git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH 11/13] wildmatch: adjust "**" behavior

From
Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Date
Oct 15, 2012, 06:26 UTC
Message-ID
<1350282362-4505-11-git-send-email-pclouds@gmail.com>
In-Reply-To
<1350282362-4505-1-git-send-email-pclouds@gmail.com>

Standard wildmatch() sees consecutive asterisks as "*" that can also match slashes. But that may be hard to explain to users as "abc/**/def" can match "abcdef", "abcxyzdef", "abc/def", "abc/x/def", "abc/x/y/def"...

This patch changes wildmatch so that users can do
- "**/def" -> all paths ending with file/directory 'def'
- "abc/**" - equivalent to "/abc/"
- "abc/**/def" -> "abc/x/def", "abc/x/y/def"...
- otherwise consider the pattern malformed if "**" is found

Basically the magic of "**" only remains if it's wrapped around by slashes.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
---
 t/t3070-wildmatch.sh |  5 +++--
 wildmatch.c          | 13 +++++++------
 wildmatch.h          |  6 ++++++
 3 files changed, 16 insertions(+), 8 deletions(-)
diff --git a/t/t3070-wildmatch.sh b/t/t3070-wildmatch.sh
index dd95b00..15848d5 100755
--- a/t/t3070-wildmatch.sh
+++ b/t/t3070-wildmatch.sh
@@ -46,7 +46,7 @@ match 0 0 foobar 'foo\*bar'
 match 1 1 'f\oo' 'f\\oo'
 match 1 1 ball '*[al]?'
 match 0 0 ten '[ten]'
-match 1 1 ten '**[!te]'
+match 0 1 ten '**[!te]'
 match 0 0 ten '**[!ten]'
 match 1 1 ten 't[a-g]n'
 match 0 0 ten 't[!a-g]n'
@@ -61,7 +61,8 @@ match 1 1 ']' ']'
 
 # Extended slash-matching features
 match 0 0 'foo/baz/bar' 'foo*bar'
-match 1 0 'foo/baz/bar' 'foo**bar'
+match 0 0 'foo/baz/bar' 'foo**bar'
+match 0 1 'foobazbar' 'foo**bar'
 match 0 0 'foo/bar' 'foo?bar'
 match 0 0 'foo/bar' 'foo[/]bar'
 match 0 0 'foo/bar' 'f[^eiu][^eiu][^eiu][^eiu][^eiu]r'
diff --git a/wildmatch.c b/wildmatch.c
index 5469866..85bc0df 100644
--- a/wildmatch.c
+++ b/wildmatch.c
@@ -21,11 +21,6 @@ typedef unsigned char uchar;
 #define FALSE 0
 #define TRUE 1
 
-#define NOMATCH 1
-#define MATCH 0
-#define ABORT_ALL -1
-#define ABORT_TO_STARSTAR -2
-
 #define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \
 				    && *(class) == *(litmatch) \
 				    && strncmp((char*)class, litmatch, len) == 0)
@@ -90,8 +85,14 @@ static int dowild(const uchar *p, const uchar *text, int force_lower_case)
 			continue;
 		case '*':
 			if (*++p == '*') {
+				const uchar *prev_p = p - 2;
 				while (*++p == '*') {}
-				special = TRUE;
+				if ((prev_p == text || *prev_p == '/') ||
+				    (*p == '\0' || *p == '/' ||
+				     (p[0] == '\\' && p[1] == '/'))) {
+					special = TRUE;
+				} else
+					return ABORT_MALFORMED;
 			} else
 				special = FALSE;
 			if (*p == '\0') {
diff --git a/wildmatch.h b/wildmatch.h
index e974f9a..984a38c 100644
--- a/wildmatch.h
+++ b/wildmatch.h
@@ -1,3 +1,9 @@
 /* wildmatch.h */
 
+#define ABORT_MALFORMED 2
+#define NOMATCH 1
+#define MATCH 0
+#define ABORT_ALL -1
+#define ABORT_TO_STARSTAR -2
+
 int wildmatch(const char *pattern, const char *text, int flags);
-- 
1.8.0.rc0.29.g1fdd78f
Previous: Nguyễn Thái Ngọc DuyNext: Nguyễn Thái Ngọc Duy
Message 18 of 55 in “nd/attr-match-more-optim, nd/wildmatch and as/check-ignore”
  1. Nguyễn Thái Ngọc DuyOct 15, 2012
  2. 1/6 exclude: stricten a length check in EXC_FLAG_ENDSWITH caseNguyễn Thái Ngọc Duy, Oct 15, 2012
  3. 2/6 exclude: split basename matching code into a separate functionNguyễn Thái Ngọc Duy, Oct 15, 2012
  4. 3/6 exclude: fix a bug in prefix compare optimizationNguyễn Thái Ngọc Duy, Oct 15, 2012
  5. 4/6 exclude: split pathname matching code into a separate functionNguyễn Thái Ngọc Duy, Oct 15, 2012
  6. 5/6 gitignore: make pattern parsing code a separate functionNguyễn Thái Ngọc Duy, Oct 15, 2012
  7. 6/6 attr: more matching optimizations from .gitignoreNguyễn Thái Ngọc Duy, Oct 15, 2012
  8. 01/13 ctype: make sane_ctype[] const arrayNguyễn Thái Ngọc Duy, Oct 15, 2012
  9. 02/13 ctype: support iscntrl, ispunct, isxdigit and isprintNguyễn Thái Ngọc Duy, Oct 15, 2012
  10. 03/13 Import wildmatch from rsyncNguyễn Thái Ngọc Duy, Oct 15, 2012
  11. 04/13 wildmatch: remove unnecessary functionsNguyễn Thái Ngọc Duy, Oct 15, 2012
  12. 05/13 wildmatch: follow Git's coding conventionNguyễn Thái Ngọc Duy, Oct 15, 2012
  13. 06/13 Integrate wildmatch to gitNguyễn Thái Ngọc Duy, Oct 15, 2012
  14. 07/13 t3070: disable unreliable fnmatch testsNguyễn Thái Ngọc Duy, Oct 15, 2012
  15. 08/13 wildmatch: make wildmatch's return value compatible with fnmatchNguyễn Thái Ngọc Duy, Oct 15, 2012
  16. 09/13 wildmatch: remove static variable force_lower_caseNguyễn Thái Ngọc Duy, Oct 15, 2012
  17. 10/13 wildmatch: fix case-insensitive matchingNguyễn Thái Ngọc Duy, Oct 15, 2012
  18. 11/13 wildmatch: adjust "**" behaviorNguyễn Thái Ngọc Duy, Oct 15, 2012
  19. 12/13 wildmatch: make /**/ match zero or more directoriesNguyễn Thái Ngọc Duy, Oct 15, 2012
  20. 13/13 Support "**" wildcard in .gitignore and .gitattributesNguyễn Thái Ngọc Duy, Oct 15, 2012
  21. 14/13 wildmatch: fix tests that fail on Windows due to path manglingJohannes Sixt, Nov 4, 2012
  22. Nguyen Thai Ngoc DuyNov 6, 2012
  23. Johannes SixtNov 7, 2012
  24. 14/13 test-wildmatch: fix tests that fail on Windows due to path manglingNguyễn Thái Ngọc Duy, Nov 11, 2012
  25. 15/13 compat/fnmatch: fix off-by-one character class's length checkNguyễn Thái Ngọc Duy, Nov 11, 2012
  26. Johannes SixtNov 13, 2012
  27. Johannes SixtNov 20, 2012
  28. Junio C HamanoNov 11, 2012
  29. 14/13 test-wildmatch: avoid Windows path manglingNguyễn Thái Ngọc Duy, Nov 13, 2012
  30. Johannes SixtNov 13, 2012
  31. 14/13 test-wildmatch: avoid Windows path manglingJohannes Sixt, Nov 20, 2012
  32. Junio C HamanoNov 20, 2012
  33. 14/13 test-wildmatch: avoid Windows path manglingJohannes Sixt, Nov 21, 2012
  34. 01/12 dir.c: rename cryptic 'which' variable to more consistent nameNguyễn Thái Ngọc Duy, Oct 15, 2012
  35. 02/12 dir.c: rename path_excluded() to is_path_excluded()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  36. 03/12 dir.c: rename excluded_from_list() to is_excluded_from_list()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  37. 04/12 dir.c: rename excluded() to is_excluded()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  38. 05/12 dir.c: refactor is_excluded_from_list()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  39. 06/12 dir.c: refactor is_excluded()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  40. 07/12 dir.c: refactor is_path_excluded()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  41. 08/12 dir.c: keep track of where patterns came fromNguyễn Thái Ngọc Duy, Oct 15, 2012
  42. 09/12 dir.c: refactor treat_gitlinks()Nguyễn Thái Ngọc Duy, Oct 15, 2012
  43. 10/12 pathspec.c: move reusable code from builtin/add.cNguyễn Thái Ngọc Duy, Oct 15, 2012
  44. 11/12 dir.c: provide free_directory() for reclaiming dir_struct memoryNguyễn Thái Ngọc Duy, Oct 15, 2012
  45. 12/12 Add git-check-ignore sub-commandNguyễn Thái Ngọc Duy, Oct 15, 2012
  46. Junio C HamanoOct 15, 2012
  47. Nguyen Thai Ngoc DuyOct 16, 2012
  48. Adam SpiersOct 16, 2012
  49. Nguyen Thai Ngoc DuyOct 16, 2012
  50. Adam SpiersOct 16, 2012
  51. Junio C HamanoOct 16, 2012
  52. Adam SpiersDec 17, 2012
  53. t0007: fix tests on WindowsJohannes Sixt, Nov 4, 2012
  54. Jeff KingNov 8, 2012
  55. Junio C HamanoOct 15, 2012

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.