4 * Copyright (c) 2006 Junio C Hamano
11 #include "tree-walk.h"
15 #ifndef NO_EXTERNAL_GREP
17 #define NO_EXTERNAL_GREP 0
19 #define NO_EXTERNAL_GREP 1
24 * git grep pathspecs are somewhat different from diff-tree pathspecs;
25 * pathname wildcards are allowed.
27 static int pathspec_matches(const char **paths
, const char *name
)
30 if (!paths
|| !*paths
)
32 namelen
= strlen(name
);
33 for (i
= 0; paths
[i
]; i
++) {
34 const char *match
= paths
[i
];
35 int matchlen
= strlen(match
);
36 const char *cp
, *meta
;
39 ((matchlen
<= namelen
) &&
40 !strncmp(name
, match
, matchlen
) &&
41 (match
[matchlen
-1] == '/' ||
42 name
[matchlen
] == '\0' || name
[matchlen
] == '/')))
44 if (!fnmatch(match
, name
, 0))
46 if (name
[namelen
-1] != '/')
49 /* We are being asked if the directory ("name") is worth
52 * Find the longest leading directory name that does
53 * not have metacharacter in the pathspec; the name
54 * we are looking at must overlap with that directory.
56 for (cp
= match
, meta
= NULL
; cp
- match
< matchlen
; cp
++) {
58 if (ch
== '*' || ch
== '[' || ch
== '?') {
64 meta
= cp
; /* fully literal */
66 if (namelen
<= meta
- match
) {
67 /* Looking at "Documentation/" and
68 * the pattern says "Documentation/howto/", or
69 * "Documentation/diff*.txt". The name we
70 * have should match prefix.
72 if (!memcmp(match
, name
, namelen
))
77 if (meta
- match
< namelen
) {
78 /* Looking at "Documentation/howto/" and
79 * the pattern says "Documentation/h*";
80 * match up to "Do.../h"; this avoids descending
81 * into "Documentation/technical/".
83 if (!memcmp(match
, name
, meta
- match
))
91 static int grep_sha1(struct grep_opt
*opt
, const unsigned char *sha1
, const char *name
, int tree_name_len
)
95 enum object_type type
;
99 data
= read_sha1_file(sha1
, &type
, &size
);
101 error("'%s': unable to read %s", name
, sha1_to_hex(sha1
));
104 if (opt
->relative
&& opt
->prefix_length
) {
105 static char name_buf
[PATH_MAX
];
107 int name_len
= strlen(name
) - opt
->prefix_length
+ 1;
110 name
+= opt
->prefix_length
;
112 if (ARRAY_SIZE(name_buf
) <= name_len
)
113 cp
= to_free
= xmalloc(name_len
);
116 memcpy(cp
, name
, tree_name_len
);
117 strcpy(cp
+ tree_name_len
,
118 name
+ tree_name_len
+ opt
->prefix_length
);
122 hit
= grep_buffer(opt
, name
, data
, size
);
128 static int grep_file(struct grep_opt
*opt
, const char *filename
)
135 if (lstat(filename
, &st
) < 0) {
138 error("'%s': %s", filename
, strerror(errno
));
142 return 0; /* empty file -- no grep hit */
143 if (!S_ISREG(st
.st_mode
))
145 sz
= xsize_t(st
.st_size
);
146 i
= open(filename
, O_RDONLY
);
149 data
= xmalloc(sz
+ 1);
150 if (st
.st_size
!= read_in_full(i
, data
, sz
)) {
151 error("'%s': short read %s", filename
, strerror(errno
));
157 if (opt
->relative
&& opt
->prefix_length
)
158 filename
+= opt
->prefix_length
;
159 i
= grep_buffer(opt
, filename
, data
, sz
);
164 #if !NO_EXTERNAL_GREP
165 static int exec_grep(int argc
, const char **argv
)
175 execvp("grep", (char **) argv
);
178 while (waitpid(pid
, &status
, 0) < 0) {
183 if (WIFEXITED(status
)) {
184 if (!WEXITSTATUS(status
))
193 #define push_arg(a) do { \
194 if (nr < MAXARGS) argv[nr++] = (a); \
195 else die("maximum number of args exceeded"); \
199 * If you send a singleton filename to grep, it does not give
200 * the name of the file. GNU grep has "-H" but we would want
201 * that behaviour in a portable way.
203 * So we keep two pathnames in argv buffer unsent to grep in
204 * the main loop if we need to do more than one grep.
206 static int flush_grep(struct grep_opt
*opt
,
207 int argc
, int arg0
, const char **argv
, int *kept
)
210 int count
= argc
- arg0
;
211 const char *kept_0
= NULL
;
215 * Because we keep at least 2 paths in the call from
216 * the main loop (i.e. kept != NULL), and MAXARGS is
217 * far greater than 2, this usually is a call to
218 * conclude the grep. However, the user could attempt
219 * to overflow the argv buffer by giving too many
220 * options to leave very small number of real
221 * arguments even for the call in the main loop.
224 die("insanely many options to grep");
227 * If we have two or more paths, we do not have to do
228 * anything special, but we need to push /dev/null to
229 * get "-H" behaviour of GNU grep portably but when we
230 * are not doing "-l" nor "-L" nor "-c".
234 !opt
->unmatch_name_only
&&
236 argv
[argc
++] = "/dev/null";
243 * Called because we found many paths and haven't finished
244 * iterating over the cache yet. We keep two paths
245 * for the concluding call. argv[argc-2] and argv[argc-1]
246 * has the last two paths, so save the first one away,
247 * replace it with NULL while sending the list to grep,
248 * and recover them after we are done.
251 kept_0
= argv
[argc
-2];
256 status
= exec_grep(argc
, argv
);
260 * Then recover them. Now the last arg is beyond the
261 * terminating NULL which is at argc, and the second
262 * from the last is what we saved away in kept_0
264 argv
[arg0
++] = kept_0
;
265 argv
[arg0
] = argv
[argc
+1];
270 static int external_grep(struct grep_opt
*opt
, const char **paths
, int cached
)
272 int i
, nr
, argc
, hit
, len
, status
;
273 const char *argv
[MAXARGS
+1];
274 char randarg
[ARGBUF
];
275 char *argptr
= randarg
;
278 if (opt
->extended
|| (opt
->relative
&& opt
->prefix_length
))
288 if (opt
->regflags
& REG_EXTENDED
)
290 if (opt
->regflags
& REG_ICASE
)
292 if (opt
->word_regexp
)
296 if (opt
->unmatch_name_only
)
300 if (opt
->post_context
|| opt
->pre_context
) {
301 if (opt
->post_context
!= opt
->pre_context
) {
302 if (opt
->pre_context
) {
304 len
+= snprintf(argptr
, sizeof(randarg
)-len
,
305 "%u", opt
->pre_context
) + 1;
306 if (sizeof(randarg
) <= len
)
307 die("maximum length of args exceeded");
311 if (opt
->post_context
) {
313 len
+= snprintf(argptr
, sizeof(randarg
)-len
,
314 "%u", opt
->post_context
) + 1;
315 if (sizeof(randarg
) <= len
)
316 die("maximum length of args exceeded");
323 len
+= snprintf(argptr
, sizeof(randarg
)-len
,
324 "%u", opt
->post_context
) + 1;
325 if (sizeof(randarg
) <= len
)
326 die("maximum length of args exceeded");
331 for (p
= opt
->pattern_list
; p
; p
= p
->next
) {
333 push_arg(p
->pattern
);
338 for (i
= 0; i
< active_nr
; i
++) {
339 struct cache_entry
*ce
= active_cache
[i
];
342 if (!S_ISREG(ce
->ce_mode
))
344 if (!pathspec_matches(paths
, ce
->name
))
347 if (name
[0] == '-') {
348 int len
= ce_namelen(ce
);
349 name
= xmalloc(len
+ 3);
350 memcpy(name
, "./", 2);
351 memcpy(name
+ 2, ce
->name
, len
+ 1);
354 if (MAXARGS
<= argc
) {
355 status
= flush_grep(opt
, argc
, nr
, argv
, &kept
);
363 } while (i
< active_nr
&&
364 !strcmp(ce
->name
, active_cache
[i
]->name
));
365 i
--; /* compensate for loop control */
369 status
= flush_grep(opt
, argc
, nr
, argv
, NULL
);
377 static int grep_cache(struct grep_opt
*opt
, const char **paths
, int cached
)
383 #if !NO_EXTERNAL_GREP
385 * Use the external "grep" command for the case where
386 * we grep through the checked-out files. It tends to
387 * be a lot more optimized
390 hit
= external_grep(opt
, paths
, cached
);
396 for (nr
= 0; nr
< active_nr
; nr
++) {
397 struct cache_entry
*ce
= active_cache
[nr
];
398 if (!S_ISREG(ce
->ce_mode
))
400 if (!pathspec_matches(paths
, ce
->name
))
405 hit
|= grep_sha1(opt
, ce
->sha1
, ce
->name
, 0);
408 hit
|= grep_file(opt
, ce
->name
);
412 } while (nr
< active_nr
&&
413 !strcmp(ce
->name
, active_cache
[nr
]->name
));
414 nr
--; /* compensate for loop control */
417 free_grep_patterns(opt
);
421 static int grep_tree(struct grep_opt
*opt
, const char **paths
,
422 struct tree_desc
*tree
,
423 const char *tree_name
, const char *base
)
427 struct name_entry entry
;
429 int tn_len
= strlen(tree_name
);
430 char *path_buf
= xmalloc(PATH_MAX
+ tn_len
+ 100);
433 tn_len
= sprintf(path_buf
, "%s:", tree_name
);
434 down
= path_buf
+ tn_len
;
441 len
= strlen(path_buf
);
443 while (tree_entry(tree
, &entry
)) {
444 strcpy(path_buf
+ len
, entry
.path
);
446 if (S_ISDIR(entry
.mode
))
447 /* Match "abc/" against pathspec to
448 * decide if we want to descend into "abc"
451 strcpy(path_buf
+ len
+ tree_entry_len(entry
.path
, entry
.sha1
), "/");
453 if (!pathspec_matches(paths
, down
))
455 else if (S_ISREG(entry
.mode
))
456 hit
|= grep_sha1(opt
, entry
.sha1
, path_buf
, tn_len
);
457 else if (S_ISDIR(entry
.mode
)) {
458 enum object_type type
;
459 struct tree_desc sub
;
463 data
= read_sha1_file(entry
.sha1
, &type
, &size
);
465 die("unable to read tree (%s)",
466 sha1_to_hex(entry
.sha1
));
467 init_tree_desc(&sub
, data
, size
);
468 hit
|= grep_tree(opt
, paths
, &sub
, tree_name
, down
);
475 static int grep_object(struct grep_opt
*opt
, const char **paths
,
476 struct object
*obj
, const char *name
)
478 if (obj
->type
== OBJ_BLOB
)
479 return grep_sha1(opt
, obj
->sha1
, name
, 0);
480 if (obj
->type
== OBJ_COMMIT
|| obj
->type
== OBJ_TREE
) {
481 struct tree_desc tree
;
485 data
= read_object_with_reference(obj
->sha1
, tree_type
,
488 die("unable to read tree (%s)", sha1_to_hex(obj
->sha1
));
489 init_tree_desc(&tree
, data
, size
);
490 hit
= grep_tree(opt
, paths
, &tree
, name
, "");
494 die("unable to grep from object of type %s", typename(obj
->type
));
497 static const char builtin_grep_usage
[] =
498 "git-grep <option>* <rev>* [-e] <pattern> [<path>...]";
500 static const char emsg_invalid_context_len
[] =
501 "%s: invalid context length argument";
502 static const char emsg_missing_context_len
[] =
503 "missing context length argument";
504 static const char emsg_missing_argument
[] =
505 "option requires an argument -%s";
507 int cmd_grep(int argc
, const char **argv
, const char *prefix
)
511 int seen_dashdash
= 0;
513 struct object_array list
= { 0, 0, NULL
};
514 const char **paths
= NULL
;
517 memset(&opt
, 0, sizeof(opt
));
518 opt
.prefix_length
= (prefix
&& *prefix
) ? strlen(prefix
) : 0;
521 opt
.pattern_tail
= &opt
.pattern_list
;
522 opt
.regflags
= REG_NEWLINE
;
525 * If there is no -- then the paths must exist in the working
526 * tree. If there is no explicit pattern specified with -e or
527 * -f, we take the first unrecognized non option to be the
528 * pattern, but then what follows it must be zero or more
529 * valid refs up to the -- (if exists), and then existing
530 * paths. If there is an explicit pattern, then the first
531 * unrecognized non option is the beginning of the refs list
532 * that continues up to the -- (if exists), and then paths.
536 const char *arg
= argv
[1];
538 if (!strcmp("--cached", arg
)) {
542 if (!strcmp("-a", arg
) ||
543 !strcmp("--text", arg
)) {
544 opt
.binary
= GREP_BINARY_TEXT
;
547 if (!strcmp("-i", arg
) ||
548 !strcmp("--ignore-case", arg
)) {
549 opt
.regflags
|= REG_ICASE
;
552 if (!strcmp("-I", arg
)) {
553 opt
.binary
= GREP_BINARY_NOMATCH
;
556 if (!strcmp("-v", arg
) ||
557 !strcmp("--invert-match", arg
)) {
561 if (!strcmp("-E", arg
) ||
562 !strcmp("--extended-regexp", arg
)) {
563 opt
.regflags
|= REG_EXTENDED
;
566 if (!strcmp("-F", arg
) ||
567 !strcmp("--fixed-strings", arg
)) {
571 if (!strcmp("-G", arg
) ||
572 !strcmp("--basic-regexp", arg
)) {
573 opt
.regflags
&= ~REG_EXTENDED
;
576 if (!strcmp("-n", arg
)) {
580 if (!strcmp("-h", arg
)) {
584 if (!strcmp("-H", arg
)) {
588 if (!strcmp("-l", arg
) ||
589 !strcmp("--name-only", arg
) ||
590 !strcmp("--files-with-matches", arg
)) {
594 if (!strcmp("-L", arg
) ||
595 !strcmp("--files-without-match", arg
)) {
596 opt
.unmatch_name_only
= 1;
599 if (!strcmp("-c", arg
) ||
600 !strcmp("--count", arg
)) {
604 if (!strcmp("-w", arg
) ||
605 !strcmp("--word-regexp", arg
)) {
609 if (!prefixcmp(arg
, "-A") ||
610 !prefixcmp(arg
, "-B") ||
611 !prefixcmp(arg
, "-C") ||
612 (arg
[0] == '-' && '1' <= arg
[1] && arg
[1] <= '9')) {
616 case 'A': case 'B': case 'C':
619 die(emsg_missing_context_len
);
630 if (strtoul_ui(scan
, 10, &num
))
631 die(emsg_invalid_context_len
, scan
);
634 opt
.post_context
= num
;
638 opt
.post_context
= num
;
640 opt
.pre_context
= num
;
645 if (!strcmp("-f", arg
)) {
650 die(emsg_missing_argument
, arg
);
651 patterns
= fopen(argv
[1], "r");
653 die("'%s': %s", argv
[1], strerror(errno
));
654 while (fgets(buf
, sizeof(buf
), patterns
)) {
655 int len
= strlen(buf
);
656 if (len
&& buf
[len
-1] == '\n')
658 /* ignore empty line like grep does */
661 append_grep_pattern(&opt
, xstrdup(buf
),
670 if (!strcmp("--not", arg
)) {
671 append_grep_pattern(&opt
, arg
, "command line", 0,
675 if (!strcmp("--and", arg
)) {
676 append_grep_pattern(&opt
, arg
, "command line", 0,
680 if (!strcmp("--or", arg
))
681 continue; /* no-op */
682 if (!strcmp("(", arg
)) {
683 append_grep_pattern(&opt
, arg
, "command line", 0,
687 if (!strcmp(")", arg
)) {
688 append_grep_pattern(&opt
, arg
, "command line", 0,
692 if (!strcmp("--all-match", arg
)) {
696 if (!strcmp("-e", arg
)) {
698 append_grep_pattern(&opt
, argv
[1],
705 die(emsg_missing_argument
, arg
);
707 if (!strcmp("--full-name", arg
)) {
711 if (!strcmp("--", arg
)) {
712 /* later processing wants to have this at argv[1] */
718 usage(builtin_grep_usage
);
720 /* First unrecognized non-option token */
721 if (!opt
.pattern_list
) {
722 append_grep_pattern(&opt
, arg
, "command line", 0,
727 /* We are looking at the first path or rev;
728 * it is found at argv[1] after leaving the
736 if (!opt
.pattern_list
)
737 die("no pattern given.");
738 if ((opt
.regflags
!= REG_NEWLINE
) && opt
.fixed
)
739 die("cannot mix --fixed-strings and regexp");
740 compile_grep_patterns(&opt
);
742 /* Check revs and then paths */
743 for (i
= 1; i
< argc
; i
++) {
744 const char *arg
= argv
[i
];
745 unsigned char sha1
[20];
747 if (!get_sha1(arg
, sha1
)) {
748 struct object
*object
= parse_object(sha1
);
750 die("bad object %s", arg
);
751 add_object_array(object
, arg
, &list
);
754 if (!strcmp(arg
, "--")) {
761 /* The rest are paths */
762 if (!seen_dashdash
) {
764 for (j
= i
; j
< argc
; j
++)
765 verify_filename(prefix
, argv
[j
]);
769 paths
= get_pathspec(prefix
, argv
+ i
);
770 if (opt
.prefix_length
&& opt
.relative
) {
771 /* Make sure we do not get outside of paths */
772 for (i
= 0; paths
[i
]; i
++)
773 if (strncmp(prefix
, paths
[i
], opt
.prefix_length
))
774 die("git-grep: cannot generate relative filenames containing '..'");
778 paths
= xcalloc(2, sizeof(const char *));
784 return !grep_cache(&opt
, paths
, cached
);
787 die("both --cached and trees are given.");
789 for (i
= 0; i
< list
.nr
; i
++) {
790 struct object
*real_obj
;
791 real_obj
= deref_tag(list
.objects
[i
].item
, NULL
, 0);
792 if (grep_object(&opt
, paths
, real_obj
, list
.objects
[i
].name
))
795 free_grep_patterns(&opt
);