Fossil SCM

Enhance the function to find case-preserved filenames on Windows to deal with non-ASCII filenames.

florian 2024-10-15 05:17 trunk

Commit 9919dfbbaa2019e7eb7bbfdb2f821058e25553478e27079f95122d672a6fbdff

Parent 2d5a23e919dff2f…

2 files changed +4 +118

M src/file.c

		--- src/file.c
		+++ src/file.c
		@@ -1372,10 +1372,11 @@
1372	1372	** just fossil_strdup(). But for case-insenstiive but "case preserving"
1373	1373	** filesystems, such as on MacOS or Windows, we want the filename to be
1374	1374	** in the preserved casing. That's what this routine does.
1375	1375	*/
1376	1376	char file_case_preferred_name(const char zDir, const char *zPath){
	1377	+#ifndef _WIN32 /* Call win32_file_case_preferred_name() on Windows. */
1377	1378	DIR *d;
1378	1379	int i;
1379	1380	char *zResult = 0;
1380	1381	void *zNative = 0;
1381	1382
		@@ -1407,10 +1408,13 @@
1407	1408	closedir(d);
1408	1409	}
1409	1410	fossil_path_free(zNative);
1410	1411	if( zResult==0 ) zResult = fossil_strdup(zPath);
1411	1412	return zResult;
	1413	+#else /* _WIN32 */
	1414	+ return win32_file_case_preferred_name(zDir,zPath);
	1415	+#endif /* _WIN32 */
1412	1416	}
1413	1417
1414	1418	/*
1415	1419	** COMMAND: test-case-filename
1416	1420	**
1417	1421

	--- src/file.c
	+++ src/file.c
	@@ -1372,10 +1372,11 @@
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 +1408,13 @@
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	**
1417

	--- src/file.c
	+++ src/file.c
	@@ -1372,10 +1372,11 @@
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	#ifndef _WIN32 /* Call win32_file_case_preferred_name() on Windows. */
1378	DIR *d;
1379	int i;
1380	char *zResult = 0;
1381	void *zNative = 0;
1382
	@@ -1407,10 +1408,13 @@
1408	closedir(d);
1409	}
1410	fossil_path_free(zNative);
1411	if( zResult==0 ) zResult = fossil_strdup(zPath);
1412	return zResult;
1413	#else /* _WIN32 */
1414	return win32_file_case_preferred_name(zDir,zPath);
1415	#endif /* _WIN32 */
1416	}
1417
1418	/*
1419	** COMMAND: test-case-filename
1420	**
1421

M src/winfile.c

+118

		--- src/winfile.c
		+++ src/winfile.c
		@@ -290,6 +290,124 @@
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	+** lstrcmpiW() function on Windows XP.
	299	+*/
	300	+int win32_compare_filenames_nocase(
	301	+ const wchar_t *fn1,
	302	+ const wchar_t *fn2
	303	+){
	304	+ static FARPROC fnCompareStringOrdinal;
	305	+ static int try_fnCompareStringOrdinal;
	306	+ if( !try_fnCompareStringOrdinal ){
	307	+ fnCompareStringOrdinal =
	308	+ GetProcAddress(GetModuleHandleA("kernel32"),"CompareStringOrdinal");
	309	+ try_fnCompareStringOrdinal = 1;
	310	+ }
	311	+ if( fnCompareStringOrdinal ){
	312	+ return -2 + fnCompareStringOrdinal(fn1,-1,fn2,-1,1);
	313	+ }else{
	314	+ return lstrcmpiW(fn1,fn2);
	315	+ }
	316	+}
	317	+
	318	+/* Helper macros to deal with directory separators. */
	319	+#define IS_DIRSEP(s,i) ( s[i]=='/' \|\| s[i]=='\\' )
	320	+#define NEXT_DIRSEP(s,i) while( s[i] && s[i]!='/' && s[i]!='\\' ){i++;}
	321	+
	322	+/* The Win32 version of file_case_preferred_name() from file.c, which is able to
	323	+** find case-preserved file names containing non-ASCII characters. The result is
	324	+** allocated by fossil_malloc() and should be free'd by tha caller. While this
	325	+** function usually gets canonicalized paths, it is able to handle any input and
	326	+** figure out more cases than the original:
	327	+**
	328	+** fossil test-case-filename C:/ .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
	329	+** → Original: .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
	330	+** → Modified: .//..\Windows\/.//.\System32\.\notepad.exe
	331	+**
	332	+** md ÄÖÜ
	333	+** fossil test-case-filename ./\ .\äöü\/[empty]\\/
	334	+** → Original: ./äöü\/[empty]\\/
	335	+** → Modified: .\ÄÖÜ\/[empty]\\/
	336	+**
	337	+** The function preserves slashes and backslashes: only single file or directory
	338	+** components without directory separators ("basenames") are converted to UTF-16
	339	+** using fossil_utf8_to_path(), so bypassing its slash ↔ backslash translations.
	340	+** Note that the original function doesn't preserve all slashes and backslashes,
	341	+** for example in the second example above.
	342	+**
	343	+** NOTE: As of Windows 10, version 1803, case sensitivity may be enabled on a
	344	+** per-directory basis, as returned by NtQueryInformationFile() with the file
	345	+** information class FILE_CASE_SENSITIVE_INFORMATION. So this function may be
	346	+** changed to act like fossil_strdup() for files located in such directories.
	347	+*/
	348	+char *win32_file_case_preferred_name(
	349	+ const char *zBase,
	350	+ const char *zPath
	351	+){
	352	+ int cchBase = strlen(zBase);
	353	+ int cchPath = strlen(zPath);
	354	+ int cchBuf = cchBase + cchPath + 1;
	355	+ int cchRes = cchPath + 1;
	356	+ char *zBuf = fossil_malloc(cchBuf);
	357	+ char *zRes = fossil_malloc(cchRes);
	358	+ int i, j;
	359	+ memcpy(zBuf,zBase,cchBase);
	360	+ cchRes = 0;
	361	+ if( !IS_DIRSEP(zBuf,cchBase-1) ){
	362	+ zBuf[cchBase++]=L'/';
	363	+ }
	364	+ memcpy(zBuf+cchBase,zPath,cchPath+1);
	365	+ i = j = cchBase;
	366	+ while( 1 ){
	367	+ WIN32_FIND_DATAW fd;
	368	+ HANDLE hFind;
	369	+ wchar_t *wzBuf;
	370	+ char *zCompBuf = 0;
	371	+ char *zComp = &zBuf[i];
	372	+ int cchComp;
	373	+ char chSep;
	374	+ int fDone;
	375	+ if( IS_DIRSEP(zBuf,i) ){
	376	+ zRes[cchRes++] = zBuf[i];
	377	+ i = j = i+1;
	378	+ continue;
	379	+ }
	380	+ NEXT_DIRSEP(zBuf,j);
	381	+ fDone = zBuf[j]==0;
	382	+ chSep = zBuf[j];
	383	+ zBuf[j] = 0; /* Truncate working buffer. */
	384	+ wzBuf = fossil_utf8_to_path(zBuf,0);
	385	+ hFind = FindFirstFileW(wzBuf,&fd);
	386	+ if( hFind!= INVALID_HANDLE_VALUE ){
	387	+ wchar_t *wzComp = fossil_utf8_to_path(zComp,0);
	388	+ FindClose(hFind);
	389	+ /* Test fd.cFileName, not fd.cAlternateFileName (classic 8.3 format). */
	390	+ if( win32_compare_filenames_nocase(wzComp,fd.cFileName)==0 ){
	391	+ zCompBuf = fossil_path_to_utf8(fd.cFileName);
	392	+ zComp = zCompBuf;
	393	+ }
	394	+ fossil_path_free(wzComp);
	395	+ }
	396	+ fossil_path_free(wzBuf);
	397	+ cchComp = strlen(zComp);
	398	+ memcpy(zRes+cchRes,zComp,cchComp);
	399	+ cchRes += cchComp;
	400	+ if( zCompBuf ){
	401	+ fossil_path_free(zCompBuf);
	402	+ }
	403	+ if( fDone ){
	404	+ zRes[cchRes] = 0;
	405	+ break;
	406	+ }
	407	+ zBuf[j] = chSep; /* Undo working buffer truncation. */
	408	+ i = j;
	409	+ }
	410	+ fossil_free(zBuf);
	411	+ return zRes;
	412	+}
295	413	#endif /* _WIN32 -- This code is for win32 only */
296	414

	--- src/winfile.c
	+++ src/winfile.c
	@@ -290,6 +290,124 @@
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 +290,124 @@
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	** lstrcmpiW() function on Windows XP.
299	*/
300	int win32_compare_filenames_nocase(
301	const wchar_t *fn1,
302	const wchar_t *fn2
303	){
304	static FARPROC fnCompareStringOrdinal;
305	static int try_fnCompareStringOrdinal;
306	if( !try_fnCompareStringOrdinal ){
307	fnCompareStringOrdinal =
308	GetProcAddress(GetModuleHandleA("kernel32"),"CompareStringOrdinal");
309	try_fnCompareStringOrdinal = 1;
310	}
311	if( fnCompareStringOrdinal ){
312	return -2 + fnCompareStringOrdinal(fn1,-1,fn2,-1,1);
313	}else{
314	return lstrcmpiW(fn1,fn2);
315	}
316	}
317
318	/* Helper macros to deal with directory separators. */
319	#define IS_DIRSEP(s,i) ( s[i]=='/' \|\| s[i]=='\\' )
320	#define NEXT_DIRSEP(s,i) while( s[i] && s[i]!='/' && s[i]!='\\' ){i++;}
321
322	/* The Win32 version of file_case_preferred_name() from file.c, which is able to
323	** find case-preserved file names containing non-ASCII characters. The result is
324	** allocated by fossil_malloc() and should be free'd by tha caller. While this
325	** function usually gets canonicalized paths, it is able to handle any input and
326	** figure out more cases than the original:
327	**
328	** fossil test-case-filename C:/ .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
329	** → Original: .//..\WINDOWS\/.//.\SYSTEM32\.\NOTEPAD.EXE
330	** → Modified: .//..\Windows\/.//.\System32\.\notepad.exe
331	**
332	** md ÄÖÜ
333	** fossil test-case-filename ./\ .\äöü\/[empty]\\/
334	** → Original: ./äöü\/[empty]\\/
335	** → Modified: .\ÄÖÜ\/[empty]\\/
336	**
337	** The function preserves slashes and backslashes: only single file or directory
338	** components without directory separators ("basenames") are converted to UTF-16
339	** using fossil_utf8_to_path(), so bypassing its slash ↔ backslash translations.
340	** Note that the original function doesn't preserve all slashes and backslashes,
341	** for example in the second example above.
342	**
343	** NOTE: As of Windows 10, version 1803, case sensitivity may be enabled on a
344	** per-directory basis, as returned by NtQueryInformationFile() with the file
345	** information class FILE_CASE_SENSITIVE_INFORMATION. So this function may be
346	** changed to act like fossil_strdup() for files located in such directories.
347	*/
348	char *win32_file_case_preferred_name(
349	const char *zBase,
350	const char *zPath
351	){
352	int cchBase = strlen(zBase);
353	int cchPath = strlen(zPath);
354	int cchBuf = cchBase + cchPath + 1;
355	int cchRes = cchPath + 1;
356	char *zBuf = fossil_malloc(cchBuf);
357	char *zRes = fossil_malloc(cchRes);
358	int i, j;
359	memcpy(zBuf,zBase,cchBase);
360	cchRes = 0;
361	if( !IS_DIRSEP(zBuf,cchBase-1) ){
362	zBuf[cchBase++]=L'/';
363	}
364	memcpy(zBuf+cchBase,zPath,cchPath+1);
365	i = j = cchBase;
366	while( 1 ){
367	WIN32_FIND_DATAW fd;
368	HANDLE hFind;
369	wchar_t *wzBuf;
370	char *zCompBuf = 0;
371	char *zComp = &zBuf[i];
372	int cchComp;
373	char chSep;
374	int fDone;
375	if( IS_DIRSEP(zBuf,i) ){
376	zRes[cchRes++] = zBuf[i];
377	i = j = i+1;
378	continue;
379	}
380	NEXT_DIRSEP(zBuf,j);
381	fDone = zBuf[j]==0;
382	chSep = zBuf[j];
383	zBuf[j] = 0; /* Truncate working buffer. */
384	wzBuf = fossil_utf8_to_path(zBuf,0);
385	hFind = FindFirstFileW(wzBuf,&fd);
386	if( hFind!= INVALID_HANDLE_VALUE ){
387	wchar_t *wzComp = fossil_utf8_to_path(zComp,0);
388	FindClose(hFind);
389	/* Test fd.cFileName, not fd.cAlternateFileName (classic 8.3 format). */
390	if( win32_compare_filenames_nocase(wzComp,fd.cFileName)==0 ){
391	zCompBuf = fossil_path_to_utf8(fd.cFileName);
392	zComp = zCompBuf;
393	}
394	fossil_path_free(wzComp);
395	}
396	fossil_path_free(wzBuf);
397	cchComp = strlen(zComp);
398	memcpy(zRes+cchRes,zComp,cchComp);
399	cchRes += cchComp;
400	if( zCompBuf ){
401	fossil_path_free(zCompBuf);
402	}
403	if( fDone ){
404	zRes[cchRes] = 0;
405	break;
406	}
407	zBuf[j] = chSep; /* Undo working buffer truncation. */
408	i = j;
409	}
410	fossil_free(zBuf);
411	return zRes;
412	}
413	#endif /* _WIN32 -- This code is for win32 only */
414

Fossil SCM

Keyboard Shortcuts