[rsync/rsync-patches.git] / ignore-case.diff

This adds the --ignore-case option, which makes rsync compare filenames
in a case-insensitive manner.

To use this patch, run these commands for a successful build:

    patch -p1 <patches/ignore-case.diff
    ./configure                            (optional if already run)
    make

TODO:

 - Make this code handle multibyte character encodings, and honor the
   --iconv setting when converting case.

--- old/exclude.c
+++ new/exclude.c
@@ -583,16 +583,15 @@ static int rule_matches(char *name, stru
 		if (litmatch_array(pattern, strings, slash_handling))
 			return ret_match;
 	} else if (anchored_match) {
-		if (strcmp(name, pattern) == 0)
+		if (ic_strEQ(name, pattern))
 			return ret_match;
 	} else {
 		int l1 = strlen(name);
 		int l2 = strlen(pattern);
-		if (l2 <= l1 &&
-		    strcmp(name+(l1-l2),pattern) == 0 &&
-		    (l1==l2 || name[l1-(l2+1)] == '/')) {
+		if (l2 <= l1
+		 && ic_strEQ(name + (l1-l2), pattern)
+		 && (l1 == l2 || name[l1 - (l2+1)] == '/'))
 			return ret_match;
-		}
 	}
 
 	return !ret_match;
--- old/flist.c
+++ new/flist.c
@@ -35,6 +35,7 @@ extern int inc_recurse;
 extern int do_progress;
 extern int always_checksum;
 extern int module_id;
+extern int ignore_case;
 extern int ignore_errors;
 extern int numeric_ids;
 extern int recurse;
@@ -2570,6 +2571,7 @@ int f_name_cmp(const struct file_struct 
 {
 	int dif;
 	const uchar *c1, *c2;
+	uchar ch1, ch2;
 	enum fnc_state state1, state2;
 	enum fnc_type type1, type2;
 	enum fnc_type t_path = protocol_version >= 29 ? t_PATH : t_ITEM;
@@ -2680,7 +2682,15 @@ int f_name_cmp(const struct file_struct 
 			if (type1 != type2)
 				return type1 == t_PATH ? 1 : -1;
 		}
-	} while ((dif = (int)*c1++ - (int)*c2++) == 0);
+		ch1 = *c1++;
+		ch2 = *c2++;
+		if (ignore_case) {
+			if (isupper(ch1))
+				ch1 = tolower(ch1);
+			if (isupper(ch2))
+				ch2 = tolower(ch2);
+		}
+	} while ((dif = (int)ch1 - (int)ch2) == 0);
 
 	return dif;
 }
--- old/ifuncs.h
+++ new/ifuncs.h
@@ -98,3 +98,38 @@ toUpper(const char *ptr)
 {
 	return toupper(*(unsigned char *)ptr);
 }
+
+static inline int
+strEQ(const char *s1, const char *s2)
+{
+	return strcmp(s1, s2) == 0;
+}
+
+static inline int
+strnEQ(const char *s1, const char *s2, size_t n)
+{
+	return strncmp(s1, s2, n) == 0;
+}
+
+static inline int
+ic_strEQ(const char *s1, const char *s2)
+{
+	extern int ignore_case;
+	if (ignore_case)
+		return strcasecmp(s1, s2) == 0;
+	return strcmp(s1, s2) == 0;
+}
+
+static inline int
+ic_strnEQ(const char *s1, const char *s2, size_t n)
+{
+	extern int ignore_case;
+	if (ignore_case)
+		return strcasecmp(s1, s2, n) == 0;
+	return strcmp(s1, s2, n) == 0;
+}
+
+#define strNE(s1,s2) (!strEQ(s1,s2))
+#define strnNE(s1,s2,n) (!strnEQ(s1,s2,n))
+#define ic_strNE(s1,s2) (!ic_strEQ(s1,s2))
+#define ic_strnNE(s1,s2) (!ic_strnEQ(s1,s2,n))
--- old/lib/wildmatch.c
+++ new/lib/wildmatch.c
@@ -53,6 +53,8 @@
 #define ISUPPER(c) (ISASCII(c) && isupper(c))
 #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
 
+extern int ignore_case;
+
 #ifdef WILD_TEST_ITERATIONS
 int wildmatch_iteration_count;
 #endif
@@ -72,6 +74,8 @@ static int dowild(const uchar *p, const 
     for ( ; (p_ch = *p) != '\0'; text++, p++) {
 	int matched, special;
 	uchar t_ch, prev_ch;
+	if (ignore_case && ISUPPER(p_ch))
+	    p_ch = tolower(p_ch);
 	while ((t_ch = *text) == '\0') {
 	    if (*a == NULL) {
 		if (p_ch != '*')
@@ -237,12 +241,21 @@ static int dowild(const uchar *p, const 
  * of "text" and any strings in array "a". */
 static int doliteral(const uchar *s, const uchar *text, const uchar*const *a)
 {
+    uchar s_ch, t_ch;
     for ( ; *s != '\0'; text++, s++) {
 	while (*text == '\0') {
 	    if ((text = *a++) == NULL)
 		return FALSE;
 	}
-	if (*text != *s)
+	s_ch = *s;
+	t_ch = *text;
+	if (ignore_case) {
+	    if (ISUPPER(s_ch))
+		s_ch = tolower(s_ch);
+	    if (ISUPPER(t_ch))
+		t_ch = tolower(t_ch);
+	}
+	if (t_ch != s_ch)
 	    return FALSE;
     }
 
@@ -288,10 +301,14 @@ static const uchar *trailing_N_elements(
 int wildmatch(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
-    return dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE;
+    force_lower_case = ignore_case;
+    ret = dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE;
+    force_lower_case = 0;
+    return ret;
 }
 
 /* Match the "pattern" against the forced-to-lower-case "text" string. */
@@ -331,12 +348,14 @@ int wildmatch_array(const char *pattern,
     if (!text)
 	return FALSE;
 
+    force_lower_case = ignore_case;
+
     if ((matched = dowild(p, text, a)) != TRUE && where < 0
      && matched != ABORT_ALL) {
 	while (1) {
 	    if (*text == '\0') {
 		if ((text = (uchar*)*a++) == NULL)
-		    return FALSE;
+		    break;
 		continue;
 	    }
 	    if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE
@@ -344,6 +363,9 @@ int wildmatch_array(const char *pattern,
 		break;
 	}
     }
+
+    force_lower_case = 0;
+
     return matched == TRUE;
 }
 
--- old/options.c
+++ new/options.c
@@ -115,6 +115,7 @@ OFF_T max_size = 0;
 OFF_T min_size = 0;
 int ignore_errors = 0;
 int modify_window = 0;
+int ignore_case = 0;
 int blocking_io = -1;
 int checksum_seed = 0;
 int inplace = 0;
@@ -400,6 +401,7 @@ void usage(enum logcode F)
   rprintf(F,"     --files-from=FILE       read list of source-file names from FILE\n");
   rprintf(F," -0, --from0                 all *-from/filter files are delimited by 0s\n");
   rprintf(F," -s, --protect-args          no space-splitting; only wildcard special-chars\n");
+  rprintf(F,"     --ignore-case           ignore case when comparing filenames\n");
   rprintf(F,"     --address=ADDRESS       bind address for outgoing socket to daemon\n");
   rprintf(F,"     --port=PORT             specify double-colon alternate port number\n");
   rprintf(F,"     --sockopts=OPTIONS      specify custom TCP options\n");
@@ -593,6 +595,8 @@ static struct poptOption long_options[] 
   {"read-batch",       0,  POPT_ARG_STRING, &batch_name, OPT_READ_BATCH, 0, 0 },
   {"write-batch",      0,  POPT_ARG_STRING, &batch_name, OPT_WRITE_BATCH, 0, 0 },
   {"only-write-batch", 0,  POPT_ARG_STRING, &batch_name, OPT_ONLY_WRITE_BATCH, 0, 0 },
+  {"ignore-case",      0,  POPT_ARG_VAL,    &ignore_case, 1, 0, 0 },
+  {"no-ignore-case",   0,  POPT_ARG_VAL,    &ignore_case, 0, 0, 0 },
   {"files-from",       0,  POPT_ARG_STRING, &files_from, 0, 0, 0 },
   {"from0",           '0', POPT_ARG_VAL,    &eol_nulls, 1, 0, 0},
   {"no-from0",         0,  POPT_ARG_VAL,    &eol_nulls, 0, 0, 0},
@@ -1909,6 +1913,9 @@ void server_options(char **args, int *ar
 		args[ac++] = arg;
 	}
 
+	if (ignore_case)
+		args[ac++] = "--ignore-case";
+
 	if (partial_dir && am_sender) {
 		if (partial_dir != tmp_partialdir) {
 			args[ac++] = "--partial-dir";
--- old/rsync.yo
+++ new/rsync.yo
@@ -402,6 +402,7 @@ to the detailed description below for a 
      --files-from=FILE       read list of source-file names from FILE
  -0, --from0                 all *from/filter files are delimited by 0s
  -s, --protect-args          no space-splitting; wildcard chars only
+     --ignore-case           ignore case when comparing filenames
      --address=ADDRESS       bind address for outgoing socket to daemon
      --port=PORT             specify double-colon alternate port number
      --sockopts=OPTIONS      specify custom TCP options
@@ -1383,6 +1384,10 @@ If you use this option with bf(--iconv),
 from the local to the remote character set.  The translation happens before
 wild-cards are expanded.  See also the bf(--files-from) option.
 
+dit(bf(--ignore-case)) This option tells rsync to ignore upper-/lower-case
+differences when comparing filenames.  This can avoid problems when sending
+files to a filesystem that ignores these differences.
+
 dit(bf(-T, --temp-dir=DIR)) This option instructs rsync to use DIR as a
 scratch directory when creating temporary copies of the files transferred
 on the receiving side.  The default behavior is to create each temporary
--- old/wildtest.c
+++ new/wildtest.c
@@ -31,6 +31,7 @@ int fnmatch_errors = 0;
 #endif
 
 int wildmatch_errors = 0;
+int ignore_case = 0;
 
 typedef char bool;
Commit	Line	Data
992633a5 WD	1	This adds the --ignore-case option, which makes rsync compare filenames
992633a5 WD	2	in a case-insensitive manner.
8a524ae0	3
03019e41 WD	4	To use this patch, run these commands for a successful build:
	5
	6	patch -p1 <patches/ignore-case.diff
	7	./configure (optional if already run)
	8	make
	9
0f8eae47 WD	10	TODO:
	11
	12	- Make this code handle multibyte character encodings, and honor the
	13	--iconv setting when converting case.
	14
6100b16e WD	15	--- old/exclude.c
6100b16e WD	16	+++ new/exclude.c
5befb079	17	@@ -583,16 +583,15 @@ static int rule_matches(char *name, stru
6100b16e WD	18	if (litmatch_array(pattern, strings, slash_handling))
	19	return ret_match;
	20	} else if (anchored_match) {
	21	- if (strcmp(name, pattern) == 0)
5befb079	22	+ if (ic_strEQ(name, pattern))
6100b16e WD	23	return ret_match;
	24	} else {
	25	int l1 = strlen(name);
	26	int l2 = strlen(pattern);
	27	- if (l2 <= l1 &&
	28	- strcmp(name+(l1-l2),pattern) == 0 &&
	29	- (l1==l2 \|\| name[l1-(l2+1)] == '/')) {
	30	+ if (l2 <= l1
5befb079	31	+ && ic_strEQ(name + (l1-l2), pattern)
6100b16e WD	32	+ && (l1 == l2 \|\| name[l1 - (l2+1)] == '/'))
	33	return ret_match;
	34	- }
	35	}
	36
	37	return !ret_match;
9a7eef96 WD	38	--- old/flist.c
9a7eef96 WD	39	+++ new/flist.c
f80ebf7e	40	@@ -35,6 +35,7 @@ extern int inc_recurse;
9c657c9f WD	41	extern int do_progress;
	42	extern int always_checksum;
	43	extern int module_id;
	44	+extern int ignore_case;
	45	extern int ignore_errors;
	46	extern int numeric_ids;
	47	extern int recurse;
5befb079	48	@@ -2570,6 +2571,7 @@ int f_name_cmp(const struct file_struct
f80ebf7e WD	49	{
f80ebf7e WD	50	int dif;
7bfcb297 WD	51	const uchar c1, c2;
7bfcb297 WD	52	+ uchar ch1, ch2;
f80ebf7e WD	53	enum fnc_state state1, state2;
	54	enum fnc_type type1, type2;
	55	enum fnc_type t_path = protocol_version >= 29 ? t_PATH : t_ITEM;
5befb079	56	@@ -2680,7 +2682,15 @@ int f_name_cmp(const struct file_struct
3f053c45 WD	57	if (type1 != type2)
	58	return type1 == t_PATH ? 1 : -1;
	59	}
	60	- } while ((dif = (int)c1++ - (int)c2++) == 0);
5befb079 WD	61	+ ch1 = *c1++;
5befb079 WD	62	+ ch2 = *c2++;
9c657c9f	63	+ if (ignore_case) {
5befb079 WD	64	+ if (isupper(ch1))
	65	+ ch1 = tolower(ch1);
	66	+ if (isupper(ch2))
	67	+ ch2 = tolower(ch2);
f80ebf7e	68	+ }
5befb079	69	+ } while ((dif = (int)ch1 - (int)ch2) == 0);
3f053c45 WD	70
	71	return dif;
	72	}
5befb079 WD	73	--- old/ifuncs.h
	74	+++ new/ifuncs.h
	75	@@ -98,3 +98,38 @@ toUpper(const char *ptr)
	76	{
	77	return toupper((unsigned char )ptr);
	78	}
	79	+
	80	+static inline int
	81	+strEQ(const char s1, const char s2)
	82	+{
	83	+ return strcmp(s1, s2) == 0;
	84	+}
	85	+
	86	+static inline int
	87	+strnEQ(const char s1, const char s2, size_t n)
	88	+{
	89	+ return strncmp(s1, s2, n) == 0;
	90	+}
	91	+
	92	+static inline int
	93	+ic_strEQ(const char s1, const char s2)
	94	+{
	95	+ extern int ignore_case;
	96	+ if (ignore_case)
	97	+ return strcasecmp(s1, s2) == 0;
	98	+ return strcmp(s1, s2) == 0;
	99	+}
	100	+
	101	+static inline int
	102	+ic_strnEQ(const char s1, const char s2, size_t n)
	103	+{
	104	+ extern int ignore_case;
	105	+ if (ignore_case)
	106	+ return strcasecmp(s1, s2, n) == 0;
	107	+ return strcmp(s1, s2, n) == 0;
	108	+}
	109	+
	110	+#define strNE(s1,s2) (!strEQ(s1,s2))
	111	+#define strnNE(s1,s2,n) (!strnEQ(s1,s2,n))
	112	+#define ic_strNE(s1,s2) (!ic_strEQ(s1,s2))
	113	+#define ic_strnNE(s1,s2) (!ic_strnEQ(s1,s2,n))
9a7eef96 WD	114	--- old/lib/wildmatch.c
9a7eef96 WD	115	+++ new/lib/wildmatch.c
992633a5 WD	116	@@ -53,6 +53,8 @@
	117	#define ISUPPER(c) (ISASCII(c) && isupper(c))
	118	#define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
	119
	120	+extern int ignore_case;
	121	+
	122	#ifdef WILD_TEST_ITERATIONS
	123	int wildmatch_iteration_count;
	124	#endif
66e2df0e	125	@@ -72,6 +74,8 @@ static int dowild(const uchar *p, const
ffadcb36 WD	126	for ( ; (p_ch = *p) != '\0'; text++, p++) {
	127	int matched, special;
	128	uchar t_ch, prev_ch;
	129	+ if (ignore_case && ISUPPER(p_ch))
	130	+ p_ch = tolower(p_ch);
66e2df0e WD	131	while ((t_ch = *text) == '\0') {
	132	if (*a == NULL) {
	133	if (p_ch != '*')
6100b16e WD	134	@@ -237,12 +241,21 @@ static int dowild(const uchar *p, const
	135	* of "text" and any strings in array "a". */
	136	static int doliteral(const uchar s, const uchar text, const ucharconst a)
	137	{
	138	+ uchar s_ch, t_ch;
	139	for ( ; *s != '\0'; text++, s++) {
	140	while (*text == '\0') {
	141	if ((text = *a++) == NULL)
	142	return FALSE;
	143	}
	144	- if (text != s)
	145	+ s_ch = *s;
	146	+ t_ch = *text;
	147	+ if (ignore_case) {
	148	+ if (ISUPPER(s_ch))
	149	+ s_ch = tolower(s_ch);
	150	+ if (ISUPPER(t_ch))
	151	+ t_ch = tolower(t_ch);
	152	+ }
	153	+ if (t_ch != s_ch)
	154	return FALSE;
	155	}
	156
	157	@@ -288,10 +301,14 @@ static const uchar *trailing_N_elements(
ffadcb36 WD	158	int wildmatch(const char pattern, const char text)
ffadcb36 WD	159	{
66e2df0e	160	static const uchar nomore[1]; / A NULL pointer. */
ffadcb36 WD	161	+ int ret;
	162	#ifdef WILD_TEST_ITERATIONS
	163	wildmatch_iteration_count = 0;
	164	#endif
66e2df0e	165	- return dowild((const uchar)pattern, (const uchar)text, nomore) == TRUE;
ffadcb36	166	+ force_lower_case = ignore_case;
66e2df0e	167	+ ret = dowild((const uchar)pattern, (const uchar)text, nomore) == TRUE;
ffadcb36	168	+ force_lower_case = 0;
66e2df0e	169	+ return ret;
ffadcb36 WD	170	}
	171
	172	/* Match the "pattern" against the forced-to-lower-case "text" string. */
c05e712e WD	173	@@ -331,12 +348,14 @@ int wildmatch_array(const char *pattern,
	174	if (!text)
	175	return FALSE;
	176
	177	+ force_lower_case = ignore_case;
	178	+
	179	if ((matched = dowild(p, text, a)) != TRUE && where < 0
	180	&& matched != ABORT_ALL) {
	181	while (1) {
	182	if (*text == '\0') {
	183	if ((text = (uchar)a++) == NULL)
	184	- return FALSE;
	185	+ break;
	186	continue;
	187	}
	188	if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE
	189	@@ -344,6 +363,9 @@ int wildmatch_array(const char *pattern,
	190	break;
	191	}
	192	}
	193	+
	194	+ force_lower_case = 0;
	195	+
	196	return matched == TRUE;
	197	}
	198
9a7eef96 WD	199	--- old/options.c
9a7eef96 WD	200	+++ new/options.c
5befb079	201	@@ -115,6 +115,7 @@ OFF_T max_size = 0;
56208f7b	202	OFF_T min_size = 0;
5f7bb027 WD	203	int ignore_errors = 0;
	204	int modify_window = 0;
	205	+int ignore_case = 0;
	206	int blocking_io = -1;
	207	int checksum_seed = 0;
f6c3b300	208	int inplace = 0;
5befb079	209	@@ -400,6 +401,7 @@ void usage(enum logcode F)
a7219d20	210	rprintf(F," --files-from=FILE read list of source-file names from FILE\n");
896871f8	211	rprintf(F," -0, --from0 all *-from/filter files are delimited by 0s\n");
790ba11a	212	rprintf(F," -s, --protect-args no space-splitting; only wildcard special-chars\n");
5befb079 WD	213	+ rprintf(F," --ignore-case ignore case when comparing filenames\n");
	214	rprintf(F," --address=ADDRESS bind address for outgoing socket to daemon\n");
	215	rprintf(F," --port=PORT specify double-colon alternate port number\n");
	216	rprintf(F," --sockopts=OPTIONS specify custom TCP options\n");
7bfcb297	217	@@ -593,6 +595,8 @@ static struct poptOption long_options[]
6cbbe66d WD	218	{"read-batch", 0, POPT_ARG_STRING, &batch_name, OPT_READ_BATCH, 0, 0 },
6cbbe66d WD	219	{"write-batch", 0, POPT_ARG_STRING, &batch_name, OPT_WRITE_BATCH, 0, 0 },
edea1111	220	{"only-write-batch", 0, POPT_ARG_STRING, &batch_name, OPT_ONLY_WRITE_BATCH, 0, 0 },
2c20ee4e WD	221	+ {"ignore-case", 0, POPT_ARG_VAL, &ignore_case, 1, 0, 0 },
2c20ee4e WD	222	+ {"no-ignore-case", 0, POPT_ARG_VAL, &ignore_case, 0, 0, 0 },
6cbbe66d WD	223	{"files-from", 0, POPT_ARG_STRING, &files_from, 0, 0, 0 },
	224	{"from0", '0', POPT_ARG_VAL, &eol_nulls, 1, 0, 0},
	225	{"no-from0", 0, POPT_ARG_VAL, &eol_nulls, 0, 0, 0},
5befb079	226	@@ -1909,6 +1913,9 @@ void server_options(char *args, int ar
5f7bb027 WD	227	args[ac++] = arg;
5f7bb027 WD	228	}
7628f156	229
5f7bb027 WD	230	+ if (ignore_case)
5f7bb027 WD	231	+ args[ac++] = "--ignore-case";
7628f156	232	+
afbebe13	233	if (partial_dir && am_sender) {
9a7eef96	234	if (partial_dir != tmp_partialdir) {
27a7053c	235	args[ac++] = "--partial-dir";
5befb079 WD	236	--- old/rsync.yo
	237	+++ new/rsync.yo
	238	@@ -402,6 +402,7 @@ to the detailed description below for a
	239	--files-from=FILE read list of source-file names from FILE
	240	-0, --from0 all *from/filter files are delimited by 0s
	241	-s, --protect-args no space-splitting; wildcard chars only
	242	+ --ignore-case ignore case when comparing filenames
	243	--address=ADDRESS bind address for outgoing socket to daemon
	244	--port=PORT specify double-colon alternate port number
	245	--sockopts=OPTIONS specify custom TCP options
	246	@@ -1383,6 +1384,10 @@ If you use this option with bf(--iconv),
	247	from the local to the remote character set. The translation happens before
	248	wild-cards are expanded. See also the bf(--files-from) option.
	249
	250	+dit(bf(--ignore-case)) This option tells rsync to ignore upper-/lower-case
	251	+differences when comparing filenames. This can avoid problems when sending
	252	+files to a filesystem that ignores these differences.
	253	+
	254	dit(bf(-T, --temp-dir=DIR)) This option instructs rsync to use DIR as a
	255	scratch directory when creating temporary copies of the files transferred
	256	on the receiving side. The default behavior is to create each temporary
9a7eef96 WD	257	--- old/wildtest.c
9a7eef96 WD	258	+++ new/wildtest.c
6100b16e	259	@@ -31,6 +31,7 @@ int fnmatch_errors = 0;
992633a5 WD	260	#endif
	261
	262	int wildmatch_errors = 0;
	263	+int ignore_case = 0;
	264
	265	typedef char bool;
5f7bb027	266