[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
@@ -21,6 +21,7 @@
  */
 
 #include "rsync.h"
+#include "ifuncs.h"
 
 extern int verbose;
 extern int am_server;
@@ -583,16 +584,15 @@ 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 strncasecmp(s1, s2, n) == 0;
+	return strncmp(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
c82285d5 WD	17	@@ -21,6 +21,7 @@
	18	*/
	19
	20	#include "rsync.h"
	21	+#include "ifuncs.h"
	22
	23	extern int verbose;
	24	extern int am_server;
	25	@@ -583,16 +584,15 @@ static int rule_matches(char *name, stru
6100b16e WD	26	if (litmatch_array(pattern, strings, slash_handling))
	27	return ret_match;
	28	} else if (anchored_match) {
	29	- if (strcmp(name, pattern) == 0)
5befb079	30	+ if (ic_strEQ(name, pattern))
6100b16e WD	31	return ret_match;
	32	} else {
	33	int l1 = strlen(name);
	34	int l2 = strlen(pattern);
	35	- if (l2 <= l1 &&
	36	- strcmp(name+(l1-l2),pattern) == 0 &&
	37	- (l1==l2 \|\| name[l1-(l2+1)] == '/')) {
	38	+ if (l2 <= l1
5befb079	39	+ && ic_strEQ(name + (l1-l2), pattern)
6100b16e WD	40	+ && (l1 == l2 \|\| name[l1 - (l2+1)] == '/'))
	41	return ret_match;
	42	- }
	43	}
	44
	45	return !ret_match;
9a7eef96 WD	46	--- old/flist.c
9a7eef96 WD	47	+++ new/flist.c
f80ebf7e	48	@@ -35,6 +35,7 @@ extern int inc_recurse;
9c657c9f WD	49	extern int do_progress;
	50	extern int always_checksum;
	51	extern int module_id;
	52	+extern int ignore_case;
	53	extern int ignore_errors;
	54	extern int numeric_ids;
	55	extern int recurse;
5befb079	56	@@ -2570,6 +2571,7 @@ int f_name_cmp(const struct file_struct
f80ebf7e WD	57	{
f80ebf7e WD	58	int dif;
7bfcb297 WD	59	const uchar c1, c2;
7bfcb297 WD	60	+ uchar ch1, ch2;
f80ebf7e WD	61	enum fnc_state state1, state2;
	62	enum fnc_type type1, type2;
	63	enum fnc_type t_path = protocol_version >= 29 ? t_PATH : t_ITEM;
5befb079	64	@@ -2680,7 +2682,15 @@ int f_name_cmp(const struct file_struct
3f053c45 WD	65	if (type1 != type2)
	66	return type1 == t_PATH ? 1 : -1;
	67	}
	68	- } while ((dif = (int)c1++ - (int)c2++) == 0);
5befb079 WD	69	+ ch1 = *c1++;
5befb079 WD	70	+ ch2 = *c2++;
9c657c9f	71	+ if (ignore_case) {
5befb079 WD	72	+ if (isupper(ch1))
	73	+ ch1 = tolower(ch1);
	74	+ if (isupper(ch2))
	75	+ ch2 = tolower(ch2);
f80ebf7e	76	+ }
5befb079	77	+ } while ((dif = (int)ch1 - (int)ch2) == 0);
3f053c45 WD	78
	79	return dif;
	80	}
5befb079 WD	81	--- old/ifuncs.h
	82	+++ new/ifuncs.h
	83	@@ -98,3 +98,38 @@ toUpper(const char *ptr)
	84	{
	85	return toupper((unsigned char )ptr);
	86	}
	87	+
	88	+static inline int
	89	+strEQ(const char s1, const char s2)
	90	+{
	91	+ return strcmp(s1, s2) == 0;
	92	+}
	93	+
	94	+static inline int
	95	+strnEQ(const char s1, const char s2, size_t n)
	96	+{
	97	+ return strncmp(s1, s2, n) == 0;
	98	+}
	99	+
	100	+static inline int
	101	+ic_strEQ(const char s1, const char s2)
	102	+{
	103	+ extern int ignore_case;
	104	+ if (ignore_case)
	105	+ return strcasecmp(s1, s2) == 0;
	106	+ return strcmp(s1, s2) == 0;
	107	+}
	108	+
	109	+static inline int
	110	+ic_strnEQ(const char s1, const char s2, size_t n)
	111	+{
	112	+ extern int ignore_case;
	113	+ if (ignore_case)
c82285d5 WD	114	+ return strncasecmp(s1, s2, n) == 0;
c82285d5 WD	115	+ return strncmp(s1, s2, n) == 0;
5befb079 WD	116	+}
	117	+
	118	+#define strNE(s1,s2) (!strEQ(s1,s2))
	119	+#define strnNE(s1,s2,n) (!strnEQ(s1,s2,n))
	120	+#define ic_strNE(s1,s2) (!ic_strEQ(s1,s2))
	121	+#define ic_strnNE(s1,s2) (!ic_strnEQ(s1,s2,n))
9a7eef96 WD	122	--- old/lib/wildmatch.c
9a7eef96 WD	123	+++ new/lib/wildmatch.c
992633a5 WD	124	@@ -53,6 +53,8 @@
	125	#define ISUPPER(c) (ISASCII(c) && isupper(c))
	126	#define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
	127
	128	+extern int ignore_case;
	129	+
	130	#ifdef WILD_TEST_ITERATIONS
	131	int wildmatch_iteration_count;
	132	#endif
66e2df0e	133	@@ -72,6 +74,8 @@ static int dowild(const uchar *p, const
ffadcb36 WD	134	for ( ; (p_ch = *p) != '\0'; text++, p++) {
	135	int matched, special;
	136	uchar t_ch, prev_ch;
	137	+ if (ignore_case && ISUPPER(p_ch))
	138	+ p_ch = tolower(p_ch);
66e2df0e WD	139	while ((t_ch = *text) == '\0') {
	140	if (*a == NULL) {
	141	if (p_ch != '*')
6100b16e WD	142	@@ -237,12 +241,21 @@ static int dowild(const uchar *p, const
	143	* of "text" and any strings in array "a". */
	144	static int doliteral(const uchar s, const uchar text, const ucharconst a)
	145	{
	146	+ uchar s_ch, t_ch;
	147	for ( ; *s != '\0'; text++, s++) {
	148	while (*text == '\0') {
	149	if ((text = *a++) == NULL)
	150	return FALSE;
	151	}
	152	- if (text != s)
	153	+ s_ch = *s;
	154	+ t_ch = *text;
	155	+ if (ignore_case) {
	156	+ if (ISUPPER(s_ch))
	157	+ s_ch = tolower(s_ch);
	158	+ if (ISUPPER(t_ch))
	159	+ t_ch = tolower(t_ch);
	160	+ }
	161	+ if (t_ch != s_ch)
	162	return FALSE;
	163	}
	164
	165	@@ -288,10 +301,14 @@ static const uchar *trailing_N_elements(
ffadcb36 WD	166	int wildmatch(const char pattern, const char text)
ffadcb36 WD	167	{
66e2df0e	168	static const uchar nomore[1]; / A NULL pointer. */
ffadcb36 WD	169	+ int ret;
	170	#ifdef WILD_TEST_ITERATIONS
	171	wildmatch_iteration_count = 0;
	172	#endif
66e2df0e	173	- return dowild((const uchar)pattern, (const uchar)text, nomore) == TRUE;
ffadcb36	174	+ force_lower_case = ignore_case;
66e2df0e	175	+ ret = dowild((const uchar)pattern, (const uchar)text, nomore) == TRUE;
ffadcb36	176	+ force_lower_case = 0;
66e2df0e	177	+ return ret;
ffadcb36 WD	178	}
	179
	180	/* Match the "pattern" against the forced-to-lower-case "text" string. */
c05e712e WD	181	@@ -331,12 +348,14 @@ int wildmatch_array(const char *pattern,
	182	if (!text)
	183	return FALSE;
	184
	185	+ force_lower_case = ignore_case;
	186	+
	187	if ((matched = dowild(p, text, a)) != TRUE && where < 0
	188	&& matched != ABORT_ALL) {
	189	while (1) {
	190	if (*text == '\0') {
	191	if ((text = (uchar)a++) == NULL)
	192	- return FALSE;
	193	+ break;
	194	continue;
	195	}
	196	if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE
	197	@@ -344,6 +363,9 @@ int wildmatch_array(const char *pattern,
	198	break;
	199	}
	200	}
	201	+
	202	+ force_lower_case = 0;
	203	+
	204	return matched == TRUE;
	205	}
	206
9a7eef96 WD	207	--- old/options.c
9a7eef96 WD	208	+++ new/options.c
5befb079	209	@@ -115,6 +115,7 @@ OFF_T max_size = 0;
56208f7b	210	OFF_T min_size = 0;
5f7bb027 WD	211	int ignore_errors = 0;
	212	int modify_window = 0;
	213	+int ignore_case = 0;
	214	int blocking_io = -1;
	215	int checksum_seed = 0;
f6c3b300	216	int inplace = 0;
5befb079	217	@@ -400,6 +401,7 @@ void usage(enum logcode F)
a7219d20	218	rprintf(F," --files-from=FILE read list of source-file names from FILE\n");
896871f8	219	rprintf(F," -0, --from0 all *-from/filter files are delimited by 0s\n");
790ba11a	220	rprintf(F," -s, --protect-args no space-splitting; only wildcard special-chars\n");
5befb079 WD	221	+ rprintf(F," --ignore-case ignore case when comparing filenames\n");
	222	rprintf(F," --address=ADDRESS bind address for outgoing socket to daemon\n");
	223	rprintf(F," --port=PORT specify double-colon alternate port number\n");
	224	rprintf(F," --sockopts=OPTIONS specify custom TCP options\n");
7bfcb297	225	@@ -593,6 +595,8 @@ static struct poptOption long_options[]
6cbbe66d WD	226	{"read-batch", 0, POPT_ARG_STRING, &batch_name, OPT_READ_BATCH, 0, 0 },
6cbbe66d WD	227	{"write-batch", 0, POPT_ARG_STRING, &batch_name, OPT_WRITE_BATCH, 0, 0 },
edea1111	228	{"only-write-batch", 0, POPT_ARG_STRING, &batch_name, OPT_ONLY_WRITE_BATCH, 0, 0 },
2c20ee4e WD	229	+ {"ignore-case", 0, POPT_ARG_VAL, &ignore_case, 1, 0, 0 },
2c20ee4e WD	230	+ {"no-ignore-case", 0, POPT_ARG_VAL, &ignore_case, 0, 0, 0 },
6cbbe66d WD	231	{"files-from", 0, POPT_ARG_STRING, &files_from, 0, 0, 0 },
	232	{"from0", '0', POPT_ARG_VAL, &eol_nulls, 1, 0, 0},
	233	{"no-from0", 0, POPT_ARG_VAL, &eol_nulls, 0, 0, 0},
5befb079	234	@@ -1909,6 +1913,9 @@ void server_options(char *args, int ar
5f7bb027 WD	235	args[ac++] = arg;
5f7bb027 WD	236	}
7628f156	237
5f7bb027 WD	238	+ if (ignore_case)
5f7bb027 WD	239	+ args[ac++] = "--ignore-case";
7628f156	240	+
afbebe13	241	if (partial_dir && am_sender) {
9a7eef96	242	if (partial_dir != tmp_partialdir) {
27a7053c	243	args[ac++] = "--partial-dir";
5befb079 WD	244	--- old/rsync.yo
	245	+++ new/rsync.yo
	246	@@ -402,6 +402,7 @@ to the detailed description below for a
	247	--files-from=FILE read list of source-file names from FILE
	248	-0, --from0 all *from/filter files are delimited by 0s
	249	-s, --protect-args no space-splitting; wildcard chars only
	250	+ --ignore-case ignore case when comparing filenames
	251	--address=ADDRESS bind address for outgoing socket to daemon
	252	--port=PORT specify double-colon alternate port number
	253	--sockopts=OPTIONS specify custom TCP options
	254	@@ -1383,6 +1384,10 @@ If you use this option with bf(--iconv),
	255	from the local to the remote character set. The translation happens before
	256	wild-cards are expanded. See also the bf(--files-from) option.
	257
	258	+dit(bf(--ignore-case)) This option tells rsync to ignore upper-/lower-case
	259	+differences when comparing filenames. This can avoid problems when sending
	260	+files to a filesystem that ignores these differences.
	261	+
	262	dit(bf(-T, --temp-dir=DIR)) This option instructs rsync to use DIR as a
	263	scratch directory when creating temporary copies of the files transferred
	264	on the receiving side. The default behavior is to create each temporary
9a7eef96 WD	265	--- old/wildtest.c
9a7eef96 WD	266	+++ new/wildtest.c
6100b16e	267	@@ -31,6 +31,7 @@ int fnmatch_errors = 0;
992633a5 WD	268	#endif
	269
	270	int wildmatch_errors = 0;
	271	+int ignore_case = 0;
	272
	273	typedef char bool;
5f7bb027	274