Fossil SCM

Merge with trunk to get the latest web UI fixes.

florian 2022-08-14 08:54 timeline-keyboard-navigation merge

Commit e317872280cbbb9f36bbd8099a7cb0f5beb28716c95be8de579ac01039eec1aa

Parent cfed59c9d1fbe1c…

9 files changed +44 -26 +34 -8 +55 -10 +1 -1 +25 -16 +1 -1 +1 -1 +130 -76 +54 -36

~ Dockerfile ~ src/accordion.js ~ src/branch.c ~ src/json_branch.c ~ src/main.c ~ src/setup.c ~ src/stat.c ~ www/build.wiki ~ www/fossil-v-git.wiki

M Dockerfile

+44 -26

		--- Dockerfile
		+++ Dockerfile
		@@ -1,26 +1,44 @@
1		-###
2		-# Dockerfile for Fossil
3		-###
4		-FROM fedora:29
5		-
6		-### Now install some additional parts we will need for the build
7		-RUN dnf update -y && dnf install -y gcc make tcl tcl-devel zlib-devel openssl-devel tar && dnf clean all && groupadd -r fossil -g 433 && useradd -u 431 -r -g fossil -d /opt/fossil -s /sbin/nologin -c "Fossil user" fossil
8		-
9		-### If you want to build "trunk", change the next line accordingly.
10		-ENV FOSSIL_INSTALL_VERSION release
11		-
12		-RUN curl "https://fossil-scm.org/home/tarball/fossil-src.tar.gz?name=fossil-src&uuid=${FOSSIL_INSTALL_VERSION}" \| tar zx
13		-RUN cd fossil-src && ./configure --disable-fusefs --json --with-th1-docs --with-th1-hooks --with-tcl=1 --with-tcl-stubs --with-tcl-private-stubs
14		-RUN cd fossil-src/src && mv main.c main.c.orig && sed s/\"now\"/0/ <main.c.orig >main.c
15		-RUN cd fossil-src && make && strip fossil && cp fossil /usr/bin && cd .. && rm -rf fossil-src && chmod a+rx /usr/bin/fossil && mkdir -p /opt/fossil && chown fossil:fossil /opt/fossil
16		-
17		-### Build is done, remove modules no longer needed
18		-RUN dnf remove -y gcc make zlib-devel tcl-devel openssl-devel tar && dnf clean all
19		-
20		-USER fossil
21		-
22		-ENV HOME /opt/fossil
23		-
24		-EXPOSE 8080
25		-
26		-CMD ["/usr/bin/fossil", "server", "--create", "--user", "admin", "/opt/fossil/repository.fossil"]
	1	+# STAGE 1: Build a static Fossil binary atop Alpine Linux
	2	+
	3	+# Avoid the temptation to swap the wget call below out for an ADD URL
	4	+# directive. The URL is fixed for a given release tag, which triggers
	5	+# Docker's caching behavior, causing it to reuse that version as long
	6	+# as it remains in the cache. We prefer to rely on the caching of the
	7	+# server instance on fossil-scm.org, which will keep these trunk
	8	+# tarballs around until the next trunk commit.
	9	+
	10	+FROM alpine:latest AS builder
	11	+WORKDIR /tmp
	12	+RUN apk update \
	13	+ && apk upgrade --no-cache \
	14	+ && apk add --no-cache \
	15	+ busybox-static gcc make \
	16	+ musl-dev \
	17	+ openssl-dev openssl-libs-static \
	18	+ zlib-dev zlib-static \
	19	+ && wget -O - https://fossil-scm.org/home/tarball/src \| tar -xz \
	20	+ && src/configure --static CFLAGS='-Os -s' \
	21	+ && make -j
	22	+
	23	+# STAGE 2: Pare that back to the bare essentials.
	24	+
	25	+FROM scratch
	26	+WORKDIR /jail
	27	+COPY --from=builder /tmp/fossil /jail/bin/
	28	+COPY --from=builder /bin/busybox.static /bin/busybox
	29	+RUN [ "/bin/busybox", "--install", "/bin" ]
	30	+RUN mkdir -m 700 dev museum \
	31	+ && mknod -m 600 dev/null c 1 3 \
	32	+ && mknod -m 600 dev/urandom c 1 9
	33	+
	34	+# Now we can run the stripped-down environment in a chroot jail, while
	35	+# leaving open the option to debug it live via the Busybox shell.
	36	+
	37	+EXPOSE 8080/tcp
	38	+CMD [ \
	39	+ "bin/fossil", "server", \
	40	+ "--chroot", "/jail", \
	41	+ "--create", \
	42	+ "--jsmode", "bundled", \
	43	+ "--user", "admin", \
	44	+ "museum/repo.fossil"]
27	45

	--- Dockerfile
	+++ Dockerfile
	@@ -1,26 +1,44 @@
1	###
2	# Dockerfile for Fossil
3	###
4	FROM fedora:29
5
6	### Now install some additional parts we will need for the build
7	RUN dnf update -y && dnf install -y gcc make tcl tcl-devel zlib-devel openssl-devel tar && dnf clean all && groupadd -r fossil -g 433 && useradd -u 431 -r -g fossil -d /opt/fossil -s /sbin/nologin -c "Fossil user" fossil
8
9	### If you want to build "trunk", change the next line accordingly.
10	ENV FOSSIL_INSTALL_VERSION release
11
12	RUN curl "https://fossil-scm.org/home/tarball/fossil-src.tar.gz?name=fossil-src&uuid=${FOSSIL_INSTALL_VERSION}" \| tar zx
13	RUN cd fossil-src && ./configure --disable-fusefs --json --with-th1-docs --with-th1-hooks --with-tcl=1 --with-tcl-stubs --with-tcl-private-stubs
14	RUN cd fossil-src/src && mv main.c main.c.orig && sed s/\"now\"/0/ <main.c.orig >main.c
15	RUN cd fossil-src && make && strip fossil && cp fossil /usr/bin && cd .. && rm -rf fossil-src && chmod a+rx /usr/bin/fossil && mkdir -p /opt/fossil && chown fossil:fossil /opt/fossil
16
17	### Build is done, remove modules no longer needed
18	RUN dnf remove -y gcc make zlib-devel tcl-devel openssl-devel tar && dnf clean all
19
20	USER fossil
21
22	ENV HOME /opt/fossil
23
24	EXPOSE 8080
25
26	CMD ["/usr/bin/fossil", "server", "--create", "--user", "admin", "/opt/fossil/repository.fossil"]


















27

	--- Dockerfile
	+++ Dockerfile
	@@ -1,26 +1,44 @@
1	# STAGE 1: Build a static Fossil binary atop Alpine Linux
2
3	# Avoid the temptation to swap the wget call below out for an ADD URL
4	# directive. The URL is fixed for a given release tag, which triggers
5	# Docker's caching behavior, causing it to reuse that version as long
6	# as it remains in the cache. We prefer to rely on the caching of the
7	# server instance on fossil-scm.org, which will keep these trunk
8	# tarballs around until the next trunk commit.
9
10	FROM alpine:latest AS builder
11	WORKDIR /tmp
12	RUN apk update \
13	&& apk upgrade --no-cache \
14	&& apk add --no-cache \
15	busybox-static gcc make \
16	musl-dev \
17	openssl-dev openssl-libs-static \
18	zlib-dev zlib-static \
19	&& wget -O - https://fossil-scm.org/home/tarball/src \| tar -xz \
20	&& src/configure --static CFLAGS='-Os -s' \
21	&& make -j
22
23	# STAGE 2: Pare that back to the bare essentials.
24
25	FROM scratch
26	WORKDIR /jail
27	COPY --from=builder /tmp/fossil /jail/bin/
28	COPY --from=builder /bin/busybox.static /bin/busybox
29	RUN [ "/bin/busybox", "--install", "/bin" ]
30	RUN mkdir -m 700 dev museum \
31	&& mknod -m 600 dev/null c 1 3 \
32	&& mknod -m 600 dev/urandom c 1 9
33
34	# Now we can run the stripped-down environment in a chroot jail, while
35	# leaving open the option to debug it live via the Busybox shell.
36
37	EXPOSE 8080/tcp
38	CMD [ \
39	"bin/fossil", "server", \
40	"--chroot", "/jail", \
41	"--create", \
42	"--jsmode", "bundled", \
43	"--user", "admin", \
44	"museum/repo.fossil"]
45

M src/accordion.js

+34 -8

		--- src/accordion.js
		+++ src/accordion.js
		@@ -1,12 +1,34 @@
1		-/* Attach appropriate javascript to each ".accordion" button so that
2		-** it expands and contracts when clicked.
3		-** The uncompressed source code for the SVG icons can be found on the
4		-** wiki page "branch/accordion-experiments" in the Fossil repository.
	1	+/*
	2	+** Attach appropriate javascript to each ".accordion" button so that it expands
	3	+** and contracts when clicked.
	4	+**
	5	+** The uncompressed source code for the SVG icons can be found on the wiki page
	6	+** "branch/accordion-experiments" in the Fossil repository.
	7	+**
	8	+** Implementation notes:
	9	+**
	10	+** The `maxHeight' CSS property is quite restrictive for vertical resizing of
	11	+** elements, especially for dynamic-content areas like the diff panels. That's
	12	+** why `maxHeight' is set only during animation, to prevent truncated elements.
	13	+** (The diff panels may get truncated right after page loading, and other
	14	+** elements may get truncated when resizing the browser window to a smaller
	15	+** width, causing vertical growth.)
	16	+**
	17	+** Another problem is that `scrollHeight' used to calculate the expanded height
	18	+** while still in the contracted state may return values with small errors on
	19	+** some browsers, especially for large elements, presumably due to omitting the
	20	+** space required by the vertical scrollbar that may become necessary, causing
	21	+** additional horizontal shrinking and consequently more vertical growth than
	22	+** calculated. That's why setting `maxHeight' to `scrollHeight' is considered
	23	+** "good enough" only during animation, but cleared afterwards.
	24	+**
	25	+** https://fossil-scm.org/forum/forumpost/66d7075f40
	26	+** https://fossil-scm.org/home/timeline?r=accordion-fix
5	27	*/
6	28	var acc_svgdata = ["data:image/svg+xml,"+
7		- "%3Csvg xmlns='http:"+"/"+"/www.w3.org/2000/svg' viewBox='0 0 16 16'%3E"+
	29	+ "%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 16 16'%3E"+
8	30	"%3Cpath style='fill:black;opacity:0' d='M16,16H0V0h16v16z'/%3E"+
9	31	"%3Cpath style='fill:rgb(240,240,240)' d='M14,14H2V2h12v12z'/%3E"+
10	32	"%3Cpath style='fill:rgb(64,64,64)' d='M13,13H3V3h10v10z'/%3E"+
11	33	"%3Cpath style='fill:rgb(248,248,248)' d='M12,12H4V4h8v8z'/%3E"+
12	34	"%3Cpath style='fill:rgb(80,128,208)' d='", "'/%3E%3C/svg%3E",
		@@ -19,17 +41,21 @@
19	41	a[i].insertBefore(img,a[i].firstChild);
20	42	img = document.createElement("img");
21	43	img.src = acc_svgdata[0]+acc_svgdata[3]+acc_svgdata[1];
22	44	img.className = "accordion_btn accordion_btn_minus";
23	45	a[i].insertBefore(img,a[i].firstChild);
24		- var p = a[i].nextElementSibling;
25		- p.style.maxHeight = p.scrollHeight + "px";
26	46	a[i].addEventListener("click",function(){
27	47	var x = this.nextElementSibling;
28	48	if( this.classList.contains("accordion_closed") ){
29	49	x.style.maxHeight = x.scrollHeight + "px";
	50	+ setTimeout(function(){
	51	+ x.style.maxHeight = "";
	52	+ },250); // default.css: .accordion_panel { transition-duration }
30	53	}else{
31		- x.style.maxHeight = "0";
	54	+ x.style.maxHeight = x.scrollHeight + "px";
	55	+ setTimeout(function(){
	56	+ x.style.maxHeight = "0";
	57	+ },1);
32	58	}
33	59	this.classList.toggle("accordion_closed");
34	60	});
35	61	}
36	62

	--- src/accordion.js
	+++ src/accordion.js
	@@ -1,12 +1,34 @@
1	/* Attach appropriate javascript to each ".accordion" button so that
2	** it expands and contracts when clicked.
3	** The uncompressed source code for the SVG icons can be found on the
4	** wiki page "branch/accordion-experiments" in the Fossil repository.






















5	*/
6	var acc_svgdata = ["data:image/svg+xml,"+
7	"%3Csvg xmlns='http:"+"/"+"/www.w3.org/2000/svg' viewBox='0 0 16 16'%3E"+
8	"%3Cpath style='fill:black;opacity:0' d='M16,16H0V0h16v16z'/%3E"+
9	"%3Cpath style='fill:rgb(240,240,240)' d='M14,14H2V2h12v12z'/%3E"+
10	"%3Cpath style='fill:rgb(64,64,64)' d='M13,13H3V3h10v10z'/%3E"+
11	"%3Cpath style='fill:rgb(248,248,248)' d='M12,12H4V4h8v8z'/%3E"+
12	"%3Cpath style='fill:rgb(80,128,208)' d='", "'/%3E%3C/svg%3E",
	@@ -19,17 +41,21 @@
19	a[i].insertBefore(img,a[i].firstChild);
20	img = document.createElement("img");
21	img.src = acc_svgdata[0]+acc_svgdata[3]+acc_svgdata[1];
22	img.className = "accordion_btn accordion_btn_minus";
23	a[i].insertBefore(img,a[i].firstChild);
24	var p = a[i].nextElementSibling;
25	p.style.maxHeight = p.scrollHeight + "px";
26	a[i].addEventListener("click",function(){
27	var x = this.nextElementSibling;
28	if( this.classList.contains("accordion_closed") ){
29	x.style.maxHeight = x.scrollHeight + "px";



30	}else{
31	x.style.maxHeight = "0";



32	}
33	this.classList.toggle("accordion_closed");
34	});
35	}
36

	--- src/accordion.js
	+++ src/accordion.js
	@@ -1,12 +1,34 @@
1	/*
2	** Attach appropriate javascript to each ".accordion" button so that it expands
3	** and contracts when clicked.
4	**
5	** The uncompressed source code for the SVG icons can be found on the wiki page
6	** "branch/accordion-experiments" in the Fossil repository.
7	**
8	** Implementation notes:
9	**
10	** The `maxHeight' CSS property is quite restrictive for vertical resizing of
11	** elements, especially for dynamic-content areas like the diff panels. That's
12	** why `maxHeight' is set only during animation, to prevent truncated elements.
13	** (The diff panels may get truncated right after page loading, and other
14	** elements may get truncated when resizing the browser window to a smaller
15	** width, causing vertical growth.)
16	**
17	** Another problem is that `scrollHeight' used to calculate the expanded height
18	** while still in the contracted state may return values with small errors on
19	** some browsers, especially for large elements, presumably due to omitting the
20	** space required by the vertical scrollbar that may become necessary, causing
21	** additional horizontal shrinking and consequently more vertical growth than
22	** calculated. That's why setting `maxHeight' to `scrollHeight' is considered
23	** "good enough" only during animation, but cleared afterwards.
24	**
25	** https://fossil-scm.org/forum/forumpost/66d7075f40
26	** https://fossil-scm.org/home/timeline?r=accordion-fix
27	*/
28	var acc_svgdata = ["data:image/svg+xml,"+
29	"%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 16 16'%3E"+
30	"%3Cpath style='fill:black;opacity:0' d='M16,16H0V0h16v16z'/%3E"+
31	"%3Cpath style='fill:rgb(240,240,240)' d='M14,14H2V2h12v12z'/%3E"+
32	"%3Cpath style='fill:rgb(64,64,64)' d='M13,13H3V3h10v10z'/%3E"+
33	"%3Cpath style='fill:rgb(248,248,248)' d='M12,12H4V4h8v8z'/%3E"+
34	"%3Cpath style='fill:rgb(80,128,208)' d='", "'/%3E%3C/svg%3E",
	@@ -19,17 +41,21 @@
41	a[i].insertBefore(img,a[i].firstChild);
42	img = document.createElement("img");
43	img.src = acc_svgdata[0]+acc_svgdata[3]+acc_svgdata[1];
44	img.className = "accordion_btn accordion_btn_minus";
45	a[i].insertBefore(img,a[i].firstChild);


46	a[i].addEventListener("click",function(){
47	var x = this.nextElementSibling;
48	if( this.classList.contains("accordion_closed") ){
49	x.style.maxHeight = x.scrollHeight + "px";
50	setTimeout(function(){
51	x.style.maxHeight = "";
52	},250); // default.css: .accordion_panel { transition-duration }
53	}else{
54	x.style.maxHeight = x.scrollHeight + "px";
55	setTimeout(function(){
56	x.style.maxHeight = "0";
57	},1);
58	}
59	this.classList.toggle("accordion_closed");
60	});
61	}
62

M src/branch.c

+55 -10

		--- src/branch.c
		+++ src/branch.c
		@@ -289,31 +289,51 @@
289	289	** Prepare a query that will list branches.
290	290	**
291	291	** If (which<0) then the query pulls only closed branches. If
292	292	** (which>0) then the query pulls all (closed and opened)
293	293	** branches. Else the query pulls currently-opened branches.
	294	+**
	295	+** If the BRL_ORDERBY_MTIME flag is set and nLimitMRU ("Limit Most Recently Used
	296	+** style") is a non-zero number, the result is limited to nLimitMRU entries, and
	297	+** the BRL_REVERSE flag is applied in an outer query after processing the limit,
	298	+** so that it's possible to generate short lists with the most recently modified
	299	+** branches sorted chronologically in either direction, as does the "branch lsh"
	300	+** command.
	301	+** For other cases, the outer query is also generated, but works as a no-op. The
	302	+** code to build the outer query is marked with // OUTER QUERY // comments.
294	303	*/
295		-void branch_prepare_list_query(Stmt pQuery, int brFlags, const char zBrNameGlob){
	304	+void branch_prepare_list_query(
	305	+ Stmt *pQuery,
	306	+ int brFlags,
	307	+ const char *zBrNameGlob,
	308	+ int nLimitMRU
	309	+){
296	310	Blob sql;
297	311	blob_init(&sql, 0, 0);
298	312	brlist_create_temp_table();
	313	+ /* Ignore nLimitMRU if no chronological sort requested. */
	314	+ if( (brFlags & BRL_ORDERBY_MTIME)==0 ) nLimitMRU = 0;
	315	+ /* Undocumented: invert negative values for nLimitMRU, so that command-line
	316	+ ** arguments similar to `head -5' with "option numbers" are possible. */
	317	+ if( nLimitMRU<0 ) nLimitMRU = -nLimitMRU;
	318	+ blob_append_sql(&sql,"SELECT name, isprivate FROM ("); /* OUTER QUERY */
299	319	switch( brFlags & BRL_OPEN_CLOSED_MASK ){
300	320	case BRL_CLOSED_ONLY: {
301	321	blob_append_sql(&sql,
302		- "SELECT name, isprivate FROM tmp_brlist WHERE isclosed"
	322	+ "SELECT name, isprivate, mtime FROM tmp_brlist WHERE isclosed"
303	323	);
304	324	break;
305	325	}
306	326	case BRL_BOTH: {
307	327	blob_append_sql(&sql,
308		- "SELECT name, isprivate FROM tmp_brlist WHERE 1"
	328	+ "SELECT name, isprivate, mtime FROM tmp_brlist WHERE 1"
309	329	);
310	330	break;
311	331	}
312	332	case BRL_OPEN_ONLY: {
313	333	blob_append_sql(&sql,
314		- "SELECT name, isprivate FROM tmp_brlist WHERE NOT isclosed"
	334	+ "SELECT name, isprivate, mtime FROM tmp_brlist WHERE NOT isclosed"
315	335	);
316	336	break;
317	337	}
318	338	}
319	339	if( brFlags & BRL_PRIVATE ) blob_append_sql(&sql, " AND isprivate");
		@@ -321,13 +341,20 @@
321	341	if( brFlags & BRL_ORDERBY_MTIME ){
322	342	blob_append_sql(&sql, " ORDER BY -mtime");
323	343	}else{
324	344	blob_append_sql(&sql, " ORDER BY name COLLATE nocase");
325	345	}
326		- if( brFlags & BRL_REVERSE ){
	346	+ if( brFlags & BRL_REVERSE && !nLimitMRU ){
327	347	blob_append_sql(&sql," DESC");
328	348	}
	349	+ if( nLimitMRU ){
	350	+ blob_append_sql(&sql," LIMIT %d",nLimitMRU);
	351	+ }
	352	+ blob_append_sql(&sql,")"); /* OUTER QUERY */
	353	+ if( brFlags & BRL_REVERSE && nLimitMRU ){
	354	+ blob_append_sql(&sql," ORDER BY mtime"); /* OUTER QUERY */
	355	+ }
329	356	db_prepare_blob(pQuery, &sql);
330	357	blob_reset(&sql);
331	358	}
332	359
333	360	/*
		@@ -585,10 +612,11 @@
585	612	** > fossil branch info BRANCH-NAME
586	613	**
587	614	** Print information about a branch
588	615	**
589	616	** > fossil branch list\|ls ?OPTIONS? ?GLOB?
	617	+** > fossil branch lsh ?OPTIONS? ?LIMIT?
590	618	**
591	619	** List all branches. Options:
592	620	** -a\|--all List all branches. Default show only open branches
593	621	** -c\|--closed List closed branches.
594	622	** -p List only private branches.
		@@ -597,10 +625,15 @@
597	625	**
598	626	** The current branch is marked with an asterisk. Private branches are
599	627	** marked with a hash sign.
600	628	**
601	629	** If GLOB is given, show only branches matching the pattern.
	630	+**
	631	+** The "lsh" variant of this subcommand shows recently changed branches,
	632	+** and accepts an optional LIMIT argument (defaults to 5) to cap output,
	633	+** but no GLOB argument. All other options are supported, with -t being
	634	+** an implied no-op.
602	635	**
603	636	** > fossil branch new BRANCH-NAME BASIS ?OPTIONS?
604	637	**
605	638	** Create a new branch BRANCH-NAME off of check-in BASIS.
606	639	** Supported options for this subcommand include:
		@@ -651,29 +684,41 @@
651	684	"SELECT datetime(mtime,toLocal()) FROM event"
652	685	" WHERE objid=%d", rid);
653	686	fossil_print("%s: open as of %s on %.16s\n", zBrName, zDate, zUuid);
654	687	}
655	688	}
656		- }else if( (strncmp(zCmd,"list",n)==0)\|\|(strncmp(zCmd, "ls", n)==0) ){
	689	+ }else if( strncmp(zCmd,"list",n)==0 \|\|
	690	+ strncmp(zCmd, "ls", n)==0 \|\|
	691	+ strcmp(zCmd, "lsh")==0 ){
657	692	Stmt q;
658	693	int vid;
659	694	char *zCurrent = 0;
660	695	const char *zBrNameGlob = 0;
	696	+ int nLimit = 0;
661	697	int brFlags = BRL_OPEN_ONLY;
662	698	if( find_option("all","a",0)!=0 ) brFlags = BRL_BOTH;
663	699	if( find_option("closed","c",0)!=0 ) brFlags = BRL_CLOSED_ONLY;
664	700	if( find_option("t",0,0)!=0 ) brFlags \|= BRL_ORDERBY_MTIME;
665	701	if( find_option("r",0,0)!=0 ) brFlags \|= BRL_REVERSE;
666	702	if( find_option("p",0,0)!=0 ) brFlags \|= BRL_PRIVATE;
667		- if( g.argc >= 4 ) zBrNameGlob = g.argv[3];
	703	+
	704	+ if( strcmp(zCmd, "lsh")==0 ){
	705	+ nLimit = 5;
	706	+ if( g.argc>4 \|\| (g.argc==4 && (nLimit = atoi(g.argv[3]))==0) ){
	707	+ fossil_fatal("the lsh subcommand allows one optional numeric argument");
	708	+ }
	709	+ brFlags \|= BRL_ORDERBY_MTIME;
	710	+ }else{
	711	+ if( g.argc >= 4 ) zBrNameGlob = g.argv[3];
	712	+ }
668	713
669	714	if( g.localOpen ){
670	715	vid = db_lget_int("checkout", 0);
671	716	zCurrent = db_text(0, "SELECT value FROM tagxref"
672	717	" WHERE rid=%d AND tagid=%d", vid, TAG_BRANCH);
673	718	}
674		- branch_prepare_list_query(&q, brFlags, zBrNameGlob);
	719	+ branch_prepare_list_query(&q, brFlags, zBrNameGlob, nLimit);
675	720	while( db_step(&q)==SQLITE_ROW ){
676	721	const char *zBr = db_column_text(&q, 0);
677	722	int isPriv = zCurrent!=0 && db_column_int(&q, 1)==1;
678	723	int isCur = zCurrent!=0 && fossil_strcmp(zCurrent,zBr)==0;
679	724	fossil_print("%s%s%s\n",
		@@ -703,11 +748,11 @@
703	748	usage("branch unhide branch-name(s)...");
704	749	}
705	750	branch_cmd_hide(3,0);
706	751	}else{
707	752	fossil_fatal("branch subcommand should be one of: "
708		- "close current hide info list ls new reopen unhide");
	753	+ "close current hide info list ls lsh new reopen unhide");
709	754	}
710	755	}
711	756
712	757	/*
713	758	** This is the new-style branch-list page that shows the branch names
		@@ -862,11 +907,11 @@
862	907	@ reopened).</li>
863	908	@ </ol>
864	909	style_sidebox_end();
865	910	#endif
866	911
867		- branch_prepare_list_query(&q, brFlags, 0);
	912	+ branch_prepare_list_query(&q, brFlags, 0, 0);
868	913	cnt = 0;
869	914	while( db_step(&q)==SQLITE_ROW ){
870	915	const char *zBr = db_column_text(&q, 0);
871	916	if( cnt==0 ){
872	917	if( colorTest ){
873	918

	--- src/branch.c
	+++ src/branch.c
	@@ -289,31 +289,51 @@
289	** Prepare a query that will list branches.
290	**
291	** If (which<0) then the query pulls only closed branches. If
292	** (which>0) then the query pulls all (closed and opened)
293	** branches. Else the query pulls currently-opened branches.









294	*/
295	void branch_prepare_list_query(Stmt pQuery, int brFlags, const char zBrNameGlob){





296	Blob sql;
297	blob_init(&sql, 0, 0);
298	brlist_create_temp_table();






299	switch( brFlags & BRL_OPEN_CLOSED_MASK ){
300	case BRL_CLOSED_ONLY: {
301	blob_append_sql(&sql,
302	"SELECT name, isprivate FROM tmp_brlist WHERE isclosed"
303	);
304	break;
305	}
306	case BRL_BOTH: {
307	blob_append_sql(&sql,
308	"SELECT name, isprivate FROM tmp_brlist WHERE 1"
309	);
310	break;
311	}
312	case BRL_OPEN_ONLY: {
313	blob_append_sql(&sql,
314	"SELECT name, isprivate FROM tmp_brlist WHERE NOT isclosed"
315	);
316	break;
317	}
318	}
319	if( brFlags & BRL_PRIVATE ) blob_append_sql(&sql, " AND isprivate");
	@@ -321,13 +341,20 @@
321	if( brFlags & BRL_ORDERBY_MTIME ){
322	blob_append_sql(&sql, " ORDER BY -mtime");
323	}else{
324	blob_append_sql(&sql, " ORDER BY name COLLATE nocase");
325	}
326	if( brFlags & BRL_REVERSE ){
327	blob_append_sql(&sql," DESC");
328	}







329	db_prepare_blob(pQuery, &sql);
330	blob_reset(&sql);
331	}
332
333	/*
	@@ -585,10 +612,11 @@
585	** > fossil branch info BRANCH-NAME
586	**
587	** Print information about a branch
588	**
589	** > fossil branch list\|ls ?OPTIONS? ?GLOB?

590	**
591	** List all branches. Options:
592	** -a\|--all List all branches. Default show only open branches
593	** -c\|--closed List closed branches.
594	** -p List only private branches.
	@@ -597,10 +625,15 @@
597	**
598	** The current branch is marked with an asterisk. Private branches are
599	** marked with a hash sign.
600	**
601	** If GLOB is given, show only branches matching the pattern.





602	**
603	** > fossil branch new BRANCH-NAME BASIS ?OPTIONS?
604	**
605	** Create a new branch BRANCH-NAME off of check-in BASIS.
606	** Supported options for this subcommand include:
	@@ -651,29 +684,41 @@
651	"SELECT datetime(mtime,toLocal()) FROM event"
652	" WHERE objid=%d", rid);
653	fossil_print("%s: open as of %s on %.16s\n", zBrName, zDate, zUuid);
654	}
655	}
656	}else if( (strncmp(zCmd,"list",n)==0)\|\|(strncmp(zCmd, "ls", n)==0) ){


657	Stmt q;
658	int vid;
659	char *zCurrent = 0;
660	const char *zBrNameGlob = 0;

661	int brFlags = BRL_OPEN_ONLY;
662	if( find_option("all","a",0)!=0 ) brFlags = BRL_BOTH;
663	if( find_option("closed","c",0)!=0 ) brFlags = BRL_CLOSED_ONLY;
664	if( find_option("t",0,0)!=0 ) brFlags \|= BRL_ORDERBY_MTIME;
665	if( find_option("r",0,0)!=0 ) brFlags \|= BRL_REVERSE;
666	if( find_option("p",0,0)!=0 ) brFlags \|= BRL_PRIVATE;
667	if( g.argc >= 4 ) zBrNameGlob = g.argv[3];









668
669	if( g.localOpen ){
670	vid = db_lget_int("checkout", 0);
671	zCurrent = db_text(0, "SELECT value FROM tagxref"
672	" WHERE rid=%d AND tagid=%d", vid, TAG_BRANCH);
673	}
674	branch_prepare_list_query(&q, brFlags, zBrNameGlob);
675	while( db_step(&q)==SQLITE_ROW ){
676	const char *zBr = db_column_text(&q, 0);
677	int isPriv = zCurrent!=0 && db_column_int(&q, 1)==1;
678	int isCur = zCurrent!=0 && fossil_strcmp(zCurrent,zBr)==0;
679	fossil_print("%s%s%s\n",
	@@ -703,11 +748,11 @@
703	usage("branch unhide branch-name(s)...");
704	}
705	branch_cmd_hide(3,0);
706	}else{
707	fossil_fatal("branch subcommand should be one of: "
708	"close current hide info list ls new reopen unhide");
709	}
710	}
711
712	/*
713	** This is the new-style branch-list page that shows the branch names
	@@ -862,11 +907,11 @@
862	@ reopened).</li>
863	@ </ol>
864	style_sidebox_end();
865	#endif
866
867	branch_prepare_list_query(&q, brFlags, 0);
868	cnt = 0;
869	while( db_step(&q)==SQLITE_ROW ){
870	const char *zBr = db_column_text(&q, 0);
871	if( cnt==0 ){
872	if( colorTest ){
873

	--- src/branch.c
	+++ src/branch.c
	@@ -289,31 +289,51 @@
289	** Prepare a query that will list branches.
290	**
291	** If (which<0) then the query pulls only closed branches. If
292	** (which>0) then the query pulls all (closed and opened)
293	** branches. Else the query pulls currently-opened branches.
294	**
295	** If the BRL_ORDERBY_MTIME flag is set and nLimitMRU ("Limit Most Recently Used
296	** style") is a non-zero number, the result is limited to nLimitMRU entries, and
297	** the BRL_REVERSE flag is applied in an outer query after processing the limit,
298	** so that it's possible to generate short lists with the most recently modified
299	** branches sorted chronologically in either direction, as does the "branch lsh"
300	** command.
301	** For other cases, the outer query is also generated, but works as a no-op. The
302	** code to build the outer query is marked with // OUTER QUERY // comments.
303	*/
304	void branch_prepare_list_query(
305	Stmt *pQuery,
306	int brFlags,
307	const char *zBrNameGlob,
308	int nLimitMRU
309	){
310	Blob sql;
311	blob_init(&sql, 0, 0);
312	brlist_create_temp_table();
313	/* Ignore nLimitMRU if no chronological sort requested. */
314	if( (brFlags & BRL_ORDERBY_MTIME)==0 ) nLimitMRU = 0;
315	/* Undocumented: invert negative values for nLimitMRU, so that command-line
316	** arguments similar to `head -5' with "option numbers" are possible. */
317	if( nLimitMRU<0 ) nLimitMRU = -nLimitMRU;
318	blob_append_sql(&sql,"SELECT name, isprivate FROM ("); /* OUTER QUERY */
319	switch( brFlags & BRL_OPEN_CLOSED_MASK ){
320	case BRL_CLOSED_ONLY: {
321	blob_append_sql(&sql,
322	"SELECT name, isprivate, mtime FROM tmp_brlist WHERE isclosed"
323	);
324	break;
325	}
326	case BRL_BOTH: {
327	blob_append_sql(&sql,
328	"SELECT name, isprivate, mtime FROM tmp_brlist WHERE 1"
329	);
330	break;
331	}
332	case BRL_OPEN_ONLY: {
333	blob_append_sql(&sql,
334	"SELECT name, isprivate, mtime FROM tmp_brlist WHERE NOT isclosed"
335	);
336	break;
337	}
338	}
339	if( brFlags & BRL_PRIVATE ) blob_append_sql(&sql, " AND isprivate");
	@@ -321,13 +341,20 @@
341	if( brFlags & BRL_ORDERBY_MTIME ){
342	blob_append_sql(&sql, " ORDER BY -mtime");
343	}else{
344	blob_append_sql(&sql, " ORDER BY name COLLATE nocase");
345	}
346	if( brFlags & BRL_REVERSE && !nLimitMRU ){
347	blob_append_sql(&sql," DESC");
348	}
349	if( nLimitMRU ){
350	blob_append_sql(&sql," LIMIT %d",nLimitMRU);
351	}
352	blob_append_sql(&sql,")"); /* OUTER QUERY */
353	if( brFlags & BRL_REVERSE && nLimitMRU ){
354	blob_append_sql(&sql," ORDER BY mtime"); /* OUTER QUERY */
355	}
356	db_prepare_blob(pQuery, &sql);
357	blob_reset(&sql);
358	}
359
360	/*
	@@ -585,10 +612,11 @@
612	** > fossil branch info BRANCH-NAME
613	**
614	** Print information about a branch
615	**
616	** > fossil branch list\|ls ?OPTIONS? ?GLOB?
617	** > fossil branch lsh ?OPTIONS? ?LIMIT?
618	**
619	** List all branches. Options:
620	** -a\|--all List all branches. Default show only open branches
621	** -c\|--closed List closed branches.
622	** -p List only private branches.
	@@ -597,10 +625,15 @@
625	**
626	** The current branch is marked with an asterisk. Private branches are
627	** marked with a hash sign.
628	**
629	** If GLOB is given, show only branches matching the pattern.
630	**
631	** The "lsh" variant of this subcommand shows recently changed branches,
632	** and accepts an optional LIMIT argument (defaults to 5) to cap output,
633	** but no GLOB argument. All other options are supported, with -t being
634	** an implied no-op.
635	**
636	** > fossil branch new BRANCH-NAME BASIS ?OPTIONS?
637	**
638	** Create a new branch BRANCH-NAME off of check-in BASIS.
639	** Supported options for this subcommand include:
	@@ -651,29 +684,41 @@
684	"SELECT datetime(mtime,toLocal()) FROM event"
685	" WHERE objid=%d", rid);
686	fossil_print("%s: open as of %s on %.16s\n", zBrName, zDate, zUuid);
687	}
688	}
689	}else if( strncmp(zCmd,"list",n)==0 \|\|
690	strncmp(zCmd, "ls", n)==0 \|\|
691	strcmp(zCmd, "lsh")==0 ){
692	Stmt q;
693	int vid;
694	char *zCurrent = 0;
695	const char *zBrNameGlob = 0;
696	int nLimit = 0;
697	int brFlags = BRL_OPEN_ONLY;
698	if( find_option("all","a",0)!=0 ) brFlags = BRL_BOTH;
699	if( find_option("closed","c",0)!=0 ) brFlags = BRL_CLOSED_ONLY;
700	if( find_option("t",0,0)!=0 ) brFlags \|= BRL_ORDERBY_MTIME;
701	if( find_option("r",0,0)!=0 ) brFlags \|= BRL_REVERSE;
702	if( find_option("p",0,0)!=0 ) brFlags \|= BRL_PRIVATE;
703
704	if( strcmp(zCmd, "lsh")==0 ){
705	nLimit = 5;
706	if( g.argc>4 \|\| (g.argc==4 && (nLimit = atoi(g.argv[3]))==0) ){
707	fossil_fatal("the lsh subcommand allows one optional numeric argument");
708	}
709	brFlags \|= BRL_ORDERBY_MTIME;
710	}else{
711	if( g.argc >= 4 ) zBrNameGlob = g.argv[3];
712	}
713
714	if( g.localOpen ){
715	vid = db_lget_int("checkout", 0);
716	zCurrent = db_text(0, "SELECT value FROM tagxref"
717	" WHERE rid=%d AND tagid=%d", vid, TAG_BRANCH);
718	}
719	branch_prepare_list_query(&q, brFlags, zBrNameGlob, nLimit);
720	while( db_step(&q)==SQLITE_ROW ){
721	const char *zBr = db_column_text(&q, 0);
722	int isPriv = zCurrent!=0 && db_column_int(&q, 1)==1;
723	int isCur = zCurrent!=0 && fossil_strcmp(zCurrent,zBr)==0;
724	fossil_print("%s%s%s\n",
	@@ -703,11 +748,11 @@
748	usage("branch unhide branch-name(s)...");
749	}
750	branch_cmd_hide(3,0);
751	}else{
752	fossil_fatal("branch subcommand should be one of: "
753	"close current hide info list ls lsh new reopen unhide");
754	}
755	}
756
757	/*
758	** This is the new-style branch-list page that shows the branch names
	@@ -862,11 +907,11 @@
907	@ reopened).</li>
908	@ </ol>
909	style_sidebox_end();
910	#endif
911
912	branch_prepare_list_query(&q, brFlags, 0, 0);
913	cnt = 0;
914	while( db_step(&q)==SQLITE_ROW ){
915	const char *zBr = db_column_text(&q, 0);
916	if( cnt==0 ){
917	if( colorTest ){
918

M src/json_branch.c

+1 -1

		--- src/json_branch.c
		+++ src/json_branch.c
		@@ -128,11 +128,11 @@
128	128	cson_object_set(pay,"current",json_new_string(zCurrent));
129	129	}
130	130	}
131	131
132	132
133		- branch_prepare_list_query(&q, branchListFlags, 0);
	133	+ branch_prepare_list_query(&q, branchListFlags, 0, 0);
134	134	cson_object_set(pay,"branches",listV);
135	135	while((SQLITE_ROW==db_step(&q))){
136	136	cson_value * v = cson_sqlite3_column_to_value(q.pStmt,0);
137	137	if(v){
138	138	cson_array_append(list,v);
139	139

	--- src/json_branch.c
	+++ src/json_branch.c
	@@ -128,11 +128,11 @@
128	cson_object_set(pay,"current",json_new_string(zCurrent));
129	}
130	}
131
132
133	branch_prepare_list_query(&q, branchListFlags, 0);
134	cson_object_set(pay,"branches",listV);
135	while((SQLITE_ROW==db_step(&q))){
136	cson_value * v = cson_sqlite3_column_to_value(q.pStmt,0);
137	if(v){
138	cson_array_append(list,v);
139

	--- src/json_branch.c
	+++ src/json_branch.c
	@@ -128,11 +128,11 @@
128	cson_object_set(pay,"current",json_new_string(zCurrent));
129	}
130	}
131
132
133	branch_prepare_list_query(&q, branchListFlags, 0, 0);
134	cson_object_set(pay,"branches",listV);
135	while((SQLITE_ROW==db_step(&q))){
136	cson_value * v = cson_sqlite3_column_to_value(q.pStmt,0);
137	if(v){
138	cson_array_append(list,v);
139

M src/main.c

+25 -16

		--- src/main.c
		+++ src/main.c
		@@ -1471,21 +1471,24 @@
1471	1471	/*
1472	1472	** If running as root, chroot to the directory containing the
1473	1473	** repository zRepo and then drop root privileges. Return the
1474	1474	** new repository name.
1475	1475	**
1476		-** zRepo might be a directory itself. In that case chroot into
1477		-** the directory zRepo.
	1476	+** zRepo can be a directory. If so and if the repo name was saved
	1477	+** to g.zRepositoryName before we were called, we canonicalize the
	1478	+** two paths and check that one is the prefix of the other, else you
	1479	+** won't be able to open the repo inside the jail. If it all works
	1480	+** out, we return the "jailed" version of the repo name.
1478	1481	**
1479	1482	** Assume the user-id and group-id of the repository, or if zRepo
1480	1483	** is a directory, of that directory.
1481	1484	**
1482	1485	** The noJail flag means that the chroot jail is not entered. But
1483	1486	** privileges are still lowered to that of the user-id and group-id
1484	1487	** of the repository file.
1485	1488	*/
1486		-char enter_chroot_jail(char zRepo, int noJail){
	1489	+static char enter_chroot_jail(const char zRepo, int noJail){
1487	1490	#if !defined(_WIN32)
1488	1491	if( getuid()==0 ){
1489	1492	int i;
1490	1493	struct stat sStat;
1491	1494	Blob dir;
		@@ -1496,15 +1499,27 @@
1496	1499
1497	1500	file_canonical_name(zRepo, &dir, 0);
1498	1501	zDir = blob_str(&dir);
1499	1502	if( !noJail ){
1500	1503	if( file_isdir(zDir, ExtFILE)==1 ){
	1504	+ if( g.zRepositoryName ){
	1505	+ size_t n = strlen(zDir);
	1506	+ Blob repo;
	1507	+ file_canonical_name(g.zRepositoryName, &repo, 0);
	1508	+ zRepo = blob_str(&repo);
	1509	+ if( strncmp(zRepo, zDir, n)!=0 ){
	1510	+ fossil_fatal("repo %s not under chroot dir %s", zRepo, zDir);
	1511	+ }
	1512	+ zRepo += n;
	1513	+ if( *zRepo == '\0' ) zRepo = "/";
	1514	+ }else {
	1515	+ zRepo = "/";
	1516	+ g.fJail = 1;
	1517	+ }
1501	1518	if( file_chdir(zDir, 1) ){
1502	1519	fossil_panic("unable to chroot into %s", zDir);
1503	1520	}
1504		- g.fJail = 1;
1505		- zRepo = "/";
1506	1521	}else{
1507	1522	for(i=strlen(zDir)-1; i>0 && zDir[i]!='/'; i--){}
1508	1523	if( zDir[i]!='/' ) fossil_fatal("bad repository name: %s", zRepo);
1509	1524	if( i>0 ){
1510	1525	zDir[i] = 0;
		@@ -1527,11 +1542,11 @@
1527	1542	if( g.db==0 && file_isfile(zRepo, ExtFILE) ){
1528	1543	db_open_repository(zRepo);
1529	1544	}
1530	1545	}
1531	1546	#endif
1532		- return zRepo;
	1547	+ return (char)zRepo; / no longer const: always reassigned from blob_str() */
1533	1548	}
1534	1549
1535	1550	/*
1536	1551	** Called whenever a crash is encountered while processing a webpage.
1537	1552	*/
		@@ -2814,15 +2829,12 @@
2814	2829	zIpAddr = cgi_ssh_remote_addr(0);
2815	2830	if( zIpAddr && zIpAddr[0] ){
2816	2831	g.fSshClient \|= CGI_SSH_CLIENT;
2817	2832	}
2818	2833	}
2819		- if( zChRoot ){
2820		- enter_chroot_jail((char*)zChRoot, noJail);
2821		- }else{
2822		- g.zRepositoryName = enter_chroot_jail(g.zRepositoryName, noJail);
2823		- }
	2834	+ g.zRepositoryName = enter_chroot_jail(
	2835	+ zChRoot ? zChRoot : g.zRepositoryName, noJail);
2824	2836	if( useSCGI ){
2825	2837	cgi_handle_scgi_request();
2826	2838	}else if( g.fSshClient & CGI_SSH_CLIENT ){
2827	2839	ssh_request_loop(zIpAddr, glob_create(zFileGlob));
2828	2840	}else{
		@@ -3319,15 +3331,12 @@
3319	3331	g.cgiOutput = 1;
3320	3332	find_server_repository(2, 0);
3321	3333	if( fossil_strcmp(g.zRepositoryName,"/")==0 ){
3322	3334	allowRepoList = 1;
3323	3335	}else{
3324		- if( zChRoot ){
3325		- enter_chroot_jail((char*)zChRoot, noJail);
3326		- }else{
3327		- g.zRepositoryName = enter_chroot_jail(g.zRepositoryName, noJail);
3328		- }
	3336	+ g.zRepositoryName = enter_chroot_jail(
	3337	+ zChRoot ? zChRoot : g.zRepositoryName, noJail);
3329	3338	}
3330	3339	if( flags & HTTP_SERVER_SCGI ){
3331	3340	cgi_handle_scgi_request();
3332	3341	}else if( g.httpUseSSL ){
3333	3342	#if FOSSIL_ENABLE_SSL
3334	3343

	--- src/main.c
	+++ src/main.c
	@@ -1471,21 +1471,24 @@
1471	/*
1472	** If running as root, chroot to the directory containing the
1473	** repository zRepo and then drop root privileges. Return the
1474	** new repository name.
1475	**
1476	** zRepo might be a directory itself. In that case chroot into
1477	** the directory zRepo.



1478	**
1479	** Assume the user-id and group-id of the repository, or if zRepo
1480	** is a directory, of that directory.
1481	**
1482	** The noJail flag means that the chroot jail is not entered. But
1483	** privileges are still lowered to that of the user-id and group-id
1484	** of the repository file.
1485	*/
1486	char enter_chroot_jail(char zRepo, int noJail){
1487	#if !defined(_WIN32)
1488	if( getuid()==0 ){
1489	int i;
1490	struct stat sStat;
1491	Blob dir;
	@@ -1496,15 +1499,27 @@
1496
1497	file_canonical_name(zRepo, &dir, 0);
1498	zDir = blob_str(&dir);
1499	if( !noJail ){
1500	if( file_isdir(zDir, ExtFILE)==1 ){














1501	if( file_chdir(zDir, 1) ){
1502	fossil_panic("unable to chroot into %s", zDir);
1503	}
1504	g.fJail = 1;
1505	zRepo = "/";
1506	}else{
1507	for(i=strlen(zDir)-1; i>0 && zDir[i]!='/'; i--){}
1508	if( zDir[i]!='/' ) fossil_fatal("bad repository name: %s", zRepo);
1509	if( i>0 ){
1510	zDir[i] = 0;
	@@ -1527,11 +1542,11 @@
1527	if( g.db==0 && file_isfile(zRepo, ExtFILE) ){
1528	db_open_repository(zRepo);
1529	}
1530	}
1531	#endif
1532	return zRepo;
1533	}
1534
1535	/*
1536	** Called whenever a crash is encountered while processing a webpage.
1537	*/
	@@ -2814,15 +2829,12 @@
2814	zIpAddr = cgi_ssh_remote_addr(0);
2815	if( zIpAddr && zIpAddr[0] ){
2816	g.fSshClient \|= CGI_SSH_CLIENT;
2817	}
2818	}
2819	if( zChRoot ){
2820	enter_chroot_jail((char*)zChRoot, noJail);
2821	}else{
2822	g.zRepositoryName = enter_chroot_jail(g.zRepositoryName, noJail);
2823	}
2824	if( useSCGI ){
2825	cgi_handle_scgi_request();
2826	}else if( g.fSshClient & CGI_SSH_CLIENT ){
2827	ssh_request_loop(zIpAddr, glob_create(zFileGlob));
2828	}else{
	@@ -3319,15 +3331,12 @@
3319	g.cgiOutput = 1;
3320	find_server_repository(2, 0);
3321	if( fossil_strcmp(g.zRepositoryName,"/")==0 ){
3322	allowRepoList = 1;
3323	}else{
3324	if( zChRoot ){
3325	enter_chroot_jail((char*)zChRoot, noJail);
3326	}else{
3327	g.zRepositoryName = enter_chroot_jail(g.zRepositoryName, noJail);
3328	}
3329	}
3330	if( flags & HTTP_SERVER_SCGI ){
3331	cgi_handle_scgi_request();
3332	}else if( g.httpUseSSL ){
3333	#if FOSSIL_ENABLE_SSL
3334

	--- src/main.c
	+++ src/main.c
	@@ -1471,21 +1471,24 @@
1471	/*
1472	** If running as root, chroot to the directory containing the
1473	** repository zRepo and then drop root privileges. Return the
1474	** new repository name.
1475	**
1476	** zRepo can be a directory. If so and if the repo name was saved
1477	** to g.zRepositoryName before we were called, we canonicalize the
1478	** two paths and check that one is the prefix of the other, else you
1479	** won't be able to open the repo inside the jail. If it all works
1480	** out, we return the "jailed" version of the repo name.
1481	**
1482	** Assume the user-id and group-id of the repository, or if zRepo
1483	** is a directory, of that directory.
1484	**
1485	** The noJail flag means that the chroot jail is not entered. But
1486	** privileges are still lowered to that of the user-id and group-id
1487	** of the repository file.
1488	*/
1489	static char enter_chroot_jail(const char zRepo, int noJail){
1490	#if !defined(_WIN32)
1491	if( getuid()==0 ){
1492	int i;
1493	struct stat sStat;
1494	Blob dir;
	@@ -1496,15 +1499,27 @@
1499
1500	file_canonical_name(zRepo, &dir, 0);
1501	zDir = blob_str(&dir);
1502	if( !noJail ){
1503	if( file_isdir(zDir, ExtFILE)==1 ){
1504	if( g.zRepositoryName ){
1505	size_t n = strlen(zDir);
1506	Blob repo;
1507	file_canonical_name(g.zRepositoryName, &repo, 0);
1508	zRepo = blob_str(&repo);
1509	if( strncmp(zRepo, zDir, n)!=0 ){
1510	fossil_fatal("repo %s not under chroot dir %s", zRepo, zDir);
1511	}
1512	zRepo += n;
1513	if( *zRepo == '\0' ) zRepo = "/";
1514	}else {
1515	zRepo = "/";
1516	g.fJail = 1;
1517	}
1518	if( file_chdir(zDir, 1) ){
1519	fossil_panic("unable to chroot into %s", zDir);
1520	}


1521	}else{
1522	for(i=strlen(zDir)-1; i>0 && zDir[i]!='/'; i--){}
1523	if( zDir[i]!='/' ) fossil_fatal("bad repository name: %s", zRepo);
1524	if( i>0 ){
1525	zDir[i] = 0;
	@@ -1527,11 +1542,11 @@
1542	if( g.db==0 && file_isfile(zRepo, ExtFILE) ){
1543	db_open_repository(zRepo);
1544	}
1545	}
1546	#endif
1547	return (char)zRepo; / no longer const: always reassigned from blob_str() */
1548	}
1549
1550	/*
1551	** Called whenever a crash is encountered while processing a webpage.
1552	*/
	@@ -2814,15 +2829,12 @@
2829	zIpAddr = cgi_ssh_remote_addr(0);
2830	if( zIpAddr && zIpAddr[0] ){
2831	g.fSshClient \|= CGI_SSH_CLIENT;
2832	}
2833	}
2834	g.zRepositoryName = enter_chroot_jail(
2835	zChRoot ? zChRoot : g.zRepositoryName, noJail);



2836	if( useSCGI ){
2837	cgi_handle_scgi_request();
2838	}else if( g.fSshClient & CGI_SSH_CLIENT ){
2839	ssh_request_loop(zIpAddr, glob_create(zFileGlob));
2840	}else{
	@@ -3319,15 +3331,12 @@
3331	g.cgiOutput = 1;
3332	find_server_repository(2, 0);
3333	if( fossil_strcmp(g.zRepositoryName,"/")==0 ){
3334	allowRepoList = 1;
3335	}else{
3336	g.zRepositoryName = enter_chroot_jail(
3337	zChRoot ? zChRoot : g.zRepositoryName, noJail);



3338	}
3339	if( flags & HTTP_SERVER_SCGI ){
3340	cgi_handle_scgi_request();
3341	}else if( g.httpUseSSL ){
3342	#if FOSSIL_ENABLE_SSL
3343

M src/setup.c

+1 -1

		--- src/setup.c
		+++ src/setup.c
		@@ -1102,11 +1102,11 @@
1102	1102	@ engines as well as a short RSS description.
1103	1103	@ (Property: "project-description")</p>
1104	1104	@ <hr />
1105	1105	entry_attribute("Canonical Server URL", 40, "email-url",
1106	1106	"eurl", "", 0);
1107		- @ <p>This is the URL used access this repository as a server.
	1107	+ @ <p>This is the URL used to access this repository as a server.
1108	1108	@ Other repositories use this URL to clone or sync against this repository.
1109	1109	@ This is also the basename for hyperlinks included in email alert text.
1110	1110	@ Omit the trailing "/".
1111	1111	@ If this repo will not be set up as a persistent server and will not
1112	1112	@ be sending email alerts, then leave this entry blank.
1113	1113

	--- src/setup.c
	+++ src/setup.c
	@@ -1102,11 +1102,11 @@
1102	@ engines as well as a short RSS description.
1103	@ (Property: "project-description")</p>
1104	@ <hr />
1105	entry_attribute("Canonical Server URL", 40, "email-url",
1106	"eurl", "", 0);
1107	@ <p>This is the URL used access this repository as a server.
1108	@ Other repositories use this URL to clone or sync against this repository.
1109	@ This is also the basename for hyperlinks included in email alert text.
1110	@ Omit the trailing "/".
1111	@ If this repo will not be set up as a persistent server and will not
1112	@ be sending email alerts, then leave this entry blank.
1113

	--- src/setup.c
	+++ src/setup.c
	@@ -1102,11 +1102,11 @@
1102	@ engines as well as a short RSS description.
1103	@ (Property: "project-description")</p>
1104	@ <hr />
1105	entry_attribute("Canonical Server URL", 40, "email-url",
1106	"eurl", "", 0);
1107	@ <p>This is the URL used to access this repository as a server.
1108	@ Other repositories use this URL to clone or sync against this repository.
1109	@ This is also the basename for hyperlinks included in email alert text.
1110	@ Omit the trailing "/".
1111	@ If this repo will not be set up as a persistent server and will not
1112	@ be sending email alerts, then leave this entry blank.
1113

M src/stat.c

+1 -1

		--- src/stat.c
		+++ src/stat.c
		@@ -231,11 +231,11 @@
231	231	@ <tr><th>Number Of Wiki Pages:</th><td>
232	232	n = db_int(0, "SELECT count() FROM tag /scan*/"
233	233	" WHERE +tagname GLOB 'wiki-*'");
234	234	@ %,d(n)
235	235	@ </td></tr>
236		- if( db_table_exists("repository","chat") ){
	236	+ if( g.perm.Chat && db_table_exists("repository","chat") ){
237	237	sqlite3_int64 sz = 0;
238	238	char zSz[100];
239	239	n = db_int(0, "SELECT max(msgid) FROM chat");
240	240	m = db_int(0, "SELECT count(*) FROM chat WHERE mdel IS NOT TRUE");
241	241	sz = db_int64(0, "SELECT sum(coalesce(length(xmsg),0)+"
242	242

	--- src/stat.c
	+++ src/stat.c
	@@ -231,11 +231,11 @@
231	@ <tr><th>Number Of Wiki Pages:</th><td>
232	n = db_int(0, "SELECT count() FROM tag /scan*/"
233	" WHERE +tagname GLOB 'wiki-*'");
234	@ %,d(n)
235	@ </td></tr>
236	if( db_table_exists("repository","chat") ){
237	sqlite3_int64 sz = 0;
238	char zSz[100];
239	n = db_int(0, "SELECT max(msgid) FROM chat");
240	m = db_int(0, "SELECT count(*) FROM chat WHERE mdel IS NOT TRUE");
241	sz = db_int64(0, "SELECT sum(coalesce(length(xmsg),0)+"
242

	--- src/stat.c
	+++ src/stat.c
	@@ -231,11 +231,11 @@
231	@ <tr><th>Number Of Wiki Pages:</th><td>
232	n = db_int(0, "SELECT count() FROM tag /scan*/"
233	" WHERE +tagname GLOB 'wiki-*'");
234	@ %,d(n)
235	@ </td></tr>
236	if( g.perm.Chat && db_table_exists("repository","chat") ){
237	sqlite3_int64 sz = 0;
238	char zSz[100];
239	n = db_int(0, "SELECT max(msgid) FROM chat");
240	m = db_int(0, "SELECT count(*) FROM chat WHERE mdel IS NOT TRUE");
241	sz = db_int64(0, "SELECT sum(coalesce(length(xmsg),0)+"
242

M www/build.wiki

+130 -76

		--- www/build.wiki
		+++ www/build.wiki
		@@ -112,11 +112,12 @@
112	112	<p>For more advanced use cases, see the [./ssl.wiki#openssl-bin\|OpenSSL
113	113	discussion in the "TLS and Fossil" document].</p>
114	114
115	115	<li><p>
116	116	To build a statically linked binary (suitable for use inside a chroot
117		-jail) add the <b>--static</b> option.
	117	+jail) add the <b>--static</b> option. (See the [#docker \| Docker section
	118	+below].)
118	119
119	120	<li><p>
120	121	To enable the native [./th1.md#tclEval \| Tcl integration feature] feature,
121	122	add the <b>--with-tcl=1</b> and <b>--with-tcl-private-stubs=1</b> options.
122	123
		@@ -248,85 +249,138 @@
248	249	TCC += -Dsocketlen_t=int
249	250	TCC += -DSQLITE_MAX_MMAP_SIZE=0
250	251	</pre></blockquote>
251	252	</ul>
252	253
253		-<h2>5.0 Building a Static Binary on Linux using Docker</h2>
254		-
255		-Building a static binary on Linux is not as straightforward as it
256		-could be because the GNU C library requires that certain components be
257		-dynamically loadable. That can be worked around by building against a
258		-different C library, which is simplest to do by way of a container
259		-environment like [https://www.docker.com/ \| Docker].
260		-
261		-The following instructions for building fossil using Docker
262		-were adapted from [https://fossil-scm.org/forum/forumpost/5dd2d61e5f \| forumpost/5dd2d61e5f].
263		-These instructions assume that docker is installed and that the user running
264		-these instructions has permission to do so (i.e., they are <tt>root</tt> or
265		-are a member of the <tt>docker</tt> group).
266		-
267		-First, create a file named <tt>Dockerfile</tt> with the following contents:
268		-
269		-<pre><code>
270		-FROM alpine:edge
271		-RUN apk update \
272		- && apk upgrade \
273		- && apk add --no-cache \
274		- curl gcc make tcl \
275		- musl-dev \
276		- openssl-dev zlib-dev \
277		- openssl-libs-static zlib-static \
278		- && curl \
279		- "https://fossil-scm.org/home/tarball/fossil-src.tar.gz?name=fossil-src&uuid=trunk" \
280		- -o fossil-src.tar.gz \
281		- && tar xf fossil-src.tar.gz \
282		- && cd fossil-src \
283		- && ./configure \
284		- --static \
285		- --disable-fusefs \
286		- --with-th1-docs \
287		- --with-th1-hooks \
288		- && make
289		-</code></pre>
290		-
291		-Be sure to modify the <tt>configure</tt> flags, if desired. e.g., add <tt>--json</tt>
292		-for JSON support.
293		-
294		-From the directory containing that file, build it with docker:
295		-
296		-<pre><code># docker build -t fossil_static .</code></pre>
297		-
298		-If you get permissions errors when running that as a non-root user,
299		-be sure to add the user to the <tt>docker</tt> group before trying
300		-again.
301		-
302		-That creates a docker image and builds a static fossil binary inside
303		-it. That step will take several minutes or more, depending on the
304		-speed of the build environment.
305		-
306		-Next, create a docker container to host the image we just created:
307		-
308		-<pre><code># docker create --name fossil fossil_static</code></pre>
309		-
310		-Then copy the fossil binary from that container:
311		-
312		-<pre><code># docker cp fossil:/fossil-src/fossil fossil</code></pre>
313		-
314		-The resulting binary will be <em>huge</em> because it is built with
315		-debug info. To strip that information, reducing the size greatly:
316		-
317		-<pre><code># strip fossil</code></pre>
318		-
319		-To delete the Docker container and image (if desired), run:
320		-
321		-<pre><code># docker container rm fossil
322		-# docker image ls
323		-</code></pre>
324		-
325		-Note the IDs of the images named <tt>fossil_static</tt> and <tt>alpine</tt>, then:
326		-
327		-<pre><code>docker image rm THE_FOSSIL_ID THE_ALPINE_ID</code></pre>
	254	+
	255	+<h2 id="docker">5.0 Building a Docker Container</h2>
	256	+
	257	+Fossil ships a <tt>Dockerfile</tt> at the top of its source tree which
	258	+you can build like so:
	259	+
	260	+<pre><code> $ docker build -t fossil --no-cache .</code></pre>
	261	+
	262	+If the image built successfully, you can create a container from it and
	263	+test that it runs:
	264	+
	265	+<pre><code> $ docker run --name fossil -p 9999:8080/tcp fossil</code></pre>
	266	+
	267	+This shows us remapping the internal TCP listening port as 9999 on the
	268	+host. As a rule, there's little point to using the "<tt>fossil server
	269	+--port</tt>" feature inside a Docker container. Let it default to 8080
	270	+internally, then remap it to wherever you want it on the host instead.
	271	+
	272	+Our stock <tt>Dockerfile</tt> configures Fossil with the default feature
	273	+set, so you may wish to modify the <tt>Dockerfile</tt> to add
	274	+configuration options, add APK packages to support those options, and so
	275	+forth.
	276	+
	277	+It builds tip-of-trunk. To get a release version instead, append
	278	+"<tt>?r=release</tt>" to the URL in the <tt>Dockerfile</tt>, then
	279	+(re)build it.
	280	+
	281	+
	282	+<h3>5.1 Running It in Production</h3>
	283	+
	284	+If you want the container to serve an existing repository, there are at
	285	+least two right ways to do it.
	286	+
	287	+The wrong way is to use the <tt>Dockerfile COPY</tt> command to bake it
	288	+into the container at build time. It will become one of the container's
	289	+base layers, so that each time you restart the container, it gets reset
	290	+to its prior state. This almost certainly is not what you want.
	291	+
	292	+The simplest correct method is to stop the container if it was running,
	293	+then say:
	294	+
	295	+<pre><code> $ docker cp /path/to/repo.fossil fossil:/jail/museum/repo.fossil
	296	+ $ docker start fossil
	297	+ $ docker exec fossil chown 0 /jail/museum/repo.fossil</code></pre>
	298	+
	299	+That copies the local Fossil repo into the container where the server
	300	+expects to find it, so that the "start" command causes it to serve from
	301	+that copied-in file instead. Since it lives atop the base layers, it
	302	+persists as part of the container proper, surviving restarts.
	303	+
	304	+(The same is true of the default mode of operation: the <tt>fossil
	305	+server --create</tt> flag initializes a fresh Fossil repo atop the base
	306	+layer.)
	307	+
	308	+If you skip the "chown" command and put "http://localhost:9999/" into
	309	+your browser, expecting to see the copied-in repo's home page, you will
	310	+get an opaque "Not Found" error instead. This is because the user and
	311	+group ID of the file will be that of your local user on the container's
	312	+host machine, which won't map to anything in the container's
	313	+<tt>/etc/passwd</tt> and <tt>/etc/group</tt> files, effectively
	314	+preventing the server from reading the copied-in repository file. You
	315	+don't have to restart the server after fixing this: simply reload the
	316	+browser, and Fossil will try again.
	317	+
	318	+This simple method has a problem: Docker containers are designed to be
	319	+killed off at the slightest cause, rebuilt, and redeployed. If you do
	320	+that with the repo inside the container, it gets destroyed, too. The
	321	+solution is to replace the "run" command above with the following:
	322	+
	323	+<pre><code> $ docker create \
	324	+ --name fossil -p 9999:8080/tcp \
	325	+ -v museum:/jail/museum fossil
	326	+</code></pre>
	327	+
	328	+Now when you "docker cp" the local repo into the container, it lands on
	329	+a separate [https://docs.docker.com/storage/volumes/ \| Docker volume]
	330	+mounted inside it, which has an independent lifetime. When you need to
	331	+rebuild the container — such as to upgrade to a newer version of Fossil
	332	+— the volume remains behind and gets remapped into the new contanier
	333	+when you recreate it by giving the above command again.
	334	+
	335	+
	336	+<h3>5.2 Why Chroot?</h3>
	337	+
	338	+A potentially surprising feature of this container is that it runs
	339	+Fossil as root, which causes [./chroot.md \| Fossil's chroot jail
	340	+feature] to kick in. Since a Docker container is a type of über-jail
	341	+already, you may be wondering why we don't either:
	342	+
	343	+ # run <tt>fossil server --nojail</tt> to skip the internal chroot; or
	344	+ # create a non-root user and force Docker to use that instead
	345	+
	346	+The reason is, although this container is quite stripped-down by today's
	347	+standards, it's based on the [https://www.busybox.net/BusyBox.html \|
	348	+surprisingly powerful Busybox project]. (This author made a living for
	349	+years in the early 1990s using Unix systems that were less powerful than
	350	+this container.) If someone ever figured out how to make a Fossil binary
	351	+execute arbitrary commands on the host or to open up a remote shell,
	352	+they'd likely be able to island-hop from there into the rest of your
	353	+network. We need this cute double-jail dance to keep the Fossil instance
	354	+from accessing the Busybox features.
	355	+
	356	+We deem this risk low since a) it's never happened, that we know of;
	357	+and b) we've turned off all of the risky features like TH1 docs.
	358	+Nevertheless, we believe in defense-in-depth.
	359	+
	360	+
	361	+<h3>5.3 Extracting a Static Binary</h3>
	362	+
	363	+Our 2-stage build process uses Alpine Linux only as a build host. Once
	364	+we've got everything reduced to the two key static binaries — Fossil and
	365	+Busybox — we throw all the rest of it away.
	366	+
	367	+A secondary benefit falls out of this process for free: it's arguably
	368	+the easiest way to build a purely static Fossil binary for Linux. Most
	369	+modern Linux distros make this surprisingly difficult, but Alpine's
	370	+back-to-basics nature makes static builds work the way they used to,
	371	+back in the day. If that's what you're after, you can skip the "run"
	372	+command above and create a temporary container from the image, then
	373	+extract the executable from it instead:
	374	+
	375	+<pre><code> $ docker create --name fossil-static-tmp fossil
	376	+ $ docker cp fossil-static-tmp:/jail/bin/fossil .
	377	+ $ docker container rm fossil-static-tmp
	378	+</code></pre>
	379	+
	380	+The resulting binary is the single largest file inside that container,
	381	+at about 6 MiB. (It's built stripped.)
328	382
329	383
330	384	<h2>6.0 Building on/for Android</h2>
331	385
332	386	<h3>6.1 Cross-compiling from Linux</h3>
333	387

	--- www/build.wiki
	+++ www/build.wiki
	@@ -112,11 +112,12 @@
112	<p>For more advanced use cases, see the [./ssl.wiki#openssl-bin\|OpenSSL
113	discussion in the "TLS and Fossil" document].</p>
114
115	<li><p>
116	To build a statically linked binary (suitable for use inside a chroot
117	jail) add the <b>--static</b> option.

118
119	<li><p>
120	To enable the native [./th1.md#tclEval \| Tcl integration feature] feature,
121	add the <b>--with-tcl=1</b> and <b>--with-tcl-private-stubs=1</b> options.
122
	@@ -248,85 +249,138 @@
248	TCC += -Dsocketlen_t=int
249	TCC += -DSQLITE_MAX_MMAP_SIZE=0
250	</pre></blockquote>
251	</ul>
252
253	<h2>5.0 Building a Static Binary on Linux using Docker</h2>
254
255	Building a static binary on Linux is not as straightforward as it
256	could be because the GNU C library requires that certain components be
257	dynamically loadable. That can be worked around by building against a
258	different C library, which is simplest to do by way of a container
259	environment like [https://www.docker.com/ \| Docker].
260
261	The following instructions for building fossil using Docker
262	were adapted from [https://fossil-scm.org/forum/forumpost/5dd2d61e5f \| forumpost/5dd2d61e5f].
263	These instructions assume that docker is installed and that the user running
264	these instructions has permission to do so (i.e., they are <tt>root</tt> or
265	are a member of the <tt>docker</tt> group).
266
267	First, create a file named <tt>Dockerfile</tt> with the following contents:
268
269	<pre><code>
270	FROM alpine:edge
271	RUN apk update \
272	&& apk upgrade \
273	&& apk add --no-cache \
274	curl gcc make tcl \
275	musl-dev \
276	openssl-dev zlib-dev \
277	openssl-libs-static zlib-static \
278	&& curl \
279	"https://fossil-scm.org/home/tarball/fossil-src.tar.gz?name=fossil-src&uuid=trunk" \
280	-o fossil-src.tar.gz \
281	&& tar xf fossil-src.tar.gz \
282	&& cd fossil-src \
283	&& ./configure \
284	--static \
285	--disable-fusefs \
286	--with-th1-docs \
287	--with-th1-hooks \
288	&& make
289	</code></pre>
290
291	Be sure to modify the <tt>configure</tt> flags, if desired. e.g., add <tt>--json</tt>
292	for JSON support.
293
294	From the directory containing that file, build it with docker:
295
296	<pre><code># docker build -t fossil_static .</code></pre>
297
298	If you get permissions errors when running that as a non-root user,
299	be sure to add the user to the <tt>docker</tt> group before trying
300	again.
301
302	That creates a docker image and builds a static fossil binary inside
303	it. That step will take several minutes or more, depending on the
304	speed of the build environment.
305
306	Next, create a docker container to host the image we just created:
307
308	<pre><code># docker create --name fossil fossil_static</code></pre>
309
310	Then copy the fossil binary from that container:
311
312	<pre><code># docker cp fossil:/fossil-src/fossil fossil</code></pre>
313
314	The resulting binary will be <em>huge</em> because it is built with
315	debug info. To strip that information, reducing the size greatly:
316
317	<pre><code># strip fossil</code></pre>
318
319	To delete the Docker container and image (if desired), run:
320
321	<pre><code># docker container rm fossil
322	# docker image ls
323	</code></pre>
324
325	Note the IDs of the images named <tt>fossil_static</tt> and <tt>alpine</tt>, then:
326
327	<pre><code>docker image rm THE_FOSSIL_ID THE_ALPINE_ID</code></pre>





















































328
329
330	<h2>6.0 Building on/for Android</h2>
331
332	<h3>6.1 Cross-compiling from Linux</h3>
333

	--- www/build.wiki
	+++ www/build.wiki
	@@ -112,11 +112,12 @@
112	<p>For more advanced use cases, see the [./ssl.wiki#openssl-bin\|OpenSSL
113	discussion in the "TLS and Fossil" document].</p>
114
115	<li><p>
116	To build a statically linked binary (suitable for use inside a chroot
117	jail) add the <b>--static</b> option. (See the [#docker \| Docker section
118	below].)
119
120	<li><p>
121	To enable the native [./th1.md#tclEval \| Tcl integration feature] feature,
122	add the <b>--with-tcl=1</b> and <b>--with-tcl-private-stubs=1</b> options.
123
	@@ -248,85 +249,138 @@
249	TCC += -Dsocketlen_t=int
250	TCC += -DSQLITE_MAX_MMAP_SIZE=0
251	</pre></blockquote>
252	</ul>
253
254
255	<h2 id="docker">5.0 Building a Docker Container</h2>
256
257	Fossil ships a <tt>Dockerfile</tt> at the top of its source tree which
258	you can build like so:
259
260	<pre><code> $ docker build -t fossil --no-cache .</code></pre>
261
262	If the image built successfully, you can create a container from it and
263	test that it runs:
264
265	<pre><code> $ docker run --name fossil -p 9999:8080/tcp fossil</code></pre>
266
267	This shows us remapping the internal TCP listening port as 9999 on the
268	host. As a rule, there's little point to using the "<tt>fossil server
269	--port</tt>" feature inside a Docker container. Let it default to 8080
270	internally, then remap it to wherever you want it on the host instead.
271
272	Our stock <tt>Dockerfile</tt> configures Fossil with the default feature
273	set, so you may wish to modify the <tt>Dockerfile</tt> to add
274	configuration options, add APK packages to support those options, and so
275	forth.
276
277	It builds tip-of-trunk. To get a release version instead, append
278	"<tt>?r=release</tt>" to the URL in the <tt>Dockerfile</tt>, then
279	(re)build it.
280
281
282	<h3>5.1 Running It in Production</h3>
283
284	If you want the container to serve an existing repository, there are at
285	least two right ways to do it.
286
287	The wrong way is to use the <tt>Dockerfile COPY</tt> command to bake it
288	into the container at build time. It will become one of the container's
289	base layers, so that each time you restart the container, it gets reset
290	to its prior state. This almost certainly is not what you want.
291
292	The simplest correct method is to stop the container if it was running,
293	then say:
294
295	<pre><code> $ docker cp /path/to/repo.fossil fossil:/jail/museum/repo.fossil
296	$ docker start fossil
297	$ docker exec fossil chown 0 /jail/museum/repo.fossil</code></pre>
298
299	That copies the local Fossil repo into the container where the server
300	expects to find it, so that the "start" command causes it to serve from
301	that copied-in file instead. Since it lives atop the base layers, it
302	persists as part of the container proper, surviving restarts.
303
304	(The same is true of the default mode of operation: the <tt>fossil
305	server --create</tt> flag initializes a fresh Fossil repo atop the base
306	layer.)
307
308	If you skip the "chown" command and put "http://localhost:9999/" into
309	your browser, expecting to see the copied-in repo's home page, you will
310	get an opaque "Not Found" error instead. This is because the user and
311	group ID of the file will be that of your local user on the container's
312	host machine, which won't map to anything in the container's
313	<tt>/etc/passwd</tt> and <tt>/etc/group</tt> files, effectively
314	preventing the server from reading the copied-in repository file. You
315	don't have to restart the server after fixing this: simply reload the
316	browser, and Fossil will try again.
317
318	This simple method has a problem: Docker containers are designed to be
319	killed off at the slightest cause, rebuilt, and redeployed. If you do
320	that with the repo inside the container, it gets destroyed, too. The
321	solution is to replace the "run" command above with the following:
322
323	<pre><code> $ docker create \
324	--name fossil -p 9999:8080/tcp \
325	-v museum:/jail/museum fossil
326	</code></pre>
327
328	Now when you "docker cp" the local repo into the container, it lands on
329	a separate [https://docs.docker.com/storage/volumes/ \| Docker volume]
330	mounted inside it, which has an independent lifetime. When you need to
331	rebuild the container — such as to upgrade to a newer version of Fossil
332	— the volume remains behind and gets remapped into the new contanier
333	when you recreate it by giving the above command again.
334
335
336	<h3>5.2 Why Chroot?</h3>
337
338	A potentially surprising feature of this container is that it runs
339	Fossil as root, which causes [./chroot.md \| Fossil's chroot jail
340	feature] to kick in. Since a Docker container is a type of über-jail
341	already, you may be wondering why we don't either:
342
343	# run <tt>fossil server --nojail</tt> to skip the internal chroot; or
344	# create a non-root user and force Docker to use that instead
345
346	The reason is, although this container is quite stripped-down by today's
347	standards, it's based on the [https://www.busybox.net/BusyBox.html \|
348	surprisingly powerful Busybox project]. (This author made a living for
349	years in the early 1990s using Unix systems that were less powerful than
350	this container.) If someone ever figured out how to make a Fossil binary
351	execute arbitrary commands on the host or to open up a remote shell,
352	they'd likely be able to island-hop from there into the rest of your
353	network. We need this cute double-jail dance to keep the Fossil instance
354	from accessing the Busybox features.
355
356	We deem this risk low since a) it's never happened, that we know of;
357	and b) we've turned off all of the risky features like TH1 docs.
358	Nevertheless, we believe in defense-in-depth.
359
360
361	<h3>5.3 Extracting a Static Binary</h3>
362
363	Our 2-stage build process uses Alpine Linux only as a build host. Once
364	we've got everything reduced to the two key static binaries — Fossil and
365	Busybox — we throw all the rest of it away.
366
367	A secondary benefit falls out of this process for free: it's arguably
368	the easiest way to build a purely static Fossil binary for Linux. Most
369	modern Linux distros make this surprisingly difficult, but Alpine's
370	back-to-basics nature makes static builds work the way they used to,
371	back in the day. If that's what you're after, you can skip the "run"
372	command above and create a temporary container from the image, then
373	extract the executable from it instead:
374
375	<pre><code> $ docker create --name fossil-static-tmp fossil
376	$ docker cp fossil-static-tmp:/jail/bin/fossil .
377	$ docker container rm fossil-static-tmp
378	</code></pre>
379
380	The resulting binary is the single largest file inside that container,
381	at about 6 MiB. (It's built stripped.)
382
383
384	<h2>6.0 Building on/for Android</h2>
385
386	<h3>6.1 Cross-compiling from Linux</h3>
387

M www/fossil-v-git.wiki

+54 -36

		--- www/fossil-v-git.wiki
		+++ www/fossil-v-git.wiki
		@@ -113,36 +113,41 @@
113	113	[/help?cmd=ui\|UI],
114	114	protected by [./caps/ \| a fine-grained role-based
115	115	access control system].
116	116	These additional capabilities are available for Git as 3rd-party
117	117	add-ons, but with Fossil they are integrated into
118		-the design. One way to describe Fossil is that it is
	118	+the design, to the point that it approximates
119	119	"[https://github.com/ \| GitHub]-in-a-box."
120	120
121		-Fossil can do operations over all local repo clones and check-out
122		-directories with a single command. For example, Fossil lets you say
123		-"<tt>fossil all sync</tt>" on a laptop prior to taking it off the network
124		-hosting those repos. You can sync up to all of the private repos on your
125		-company network plus those public Internet-hosted repos you use. Whether
126		-going out for a working lunch or on a transoceanic airplane trip, one
127		-command gets you in sync. This works with several other Fossil
128		-sub-commands, such as "<tt>fossil all changes</tt>" to get a list of files
	121	+Even if you only want straight version control, Fossil has affordances
	122	+not available in Git.
	123	+
	124	+For instance, Fossil can do operations over all local repo clones and
	125	+check-out directories with a single command. You can say "<tt>fossil
	126	+all sync</tt>" on a laptop prior to taking it off the network hosting
	127	+those repos, as before going on a trip. It doesn't matter if those
	128	+repos are private and restricted to your company network or public
	129	+Internet-hosted repos, you get synced up with everything you need while
	130	+off-network.
	131	+
	132	+You get the same capability with several other Fossil
	133	+sub-commands as well, such as "<tt>fossil all changes</tt>" to get a list of files
129	134	that you forgot to commit prior to the end of your working day, across
130	135	all repos.
131	136
132	137	Whenever Fossil is told to modify the local checkout in some destructive
133	138	way ([/help?cmd=rm\|fossil rm], [/help?cmd=update\|fossil update],
134	139	[/help?cmd=revert\|fossil revert], etc.) Fossil remembers the prior state
135	140	and is able to return the check-out directory to that state with a
136		-<tt>fossil undo</tt> command. You cannot undo a commit in Fossil
137		-([#history \| on purpose!]) but as long as the change remains confined to
	141	+<tt>fossil undo</tt> command. While you cannot undo a commit in Fossil
	142	+— [#history \| on purpose!] — as long as the change remains confined to
138	143	the local check-out directory only, Fossil makes undo
139	144	[https://git-scm.com/book/en/v2/Git-Basics-Undoing-Things\|easier than in
140	145	Git].
141	146
142		-For developers who choose to self-host projects (rather than using a
143		-3rd-party service such as GitHub) Fossil is much easier to set up, since
	147	+For developers who choose to self-host projects rather than rely on a
	148	+3rd-party service such as GitHub, Fossil is much easier to set up:
144	149	the stand-alone Fossil executable together with a [./server/any/cgi.md\|2-line CGI script]
145	150	suffice to instantiate a full-featured developer website. To accomplish
146	151	the same using Git requires locating, installing, configuring, integrating,
147	152	and managing a wide assortment of separate tools. Standing up a developer
148	153	website using Fossil can be done in minutes, whereas doing the same using
		@@ -165,27 +170,38 @@
165	170	so that most people end up installing it via some kind of package
166	171	manager, simply because the creation of complicated binary packages is
167	172	best delegated to people skilled in their creation. Normal Git users are
168	173	not expected to build Git from source and install it themselves.
169	174
170		-Fossil is a single self-contained stand-alone executable which by default
171		-depends only on common platform libraries. You can statically link
172		-to get an executable with no external dependencies at all — a useful
173		-feature for running inside a restrictive
174		-[https://en.wikipedia.org/wiki/Chroot\|chroot jail].
175		-
176		-The precompiled Fossil binaries are delivered as just a single
177		-executable. The precompiled Windows deliveries are just a ZIP archive
	175	+Fossil is a single self-contained stand-alone executable which
	176	+depends only on common platform libraries in its default configuration.
	177	+To install one of [https://fossil-scm.org/home/uv/download.html \| our
	178	+precompiled binaries], unpack the executable from the archive and put it
	179	+somewhere in your <tt>PATH</tt>. To uninstall it, delete the executable.
	180	+
	181	+This policy is particularly useful when running Fossil inside a
	182	+restrictive container, anything from [./chroot.md \| classic chroot
	183	+jails] to modern [https://en.wikipedia.org/wiki/OS-level_virtualization
	184	+\| OS-level virtualization mechanisms] such as
	185	+[https://en.wikipedia.org/wiki/Docker_(software) \| Docker]. Our
	186	+[/file?name=Dockerfile&ci=trunk \| stock <tt>Dockerfile</tt>]
	187	+creates a container that's under 9 MiB on 64-bit Linux, including
	188	+a capable [https://www.busybox.net/ \| Busybox] environment for live
	189	+debugging of the container's innards.
	190	+
	191	+Modern Linux systems tend to make full static linking
	192	+[https://stackoverflow.com/questions/3430400/linux-static-linking-is-dead
	193	+\| difficult], but our official executables do statically link to OpenSSL
	194	+to remove a version dependency, resulting in an executable that's around
	195	+6 MiB, depending on the platform. ([Release Build How-To \| Details].)
	196	+The result is dependent only upon widespread platform libraries with
	197	+stable ABIs such as glibc, zlib, etc.
	198	+
	199	+Full static linking is easier on Windows, so our precompiled Windows
	200	+binaries are just a ZIP archive
178	201	containing only "<tt>fossil.exe</tt>". There is no "<tt>setup.exe</tt>"
179		-to run. Linux and Mac precompiled binaries are a tarball containing
180		-just the "<tt>fossil</tt>" executable. To install, just put the
181		-executable on your PATH. To uninstall, just delete the executable.
182		-To upgrade (or downgrade) simply replace the executable.
183		-
184		-A typical Fossil executable is between 5 and 7 megabytes uncompressed
185		-(as of 2020-12-12),
186		-assuming that the executable is statically linked against OpenSSL.
	202	+to run.
187	203
188	204	Fossil is easy to build from sources. Just run
189	205	"<tt>./configure && make</tt>" on POSIX systems and
190	206	"<tt>nmake /f Makefile.msc</tt>" on Windows.
191	207
		@@ -208,18 +224,20 @@
208	224	matters is effectiveness and efficiency. We believe Fossil achieves
209	225	this.
210	226
211	227	The above size comparisons aren't apples-to-apples anyway. We've
212	228	compared the size of Fossil with all of its [#features \| many built-in
213		-features] to a fairly minimal Git installation. You must add a lot
214		-of third-party
215		-software to Git to give it a Fossil-equivalent feature set. Consider
216		-[https://about.gitlab.com/\|GitLab], a third-party extension to Git
217		-wrapping it in many features, making it roughly Fossil-equivalent,
	229	+features] to a fairly minimal Git installation. You must add a lot of
	230	+third-party software to Git to give it a Fossil-equivalent feature set.
	231	+Consider [https://about.gitlab.com/\|GitLab], a third-party extension to
	232	+Git wrapping it in many features, making it roughly Fossil-equivalent,
218	233	though [https://docs.gitlab.com/ee/install/requirements.html\|much more
219		-resource hungry] and hence more costly to run than the equivalent
220		-Fossil setup. GitLab's basic requirements are easy to accept when you're dedicating
	234	+resource hungry] and hence more costly to run than the equivalent Fossil
	235	+setup. [https://hub.docker.com/r/gitlab/gitlab-ce/ \| The official GitLab
	236	+Community Edition container] currently clocks in at 2.66 GiB!
	237	+
	238	+GitLab's requirements are easy to accept when you're dedicating
221	239	a local rack server or blade to it, since its minimum requirements are
222	240	more or less a description of the smallest
223	241	thing you could call a "server" these days, but when you go to host that
224	242	in the cloud, you can expect to pay about 8 times as much to comfortably host
225	243	GitLab as for Fossil.³ This difference is largely due to basic
226	244

	--- www/fossil-v-git.wiki
	+++ www/fossil-v-git.wiki
	@@ -113,36 +113,41 @@
113	[/help?cmd=ui\|UI],
114	protected by [./caps/ \| a fine-grained role-based
115	access control system].
116	These additional capabilities are available for Git as 3rd-party
117	add-ons, but with Fossil they are integrated into
118	the design. One way to describe Fossil is that it is
119	"[https://github.com/ \| GitHub]-in-a-box."
120
121	Fossil can do operations over all local repo clones and check-out
122	directories with a single command. For example, Fossil lets you say
123	"<tt>fossil all sync</tt>" on a laptop prior to taking it off the network
124	hosting those repos. You can sync up to all of the private repos on your
125	company network plus those public Internet-hosted repos you use. Whether
126	going out for a working lunch or on a transoceanic airplane trip, one
127	command gets you in sync. This works with several other Fossil
128	sub-commands, such as "<tt>fossil all changes</tt>" to get a list of files





129	that you forgot to commit prior to the end of your working day, across
130	all repos.
131
132	Whenever Fossil is told to modify the local checkout in some destructive
133	way ([/help?cmd=rm\|fossil rm], [/help?cmd=update\|fossil update],
134	[/help?cmd=revert\|fossil revert], etc.) Fossil remembers the prior state
135	and is able to return the check-out directory to that state with a
136	<tt>fossil undo</tt> command. You cannot undo a commit in Fossil
137	([#history \| on purpose!]) but as long as the change remains confined to
138	the local check-out directory only, Fossil makes undo
139	[https://git-scm.com/book/en/v2/Git-Basics-Undoing-Things\|easier than in
140	Git].
141
142	For developers who choose to self-host projects (rather than using a
143	3rd-party service such as GitHub) Fossil is much easier to set up, since
144	the stand-alone Fossil executable together with a [./server/any/cgi.md\|2-line CGI script]
145	suffice to instantiate a full-featured developer website. To accomplish
146	the same using Git requires locating, installing, configuring, integrating,
147	and managing a wide assortment of separate tools. Standing up a developer
148	website using Fossil can be done in minutes, whereas doing the same using
	@@ -165,27 +170,38 @@
165	so that most people end up installing it via some kind of package
166	manager, simply because the creation of complicated binary packages is
167	best delegated to people skilled in their creation. Normal Git users are
168	not expected to build Git from source and install it themselves.
169
170	Fossil is a single self-contained stand-alone executable which by default
171	depends only on common platform libraries. You can statically link
172	to get an executable with no external dependencies at all — a useful
173	feature for running inside a restrictive
174	[https://en.wikipedia.org/wiki/Chroot\|chroot jail].
175
176	The precompiled Fossil binaries are delivered as just a single
177	executable. The precompiled Windows deliveries are just a ZIP archive


















178	containing only "<tt>fossil.exe</tt>". There is no "<tt>setup.exe</tt>"
179	to run. Linux and Mac precompiled binaries are a tarball containing
180	just the "<tt>fossil</tt>" executable. To install, just put the
181	executable on your PATH. To uninstall, just delete the executable.
182	To upgrade (or downgrade) simply replace the executable.
183
184	A typical Fossil executable is between 5 and 7 megabytes uncompressed
185	(as of 2020-12-12),
186	assuming that the executable is statically linked against OpenSSL.
187
188	Fossil is easy to build from sources. Just run
189	"<tt>./configure && make</tt>" on POSIX systems and
190	"<tt>nmake /f Makefile.msc</tt>" on Windows.
191
	@@ -208,18 +224,20 @@
208	matters is effectiveness and efficiency. We believe Fossil achieves
209	this.
210
211	The above size comparisons aren't apples-to-apples anyway. We've
212	compared the size of Fossil with all of its [#features \| many built-in
213	features] to a fairly minimal Git installation. You must add a lot
214	of third-party
215	software to Git to give it a Fossil-equivalent feature set. Consider
216	[https://about.gitlab.com/\|GitLab], a third-party extension to Git
217	wrapping it in many features, making it roughly Fossil-equivalent,
218	though [https://docs.gitlab.com/ee/install/requirements.html\|much more
219	resource hungry] and hence more costly to run than the equivalent
220	Fossil setup. GitLab's basic requirements are easy to accept when you're dedicating



221	a local rack server or blade to it, since its minimum requirements are
222	more or less a description of the smallest
223	thing you could call a "server" these days, but when you go to host that
224	in the cloud, you can expect to pay about 8 times as much to comfortably host
225	GitLab as for Fossil.³ This difference is largely due to basic
226

	--- www/fossil-v-git.wiki
	+++ www/fossil-v-git.wiki
	@@ -113,36 +113,41 @@
113	[/help?cmd=ui\|UI],
114	protected by [./caps/ \| a fine-grained role-based
115	access control system].
116	These additional capabilities are available for Git as 3rd-party
117	add-ons, but with Fossil they are integrated into
118	the design, to the point that it approximates
119	"[https://github.com/ \| GitHub]-in-a-box."
120
121	Even if you only want straight version control, Fossil has affordances
122	not available in Git.
123
124	For instance, Fossil can do operations over all local repo clones and
125	check-out directories with a single command. You can say "<tt>fossil
126	all sync</tt>" on a laptop prior to taking it off the network hosting
127	those repos, as before going on a trip. It doesn't matter if those
128	repos are private and restricted to your company network or public
129	Internet-hosted repos, you get synced up with everything you need while
130	off-network.
131
132	You get the same capability with several other Fossil
133	sub-commands as well, such as "<tt>fossil all changes</tt>" to get a list of files
134	that you forgot to commit prior to the end of your working day, across
135	all repos.
136
137	Whenever Fossil is told to modify the local checkout in some destructive
138	way ([/help?cmd=rm\|fossil rm], [/help?cmd=update\|fossil update],
139	[/help?cmd=revert\|fossil revert], etc.) Fossil remembers the prior state
140	and is able to return the check-out directory to that state with a
141	<tt>fossil undo</tt> command. While you cannot undo a commit in Fossil
142	— [#history \| on purpose!] — as long as the change remains confined to
143	the local check-out directory only, Fossil makes undo
144	[https://git-scm.com/book/en/v2/Git-Basics-Undoing-Things\|easier than in
145	Git].
146
147	For developers who choose to self-host projects rather than rely on a
148	3rd-party service such as GitHub, Fossil is much easier to set up:
149	the stand-alone Fossil executable together with a [./server/any/cgi.md\|2-line CGI script]
150	suffice to instantiate a full-featured developer website. To accomplish
151	the same using Git requires locating, installing, configuring, integrating,
152	and managing a wide assortment of separate tools. Standing up a developer
153	website using Fossil can be done in minutes, whereas doing the same using
	@@ -165,27 +170,38 @@
170	so that most people end up installing it via some kind of package
171	manager, simply because the creation of complicated binary packages is
172	best delegated to people skilled in their creation. Normal Git users are
173	not expected to build Git from source and install it themselves.
174
175	Fossil is a single self-contained stand-alone executable which
176	depends only on common platform libraries in its default configuration.
177	To install one of [https://fossil-scm.org/home/uv/download.html \| our
178	precompiled binaries], unpack the executable from the archive and put it
179	somewhere in your <tt>PATH</tt>. To uninstall it, delete the executable.
180
181	This policy is particularly useful when running Fossil inside a
182	restrictive container, anything from [./chroot.md \| classic chroot
183	jails] to modern [https://en.wikipedia.org/wiki/OS-level_virtualization
184	\| OS-level virtualization mechanisms] such as
185	[https://en.wikipedia.org/wiki/Docker_(software) \| Docker]. Our
186	[/file?name=Dockerfile&ci=trunk \| stock <tt>Dockerfile</tt>]
187	creates a container that's under 9 MiB on 64-bit Linux, including
188	a capable [https://www.busybox.net/ \| Busybox] environment for live
189	debugging of the container's innards.
190
191	Modern Linux systems tend to make full static linking
192	[https://stackoverflow.com/questions/3430400/linux-static-linking-is-dead
193	\| difficult], but our official executables do statically link to OpenSSL
194	to remove a version dependency, resulting in an executable that's around
195	6 MiB, depending on the platform. ([Release Build How-To \| Details].)
196	The result is dependent only upon widespread platform libraries with
197	stable ABIs such as glibc, zlib, etc.
198
199	Full static linking is easier on Windows, so our precompiled Windows
200	binaries are just a ZIP archive
201	containing only "<tt>fossil.exe</tt>". There is no "<tt>setup.exe</tt>"
202	to run.







203
204	Fossil is easy to build from sources. Just run
205	"<tt>./configure && make</tt>" on POSIX systems and
206	"<tt>nmake /f Makefile.msc</tt>" on Windows.
207
	@@ -208,18 +224,20 @@
224	matters is effectiveness and efficiency. We believe Fossil achieves
225	this.
226
227	The above size comparisons aren't apples-to-apples anyway. We've
228	compared the size of Fossil with all of its [#features \| many built-in
229	features] to a fairly minimal Git installation. You must add a lot of
230	third-party software to Git to give it a Fossil-equivalent feature set.
231	Consider [https://about.gitlab.com/\|GitLab], a third-party extension to
232	Git wrapping it in many features, making it roughly Fossil-equivalent,

233	though [https://docs.gitlab.com/ee/install/requirements.html\|much more
234	resource hungry] and hence more costly to run than the equivalent Fossil
235	setup. [https://hub.docker.com/r/gitlab/gitlab-ce/ \| The official GitLab
236	Community Edition container] currently clocks in at 2.66 GiB!
237
238	GitLab's requirements are easy to accept when you're dedicating
239	a local rack server or blade to it, since its minimum requirements are
240	more or less a description of the smallest
241	thing you could call a "server" these days, but when you go to host that
242	in the cloud, you can expect to pay about 8 times as much to comfortably host
243	GitLab as for Fossil.³ This difference is largely due to basic
244

Fossil SCM

Keyboard Shortcuts