then all non-guest users must
159	159	send authentication info in their requests in order to be able to fetch
160	160	ticket info.
161	161
162	162	Cookie-aware clients should send the login-generated cookie with each
163		-request, in which case they do not need explicitly include the
	163	+request, in which case they do not need to explicitly include the
164	164	`authToken` in the JSON envelope/GET arguments. If submitted, the
165	165	`authToken` is used, otherwise the cookie, if set, is used. Note that
166	166	fossil uses a project-dependent cookie name in order to help thwart
167	167	attacks, so there is no simple mapping of cookie name to auth
168	168	token. That said, the cookie's value is also the auth token's value.
		@@ -200,11 +200,11 @@
200	200
201	201	The password value may be time-limited, and may eventually become
202	202	invalidated due to old age. This is unspecified.
203	203
204	204	*Potential* (low-probability) bug regarding the seed value: from
205		-what i hear, some unusual JSON platforms don't support full 32-bit
	205	+what I hear, some unusual JSON platforms don't support full 32-bit
206	206	precision. If absolutely necessary we could chop off a bit or two from
207	207	the seed value (if it ever becomes a problem and if DRH blesses it).
208	208	Or we could just make it a double.
209	209
210	210
211	211

	--- www/json-api/api-auth.md
	+++ www/json-api/api-auth.md
	@@ -33,11 +33,11 @@
33	purposes, the "auth token" and the "login cookie" are the same thing (or
34	serve the same purpose), and the auth token is in fact just the value
35	part of the login cookie (which has a project-specific key).
36
37	Note that fossil has two conventional user names which can show up in
38	various response but do not refer to specific people: nobody and
39	anonymous. The nobody user is anyone who is not logged in. The anonymous
40	user is logged in but has no persistent user data (no associated user
41	name, email address, or similar). Normally the guest (nobody) user has
42	more access restrictions. The distinction between the two is largely
43	historical - it is a mechanism to keep bots from following the
	@@ -158,11 +158,11 @@
158	tickets is disabled for the guest user then all non-guest users must
159	send authentication info in their requests in order to be able to fetch
160	ticket info.
161
162	Cookie-aware clients should send the login-generated cookie with each
163	request, in which case they do not need explicitly include the
164	`authToken` in the JSON envelope/GET arguments. If submitted, the
165	`authToken` is used, otherwise the cookie, if set, is used. Note that
166	fossil uses a project-dependent cookie name in order to help thwart
167	attacks, so there is no simple mapping of cookie name to auth
168	token. That said, the cookie's value is also the auth token's value.
	@@ -200,11 +200,11 @@
200
201	The password value may be time-limited, and may eventually become
202	invalidated due to old age. This is unspecified.
203
204	*Potential* (low-probability) bug regarding the seed value: from
205	what i hear, some unusual JSON platforms don't support full 32-bit
206	precision. If absolutely necessary we could chop off a bit or two from
207	the seed value (if it ever becomes a problem and if DRH blesses it).
208	Or we could just make it a double.
209
210
211

	--- www/json-api/api-auth.md
	+++ www/json-api/api-auth.md
	@@ -33,11 +33,11 @@
33	purposes, the "auth token" and the "login cookie" are the same thing (or
34	serve the same purpose), and the auth token is in fact just the value
35	part of the login cookie (which has a project-specific key).
36
37	Note that fossil has two conventional user names which can show up in
38	various responses but do not refer to specific people: nobody and
39	anonymous. The nobody user is anyone who is not logged in. The anonymous
40	user is logged in but has no persistent user data (no associated user
41	name, email address, or similar). Normally the guest (nobody) user has
42	more access restrictions. The distinction between the two is largely
43	historical - it is a mechanism to keep bots from following the
	@@ -158,11 +158,11 @@
158	tickets is disabled for the guest user then all non-guest users must
159	send authentication info in their requests in order to be able to fetch
160	ticket info.
161
162	Cookie-aware clients should send the login-generated cookie with each
163	request, in which case they do not need to explicitly include the
164	`authToken` in the JSON envelope/GET arguments. If submitted, the
165	`authToken` is used, otherwise the cookie, if set, is used. Note that
166	fossil uses a project-dependent cookie name in order to help thwart
167	attacks, so there is no simple mapping of cookie name to auth
168	token. That said, the cookie's value is also the auth token's value.
	@@ -200,11 +200,11 @@
200
201	The password value may be time-limited, and may eventually become
202	invalidated due to old age. This is unspecified.
203
204	*Potential* (low-probability) bug regarding the seed value: from
205	what I hear, some unusual JSON platforms don't support full 32-bit
206	precision. If absolutely necessary we could chop off a bit or two from
207	the seed value (if it ever becomes a problem and if DRH blesses it).
208	Or we could just make it a double.
209
210
211

M www/json-api/api-checkout.md

+1 -1

		--- www/json-api/api-checkout.md
		+++ www/json-api/api-checkout.md
		@@ -7,11 +7,11 @@
7	7
8	8	Required permissions: n/a (local access only)
9	9
10	10	Request: `/json/status`
11	11
12		-This command requires a local checkout and is analog to the "fossil
	12	+This command requires a local checkout and is the analog to the "fossil
13	13	status" command.
14	14
15	15	Request Options: currently none.
16	16
17	17	Payload example:
18	18

	--- www/json-api/api-checkout.md
	+++ www/json-api/api-checkout.md
	@@ -7,11 +7,11 @@
7
8	Required permissions: n/a (local access only)
9
10	Request: `/json/status`
11
12	This command requires a local checkout and is analog to the "fossil
13	status" command.
14
15	Request Options: currently none.
16
17	Payload example:
18

	--- www/json-api/api-checkout.md
	+++ www/json-api/api-checkout.md
	@@ -7,11 +7,11 @@
7
8	Required permissions: n/a (local access only)
9
10	Request: `/json/status`
11
12	This command requires a local checkout and is the analog to the "fossil
13	status" command.
14
15	Request Options: currently none.
16
17	Payload example:
18

M www/json-api/api-finfo.md

+1 -1

		--- www/json-api/api-finfo.md
		+++ www/json-api/api-finfo.md
		@@ -20,11 +20,11 @@
20	20	as opposed to listing all checkins. If set, neither "before" nor
21	21	"after" have any effect.\
22	22	CLI mode: `--checkin\|-ci`
23	23	- `before=DATETIME` only lists checkins from on or before this time.\
24	24	CLI mode: `--before\|-b`
25		-- `after=DATETIME` only lists checkins from on or before this time.
	25	+- `after=DATETIME` only lists checkins from on or after this time.
26	26	Using this option swaps the sort order, to provide reasonable
27	27	behaviour in conjunction with the limit option.\
28	28	Only one of "before" and "after" may be specified, and if both are
29	29	specified then which one takes precedence is unspecified.\
30	30	CLI mode: `--after\|-a`
31	31

	--- www/json-api/api-finfo.md
	+++ www/json-api/api-finfo.md
	@@ -20,11 +20,11 @@
20	as opposed to listing all checkins. If set, neither "before" nor
21	"after" have any effect.\
22	CLI mode: `--checkin\|-ci`
23	- `before=DATETIME` only lists checkins from on or before this time.\
24	CLI mode: `--before\|-b`
25	- `after=DATETIME` only lists checkins from on or before this time.
26	Using this option swaps the sort order, to provide reasonable
27	behaviour in conjunction with the limit option.\
28	Only one of "before" and "after" may be specified, and if both are
29	specified then which one takes precedence is unspecified.\
30	CLI mode: `--after\|-a`
31

	--- www/json-api/api-finfo.md
	+++ www/json-api/api-finfo.md
	@@ -20,11 +20,11 @@
20	as opposed to listing all checkins. If set, neither "before" nor
21	"after" have any effect.\
22	CLI mode: `--checkin\|-ci`
23	- `before=DATETIME` only lists checkins from on or before this time.\
24	CLI mode: `--before\|-b`
25	- `after=DATETIME` only lists checkins from on or after this time.
26	Using this option swaps the sort order, to provide reasonable
27	behaviour in conjunction with the limit option.\
28	Only one of "before" and "after" may be specified, and if both are
29	specified then which one takes precedence is unspecified.\
30	CLI mode: `--after\|-a`
31

M www/json-api/api-query.md

+2 -2

		--- www/json-api/api-query.md
		+++ www/json-api/api-query.md
		@@ -60,11 +60,11 @@
60	60	…
61	61	]
62	62	}
63	63	```
64	64
65		-The column names are provided in a separate field is because their order
	65	+The column names are provided in a separate field because their order
66	66	is guaranteed to match the order of the query columns, whereas object
67	67	key/value pairs might get reordered (typically sorted by key) when
68	68	travelling through different JSON implementations. In this manner,
69	69	clients can e.g. be sure to render the columns in the proper
70	70	(query-specified) order.
		@@ -76,10 +76,10 @@
76	76
77	77	Note the column names are never guaranteed to be exactly as they
78	78	appear in the SQL unless they are qualified with an AS, e.g. `SELECT
79	79	foo AS foo...`. When generating reports which need fixed column names, it
80	80	is highly recommended to use an AS qualifier for every column, even if
81		-they use the same name as the column. This is the only way to guaranty
	81	+they use the same name as the column. This is the only way to guarantee
82	82	that the result column names will be stable. (FYI: that behaviour comes
83	83	from sqlite3, not the JSON bits, and this behaviour has been known to
84	84	change between sqlite3 versions (so this is not just an idle threat of
85	85	potential future incompatibility).)
86	86

	--- www/json-api/api-query.md
	+++ www/json-api/api-query.md
	@@ -60,11 +60,11 @@
60	…
61	]
62	}
63	```
64
65	The column names are provided in a separate field is because their order
66	is guaranteed to match the order of the query columns, whereas object
67	key/value pairs might get reordered (typically sorted by key) when
68	travelling through different JSON implementations. In this manner,
69	clients can e.g. be sure to render the columns in the proper
70	(query-specified) order.
	@@ -76,10 +76,10 @@
76
77	Note the column names are never guaranteed to be exactly as they
78	appear in the SQL unless they are qualified with an AS, e.g. `SELECT
79	foo AS foo...`. When generating reports which need fixed column names, it
80	is highly recommended to use an AS qualifier for every column, even if
81	they use the same name as the column. This is the only way to guaranty
82	that the result column names will be stable. (FYI: that behaviour comes
83	from sqlite3, not the JSON bits, and this behaviour has been known to
84	change between sqlite3 versions (so this is not just an idle threat of
85	potential future incompatibility).)
86

	--- www/json-api/api-query.md
	+++ www/json-api/api-query.md
	@@ -60,11 +60,11 @@
60	…
61	]
62	}
63	```
64
65	The column names are provided in a separate field because their order
66	is guaranteed to match the order of the query columns, whereas object
67	key/value pairs might get reordered (typically sorted by key) when
68	travelling through different JSON implementations. In this manner,
69	clients can e.g. be sure to render the columns in the proper
70	(query-specified) order.
	@@ -76,10 +76,10 @@
76
77	Note the column names are never guaranteed to be exactly as they
78	appear in the SQL unless they are qualified with an AS, e.g. `SELECT
79	foo AS foo...`. When generating reports which need fixed column names, it
80	is highly recommended to use an AS qualifier for every column, even if
81	they use the same name as the column. This is the only way to guarantee
82	that the result column names will be stable. (FYI: that behaviour comes
83	from sqlite3, not the JSON bits, and this behaviour has been known to
84	change between sqlite3 versions (so this is not just an idle threat of
85	potential future incompatibility).)
86

M www/json-api/api-tag.md

+2 -2

		--- www/json-api/api-tag.md
		+++ www/json-api/api-tag.md
		@@ -98,15 +98,15 @@
98	98	path element.
99	99	- `limit=int` (defalt=0) Limits the number of results (0=no limit).
100	100	Since they are ordered from oldest to newest, the newest N results
101	101	will be returned.
102	102	- `type=string` (default=`*`) Searches only for the given type of
103		- artifact (using fossil's conventional type naming: ci, e, t, w.
	103	+ artifact (using fossil's conventional type naming: ci, e, t, w.)
104	104	- `raw=bool` (=false) If enabled, the response is an array of hashes
105	105	of the requested artifact type; otherwise,
106	106	it is an array of higher-level objects. If this is
107		- true, the "name" property is interpretted as-is. If it is false, the
	107	+ true, the "name" property is interpreted as-is. If it is false, the
108	108	name is automatically prepended with "sym-" (meaning a branch).
109	109	(FIXME: the current semantics are confusing and hard to remember.
110	110	Re-do them.)
111	111
112	112	**Response payload example, in RAW mode: (expect this format to change
113	113

	--- www/json-api/api-tag.md
	+++ www/json-api/api-tag.md
	@@ -98,15 +98,15 @@
98	path element.
99	- `limit=int` (defalt=0) Limits the number of results (0=no limit).
100	Since they are ordered from oldest to newest, the newest N results
101	will be returned.
102	- `type=string` (default=`*`) Searches only for the given type of
103	artifact (using fossil's conventional type naming: ci, e, t, w.
104	- `raw=bool` (=false) If enabled, the response is an array of hashes
105	of the requested artifact type; otherwise,
106	it is an array of higher-level objects. If this is
107	true, the "name" property is interpretted as-is. If it is false, the
108	name is automatically prepended with "sym-" (meaning a branch).
109	(FIXME: the current semantics are confusing and hard to remember.
110	Re-do them.)
111
112	**Response payload example, in RAW mode: (expect this format to change
113

	--- www/json-api/api-tag.md
	+++ www/json-api/api-tag.md
	@@ -98,15 +98,15 @@
98	path element.
99	- `limit=int` (defalt=0) Limits the number of results (0=no limit).
100	Since they are ordered from oldest to newest, the newest N results
101	will be returned.
102	- `type=string` (default=`*`) Searches only for the given type of
103	artifact (using fossil's conventional type naming: ci, e, t, w.)
104	- `raw=bool` (=false) If enabled, the response is an array of hashes
105	of the requested artifact type; otherwise,
106	it is an array of higher-level objects. If this is
107	true, the "name" property is interpreted as-is. If it is false, the
108	name is automatically prepended with "sym-" (meaning a branch).
109	(FIXME: the current semantics are confusing and hard to remember.
110	Re-do them.)
111
112	**Response payload example, in RAW mode: (expect this format to change
113

M www/json-api/conventions.md

+8 -8

		--- www/json-api/conventions.md
		+++ www/json-api/conventions.md
		@@ -100,11 +100,11 @@
100	100
101	101	The API allows most of those (normally all but the payload) to come in
102	102	as either GET parameters or properties of the top-level POSTed request
103	103	JSON envelope, with GET taking priority over POST. (Reminder to self: we
104	104	could potentially also use values from cookies. Fossil currently only
105		-uses 1 cookie (the login token), and i'd prefer to keep it that way.)
	105	+uses 1 cookie (the login token), and I'd prefer to keep it that way.)
106	106
107	107	POST requests without such an envelope will be rejected, generating a
108	108	Fossil/JSON error response (as opposed to an HTTP error response). GET
109	109	requests, by definition, never have an envelope.
110	110
		@@ -169,11 +169,11 @@
169	169	No higher-level constructs, e.g. JSON arrays or objects, are
170	170	accepted in string form. Such parameters must be set in the POST
171	171	envelope or payload, as specified by the specific API.
172	172
173	173	This API does not currently use any floating-point parameters, but
174		-does return floating-point results in a couple places.
	174	+does return floating-point results in a couple of places.
175	175
176	176	For integer parameters we use a conventional string-to-int algorithm
177	177	and assume base 10 (analog to atoi(3)). The API may err on the side of
178	178	usability when given technically invalid values. e.g. "123abc" will
179	179	likely be interpreted as the integer 123. No APIs currently rely on
		@@ -205,11 +205,11 @@
205	205	cause inconsistencies vis-a-vis the HTML interface).
206	206
207	207	<a id="response-envelope"></a>
208	208	# Response Envelope
209	209
210		-Every response comes in the form of a HTTP response or (in CLI mode)
	210	+Every response comes in the form of an HTTP response or (in CLI mode)
211	211	JSON sent to stdout. The body of the response is a JSON object following
212	212	a common envelope format. The envelope has the following properties:
213	213
214	214
215	215	- `fossil`: Fossil server version string. This property is basically
		@@ -216,11 +216,11 @@
216	216	"the official response envelope marker" - if it is set, clients can
217	217	"probably safely assume" that the object indeed came from one of the
218	218	Fossil/JSON APIs. This API never creates responses which do not
219	219	contain this property.
220	220	- `requestId`: Only set if the request contained it, and then it is
221		- echoed back to the caller as-is. This can be use to determine
	221	+ echoed back to the caller as-is. This can be used to determine
222	222	(client-side) which request a given response is coming in for
223	223	(assuming multiple asynchronous requests are pending). In practice
224	224	this generally isn’t needed because response handling tends to be
225	225	done by closures associated with the original request object (at
226	226	least in JavaScript code). In languages without closures it might
		@@ -235,11 +235,11 @@
235	235	- `payload`: Request-specific response payload (data type/structure is
236	236	request-specific). The payload is never set for error responses,
237	237	only for success responses (and only those which actually have a
238	238	payload - not all do).
239	239	- `timestamp`: Response timestamp (GMT Unix Epoch). We use seconds
240		- precision because i did not know at the time that Fossil actually
	240	+ precision because I did not know at the time that Fossil actually
241	241	records millisecond precision.
242	242	- `payloadVersion`: Not initially needed, but reserved for future use
243	243	in maintaining version compatibility when the format of a given
244	244	response type's payload changes. If needed, the "first version"
245	245	value is assumed to be 0, for semantic [near-]compatibility with the
		@@ -402,11 +402,11 @@
402	402	(they are represented by `\t`) there is no chance that such a global
403	403	replacement will corrupt JSON string contents - only the formatting will
404	404	be affected.
405	405
406	406	Potential TODO: because extraneous indention "could potentially" be used
407		-as a form DoS, the option might be subject to later removed from HTTP
	407	+as a form of DoS, the option might be subject to later removal from HTTP
408	408	mode (in CLI it's fine).
409	409
410	410	In HTTP mode no trailing newline is added to the output, whereas in CLI
411	411	mode one is normally appended (exception: in JSONP mode no newline is
412	412	appended, to (rather pedantically and arbitraily) allow the client to
		@@ -482,11 +482,11 @@
482	482	normally want to see very specific error codes when tracking down a
483	483	problem. We can offer a configuration option to "dumb down" error
484	484	codes to their generic category by simply doing a modulo 100
485	485	(or 1000) against the native error code number. e.g. FOSSIL-1271
486	486	could (via a simple modulo) be reduced to FOSSIL-1200 or
487		- FOSSIL-1000, depending on the paranoia level of the sysadmin. i have
	487	+ FOSSIL-1000, depending on the paranoia level of the sysadmin. I have
488	488	tried to order the result code numbers so that a dumb-down level of
489	489	2 provides reasonably usable results without giving away too much
490	490	detail to malicious clients.\
491	491	(TODO: `g.json.errorDetailParanoia` is used to set the
492	492	default dumb-down level, but it is currently set at compile-time.
		@@ -533,11 +533,11 @@
533	533	can actually implement this one, though.)
534	534	- `FOSSIL-1106`: Assertion failed (or would have had we
535	535	continued). Note: if an `assert()` fails in CGI/server modes, the HTTP
536	536	response will be code 500 (Internal Server Error). We want to avoid
537	537	that and return a JSON response instead. All of that said - there seems
538		- to be little reason to implementi this, since assertions are "truly
	538	+ to be little reason to implement this, since assertions are "truly
539	539	serious" errors.
540	540	- `FOSSIL-1107`: Allocation/out of memory error. This cannot be reasonably
541	541	reported because fossil aborts if an allocation fails.
542	542	- `FOSSIL-1108`: Requested API is not yet implemented.
543	543	- `FOSSIL-1109`: Panic! Fossil's `fossil_panic()` or `cgi_panic()` was
544	544

	--- www/json-api/conventions.md
	+++ www/json-api/conventions.md
	@@ -100,11 +100,11 @@
100
101	The API allows most of those (normally all but the payload) to come in
102	as either GET parameters or properties of the top-level POSTed request
103	JSON envelope, with GET taking priority over POST. (Reminder to self: we
104	could potentially also use values from cookies. Fossil currently only
105	uses 1 cookie (the login token), and i'd prefer to keep it that way.)
106
107	POST requests without such an envelope will be rejected, generating a
108	Fossil/JSON error response (as opposed to an HTTP error response). GET
109	requests, by definition, never have an envelope.
110
	@@ -169,11 +169,11 @@
169	No higher-level constructs, e.g. JSON arrays or objects, are
170	accepted in string form. Such parameters must be set in the POST
171	envelope or payload, as specified by the specific API.
172
173	This API does not currently use any floating-point parameters, but
174	does return floating-point results in a couple places.
175
176	For integer parameters we use a conventional string-to-int algorithm
177	and assume base 10 (analog to atoi(3)). The API may err on the side of
178	usability when given technically invalid values. e.g. "123abc" will
179	likely be interpreted as the integer 123. No APIs currently rely on
	@@ -205,11 +205,11 @@
205	cause inconsistencies vis-a-vis the HTML interface).
206
207	<a id="response-envelope"></a>
208	# Response Envelope
209
210	Every response comes in the form of a HTTP response or (in CLI mode)
211	JSON sent to stdout. The body of the response is a JSON object following
212	a common envelope format. The envelope has the following properties:
213
214
215	- `fossil`: Fossil server version string. This property is basically
	@@ -216,11 +216,11 @@
216	"the official response envelope marker" - if it is set, clients can
217	"probably safely assume" that the object indeed came from one of the
218	Fossil/JSON APIs. This API never creates responses which do not
219	contain this property.
220	- `requestId`: Only set if the request contained it, and then it is
221	echoed back to the caller as-is. This can be use to determine
222	(client-side) which request a given response is coming in for
223	(assuming multiple asynchronous requests are pending). In practice
224	this generally isn’t needed because response handling tends to be
225	done by closures associated with the original request object (at
226	least in JavaScript code). In languages without closures it might
	@@ -235,11 +235,11 @@
235	- `payload`: Request-specific response payload (data type/structure is
236	request-specific). The payload is never set for error responses,
237	only for success responses (and only those which actually have a
238	payload - not all do).
239	- `timestamp`: Response timestamp (GMT Unix Epoch). We use seconds
240	precision because i did not know at the time that Fossil actually
241	records millisecond precision.
242	- `payloadVersion`: Not initially needed, but reserved for future use
243	in maintaining version compatibility when the format of a given
244	response type's payload changes. If needed, the "first version"
245	value is assumed to be 0, for semantic [near-]compatibility with the
	@@ -402,11 +402,11 @@
402	(they are represented by `\t`) there is no chance that such a global
403	replacement will corrupt JSON string contents - only the formatting will
404	be affected.
405
406	Potential TODO: because extraneous indention "could potentially" be used
407	as a form DoS, the option might be subject to later removed from HTTP
408	mode (in CLI it's fine).
409
410	In HTTP mode no trailing newline is added to the output, whereas in CLI
411	mode one is normally appended (exception: in JSONP mode no newline is
412	appended, to (rather pedantically and arbitraily) allow the client to
	@@ -482,11 +482,11 @@
482	normally want to see very specific error codes when tracking down a
483	problem. We can offer a configuration option to "dumb down" error
484	codes to their generic category by simply doing a modulo 100
485	(or 1000) against the native error code number. e.g. FOSSIL-1271
486	could (via a simple modulo) be reduced to FOSSIL-1200 or
487	FOSSIL-1000, depending on the paranoia level of the sysadmin. i have
488	tried to order the result code numbers so that a dumb-down level of
489	2 provides reasonably usable results without giving away too much
490	detail to malicious clients.\
491	(TODO: `g.json.errorDetailParanoia` is used to set the
492	default dumb-down level, but it is currently set at compile-time.
	@@ -533,11 +533,11 @@
533	can actually implement this one, though.)
534	- `FOSSIL-1106`: Assertion failed (or would have had we
535	continued). Note: if an `assert()` fails in CGI/server modes, the HTTP
536	response will be code 500 (Internal Server Error). We want to avoid
537	that and return a JSON response instead. All of that said - there seems
538	to be little reason to implementi this, since assertions are "truly
539	serious" errors.
540	- `FOSSIL-1107`: Allocation/out of memory error. This cannot be reasonably
541	reported because fossil aborts if an allocation fails.
542	- `FOSSIL-1108`: Requested API is not yet implemented.
543	- `FOSSIL-1109`: Panic! Fossil's `fossil_panic()` or `cgi_panic()` was
544

	--- www/json-api/conventions.md
	+++ www/json-api/conventions.md
	@@ -100,11 +100,11 @@
100
101	The API allows most of those (normally all but the payload) to come in
102	as either GET parameters or properties of the top-level POSTed request
103	JSON envelope, with GET taking priority over POST. (Reminder to self: we
104	could potentially also use values from cookies. Fossil currently only
105	uses 1 cookie (the login token), and I'd prefer to keep it that way.)
106
107	POST requests without such an envelope will be rejected, generating a
108	Fossil/JSON error response (as opposed to an HTTP error response). GET
109	requests, by definition, never have an envelope.
110
	@@ -169,11 +169,11 @@
169	No higher-level constructs, e.g. JSON arrays or objects, are
170	accepted in string form. Such parameters must be set in the POST
171	envelope or payload, as specified by the specific API.
172
173	This API does not currently use any floating-point parameters, but
174	does return floating-point results in a couple of places.
175
176	For integer parameters we use a conventional string-to-int algorithm
177	and assume base 10 (analog to atoi(3)). The API may err on the side of
178	usability when given technically invalid values. e.g. "123abc" will
179	likely be interpreted as the integer 123. No APIs currently rely on
	@@ -205,11 +205,11 @@
205	cause inconsistencies vis-a-vis the HTML interface).
206
207	<a id="response-envelope"></a>
208	# Response Envelope
209
210	Every response comes in the form of an HTTP response or (in CLI mode)
211	JSON sent to stdout. The body of the response is a JSON object following
212	a common envelope format. The envelope has the following properties:
213
214
215	- `fossil`: Fossil server version string. This property is basically
	@@ -216,11 +216,11 @@
216	"the official response envelope marker" - if it is set, clients can
217	"probably safely assume" that the object indeed came from one of the
218	Fossil/JSON APIs. This API never creates responses which do not
219	contain this property.
220	- `requestId`: Only set if the request contained it, and then it is
221	echoed back to the caller as-is. This can be used to determine
222	(client-side) which request a given response is coming in for
223	(assuming multiple asynchronous requests are pending). In practice
224	this generally isn’t needed because response handling tends to be
225	done by closures associated with the original request object (at
226	least in JavaScript code). In languages without closures it might
	@@ -235,11 +235,11 @@
235	- `payload`: Request-specific response payload (data type/structure is
236	request-specific). The payload is never set for error responses,
237	only for success responses (and only those which actually have a
238	payload - not all do).
239	- `timestamp`: Response timestamp (GMT Unix Epoch). We use seconds
240	precision because I did not know at the time that Fossil actually
241	records millisecond precision.
242	- `payloadVersion`: Not initially needed, but reserved for future use
243	in maintaining version compatibility when the format of a given
244	response type's payload changes. If needed, the "first version"
245	value is assumed to be 0, for semantic [near-]compatibility with the
	@@ -402,11 +402,11 @@
402	(they are represented by `\t`) there is no chance that such a global
403	replacement will corrupt JSON string contents - only the formatting will
404	be affected.
405
406	Potential TODO: because extraneous indention "could potentially" be used
407	as a form of DoS, the option might be subject to later removal from HTTP
408	mode (in CLI it's fine).
409
410	In HTTP mode no trailing newline is added to the output, whereas in CLI
411	mode one is normally appended (exception: in JSONP mode no newline is
412	appended, to (rather pedantically and arbitraily) allow the client to
	@@ -482,11 +482,11 @@
482	normally want to see very specific error codes when tracking down a
483	problem. We can offer a configuration option to "dumb down" error
484	codes to their generic category by simply doing a modulo 100
485	(or 1000) against the native error code number. e.g. FOSSIL-1271
486	could (via a simple modulo) be reduced to FOSSIL-1200 or
487	FOSSIL-1000, depending on the paranoia level of the sysadmin. I have
488	tried to order the result code numbers so that a dumb-down level of
489	2 provides reasonably usable results without giving away too much
490	detail to malicious clients.\
491	(TODO: `g.json.errorDetailParanoia` is used to set the
492	default dumb-down level, but it is currently set at compile-time.
	@@ -533,11 +533,11 @@
533	can actually implement this one, though.)
534	- `FOSSIL-1106`: Assertion failed (or would have had we
535	continued). Note: if an `assert()` fails in CGI/server modes, the HTTP
536	response will be code 500 (Internal Server Error). We want to avoid
537	that and return a JSON response instead. All of that said - there seems
538	to be little reason to implement this, since assertions are "truly
539	serious" errors.
540	- `FOSSIL-1107`: Allocation/out of memory error. This cannot be reasonably
541	reported because fossil aborts if an allocation fails.
542	- `FOSSIL-1108`: Requested API is not yet implemented.
543	- `FOSSIL-1109`: Panic! Fossil's `fossil_panic()` or `cgi_panic()` was
544

M www/json-api/hacking.md

+7 -7

		--- www/json-api/hacking.md
		+++ www/json-api/hacking.md
		@@ -79,11 +79,11 @@
79	79	core. The disadvantages of this are that we lose fossil's conventional
80	80	help text mechanism (which is based on code comments in the
81	81	command/path's dispatcher impl) and the ability to write abbreviated
82	82	command names in CLI mode ("json" itself may be abbreviated, but not the
83	83	subcommands). The advantages are that we can handle CLI/HTTP modes
84		-almost identically (there are a couple minor differences) by unifying
	84	+almost identically (there are a couple of minor differences) by unifying
85	85	them under the same callback functions much more easily.
86	86
87	87	The top-level "json" command/path uses its own dispatching mechanism
88	88	which uses either the path (in HTTP mode) or CLI positional arguments to
89	89	dispatch commands (stopping at the first "flag option" (e.g. -foo) in
		@@ -125,11 +125,11 @@
125	125	(e.g. `db_prepare()` can be fatal, and callbacks may call `fossil_panic()`
126	126	if they really want to). One exception is `fossil_exit()`, which does
127	127	_not_ generate any extra output and will `exit()` the app. In the JSON
128	128	API, as a rule of thumb, `fossil_exit()` is only used when we want a
129	129	failed request to cause an HTTP 500 error, and it is reserved for
130		-allocation errors and similar truly catostrophic failures. That said...
	130	+allocation errors and similar truly catastrophic failures. That said...
131	131	libcson has been hacked to use `fossil_alloc()` and friends for memory
132	132	management, and those routines exit on error, so alloc error handling in
133	133	the JSON command handler code can afford to be a little lax (the
134	134	majority of potential errors clients get from the cson API have
135	135	allocation failure as their root cause).
		@@ -259,11 +259,11 @@
259	259	- `json_getenv()` and `json_getenv_TYPE()` search the so-called "JSON
260	260	environment," which is a superset of the GET/POST/`POST.payload` (if
261	261	`POST.payload` is-a Object).
262	262	- `json_find_option_TYPE()`: searches the CLI args (only when in CLI
263	263	mode) and the JSON environment.
264		-- The use of fossil's `P()` and `PD()` macros is discourages in JSON
	264	+- The use of fossil's `P()` and `PD()` macros is discouraged in JSON
265	265	callbacks because they can only handle String data from the CLI or
266	266	GET parameters (not POST/`POST.payload`). (Note that `P()` and `PD()`
267	267	normally also handle POSTed keys, but they only "see" values
268	268	posted as form-urlencoded fields, and not JSON format.)
269	269	- `find_option()` (from `src/main.c`) "should" also be avoided in
		@@ -280,11 +280,11 @@
280	280
281	281	<a href="creating-json-values"></a>
282	282	## Creating JSON Values
283	283
284	284	cson has a fairly rich API for creating and manipulating the various
285		-JSON-defined value types. For a detailed overview and demonstration i
	285	+JSON-defined value types. For a detailed overview and demonstration I
286	286	recommend reading:
287	287
288	288	[](https://fossil.wanderinghorse.net/wikis/cson/?page=HowTo)
289	289
290	290	That said, the Fossil/JSON API has several convenience wrappers to save
		@@ -310,15 +310,15 @@
310	310	to Arrays or Objects, or convert single columns to a JSON-compatible
311	311	form. See `json_stmt_to_array_of_obj()`,
312	312	`json_stmt_to_array_of_array()` (both in `src/json.c`), and
313	313	`cson_sqlite3_column_to_value()` and friends (in
314	314	`extsrc/cson_amalgamation.h`). They work in an intuitive way for numeric
315		-types, but they optimistically/natively assume that any fields of type
	315	+types, but they optimistically/naively assume that any fields of type
316	316	TEXT or BLOB are actually UTF8 data, and treat them as such. cson's
317	317	string class only handles UTF8 data and it is semantically illegal to
318	318	feed them anything but UTF8. Violating this will likely result in
319		-down-stream errors (e.g. when emiting the JSON string output). **The
	319	+down-stream errors (e.g. when emitting the JSON string output). **The
320	320	moral of this story is:** do not use these APIs to fetch binary data.
321	321	JSON doesn't do binary and the `cson_string` class does not
322	322	protect itself against clients feeding it non-UTF8 data.
323	323
324	324	Here's a basic example of using these features:
		@@ -347,7 +347,7 @@
347	347	AS clauses to be guaranteed that the db driver will return the column
348	348	names we want. Note that the AS clause is often used to translate column
349	349	names into something more JSON-conventional or user-friendly, e.g.
350	350	"SELECT cap AS capabilities...". Alternately, we can convert the
351	351	individual `sqlite3_stmt` column values to JSON using
352		-`cson_sqlite3_column_to_value()`, without refering directly to the
	352	+`cson_sqlite3_column_to_value()`, without referring directly to the
353	353	db-reported column name.
354	354

	--- www/json-api/hacking.md
	+++ www/json-api/hacking.md
	@@ -79,11 +79,11 @@
79	core. The disadvantages of this are that we lose fossil's conventional
80	help text mechanism (which is based on code comments in the
81	command/path's dispatcher impl) and the ability to write abbreviated
82	command names in CLI mode ("json" itself may be abbreviated, but not the
83	subcommands). The advantages are that we can handle CLI/HTTP modes
84	almost identically (there are a couple minor differences) by unifying
85	them under the same callback functions much more easily.
86
87	The top-level "json" command/path uses its own dispatching mechanism
88	which uses either the path (in HTTP mode) or CLI positional arguments to
89	dispatch commands (stopping at the first "flag option" (e.g. -foo) in
	@@ -125,11 +125,11 @@
125	(e.g. `db_prepare()` can be fatal, and callbacks may call `fossil_panic()`
126	if they really want to). One exception is `fossil_exit()`, which does
127	_not_ generate any extra output and will `exit()` the app. In the JSON
128	API, as a rule of thumb, `fossil_exit()` is only used when we want a
129	failed request to cause an HTTP 500 error, and it is reserved for
130	allocation errors and similar truly catostrophic failures. That said...
131	libcson has been hacked to use `fossil_alloc()` and friends for memory
132	management, and those routines exit on error, so alloc error handling in
133	the JSON command handler code can afford to be a little lax (the
134	majority of potential errors clients get from the cson API have
135	allocation failure as their root cause).
	@@ -259,11 +259,11 @@
259	- `json_getenv()` and `json_getenv_TYPE()` search the so-called "JSON
260	environment," which is a superset of the GET/POST/`POST.payload` (if
261	`POST.payload` is-a Object).
262	- `json_find_option_TYPE()`: searches the CLI args (only when in CLI
263	mode) and the JSON environment.
264	- The use of fossil's `P()` and `PD()` macros is discourages in JSON
265	callbacks because they can only handle String data from the CLI or
266	GET parameters (not POST/`POST.payload`). (Note that `P()` and `PD()`
267	normally also handle POSTed keys, but they only "see" values
268	posted as form-urlencoded fields, and not JSON format.)
269	- `find_option()` (from `src/main.c`) "should" also be avoided in
	@@ -280,11 +280,11 @@
280
281	<a href="creating-json-values"></a>
282	## Creating JSON Values
283
284	cson has a fairly rich API for creating and manipulating the various
285	JSON-defined value types. For a detailed overview and demonstration i
286	recommend reading:
287
288	[](https://fossil.wanderinghorse.net/wikis/cson/?page=HowTo)
289
290	That said, the Fossil/JSON API has several convenience wrappers to save
	@@ -310,15 +310,15 @@
310	to Arrays or Objects, or convert single columns to a JSON-compatible
311	form. See `json_stmt_to_array_of_obj()`,
312	`json_stmt_to_array_of_array()` (both in `src/json.c`), and
313	`cson_sqlite3_column_to_value()` and friends (in
314	`extsrc/cson_amalgamation.h`). They work in an intuitive way for numeric
315	types, but they optimistically/natively assume that any fields of type
316	TEXT or BLOB are actually UTF8 data, and treat them as such. cson's
317	string class only handles UTF8 data and it is semantically illegal to
318	feed them anything but UTF8. Violating this will likely result in
319	down-stream errors (e.g. when emiting the JSON string output). **The
320	moral of this story is:** do not use these APIs to fetch binary data.
321	JSON doesn't do binary and the `cson_string` class does not
322	protect itself against clients feeding it non-UTF8 data.
323
324	Here's a basic example of using these features:
	@@ -347,7 +347,7 @@
347	AS clauses to be guaranteed that the db driver will return the column
348	names we want. Note that the AS clause is often used to translate column
349	names into something more JSON-conventional or user-friendly, e.g.
350	"SELECT cap AS capabilities...". Alternately, we can convert the
351	individual `sqlite3_stmt` column values to JSON using
352	`cson_sqlite3_column_to_value()`, without refering directly to the
353	db-reported column name.
354

	--- www/json-api/hacking.md
	+++ www/json-api/hacking.md
	@@ -79,11 +79,11 @@
79	core. The disadvantages of this are that we lose fossil's conventional
80	help text mechanism (which is based on code comments in the
81	command/path's dispatcher impl) and the ability to write abbreviated
82	command names in CLI mode ("json" itself may be abbreviated, but not the
83	subcommands). The advantages are that we can handle CLI/HTTP modes
84	almost identically (there are a couple of minor differences) by unifying
85	them under the same callback functions much more easily.
86
87	The top-level "json" command/path uses its own dispatching mechanism
88	which uses either the path (in HTTP mode) or CLI positional arguments to
89	dispatch commands (stopping at the first "flag option" (e.g. -foo) in
	@@ -125,11 +125,11 @@
125	(e.g. `db_prepare()` can be fatal, and callbacks may call `fossil_panic()`
126	if they really want to). One exception is `fossil_exit()`, which does
127	_not_ generate any extra output and will `exit()` the app. In the JSON
128	API, as a rule of thumb, `fossil_exit()` is only used when we want a
129	failed request to cause an HTTP 500 error, and it is reserved for
130	allocation errors and similar truly catastrophic failures. That said...
131	libcson has been hacked to use `fossil_alloc()` and friends for memory
132	management, and those routines exit on error, so alloc error handling in
133	the JSON command handler code can afford to be a little lax (the
134	majority of potential errors clients get from the cson API have
135	allocation failure as their root cause).
	@@ -259,11 +259,11 @@
259	- `json_getenv()` and `json_getenv_TYPE()` search the so-called "JSON
260	environment," which is a superset of the GET/POST/`POST.payload` (if
261	`POST.payload` is-a Object).
262	- `json_find_option_TYPE()`: searches the CLI args (only when in CLI
263	mode) and the JSON environment.
264	- The use of fossil's `P()` and `PD()` macros is discouraged in JSON
265	callbacks because they can only handle String data from the CLI or
266	GET parameters (not POST/`POST.payload`). (Note that `P()` and `PD()`
267	normally also handle POSTed keys, but they only "see" values
268	posted as form-urlencoded fields, and not JSON format.)
269	- `find_option()` (from `src/main.c`) "should" also be avoided in
	@@ -280,11 +280,11 @@
280
281	<a href="creating-json-values"></a>
282	## Creating JSON Values
283
284	cson has a fairly rich API for creating and manipulating the various
285	JSON-defined value types. For a detailed overview and demonstration I
286	recommend reading:
287
288	[](https://fossil.wanderinghorse.net/wikis/cson/?page=HowTo)
289
290	That said, the Fossil/JSON API has several convenience wrappers to save
	@@ -310,15 +310,15 @@
310	to Arrays or Objects, or convert single columns to a JSON-compatible
311	form. See `json_stmt_to_array_of_obj()`,
312	`json_stmt_to_array_of_array()` (both in `src/json.c`), and
313	`cson_sqlite3_column_to_value()` and friends (in
314	`extsrc/cson_amalgamation.h`). They work in an intuitive way for numeric
315	types, but they optimistically/naively assume that any fields of type
316	TEXT or BLOB are actually UTF8 data, and treat them as such. cson's
317	string class only handles UTF8 data and it is semantically illegal to
318	feed them anything but UTF8. Violating this will likely result in
319	down-stream errors (e.g. when emitting the JSON string output). **The
320	moral of this story is:** do not use these APIs to fetch binary data.
321	JSON doesn't do binary and the `cson_string` class does not
322	protect itself against clients feeding it non-UTF8 data.
323
324	Here's a basic example of using these features:
	@@ -347,7 +347,7 @@
347	AS clauses to be guaranteed that the db driver will return the column
348	names we want. Note that the AS clause is often used to translate column
349	names into something more JSON-conventional or user-friendly, e.g.
350	"SELECT cap AS capabilities...". Alternately, we can convert the
351	individual `sqlite3_stmt` column values to JSON using
352	`cson_sqlite3_column_to_value()`, without referring directly to the
353	db-reported column name.
354

M www/makefile.wiki

+10 -9

		--- www/makefile.wiki
		+++ www/makefile.wiki
		@@ -7,11 +7,11 @@
7	7	before it is compiled. Most users will download a
8	8	[https://fossil-scm.org/home/uv/download.html \| precompiled binary]
9	9	so this is of no consequence to them, and even those who
10	10	want to compile the code themselves can use one of the
11	11	[./build.wiki \| existing makefiles].
12		-So must people do not need to be concerned with the
	12	+So most people do not need to be concerned with the
13	13	build complexities of Fossil. But hard-core developers who desire
14	14	a deep understanding of how Fossil is put together can benefit
15	15	from reviewing this article.
16	16
17	17	<h1 id="srctour">2.0 Source Code Tour</h1>
		@@ -56,21 +56,21 @@
56	56	The TH1 script engine is implemented using files:
57	57
58	58	9. th.c
59	59	10. th.h
60	60
61		-The proprocessing steps are omitted for all of these imported
	61	+The preprocessing steps are omitted for all of these imported
62	62	files.
63	63
64	64	The VERSION.h header file is generated from other information sources
65	65	using a small program called:
66	66
67	67	11. [/file/tools/mkversion.c \| mkversion.c]
68	68
69	69	The builtin_data.h header file contains the definitions of C-language
70	70	byte-array constants that contain various resources such as scripts and
71		-images. The builtin_data.h header file is generate from the original
	71	+images. The builtin_data.h header file is generated from the original
72	72	resource files using a small program called:
73	73
74	74	12 [/file/tools/mkbuiltin.c \| mkbuiltin.c]
75	75
76	76	Examples of built-in resources include the [/file/src/diff.tcl \| diff.tcl]
		@@ -132,11 +132,11 @@
132	132	the exceptions described above.
133	133
134	134	<h1>3.0 Automatically generated files</h1>
135	135
136	136	The "VERSION.h" header file contains some C preprocessor macros that
137		-identify the version of Fossil that is to be build. The VERSION.h file is
	137	+identify the version of Fossil that is to be built. The VERSION.h file is
138	138	generated automatically from information extracted from the "manifest",
139	139	"manifest.uuid", and "VERSION" source files in the root directory of the
140	140	source tree.
141	141	(The "manifest" and "manifest.uuid" files are automatically generated and
142	142	updated by Fossil itself. See the [/help/setting \| fossil set manifest]
		@@ -294,12 +294,12 @@
294	294	link against the standard C library. No other libraries or external
295	295	dependences are used.
296	296
297	297	<h1>7.0 Debugging</h1>
298	298
299		-Debug mode is controlled via FOSSIL_DEBUG preprocessor macro which could be
300		-set explicitly at the make command for the target platform.
	299	+Debug mode is controlled via a FOSSIL_DEBUG preprocessor macro. This can be
	300	+set explicitly with the make command for the target platform.
301	301
302	302	However, in practice it is instead recommended to add a respective configure
303	303	option for the target platform and then perform a clean build. This way the
304	304	Debug flags are consistently applied across the whole build process. For
305	305	example, use these Debug flags in addition to other flags passed to the
		@@ -313,14 +313,15 @@
313	313	On Windows:
314	314	<pre>
315	315	win\buildmsvc.bat FOSSIL_DEBUG=1
316	316	</pre>
317	317
318		-The resulting fossil binary could then be loaded into a platform-specific
	318	+The resulting fossil binary can then be loaded into a platform-specific
319	319	debugger. Source files displayed in the debugger correspond to the ones
320		-generated from the translation stage of the build process, that is what was
321		-actually compiled into the object files.
	320	+generated from the translation stage of the build process, as that is what
	321	+was actually compiled into the respective object files that make up the
	322	+fossil binary.
322	323
323	324	<h1>8.0 See Also</h1>
324	325
325	326	* [./tech_overview.wiki \| A Technical Overview Of Fossil]
326	327	* [./adding_code.wiki \| How To Add Features To Fossil]
327	328

	--- www/makefile.wiki
	+++ www/makefile.wiki
	@@ -7,11 +7,11 @@
7	before it is compiled. Most users will download a
8	[https://fossil-scm.org/home/uv/download.html \| precompiled binary]
9	so this is of no consequence to them, and even those who
10	want to compile the code themselves can use one of the
11	[./build.wiki \| existing makefiles].
12	So must people do not need to be concerned with the
13	build complexities of Fossil. But hard-core developers who desire
14	a deep understanding of how Fossil is put together can benefit
15	from reviewing this article.
16
17	<h1 id="srctour">2.0 Source Code Tour</h1>
	@@ -56,21 +56,21 @@
56	The TH1 script engine is implemented using files:
57
58	9. th.c
59	10. th.h
60
61	The proprocessing steps are omitted for all of these imported
62	files.
63
64	The VERSION.h header file is generated from other information sources
65	using a small program called:
66
67	11. [/file/tools/mkversion.c \| mkversion.c]
68
69	The builtin_data.h header file contains the definitions of C-language
70	byte-array constants that contain various resources such as scripts and
71	images. The builtin_data.h header file is generate from the original
72	resource files using a small program called:
73
74	12 [/file/tools/mkbuiltin.c \| mkbuiltin.c]
75
76	Examples of built-in resources include the [/file/src/diff.tcl \| diff.tcl]
	@@ -132,11 +132,11 @@
132	the exceptions described above.
133
134	<h1>3.0 Automatically generated files</h1>
135
136	The "VERSION.h" header file contains some C preprocessor macros that
137	identify the version of Fossil that is to be build. The VERSION.h file is
138	generated automatically from information extracted from the "manifest",
139	"manifest.uuid", and "VERSION" source files in the root directory of the
140	source tree.
141	(The "manifest" and "manifest.uuid" files are automatically generated and
142	updated by Fossil itself. See the [/help/setting \| fossil set manifest]
	@@ -294,12 +294,12 @@
294	link against the standard C library. No other libraries or external
295	dependences are used.
296
297	<h1>7.0 Debugging</h1>
298
299	Debug mode is controlled via FOSSIL_DEBUG preprocessor macro which could be
300	set explicitly at the make command for the target platform.
301
302	However, in practice it is instead recommended to add a respective configure
303	option for the target platform and then perform a clean build. This way the
304	Debug flags are consistently applied across the whole build process. For
305	example, use these Debug flags in addition to other flags passed to the
	@@ -313,14 +313,15 @@
313	On Windows:
314	<pre>
315	win\buildmsvc.bat FOSSIL_DEBUG=1
316	</pre>
317
318	The resulting fossil binary could then be loaded into a platform-specific
319	debugger. Source files displayed in the debugger correspond to the ones
320	generated from the translation stage of the build process, that is what was
321	actually compiled into the object files.

322
323	<h1>8.0 See Also</h1>
324
325	* [./tech_overview.wiki \| A Technical Overview Of Fossil]
326	* [./adding_code.wiki \| How To Add Features To Fossil]
327

	--- www/makefile.wiki
	+++ www/makefile.wiki
	@@ -7,11 +7,11 @@
7	before it is compiled. Most users will download a
8	[https://fossil-scm.org/home/uv/download.html \| precompiled binary]
9	so this is of no consequence to them, and even those who
10	want to compile the code themselves can use one of the
11	[./build.wiki \| existing makefiles].
12	So most people do not need to be concerned with the
13	build complexities of Fossil. But hard-core developers who desire
14	a deep understanding of how Fossil is put together can benefit
15	from reviewing this article.
16
17	<h1 id="srctour">2.0 Source Code Tour</h1>
	@@ -56,21 +56,21 @@
56	The TH1 script engine is implemented using files:
57
58	9. th.c
59	10. th.h
60
61	The preprocessing steps are omitted for all of these imported
62	files.
63
64	The VERSION.h header file is generated from other information sources
65	using a small program called:
66
67	11. [/file/tools/mkversion.c \| mkversion.c]
68
69	The builtin_data.h header file contains the definitions of C-language
70	byte-array constants that contain various resources such as scripts and
71	images. The builtin_data.h header file is generated from the original
72	resource files using a small program called:
73
74	12 [/file/tools/mkbuiltin.c \| mkbuiltin.c]
75
76	Examples of built-in resources include the [/file/src/diff.tcl \| diff.tcl]
	@@ -132,11 +132,11 @@
132	the exceptions described above.
133
134	<h1>3.0 Automatically generated files</h1>
135
136	The "VERSION.h" header file contains some C preprocessor macros that
137	identify the version of Fossil that is to be built. The VERSION.h file is
138	generated automatically from information extracted from the "manifest",
139	"manifest.uuid", and "VERSION" source files in the root directory of the
140	source tree.
141	(The "manifest" and "manifest.uuid" files are automatically generated and
142	updated by Fossil itself. See the [/help/setting \| fossil set manifest]
	@@ -294,12 +294,12 @@
294	link against the standard C library. No other libraries or external
295	dependences are used.
296
297	<h1>7.0 Debugging</h1>
298
299	Debug mode is controlled via a FOSSIL_DEBUG preprocessor macro. This can be
300	set explicitly with the make command for the target platform.
301
302	However, in practice it is instead recommended to add a respective configure
303	option for the target platform and then perform a clean build. This way the
304	Debug flags are consistently applied across the whole build process. For
305	example, use these Debug flags in addition to other flags passed to the
	@@ -313,14 +313,15 @@
313	On Windows:
314	<pre>
315	win\buildmsvc.bat FOSSIL_DEBUG=1
316	</pre>
317
318	The resulting fossil binary can then be loaded into a platform-specific
319	debugger. Source files displayed in the debugger correspond to the ones
320	generated from the translation stage of the build process, as that is what
321	was actually compiled into the respective object files that make up the
322	fossil binary.
323
324	<h1>8.0 See Also</h1>
325
326	* [./tech_overview.wiki \| A Technical Overview Of Fossil]
327	* [./adding_code.wiki \| How To Add Features To Fossil]
328

M www/mdtest/test1.md

+1 -1

		--- www/mdtest/test1.md
		+++ www/mdtest/test1.md
		@@ -1,8 +1,8 @@
1	1	# Markdown Link-test
2	2
3		-This document exist solely as a test for some of the hyperlinking
	3	+This document exists solely as a test for some of the hyperlinking
4	4	capabilities of Markdown as implemented by Fossil.
5	5
6	6	## Relative-Path Links
7	7
8	8	* The index: [](../index.wiki)
9	9

	--- www/mdtest/test1.md
	+++ www/mdtest/test1.md
	@@ -1,8 +1,8 @@
1	# Markdown Link-test
2
3	This document exist solely as a test for some of the hyperlinking
4	capabilities of Markdown as implemented by Fossil.
5
6	## Relative-Path Links
7
8	* The index: [](../index.wiki)
9

	--- www/mdtest/test1.md
	+++ www/mdtest/test1.md
	@@ -1,8 +1,8 @@
1	# Markdown Link-test
2
3	This document exists solely as a test for some of the hyperlinking
4	capabilities of Markdown as implemented by Fossil.
5
6	## Relative-Path Links
7
8	* The index: [](../index.wiki)
9

M www/newrepo.wiki

+2 -1

		--- www/newrepo.wiki
		+++ www/newrepo.wiki
		@@ -57,11 +57,12 @@
57	57	$ fossil open ../demo.fossil
58	58	</verbatim>
59	59
60	60	That creates a file called <tt>_FOSSIL_</tt> in the current
61	61	directory, and this file contains all kinds of fossil-related
62		-information about your local repository. You can ignore it
	62	+information about your local repository. Under Linux, the BSDs or
	63	+macOS, this will instead be called <tt>.fslckout</tt>. You can ignore it
63	64	for all purposes, but be sure not to accidentally remove it
64	65	or otherwise damage it - it belongs to fossil, not you.
65	66
66	67	The next thing we need to do is add files to our repository. As it
67	68	happens, we have a few C source files lying around, which we'll
68	69

	--- www/newrepo.wiki
	+++ www/newrepo.wiki
	@@ -57,11 +57,12 @@
57	$ fossil open ../demo.fossil
58	</verbatim>
59
60	That creates a file called <tt>_FOSSIL_</tt> in the current
61	directory, and this file contains all kinds of fossil-related
62	information about your local repository. You can ignore it

63	for all purposes, but be sure not to accidentally remove it
64	or otherwise damage it - it belongs to fossil, not you.
65
66	The next thing we need to do is add files to our repository. As it
67	happens, we have a few C source files lying around, which we'll
68

	--- www/newrepo.wiki
	+++ www/newrepo.wiki
	@@ -57,11 +57,12 @@
57	$ fossil open ../demo.fossil
58	</verbatim>
59
60	That creates a file called <tt>_FOSSIL_</tt> in the current
61	directory, and this file contains all kinds of fossil-related
62	information about your local repository. Under Linux, the BSDs or
63	macOS, this will instead be called <tt>.fslckout</tt>. You can ignore it
64	for all purposes, but be sure not to accidentally remove it
65	or otherwise damage it - it belongs to fossil, not you.
66
67	The next thing we need to do is add files to our repository. As it
68	happens, we have a few C source files lying around, which we'll
69

M www/password.wiki

+6 -2

		--- www/password.wiki
		+++ www/password.wiki
		@@ -63,17 +63,21 @@
63	63	for "anonymous" uses one-time captchas not persistent passwords.
64	64
65	65	<h2>Web Interface Authentication</h2>
66	66
67	67	When a user logs into Fossil using the web interface, the login name
68		-and password are sent in the clear to the server. The server then
	68	+and password are sent in the clear to the server. For most modern fossil
	69	+server setups with [/help?cmd=redirect-to-https\|redirect-to-https] enabled,
	70	+this will be protected by the
	71	+SSL connection over HTTPS so it cannot be easily viewed. The server then
69	72	hashes the password and compares it against the value stored in USER.PW.
70	73	If they match, the server sets a cookie on the client to record the
71	74	login. This cookie contains a large amount of high-quality randomness
72	75	and is thus intractable to guess. The value of the cookie and the IP
73	76	address of the client is stored in the USER.COOKIE and USER.IPADDR fields
74	77	of the USER table on the server.
	78	+
75	79	The USER.CEXPIRE field holds an expiration date
76	80	for the cookie, encoded as a Julian day number. On all subsequent
77	81	HTTP requests, the cookie value is matched against the USER table to
78	82	enable access to the repository.
79	83
		@@ -123,11 +127,11 @@
123	127	</pre>
124	128
125	129	For older clients, the password is used for the shared secret as stated
126	130	in the URL and with no encoding.
127	131	For newer clients, the shared secret is derived from the password
128		-by transformed the password using the SHA1 hash encoding
	132	+by transforming the password using the SHA1 hash encoding
129	133	described above. However, if the first character of the password is
130	134	"" (ASCII 0x2a) then the "" is skipped and the rest of the password
131	135	is used directly as the share secret without the SHA1 encoding.
132	136
133	137	<pre>
134	138

	--- www/password.wiki
	+++ www/password.wiki
	@@ -63,17 +63,21 @@
63	for "anonymous" uses one-time captchas not persistent passwords.
64
65	<h2>Web Interface Authentication</h2>
66
67	When a user logs into Fossil using the web interface, the login name
68	and password are sent in the clear to the server. The server then



69	hashes the password and compares it against the value stored in USER.PW.
70	If they match, the server sets a cookie on the client to record the
71	login. This cookie contains a large amount of high-quality randomness
72	and is thus intractable to guess. The value of the cookie and the IP
73	address of the client is stored in the USER.COOKIE and USER.IPADDR fields
74	of the USER table on the server.

75	The USER.CEXPIRE field holds an expiration date
76	for the cookie, encoded as a Julian day number. On all subsequent
77	HTTP requests, the cookie value is matched against the USER table to
78	enable access to the repository.
79
	@@ -123,11 +127,11 @@
123	</pre>
124
125	For older clients, the password is used for the shared secret as stated
126	in the URL and with no encoding.
127	For newer clients, the shared secret is derived from the password
128	by transformed the password using the SHA1 hash encoding
129	described above. However, if the first character of the password is
130	"" (ASCII 0x2a) then the "" is skipped and the rest of the password
131	is used directly as the share secret without the SHA1 encoding.
132
133	<pre>
134

	--- www/password.wiki
	+++ www/password.wiki
	@@ -63,17 +63,21 @@
63	for "anonymous" uses one-time captchas not persistent passwords.
64
65	<h2>Web Interface Authentication</h2>
66
67	When a user logs into Fossil using the web interface, the login name
68	and password are sent in the clear to the server. For most modern fossil
69	server setups with [/help?cmd=redirect-to-https\|redirect-to-https] enabled,
70	this will be protected by the
71	SSL connection over HTTPS so it cannot be easily viewed. The server then
72	hashes the password and compares it against the value stored in USER.PW.
73	If they match, the server sets a cookie on the client to record the
74	login. This cookie contains a large amount of high-quality randomness
75	and is thus intractable to guess. The value of the cookie and the IP
76	address of the client is stored in the USER.COOKIE and USER.IPADDR fields
77	of the USER table on the server.
78
79	The USER.CEXPIRE field holds an expiration date
80	for the cookie, encoded as a Julian day number. On all subsequent
81	HTTP requests, the cookie value is matched against the USER table to
82	enable access to the repository.
83
	@@ -123,11 +127,11 @@
127	</pre>
128
129	For older clients, the password is used for the shared secret as stated
130	in the URL and with no encoding.
131	For newer clients, the shared secret is derived from the password
132	by transforming the password using the SHA1 hash encoding
133	described above. However, if the first character of the password is
134	"" (ASCII 0x2a) then the "" is skipped and the rest of the password
135	is used directly as the share secret without the SHA1 encoding.
136
137	<pre>
138

M www/patchcmd.md

+4 -4

		--- www/patchcmd.md
		+++ www/patchcmd.md
		@@ -9,11 +9,11 @@
9	9	"fossil patch push" command to make a copy of all your changes on the
10	10	remote Linux server:
11	11
12	12	fossil patch push linuxserver:/path/to/checkout
13	13
14		-In the previous "linuxserver" is the name of the remote machine and
	14	+In the previous line "linuxserver" is the name of the remote machine and
15	15	"/path/to/checkout" is an existing checkout directory for the same project
16	16	on the remote machine.
17	17
18	18	The "fossil patch push" command works by first creating a patch file,
19	19	then transfering that patch file to the remote machine using "ssh", then
		@@ -85,11 +85,11 @@
85	85	The "fossil patch create" command records all of the local, uncommitted
86	86	changes in an SQLite database file. If the argument to "fossil patch create"
87	87	is a filename, then the patch-file database is written into that file.
88	88	If the argument is "-" then the database is written on standard output.
89	89
90		-The "fossil patch apply" command reads the database that is the patch file
	90	+The "fossil patch apply" command reads the patch-file database
91	91	and applies it to the local check-out. If a filename is given as an
92	92	argument, then the database is read from that file. If the argument is "-"
93	93	then the database is read from standard input.
94	94
95	95	Hence the command:
		@@ -102,11 +102,11 @@
102	102
103	103	Likewise, a command like this:
104	104
105	105	fossil patch pull remote:projB
106	106
107		-Could be entered like this:
	107	+could be entered like this:
108	108
109	109	ssh -T remote 'cd projB;fossil patch create -' \| fossil patch apply -
110	110
111		-The "fossil patch view" command just opens the database file and prints
	111	+The "fossil patch view" command just opens the patch-file database and prints
112	112	a summary of its contents on standard output.
113	113

	--- www/patchcmd.md
	+++ www/patchcmd.md
	@@ -9,11 +9,11 @@
9	"fossil patch push" command to make a copy of all your changes on the
10	remote Linux server:
11
12	fossil patch push linuxserver:/path/to/checkout
13
14	In the previous "linuxserver" is the name of the remote machine and
15	"/path/to/checkout" is an existing checkout directory for the same project
16	on the remote machine.
17
18	The "fossil patch push" command works by first creating a patch file,
19	then transfering that patch file to the remote machine using "ssh", then
	@@ -85,11 +85,11 @@
85	The "fossil patch create" command records all of the local, uncommitted
86	changes in an SQLite database file. If the argument to "fossil patch create"
87	is a filename, then the patch-file database is written into that file.
88	If the argument is "-" then the database is written on standard output.
89
90	The "fossil patch apply" command reads the database that is the patch file
91	and applies it to the local check-out. If a filename is given as an
92	argument, then the database is read from that file. If the argument is "-"
93	then the database is read from standard input.
94
95	Hence the command:
	@@ -102,11 +102,11 @@
102
103	Likewise, a command like this:
104
105	fossil patch pull remote:projB
106
107	Could be entered like this:
108
109	ssh -T remote 'cd projB;fossil patch create -' \| fossil patch apply -
110
111	The "fossil patch view" command just opens the database file and prints
112	a summary of its contents on standard output.
113

	--- www/patchcmd.md
	+++ www/patchcmd.md
	@@ -9,11 +9,11 @@
9	"fossil patch push" command to make a copy of all your changes on the
10	remote Linux server:
11
12	fossil patch push linuxserver:/path/to/checkout
13
14	In the previous line "linuxserver" is the name of the remote machine and
15	"/path/to/checkout" is an existing checkout directory for the same project
16	on the remote machine.
17
18	The "fossil patch push" command works by first creating a patch file,
19	then transfering that patch file to the remote machine using "ssh", then
	@@ -85,11 +85,11 @@
85	The "fossil patch create" command records all of the local, uncommitted
86	changes in an SQLite database file. If the argument to "fossil patch create"
87	is a filename, then the patch-file database is written into that file.
88	If the argument is "-" then the database is written on standard output.
89
90	The "fossil patch apply" command reads the patch-file database
91	and applies it to the local check-out. If a filename is given as an
92	argument, then the database is read from that file. If the argument is "-"
93	then the database is read from standard input.
94
95	Hence the command:
	@@ -102,11 +102,11 @@
102
103	Likewise, a command like this:
104
105	fossil patch pull remote:projB
106
107	could be entered like this:
108
109	ssh -T remote 'cd projB;fossil patch create -' \| fossil patch apply -
110
111	The "fossil patch view" command just opens the patch-file database and prints
112	a summary of its contents on standard output.
113

M www/pikchr.md

+2 -2

		--- www/pikchr.md
		+++ www/pikchr.md
		@@ -1,9 +1,9 @@
1	1	# The Pikchr Diagram Language
2	2
3	3	Pikchr (pronounced "picture") is a [PIC][1]-like markup language for creating
4		-diagrams in technical documentation. Pikchr diagrams source text
	4	+diagrams in technical documentation. Source text for Pikchr diagrams
5	5	can be embedded directly in either [Markdown][2] or [Fossil Wiki][3].
6	6	Fossil translates the Pikchr source text into SVG which is displayed as
7	7	part of the rendered wiki.
8	8
9	9	[1]: wikipedia:/wiki/Pic_language
		@@ -134,6 +134,6 @@
134	134
135	135	* float-right → The diagram is shown at the right margin and
136	136	text fills in around the diagram.
137	137
138	138	* source → The display starts out showing the Pikchr source text.
139		- The reader must click (or Alt-click or Ctrl-click) to set the diagram.
	139	+ The reader must click (or Alt-click or Ctrl-click) to show the diagram.
140	140

	--- www/pikchr.md
	+++ www/pikchr.md
	@@ -1,9 +1,9 @@
1	# The Pikchr Diagram Language
2
3	Pikchr (pronounced "picture") is a [PIC][1]-like markup language for creating
4	diagrams in technical documentation. Pikchr diagrams source text
5	can be embedded directly in either [Markdown][2] or [Fossil Wiki][3].
6	Fossil translates the Pikchr source text into SVG which is displayed as
7	part of the rendered wiki.
8
9	[1]: wikipedia:/wiki/Pic_language
	@@ -134,6 +134,6 @@
134
135	* float-right → The diagram is shown at the right margin and
136	text fills in around the diagram.
137
138	* source → The display starts out showing the Pikchr source text.
139	The reader must click (or Alt-click or Ctrl-click) to set the diagram.
140

	--- www/pikchr.md
	+++ www/pikchr.md
	@@ -1,9 +1,9 @@
1	# The Pikchr Diagram Language
2
3	Pikchr (pronounced "picture") is a [PIC][1]-like markup language for creating
4	diagrams in technical documentation. Source text for Pikchr diagrams
5	can be embedded directly in either [Markdown][2] or [Fossil Wiki][3].
6	Fossil translates the Pikchr source text into SVG which is displayed as
7	part of the rendered wiki.
8
9	[1]: wikipedia:/wiki/Pic_language
	@@ -134,6 +134,6 @@
134
135	* float-right → The diagram is shown at the right margin and
136	text fills in around the diagram.
137
138	* source → The display starts out showing the Pikchr source text.
139	The reader must click (or Alt-click or Ctrl-click) to show the diagram.
140

M www/qandc.wiki

+1 -1

		--- www/qandc.wiki
		+++ www/qandc.wiki
		@@ -142,11 +142,11 @@
142	142	fossil supports disconnected operation.
143	143
144	144	As for bloat: Fossil is a single self-contained executable.
145	145	You do not need any other packages
146	146	(diff, patch, merge, cvs, svn, rcs, git, python, perl, tcl, apache,
147		-sqlite, and so forth)
	147	+SQLite, and so forth)
148	148	in order to run fossil. Fossil runs just fine in a chroot jail all
149	149	by itself. And the self-contained fossil
150	150	executable is much less than 1MB in size. (Update 2015-01-12: Fossil has
151	151	grown in the years since the previous sentence was written but is still
152	152	much less than 2MB according to "size" when compiled using -Os on x64 Linux.)
153	153

	--- www/qandc.wiki
	+++ www/qandc.wiki
	@@ -142,11 +142,11 @@
142	fossil supports disconnected operation.
143
144	As for bloat: Fossil is a single self-contained executable.
145	You do not need any other packages
146	(diff, patch, merge, cvs, svn, rcs, git, python, perl, tcl, apache,
147	sqlite, and so forth)
148	in order to run fossil. Fossil runs just fine in a chroot jail all
149	by itself. And the self-contained fossil
150	executable is much less than 1MB in size. (Update 2015-01-12: Fossil has
151	grown in the years since the previous sentence was written but is still
152	much less than 2MB according to "size" when compiled using -Os on x64 Linux.)
153

	--- www/qandc.wiki
	+++ www/qandc.wiki
	@@ -142,11 +142,11 @@
142	fossil supports disconnected operation.
143
144	As for bloat: Fossil is a single self-contained executable.
145	You do not need any other packages
146	(diff, patch, merge, cvs, svn, rcs, git, python, perl, tcl, apache,
147	SQLite, and so forth)
148	in order to run fossil. Fossil runs just fine in a chroot jail all
149	by itself. And the self-contained fossil
150	executable is much less than 1MB in size. (Update 2015-01-12: Fossil has
151	grown in the years since the previous sentence was written but is still
152	much less than 2MB according to "size" when compiled using -Os on x64 Linux.)
153

M www/quickstart.wiki

+1 -1

		--- www/quickstart.wiki
		+++ www/quickstart.wiki
		@@ -364,11 +364,11 @@
364	364	abbreviation to the 40-character
365	365	artifact identifier for a particular check-in, or it can be a
366	366	date/time stamp. ([./checkin_names.wiki \| more info])
367	367	If you omit
368	368	the <i>VERSION</i>, then fossil moves you to the
369		-latest version of the branch your are currently on.
	369	+latest version of the branch you are currently on.
370	370
371	371	The default behavior is for [./concepts.wiki#workflow\|autosync] to
372	372	be turned on. That means that a [/help/pull\|pull] automatically occurs
373	373	when you run [/help/update\|update] and a [/help/push\|push] happens
374	374	automatically after you [/help/commit\|commit]. So in normal practice,
375	375

	--- www/quickstart.wiki
	+++ www/quickstart.wiki
	@@ -364,11 +364,11 @@
364	abbreviation to the 40-character
365	artifact identifier for a particular check-in, or it can be a
366	date/time stamp. ([./checkin_names.wiki \| more info])
367	If you omit
368	the <i>VERSION</i>, then fossil moves you to the
369	latest version of the branch your are currently on.
370
371	The default behavior is for [./concepts.wiki#workflow\|autosync] to
372	be turned on. That means that a [/help/pull\|pull] automatically occurs
373	when you run [/help/update\|update] and a [/help/push\|push] happens
374	automatically after you [/help/commit\|commit]. So in normal practice,
375

	--- www/quickstart.wiki
	+++ www/quickstart.wiki
	@@ -364,11 +364,11 @@
364	abbreviation to the 40-character
365	artifact identifier for a particular check-in, or it can be a
366	date/time stamp. ([./checkin_names.wiki \| more info])
367	If you omit
368	the <i>VERSION</i>, then fossil moves you to the
369	latest version of the branch you are currently on.
370
371	The default behavior is for [./concepts.wiki#workflow\|autosync] to
372	be turned on. That means that a [/help/pull\|pull] automatically occurs
373	when you run [/help/update\|update] and a [/help/push\|push] happens
374	automatically after you [/help/commit\|commit]. So in normal practice,
375

M www/rebaseharm.md

+2 -2

		--- www/rebaseharm.md
		+++ www/rebaseharm.md
		@@ -224,16 +224,16 @@
224	224	you are keeping private branches. Or, to put it another way, you are
225	225	doing siloed development. You are not sharing your intermediate work
226	226	with collaborators. This is not good for product quality.
227	227
228	228	[Nagappan, et. al][nagappan] studied bugs in Windows Vista and found
229		-that best predictor of bugs is the distance on the org-chart between
	229	+that the best predictor of bugs is the distance on the org-chart between
230	230	the stake-holders. The bug rate is inversely related to the
231	231	amount of communication among the engineers.
232	232	Similar findings arise in other disciplines. Keeping
233	233	private branches does not prove that developers are communicating
234		-insufficiently, but it is a key symptom that problem.
	234	+insufficiently, but it is a key symptom of that problem.
235	235
236	236	[Weinberg][weinberg] argues programming should be "egoless." That
237	237	is to say, programmers should avoid linking their code with their sense of
238	238	self, as that makes it more difficult for them to find and respond
239	239	to bugs, and hence makes them less productive. Many developers are
240	240

	--- www/rebaseharm.md
	+++ www/rebaseharm.md
	@@ -224,16 +224,16 @@
224	you are keeping private branches. Or, to put it another way, you are
225	doing siloed development. You are not sharing your intermediate work
226	with collaborators. This is not good for product quality.
227
228	[Nagappan, et. al][nagappan] studied bugs in Windows Vista and found
229	that best predictor of bugs is the distance on the org-chart between
230	the stake-holders. The bug rate is inversely related to the
231	amount of communication among the engineers.
232	Similar findings arise in other disciplines. Keeping
233	private branches does not prove that developers are communicating
234	insufficiently, but it is a key symptom that problem.
235
236	[Weinberg][weinberg] argues programming should be "egoless." That
237	is to say, programmers should avoid linking their code with their sense of
238	self, as that makes it more difficult for them to find and respond
239	to bugs, and hence makes them less productive. Many developers are
240

	--- www/rebaseharm.md
	+++ www/rebaseharm.md
	@@ -224,16 +224,16 @@
224	you are keeping private branches. Or, to put it another way, you are
225	doing siloed development. You are not sharing your intermediate work
226	with collaborators. This is not good for product quality.
227
228	[Nagappan, et. al][nagappan] studied bugs in Windows Vista and found
229	that the best predictor of bugs is the distance on the org-chart between
230	the stake-holders. The bug rate is inversely related to the
231	amount of communication among the engineers.
232	Similar findings arise in other disciplines. Keeping
233	private branches does not prove that developers are communicating
234	insufficiently, but it is a key symptom of that problem.
235
236	[Weinberg][weinberg] argues programming should be "egoless." That
237	is to say, programmers should avoid linking their code with their sense of
238	self, as that makes it more difficult for them to find and respond
239	to bugs, and hence makes them less productive. Many developers are
240

M www/selfcheck.wiki

+1 -1

		--- www/selfcheck.wiki
		+++ www/selfcheck.wiki
		@@ -13,11 +13,11 @@
13	13
14	14	<h2>Atomic Check-ins With Rollback</h2>
15	15
16	16	The Fossil repository is stored in an
17	17	<a href="http://www.sqlite.org/">SQLite</a> database file.
18		-([./tech_overview.wiki \| Addition information] about the repository
	18	+([./tech_overview.wiki \| Additional information] about the repository
19	19	file format.)
20	20	SQLite is very mature and stable and has been in wide-spread use for many
21	21	years, so we are confident it will not cause repository
22	22	corruption. SQLite
23	23	databases do not corrupt even if a program or system crash or power
24	24

	--- www/selfcheck.wiki
	+++ www/selfcheck.wiki
	@@ -13,11 +13,11 @@
13
14	<h2>Atomic Check-ins With Rollback</h2>
15
16	The Fossil repository is stored in an
17	<a href="http://www.sqlite.org/">SQLite</a> database file.
18	([./tech_overview.wiki \| Addition information] about the repository
19	file format.)
20	SQLite is very mature and stable and has been in wide-spread use for many
21	years, so we are confident it will not cause repository
22	corruption. SQLite
23	databases do not corrupt even if a program or system crash or power
24

	--- www/selfcheck.wiki
	+++ www/selfcheck.wiki
	@@ -13,11 +13,11 @@
13
14	<h2>Atomic Check-ins With Rollback</h2>
15
16	The Fossil repository is stored in an
17	<a href="http://www.sqlite.org/">SQLite</a> database file.
18	([./tech_overview.wiki \| Additional information] about the repository
19	file format.)
20	SQLite is very mature and stable and has been in wide-spread use for many
21	years, so we are confident it will not cause repository
22	corruption. SQLite
23	databases do not corrupt even if a program or system crash or power
24

M www/selfhost.wiki

+1 -1

		--- www/selfhost.wiki
		+++ www/selfhost.wiki
		@@ -27,11 +27,11 @@
27	27	hosts <a href="http://www.sqlite.org/">SQLite</a> and over a
28	28	dozen other smaller projects. This demonstrates that Fossil can run on
29	29	a low-power host processor.
30	30	Multiple fossil-based projects can easily be hosted on the same machine,
31	31	even if that machine is itself one of several dozen virtual machines on
32		-single physical box. The CGI script that runs the canonical Fossil
	32	+a single physical box. The CGI script that runs the canonical Fossil
33	33	self-hosting repository is as follows:
34	34
35	35	<pre>
36	36	#!/usr/bin/fossil
37	37	repository: /fossil/fossil.fossil
38	38

	--- www/selfhost.wiki
	+++ www/selfhost.wiki
	@@ -27,11 +27,11 @@
27	hosts <a href="http://www.sqlite.org/">SQLite</a> and over a
28	dozen other smaller projects. This demonstrates that Fossil can run on
29	a low-power host processor.
30	Multiple fossil-based projects can easily be hosted on the same machine,
31	even if that machine is itself one of several dozen virtual machines on
32	single physical box. The CGI script that runs the canonical Fossil
33	self-hosting repository is as follows:
34
35	<pre>
36	#!/usr/bin/fossil
37	repository: /fossil/fossil.fossil
38

	--- www/selfhost.wiki
	+++ www/selfhost.wiki
	@@ -27,11 +27,11 @@
27	hosts <a href="http://www.sqlite.org/">SQLite</a> and over a
28	dozen other smaller projects. This demonstrates that Fossil can run on
29	a low-power host processor.
30	Multiple fossil-based projects can easily be hosted on the same machine,
31	even if that machine is itself one of several dozen virtual machines on
32	a single physical box. The CGI script that runs the canonical Fossil
33	self-hosting repository is as follows:
34
35	<pre>
36	#!/usr/bin/fossil
37	repository: /fossil/fossil.fossil
38

M www/server/debian/service.md

+1 -1

		--- www/server/debian/service.md
		+++ www/server/debian/service.md
		@@ -183,11 +183,11 @@
183	183	to the system level. We need to start this socket listener at the root
184	184	level because of the low-numbered TCP port restriction we brought up
185	185	above.
186	186
187	187	This configuration says more or less the same thing as the socket part
188		-of an `inted` entry [exemplified elsewhere in this
	188	+of an `inetd` entry [exemplified elsewhere in this
189	189	documentation](../any/inetd.md).
190	190
191	191	Next, create the service definition file in that same directory as
192	192	`[email protected]`:
193	193
194	194

	--- www/server/debian/service.md
	+++ www/server/debian/service.md
	@@ -183,11 +183,11 @@
183	to the system level. We need to start this socket listener at the root
184	level because of the low-numbered TCP port restriction we brought up
185	above.
186
187	This configuration says more or less the same thing as the socket part
188	of an `inted` entry [exemplified elsewhere in this
189	documentation](../any/inetd.md).
190
191	Next, create the service definition file in that same directory as
192	`[email protected]`:
193
194

	--- www/server/debian/service.md
	+++ www/server/debian/service.md
	@@ -183,11 +183,11 @@
183	to the system level. We need to start this socket listener at the root
184	level because of the low-numbered TCP port restriction we brought up
185	above.
186
187	This configuration says more or less the same thing as the socket part
188	of an `inetd` entry [exemplified elsewhere in this
189	documentation](../any/inetd.md).
190
191	Next, create the service definition file in that same directory as
192	`[email protected]`:
193
194

M www/server/openbsd/fastcgi.md

+1 -1

		--- www/server/openbsd/fastcgi.md
		+++ www/server/openbsd/fastcgi.md
		@@ -172,11 +172,11 @@
172	172	setting, raising the limit to 100 MiB.
173	173
174	174	[dlim]: https://man.openbsd.org/httpd.conf.5#connection
175	175	[uv]: ../../unvers.wiki
176	176
177		-NOTE: If not already in possession of a HTTPS certificate, comment
	177	+NOTE: If not already in possession of an HTTPS certificate, comment
178	178	out the `https` server block and proceed to securing a free
179	179	[Let's Encrypt Certificate](#letsencrypt); otherwise skip to
180	180	[Start `httpd`](#starthttpd).
181	181
182	182
183	183

	--- www/server/openbsd/fastcgi.md
	+++ www/server/openbsd/fastcgi.md
	@@ -172,11 +172,11 @@
172	setting, raising the limit to 100 MiB.
173
174	[dlim]: https://man.openbsd.org/httpd.conf.5#connection
175	[uv]: ../../unvers.wiki
176
177	NOTE: If not already in possession of a HTTPS certificate, comment
178	out the `https` server block and proceed to securing a free
179	[Let's Encrypt Certificate](#letsencrypt); otherwise skip to
180	[Start `httpd`](#starthttpd).
181
182
183

	--- www/server/openbsd/fastcgi.md
	+++ www/server/openbsd/fastcgi.md
	@@ -172,11 +172,11 @@
172	setting, raising the limit to 100 MiB.
173
174	[dlim]: https://man.openbsd.org/httpd.conf.5#connection
175	[uv]: ../../unvers.wiki
176
177	NOTE: If not already in possession of an HTTPS certificate, comment
178	out the `https` server block and proceed to securing a free
179	[Let's Encrypt Certificate](#letsencrypt); otherwise skip to
180	[Start `httpd`](#starthttpd).
181
182
183

M www/server/windows/service.md

+1 -1

		--- www/server/windows/service.md
		+++ www/server/windows/service.md
		@@ -39,11 +39,11 @@
39	39	This will create a windows service named 'Fossil-DSCM' running under the local
40	40	system account and accessible on port 8080 by default. `fossil winsrv` can also
41	41	start, stop, and delete the service. For all available options, please execute
42	42	`fossil help winsrv` on a windows install of Fossil.
43	43
44		-If you wish to server a directory of repositories, the `fossil winsrv` command
	44	+If you wish to serve a directory of repositories, the `fossil winsrv` command
45	45	requires a slightly different set of options vs. `fossil server`:
46	46
47	47	```
48	48	fossil winsrv create --repository D:/Path/to/Repos --repolist
49	49	```
50	50

	--- www/server/windows/service.md
	+++ www/server/windows/service.md
	@@ -39,11 +39,11 @@
39	This will create a windows service named 'Fossil-DSCM' running under the local
40	system account and accessible on port 8080 by default. `fossil winsrv` can also
41	start, stop, and delete the service. For all available options, please execute
42	`fossil help winsrv` on a windows install of Fossil.
43
44	If you wish to server a directory of repositories, the `fossil winsrv` command
45	requires a slightly different set of options vs. `fossil server`:
46
47	```
48	fossil winsrv create --repository D:/Path/to/Repos --repolist
49	```
50

	--- www/server/windows/service.md
	+++ www/server/windows/service.md
	@@ -39,11 +39,11 @@
39	This will create a windows service named 'Fossil-DSCM' running under the local
40	system account and accessible on port 8080 by default. `fossil winsrv` can also
41	start, stop, and delete the service. For all available options, please execute
42	`fossil help winsrv` on a windows install of Fossil.
43
44	If you wish to serve a directory of repositories, the `fossil winsrv` command
45	requires a slightly different set of options vs. `fossil server`:
46
47	```
48	fossil winsrv create --repository D:/Path/to/Repos --repolist
49	```
50

M www/settings.wiki

+2 -3

		--- www/settings.wiki
		+++ www/settings.wiki
		@@ -44,13 +44,12 @@
44	44	Because these options can change over time, and the inconvenience of
45	45	replicating changes, these settings are "versionable". As well as being
46	46	able to be set using the <tt>settings</tt> command or the web interface,
47	47	you can create versioned files in the <tt>.fossil-settings</tt>
48	48	subdirectory of the check-out root, named with the setting name.
49		-The contents of the file is the
50		-value of the setting, and these files are checked in, committed, merged,
51		-and so on, as with any other file.
	49	+Each file holds the value of a setting, and these files are checked in,
	50	+committed, merged, and so on, as with any other file.
52	51
53	52	Where a setting is a list of values, such as <tt>ignore-glob</tt>, you
54	53	can use a newline as a separator as well as a comma.
55	54
56	55	For example, to set the list of ignored files, create a
57	56

	--- www/settings.wiki
	+++ www/settings.wiki
	@@ -44,13 +44,12 @@
44	Because these options can change over time, and the inconvenience of
45	replicating changes, these settings are "versionable". As well as being
46	able to be set using the <tt>settings</tt> command or the web interface,
47	you can create versioned files in the <tt>.fossil-settings</tt>
48	subdirectory of the check-out root, named with the setting name.
49	The contents of the file is the
50	value of the setting, and these files are checked in, committed, merged,
51	and so on, as with any other file.
52
53	Where a setting is a list of values, such as <tt>ignore-glob</tt>, you
54	can use a newline as a separator as well as a comma.
55
56	For example, to set the list of ignored files, create a
57

	--- www/settings.wiki
	+++ www/settings.wiki
	@@ -44,13 +44,12 @@
44	Because these options can change over time, and the inconvenience of
45	replicating changes, these settings are "versionable". As well as being
46	able to be set using the <tt>settings</tt> command or the web interface,
47	you can create versioned files in the <tt>.fossil-settings</tt>
48	subdirectory of the check-out root, named with the setting name.
49	Each file holds the value of a setting, and these files are checked in,
50	committed, merged, and so on, as with any other file.

51
52	Where a setting is a list of values, such as <tt>ignore-glob</tt>, you
53	can use a newline as a separator as well as a comma.
54
55	For example, to set the list of ignored files, create a
56

M www/shunning.wiki

+2 -2

		--- www/shunning.wiki
		+++ www/shunning.wiki
		@@ -33,11 +33,11 @@
33	33
34	34	All of these are rare cases: Fossil is [./antibot.wiki \| designed to
35	35	foil spammers up front], legally problematic check-ins should range from
36	36	rare to nonexistent, and you have to go way out of your way to force
37	37	Fossil to insert bad control artifacts. Therefore, before we get to
38		-methods of permanently deleting content from a Fossil repos, let's give
	38	+methods of permanently deleting content from a Fossil repo, let's give
39	39	some alternatives that usually suffice, which don't damage the project's
40	40	fossil record:
41	41
42	42	* When a forum post or wiki article is "deleted," what actually
43	43	happens is that a new empty version is added to the Fossil repository.
		@@ -133,11 +133,11 @@
133	133	using the "configuration" command:
134	134
135	135	<b>fossil configuration pull shun</b> <i>remote-url</i><br>
136	136	<b>fossil configuration push shun</b> <i>remote-url</i>
137	137
138		-The two command above will pull or push shunning lists from or to
	138	+The two commands above will pull or push shunning lists from or to
139	139	the <i>remote-url</i> indicated and merge the lists on the receiving
140	140	end. "Admin" privilege on the remote server is required in order to
141	141	push a shun list. In contrast, the shunning list will be automatically
142	142	received by default as part of a normal client "pull" operation unless
143	143	disabled by the "<tt>auto-shun</tt>" setting.
144	144

	--- www/shunning.wiki
	+++ www/shunning.wiki
	@@ -33,11 +33,11 @@
33
34	All of these are rare cases: Fossil is [./antibot.wiki \| designed to
35	foil spammers up front], legally problematic check-ins should range from
36	rare to nonexistent, and you have to go way out of your way to force
37	Fossil to insert bad control artifacts. Therefore, before we get to
38	methods of permanently deleting content from a Fossil repos, let's give
39	some alternatives that usually suffice, which don't damage the project's
40	fossil record:
41
42	* When a forum post or wiki article is "deleted," what actually
43	happens is that a new empty version is added to the Fossil repository.
	@@ -133,11 +133,11 @@
133	using the "configuration" command:
134
135	<b>fossil configuration pull shun</b> <i>remote-url</i><br>
136	<b>fossil configuration push shun</b> <i>remote-url</i>
137
138	The two command above will pull or push shunning lists from or to
139	the <i>remote-url</i> indicated and merge the lists on the receiving
140	end. "Admin" privilege on the remote server is required in order to
141	push a shun list. In contrast, the shunning list will be automatically
142	received by default as part of a normal client "pull" operation unless
143	disabled by the "<tt>auto-shun</tt>" setting.
144

	--- www/shunning.wiki
	+++ www/shunning.wiki
	@@ -33,11 +33,11 @@
33
34	All of these are rare cases: Fossil is [./antibot.wiki \| designed to
35	foil spammers up front], legally problematic check-ins should range from
36	rare to nonexistent, and you have to go way out of your way to force
37	Fossil to insert bad control artifacts. Therefore, before we get to
38	methods of permanently deleting content from a Fossil repo, let's give
39	some alternatives that usually suffice, which don't damage the project's
40	fossil record:
41
42	* When a forum post or wiki article is "deleted," what actually
43	happens is that a new empty version is added to the Fossil repository.
	@@ -133,11 +133,11 @@
133	using the "configuration" command:
134
135	<b>fossil configuration pull shun</b> <i>remote-url</i><br>
136	<b>fossil configuration push shun</b> <i>remote-url</i>
137
138	The two commands above will pull or push shunning lists from or to
139	the <i>remote-url</i> indicated and merge the lists on the receiving
140	end. "Admin" privilege on the remote server is required in order to
141	push a shun list. In contrast, the shunning list will be automatically
142	received by default as part of a normal client "pull" operation unless
143	disabled by the "<tt>auto-shun</tt>" setting.
144

M www/stats.wiki

+2 -2

		--- www/stats.wiki
		+++ www/stats.wiki
		@@ -1,11 +1,11 @@
1	1	<title>Fossil Performance</title>
2	2
3	3	The questions will inevitably arise: How does Fossil perform?
4	4	Does it use a lot of disk space or bandwidth? Is it scalable?
5	5
6		-In an attempt to answers these questions, this report looks at several
	6	+In an attempt to answer these questions, this report looks at several
7	7	projects that use fossil for configuration management and examines how
8	8	well they are working. The following table is a summary of the results.
9	9	(Last updated on 2018-06-04.)
10	10	Explanation and analysis follows the table.
11	11
		@@ -100,11 +100,11 @@
100	100	is the unordered collection of artifacts. In fact, one of the key
101	101	characteristics of Fossil is that the entire project history can be
102	102	reconstructed simply by scanning the artifacts in an arbitrary order.
103	103
104	104	The number of check-ins is the number of times that the "commit" command
105		-has been run. A single check-in might change a 3 or 4 files, or it might
	105	+has been run. A single check-in might change 3 or 4 files, or it might
106	106	change dozens or hundreds of files. Regardless of the number of files
107	107	changed, it still only counts as one check-in.
108	108
109	109	The "Uncompressed Size" is the total size of all the artifacts within
110	110	the repository assuming they were all uncompressed and stored
111	111

	--- www/stats.wiki
	+++ www/stats.wiki
	@@ -1,11 +1,11 @@
1	<title>Fossil Performance</title>
2
3	The questions will inevitably arise: How does Fossil perform?
4	Does it use a lot of disk space or bandwidth? Is it scalable?
5
6	In an attempt to answers these questions, this report looks at several
7	projects that use fossil for configuration management and examines how
8	well they are working. The following table is a summary of the results.
9	(Last updated on 2018-06-04.)
10	Explanation and analysis follows the table.
11
	@@ -100,11 +100,11 @@
100	is the unordered collection of artifacts. In fact, one of the key
101	characteristics of Fossil is that the entire project history can be
102	reconstructed simply by scanning the artifacts in an arbitrary order.
103
104	The number of check-ins is the number of times that the "commit" command
105	has been run. A single check-in might change a 3 or 4 files, or it might
106	change dozens or hundreds of files. Regardless of the number of files
107	changed, it still only counts as one check-in.
108
109	The "Uncompressed Size" is the total size of all the artifacts within
110	the repository assuming they were all uncompressed and stored
111

	--- www/stats.wiki
	+++ www/stats.wiki
	@@ -1,11 +1,11 @@
1	<title>Fossil Performance</title>
2
3	The questions will inevitably arise: How does Fossil perform?
4	Does it use a lot of disk space or bandwidth? Is it scalable?
5
6	In an attempt to answer these questions, this report looks at several
7	projects that use fossil for configuration management and examines how
8	well they are working. The following table is a summary of the results.
9	(Last updated on 2018-06-04.)
10	Explanation and analysis follows the table.
11
	@@ -100,11 +100,11 @@
100	is the unordered collection of artifacts. In fact, one of the key
101	characteristics of Fossil is that the entire project history can be
102	reconstructed simply by scanning the artifacts in an arbitrary order.
103
104	The number of check-ins is the number of times that the "commit" command
105	has been run. A single check-in might change 3 or 4 files, or it might
106	change dozens or hundreds of files. Regardless of the number of files
107	changed, it still only counts as one check-in.
108
109	The "Uncompressed Size" is the total size of all the artifacts within
110	the repository assuming they were all uncompressed and stored
111

M www/style.wiki

+6 -5

		--- www/style.wiki
		+++ www/style.wiki
		@@ -1,8 +1,8 @@
1	1	<title>Coding Style</title>
2	2
3		-Fossil source code should following the style guidelines below.
	3	+Fossil source code should follow the style guidelines below.
4	4
5	5	<em> The Fossil source tree includes a few files taken from external
6	6	sources
7	7	(examples: [https://github.com/antirez/linenoise\|linenoise] and
8	8	[http://zlib.net/\|zLib])
		@@ -38,13 +38,14 @@
38	38
39	39	<li> -Wno-long-long: Fossil uses the 'long long' integer type, which is not strictly ANSI C-89 (defined in C99).
40	40	The use of 'long long' resolves many problems with 64-bit arithmetics, especially on 32-bit machines.
41	41	(http_ssl.c, sha3.c, shell.c, util.c)
42	42
43		- <li> alloca(): By default, sqlite3.c is compiled with the -DSQLITE_USE_ALLOCA flag to use the alloca() function.
44		- alloca() is not considered ANSI C, and normally not recommended due to portability issues, but
45		- performance and/or memory consumption improvement may be a stronger argument in favor of its usage.
	43	+ <li> alloca(): By default, sqlite3.c was compiled with the -DSQLITE_USE_ALLOCA flag to use the alloca() function.
	44	+ This is no longer the case as of 20220119. alloca() is not considered ANSI C, and normally not
	45	+ recommended due to portability issues, but performance and/or memory consumption
	46	+ improvement may have been a stronger argument in favor of its usage.
46	47	(sqlite3.c)
47	48	</ol>
48	49
49	50	<li> All comments and identifiers are in English.
50	51
		@@ -75,11 +76,11 @@
75	76
76	77	<ol>
77	78	<li value=30> Every function has a header comment describing the purpose and use
78	79	of the function.
79	80
80		- <li> Function header comment defines the behavior of the function in
	81	+ <li> A function header comment defines the behavior of the function in
81	82	sufficient detail to allow the function to be re-implemented from
82	83	scratch without reference to the original code.
83	84
84	85	<li> Functions that perform dynamic memory allocation (either directly
85	86	or indirectly via subfunctions) say so in their header comments.
86	87

	--- www/style.wiki
	+++ www/style.wiki
	@@ -1,8 +1,8 @@
1	<title>Coding Style</title>
2
3	Fossil source code should following the style guidelines below.
4
5	<em> The Fossil source tree includes a few files taken from external
6	sources
7	(examples: [https://github.com/antirez/linenoise\|linenoise] and
8	[http://zlib.net/\|zLib])
	@@ -38,13 +38,14 @@
38
39	<li> -Wno-long-long: Fossil uses the 'long long' integer type, which is not strictly ANSI C-89 (defined in C99).
40	The use of 'long long' resolves many problems with 64-bit arithmetics, especially on 32-bit machines.
41	(http_ssl.c, sha3.c, shell.c, util.c)
42
43	<li> alloca(): By default, sqlite3.c is compiled with the -DSQLITE_USE_ALLOCA flag to use the alloca() function.
44	alloca() is not considered ANSI C, and normally not recommended due to portability issues, but
45	performance and/or memory consumption improvement may be a stronger argument in favor of its usage.

46	(sqlite3.c)
47	</ol>
48
49	<li> All comments and identifiers are in English.
50
	@@ -75,11 +76,11 @@
75
76	<ol>
77	<li value=30> Every function has a header comment describing the purpose and use
78	of the function.
79
80	<li> Function header comment defines the behavior of the function in
81	sufficient detail to allow the function to be re-implemented from
82	scratch without reference to the original code.
83
84	<li> Functions that perform dynamic memory allocation (either directly
85	or indirectly via subfunctions) say so in their header comments.
86

	--- www/style.wiki
	+++ www/style.wiki
	@@ -1,8 +1,8 @@
1	<title>Coding Style</title>
2
3	Fossil source code should follow the style guidelines below.
4
5	<em> The Fossil source tree includes a few files taken from external
6	sources
7	(examples: [https://github.com/antirez/linenoise\|linenoise] and
8	[http://zlib.net/\|zLib])
	@@ -38,13 +38,14 @@
38
39	<li> -Wno-long-long: Fossil uses the 'long long' integer type, which is not strictly ANSI C-89 (defined in C99).
40	The use of 'long long' resolves many problems with 64-bit arithmetics, especially on 32-bit machines.
41	(http_ssl.c, sha3.c, shell.c, util.c)
42
43	<li> alloca(): By default, sqlite3.c was compiled with the -DSQLITE_USE_ALLOCA flag to use the alloca() function.
44	This is no longer the case as of 20220119. alloca() is not considered ANSI C, and normally not
45	recommended due to portability issues, but performance and/or memory consumption
46	improvement may have been a stronger argument in favor of its usage.
47	(sqlite3.c)
48	</ol>
49
50	<li> All comments and identifiers are in English.
51
	@@ -75,11 +76,11 @@
76
77	<ol>
78	<li value=30> Every function has a header comment describing the purpose and use
79	of the function.
80
81	<li> A function header comment defines the behavior of the function in
82	sufficient detail to allow the function to be re-implemented from
83	scratch without reference to the original code.
84
85	<li> Functions that perform dynamic memory allocation (either directly
86	or indirectly via subfunctions) say so in their header comments.
87

M www/sync.wiki

+33 -13

		--- www/sync.wiki
		+++ www/sync.wiki
		@@ -518,11 +518,11 @@
518	518	pronounced as if it were a single word "gimme" in some dialects of
519	519	English (including the dialect spoken by the original author of Fossil).
520	520
521	521	<h4>3.7.1 Unversioned Gimme Cards</h4>
522	522
523		-Sync synchronizing unversioned content, the client may send "uvgimme"
	523	+When synchronizing unversioned content, the client may send "uvgimme"
524	524	cards to the server. A uvgimme card requests that the server send
525	525	unversioned content to the client. The format of a uvgimme card is
526	526	as follows:
527	527
528	528	<pre>
		@@ -570,77 +570,97 @@
570	570
571	571	<pre>
572	572	<b>reqconfig</b> <i>configuration-name</i>
573	573	</pre>
574	574
575		-As of 2018-06-04, the configuration-name must be one of the
	575	+As of 2024-10-22, the configuration-name must be one of the
576	576	following values:
577	577
578	578	<table border=0 align="center">
579	579	<tr><td valign="top">
580	580	<ul>
581	581	<li> css
582	582	<li> header
	583	+<li> mainmenu
583	584	<li> footer
584	585	<li> details
	586	+<li> js
	587	+<li> default-skin
585	588	<li> logo-mimetype
586	589	<li> logo-image
587	590	<li> background-mimetype
588	591	<li> background-image
589		-<li> index-page
	592	+<li> icon-mimetype
	593	+<li> icon-image
590	594	<li> timeline-block-markup
	595	+<li> timeline-date-format
	596	+<li> timeline-default-style
	597	+<ul></td><td valign="top"><ul>
	598	+<li> timeline-dwelltime
	599	+<li> timeline-closetime
	600	+<li> timeline-hard-newlines
591	601	<li> timeline-max-comment
592	602	<li> timeline-plaintext
	603	+<li> timeline-truncate-at-blank
	604	+<li> timeline-tslink-info
	605	+<li> timeline-utc
593	606	<li> adunit
594	607	<li> adunit-omit-if-admin
595	608	<li> adunit-omit-if-user
596		-<ul></td><td valign="top"><ul>
597		-<li> th1-docs
	609	+<li> default-csp
	610	+<li> sitemap-extra
	611	+<li> safe-html
598	612	<li> th1-hooks
599		-<li> th1-setup
600		-<li> tcl
601		-<li> tcl-setup
	613	+<li> th1-uri-regexp
602	614	<li> project-name
	615	+<ul></td><td valign="top"><ul>
603	616	<li> short-project-name
604	617	<li> project-description
605	618	<li> index-page
606	619	<li> manifest
607	620	<li> binary-glob
608	621	<li> clean-glob
609	622	<li> ignore-glob
610	623	<li> keep-glob
611	624	<li> crlf-glob
612		-<ul></td><td valign="top"><ul>
613	625	<li> crnl-glob
614	626	<li> encoding-glob
615	627	<li> empty-dirs
616	628	<li> <s title="removed 2020-08, version 2.12.1">allow-symlinks</s>
617	629	<li> dotfiles
618	630	<li> parent-project-code
619		-<li> parent-projet-name
	631	+<li> parent-project-name
	632	+<ul></td><td valign="top"><ul>
620	633	<li> hash-policy
	634	+<li> comment-format
	635	+<li> mimetypes
	636	+<li> forbid-delta-manifests
621	637	<li> mv-rm-files
622	638	<li> ticket-table
623	639	<li> ticket-common
624	640	<li> ticket-change
625	641	<li> ticket-newpage
626	642	<li> ticket-viewpage
627	643	<li> ticket-editpage
628		-<ul></td><td valign="top"><ul>
629	644	<li> ticket-reportlist
630	645	<li> ticket-report-template
631	646	<li> ticket-key-template
632	647	<li> ticket-title-expr
633	648	<li> ticket-closed-expr
	649	+<ul></td><td valign="top"><ul>
	650	+<li> user-color-map
634	651	<li> xfer-common-script
635	652	<li> xfer-push-script
636	653	<li> xfer-commit-script
637	654	<li> xfer-ticket-script
638	655	<li> @reportfmt
639	656	<li> @user
640	657	<li> @concealed
641	658	<li> @shun
	659	+<li> @alias
	660	+<li> @subscriber
	661	+<li> @interwiki
642	662	</ul></td></tr>
643	663	</table>
644	664
645	665	New configuration-names are likely to be added in future releases of
646	666	Fossil. If the server receives a configuration-name that it does not
		@@ -861,11 +881,11 @@
861	881	<h2 id="strategies">5.0 Synchronization Strategies</h2>
862	882
863	883	<h3 id="pull-strategy">5.1 Pull</h3>
864	884
865	885	A typical pull operation proceeds as shown below. Details
866		-of the actual implementation may very slightly but the gist of
	886	+of the actual implementation may vary slightly but the gist of
867	887	a pull is captured in the following steps:
868	888
869	889	<ol>
870	890	<li>The client sends login and pull cards.
871	891	<li>The client sends a cookie card if it has previously received a cookie.
		@@ -1071,6 +1091,6 @@
1071	1091	</ol>
1072	1092
1073	1093	In a complex debugging situation, you can run the command
1074	1094	"fossil sync --transport-command ./debugging_script" where
1075	1095	"debugging_script" is some script of your own that invokes
1076		-the anomolous behavior your are trying to debug.
	1096	+the anomalous behavior you are trying to debug.
1077	1097

	--- www/sync.wiki
	+++ www/sync.wiki
	@@ -518,11 +518,11 @@
518	pronounced as if it were a single word "gimme" in some dialects of
519	English (including the dialect spoken by the original author of Fossil).
520
521	<h4>3.7.1 Unversioned Gimme Cards</h4>
522
523	Sync synchronizing unversioned content, the client may send "uvgimme"
524	cards to the server. A uvgimme card requests that the server send
525	unversioned content to the client. The format of a uvgimme card is
526	as follows:
527
528	<pre>
	@@ -570,77 +570,97 @@
570
571	<pre>
572	<b>reqconfig</b> <i>configuration-name</i>
573	</pre>
574
575	As of 2018-06-04, the configuration-name must be one of the
576	following values:
577
578	<table border=0 align="center">
579	<tr><td valign="top">
580	<ul>
581	<li> css
582	<li> header

583	<li> footer
584	<li> details


585	<li> logo-mimetype
586	<li> logo-image
587	<li> background-mimetype
588	<li> background-image
589	<li> index-page

590	<li> timeline-block-markup






591	<li> timeline-max-comment
592	<li> timeline-plaintext



593	<li> adunit
594	<li> adunit-omit-if-admin
595	<li> adunit-omit-if-user
596	<ul></td><td valign="top"><ul>
597	<li> th1-docs

598	<li> th1-hooks
599	<li> th1-setup
600	<li> tcl
601	<li> tcl-setup
602	<li> project-name

603	<li> short-project-name
604	<li> project-description
605	<li> index-page
606	<li> manifest
607	<li> binary-glob
608	<li> clean-glob
609	<li> ignore-glob
610	<li> keep-glob
611	<li> crlf-glob
612	<ul></td><td valign="top"><ul>
613	<li> crnl-glob
614	<li> encoding-glob
615	<li> empty-dirs
616	<li> <s title="removed 2020-08, version 2.12.1">allow-symlinks</s>
617	<li> dotfiles
618	<li> parent-project-code
619	<li> parent-projet-name

620	<li> hash-policy



621	<li> mv-rm-files
622	<li> ticket-table
623	<li> ticket-common
624	<li> ticket-change
625	<li> ticket-newpage
626	<li> ticket-viewpage
627	<li> ticket-editpage
628	<ul></td><td valign="top"><ul>
629	<li> ticket-reportlist
630	<li> ticket-report-template
631	<li> ticket-key-template
632	<li> ticket-title-expr
633	<li> ticket-closed-expr


634	<li> xfer-common-script
635	<li> xfer-push-script
636	<li> xfer-commit-script
637	<li> xfer-ticket-script
638	<li> @reportfmt
639	<li> @user
640	<li> @concealed
641	<li> @shun



642	</ul></td></tr>
643	</table>
644
645	New configuration-names are likely to be added in future releases of
646	Fossil. If the server receives a configuration-name that it does not
	@@ -861,11 +881,11 @@
861	<h2 id="strategies">5.0 Synchronization Strategies</h2>
862
863	<h3 id="pull-strategy">5.1 Pull</h3>
864
865	A typical pull operation proceeds as shown below. Details
866	of the actual implementation may very slightly but the gist of
867	a pull is captured in the following steps:
868
869	<ol>
870	<li>The client sends login and pull cards.
871	<li>The client sends a cookie card if it has previously received a cookie.
	@@ -1071,6 +1091,6 @@
1071	</ol>
1072
1073	In a complex debugging situation, you can run the command
1074	"fossil sync --transport-command ./debugging_script" where
1075	"debugging_script" is some script of your own that invokes
1076	the anomolous behavior your are trying to debug.
1077

	--- www/sync.wiki
	+++ www/sync.wiki
	@@ -518,11 +518,11 @@
518	pronounced as if it were a single word "gimme" in some dialects of
519	English (including the dialect spoken by the original author of Fossil).
520
521	<h4>3.7.1 Unversioned Gimme Cards</h4>
522
523	When synchronizing unversioned content, the client may send "uvgimme"
524	cards to the server. A uvgimme card requests that the server send
525	unversioned content to the client. The format of a uvgimme card is
526	as follows:
527
528	<pre>
	@@ -570,77 +570,97 @@
570
571	<pre>
572	<b>reqconfig</b> <i>configuration-name</i>
573	</pre>
574
575	As of 2024-10-22, the configuration-name must be one of the
576	following values:
577
578	<table border=0 align="center">
579	<tr><td valign="top">
580	<ul>
581	<li> css
582	<li> header
583	<li> mainmenu
584	<li> footer
585	<li> details
586	<li> js
587	<li> default-skin
588	<li> logo-mimetype
589	<li> logo-image
590	<li> background-mimetype
591	<li> background-image
592	<li> icon-mimetype
593	<li> icon-image
594	<li> timeline-block-markup
595	<li> timeline-date-format
596	<li> timeline-default-style
597	<ul></td><td valign="top"><ul>
598	<li> timeline-dwelltime
599	<li> timeline-closetime
600	<li> timeline-hard-newlines
601	<li> timeline-max-comment
602	<li> timeline-plaintext
603	<li> timeline-truncate-at-blank
604	<li> timeline-tslink-info
605	<li> timeline-utc
606	<li> adunit
607	<li> adunit-omit-if-admin
608	<li> adunit-omit-if-user
609	<li> default-csp
610	<li> sitemap-extra
611	<li> safe-html
612	<li> th1-hooks
613	<li> th1-uri-regexp


614	<li> project-name
615	<ul></td><td valign="top"><ul>
616	<li> short-project-name
617	<li> project-description
618	<li> index-page
619	<li> manifest
620	<li> binary-glob
621	<li> clean-glob
622	<li> ignore-glob
623	<li> keep-glob
624	<li> crlf-glob

625	<li> crnl-glob
626	<li> encoding-glob
627	<li> empty-dirs
628	<li> <s title="removed 2020-08, version 2.12.1">allow-symlinks</s>
629	<li> dotfiles
630	<li> parent-project-code
631	<li> parent-project-name
632	<ul></td><td valign="top"><ul>
633	<li> hash-policy
634	<li> comment-format
635	<li> mimetypes
636	<li> forbid-delta-manifests
637	<li> mv-rm-files
638	<li> ticket-table
639	<li> ticket-common
640	<li> ticket-change
641	<li> ticket-newpage
642	<li> ticket-viewpage
643	<li> ticket-editpage

644	<li> ticket-reportlist
645	<li> ticket-report-template
646	<li> ticket-key-template
647	<li> ticket-title-expr
648	<li> ticket-closed-expr
649	<ul></td><td valign="top"><ul>
650	<li> user-color-map
651	<li> xfer-common-script
652	<li> xfer-push-script
653	<li> xfer-commit-script
654	<li> xfer-ticket-script
655	<li> @reportfmt
656	<li> @user
657	<li> @concealed
658	<li> @shun
659	<li> @alias
660	<li> @subscriber
661	<li> @interwiki
662	</ul></td></tr>
663	</table>
664
665	New configuration-names are likely to be added in future releases of
666	Fossil. If the server receives a configuration-name that it does not
	@@ -861,11 +881,11 @@
881	<h2 id="strategies">5.0 Synchronization Strategies</h2>
882
883	<h3 id="pull-strategy">5.1 Pull</h3>
884
885	A typical pull operation proceeds as shown below. Details
886	of the actual implementation may vary slightly but the gist of
887	a pull is captured in the following steps:
888
889	<ol>
890	<li>The client sends login and pull cards.
891	<li>The client sends a cookie card if it has previously received a cookie.
	@@ -1071,6 +1091,6 @@
1091	</ol>
1092
1093	In a complex debugging situation, you can run the command
1094	"fossil sync --transport-command ./debugging_script" where
1095	"debugging_script" is some script of your own that invokes
1096	the anomalous behavior you are trying to debug.
1097

M www/tech_overview.wiki

+1 -1

		--- www/tech_overview.wiki
		+++ www/tech_overview.wiki
		@@ -142,11 +142,11 @@
142	142	if the ~/.fossil file does not already exist
143	143	* Otherwise, use the traditional unix name of "~/.fossil"
144	144
145	145	This algorithm is complex due to the need for historical compatibility.
146	146	Originally, the database was always just "~/.fossil". Then support
147		-for the FOSSIL_HOME environment variable as added. Later, support for the
	147	+for the FOSSIL_HOME environment variable was added. Later, support for the
148	148	[https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html\|XDG-compatible configation filenames]
149	149	was added. Each of these changes needed to continue to support legacy
150	150	installations.
151	151
152	152	On Windows, the configuration database is the first of the following
153	153

	--- www/tech_overview.wiki
	+++ www/tech_overview.wiki
	@@ -142,11 +142,11 @@
142	if the ~/.fossil file does not already exist
143	* Otherwise, use the traditional unix name of "~/.fossil"
144
145	This algorithm is complex due to the need for historical compatibility.
146	Originally, the database was always just "~/.fossil". Then support
147	for the FOSSIL_HOME environment variable as added. Later, support for the
148	[https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html\|XDG-compatible configation filenames]
149	was added. Each of these changes needed to continue to support legacy
150	installations.
151
152	On Windows, the configuration database is the first of the following
153

	--- www/tech_overview.wiki
	+++ www/tech_overview.wiki
	@@ -142,11 +142,11 @@
142	if the ~/.fossil file does not already exist
143	* Otherwise, use the traditional unix name of "~/.fossil"
144
145	This algorithm is complex due to the need for historical compatibility.
146	Originally, the database was always just "~/.fossil". Then support
147	for the FOSSIL_HOME environment variable was added. Later, support for the
148	[https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html\|XDG-compatible configation filenames]
149	was added. Each of these changes needed to continue to support legacy
150	installations.
151
152	On Windows, the configuration database is the first of the following
153

M www/th1.md

+5 -5

		--- www/th1.md
		+++ www/th1.md
		@@ -32,11 +32,11 @@
32	32	the computation, then converting the result back into a string. (This might
33	33	seem inefficient, but it is faster than people imagine, and numeric
34	34	computations do not come up very often for the kinds of work that TH1 does,
35	35	so it has never been a factor.)
36	36
37		-A TH1 script consist of a sequence of commands.
	37	+A TH1 script consists of a sequence of commands.
38	38	Each command is terminated by the first unescaped newline or ";" character.
39	39	The text of the command (excluding the newline or semicolon terminator)
40	40	is broken into space-separated tokens. The first token is the command
41	41	name and subsequent tokens are the arguments. In this sense, TH1 syntax
42	42	is similar to the familiar command-line shell syntax.
		@@ -288,11 +288,11 @@
288	288	The capability expression is a list. Each term of the list is a
289	289	cluster of [capability letters](./caps/ref.html).
290	290	The overall expression is true if any
291	291	one term is true. A single term is true if all letters within that
292	292	term are true. Or, if the term begins with "!", then the term is true
293		-if none of the terms or true. Or, if the term begins with "@" then
	293	+if none of the terms are true. Or, if the term begins with "@" then
294	294	the term is true if all of the capability letters in that term are
295	295	available to the "anonymous" user. Or, if the term is "*" then it is
296	296	always true.
297	297
298	298	Examples:
		@@ -383,11 +383,11 @@
383	383	<a id="date"></a>TH1 date Command
384	384	-----------------------------------
385	385
386	386	* date ?-local?
387	387
388		-Return a strings which is the current time and date. If the -local
	388	+Return a string which is the current time and date. If the -local
389	389	option is used, the date appears using localtime instead of UTC.
390	390
391	391	<a id="decorate"></a>TH1 decorate Command
392	392	-------------------------------------------
393	393
		@@ -570,11 +570,11 @@
570	570	<a id="linecount"></a>TH1 linecount Command
571	571	---------------------------------------------
572	572
573	573	* linecount STRING MAX MIN
574	574
575		-Returns one more than the number of \n characters in STRING. But
	575	+Returns one more than the number of `\n` characters in STRING. But
576	576	never returns less than MIN or more than MAX.
577	577
578	578	<a id="markdown"></a>TH1 markdown Command
579	579	-------------------------------------------
580	580
		@@ -842,11 +842,11 @@
842	842	-----------------------------------------------
843	843
844	844	* verifyCsrf
845	845
846	846	Before using the results of a form, first call this command to verify
847		-that this Anti-CSRF token is present and is valid. If the Anti-CSRF token
	847	+that the Anti-CSRF token is present and is valid. If the Anti-CSRF token
848	848	is missing or is incorrect, that indicates a cross-site scripting attack.
849	849	If the event of an attack is detected, an error message is generated and
850	850	all further processing is aborted.
851	851
852	852	<a id="verifyLogin"></a>TH1 verifyLogin Command
853	853

	--- www/th1.md
	+++ www/th1.md
	@@ -32,11 +32,11 @@
32	the computation, then converting the result back into a string. (This might
33	seem inefficient, but it is faster than people imagine, and numeric
34	computations do not come up very often for the kinds of work that TH1 does,
35	so it has never been a factor.)
36
37	A TH1 script consist of a sequence of commands.
38	Each command is terminated by the first unescaped newline or ";" character.
39	The text of the command (excluding the newline or semicolon terminator)
40	is broken into space-separated tokens. The first token is the command
41	name and subsequent tokens are the arguments. In this sense, TH1 syntax
42	is similar to the familiar command-line shell syntax.
	@@ -288,11 +288,11 @@
288	The capability expression is a list. Each term of the list is a
289	cluster of [capability letters](./caps/ref.html).
290	The overall expression is true if any
291	one term is true. A single term is true if all letters within that
292	term are true. Or, if the term begins with "!", then the term is true
293	if none of the terms or true. Or, if the term begins with "@" then
294	the term is true if all of the capability letters in that term are
295	available to the "anonymous" user. Or, if the term is "*" then it is
296	always true.
297
298	Examples:
	@@ -383,11 +383,11 @@
383	<a id="date"></a>TH1 date Command
384	-----------------------------------
385
386	* date ?-local?
387
388	Return a strings which is the current time and date. If the -local
389	option is used, the date appears using localtime instead of UTC.
390
391	<a id="decorate"></a>TH1 decorate Command
392	-------------------------------------------
393
	@@ -570,11 +570,11 @@
570	<a id="linecount"></a>TH1 linecount Command
571	---------------------------------------------
572
573	* linecount STRING MAX MIN
574
575	Returns one more than the number of \n characters in STRING. But
576	never returns less than MIN or more than MAX.
577
578	<a id="markdown"></a>TH1 markdown Command
579	-------------------------------------------
580
	@@ -842,11 +842,11 @@
842	-----------------------------------------------
843
844	* verifyCsrf
845
846	Before using the results of a form, first call this command to verify
847	that this Anti-CSRF token is present and is valid. If the Anti-CSRF token
848	is missing or is incorrect, that indicates a cross-site scripting attack.
849	If the event of an attack is detected, an error message is generated and
850	all further processing is aborted.
851
852	<a id="verifyLogin"></a>TH1 verifyLogin Command
853

	--- www/th1.md
	+++ www/th1.md
	@@ -32,11 +32,11 @@
32	the computation, then converting the result back into a string. (This might
33	seem inefficient, but it is faster than people imagine, and numeric
34	computations do not come up very often for the kinds of work that TH1 does,
35	so it has never been a factor.)
36
37	A TH1 script consists of a sequence of commands.
38	Each command is terminated by the first unescaped newline or ";" character.
39	The text of the command (excluding the newline or semicolon terminator)
40	is broken into space-separated tokens. The first token is the command
41	name and subsequent tokens are the arguments. In this sense, TH1 syntax
42	is similar to the familiar command-line shell syntax.
	@@ -288,11 +288,11 @@
288	The capability expression is a list. Each term of the list is a
289	cluster of [capability letters](./caps/ref.html).
290	The overall expression is true if any
291	one term is true. A single term is true if all letters within that
292	term are true. Or, if the term begins with "!", then the term is true
293	if none of the terms are true. Or, if the term begins with "@" then
294	the term is true if all of the capability letters in that term are
295	available to the "anonymous" user. Or, if the term is "*" then it is
296	always true.
297
298	Examples:
	@@ -383,11 +383,11 @@
383	<a id="date"></a>TH1 date Command
384	-----------------------------------
385
386	* date ?-local?
387
388	Return a string which is the current time and date. If the -local
389	option is used, the date appears using localtime instead of UTC.
390
391	<a id="decorate"></a>TH1 decorate Command
392	-------------------------------------------
393
	@@ -570,11 +570,11 @@
570	<a id="linecount"></a>TH1 linecount Command
571	---------------------------------------------
572
573	* linecount STRING MAX MIN
574
575	Returns one more than the number of `\n` characters in STRING. But
576	never returns less than MIN or more than MAX.
577
578	<a id="markdown"></a>TH1 markdown Command
579	-------------------------------------------
580
	@@ -842,11 +842,11 @@
842	-----------------------------------------------
843
844	* verifyCsrf
845
846	Before using the results of a form, first call this command to verify
847	that the Anti-CSRF token is present and is valid. If the Anti-CSRF token
848	is missing or is incorrect, that indicates a cross-site scripting attack.
849	If the event of an attack is detected, an error message is generated and
850	all further processing is aborted.
851
852	<a id="verifyLogin"></a>TH1 verifyLogin Command
853

M www/theory1.wiki

+2 -2

		--- www/theory1.wiki
		+++ www/theory1.wiki
		@@ -73,14 +73,14 @@
73	73	implementation detail which currently happens to use SQLite.
74	74
75	75	Another way to think of the relational tables in a Fossil repository is
76	76	as an index for the artifacts. Without the relational tables,
77	77	to generate a report like a timeline would require scanning every artifact -
78		-the equivalent of a full table scan. The relational tables hold pointers
	78	+the equivalent of a full table scan. The relational tables hold pointers to
79	79	the relevant artifacts in presorted order so that generating a timeline
80	80	is much more efficient. So like an index in a relational database, the
81		-relational tables in an Fossil repository do not add any new information,
	81	+relational tables in a Fossil repository do not add any new information,
82	82	they merely make the information in the artifacts faster and easier to
83	83	look up.
84	84
85	85	Fossil is not "based" on SQLite. Fossil simply exploits SQLite as
86	86	a powerful tool to make the implementation easier.
87	87

	--- www/theory1.wiki
	+++ www/theory1.wiki
	@@ -73,14 +73,14 @@
73	implementation detail which currently happens to use SQLite.
74
75	Another way to think of the relational tables in a Fossil repository is
76	as an index for the artifacts. Without the relational tables,
77	to generate a report like a timeline would require scanning every artifact -
78	the equivalent of a full table scan. The relational tables hold pointers
79	the relevant artifacts in presorted order so that generating a timeline
80	is much more efficient. So like an index in a relational database, the
81	relational tables in an Fossil repository do not add any new information,
82	they merely make the information in the artifacts faster and easier to
83	look up.
84
85	Fossil is not "based" on SQLite. Fossil simply exploits SQLite as
86	a powerful tool to make the implementation easier.
87

	--- www/theory1.wiki
	+++ www/theory1.wiki
	@@ -73,14 +73,14 @@
73	implementation detail which currently happens to use SQLite.
74
75	Another way to think of the relational tables in a Fossil repository is
76	as an index for the artifacts. Without the relational tables,
77	to generate a report like a timeline would require scanning every artifact -
78	the equivalent of a full table scan. The relational tables hold pointers to
79	the relevant artifacts in presorted order so that generating a timeline
80	is much more efficient. So like an index in a relational database, the
81	relational tables in a Fossil repository do not add any new information,
82	they merely make the information in the artifacts faster and easier to
83	look up.
84
85	Fossil is not "based" on SQLite. Fossil simply exploits SQLite as
86	a powerful tool to make the implementation easier.
87

M www/tickets.wiki

+2 -2

		--- www/tickets.wiki
		+++ www/tickets.wiki
		@@ -27,11 +27,11 @@
27	27	that key, or the value may be appended to the prior value.
28	28
29	29	<h2>2.0 Ticket Tables</h2>
30	30
31	31	The low-level artifact format for ticket content is tedious and
32		-cumbersome to access in real time. To facility reporting and display
	32	+cumbersome to access in real time. To facilitate reporting and display
33	33	of tickets, the low-level artifact information is collected and
34	34	summarized in a pair of SQL tables in each local repository. Display
35	35	and reporting of tickets is accomplished by querying these two tables.
36	36
37	37	Note that only the low-level ticket change artifacts are synced. The
		@@ -191,9 +191,9 @@
191	191	comment according to its chosen format. Hence, Fossil was enhanced to
192	192	support the "new-style" tickets.
193	193
194	194	The TICKETCHNG table was added to support new-style tickets. In the new
195	195	style, comment text is stored with the "icomment" (for "Incremental Comment")
196		-key and appears separately, and with its on mimetype, in multiple rows
	196	+key and appears separately, and with its own mimetype, in multiple rows
197	197	of the TICKETCHNG table. It then falls to the TH1 script code on the
198	198	View Ticket Page to query the TICKETCHNG table and extract and format
199	199	the various comments in time stamp order.
200	200

	--- www/tickets.wiki
	+++ www/tickets.wiki
	@@ -27,11 +27,11 @@
27	that key, or the value may be appended to the prior value.
28
29	<h2>2.0 Ticket Tables</h2>
30
31	The low-level artifact format for ticket content is tedious and
32	cumbersome to access in real time. To facility reporting and display
33	of tickets, the low-level artifact information is collected and
34	summarized in a pair of SQL tables in each local repository. Display
35	and reporting of tickets is accomplished by querying these two tables.
36
37	Note that only the low-level ticket change artifacts are synced. The
	@@ -191,9 +191,9 @@
191	comment according to its chosen format. Hence, Fossil was enhanced to
192	support the "new-style" tickets.
193
194	The TICKETCHNG table was added to support new-style tickets. In the new
195	style, comment text is stored with the "icomment" (for "Incremental Comment")
196	key and appears separately, and with its on mimetype, in multiple rows
197	of the TICKETCHNG table. It then falls to the TH1 script code on the
198	View Ticket Page to query the TICKETCHNG table and extract and format
199	the various comments in time stamp order.
200

	--- www/tickets.wiki
	+++ www/tickets.wiki
	@@ -27,11 +27,11 @@
27	that key, or the value may be appended to the prior value.
28
29	<h2>2.0 Ticket Tables</h2>
30
31	The low-level artifact format for ticket content is tedious and
32	cumbersome to access in real time. To facilitate reporting and display
33	of tickets, the low-level artifact information is collected and
34	summarized in a pair of SQL tables in each local repository. Display
35	and reporting of tickets is accomplished by querying these two tables.
36
37	Note that only the low-level ticket change artifacts are synced. The
	@@ -191,9 +191,9 @@
191	comment according to its chosen format. Hence, Fossil was enhanced to
192	support the "new-style" tickets.
193
194	The TICKETCHNG table was added to support new-style tickets. In the new
195	style, comment text is stored with the "icomment" (for "Incremental Comment")
196	key and appears separately, and with its own mimetype, in multiple rows
197	of the TICKETCHNG table. It then falls to the TH1 script code on the
198	View Ticket Page to query the TICKETCHNG table and extract and format
199	the various comments in time stamp order.
200

M www/unvers.wiki

+2 -2

		--- www/unvers.wiki
		+++ www/unvers.wiki
		@@ -53,12 +53,12 @@
53	53	[/help?cmd=push\|fossil push] or [/help?cmd=pull\|fossil pull].
54	54	The "-u" option is only available on "sync" and "clone".
55	55	A rough equivalent of an unversioned pull would be the
56	56	[/help?cmd=unversioned\|fossil unversioned revert] command. The
57	57	"unversioned revert"
58		-command causes the unversioned content on the local repository to overwritten
59		-by the unversioned content found on the remote repository.
	58	+command causes the unversioned content on the local repository to be
	59	+overwritten by the unversioned content found on the remote repository.
60	60
61	61	Beware that because unversioned file sync is an uncommonly dangerous
62	62	capability — there being no history to revert to in the case of human
63	63	error — even the all-powerful Fossil "setup" user does not get
64	64	unversioned file sync capability by default. See
65	65

	--- www/unvers.wiki
	+++ www/unvers.wiki
	@@ -53,12 +53,12 @@
53	[/help?cmd=push\|fossil push] or [/help?cmd=pull\|fossil pull].
54	The "-u" option is only available on "sync" and "clone".
55	A rough equivalent of an unversioned pull would be the
56	[/help?cmd=unversioned\|fossil unversioned revert] command. The
57	"unversioned revert"
58	command causes the unversioned content on the local repository to overwritten
59	by the unversioned content found on the remote repository.
60
61	Beware that because unversioned file sync is an uncommonly dangerous
62	capability — there being no history to revert to in the case of human
63	error — even the all-powerful Fossil "setup" user does not get
64	unversioned file sync capability by default. See
65

	--- www/unvers.wiki
	+++ www/unvers.wiki
	@@ -53,12 +53,12 @@
53	[/help?cmd=push\|fossil push] or [/help?cmd=pull\|fossil pull].
54	The "-u" option is only available on "sync" and "clone".
55	A rough equivalent of an unversioned pull would be the
56	[/help?cmd=unversioned\|fossil unversioned revert] command. The
57	"unversioned revert"
58	command causes the unversioned content on the local repository to be
59	overwritten by the unversioned content found on the remote repository.
60
61	Beware that because unversioned file sync is an uncommonly dangerous
62	capability — there being no history to revert to in the case of human
63	error — even the all-powerful Fossil "setup" user does not get
64	unversioned file sync capability by default. See
65

M www/userlinks.wiki

+1 -1

		--- www/userlinks.wiki
		+++ www/userlinks.wiki
		@@ -23,11 +23,11 @@
23	23	* Fossil contains a [./wikitheory.wiki \| built-in wiki].
24	24	* An [./event.wiki \| Event] is a special kind of wiki page associated
25	25	with a point in time rather than a name.
26	26	* [./settings.wiki \| Settings] control the behaviour of Fossil.
27	27	* [./ssl.wiki \| Use SSL] to encrypt communication with the server.
28		- * The [https://fossil-scm.org/forum\|Fossil forum] is, as of mid-2018,
	28	+ * The [https://fossil-scm.org/forum\|Fossil forum] is, as of late-2024,
29	29	the project's central communication channel. The
30	30	[https://www.mail-archive.com/[email protected]
31	31	\| read-only mailing list archives] house discussions spanning Fossil's
32	32	first decade.
33	33	* [./stats.wiki \| Performance statistics] taken from real-world projects
34	34

	--- www/userlinks.wiki
	+++ www/userlinks.wiki
	@@ -23,11 +23,11 @@
23	* Fossil contains a [./wikitheory.wiki \| built-in wiki].
24	* An [./event.wiki \| Event] is a special kind of wiki page associated
25	with a point in time rather than a name.
26	* [./settings.wiki \| Settings] control the behaviour of Fossil.
27	* [./ssl.wiki \| Use SSL] to encrypt communication with the server.
28	* The [https://fossil-scm.org/forum\|Fossil forum] is, as of mid-2018,
29	the project's central communication channel. The
30	[https://www.mail-archive.com/[email protected]
31	\| read-only mailing list archives] house discussions spanning Fossil's
32	first decade.
33	* [./stats.wiki \| Performance statistics] taken from real-world projects
34

	--- www/userlinks.wiki
	+++ www/userlinks.wiki
	@@ -23,11 +23,11 @@
23	* Fossil contains a [./wikitheory.wiki \| built-in wiki].
24	* An [./event.wiki \| Event] is a special kind of wiki page associated
25	with a point in time rather than a name.
26	* [./settings.wiki \| Settings] control the behaviour of Fossil.
27	* [./ssl.wiki \| Use SSL] to encrypt communication with the server.
28	* The [https://fossil-scm.org/forum\|Fossil forum] is, as of late-2024,
29	the project's central communication channel. The
30	[https://www.mail-archive.com/[email protected]
31	\| read-only mailing list archives] house discussions spanning Fossil's
32	first decade.
33	* [./stats.wiki \| Performance statistics] taken from real-world projects
34

M www/webui.wiki

+5 -5

		--- www/webui.wiki
		+++ www/webui.wiki
		@@ -116,15 +116,15 @@
116	116	The "Files" link on the menu allows you to browse through the <b>file
117	117	hierarchy</b> of the project and to view complete changes histories on
118	118	individual files, with hyperlinks to the check-ins that made those
119	119	changes, and with diffs and annotated diffs between versions.
120	120
121		-The web interface supports [./embeddeddoc.wiki \| embedded documentation].
122		-Embedded documentation is documentation files (usually in wiki format)
123		-that are checked into project as part of the source tree. Such files
124		-can be viewed as if they were ordinary web pages. This document that
125		-you are now reading is an example of embedded documentation.
	121	+The web interface supports [./embeddeddoc.wiki \| embedded documentation]
	122	+files (usually in wiki format) that are checked into the project as
	123	+part of the source tree. Such files can be viewed as if they were
	124	+ordinary web pages. This document that you are now reading is an
	125	+example of embedded documentation.
126	126
127	127	<h2>Customizing The Web Interface Appearance</h2>
128	128
129	129	Users with appropriate permissions can customize the look and feel of
130	130	the web interface using the "Admin" link on the main menu of the web
131	131

	--- www/webui.wiki
	+++ www/webui.wiki
	@@ -116,15 +116,15 @@
116	The "Files" link on the menu allows you to browse through the <b>file
117	hierarchy</b> of the project and to view complete changes histories on
118	individual files, with hyperlinks to the check-ins that made those
119	changes, and with diffs and annotated diffs between versions.
120
121	The web interface supports [./embeddeddoc.wiki \| embedded documentation].
122	Embedded documentation is documentation files (usually in wiki format)
123	that are checked into project as part of the source tree. Such files
124	can be viewed as if they were ordinary web pages. This document that
125	you are now reading is an example of embedded documentation.
126
127	<h2>Customizing The Web Interface Appearance</h2>
128
129	Users with appropriate permissions can customize the look and feel of
130	the web interface using the "Admin" link on the main menu of the web
131

	--- www/webui.wiki
	+++ www/webui.wiki
	@@ -116,15 +116,15 @@
116	The "Files" link on the menu allows you to browse through the <b>file
117	hierarchy</b> of the project and to view complete changes histories on
118	individual files, with hyperlinks to the check-ins that made those
119	changes, and with diffs and annotated diffs between versions.
120
121	The web interface supports [./embeddeddoc.wiki \| embedded documentation]
122	files (usually in wiki format) that are checked into the project as
123	part of the source tree. Such files can be viewed as if they were
124	ordinary web pages. This document that you are now reading is an
125	example of embedded documentation.
126
127	<h2>Customizing The Web Interface Appearance</h2>
128
129	Users with appropriate permissions can customize the look and feel of
130	the web interface using the "Admin" link on the main menu of the web
131

M www/whyallinone.md

+7 -4

		--- www/whyallinone.md
		+++ www/whyallinone.md
		@@ -141,11 +141,11 @@
141	141
142	142	7. Hosting all of these elements within a single service gives a
143	143	consistent look-and-feel across all aspects of the project.
144	144
145	145	Skinning independent software packages’ web interfaces to make
146		- them appear unified is more work than skinning everything once, as
	146	+ them appear unified is more work than [skinning] everything once, as
147	147	in Fossil, and even then, you can’t make independently-developed
148	148	software look like it was produced by a single entity without
149	149	resorting to heroic levels of customization. If you use a separate
150	150	DVCS web front end, chat system, forum manager, documentation
151	151	system, ticket tracker, and so on, you are likely to be relegated
		@@ -217,16 +217,19 @@
217	217	configuration change to its [role-based access control system](./caps/).
218	218	When you’re ready to turn these additional features on, you can do so
219	219	with a few mouse clicks.
220	220
221	221	Because Fossil is web-native out of the box, if you’ve delegated these
222		-features to outside systems to flesh out Git’s DVCS-only nature, Fossil
223		-can link out to these systems, and they back into Fossil, letting you
224		-use Fossil in the same DVCS-only mode.
	222	+features to outside systems to flesh out Git’s DVCS-only nature, you are
	223	+free to do the same with Fossil. One of the many things the [skinning]
	224	+facility allows is replacing the built-in links to the wiki, forum,
	225	+ticket system, etc. with links to external systems. How easy those
	226	+systems make it to link back into Fossil is up to their developers.
225	227
226	228	[Discord]: https://discord.com/
227	229	[edoc]: ./embeddeddoc.wiki
228	230	[Jira]: https://www.atlassian.com/software/jira
229	231	[MediaWiki]: https://www.mediawiki.org/
	232	+[skinning]: ./customskin.md
230	233	[Sphinx]: https://www.sphinx-doc.org/en/master/
231	234	[SSO]: https://en.wikipedia.org/wiki/Single_sign-on
232	235	[srckl]: https://www.sqlite.org/src/ext/checklist/top/index
233	236

	--- www/whyallinone.md
	+++ www/whyallinone.md
	@@ -141,11 +141,11 @@
141
142	7. Hosting all of these elements within a single service gives a
143	consistent look-and-feel across all aspects of the project.
144
145	Skinning independent software packages’ web interfaces to make
146	them appear unified is more work than skinning everything once, as
147	in Fossil, and even then, you can’t make independently-developed
148	software look like it was produced by a single entity without
149	resorting to heroic levels of customization. If you use a separate
150	DVCS web front end, chat system, forum manager, documentation
151	system, ticket tracker, and so on, you are likely to be relegated
	@@ -217,16 +217,19 @@
217	configuration change to its [role-based access control system](./caps/).
218	When you’re ready to turn these additional features on, you can do so
219	with a few mouse clicks.
220
221	Because Fossil is web-native out of the box, if you’ve delegated these
222	features to outside systems to flesh out Git’s DVCS-only nature, Fossil
223	can link out to these systems, and they back into Fossil, letting you
224	use Fossil in the same DVCS-only mode.


225
226	[Discord]: https://discord.com/
227	[edoc]: ./embeddeddoc.wiki
228	[Jira]: https://www.atlassian.com/software/jira
229	[MediaWiki]: https://www.mediawiki.org/

230	[Sphinx]: https://www.sphinx-doc.org/en/master/
231	[SSO]: https://en.wikipedia.org/wiki/Single_sign-on
232	[srckl]: https://www.sqlite.org/src/ext/checklist/top/index
233

	--- www/whyallinone.md
	+++ www/whyallinone.md
	@@ -141,11 +141,11 @@
141
142	7. Hosting all of these elements within a single service gives a
143	consistent look-and-feel across all aspects of the project.
144
145	Skinning independent software packages’ web interfaces to make
146	them appear unified is more work than [skinning] everything once, as
147	in Fossil, and even then, you can’t make independently-developed
148	software look like it was produced by a single entity without
149	resorting to heroic levels of customization. If you use a separate
150	DVCS web front end, chat system, forum manager, documentation
151	system, ticket tracker, and so on, you are likely to be relegated
	@@ -217,16 +217,19 @@
217	configuration change to its [role-based access control system](./caps/).
218	When you’re ready to turn these additional features on, you can do so
219	with a few mouse clicks.
220
221	Because Fossil is web-native out of the box, if you’ve delegated these
222	features to outside systems to flesh out Git’s DVCS-only nature, you are
223	free to do the same with Fossil. One of the many things the [skinning]
224	facility allows is replacing the built-in links to the wiki, forum,
225	ticket system, etc. with links to external systems. How easy those
226	systems make it to link back into Fossil is up to their developers.
227
228	[Discord]: https://discord.com/
229	[edoc]: ./embeddeddoc.wiki
230	[Jira]: https://www.atlassian.com/software/jira
231	[MediaWiki]: https://www.mediawiki.org/
232	[skinning]: ./customskin.md
233	[Sphinx]: https://www.sphinx-doc.org/en/master/
234	[SSO]: https://en.wikipedia.org/wiki/Single_sign-on
235	[srckl]: https://www.sqlite.org/src/ext/checklist/top/index
236

M www/wikitheory.wiki

+1 -1

		--- www/wikitheory.wiki
		+++ www/wikitheory.wiki
		@@ -84,7 +84,7 @@
84	84	wiki page at the top of the timeline
85	85	* [/info/19c60b7fc9e2] shows the text of the
86	86	[/wiki?name=checkin/19c60b7fc9e2400e56a6f938bbad0e34ca746ca2eabdecac10945539f1f5e8c6&p\|checkin/19c60b7fc9e2...]
87	87	wiki page in the "About" section.
88	88
89		-This special wiki pages are very useful for recording historical
	89	+These special wiki pages are very useful for recording historical
90	90	notes.
91	91

	--- www/wikitheory.wiki
	+++ www/wikitheory.wiki
	@@ -84,7 +84,7 @@
84	wiki page at the top of the timeline
85	* [/info/19c60b7fc9e2] shows the text of the
86	[/wiki?name=checkin/19c60b7fc9e2400e56a6f938bbad0e34ca746ca2eabdecac10945539f1f5e8c6&p\|checkin/19c60b7fc9e2...]
87	wiki page in the "About" section.
88
89	This special wiki pages are very useful for recording historical
90	notes.
91

	--- www/wikitheory.wiki
	+++ www/wikitheory.wiki
	@@ -84,7 +84,7 @@
84	wiki page at the top of the timeline
85	* [/info/19c60b7fc9e2] shows the text of the
86	[/wiki?name=checkin/19c60b7fc9e2400e56a6f938bbad0e34ca746ca2eabdecac10945539f1f5e8c6&p\|checkin/19c60b7fc9e2...]
87	wiki page in the "About" section.
88
89	These special wiki pages are very useful for recording historical
90	notes.
91

Fossil SCM

Keyboard Shortcuts