Skip to contentSkip to conversation

nd/wildmatch

PATCH In review Nguyễn Thái Ngọc Duy 7 participants 37 messages Jun 15, 2016

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v5 cover nd/wildmatch

This version splits fnmatch/wildmatch tests separately in t3070 and disables a lot more fnmatch tests. It also fixes the "cd .." in t0003 test and a comment in front of dowild(). No functional changes.

Nguyễn Thái Ngọc Duy (12):
ctype: make sane_ctype[] const array
ctype: support iscntrl, ispunct, isxdigit and isprint Import wildmatch from rsync
wildmatch: remove unnecessary functions
Integrate wildmatch to git
t3070: disable unreliable fnmatch tests
wildmatch: make wildmatch's return value compatible with fnmatch
wildmatch: remove static variable force_lower_case
wildmatch: fix case-insensitive matching
wildmatch: adjust "**" behavior
wildmatch: make /**/ match zero or more directories Support "**" wildcard in .gitignore and .gitattributes

 .gitignore                         |   1 +
 Documentation/gitignore.txt        |  19 +++
 Makefile                           |   3 +
 attr.c                             |   4 +-
 ctype.c                            |  20 +++-
 dir.c                              |   4 +-
 git-compat-util.h                  |  15 ++-
 t/t0003-attributes.sh              |  37 ++++++
 t/t3001-ls-files-others-exclude.sh |  19 +++
 t/t3070-wildmatch.sh               | 195 ++++++++++++++++++++++++++++++
 test-wildmatch.c                   |  14 +++
 wildmatch.c                        | 239 +++++++++++++++++++++++++++++++++++++
 wildmatch.h                        |   9 ++
 13 files changed, 575 insertions(+), 4 deletions(-)
 create mode 100755 t/t3070-wildmatch.sh
 create mode 100644 test-wildmatch.c
 create mode 100644 wildmatch.c
 create mode 100644 wildmatch.h
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#1 lore raw

PATCH v5 1/12 ctype: make sane_ctype[] const array

---
 ctype.c           | 2 +-
 git-compat-util.h | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
ctype.c +1 -1
index 9353271..faeaf34 100644
@@ -14,7 +14,7 @@ enum {
 	P = GIT_PATHSPEC_MAGIC  /* other non-alnum, except for ] and } */
 };
 
-unsigned char sane_ctype[256] = {
+const unsigned char sane_ctype[256] = {
 	0, 0, 0, 0, 0, 0, 0, 0, 0, S, S, 0, 0, S, 0, 0,		/*   0.. 15 */
 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,		/*  16.. 31 */
 	S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P,		/*  32.. 47 */
git-compat-util.h +1 -1
index 2fbf1fd..f8b859c 100644
@@ -510,7 +510,7 @@ extern const char tolower_trans_tbl[256];
 #undef isupper
 #undef tolower
 #undef toupper
-extern unsigned char sane_ctype[256];
+extern const unsigned char sane_ctype[256];
 #define GIT_SPACE 0x01
 #define GIT_DIGIT 0x02
 #define GIT_ALPHA 0x04
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#2 lore raw

PATCH v5 2/12 ctype: support iscntrl, ispunct, isxdigit and isprint

In reply to Nguyễn Thái Ngọc Duy
---
 ctype.c           | 18 ++++++++++++++++++
 git-compat-util.h | 13 +++++++++++++
 2 files changed, 31 insertions(+)
ctype.c +18 -0
index faeaf34..b4bf48a 100644
@@ -26,6 +26,24 @@ const unsigned char sane_ctype[256] = {
 	/* Nothing in the 128.. range */
 };
 
+enum {
+	CN = GIT_CNTRL,
+	PU = GIT_PUNCT,
+	XD = GIT_XDIGIT,
+};
+
+const unsigned char sane_ctype2[256] = {
+	CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*    0..15 */
+	CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*   16..31 */
+	0,  PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, /*   32..47 */
+	XD, XD, XD, XD, XD, XD, XD, XD, XD, XD, PU, PU, PU, PU, PU, PU, /*   48..63 */
+	PU, 0,	XD, 0,	XD, 0,	XD, 0,	0,  0,	0,  0,	0,  0,	0,  0,	/*   64..79 */
+	0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  PU, PU, PU, PU, PU, /*   80..95 */
+	PU, 0,	XD, 0,	XD, 0,	XD, 0,	0,  0,	0,  0,	0,  0,	0,  0,	/*  96..111 */
+	0,  0,	0,  0,	0,  0,	0,  0,	0,  0,	0,  PU, PU, PU, PU, CN, /* 112..127 */
+	/* Nothing in the 128.. range */
+};
+
 /* For case-insensitive kwset */
 const char tolower_trans_tbl[256] = {
 	0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
git-compat-util.h +13 -0
index f8b859c..ea11694 100644
@@ -510,14 +510,23 @@ extern const char tolower_trans_tbl[256];
 #undef isupper
 #undef tolower
 #undef toupper
+#undef iscntrl
+#undef ispunct
+#undef isxdigit
+#undef isprint
 extern const unsigned char sane_ctype[256];
+extern const unsigned char sane_ctype2[256];
 #define GIT_SPACE 0x01
 #define GIT_DIGIT 0x02
 #define GIT_ALPHA 0x04
 #define GIT_GLOB_SPECIAL 0x08
 #define GIT_REGEX_SPECIAL 0x10
 #define GIT_PATHSPEC_MAGIC 0x20
+#define GIT_CNTRL 0x01
+#define GIT_PUNCT 0x02
+#define GIT_XDIGIT 0x04
 #define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
+#define sane_istest2(x,mask) ((sane_ctype2[(unsigned char)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
 #define isspace(x) sane_istest(x,GIT_SPACE)
 #define isdigit(x) sane_istest(x,GIT_DIGIT)
@@ -527,6 +536,10 @@ extern const unsigned char sane_ctype[256];
 #define isupper(x) sane_iscase(x, 0)
 #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL)
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
+#define iscntrl(x) sane_istest2(x, GIT_CNTRL)
+#define ispunct(x) sane_istest2(x, GIT_PUNCT)
+#define isxdigit(x) sane_istest2(x, GIT_XDIGIT)
+#define isprint(x) (isalnum(x) || isspace(x) || ispunct(x))
 #define tolower(x) sane_case((unsigned char)(x), 0x20)
 #define toupper(x) sane_case((unsigned char)(x), 0)
 #define is_pathspec_magic(x) sane_istest(x,GIT_PATHSPEC_MAGIC)
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#3 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Quoted 2 lines · Nguyễn Thái Ngọc Duy

---

The description to justify why it is ctype2[] seems to have been lost. Intended?

Quoted 72 lines
 ctype.c           | 18 ++++++++++++++++++
 git-compat-util.h | 13 +++++++++++++
 2 files changed, 31 insertions(+)
ctype.c +18 -0
index faeaf34..b4bf48a 100644
@@ -26,6 +26,24 @@ const unsigned char sane_ctype[256] = {
 	/* Nothing in the 128.. range */
 };
 
+enum {
+	CN = GIT_CNTRL,
+	PU = GIT_PUNCT,
+	XD = GIT_XDIGIT,
+};
+
+const unsigned char sane_ctype2[256] = {
+	CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*    0..15 */
+	CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*   16..31 */
+	0,  PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, /*   32..47 */
+	XD, XD, XD, XD, XD, XD, XD, XD, XD, XD, PU, PU, PU, PU, PU, PU, /*   48..63 */
+	PU, 0,	XD, 0,	XD, 0,	XD, 0,	0,  0,	0,  0,	0,  0,	0,  0,	/*   64..79 */
+	0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  PU, PU, PU, PU, PU, /*   80..95 */
+	PU, 0,	XD, 0,	XD, 0,	XD, 0,	0,  0,	0,  0,	0,  0,	0,  0,	/*  96..111 */
+	0,  0,	0,  0,	0,  0,	0,  0,	0,  0,	0,  PU, PU, PU, PU, CN, /* 112..127 */
+	/* Nothing in the 128.. range */
+};
+
 /* For case-insensitive kwset */
 const char tolower_trans_tbl[256] = {
 	0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
git-compat-util.h +13 -0
index f8b859c..ea11694 100644
@@ -510,14 +510,23 @@ extern const char tolower_trans_tbl[256];
 #undef isupper
 #undef tolower
 #undef toupper
+#undef iscntrl
+#undef ispunct
+#undef isxdigit
+#undef isprint
 extern const unsigned char sane_ctype[256];
+extern const unsigned char sane_ctype2[256];
 #define GIT_SPACE 0x01
 #define GIT_DIGIT 0x02
 #define GIT_ALPHA 0x04
 #define GIT_GLOB_SPECIAL 0x08
 #define GIT_REGEX_SPECIAL 0x10
 #define GIT_PATHSPEC_MAGIC 0x20
+#define GIT_CNTRL 0x01
+#define GIT_PUNCT 0x02
+#define GIT_XDIGIT 0x04
 #define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
+#define sane_istest2(x,mask) ((sane_ctype2[(unsigned char)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
 #define isspace(x) sane_istest(x,GIT_SPACE)
 #define isdigit(x) sane_istest(x,GIT_DIGIT)
@@ -527,6 +536,10 @@ extern const unsigned char sane_ctype[256];
 #define isupper(x) sane_iscase(x, 0)
 #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL)
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
+#define iscntrl(x) sane_istest2(x, GIT_CNTRL)
+#define ispunct(x) sane_istest2(x, GIT_PUNCT)
+#define isxdigit(x) sane_istest2(x, GIT_XDIGIT)
+#define isprint(x) (isalnum(x) || isspace(x) || ispunct(x))
 #define tolower(x) sane_case((unsigned char)(x), 0x20)
 #define toupper(x) sane_case((unsigned char)(x), 0)
 #define is_pathspec_magic(x) sane_istest(x,GIT_PATHSPEC_MAGIC)
#4 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Quoted 7 lines · Junio C Hamano
Quoted 2 lines · Nguyễn Thái Ngọc Duy

---

The description to justify why it is ctype2[] seems to have been lost. Intended?

Nope. I added the description after generating patches and forgot to update the same to my branch. Thanks for catching.

Signature · 1 line
Duy
#5 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

In reply to Nguyễn Thái Ngọc Duy

Am 14.10.2012 04:35, schrieb Nguyễn Thái Ngọc Duy:

Quoted 20 lines
---
  ctype.c           | 18 ++++++++++++++++++
  git-compat-util.h | 13 +++++++++++++
  2 files changed, 31 insertions(+)
ctype.c +5 -0
index faeaf34..b4bf48a 100644
@@ -26,6 +26,24 @@ const unsigned char sane_ctype[256] = {
  	/* Nothing in the 128.. range */
  };
 
+enum {
+	CN = GIT_CNTRL,
+	PU = GIT_PUNCT,
+	XD = GIT_XDIGIT,
+};

ctype.c const unsigned char sane_ctype[256] = { · line 43

Quoted 10 lines
ctype.c +10 -0
+
+const unsigned char sane_ctype2[256] = {
+	CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*    0..15 */
+	CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*   16..31 */
+	0,  PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, /*   32..47 */
+	XD, XD, XD, XD, XD, XD, XD, XD, XD, XD, PU, PU, PU, PU, PU, PU, /*   48..63 */
+	PU, 0,	XD, 0,	XD, 0,	XD, 0,	0,  0,	0,  0,	0,  0,	0,  0,	/*   64..79 */
+	0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  PU, PU, PU, PU, PU, /*   80..95 */
+	PU, 0,	XD, 0,	XD, 0,	XD, 0,	0,  0,	0,  0,	0,  0,	0,  0,	/*  96..111 */
+	0,  0,	0,  0,	0,  0,	0,  0,	0,  0,	0,  PU, PU, PU, PU, CN, /* 112..127 */

Shouldn't [ace] (65, 67, 69) and [ACE] (97, 99, 101) be xdigits as well?

But how about using the existing hexval_table instead, like this:

	#define isxdigit(x) (hexval_table[(x)] != -1)

With that, couldn't you squeeze the other two classes into the existing sane_type?

By the way, I'm working on a patch series for implementing a lot more character classes with table lookups. It grew out of a desire to make bad_ref_char() faster but perhaps got a bit out of hand by now; it's at 24 patches and still not finished. I'm curious how long we have until it escapes. ;-)

ctype.c const unsigned char sane_ctype[256] = { · line 48

Quoted 5 lines
ctype.c +4 -0
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
+#define iscntrl(x) sane_istest2(x, GIT_CNTRL)
+#define ispunct(x) sane_istest2(x, GIT_PUNCT)
+#define isxdigit(x) sane_istest2(x, GIT_XDIGIT)
+#define isprint(x) (isalnum(x) || isspace(x) || ispunct(x))

If a single table is used, you can do with a single table lookup by adding the bits for the component classes, like isalnum and is_regex_special do.

René

#6 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Quoted 19 lines · René Scharfe
Quoted 9 lines
(quoted diff) +2 -0
+const unsigned char sane_ctype2[256] = {
+       CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*

0..15 */
+ CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, CN, /*
16..31 */
+ 0, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, PU, /*
32..47 */
+ XD, XD, XD, XD, XD, XD, XD, XD, XD, XD, PU, PU, PU, PU, PU, PU, /*
48..63 */

Quoted 8 lines

+ PU, 0, XD, 0, XD, 0, XD, 0, 0, 0, 0, 0, 0, 0, 0, 0, /*
64..79 */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, PU, PU, PU, PU, PU, /*
80..95 */
+ PU, 0, XD, 0, XD, 0, XD, 0, 0, 0, 0, 0, 0, 0, 0, 0, /*
96..111 */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, PU, PU, PU, PU, CN, /*
112..127 */

Quoted 1 line

Shouldn't [ace] (65, 67, 69) and [ACE] (97, 99, 101) be xdigits as well?

Hmm.. I generated it from LANG=C. I wonder where I got it wrong..

Quoted 6 lines

But how about using the existing hexval_table instead, like this:

        #define isxdigit(x) (hexval_table[(x)] != -1)

With that, couldn't you squeeze the other two classes into the existing sane_type?

No there are still conflicts: 9, 10 and 13 as spaces (vs controls) and 123, 124 and 126 as regex/pathspec special (vs punctuation).

Quoted 5 lines

By the way, I'm working on a patch series for implementing a lot more character classes with table lookups. It grew out of a desire to make bad_ref_char() faster but perhaps got a bit out of hand by now; it's at 24 patches and still not finished. I'm curious how long we have until it escapes. ;-)

I don't think the series is going to graduate any time soon :)

Signature · 1 line
Duy
#7 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Am 14.10.2012 15:25, schrieb Nguyen Thai Ngoc Duy:

Quoted 7 lines
Quoted 2 lines · René Scharfe

With that, couldn't you squeeze the other two classes into the existing sane_type?

No there are still conflicts: 9, 10 and 13 as spaces (vs controls) and 123, 124 and 126 as regex/pathspec special (vs punctuation).

That's not a problem, an entry in the table can have more than one bit set -- just OR them together in ctype.c. It may not look as nice, but that's OK. You could also define a character for GIT_SPACE | GIT_CNTRL etc. for cosmetic reasons.

René

#8 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Quoted 6 lines · René Scharfe

Am 14.10.2012 15:25, schrieb Nguyen Thai Ngoc Duy:

Quoted 5 lines
Quoted 2 lines · René Scharfe

With that, couldn't you squeeze the other two classes into the existing sane_type?

Quoted 7 lines
Quoted 2 lines

No there are still conflicts: 9, 10 and 13 as spaces (vs controls) and 123, 124 and 126 as regex/pathspec special (vs punctuation).

That's not a problem, an entry in the table can have more than one bit set -- just OR them together in ctype.c. It may not look as nice, but that's OK. You could also define a character for GIT_SPACE | GIT_CNTRL etc. for cosmetic reasons.

Only space chars is not a subset of control chars, which needs a new combination. So the result does not look as bad as I thought:

-- 8< --

ctype.c +8 -5
index faeaf34..0bfebb4 100644
@@ -11,18 +11,21 @@ enum {
 	D = GIT_DIGIT,
 	G = GIT_GLOB_SPECIAL,	/* *, ?, [, \\ */
 	R = GIT_REGEX_SPECIAL,	/* $, (, ), +, ., ^, {, | */
-	P = GIT_PATHSPEC_MAGIC  /* other non-alnum, except for ] and } */
+	P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */
+	X = GIT_CNTRL,
+	U = GIT_PUNCT,
+	Z = GIT_CNTRL | GIT_SPACE
 };
 
 const unsigned char sane_ctype[256] = {
-	0, 0, 0, 0, 0, 0, 0, 0, 0, S, S, 0, 0, S, 0, 0,		/*   0.. 15 */
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,		/*  16.. 31 */
+	X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X,		/*   0.. 15 */
+	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,		/*  16.. 31 */
 	S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P,		/*  32.. 47 */
 	D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G,		/*  48.. 63 */
 	P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A,		/*  64.. 79 */
-	A, A, A, A, A, A, A, A, A, A, A, G, G, 0, R, P,		/*  80.. 95 */
+	A, A, A, A, A, A, A, A, A, A, A, G, G, U, R, P,		/*  80.. 95 */
 	P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A,		/*  96..111 */
-	A, A, A, A, A, A, A, A, A, A, A, R, R, 0, P, 0,		/* 112..127 */
+	A, A, A, A, A, A, A, A, A, A, A, R, R, U, P, X,		/* 112..127 */
 	/* Nothing in the 128.. range */
 };
 
git-compat-util.h +13 -0
index f8b859c..db77f3e 100644
@@ -510,6 +510,10 @@ extern const char tolower_trans_tbl[256];
 #undef isupper
 #undef tolower
 #undef toupper
+#undef iscntrl
+#undef ispunct
+#undef isxdigit
+#undef isprint
 extern const unsigned char sane_ctype[256];
 #define GIT_SPACE 0x01
 #define GIT_DIGIT 0x02
@@ -517,6 +521,8 @@ extern const unsigned char sane_ctype[256];
 #define GIT_GLOB_SPECIAL 0x08
 #define GIT_REGEX_SPECIAL 0x10
 #define GIT_PATHSPEC_MAGIC 0x20
+#define GIT_CNTRL 0x40
+#define GIT_PUNCT 0x80
 #define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
 #define isspace(x) sane_istest(x,GIT_SPACE)
@@ -527,6 +533,13 @@ extern const unsigned char sane_ctype[256];
 #define isupper(x) sane_iscase(x, 0)
 #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL)
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
+#define iscntrl(x) (sane_istest(x,GIT_CNTRL))
+#define ispunct(x) sane_istest(x, GIT_PUNCT | GIT_REGEX_SPECIAL | \
+		GIT_GLOB_SPECIAL | GIT_PATHSPEC_MAGIC)
+#define isxdigit(x) (hexval_table[x] != -1)
+#define isprint(x) (sane_istest(x, GIT_ALPHA | GIT_DIGIT | GIT_SPACE | \
+		GIT_PUNCT | GIT_REGEX_SPECIAL | GIT_GLOB_SPECIAL | \
+		GIT_PATHSPEC_MAGIC))
 #define tolower(x) sane_case((unsigned char)(x), 0x20)
 #define toupper(x) sane_case((unsigned char)(x), 0)
 #define is_pathspec_magic(x) sane_istest(x,GIT_PATHSPEC_MAGIC)

-- 8< --

Signature · 1 line
Duy
#9 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Hi Nguyen.

I just had a need for isprint() myself, and then I found your code here.

I had a look at the POSIX locale as describe here:

http://sourceware.org/git/?p=glibc.git;a=blob;f=localedata/locales/POSIX

Some remarks below.

Am 14.10.2012 16:26, schrieb Nguyen Thai Ngoc Duy:

Quoted 11 lines

-- 8< --

ctype.c +1 -1
index faeaf34..0bfebb4 100644
@@ -11,18 +11,21 @@ enum {
 	D = GIT_DIGIT,
 	G = GIT_GLOB_SPECIAL,	/* *, ?, [, \\ */
 	R = GIT_REGEX_SPECIAL,	/* $, (, ), +, ., ^, {, | */
-	P = GIT_PATHSPEC_MAGIC  /* other non-alnum, except for ] and } */
+	P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */

ctype.c enum { · line 22

Quoted 10 lines
ctype.c +5 -2
+	X = GIT_CNTRL,
+	U = GIT_PUNCT,
+	Z = GIT_CNTRL | GIT_SPACE
 };
 
 const unsigned char sane_ctype[256] = {
-	0, 0, 0, 0, 0, 0, 0, 0, 0, S, S, 0, 0, S, 0, 0,		/*   0.. 15 */
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,		/*  16.. 31 */
+	X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X,		/*   0.. 15 */
+	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,		/*  16.. 31 */

"Normal" isspace() also includes vertical tab (11) and form-feed (12) as white-space characters. Is there a reason, why they are not included here?

Quoted 15 lines
ctype.c +2 -2
 	S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P,		/*  32.. 47 */
 	D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G,		/*  48.. 63 */
 	P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A,		/*  64.. 79 */
-	A, A, A, A, A, A, A, A, A, A, A, G, G, 0, R, P,		/*  80.. 95 */
+	A, A, A, A, A, A, A, A, A, A, A, G, G, U, R, P,		/*  80.. 95 */
 	P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A,		/*  96..111 */
-	A, A, A, A, A, A, A, A, A, A, A, R, R, 0, P, 0,		/* 112..127 */
+	A, A, A, A, A, A, A, A, A, A, A, R, R, U, P, X,		/* 112..127 */
 	/* Nothing in the 128.. range */
 };
 
git-compat-util.h +0 -0
index f8b859c..db77f3e 100644

[...]

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 542

Quoted 11 lines
git-compat-util.h +7 -0
@@ -527,6 +533,13 @@ extern const unsigned char sane_ctype[256];
 #define isupper(x) sane_iscase(x, 0)
 #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL)
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
+#define iscntrl(x) (sane_istest(x,GIT_CNTRL))
+#define ispunct(x) sane_istest(x, GIT_PUNCT | GIT_REGEX_SPECIAL | \
+		GIT_GLOB_SPECIAL | GIT_PATHSPEC_MAGIC)
+#define isxdigit(x) (hexval_table[x] != -1)
+#define isprint(x) (sane_istest(x, GIT_ALPHA | GIT_DIGIT | GIT_SPACE | \
+		GIT_PUNCT | GIT_REGEX_SPECIAL | GIT_GLOB_SPECIAL | \
+		GIT_PATHSPEC_MAGIC))

"Normal" isprint() only includes space (32) from the white-space characters. The other white-space characters are not considered printable.

Do we want to stay close to the "original", or not?

Regards
Jan

#10 lore raw

Re: [PATCH v5 02/12] ctype: support iscntrl, ispunct, isxdigit and isprint

Quoted 8 lines · Jan H. Schönherr

ctype.c enum { · line 22

Quoted 5 lines

const unsigned char sane_ctype[256] = {

ctype.c +2 -2
-     0, 0, 0, 0, 0, 0, 0, 0, 0, S, S, 0, 0, S, 0, 0,         /*   0.. 15 */
-     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,         /*  16.. 31 */
+     X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X,         /*   0.. 15 */
+     X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,         /*  16.. 31 */

"Normal" isspace() also includes vertical tab (11) and form-feed (12) as white-space characters. Is there a reason, why they are not included here?

I'm not sure. They were not classified as spaces in the very first version in 4546738 (Unlocalized isspace and friends - 2005-10-13). Maybe Linus had a reason to do so.

Quoted 8 lines

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 542

Quoted 3 lines
git-compat-util.h +3 -0
+#define isprint(x) (sane_istest(x, GIT_ALPHA | GIT_DIGIT | GIT_SPACE | \
+             GIT_PUNCT | GIT_REGEX_SPECIAL | GIT_GLOB_SPECIAL | \
+             GIT_PATHSPEC_MAGIC))

"Normal" isprint() only includes space (32) from the white-space characters. The other white-space characters are not considered printable.

Do we want to stay close to the "original", or not?

We do. I followed [1] but obvious missed the last sentence in "print"
description: "No characters specified for the keyword cntrl shall be specified". Thanks for catching. I'll fix it soon.

[1] http://pubs.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap07.html

Signature · 1 line
Duy
#11 lore raw

PATCH 1/1 Correct Git's version of isprint and isspace

In reply to Jan H. Schönherr

Git's ispace does not include 11 and 12. Git's isprint includes control space characters (10-13). According to glibc-2.14.1 on C locale on Linux, this is wrong. This patch fixes it.

---
I wrote a small C program to compare the result of all is* functions that Git replaces against the libc version. These are the only ones that differ. Which matches what Jan Schönherr commented.

 ctype.c           |  6 +++---
 git-compat-util.h | 11 ++++++-----
 2 files changed, 9 insertions(+), 8 deletions(-)
ctype.c +3 -3
index 0bfebb4..71311a3 100644
@@ -14,11 +14,11 @@ enum {
 	P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */
 	X = GIT_CNTRL,
 	U = GIT_PUNCT,
-	Z = GIT_CNTRL | GIT_SPACE
+	Z = GIT_CNTRL_SPACE
 };
 
-const unsigned char sane_ctype[256] = {
-	X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X,		/*   0.. 15 */
+const unsigned int sane_ctype[256] = {
+	X, X, X, X, X, X, X, X, X, Z, Z, Z, Z, Z, X, X,		/*   0.. 15 */
 	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,		/*  16.. 31 */
 	S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P,		/*  32.. 47 */
 	D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G,		/*  48.. 63 */
git-compat-util.h +6 -5
index 02f48f6..4ed3f94 100644
@@ -474,8 +474,8 @@ extern const char tolower_trans_tbl[256];
 #undef ispunct
 #undef isxdigit
 #undef isprint
-extern const unsigned char sane_ctype[256];
-#define GIT_SPACE 0x01
+extern const unsigned int sane_ctype[256];
+#define GIT_CNTRL_SPACE 0x01
 #define GIT_DIGIT 0x02
 #define GIT_ALPHA 0x04
 #define GIT_GLOB_SPECIAL 0x08
@@ -483,9 +483,10 @@ extern const unsigned char sane_ctype[256];
 #define GIT_PATHSPEC_MAGIC 0x20
 #define GIT_CNTRL 0x40
 #define GIT_PUNCT 0x80
-#define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
+#define GIT_SPACE 0x100
+#define sane_istest(x,mask) ((sane_ctype[(unsigned int)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
-#define isspace(x) sane_istest(x,GIT_SPACE)
+#define isspace(x) sane_istest(x,GIT_SPACE | GIT_CNTRL_SPACE)
 #define isdigit(x) sane_istest(x,GIT_DIGIT)
 #define isalpha(x) sane_istest(x,GIT_ALPHA)
 #define isalnum(x) sane_istest(x,GIT_ALPHA | GIT_DIGIT)
@@ -493,7 +494,7 @@ extern const unsigned char sane_ctype[256];
 #define isupper(x) sane_iscase(x, 0)
 #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL)
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
-#define iscntrl(x) (sane_istest(x,GIT_CNTRL))
+#define iscntrl(x) (sane_istest(x,GIT_CNTRL | GIT_CNTRL_SPACE))
 #define ispunct(x) sane_istest(x, GIT_PUNCT | GIT_REGEX_SPECIAL | \
 		GIT_GLOB_SPECIAL | GIT_PATHSPEC_MAGIC)
 #define isxdigit(x) (hexval_table[x] != -1)
Signature · 1 line
1.8.0.rc2.23.g1fb49df
#12 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

Hi.

Am 13.11.2012 11:46, schrieb Nguyễn Thái Ngọc Duy:

Quoted 23 lines

Git's ispace does not include 11 and 12. Git's isprint includes control space characters (10-13). According to glibc-2.14.1 on C locale on Linux, this is wrong. This patch fixes it.

---
I wrote a small C program to compare the result of all is* functions that Git replaces against the libc version. These are the only ones that differ. Which matches what Jan Schönherr commented.

 ctype.c           |  6 +++---
 git-compat-util.h | 11 ++++++-----
 2 files changed, 9 insertions(+), 8 deletions(-)
ctype.c +0 -1
index 0bfebb4..71311a3 100644
@@ -14,11 +14,11 @@ enum {
 	P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */
 	X = GIT_CNTRL,
 	U = GIT_PUNCT,
-	Z = GIT_CNTRL | GIT_SPACE

ctype.c enum { · line 24

Quoted 10 lines
ctype.c +3 -2
+	Z = GIT_CNTRL_SPACE
 };
 
-const unsigned char sane_ctype[256] = {
-	X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X,		/*   0.. 15 */
+const unsigned int sane_ctype[256] = {
+	X, X, X, X, X, X, X, X, X, Z, Z, Z, Z, Z, X, X,		/*   0.. 15 */
 	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,		/*  16.. 31 */
 	S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P,		/*  32.. 47 */
 	D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G,		/*  48.. 63 */

An alternative to switching from 1-byte to 4-byte values (don't we have a 2-byte datatype?), would be to free up GIT_CNTRL and simply do:

#define iscntrl(x) ((x) < 0x20)

Quoted 4 lines
git-compat-util.h +0 -0
index 02f48f6..4ed3f94 100644

[...]

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 487

Quoted 7 lines
git-compat-util.h +2 -1
@@ -483,9 +483,10 @@ extern const unsigned char sane_ctype[256];
 #define GIT_PATHSPEC_MAGIC 0x20
 #define GIT_CNTRL 0x40
 #define GIT_PUNCT 0x80
-#define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
+#define GIT_SPACE 0x100
+#define sane_istest(x,mask) ((sane_ctype[(unsigned int)(x)] & (mask)) != 0)

That should better be left "(unsigned char)"? We might access values after the array otherwise.

(That said, it wasn't really correct before either, when there really is a possibility that x >= 0x100.)

Regards
Jan

PS: It looks like my isprint() version was given precedence over your isprint() version during the merge into next. That should also be sorted out, but I've no idea which one is actually better: two comparisons versus one cache lookup and a bitop... (though my guess is that comparisons are cheaper, but then we should also convert isdigit()...)

#13 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

In reply to Nguyễn Thái Ngọc Duy

Am 13.11.2012 11:46, schrieb Nguyễn Thái Ngọc Duy:

Quoted 3 lines

Git's isprint includes
control space characters (10-13). According to glibc-2.14.1 on C locale on Linux, this is wrong. This patch fixes it.

isprint() is not in master, yet. Can we perhaps still introduce it in such a way that we never have an incorrect version in master's history?

And could you please update test-ctype.c to match the change to isspace()? The tests there just documented the status quo before I made changes to ctype.c long ago, so it's definitions are just as correct (or wrong) as the original implementation.

Thanks,
René

#14 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

In reply to Nguyễn Thái Ngọc Duy

Am 13.11.2012 11:46, schrieb Nguyễn Thái Ngọc Duy:

Quoted 2 lines

Git's ispace does not include 11 and 12. [...] According to glibc-2.14.1 on C locale on Linux, this is wrong.

11 and 12 being vertical tab (\v) and form-feed (\f). This lack goes back to the introduction of git's own character classifier macros seven years ago in 4546738b (Unlocalized isspace and friends).

Linus, do you remember if you left them out on purpose?

Thanks,
René

#15 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

Quoted 2 lines · René Scharfe

Linus, do you remember if you left them out on purpose?

Umm, no.

I have to wonder why you care? As far as I'm concerned, the only valid space is space, TAB and CR/LF.

Anything else is *noise*, not space. What's the reason for even caring?

Linus

#16 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

Quoted 5 lines · Linus Torvalds

I have to wonder why you care? As far as I'm concerned, the only valid space is space, TAB and CR/LF.

Anything else is *noise*, not space. What's the reason for even caring?

Btw, expanding the whitespace selection may actually be very counter-productive. It is used primarily for things like removing extraneous space at the end of lines etc, and for that, the current selection of SPACE, TAB and LF/CR is the right thing to do.

Adding things like FF etc - that are *technically* whitespace, but aren't the normal kind of silent whitespace - is potentially going to change things too much. People might *want* a form-feed in their messages, for all we know.

So I really object to changing things "just because". There's a reason we do our own ctype.c: it avoids the crazy crap. It avoids the idiotic localization issues, and it avoids the ambiguous cases.

So just let it be, unless you have some major real reason to actually care about a real-world case. And if you do, please explain it. Don't change things just because.

Linus

#17 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

Am 13.11.2012 20:50, schrieb Linus Torvalds:

Quoted 13 lines
Quoted 5 lines · Linus Torvalds

I have to wonder why you care? As far as I'm concerned, the only valid space is space, TAB and CR/LF.

Anything else is *noise*, not space. What's the reason for even caring?

Btw, expanding the whitespace selection may actually be very counter-productive. It is used primarily for things like removing extraneous space at the end of lines etc, and for that, the current selection of SPACE, TAB and LF/CR is the right thing to do.

Quoted 4 lines

Adding things like FF etc - that are *technically* whitespace, but aren't the normal kind of silent whitespace - is potentially going to change things too much. People might *want* a form-feed in their messages, for all we know.

The patch was motivated by the integration of the wildmatch library, which exposes named character classes to users. It replaces a call of fnmatch in match_pathname. Users probably expect [:space:] to mean the same in git as in other programs.

I never saw a vertical tab and I can't imagine what it's used for. I'd expect form-feeds to be matched as space, though. Didn't see them very often, admittedly.

Nevertheless, it's unfortunate that we have an isspace() that *almost* does what the widely known thing of the same name does. I'd shy away from changing git's version directly, because it's used more than a hundred times in the code, and estimating the impact of adding \v and \f to it. Perhaps renaming it to isgitspace() is a good first step, followed by adding a "standard" version of isspace() for wildmatch?

René

#18 lore raw

Re: [PATCH nd/wildmatch] Correct Git's version of isprint and isspace

In reply to Nguyễn Thái Ngọc Duy

Am 13.11.2012 11:46, schrieb Nguyễn Thái Ngọc Duy:

Quoted 44 lines
ctype.c +3 -3
@@ -14,11 +14,11 @@ enum {
 	P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */
 	X = GIT_CNTRL,
 	U = GIT_PUNCT,
-	Z = GIT_CNTRL | GIT_SPACE
+	Z = GIT_CNTRL_SPACE
 };
 
-const unsigned char sane_ctype[256] = {
-	X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X,		/*   0.. 15 */
+const unsigned int sane_ctype[256] = {
+	X, X, X, X, X, X, X, X, X, Z, Z, Z, Z, Z, X, X,		/*   0.. 15 */
 	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,		/*  16.. 31 */
 	S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P,		/*  32.. 47 */
 	D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G,		/*  48.. 63 */
git-compat-util.h +5 -4
index 02f48f6..4ed3f94 100644
@@ -474,8 +474,8 @@ extern const char tolower_trans_tbl[256];
 #undef ispunct
 #undef isxdigit
 #undef isprint
-extern const unsigned char sane_ctype[256];
-#define GIT_SPACE 0x01
+extern const unsigned int sane_ctype[256];
+#define GIT_CNTRL_SPACE 0x01
 #define GIT_DIGIT 0x02
 #define GIT_ALPHA 0x04
 #define GIT_GLOB_SPECIAL 0x08
@@ -483,9 +483,10 @@ extern const unsigned char sane_ctype[256];
 #define GIT_PATHSPEC_MAGIC 0x20
 #define GIT_CNTRL 0x40
 #define GIT_PUNCT 0x80
-#define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
+#define GIT_SPACE 0x100
+#define sane_istest(x,mask) ((sane_ctype[(unsigned int)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
-#define isspace(x) sane_istest(x,GIT_SPACE)
+#define isspace(x) sane_istest(x,GIT_SPACE | GIT_CNTRL_SPACE)
 #define isdigit(x) sane_istest(x,GIT_DIGIT)
 #define isalpha(x) sane_istest(x,GIT_ALPHA)
 #define isalnum(x) sane_istest(x,GIT_ALPHA | GIT_DIGIT)

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 500

Quoted 8 lines
git-compat-util.h +1 -1
@@ -493,7 +494,7 @@ extern const unsigned char sane_ctype[256];
 #define isupper(x) sane_iscase(x, 0)
 #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL)
 #define is_regex_special(x) sane_istest(x,GIT_GLOB_SPECIAL | GIT_REGEX_SPECIAL)
-#define iscntrl(x) (sane_istest(x,GIT_CNTRL))
+#define iscntrl(x) (sane_istest(x,GIT_CNTRL | GIT_CNTRL_SPACE))
 #define ispunct(x) sane_istest(x, GIT_PUNCT | GIT_REGEX_SPECIAL | \
 		GIT_GLOB_SPECIAL | GIT_PATHSPEC_MAGIC)
 #define isxdigit(x) (hexval_table[x] != -1)

So we have two properties that overlap:

SSSSSSSSSS
CCCCCCCC

You seem to generate partions:

XXXYYYYYZZZZZ

then assign individual bits to each partition. Now each entry in the lookup table has only one bit set. Then you define isxxx() to check for one of the two possible bits:

iscntrl is X or Y
isspace is Y or Z

But shouldn't you just assign one bit for S and another one for C, have entries in the lookup table with more than one bit set, and check for only one bit in the isxxx macro?

That way you don't run out of bits as easily as you do with this patch.

-- Hannes

#19 lore raw

PATCH 1/1 wildmatch: correct isprint and isspace

In reply to Nguyễn Thái Ngọc Duy

Current isprint() incorrectly includes control characters 9, 10 and 13, which is fixed by this patch.

Current isspace() lacks 11 and 12. But Git's isspace() has been designed this way since the beginning and has over 100 call sites relying on this. Instead of updating isspace() behavior (which could be tricky as patches from other topics may come in parallel that assume the old isspace()), a new isspace_posix() is introduced and used by wildmatch.c. Other part of Git can be converted to use this new function if it seems appropriate.

---
Sorry for the late response. I'll reply to everybody in one mail.

Quoted 4 lines · Jan H. Schönherr

An alternative to switching from 1-byte to 4-byte values (don't we have a 2-byte datatype?), would be to free up GIT_CNTRL and simply do:

#define iscntrl(x) ((x) < 0x20)

No. 127 is also a control character.

Quoted 13 lines · Johannes Sixt

So we have two properties that overlap:

SSSSSSSSSS
CCCCCCCC

You seem to generate partions:

XXXYYYYYZZZZZ

then assign individual bits to each partition. Now each entry in the lookup table has only one bit set. Then you define isxxx() to check for one of the two possible bits:

Quoted 8 lines

iscntrl is X or Y
isspace is Y or Z

But shouldn't you just assign one bit for S and another one for C, have entries in the lookup table with more than one bit set, and check for only one bit in the isxxx macro?

That way you don't run out of bits as easily as you do with this patch.

I need three sets of characters actually: control, spaces and printable (which contains non-control spaces). Making it (isspace(x) && (x) >= 32) is simpler and because isprint() is only used in wildmatch, I don't need to think about performance penalty (yet).

Quoted 6 lines · René Scharfe

Nevertheless, it's unfortunate that we have an isspace() that *almost* does what the widely known thing of the same name does. I'd shy away from changing git's version directly, because it's used more than a hundred times in the code, and estimating the impact of adding \v and \f to it. Perhaps renaming it to isgitspace() is a good first step, followed by adding a "standard" version of isspace() for wildmatch?

There are just too many call sites of isspace() and there is a risk of new call sites coming in independently. So I think keeping isspace() as-is and using a different name for the standard version is probably a better choice.

As the new isspace_posix() is only used by wildmatch, its performance as of now is not critical and a simple macro like in this patch is probably enough. We can optimize it later if we need to.

 git-compat-util.h | 4 +++-
 wildmatch.c       | 8 ++------
 2 files changed, 5 insertions(+), 7 deletions(-)
git-compat-util.h +3 -1
index 02f48f6..d4c3fda 100644
@@ -486,6 +486,7 @@ extern const unsigned char sane_ctype[256];
 #define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
 #define isspace(x) sane_istest(x,GIT_SPACE)
+#define isspace_posix(x) (((x) >= 9 && (x) <= 13) || (x) == 32)
 #define isdigit(x) sane_istest(x,GIT_DIGIT)
 #define isalpha(x) sane_istest(x,GIT_ALPHA)
 #define isalnum(x) sane_istest(x,GIT_ALPHA | GIT_DIGIT)
@@ -499,7 +500,8 @@ extern const unsigned char sane_ctype[256];
 #define isxdigit(x) (hexval_table[x] != -1)
 #define isprint(x) (sane_istest(x, GIT_ALPHA | GIT_DIGIT | GIT_SPACE | \
 		GIT_PUNCT | GIT_REGEX_SPECIAL | GIT_GLOB_SPECIAL | \
-		GIT_PATHSPEC_MAGIC))
+		GIT_PATHSPEC_MAGIC) && \
+		(x) >= 32)
 #define tolower(x) sane_case((unsigned char)(x), 0x20)
 #define toupper(x) sane_case((unsigned char)(x), 0)
 #define is_pathspec_magic(x) sane_istest(x,GIT_PATHSPEC_MAGIC)
wildmatch.c +2 -6
index 3972e26..fd74efd 100644
@@ -37,11 +37,7 @@ typedef unsigned char uchar;
 # define ISBLANK(c) ((c) == ' ' || (c) == '\t')
 #endif
 
-#ifdef isgraph
-# define ISGRAPH(c) (ISASCII(c) && isgraph(c))
-#else
-# define ISGRAPH(c) (ISASCII(c) && isprint(c) && !isspace(c))
-#endif
+#define ISGRAPH(c) (ISASCII(c) && isprint(c) && !isspace_posix(c))
 
 #define ISPRINT(c) (ISASCII(c) && isprint(c))
 #define ISDIGIT(c) (ISASCII(c) && isdigit(c))
@@ -50,7 +46,7 @@ typedef unsigned char uchar;
 #define ISCNTRL(c) (ISASCII(c) && iscntrl(c))
 #define ISLOWER(c) (ISASCII(c) && islower(c))
 #define ISPUNCT(c) (ISASCII(c) && ispunct(c))
-#define ISSPACE(c) (ISASCII(c) && isspace(c))
+#define ISSPACE(c) (ISASCII(c) && isspace_posix(c))
 #define ISUPPER(c) (ISASCII(c) && isupper(c))
 #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
 
Signature · 1 line
1.8.0.rc2.23.g1fb49df
#20 lore raw

Re: [PATCH] wildmatch: correct isprint and isspace

Am 15.11.2012 13:19, schrieb Nguyễn Thái Ngọc Duy:

Quoted 8 lines · René Scharfe
Quoted 6 lines

Nevertheless, it's unfortunate that we have an isspace() that *almost* does what the widely known thing of the same name does. I'd shy away from changing git's version directly, because it's used more than a hundred times in the code, and estimating the impact of adding \v and \f to it. Perhaps renaming it to isgitspace() is a good first step, followed by adding a "standard" version of isspace() for wildmatch?

Quoted 4 lines

There are just too many call sites of isspace() and there is a risk of new call sites coming in independently. So I think keeping isspace() as-is and using a different name for the standard version is probably a better choice.

After having a closer look, where wildmatch is actually used -- matching filenames -- and I've not yet seen \v or \f in a filename, it's possibly unnecessary to do anything about isspace() right now.

(It's probably more an issue that filenames can be localized, and we only support unlocalized character classes.)

Quoted 13 lines
git-compat-util.h +1 -0
index 02f48f6..d4c3fda 100644
@@ -486,6 +486,7 @@ extern const unsigned char sane_ctype[256];
 #define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0)
 #define isascii(x) (((x) & ~0x7f) == 0)
 #define isspace(x) sane_istest(x,GIT_SPACE)
+#define isspace_posix(x) (((x) >= 9 && (x) <= 13) || (x) == 32)
 #define isdigit(x) sane_istest(x,GIT_DIGIT)
 #define isalpha(x) sane_istest(x,GIT_ALPHA)
 #define isalnum(x) sane_istest(x,GIT_ALPHA | GIT_DIGIT)

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 500

Quoted 1 line
git-compat-util.h +0 -0
@@ -499,7 +500,8 @@ extern const unsigned char sane_ctype[256];
 #define isxdigit(x) (hexval_table[x] != -1)

This was from a previous patch, but maybe: "hexval_table[(unsigned char)x]"

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 504

Quoted 5 lines
git-compat-util.h +2 -1
 #define isprint(x) (sane_istest(x, GIT_ALPHA | GIT_DIGIT | GIT_SPACE | \
 		GIT_PUNCT | GIT_REGEX_SPECIAL | GIT_GLOB_SPECIAL | \
-		GIT_PATHSPEC_MAGIC))
+		GIT_PATHSPEC_MAGIC) && \
+		(x) >= 32)

May I suggest the current is_print() implementation in master:

#define isprint(x) ((x) >= 0x20 && (x) <= 0x7e)

To summarize my opinion:

I no longer see a reason to correct isspace() (unless somebody with an actual use case complains), and a more POSIXly isprint() is already in master.

=> Nothing to do. :)

Regards
Jan

#21 lore raw

Re: [PATCH] wildmatch: correct isprint and isspace

Quoted 11 lines · Jan H. Schönherr

git-compat-util.h extern const unsigned char sane_ctype[256]; · line 504

Quoted 5 lines
 #define isprint(x) (sane_istest(x, GIT_ALPHA | GIT_DIGIT | GIT_SPACE | \
              GIT_PUNCT | GIT_REGEX_SPECIAL | GIT_GLOB_SPECIAL | \
git-compat-util.h +2 -1
-             GIT_PATHSPEC_MAGIC))
+             GIT_PATHSPEC_MAGIC) && \
+             (x) >= 32)

May I suggest the current is_print() implementation in master:

#define isprint(x) ((x) >= 0x20 && (x) <= 0x7e)

Quoted 6 lines

To summarize my opinion:

I no longer see a reason to correct isspace() (unless somebody with an actual use case complains), and a more POSIXly isprint() is already in master.

=> Nothing to do. :)

Yeah. I remember to remind myself to check "the implementation in master" you mentioned but I probably failed at that. Just checked that isprint() is already in master, and your comment about isspace() use in wildmatch.c makes sense too. So I'm all for doing nothing.

Signature · 1 line
Duy
#22 lore raw

PATCH v5 3/12 Import wildmatch from rsync

In reply to Nguyễn Thái Ngọc Duy

These files are from rsync.git commit f92f5b166e3019db42bc7fe1aa2f1a9178cd215d, which was the last commit before rsync turned GPL-3. All files are imported as-is and no-op. Adaptation is done in a separate patch.

rsync.git           ->  git.git
lib/wildmatch.[ch]      wildmatch.[ch]
wildtest.txt            t/t3070/wildtest.txt
---
 t/t3070/wildtest.txt | 165 +++++++++++++++++++++++
 wildmatch.c          | 368 +++++++++++++++++++++++++++++++++++++++++++++++++++
 wildmatch.h          |   6 +
 3 files changed, 539 insertions(+)
 create mode 100644 t/t3070/wildtest.txt
 create mode 100644 wildmatch.c
 create mode 100644 wildmatch.h
t/t3070/wildtest.txt new file +165 -0
new file mode 100644
index 0000000..42c1678
@@ -0,0 +1,165 @@
+# Input is in the following format (all items white-space separated):
+#
+# The first two items are 1 or 0 indicating if the wildmat call is expected to
+# succeed and if fnmatch works the same way as wildmat, respectively.  After
+# that is a text string for the match, and a pattern string.  Strings can be
+# quoted (if desired) in either double or single quotes, as well as backticks.
+#
+# MATCH FNMATCH_SAME "text to match" 'pattern to use'
+
+# Basic wildmat features
+1 1 foo			foo
+0 1 foo			bar
+1 1 ''			""
+1 1 foo			???
+0 1 foo			??
+1 1 foo			*
+1 1 foo			f*
+0 1 foo			*f
+1 1 foo			*foo*
+1 1 foobar		*ob*a*r*
+1 1 aaaaaaabababab	*ab
+1 1 foo*		foo\*
+0 1 foobar		foo\*bar
+1 1 f\oo		f\\oo
+1 1 ball		*[al]?
+0 1 ten			[ten]
+1 1 ten			**[!te]
+0 1 ten			**[!ten]
+1 1 ten			t[a-g]n
+0 1 ten			t[!a-g]n
+1 1 ton			t[!a-g]n
+1 1 ton			t[^a-g]n
+1 1 a]b			a[]]b
+1 1 a-b			a[]-]b
+1 1 a]b			a[]-]b
+0 1 aab			a[]-]b
+1 1 aab			a[]a-]b
+1 1 ]			]
+
+# Extended slash-matching features
+0 1 foo/baz/bar		foo*bar
+1 1 foo/baz/bar		foo**bar
+0 1 foo/bar		foo?bar
+0 1 foo/bar		foo[/]bar
+0 1 foo/bar		f[^eiu][^eiu][^eiu][^eiu][^eiu]r
+1 1 foo-bar		f[^eiu][^eiu][^eiu][^eiu][^eiu]r
+0 1 foo			**/foo
+1 1 /foo		**/foo
+1 1 bar/baz/foo		**/foo
+0 1 bar/baz/foo		*/foo
+0 0 foo/bar/baz		**/bar*
+1 1 deep/foo/bar/baz	**/bar/*
+0 1 deep/foo/bar/baz/	**/bar/*
+1 1 deep/foo/bar/baz/	**/bar/**
+0 1 deep/foo/bar	**/bar/*
+1 1 deep/foo/bar/	**/bar/**
+1 1 foo/bar/baz		**/bar**
+1 1 foo/bar/baz/x	*/bar/**
+0 0 deep/foo/bar/baz/x	*/bar/**
+1 1 deep/foo/bar/baz/x	**/bar/*/*
+
+# Various additional tests
+0 1 acrt		a[c-c]st
+1 1 acrt		a[c-c]rt
+0 1 ]			[!]-]
+1 1 a			[!]-]
+0 1 ''			\
+0 1 \			\
+0 1 /\			*/\
+1 1 /\			*/\\
+1 1 foo			foo
+1 1 @foo		@foo
+0 1 foo			@foo
+1 1 [ab]		\[ab]
+1 1 [ab]		[[]ab]
+1 1 [ab]		[[:]ab]
+0 1 [ab]		[[::]ab]
+1 1 [ab]		[[:digit]ab]
+1 1 [ab]		[\[:]ab]
+1 1 ?a?b		\??\?b
+1 1 abc			\a\b\c
+0 1 foo			''
+1 1 foo/bar/baz/to	**/t[o]
+
+# Character class tests
+1 1 a1B		[[:alpha:]][[:digit:]][[:upper:]]
+0 1 a		[[:digit:][:upper:][:space:]]
+1 1 A		[[:digit:][:upper:][:space:]]
+1 1 1		[[:digit:][:upper:][:space:]]
+0 1 1		[[:digit:][:upper:][:spaci:]]
+1 1 ' '		[[:digit:][:upper:][:space:]]
+0 1 .		[[:digit:][:upper:][:space:]]
+1 1 .		[[:digit:][:punct:][:space:]]
+1 1 5		[[:xdigit:]]
+1 1 f		[[:xdigit:]]
+1 1 D		[[:xdigit:]]
+1 1 _		[[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]
+#1 1 �		[^[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]
+1 1 		[^[:alnum:][:alpha:][:blank:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]
+1 1 .		[^[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:lower:][:space:][:upper:][:xdigit:]]
+1 1 5		[a-c[:digit:]x-z]
+1 1 b		[a-c[:digit:]x-z]
+1 1 y		[a-c[:digit:]x-z]
+0 1 q		[a-c[:digit:]x-z]
+
+# Additional tests, including some malformed wildmats
+1 1 ]		[\\-^]
+0 1 [		[\\-^]
+1 1 -		[\-_]
+1 1 ]		[\]]
+0 1 \]		[\]]
+0 1 \		[\]]
+0 1 ab		a[]b
+0 1 a[]b	a[]b
+0 1 ab[		ab[
+0 1 ab		[!
+0 1 ab		[-
+1 1 -		[-]
+0 1 -		[a-
+0 1 -		[!a-
+1 1 -		[--A]
+1 1 5		[--A]
+1 1 ' '		'[ --]'
+1 1 $		'[ --]'
+1 1 -		'[ --]'
+0 1 0		'[ --]'
+1 1 -		[---]
+1 1 -		[------]
+0 1 j		[a-e-n]
+1 1 -		[a-e-n]
+1 1 a		[!------]
+0 1 [		[]-a]
+1 1 ^		[]-a]
+0 1 ^		[!]-a]
+1 1 [		[!]-a]
+1 1 ^		[a^bc]
+1 1 -b]		[a-]b]
+0 1 \		[\]
+1 1 \		[\\]
+0 1 \		[!\\]
+1 1 G		[A-\\]
+0 1 aaabbb	b*a
+0 1 aabcaa	*ba*
+1 1 ,		[,]
+1 1 ,		[\\,]
+1 1 \		[\\,]
+1 1 -		[,-.]
+0 1 +		[,-.]
+0 1 -.]		[,-.]
+1 1 2		[\1-\3]
+1 1 3		[\1-\3]
+0 1 4		[\1-\3]
+1 1 \		[[-\]]
+1 1 [		[[-\]]
+1 1 ]		[[-\]]
+0 1 -		[[-\]]
+
+# Test recursion and the abort code (use "wildtest -i" to see iteration counts)
+1 1 -adobe-courier-bold-o-normal--12-120-75-75-m-70-iso8859-1	-*-*-*-*-*-*-12-*-*-*-m-*-*-*
+0 1 -adobe-courier-bold-o-normal--12-120-75-75-X-70-iso8859-1	-*-*-*-*-*-*-12-*-*-*-m-*-*-*
+0 1 -adobe-courier-bold-o-normal--12-120-75-75-/-70-iso8859-1	-*-*-*-*-*-*-12-*-*-*-m-*-*-*
+1 1 /adobe/courier/bold/o/normal//12/120/75/75/m/70/iso8859/1	/*/*/*/*/*/*/12/*/*/*/m/*/*/*
+0 1 /adobe/courier/bold/o/normal//12/120/75/75/X/70/iso8859/1	/*/*/*/*/*/*/12/*/*/*/m/*/*/*
+1 1 abcd/abcdefg/abcdefghijk/abcdefghijklmnop.txt		**/*a*b*g*n*t
+0 1 abcd/abcdefg/abcdefghijk/abcdefghijklmnop.txtz		**/*a*b*g*n*t
wildmatch.c new file +368 -0

368 lines not on this page · raw message

wildmatch.h new file +6 -0
new file mode 100644
index 0000000..e7f1a35
@@ -0,0 +1,6 @@
+/* wildmatch.h */
+
+int wildmatch(const char *pattern, const char *text);
+int iwildmatch(const char *pattern, const char *text);
+int wildmatch_array(const char *pattern, const char*const *texts, int where);
+int litmatch_array(const char *string, const char*const *texts, int where);
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#23 lore raw

PATCH v5 4/12 wildmatch: remove unnecessary functions

In reply to Nguyễn Thái Ngọc Duy
---
 wildmatch.c | 164 ++++--------------------------------------------------------
 wildmatch.h |   2 -
 2 files changed, 10 insertions(+), 156 deletions(-)
wildmatch.c +10 -154
index f3a1731..fae7397 100644
@@ -53,33 +53,18 @@
 #define ISUPPER(c) (ISASCII(c) && isupper(c))
 #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
 
-#ifdef WILD_TEST_ITERATIONS
-int wildmatch_iteration_count;
-#endif
-
 static int force_lower_case = 0;
 
-/* Match pattern "p" against the a virtually-joined string consisting
- * of "text" and any strings in array "a". */
-static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
+/* Match pattern "p" against "text" */
+static int dowild(const uchar *p, const uchar *text)
 {
     uchar p_ch;
 
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count++;
-#endif
-
     for ( ; (p_ch = *p) != '\0'; text++, p++) {
 	int matched, special;
 	uchar t_ch, prev_ch;
-	while ((t_ch = *text) == '\0') {
-	    if (*a == NULL) {
-		if (p_ch != '*')
-		    return ABORT_ALL;
-		break;
-	    }
-	    text = *a++;
-	}
+	if ((t_ch = *text) == '\0' && p_ch != '*')
+		return ABORT_ALL;
 	if (force_lower_case && ISUPPER(t_ch))
 	    t_ch = tolower(t_ch);
 	switch (p_ch) {
@@ -107,21 +92,15 @@ static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
 		/* Trailing "**" matches everything.  Trailing "*" matches
 		 * only if there are no more slash characters. */
 		if (!special) {
-		    do {
 			if (strchr((char*)text, '/') != NULL)
 			    return FALSE;
-		    } while ((text = *a++) != NULL);
 		}
 		return TRUE;
 	    }
 	    while (1) {
-		if (t_ch == '\0') {
-		    if ((text = *a++) == NULL)
-			break;
-		    t_ch = *text;
-		    continue;
-		}
-		if ((matched = dowild(p, text, a)) != FALSE) {
+		if (t_ch == '\0')
+		    break;
+		if ((matched = dowild(p, text)) != FALSE) {
 		    if (!special || matched != ABORT_TO_STARSTAR)
 			return matched;
 		} else if (!special && t_ch == '/')
@@ -225,144 +204,21 @@ static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
 	}
     }
 
-    do {
-	if (*text)
-	    return FALSE;
-    } while ((text = *a++) != NULL);
-
-    return TRUE;
-}
-
-/* Match literal string "s" against the a virtually-joined string consisting
- * of "text" and any strings in array "a". */
-static int doliteral(const uchar *s, const uchar *text, const uchar*const *a)
-{
-    for ( ; *s != '\0'; text++, s++) {
-	while (*text == '\0') {
-	    if ((text = *a++) == NULL)
-		return FALSE;
-	}
-	if (*text != *s)
-	    return FALSE;
-    }
-
-    do {
-	if (*text)
-	    return FALSE;
-    } while ((text = *a++) != NULL);
-
-    return TRUE;
-}
-
-/* Return the last "count" path elements from the concatenated string.
- * We return a string pointer to the start of the string, and update the
- * array pointer-pointer to point to any remaining string elements. */
-static const uchar *trailing_N_elements(const uchar*const **a_ptr, int count)
-{
-    const uchar*const *a = *a_ptr;
-    const uchar*const *first_a = a;
-
-    while (*a)
-	    a++;
-
-    while (a != first_a) {
-	const uchar *s = *--a;
-	s += strlen((char*)s);
-	while (--s >= *a) {
-	    if (*s == '/' && !--count) {
-		*a_ptr = a+1;
-		return s+1;
-	    }
-	}
-    }
-
-    if (count == 1) {
-	*a_ptr = a+1;
-	return *a;
-    }
-
-    return NULL;
+    return *text ? FALSE : TRUE;
 }
 
 /* Match the "pattern" against the "text" string. */
 int wildmatch(const char *pattern, const char *text)
 {
-    static const uchar *nomore[1]; /* A NULL pointer. */
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count = 0;
-#endif
-    return dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE;
+    return dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
 }
 
 /* Match the "pattern" against the forced-to-lower-case "text" string. */
 int iwildmatch(const char *pattern, const char *text)
 {
-    static const uchar *nomore[1]; /* A NULL pointer. */
     int ret;
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count = 0;
-#endif
     force_lower_case = 1;
-    ret = dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE;
+    ret = dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
     force_lower_case = 0;
     return ret;
 }
-
-/* Match pattern "p" against the a virtually-joined string consisting
- * of all the pointers in array "texts" (which has a NULL pointer at the
- * end).  The int "where" can be 0 (normal matching), > 0 (match only
- * the trailing N slash-separated filename components of "texts"), or < 0
- * (match the "pattern" at the start or after any slash in "texts"). */
-int wildmatch_array(const char *pattern, const char*const *texts, int where)
-{
-    const uchar *p = (const uchar*)pattern;
-    const uchar*const *a = (const uchar*const*)texts;
-    const uchar *text;
-    int matched;
-
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count = 0;
-#endif
-
-    if (where > 0)
-	text = trailing_N_elements(&a, where);
-    else
-	text = *a++;
-    if (!text)
-	return FALSE;
-
-    if ((matched = dowild(p, text, a)) != TRUE && where < 0
-     && matched != ABORT_ALL) {
-	while (1) {
-	    if (*text == '\0') {
-		if ((text = (uchar*)*a++) == NULL)
-		    return FALSE;
-		continue;
-	    }
-	    if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE
-	     && matched != ABORT_TO_STARSTAR)
-		break;
-	}
-    }
-    return matched == TRUE;
-}
-
-/* Match literal string "s" against the a virtually-joined string consisting
- * of all the pointers in array "texts" (which has a NULL pointer at the
- * end).  The int "where" can be 0 (normal matching), or > 0 (match
- * only the trailing N slash-separated filename components of "texts"). */
-int litmatch_array(const char *string, const char*const *texts, int where)
-{
-    const uchar *s = (const uchar*)string;
-    const uchar*const *a = (const uchar* const*)texts;
-    const uchar *text;
-
-    if (where > 0)
-	text = trailing_N_elements(&a, where);
-    else
-	text = *a++;
-    if (!text)
-	return FALSE;
-
-    return doliteral(s, text, a) == TRUE;
-}
wildmatch.h +0 -2
index e7f1a35..562faa3 100644
@@ -2,5 +2,3 @@
 
 int wildmatch(const char *pattern, const char *text);
 int iwildmatch(const char *pattern, const char *text);
-int wildmatch_array(const char *pattern, const char*const *texts, int where);
-int litmatch_array(const char *string, const char*const *texts, int where);
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#24 lore raw

Re: [PATCH v5 04/12] wildmatch: remove unnecessary functions

Quoted 2 lines · Nguyễn Thái Ngọc Duy

---

The comment-fix seems to be new but otherwise this is unchanged, right?

Quoted 229 lines
 wildmatch.c | 164 ++++--------------------------------------------------------
 wildmatch.h |   2 -
 2 files changed, 10 insertions(+), 156 deletions(-)
wildmatch.c +10 -154
index f3a1731..fae7397 100644
@@ -53,33 +53,18 @@
 #define ISUPPER(c) (ISASCII(c) && isupper(c))
 #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
 
-#ifdef WILD_TEST_ITERATIONS
-int wildmatch_iteration_count;
-#endif
-
 static int force_lower_case = 0;
 
-/* Match pattern "p" against the a virtually-joined string consisting
- * of "text" and any strings in array "a". */
-static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
+/* Match pattern "p" against "text" */
+static int dowild(const uchar *p, const uchar *text)
 {
     uchar p_ch;
 
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count++;
-#endif
-
     for ( ; (p_ch = *p) != '\0'; text++, p++) {
 	int matched, special;
 	uchar t_ch, prev_ch;
-	while ((t_ch = *text) == '\0') {
-	    if (*a == NULL) {
-		if (p_ch != '*')
-		    return ABORT_ALL;
-		break;
-	    }
-	    text = *a++;
-	}
+	if ((t_ch = *text) == '\0' && p_ch != '*')
+		return ABORT_ALL;
 	if (force_lower_case && ISUPPER(t_ch))
 	    t_ch = tolower(t_ch);
 	switch (p_ch) {
@@ -107,21 +92,15 @@ static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
 		/* Trailing "**" matches everything.  Trailing "*" matches
 		 * only if there are no more slash characters. */
 		if (!special) {
-		    do {
 			if (strchr((char*)text, '/') != NULL)
 			    return FALSE;
-		    } while ((text = *a++) != NULL);
 		}
 		return TRUE;
 	    }
 	    while (1) {
-		if (t_ch == '\0') {
-		    if ((text = *a++) == NULL)
-			break;
-		    t_ch = *text;
-		    continue;
-		}
-		if ((matched = dowild(p, text, a)) != FALSE) {
+		if (t_ch == '\0')
+		    break;
+		if ((matched = dowild(p, text)) != FALSE) {
 		    if (!special || matched != ABORT_TO_STARSTAR)
 			return matched;
 		} else if (!special && t_ch == '/')
@@ -225,144 +204,21 @@ static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
 	}
     }
 
-    do {
-	if (*text)
-	    return FALSE;
-    } while ((text = *a++) != NULL);
-
-    return TRUE;
-}
-
-/* Match literal string "s" against the a virtually-joined string consisting
- * of "text" and any strings in array "a". */
-static int doliteral(const uchar *s, const uchar *text, const uchar*const *a)
-{
-    for ( ; *s != '\0'; text++, s++) {
-	while (*text == '\0') {
-	    if ((text = *a++) == NULL)
-		return FALSE;
-	}
-	if (*text != *s)
-	    return FALSE;
-    }
-
-    do {
-	if (*text)
-	    return FALSE;
-    } while ((text = *a++) != NULL);
-
-    return TRUE;
-}
-
-/* Return the last "count" path elements from the concatenated string.
- * We return a string pointer to the start of the string, and update the
- * array pointer-pointer to point to any remaining string elements. */
-static const uchar *trailing_N_elements(const uchar*const **a_ptr, int count)
-{
-    const uchar*const *a = *a_ptr;
-    const uchar*const *first_a = a;
-
-    while (*a)
-	    a++;
-
-    while (a != first_a) {
-	const uchar *s = *--a;
-	s += strlen((char*)s);
-	while (--s >= *a) {
-	    if (*s == '/' && !--count) {
-		*a_ptr = a+1;
-		return s+1;
-	    }
-	}
-    }
-
-    if (count == 1) {
-	*a_ptr = a+1;
-	return *a;
-    }
-
-    return NULL;
+    return *text ? FALSE : TRUE;
 }
 
 /* Match the "pattern" against the "text" string. */
 int wildmatch(const char *pattern, const char *text)
 {
-    static const uchar *nomore[1]; /* A NULL pointer. */
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count = 0;
-#endif
-    return dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE;
+    return dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
 }
 
 /* Match the "pattern" against the forced-to-lower-case "text" string. */
 int iwildmatch(const char *pattern, const char *text)
 {
-    static const uchar *nomore[1]; /* A NULL pointer. */
     int ret;
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count = 0;
-#endif
     force_lower_case = 1;
-    ret = dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE;
+    ret = dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
     force_lower_case = 0;
     return ret;
 }
-
-/* Match pattern "p" against the a virtually-joined string consisting
- * of all the pointers in array "texts" (which has a NULL pointer at the
- * end).  The int "where" can be 0 (normal matching), > 0 (match only
- * the trailing N slash-separated filename components of "texts"), or < 0
- * (match the "pattern" at the start or after any slash in "texts"). */
-int wildmatch_array(const char *pattern, const char*const *texts, int where)
-{
-    const uchar *p = (const uchar*)pattern;
-    const uchar*const *a = (const uchar*const*)texts;
-    const uchar *text;
-    int matched;
-
-#ifdef WILD_TEST_ITERATIONS
-    wildmatch_iteration_count = 0;
-#endif
-
-    if (where > 0)
-	text = trailing_N_elements(&a, where);
-    else
-	text = *a++;
-    if (!text)
-	return FALSE;
-
-    if ((matched = dowild(p, text, a)) != TRUE && where < 0
-     && matched != ABORT_ALL) {
-	while (1) {
-	    if (*text == '\0') {
-		if ((text = (uchar*)*a++) == NULL)
-		    return FALSE;
-		continue;
-	    }
-	    if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE
-	     && matched != ABORT_TO_STARSTAR)
-		break;
-	}
-    }
-    return matched == TRUE;
-}
-
-/* Match literal string "s" against the a virtually-joined string consisting
- * of all the pointers in array "texts" (which has a NULL pointer at the
- * end).  The int "where" can be 0 (normal matching), or > 0 (match
- * only the trailing N slash-separated filename components of "texts"). */
-int litmatch_array(const char *string, const char*const *texts, int where)
-{
-    const uchar *s = (const uchar*)string;
-    const uchar*const *a = (const uchar* const*)texts;
-    const uchar *text;
-
-    if (where > 0)
-	text = trailing_N_elements(&a, where);
-    else
-	text = *a++;
-    if (!text)
-	return FALSE;
-
-    return doliteral(s, text, a) == TRUE;
-}
wildmatch.h +0 -2
index e7f1a35..562faa3 100644
@@ -2,5 +2,3 @@
 
 int wildmatch(const char *pattern, const char *text);
 int iwildmatch(const char *pattern, const char *text);
-int wildmatch_array(const char *pattern, const char*const *texts, int where);
-int litmatch_array(const char *string, const char*const *texts, int where);
#26 lore raw

PATCH v5 5/12 Integrate wildmatch to git

In reply to Nguyễn Thái Ngọc Duy
---
 .gitignore           |   1 +
 Makefile             |   3 +
 t/t3070-wildmatch.sh | 188 +++++++++++++++++++++++++++++++++++++++++++++++++++
 t/t3070/wildtest.txt | 165 --------------------------------------------
 test-wildmatch.c     |  14 ++++
 wildmatch.c          |   5 +-
 6 files changed, 210 insertions(+), 166 deletions(-)
 create mode 100755 t/t3070-wildmatch.sh
 delete mode 100644 t/t3070/wildtest.txt
 create mode 100644 test-wildmatch.c
.gitignore +1 -0
index a188a82..37c3507 100644
@@ -197,6 +197,7 @@
 /test-string-list
 /test-subprocess
 /test-svn-fe
+/test-wildmatch
 /common-cmds.h
 *.tar.gz
 *.dsc
Makefile +3 -0
index f69979e..c752673 100644
@@ -523,6 +523,7 @@ TEST_PROGRAMS_NEED_X += test-sigchain
 TEST_PROGRAMS_NEED_X += test-string-list
 TEST_PROGRAMS_NEED_X += test-subprocess
 TEST_PROGRAMS_NEED_X += test-svn-fe
+TEST_PROGRAMS_NEED_X += test-wildmatch
 
 TEST_PROGRAMS = $(patsubst %,%$X,$(TEST_PROGRAMS_NEED_X))
 
@@ -695,6 +696,7 @@ LIB_H += userdiff.h
 LIB_H += utf8.h
 LIB_H += varint.h
 LIB_H += walker.h
+LIB_H += wildmatch.h
 LIB_H += wt-status.h
 LIB_H += xdiff-interface.h
 LIB_H += xdiff/xdiff.h
@@ -826,6 +828,7 @@ LIB_OBJS += utf8.o
 LIB_OBJS += varint.o
 LIB_OBJS += version.o
 LIB_OBJS += walker.o
+LIB_OBJS += wildmatch.o
 LIB_OBJS += wrapper.o
 LIB_OBJS += write_or_die.o
 LIB_OBJS += ws.o
t/t3070-wildmatch.sh new file +188 -0

188 lines not on this page · raw message

t/t3070/wildtest.txt deleted +0 -165

165 lines not on this page · raw message

test-wildmatch.c new file +14 -0
new file mode 100644
index 0000000..ac56420
@@ -0,0 +1,14 @@
+#include "cache.h"
+#include "wildmatch.h"
+
+int main(int argc, char **argv)
+{
+	if (!strcmp(argv[1], "wildmatch"))
+		return wildmatch(argv[3], argv[2]) ? 0 : 1;
+	else if (!strcmp(argv[1], "iwildmatch"))
+		return iwildmatch(argv[3], argv[2]) ? 0 : 1;
+	else if (!strcmp(argv[1], "fnmatch"))
+		return !!fnmatch(argv[3], argv[2], FNM_PATHNAME);
+	else
+		return 1;
+}
wildmatch.c +4 -1
index fae7397..d0b906a 100644
@@ -9,7 +9,10 @@
 **  work differently than '*', and to fix the character-class code.
 */
 
-#include "rsync.h"
+#include "cache.h"
+#include "wildmatch.h"
+
+typedef unsigned char uchar;
 
 /* What character marks an inverted character class? */
 #define NEGATE_CLASS	'!'
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#27 lore raw

Re: [PATCH v5 05/12] Integrate wildmatch to git

Quoted 21 lines · Nguyễn Thái Ngọc Duy

+++ b/t/t3070-wildmatch.sh

t/t3070-wildmatch.sh new file +19 -0
@@ -0,0 +1,188 @@
+#!/bin/sh
+
+test_description='wildmatch tests'
+
+. ./test-lib.sh
+
+match() {
+    if [ $1 = 1 ]; then
+	test_expect_success "wildmatch:    match '$3' '$4'" "
+	    test-wildmatch wildmatch '$3' '$4'
+	"
+    else
+	test_expect_success "wildmatch: no match '$3' '$4'" "
+	    ! test-wildmatch wildmatch '$3' '$4'
+	"
+    fi
+    if [ $2 = 1 ]; then
+	test_expect_success "fnmatch:      match '$3' '$4'" "
+	    test-wildmatch fnmatch '$3' '$4'

t/t3070-wildmatch.sh line 29

Quoted 10 lines
t/t3070-wildmatch.sh new file +10 -0
+	"
+    elif [ $2 = 0 ]; then
+	test_expect_success "fnmatch:   no match '$3' '$4'" "
+	    ! test-wildmatch fnmatch '$3' '$4'
+	"
+#    else
+#	test_expect_success BROKEN_FNMATCH "fnmatch:       '$3' '$4'" "
+#	    ! test-wildmatch fnmatch '$3' '$4'
+#	"
+    fi

Heh, broken can be two-way. Either it may succeed matching what it shouldn't, or it may not match what it should.

#28 lore raw

Re: [PATCH v5 05/12] Integrate wildmatch to git

In reply to Nguyễn Thái Ngọc Duy
t/t3070-wildmatch.sh new file +8 -0
new file mode 100755
index 0000000..dbd3c8b
@@ -0,0 +1,188 @@
+#!/bin/sh
+#    else
+#    test_expect_success BROKEN_FNMATCH "fnmatch:       '$3' '$4'" "
+#        ! test-wildmatch fnmatch '$3' '$4'
+#    "
+    fi
+}
+

Thanks:
On my Mac OS X box:

# passed all 259 test(s)

And a quick test on cygwin:

$ ./t3070-wildmatch.sh  2>&1 | grep "not ok"
not ok - 148 fnmatch:      match '5' '[[:xdigit:]]'
not ok - 150 fnmatch:      match 'f' '[[:xdigit:]]'
not ok - 152 fnmatch:      match 'D' '[[:xdigit:]]'

And 2 micronits:
a) Commented out code
b) Whithespace damage
( 4 spaces used for an indent of 1, TAB for indent of 2)

/Torsten

#29 lore raw

PATCH v5 6/12 t3070: disable unreliable fnmatch tests

In reply to Nguyễn Thái Ngọc Duy

These tests show different results on different fnmatch() versions. We don't want to test fnmatch here. We want to make sure wildmatch behavior matches fnmatch and that only makes sense in cases when fnmatch() behaves consistently.

---
 t/t3070-wildmatch.sh | 86 ++++++++++++++++++++++++++--------------------------
 1 file changed, 43 insertions(+), 43 deletions(-)
t/t3070-wildmatch.sh +43 -43
index dbd3c8b..dd95b00 100755
@@ -52,11 +52,11 @@ match 1 1 ten 't[a-g]n'
 match 0 0 ten 't[!a-g]n'
 match 1 1 ton 't[!a-g]n'
 match 1 1 ton 't[^a-g]n'
-match 1 1 'a]b' 'a[]]b'
-match 1 1 a-b 'a[]-]b'
-match 1 1 'a]b' 'a[]-]b'
-match 0 0 aab 'a[]-]b'
-match 1 1 aab 'a[]a-]b'
+match 1 x 'a]b' 'a[]]b'
+match 1 x a-b 'a[]-]b'
+match 1 x 'a]b' 'a[]-]b'
+match 0 x aab 'a[]-]b'
+match 1 x aab 'a[]a-]b'
 match 1 1 ']' ']'
 
 # Extended slash-matching features
@@ -67,7 +67,7 @@ match 0 0 'foo/bar' 'foo[/]bar'
 match 0 0 'foo/bar' 'f[^eiu][^eiu][^eiu][^eiu][^eiu]r'
 match 1 1 'foo-bar' 'f[^eiu][^eiu][^eiu][^eiu][^eiu]r'
 match 0 0 'foo' '**/foo'
-match 1 1 '/foo' '**/foo'
+match 1 x '/foo' '**/foo'
 match 1 0 'bar/baz/foo' '**/foo'
 match 0 0 'bar/baz/foo' '*/foo'
 match 0 0 'foo/bar/baz' '**/bar*'
@@ -85,77 +85,77 @@ match 1 0 'deep/foo/bar/baz/x' '**/bar/*/*'
 match 0 0 'acrt' 'a[c-c]st'
 match 1 1 'acrt' 'a[c-c]rt'
 match 0 0 ']' '[!]-]'
-match 1 1 'a' '[!]-]'
+match 1 x 'a' '[!]-]'
 match 0 0 '' '\'
-match 0 0 '\' '\'
-match 0 0 '/\' '*/\'
-match 1 1 '/\' '*/\\'
+match 0 x '\' '\'
+match 0 x '/\' '*/\'
+match 1 x '/\' '*/\\'
 match 1 1 'foo' 'foo'
 match 1 1 '@foo' '@foo'
 match 0 0 'foo' '@foo'
 match 1 1 '[ab]' '\[ab]'
 match 1 1 '[ab]' '[[]ab]'
-match 1 1 '[ab]' '[[:]ab]'
-match 0 0 '[ab]' '[[::]ab]'
-match 1 1 '[ab]' '[[:digit]ab]'
-match 1 1 '[ab]' '[\[:]ab]'
+match 1 x '[ab]' '[[:]ab]'
+match 0 x '[ab]' '[[::]ab]'
+match 1 x '[ab]' '[[:digit]ab]'
+match 1 x '[ab]' '[\[:]ab]'
 match 1 1 '?a?b' '\??\?b'
 match 1 1 'abc' '\a\b\c'
 match 0 0 'foo' ''
 match 1 0 'foo/bar/baz/to' '**/t[o]'
 
 # Character class tests
-match 1 1 'a1B' '[[:alpha:]][[:digit:]][[:upper:]]'
-match 0 0 'a' '[[:digit:][:upper:][:space:]]'
-match 1 1 'A' '[[:digit:][:upper:][:space:]]'
-match 1 0 '1' '[[:digit:][:upper:][:space:]]'
-match 0 0 '1' '[[:digit:][:upper:][:spaci:]]'
-match 1 1 ' ' '[[:digit:][:upper:][:space:]]'
-match 0 0 '.' '[[:digit:][:upper:][:space:]]'
-match 1 1 '.' '[[:digit:][:punct:][:space:]]'
+match 1 x 'a1B' '[[:alpha:]][[:digit:]][[:upper:]]'
+match 0 x 'a' '[[:digit:][:upper:][:space:]]'
+match 1 x 'A' '[[:digit:][:upper:][:space:]]'
+match 1 x '1' '[[:digit:][:upper:][:space:]]'
+match 0 x '1' '[[:digit:][:upper:][:spaci:]]'
+match 1 x ' ' '[[:digit:][:upper:][:space:]]'
+match 0 x '.' '[[:digit:][:upper:][:space:]]'
+match 1 x '.' '[[:digit:][:punct:][:space:]]'
 match 1 1 '5' '[[:xdigit:]]'
 match 1 1 'f' '[[:xdigit:]]'
 match 1 1 'D' '[[:xdigit:]]'
-match 1 0 '_' '[[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]'
-match 1 0 '_' '[[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]'
-match 1 1 '.' '[^[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:lower:][:space:][:upper:][:xdigit:]]'
-match 1 1 '5' '[a-c[:digit:]x-z]'
-match 1 1 'b' '[a-c[:digit:]x-z]'
-match 1 1 'y' '[a-c[:digit:]x-z]'
-match 0 0 'q' '[a-c[:digit:]x-z]'
+match 1 x '_' '[[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]'
+match 1 x '_' '[[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:graph:][:lower:][:print:][:punct:][:space:][:upper:][:xdigit:]]'
+match 1 x '.' '[^[:alnum:][:alpha:][:blank:][:cntrl:][:digit:][:lower:][:space:][:upper:][:xdigit:]]'
+match 1 x '5' '[a-c[:digit:]x-z]'
+match 1 x 'b' '[a-c[:digit:]x-z]'
+match 1 x 'y' '[a-c[:digit:]x-z]'
+match 0 x 'q' '[a-c[:digit:]x-z]'
 
 # Additional tests, including some malformed wildmats
-match 1 1 ']' '[\\-^]'
+match 1 x ']' '[\\-^]'
 match 0 0 '[' '[\\-^]'
-match 1 1 '-' '[\-_]'
-match 1 1 ']' '[\]]'
+match 1 x '-' '[\-_]'
+match 1 x ']' '[\]]'
 match 0 0 '\]' '[\]]'
 match 0 0 '\' '[\]]'
 match 0 0 'ab' 'a[]b'
-match 0 1 'a[]b' 'a[]b'
-match 0 1 'ab[' 'ab['
+match 0 x 'a[]b' 'a[]b'
+match 0 x 'ab[' 'ab['
 match 0 0 'ab' '[!'
 match 0 0 'ab' '[-'
 match 1 1 '-' '[-]'
 match 0 0 '-' '[a-'
 match 0 0 '-' '[!a-'
-match 1 1 '-' '[--A]'
-match 1 1 '5' '[--A]'
+match 1 x '-' '[--A]'
+match 1 x '5' '[--A]'
 match 1 1 ' ' '[ --]'
 match 1 1 '$' '[ --]'
 match 1 1 '-' '[ --]'
 match 0 0 '0' '[ --]'
-match 1 1 '-' '[---]'
-match 1 1 '-' '[------]'
+match 1 x '-' '[---]'
+match 1 x '-' '[------]'
 match 0 0 'j' '[a-e-n]'
-match 1 1 '-' '[a-e-n]'
-match 1 1 'a' '[!------]'
+match 1 x '-' '[a-e-n]'
+match 1 x 'a' '[!------]'
 match 0 0 '[' '[]-a]'
-match 1 1 '^' '[]-a]'
+match 1 x '^' '[]-a]'
 match 0 0 '^' '[!]-a]'
-match 1 1 '[' '[!]-a]'
+match 1 x '[' '[!]-a]'
 match 1 1 '^' '[a^bc]'
-match 1 1 '-b]' '[a-]b]'
+match 1 x '-b]' '[a-]b]'
 match 0 0 '\' '[\]'
 match 1 1 '\' '[\\]'
 match 0 0 '\' '[!\\]'
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#30 lore raw

PATCH v5 7/12 wildmatch: make wildmatch's return value compatible with fnmatch

In reply to Nguyễn Thái Ngọc Duy

wildmatch returns non-zero if matched, zero otherwise. This patch makes it return zero if matches, non-zero otherwise, like fnmatch().

---
 test-wildmatch.c |  4 ++--
 wildmatch.c      | 21 ++++++++++++---------
 2 files changed, 14 insertions(+), 11 deletions(-)
test-wildmatch.c +2 -2
index ac56420..77014e9 100644
@@ -4,9 +4,9 @@
 int main(int argc, char **argv)
 {
 	if (!strcmp(argv[1], "wildmatch"))
-		return wildmatch(argv[3], argv[2]) ? 0 : 1;
+		return !!wildmatch(argv[3], argv[2]);
 	else if (!strcmp(argv[1], "iwildmatch"))
-		return iwildmatch(argv[3], argv[2]) ? 0 : 1;
+		return !!iwildmatch(argv[3], argv[2]);
 	else if (!strcmp(argv[1], "fnmatch"))
 		return !!fnmatch(argv[3], argv[2], FNM_PATHNAME);
 	else
wildmatch.c +12 -9
index d0b906a..e3ac6cc 100644
@@ -20,6 +20,9 @@ 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
 
@@ -78,12 +81,12 @@ static int dowild(const uchar *p, const uchar *text)
 	    /* FALLTHROUGH */
 	  default:
 	    if (t_ch != p_ch)
-		return FALSE;
+		return NOMATCH;
 	    continue;
 	  case '?':
 	    /* Match anything but '/'. */
 	    if (t_ch == '/')
-		return FALSE;
+		return NOMATCH;
 	    continue;
 	  case '*':
 	    if (*++p == '*') {
@@ -96,14 +99,14 @@ static int dowild(const uchar *p, const uchar *text)
 		 * only if there are no more slash characters. */
 		if (!special) {
 			if (strchr((char*)text, '/') != NULL)
-			    return FALSE;
+			    return NOMATCH;
 		}
-		return TRUE;
+		return MATCH;
 	    }
 	    while (1) {
 		if (t_ch == '\0')
 		    break;
-		if ((matched = dowild(p, text)) != FALSE) {
+		if ((matched = dowild(p, text)) != NOMATCH) {
 		    if (!special || matched != ABORT_TO_STARSTAR)
 			return matched;
 		} else if (!special && t_ch == '/')
@@ -202,18 +205,18 @@ static int dowild(const uchar *p, const uchar *text)
 		    matched = TRUE;
 	    } while (prev_ch = p_ch, (p_ch = *++p) != ']');
 	    if (matched == special || t_ch == '/')
-		return FALSE;
+		return NOMATCH;
 	    continue;
 	}
     }
 
-    return *text ? FALSE : TRUE;
+    return *text ? NOMATCH : MATCH;
 }
 
 /* Match the "pattern" against the "text" string. */
 int wildmatch(const char *pattern, const char *text)
 {
-    return dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
+    return dowild((const uchar*)pattern, (const uchar*)text);
 }
 
 /* Match the "pattern" against the forced-to-lower-case "text" string. */
@@ -221,7 +224,7 @@ int iwildmatch(const char *pattern, const char *text)
 {
     int ret;
     force_lower_case = 1;
-    ret = dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
+    ret = dowild((const uchar*)pattern, (const uchar*)text);
     force_lower_case = 0;
     return ret;
 }
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#31 lore raw

Re: [PATCH v5 07/12] wildmatch: make wildmatch's return value compatible with fnmatch

Quoted 5 lines · Nguyễn Thái Ngọc Duy

wildmatch returns non-zero if matched, zero otherwise. This patch makes it return zero if matches, non-zero otherwise, like fnmatch().

---

OK, so ABORT cases where the patterns are either broken or impossible to match are also taken as not-matching, which sounds like the right thing to do.

Quoted 98 lines
 test-wildmatch.c |  4 ++--
 wildmatch.c      | 21 ++++++++++++---------
 2 files changed, 14 insertions(+), 11 deletions(-)
test-wildmatch.c +2 -2
index ac56420..77014e9 100644
@@ -4,9 +4,9 @@
 int main(int argc, char **argv)
 {
 	if (!strcmp(argv[1], "wildmatch"))
-		return wildmatch(argv[3], argv[2]) ? 0 : 1;
+		return !!wildmatch(argv[3], argv[2]);
 	else if (!strcmp(argv[1], "iwildmatch"))
-		return iwildmatch(argv[3], argv[2]) ? 0 : 1;
+		return !!iwildmatch(argv[3], argv[2]);
 	else if (!strcmp(argv[1], "fnmatch"))
 		return !!fnmatch(argv[3], argv[2], FNM_PATHNAME);
 	else
wildmatch.c +12 -9
index d0b906a..e3ac6cc 100644
@@ -20,6 +20,9 @@ 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
 
@@ -78,12 +81,12 @@ static int dowild(const uchar *p, const uchar *text)
 	    /* FALLTHROUGH */
 	  default:
 	    if (t_ch != p_ch)
-		return FALSE;
+		return NOMATCH;
 	    continue;
 	  case '?':
 	    /* Match anything but '/'. */
 	    if (t_ch == '/')
-		return FALSE;
+		return NOMATCH;
 	    continue;
 	  case '*':
 	    if (*++p == '*') {
@@ -96,14 +99,14 @@ static int dowild(const uchar *p, const uchar *text)
 		 * only if there are no more slash characters. */
 		if (!special) {
 			if (strchr((char*)text, '/') != NULL)
-			    return FALSE;
+			    return NOMATCH;
 		}
-		return TRUE;
+		return MATCH;
 	    }
 	    while (1) {
 		if (t_ch == '\0')
 		    break;
-		if ((matched = dowild(p, text)) != FALSE) {
+		if ((matched = dowild(p, text)) != NOMATCH) {
 		    if (!special || matched != ABORT_TO_STARSTAR)
 			return matched;
 		} else if (!special && t_ch == '/')
@@ -202,18 +205,18 @@ static int dowild(const uchar *p, const uchar *text)
 		    matched = TRUE;
 	    } while (prev_ch = p_ch, (p_ch = *++p) != ']');
 	    if (matched == special || t_ch == '/')
-		return FALSE;
+		return NOMATCH;
 	    continue;
 	}
     }
 
-    return *text ? FALSE : TRUE;
+    return *text ? NOMATCH : MATCH;
 }
 
 /* Match the "pattern" against the "text" string. */
 int wildmatch(const char *pattern, const char *text)
 {
-    return dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
+    return dowild((const uchar*)pattern, (const uchar*)text);
 }
 
 /* Match the "pattern" against the forced-to-lower-case "text" string. */
@@ -221,7 +224,7 @@ int iwildmatch(const char *pattern, const char *text)
 {
     int ret;
     force_lower_case = 1;
-    ret = dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
+    ret = dowild((const uchar*)pattern, (const uchar*)text);
     force_lower_case = 0;
     return ret;
 }
#32 lore raw

PATCH v5 8/12 wildmatch: remove static variable force_lower_case

In reply to Nguyễn Thái Ngọc Duy

One place less to worry about thread safety. Also combine wildmatch and iwildmatch into one.

---
 test-wildmatch.c |  4 ++--
 wildmatch.c      | 21 +++++----------------
 wildmatch.h      |  3 +--
 3 files changed, 8 insertions(+), 20 deletions(-)
test-wildmatch.c +2 -2
index 77014e9..74c0864 100644
@@ -4,9 +4,9 @@
 int main(int argc, char **argv)
 {
 	if (!strcmp(argv[1], "wildmatch"))
-		return !!wildmatch(argv[3], argv[2]);
+		return !!wildmatch(argv[3], argv[2], 0);
 	else if (!strcmp(argv[1], "iwildmatch"))
-		return !!iwildmatch(argv[3], argv[2]);
+		return !!wildmatch(argv[3], argv[2], FNM_CASEFOLD);
 	else if (!strcmp(argv[1], "fnmatch"))
 		return !!fnmatch(argv[3], argv[2], FNM_PATHNAME);
 	else
wildmatch.c +5 -16
index e3ac6cc..20c5ef6 100644
@@ -59,10 +59,8 @@ typedef unsigned char uchar;
 #define ISUPPER(c) (ISASCII(c) && isupper(c))
 #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
 
-static int force_lower_case = 0;
-
 /* Match pattern "p" against "text" */
-static int dowild(const uchar *p, const uchar *text)
+static int dowild(const uchar *p, const uchar *text, int force_lower_case)
 {
     uchar p_ch;
 
@@ -106,7 +104,7 @@ static int dowild(const uchar *p, const uchar *text)
 	    while (1) {
 		if (t_ch == '\0')
 		    break;
-		if ((matched = dowild(p, text)) != NOMATCH) {
+		if ((matched = dowild(p, text, force_lower_case)) != NOMATCH) {
 		    if (!special || matched != ABORT_TO_STARSTAR)
 			return matched;
 		} else if (!special && t_ch == '/')
@@ -214,17 +212,8 @@ static int dowild(const uchar *p, const uchar *text)
 }
 
 /* Match the "pattern" against the "text" string. */
-int wildmatch(const char *pattern, const char *text)
-{
-    return dowild((const uchar*)pattern, (const uchar*)text);
-}
-
-/* Match the "pattern" against the forced-to-lower-case "text" string. */
-int iwildmatch(const char *pattern, const char *text)
+int wildmatch(const char *pattern, const char *text, int flags)
 {
-    int ret;
-    force_lower_case = 1;
-    ret = dowild((const uchar*)pattern, (const uchar*)text);
-    force_lower_case = 0;
-    return ret;
+    return dowild((const uchar*)pattern, (const uchar*)text,
+		  flags & FNM_CASEFOLD ? 1 : 0);
 }
wildmatch.h +1 -2
index 562faa3..e974f9a 100644
@@ -1,4 +1,3 @@
 /* wildmatch.h */
 
-int wildmatch(const char *pattern, const char *text);
-int iwildmatch(const char *pattern, const char *text);
+int wildmatch(const char *pattern, const char *text, int flags);
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#33 lore raw

PATCH v5 9/12 wildmatch: fix case-insensitive matching

In reply to Nguyễn Thái Ngọc Duy

dowild() does case insensitive matching by lower-casing the text. That means lower case letters in patterns imply case-insensitive matching, but upper case means exact matching.

We do not want that subtlety. Lower case pattern too so iwildmatch() always does what we expect it to do.

---
 wildmatch.c | 2 ++
 1 file changed, 2 insertions(+)
wildmatch.c +2 -0
index 20c5ef6..6542524 100644
@@ -71,6 +71,8 @@ static int dowild(const uchar *p, const uchar *text, int force_lower_case)
 		return ABORT_ALL;
 	if (force_lower_case && ISUPPER(t_ch))
 	    t_ch = tolower(t_ch);
+	if (force_lower_case && ISUPPER(p_ch))
+	    p_ch = tolower(p_ch);
 	switch (p_ch) {
 	  case '\\':
 	    /* Literal match with following character.  Note that the test
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#34 lore raw

PATCH v5 10/12 wildmatch: adjust "**" behavior

In reply to Nguyễn Thái Ngọc Duy

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.

---
 t/t3070-wildmatch.sh |  5 +++--
 wildmatch.c          | 13 +++++++------
 wildmatch.h          |  6 ++++++
 3 files changed, 16 insertions(+), 8 deletions(-)
t/t3070-wildmatch.sh +3 -2
index dd95b00..15848d5 100755
@@ -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'
wildmatch.c +7 -6
index 6542524..7209f26 100644
@@ -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') {
wildmatch.h +6 -0
index e974f9a..984a38c 100644
@@ -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);
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#35 lore raw

PATCH v5 11/12 wildmatch: make /**/ match zero or more directories

In reply to Nguyễn Thái Ngọc Duy

"foo/**/bar" matches "foo/x/bar", "foo/x/y/bar"... but not "foo/bar". We make a special case, when foo/**/ is detected (and "foo/" part is already matched), try matching "bar" with the rest of the string.

"Match one or more directories" semantics can be easily achieved using "foo/*/**/bar".

This also makes "**/foo" match "foo" in addition to "x/foo", "x/y/foo"..

---
 t/t3070-wildmatch.sh |  8 +++++++-
 wildmatch.c          | 17 +++++++++++++++++
 2 files changed, 24 insertions(+), 1 deletion(-)
t/t3070-wildmatch.sh +7 -1
index 15848d5..e6ad6f4 100755
@@ -63,11 +63,17 @@ match 1 1 ']' ']'
 match 0 0 'foo/baz/bar' 'foo*bar'
 match 0 0 'foo/baz/bar' 'foo**bar'
 match 0 1 'foobazbar' 'foo**bar'
+match 1 1 'foo/baz/bar' 'foo/**/bar'
+match 1 0 'foo/baz/bar' 'foo/**/**/bar'
+match 1 0 'foo/b/a/z/bar' 'foo/**/bar'
+match 1 0 'foo/b/a/z/bar' 'foo/**/**/bar'
+match 1 0 'foo/bar' 'foo/**/bar'
+match 1 0 'foo/bar' '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'
 match 1 1 'foo-bar' 'f[^eiu][^eiu][^eiu][^eiu][^eiu]r'
-match 0 0 'foo' '**/foo'
+match 1 0 'foo' '**/foo'
 match 1 x '/foo' '**/foo'
 match 1 0 'bar/baz/foo' '**/foo'
 match 0 0 'bar/baz/foo' '*/foo'
wildmatch.c +17 -0
index 7209f26..35c34ac 100644
@@ -90,6 +90,23 @@ static int dowild(const uchar *p, const uchar *text, int force_lower_case)
 		if ((prev_p == text || *prev_p == '/') ||
 		    (*p == '\0' || *p == '/' ||
 		     (p[0] == '\\' && p[1] == '/'))) {
+			/*
+			 * Assuming we already match 'foo/' and are at
+			 * <star star slash>, just assume it matches
+			 * nothing and go ahead match the rest of the
+			 * pattern with the remaining string. This
+			 * helps make foo/<*><*>/bar (<> because
+			 * otherwise it breaks C comment syntax) match
+			 * both foo/bar and foo/a/bar.
+			 *
+			 * Crazy patterns like /<*><*>/<*><*>/ are
+			 * treated like /<*><*>/. But undefined
+			 * behavior is even appropriate for people
+			 * writing such a pattern.
+			 */
+			if (p[0] == '/' &&
+			    dowild(p + 1, text, force_lower_case) == MATCH)
+				return MATCH;
 		    special = TRUE;
 		} else
 		    return ABORT_MALFORMED;
Signature · 1 line
1.8.0.rc2.11.g2b79d01
#36 lore raw

PATCH v5 12/12 Support "**" wildcard in .gitignore and .gitattributes

In reply to Nguyễn Thái Ngọc Duy
---
 Documentation/gitignore.txt        | 19 +++++++++++++++++++
 attr.c                             |  4 +++-
 dir.c                              |  4 +++-
 t/t0003-attributes.sh              | 37 +++++++++++++++++++++++++++++++++++++
 t/t3001-ls-files-others-exclude.sh | 19 +++++++++++++++++++
 5 files changed, 81 insertions(+), 2 deletions(-)
Documentation/gitignore.txt +19 -0
index 1b82fe1..91a6438 100644
@@ -108,6 +108,25 @@ PATTERN FORMAT
    For example, "/{asterisk}.c" matches "cat-file.c" but not
    "mozilla-sha1/sha1.c".
 
+Two consecutive asterisks ("`**`") in patterns matched against
+full pathname may have special meaning:
+
+ - A leading "`**`" followed by a slash means match in all
+   directories. For example, "`**/foo`" matches file or directory
+   "`foo`" anywhere, the same as pattern "`foo`". "**/foo/bar"
+   matches file or directory "`bar`" anywhere that is directly
+   under directory "`foo`".
+
+ - A trailing "/**" matches everything inside. For example,
+   "abc/**" matches all files inside directory "abc", relative
+   to the location of the `.gitignore` file, with infinite depth.
+
+ - A slash followed by two consecutive asterisks then a slash
+   matches zero or more directories. For example, "`a/**/b`"
+   matches "`a/b`", "`a/x/b`", "`a/x/y/b`" and so on.
+
+ - Other consecutive asterisks are considered invalid.
+
 NOTES
 -----
 
attr.c +3 -1
index 887a9ae..8010429 100644
@@ -12,6 +12,7 @@
 #include "exec_cmd.h"
 #include "attr.h"
 #include "dir.h"
+#include "wildmatch.h"
 
 const char git_attr__true[] = "(builtin)true";
 const char git_attr__false[] = "\0(builtin)false";
@@ -666,7 +667,8 @@ static int path_matches(const char *pathname, int pathlen,
 		return 0;
 	if (baselen != 0)
 		baselen++;
-	return fnmatch_icase(pattern, pathname + baselen, FNM_PATHNAME) == 0;
+	return wildmatch(pattern, pathname + baselen,
+			 ignore_case ? FNM_CASEFOLD : 0) == 0;
 }
 
 static int macroexpand_one(int attr_nr, int rem);
dir.c +3 -1
index 4868339..442db1c 100644
@@ -8,6 +8,7 @@
 #include "cache.h"
 #include "dir.h"
 #include "refs.h"
+#include "wildmatch.h"
 
 struct path_simplify {
 	int len;
@@ -575,7 +576,8 @@ int excluded_from_list(const char *pathname,
 			namelen -= prefix;
 		}
 
-		if (!namelen || !fnmatch_icase(exclude, name, FNM_PATHNAME))
+		if (!namelen ||
+		    wildmatch(exclude, name, ignore_case ? FNM_CASEFOLD : 0) == 0)
 			return to_exclude;
 	}
 	return -1; /* undecided */
t/t0003-attributes.sh +37 -0
index febc45c..d5a6946 100755
@@ -196,6 +196,43 @@ test_expect_success 'root subdir attribute test' '
 	attr_check subdir/a/i unspecified
 '
 
+test_expect_success '"**" test' '
+	echo "**/f foo=bar" >.gitattributes &&
+	cat <<\EOF >expect &&
+f: foo: bar
+a/f: foo: bar
+a/b/f: foo: bar
+a/b/c/f: foo: bar
+EOF
+	git check-attr foo -- "f" >actual 2>err &&
+	git check-attr foo -- "a/f" >>actual 2>>err &&
+	git check-attr foo -- "a/b/f" >>actual 2>>err &&
+	git check-attr foo -- "a/b/c/f" >>actual 2>>err &&
+	test_cmp expect actual &&
+	test_line_count = 0 err
+'
+
+test_expect_success '"**" with no slashes test' '
+	echo "a**f foo=bar" >.gitattributes &&
+	git check-attr foo -- "f" >actual &&
+	cat <<\EOF >expect &&
+f: foo: unspecified
+af: foo: bar
+axf: foo: bar
+a/f: foo: unspecified
+a/b/f: foo: unspecified
+a/b/c/f: foo: unspecified
+EOF
+	git check-attr foo -- "f" >actual 2>err &&
+	git check-attr foo -- "af" >>actual 2>err &&
+	git check-attr foo -- "axf" >>actual 2>err &&
+	git check-attr foo -- "a/f" >>actual 2>>err &&
+	git check-attr foo -- "a/b/f" >>actual 2>>err &&
+	git check-attr foo -- "a/b/c/f" >>actual 2>>err &&
+	test_cmp expect actual &&
+	test_line_count = 0 err
+'
+
 test_expect_success 'setup bare' '
 	git clone --bare . bare.git &&
 	cd bare.git
t/t3001-ls-files-others-exclude.sh +19 -0
index c8fe978..278315d 100755
@@ -214,4 +214,23 @@ test_expect_success 'subdirectory ignore (l1)' '
 	test_cmp expect actual
 '
 
+
+test_expect_success 'ls-files with "**" patterns' '
+	cat <<\EOF >expect &&
+a.1
+one/a.1
+one/two/a.1
+three/a.1
+EOF
+	git ls-files -o -i --exclude "**/a.1" >actual
+	test_cmp expect actual
+'
+
+
+test_expect_success 'ls-files with "**" patterns and no slashes' '
+	: >expect &&
+	git ls-files -o -i --exclude "one**a.1" >actual &&
+	test_cmp expect actual
+'
+
 test_done
Signature · 1 line
1.8.0.rc2.11.g2b79d01