4 * Copyright (c) 2006 Junio C Hamano
11 #include "tree-walk.h"
16 * git grep pathspecs are somewhat different from diff-tree pathspecs;
17 * pathname wildcards are allowed.
19 static int pathspec_matches(const char **paths
, const char *name
)
22 if (!paths
|| !*paths
)
24 namelen
= strlen(name
);
25 for (i
= 0; paths
[i
]; i
++) {
26 const char *match
= paths
[i
];
27 int matchlen
= strlen(match
);
28 const char *cp
, *meta
;
31 ((matchlen
<= namelen
) &&
32 !strncmp(name
, match
, matchlen
) &&
33 (match
[matchlen
-1] == '/' ||
34 name
[matchlen
] == '\0' || name
[matchlen
] == '/')))
36 if (!fnmatch(match
, name
, 0))
38 if (name
[namelen
-1] != '/')
41 /* We are being asked if the directory ("name") is worth
44 * Find the longest leading directory name that does
45 * not have metacharacter in the pathspec; the name
46 * we are looking at must overlap with that directory.
48 for (cp
= match
, meta
= NULL
; cp
- match
< matchlen
; cp
++) {
50 if (ch
== '*' || ch
== '[' || ch
== '?') {
56 meta
= cp
; /* fully literal */
58 if (namelen
<= meta
- match
) {
59 /* Looking at "Documentation/" and
60 * the pattern says "Documentation/howto/", or
61 * "Documentation/diff*.txt". The name we
62 * have should match prefix.
64 if (!memcmp(match
, name
, namelen
))
69 if (meta
- match
< namelen
) {
70 /* Looking at "Documentation/howto/" and
71 * the pattern says "Documentation/h*";
72 * match up to "Do.../h"; this avoids descending
73 * into "Documentation/technical/".
75 if (!memcmp(match
, name
, meta
- match
))
83 static int grep_sha1(struct grep_opt
*opt
, const unsigned char *sha1
, const char *name
, int tree_name_len
)
87 enum object_type type
;
91 data
= read_sha1_file(sha1
, &type
, &size
);
93 error("'%s': unable to read %s", name
, sha1_to_hex(sha1
));
96 if (opt
->relative
&& opt
->prefix_length
) {
97 static char name_buf
[PATH_MAX
];
99 int name_len
= strlen(name
) - opt
->prefix_length
+ 1;
102 name
+= opt
->prefix_length
;
104 if (ARRAY_SIZE(name_buf
) <= name_len
)
105 cp
= to_free
= xmalloc(name_len
);
108 memcpy(cp
, name
, tree_name_len
);
109 strcpy(cp
+ tree_name_len
,
110 name
+ tree_name_len
+ opt
->prefix_length
);
114 hit
= grep_buffer(opt
, name
, data
, size
);
120 static int grep_file(struct grep_opt
*opt
, const char *filename
)
127 if (lstat(filename
, &st
) < 0) {
130 error("'%s': %s", filename
, strerror(errno
));
134 return 0; /* empty file -- no grep hit */
135 if (!S_ISREG(st
.st_mode
))
137 sz
= xsize_t(st
.st_size
);
138 i
= open(filename
, O_RDONLY
);
141 data
= xmalloc(sz
+ 1);
142 if (st
.st_size
!= read_in_full(i
, data
, sz
)) {
143 error("'%s': short read %s", filename
, strerror(errno
));
149 if (opt
->relative
&& opt
->prefix_length
)
150 filename
+= opt
->prefix_length
;
151 i
= grep_buffer(opt
, filename
, data
, sz
);
157 static int exec_grep(int argc
, const char **argv
)
167 execvp("grep", (char **) argv
);
170 while (waitpid(pid
, &status
, 0) < 0) {
175 if (WIFEXITED(status
)) {
176 if (!WEXITSTATUS(status
))
185 #define push_arg(a) do { \
186 if (nr < MAXARGS) argv[nr++] = (a); \
187 else die("maximum number of args exceeded"); \
191 * If you send a singleton filename to grep, it does not give
192 * the name of the file. GNU grep has "-H" but we would want
193 * that behaviour in a portable way.
195 * So we keep two pathnames in argv buffer unsent to grep in
196 * the main loop if we need to do more than one grep.
198 static int flush_grep(struct grep_opt
*opt
,
199 int argc
, int arg0
, const char **argv
, int *kept
)
202 int count
= argc
- arg0
;
203 const char *kept_0
= NULL
;
207 * Because we keep at least 2 paths in the call from
208 * the main loop (i.e. kept != NULL), and MAXARGS is
209 * far greater than 2, this usually is a call to
210 * conclude the grep. However, the user could attempt
211 * to overflow the argv buffer by giving too many
212 * options to leave very small number of real
213 * arguments even for the call in the main loop.
216 die("insanely many options to grep");
219 * If we have two or more paths, we do not have to do
220 * anything special, but we need to push /dev/null to
221 * get "-H" behaviour of GNU grep portably but when we
222 * are not doing "-l" nor "-L" nor "-c".
226 !opt
->unmatch_name_only
&&
228 argv
[argc
++] = "/dev/null";
235 * Called because we found many paths and haven't finished
236 * iterating over the cache yet. We keep two paths
237 * for the concluding call. argv[argc-2] and argv[argc-1]
238 * has the last two paths, so save the first one away,
239 * replace it with NULL while sending the list to grep,
240 * and recover them after we are done.
243 kept_0
= argv
[argc
-2];
248 status
= exec_grep(argc
, argv
);
252 * Then recover them. Now the last arg is beyond the
253 * terminating NULL which is at argc, and the second
254 * from the last is what we saved away in kept_0
256 argv
[arg0
++] = kept_0
;
257 argv
[arg0
] = argv
[argc
+1];
262 static int external_grep(struct grep_opt
*opt
, const char **paths
, int cached
)
264 int i
, nr
, argc
, hit
, len
, status
;
265 const char *argv
[MAXARGS
+1];
266 char randarg
[ARGBUF
];
267 char *argptr
= randarg
;
270 if (opt
->extended
|| (opt
->relative
&& opt
->prefix_length
))
280 if (opt
->regflags
& REG_EXTENDED
)
282 if (opt
->regflags
& REG_ICASE
)
284 if (opt
->word_regexp
)
288 if (opt
->unmatch_name_only
)
292 if (opt
->post_context
|| opt
->pre_context
) {
293 if (opt
->post_context
!= opt
->pre_context
) {
294 if (opt
->pre_context
) {
296 len
+= snprintf(argptr
, sizeof(randarg
)-len
,
297 "%u", opt
->pre_context
) + 1;
298 if (sizeof(randarg
) <= len
)
299 die("maximum length of args exceeded");
303 if (opt
->post_context
) {
305 len
+= snprintf(argptr
, sizeof(randarg
)-len
,
306 "%u", opt
->post_context
) + 1;
307 if (sizeof(randarg
) <= len
)
308 die("maximum length of args exceeded");
315 len
+= snprintf(argptr
, sizeof(randarg
)-len
,
316 "%u", opt
->post_context
) + 1;
317 if (sizeof(randarg
) <= len
)
318 die("maximum length of args exceeded");
323 for (p
= opt
->pattern_list
; p
; p
= p
->next
) {
325 push_arg(p
->pattern
);
330 for (i
= 0; i
< active_nr
; i
++) {
331 struct cache_entry
*ce
= active_cache
[i
];
334 if (!S_ISREG(ntohl(ce
->ce_mode
)))
336 if (!pathspec_matches(paths
, ce
->name
))
339 if (name
[0] == '-') {
340 int len
= ce_namelen(ce
);
341 name
= xmalloc(len
+ 3);
342 memcpy(name
, "./", 2);
343 memcpy(name
+ 2, ce
->name
, len
+ 1);
346 if (MAXARGS
<= argc
) {
347 status
= flush_grep(opt
, argc
, nr
, argv
, &kept
);
355 } while (i
< active_nr
&&
356 !strcmp(ce
->name
, active_cache
[i
]->name
));
357 i
--; /* compensate for loop control */
361 status
= flush_grep(opt
, argc
, nr
, argv
, NULL
);
369 static int grep_cache(struct grep_opt
*opt
, const char **paths
, int cached
)
377 * Use the external "grep" command for the case where
378 * we grep through the checked-out files. It tends to
379 * be a lot more optimized
382 hit
= external_grep(opt
, paths
, cached
);
388 for (nr
= 0; nr
< active_nr
; nr
++) {
389 struct cache_entry
*ce
= active_cache
[nr
];
390 if (!S_ISREG(ntohl(ce
->ce_mode
)))
392 if (!pathspec_matches(paths
, ce
->name
))
397 hit
|= grep_sha1(opt
, ce
->sha1
, ce
->name
, 0);
400 hit
|= grep_file(opt
, ce
->name
);
404 } while (nr
< active_nr
&&
405 !strcmp(ce
->name
, active_cache
[nr
]->name
));
406 nr
--; /* compensate for loop control */
409 free_grep_patterns(opt
);
413 static int grep_tree(struct grep_opt
*opt
, const char **paths
,
414 struct tree_desc
*tree
,
415 const char *tree_name
, const char *base
)
419 struct name_entry entry
;
421 int tn_len
= strlen(tree_name
);
422 char *path_buf
= xmalloc(PATH_MAX
+ tn_len
+ 100);
425 tn_len
= sprintf(path_buf
, "%s:", tree_name
);
426 down
= path_buf
+ tn_len
;
433 len
= strlen(path_buf
);
435 while (tree_entry(tree
, &entry
)) {
436 strcpy(path_buf
+ len
, entry
.path
);
438 if (S_ISDIR(entry
.mode
))
439 /* Match "abc/" against pathspec to
440 * decide if we want to descend into "abc"
443 strcpy(path_buf
+ len
+ tree_entry_len(entry
.path
, entry
.sha1
), "/");
445 if (!pathspec_matches(paths
, down
))
447 else if (S_ISREG(entry
.mode
))
448 hit
|= grep_sha1(opt
, entry
.sha1
, path_buf
, tn_len
);
449 else if (S_ISDIR(entry
.mode
)) {
450 enum object_type type
;
451 struct tree_desc sub
;
455 data
= read_sha1_file(entry
.sha1
, &type
, &size
);
457 die("unable to read tree (%s)",
458 sha1_to_hex(entry
.sha1
));
459 init_tree_desc(&sub
, data
, size
);
460 hit
|= grep_tree(opt
, paths
, &sub
, tree_name
, down
);
467 static int grep_object(struct grep_opt
*opt
, const char **paths
,
468 struct object
*obj
, const char *name
)
470 if (obj
->type
== OBJ_BLOB
)
471 return grep_sha1(opt
, obj
->sha1
, name
, 0);
472 if (obj
->type
== OBJ_COMMIT
|| obj
->type
== OBJ_TREE
) {
473 struct tree_desc tree
;
477 data
= read_object_with_reference(obj
->sha1
, tree_type
,
480 die("unable to read tree (%s)", sha1_to_hex(obj
->sha1
));
481 init_tree_desc(&tree
, data
, size
);
482 hit
= grep_tree(opt
, paths
, &tree
, name
, "");
486 die("unable to grep from object of type %s", typename(obj
->type
));
489 static const char builtin_grep_usage
[] =
490 "git-grep <option>* <rev>* [-e] <pattern> [<path>...]";
492 static const char emsg_invalid_context_len
[] =
493 "%s: invalid context length argument";
494 static const char emsg_missing_context_len
[] =
495 "missing context length argument";
496 static const char emsg_missing_argument
[] =
497 "option requires an argument -%s";
499 int cmd_grep(int argc
, const char **argv
, const char *prefix
)
503 int seen_dashdash
= 0;
505 struct object_array list
= { 0, 0, NULL
};
506 const char **paths
= NULL
;
509 memset(&opt
, 0, sizeof(opt
));
510 opt
.prefix_length
= (prefix
&& *prefix
) ? strlen(prefix
) : 0;
513 opt
.pattern_tail
= &opt
.pattern_list
;
514 opt
.regflags
= REG_NEWLINE
;
517 * If there is no -- then the paths must exist in the working
518 * tree. If there is no explicit pattern specified with -e or
519 * -f, we take the first unrecognized non option to be the
520 * pattern, but then what follows it must be zero or more
521 * valid refs up to the -- (if exists), and then existing
522 * paths. If there is an explicit pattern, then the first
523 * unrecognized non option is the beginning of the refs list
524 * that continues up to the -- (if exists), and then paths.
528 const char *arg
= argv
[1];
530 if (!strcmp("--cached", arg
)) {
534 if (!strcmp("-a", arg
) ||
535 !strcmp("--text", arg
)) {
536 opt
.binary
= GREP_BINARY_TEXT
;
539 if (!strcmp("-i", arg
) ||
540 !strcmp("--ignore-case", arg
)) {
541 opt
.regflags
|= REG_ICASE
;
544 if (!strcmp("-I", arg
)) {
545 opt
.binary
= GREP_BINARY_NOMATCH
;
548 if (!strcmp("-v", arg
) ||
549 !strcmp("--invert-match", arg
)) {
553 if (!strcmp("-E", arg
) ||
554 !strcmp("--extended-regexp", arg
)) {
555 opt
.regflags
|= REG_EXTENDED
;
558 if (!strcmp("-F", arg
) ||
559 !strcmp("--fixed-strings", arg
)) {
563 if (!strcmp("-G", arg
) ||
564 !strcmp("--basic-regexp", arg
)) {
565 opt
.regflags
&= ~REG_EXTENDED
;
568 if (!strcmp("-n", arg
)) {
572 if (!strcmp("-h", arg
)) {
576 if (!strcmp("-H", arg
)) {
580 if (!strcmp("-l", arg
) ||
581 !strcmp("--files-with-matches", arg
)) {
585 if (!strcmp("-L", arg
) ||
586 !strcmp("--files-without-match", arg
)) {
587 opt
.unmatch_name_only
= 1;
590 if (!strcmp("-c", arg
) ||
591 !strcmp("--count", arg
)) {
595 if (!strcmp("-w", arg
) ||
596 !strcmp("--word-regexp", arg
)) {
600 if (!prefixcmp(arg
, "-A") ||
601 !prefixcmp(arg
, "-B") ||
602 !prefixcmp(arg
, "-C") ||
603 (arg
[0] == '-' && '1' <= arg
[1] && arg
[1] <= '9')) {
607 case 'A': case 'B': case 'C':
610 die(emsg_missing_context_len
);
621 if (strtoul_ui(scan
, 10, &num
))
622 die(emsg_invalid_context_len
, scan
);
625 opt
.post_context
= num
;
629 opt
.post_context
= num
;
631 opt
.pre_context
= num
;
636 if (!strcmp("-f", arg
)) {
641 die(emsg_missing_argument
, arg
);
642 patterns
= fopen(argv
[1], "r");
644 die("'%s': %s", argv
[1], strerror(errno
));
645 while (fgets(buf
, sizeof(buf
), patterns
)) {
646 int len
= strlen(buf
);
647 if (buf
[len
-1] == '\n')
649 /* ignore empty line like grep does */
652 append_grep_pattern(&opt
, xstrdup(buf
),
661 if (!strcmp("--not", arg
)) {
662 append_grep_pattern(&opt
, arg
, "command line", 0,
666 if (!strcmp("--and", arg
)) {
667 append_grep_pattern(&opt
, arg
, "command line", 0,
671 if (!strcmp("--or", arg
))
672 continue; /* no-op */
673 if (!strcmp("(", arg
)) {
674 append_grep_pattern(&opt
, arg
, "command line", 0,
678 if (!strcmp(")", arg
)) {
679 append_grep_pattern(&opt
, arg
, "command line", 0,
683 if (!strcmp("--all-match", arg
)) {
687 if (!strcmp("-e", arg
)) {
689 append_grep_pattern(&opt
, argv
[1],
696 die(emsg_missing_argument
, arg
);
698 if (!strcmp("--full-name", arg
)) {
702 if (!strcmp("--", arg
)) {
703 /* later processing wants to have this at argv[1] */
709 usage(builtin_grep_usage
);
711 /* First unrecognized non-option token */
712 if (!opt
.pattern_list
) {
713 append_grep_pattern(&opt
, arg
, "command line", 0,
718 /* We are looking at the first path or rev;
719 * it is found at argv[1] after leaving the
727 if (!opt
.pattern_list
)
728 die("no pattern given.");
729 if ((opt
.regflags
!= REG_NEWLINE
) && opt
.fixed
)
730 die("cannot mix --fixed-strings and regexp");
731 compile_grep_patterns(&opt
);
733 /* Check revs and then paths */
734 for (i
= 1; i
< argc
; i
++) {
735 const char *arg
= argv
[i
];
736 unsigned char sha1
[20];
738 if (!get_sha1(arg
, sha1
)) {
739 struct object
*object
= parse_object(sha1
);
741 die("bad object %s", arg
);
742 add_object_array(object
, arg
, &list
);
745 if (!strcmp(arg
, "--")) {
752 /* The rest are paths */
753 if (!seen_dashdash
) {
755 for (j
= i
; j
< argc
; j
++)
756 verify_filename(prefix
, argv
[j
]);
760 paths
= get_pathspec(prefix
, argv
+ i
);
761 if (opt
.prefix_length
&& opt
.relative
) {
762 /* Make sure we do not get outside of paths */
763 for (i
= 0; paths
[i
]; i
++)
764 if (strncmp(prefix
, paths
[i
], opt
.prefix_length
))
765 die("git-grep: cannot generate relative filenames containing '..'");
769 paths
= xcalloc(2, sizeof(const char *));
775 return !grep_cache(&opt
, paths
, cached
);
778 die("both --cached and trees are given.");
780 for (i
= 0; i
< list
.nr
; i
++) {
781 struct object
*real_obj
;
782 real_obj
= deref_tag(list
.objects
[i
].item
, NULL
, 0);
783 if (grep_object(&opt
, paths
, real_obj
, list
.objects
[i
].name
))
786 free_grep_patterns(&opt
);