4 * Copyright (c) 2006 Junio C Hamano
11 #include "tree-walk.h"
13 #include "parse-options.h"
18 #ifndef NO_EXTERNAL_GREP
20 #define NO_EXTERNAL_GREP 0
22 #define NO_EXTERNAL_GREP 1
26 static char const * const grep_usage[] = {
27 "git grep [options] [-e] <pattern> [<rev>...] [[--] path...]",
31 static int grep_config(const char *var, const char *value, void *cb)
33 struct grep_opt *opt = cb;
35 switch (userdiff_config(var, value)) {
41 if (!strcmp(var, "color.grep")) {
42 opt->color = git_config_colorbool(var, value, -1);
45 if (!strcmp(var, "color.grep.external"))
46 return git_config_string(&(opt->color_external), var, value);
47 if (!strcmp(var, "color.grep.match")) {
49 return config_error_nonbool(var);
50 color_parse(value, var, opt->color_match);
53 return git_color_default_config(var, value, cb);
57 * Return non-zero if max_depth is negative or path has no more then max_depth
60 static int accept_subdir(const char *path, int max_depth)
65 while ((path = strchr(path, '/')) != NULL) {
75 * Return non-zero if name is a subdirectory of match and is not too deep.
77 static int is_subdir(const char *name, int namelen,
78 const char *match, int matchlen, int max_depth)
80 if (matchlen > namelen || strncmp(name, match, matchlen))
83 if (name[matchlen] == '\0') /* exact match */
86 if (!matchlen || match[matchlen-1] == '/' || name[matchlen] == '/')
87 return accept_subdir(name + matchlen + 1, max_depth);
93 * git grep pathspecs are somewhat different from diff-tree pathspecs;
94 * pathname wildcards are allowed.
96 static int pathspec_matches(const char **paths, const char *name, int max_depth)
99 if (!paths || !*paths)
100 return accept_subdir(name, max_depth);
101 namelen = strlen(name);
102 for (i = 0; paths[i]; i++) {
103 const char *match = paths[i];
104 int matchlen = strlen(match);
105 const char *cp, *meta;
107 if (is_subdir(name, namelen, match, matchlen, max_depth))
109 if (!fnmatch(match, name, 0))
111 if (name[namelen-1] != '/')
114 /* We are being asked if the directory ("name") is worth
117 * Find the longest leading directory name that does
118 * not have metacharacter in the pathspec; the name
119 * we are looking at must overlap with that directory.
121 for (cp = match, meta = NULL; cp - match < matchlen; cp++) {
123 if (ch == '*' || ch == '[' || ch == '?') {
129 meta = cp; /* fully literal */
131 if (namelen <= meta - match) {
132 /* Looking at "Documentation/" and
133 * the pattern says "Documentation/howto/", or
134 * "Documentation/diff*.txt". The name we
135 * have should match prefix.
137 if (!memcmp(match, name, namelen))
142 if (meta - match < namelen) {
143 /* Looking at "Documentation/howto/" and
144 * the pattern says "Documentation/h*";
145 * match up to "Do.../h"; this avoids descending
146 * into "Documentation/technical/".
148 if (!memcmp(match, name, meta - match))
156 static int grep_sha1(struct grep_opt *opt, const unsigned char *sha1, const char *name, int tree_name_len)
160 enum object_type type;
162 struct strbuf pathbuf = STRBUF_INIT;
164 data = read_sha1_file(sha1, &type, &size);
166 error("'%s': unable to read %s", name, sha1_to_hex(sha1));
169 if (opt->relative && opt->prefix_length) {
170 quote_path_relative(name + tree_name_len, -1, &pathbuf, opt->prefix);
171 strbuf_insert(&pathbuf, 0, name, tree_name_len);
174 hit = grep_buffer(opt, name, data, size);
175 strbuf_release(&pathbuf);
180 static int grep_file(struct grep_opt *opt, const char *filename)
186 struct strbuf buf = STRBUF_INIT;
188 if (lstat(filename, &st) < 0) {
191 error("'%s': %s", filename, strerror(errno));
195 return 0; /* empty file -- no grep hit */
196 if (!S_ISREG(st.st_mode))
198 sz = xsize_t(st.st_size);
199 i = open(filename, O_RDONLY);
202 data = xmalloc(sz + 1);
203 if (st.st_size != read_in_full(i, data, sz)) {
204 error("'%s': short read %s", filename, strerror(errno));
210 if (opt->relative && opt->prefix_length)
211 filename = quote_path_relative(filename, -1, &buf, opt->prefix);
212 i = grep_buffer(opt, filename, data, sz);
213 strbuf_release(&buf);
218 #if !NO_EXTERNAL_GREP
219 static int exec_grep(int argc, const char **argv)
229 execvp("grep", (char **) argv);
232 while (waitpid(pid, &status, 0) < 0) {
237 if (WIFEXITED(status)) {
238 if (!WEXITSTATUS(status))
247 #define push_arg(a) do { \
248 if (nr < MAXARGS) argv[nr++] = (a); \
249 else die("maximum number of args exceeded"); \
253 * If you send a singleton filename to grep, it does not give
254 * the name of the file. GNU grep has "-H" but we would want
255 * that behaviour in a portable way.
257 * So we keep two pathnames in argv buffer unsent to grep in
258 * the main loop if we need to do more than one grep.
260 static int flush_grep(struct grep_opt *opt,
261 int argc, int arg0, const char **argv, int *kept)
264 int count = argc - arg0;
265 const char *kept_0 = NULL;
269 * Because we keep at least 2 paths in the call from
270 * the main loop (i.e. kept != NULL), and MAXARGS is
271 * far greater than 2, this usually is a call to
272 * conclude the grep. However, the user could attempt
273 * to overflow the argv buffer by giving too many
274 * options to leave very small number of real
275 * arguments even for the call in the main loop.
278 die("insanely many options to grep");
281 * If we have two or more paths, we do not have to do
282 * anything special, but we need to push /dev/null to
283 * get "-H" behaviour of GNU grep portably but when we
284 * are not doing "-l" nor "-L" nor "-c".
288 !opt->unmatch_name_only &&
290 argv[argc++] = "/dev/null";
297 * Called because we found many paths and haven't finished
298 * iterating over the cache yet. We keep two paths
299 * for the concluding call. argv[argc-2] and argv[argc-1]
300 * has the last two paths, so save the first one away,
301 * replace it with NULL while sending the list to grep,
302 * and recover them after we are done.
305 kept_0 = argv[argc-2];
310 if (opt->pre_context || opt->post_context) {
312 * grep handles hunk marks between files, but we need to
313 * do that ourselves between multiple calls.
315 if (opt->show_hunk_mark)
316 write_or_die(1, "--\n", 3);
318 opt->show_hunk_mark = 1;
321 status = exec_grep(argc, argv);
325 * Then recover them. Now the last arg is beyond the
326 * terminating NULL which is at argc, and the second
327 * from the last is what we saved away in kept_0
329 argv[arg0++] = kept_0;
330 argv[arg0] = argv[argc+1];
335 static void grep_add_color(struct strbuf *sb, const char *escape_seq)
337 size_t orig_len = sb->len;
339 while (*escape_seq) {
340 if (*escape_seq == 'm')
341 strbuf_addch(sb, ';');
342 else if (*escape_seq != '\033' && *escape_seq != '[')
343 strbuf_addch(sb, *escape_seq);
346 if (sb->len > orig_len && sb->buf[sb->len - 1] == ';')
347 strbuf_setlen(sb, sb->len - 1);
350 static int external_grep(struct grep_opt *opt, const char **paths, int cached)
352 int i, nr, argc, hit, len, status;
353 const char *argv[MAXARGS+1];
354 char randarg[ARGBUF];
355 char *argptr = randarg;
358 if (opt->extended || (opt->relative && opt->prefix_length))
368 if (opt->regflags & REG_EXTENDED)
370 if (opt->regflags & REG_ICASE)
372 if (opt->binary == GREP_BINARY_NOMATCH)
374 if (opt->word_regexp)
378 if (opt->unmatch_name_only)
380 if (opt->null_following_name)
381 /* in GNU grep git's "-z" translates to "-Z" */
385 if (opt->post_context || opt->pre_context) {
386 if (opt->post_context != opt->pre_context) {
387 if (opt->pre_context) {
389 len += snprintf(argptr, sizeof(randarg)-len,
390 "%u", opt->pre_context) + 1;
391 if (sizeof(randarg) <= len)
392 die("maximum length of args exceeded");
396 if (opt->post_context) {
398 len += snprintf(argptr, sizeof(randarg)-len,
399 "%u", opt->post_context) + 1;
400 if (sizeof(randarg) <= len)
401 die("maximum length of args exceeded");
408 len += snprintf(argptr, sizeof(randarg)-len,
409 "%u", opt->post_context) + 1;
410 if (sizeof(randarg) <= len)
411 die("maximum length of args exceeded");
416 for (p = opt->pattern_list; p; p = p->next) {
418 push_arg(p->pattern);
421 struct strbuf sb = STRBUF_INIT;
423 grep_add_color(&sb, opt->color_match);
424 setenv("GREP_COLOR", sb.buf, 1);
427 strbuf_addstr(&sb, "mt=");
428 grep_add_color(&sb, opt->color_match);
429 strbuf_addstr(&sb, ":sl=:cx=:fn=:ln=:bn=:se=");
430 setenv("GREP_COLORS", sb.buf, 1);
434 if (opt->color_external && strlen(opt->color_external) > 0)
435 push_arg(opt->color_external);
440 for (i = 0; i < active_nr; i++) {
441 struct cache_entry *ce = active_cache[i];
444 if (!S_ISREG(ce->ce_mode))
446 if (!pathspec_matches(paths, ce->name, opt->max_depth))
449 if (name[0] == '-') {
450 int len = ce_namelen(ce);
451 name = xmalloc(len + 3);
452 memcpy(name, "./", 2);
453 memcpy(name + 2, ce->name, len + 1);
456 if (MAXARGS <= argc) {
457 status = flush_grep(opt, argc, nr, argv, &kept);
465 } while (i < active_nr &&
466 !strcmp(ce->name, active_cache[i]->name));
467 i--; /* compensate for loop control */
471 status = flush_grep(opt, argc, nr, argv, NULL);
479 static int grep_cache(struct grep_opt *opt, const char **paths, int cached,
480 int external_grep_allowed)
486 #if !NO_EXTERNAL_GREP
488 * Use the external "grep" command for the case where
489 * we grep through the checked-out files. It tends to
490 * be a lot more optimized
492 if (!cached && external_grep_allowed) {
493 hit = external_grep(opt, paths, cached);
500 for (nr = 0; nr < active_nr; nr++) {
501 struct cache_entry *ce = active_cache[nr];
502 if (!S_ISREG(ce->ce_mode))
504 if (!pathspec_matches(paths, ce->name, opt->max_depth))
507 * If CE_VALID is on, we assume worktree file and its cache entry
508 * are identical, even if worktree file has been modified, so use
509 * cache version instead
511 if (cached || (ce->ce_flags & CE_VALID)) {
514 hit |= grep_sha1(opt, ce->sha1, ce->name, 0);
517 hit |= grep_file(opt, ce->name);
521 } while (nr < active_nr &&
522 !strcmp(ce->name, active_cache[nr]->name));
523 nr--; /* compensate for loop control */
526 free_grep_patterns(opt);
530 static int grep_tree(struct grep_opt *opt, const char **paths,
531 struct tree_desc *tree,
532 const char *tree_name, const char *base)
536 struct name_entry entry;
538 int tn_len = strlen(tree_name);
539 struct strbuf pathbuf;
541 strbuf_init(&pathbuf, PATH_MAX + tn_len);
544 strbuf_add(&pathbuf, tree_name, tn_len);
545 strbuf_addch(&pathbuf, ':');
546 tn_len = pathbuf.len;
548 strbuf_addstr(&pathbuf, base);
551 while (tree_entry(tree, &entry)) {
552 int te_len = tree_entry_len(entry.path, entry.sha1);
554 strbuf_add(&pathbuf, entry.path, te_len);
556 if (S_ISDIR(entry.mode))
557 /* Match "abc/" against pathspec to
558 * decide if we want to descend into "abc"
561 strbuf_addch(&pathbuf, '/');
563 down = pathbuf.buf + tn_len;
564 if (!pathspec_matches(paths, down, opt->max_depth))
566 else if (S_ISREG(entry.mode))
567 hit |= grep_sha1(opt, entry.sha1, pathbuf.buf, tn_len);
568 else if (S_ISDIR(entry.mode)) {
569 enum object_type type;
570 struct tree_desc sub;
574 data = read_sha1_file(entry.sha1, &type, &size);
576 die("unable to read tree (%s)",
577 sha1_to_hex(entry.sha1));
578 init_tree_desc(&sub, data, size);
579 hit |= grep_tree(opt, paths, &sub, tree_name, down);
583 strbuf_release(&pathbuf);
587 static int grep_object(struct grep_opt *opt, const char **paths,
588 struct object *obj, const char *name)
590 if (obj->type == OBJ_BLOB)
591 return grep_sha1(opt, obj->sha1, name, 0);
592 if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
593 struct tree_desc tree;
597 data = read_object_with_reference(obj->sha1, tree_type,
600 die("unable to read tree (%s)", sha1_to_hex(obj->sha1));
601 init_tree_desc(&tree, data, size);
602 hit = grep_tree(opt, paths, &tree, name, "");
606 die("unable to grep from object of type %s", typename(obj->type));
609 static int context_callback(const struct option *opt, const char *arg,
612 struct grep_opt *grep_opt = opt->value;
617 grep_opt->pre_context = grep_opt->post_context = 0;
620 value = strtol(arg, (char **)&endp, 10);
622 return error("switch `%c' expects a numerical value",
625 grep_opt->pre_context = grep_opt->post_context = value;
629 static int file_callback(const struct option *opt, const char *arg, int unset)
631 struct grep_opt *grep_opt = opt->value;
634 struct strbuf sb = STRBUF_INIT;
636 patterns = fopen(arg, "r");
638 die_errno("cannot open '%s'", arg);
639 while (strbuf_getline(&sb, patterns, '\n') == 0) {
640 /* ignore empty line like grep does */
643 append_grep_pattern(grep_opt, strbuf_detach(&sb, NULL), arg,
644 ++lno, GREP_PATTERN);
651 static int not_callback(const struct option *opt, const char *arg, int unset)
653 struct grep_opt *grep_opt = opt->value;
654 append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
658 static int and_callback(const struct option *opt, const char *arg, int unset)
660 struct grep_opt *grep_opt = opt->value;
661 append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
665 static int open_callback(const struct option *opt, const char *arg, int unset)
667 struct grep_opt *grep_opt = opt->value;
668 append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
672 static int close_callback(const struct option *opt, const char *arg, int unset)
674 struct grep_opt *grep_opt = opt->value;
675 append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
679 static int pattern_callback(const struct option *opt, const char *arg,
682 struct grep_opt *grep_opt = opt->value;
683 append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
687 static int help_callback(const struct option *opt, const char *arg, int unset)
692 int cmd_grep(int argc, const char **argv, const char *prefix)
696 int external_grep_allowed = 1;
697 int seen_dashdash = 0;
699 struct object_array list = { 0, 0, NULL };
700 const char **paths = NULL;
703 struct option options[] = {
704 OPT_BOOLEAN(0, "cached", &cached,
705 "search in index instead of in the work tree"),
707 OPT_BOOLEAN('v', "invert-match", &opt.invert,
708 "show non-matching lines"),
709 OPT_BIT('i', "ignore-case", &opt.regflags,
710 "case insensitive matching", REG_ICASE),
711 OPT_BOOLEAN('w', "word-regexp", &opt.word_regexp,
712 "match patterns only at word boundaries"),
713 OPT_SET_INT('a', "text", &opt.binary,
714 "process binary files as text", GREP_BINARY_TEXT),
715 OPT_SET_INT('I', NULL, &opt.binary,
716 "don't match patterns in binary files",
717 GREP_BINARY_NOMATCH),
718 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, "depth",
719 "descend at most <depth> levels", PARSE_OPT_NONEG,
722 OPT_BIT('E', "extended-regexp", &opt.regflags,
723 "use extended POSIX regular expressions", REG_EXTENDED),
724 OPT_NEGBIT('G', "basic-regexp", &opt.regflags,
725 "use basic POSIX regular expressions (default)",
727 OPT_BOOLEAN('F', "fixed-strings", &opt.fixed,
728 "interpret patterns as fixed strings"),
730 OPT_BOOLEAN('n', NULL, &opt.linenum, "show line numbers"),
731 OPT_NEGBIT('h', NULL, &opt.pathname, "don't show filenames", 1),
732 OPT_BIT('H', NULL, &opt.pathname, "show filenames", 1),
733 OPT_NEGBIT(0, "full-name", &opt.relative,
734 "show filenames relative to top directory", 1),
735 OPT_BOOLEAN('l', "files-with-matches", &opt.name_only,
736 "show only filenames instead of matching lines"),
737 OPT_BOOLEAN(0, "name-only", &opt.name_only,
738 "synonym for --files-with-matches"),
739 OPT_BOOLEAN('L', "files-without-match",
740 &opt.unmatch_name_only,
741 "show only the names of files without match"),
742 OPT_BOOLEAN('z', "null", &opt.null_following_name,
743 "print NUL after filenames"),
744 OPT_BOOLEAN('c', "count", &opt.count,
745 "show the number of matches instead of matching lines"),
746 OPT_SET_INT(0, "color", &opt.color, "highlight matches", 1),
748 OPT_CALLBACK('C', NULL, &opt, "n",
749 "show <n> context lines before and after matches",
751 OPT_INTEGER('B', NULL, &opt.pre_context,
752 "show <n> context lines before matches"),
753 OPT_INTEGER('A', NULL, &opt.post_context,
754 "show <n> context lines after matches"),
755 OPT_NUMBER_CALLBACK(&opt, "shortcut for -C NUM",
757 OPT_BOOLEAN('p', "show-function", &opt.funcname,
758 "show a line with the function name before matches"),
760 OPT_CALLBACK('f', NULL, &opt, "file",
761 "read patterns from file", file_callback),
762 { OPTION_CALLBACK, 'e', NULL, &opt, "pattern",
763 "match <pattern>", PARSE_OPT_NONEG, pattern_callback },
764 { OPTION_CALLBACK, 0, "and", &opt, NULL,
765 "combine patterns specified with -e",
766 PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
767 OPT_BOOLEAN(0, "or", &dummy, ""),
768 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
769 PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
770 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
771 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
773 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
774 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
776 OPT_BOOLEAN(0, "all-match", &opt.all_match,
777 "show only matches from files that match all patterns"),
780 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
781 "allow calling of grep(1) (ignored by this build)"),
783 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
784 "allow calling of grep(1) (default)"),
786 { OPTION_CALLBACK, 0, "help-all", &options, NULL, "show usage",
787 PARSE_OPT_HIDDEN | PARSE_OPT_NOARG, help_callback },
791 memset(&opt, 0, sizeof(opt));
793 opt.prefix_length = (prefix && *prefix) ? strlen(prefix) : 0;
796 opt.pattern_tail = &opt.pattern_list;
797 opt.regflags = REG_NEWLINE;
800 strcpy(opt.color_match, GIT_COLOR_RED GIT_COLOR_BOLD);
802 git_config(grep_config, &opt);
804 opt.color = git_use_color_default;
807 * If there is no -- then the paths must exist in the working
808 * tree. If there is no explicit pattern specified with -e or
809 * -f, we take the first unrecognized non option to be the
810 * pattern, but then what follows it must be zero or more
811 * valid refs up to the -- (if exists), and then existing
812 * paths. If there is an explicit pattern, then the first
813 * unrecognized non option is the beginning of the refs list
814 * that continues up to the -- (if exists), and then paths.
816 argc = parse_options(argc, argv, prefix, options, grep_usage,
817 PARSE_OPT_KEEP_DASHDASH |
818 PARSE_OPT_STOP_AT_NON_OPTION |
819 PARSE_OPT_NO_INTERNAL_HELP);
821 /* First unrecognized non-option token */
822 if (argc > 0 && !opt.pattern_list) {
823 append_grep_pattern(&opt, argv[0], "command line", 0,
829 if ((opt.color && !opt.color_external) || opt.funcname)
830 external_grep_allowed = 0;
831 if (!opt.pattern_list)
832 die("no pattern given.");
833 if ((opt.regflags != REG_NEWLINE) && opt.fixed)
834 die("cannot mix --fixed-strings and regexp");
835 compile_grep_patterns(&opt);
837 /* Check revs and then paths */
838 for (i = 0; i < argc; i++) {
839 const char *arg = argv[i];
840 unsigned char sha1[20];
842 if (!get_sha1(arg, sha1)) {
843 struct object *object = parse_object(sha1);
845 die("bad object %s", arg);
846 add_object_array(object, arg, &list);
849 if (!strcmp(arg, "--")) {
856 /* The rest are paths */
857 if (!seen_dashdash) {
859 for (j = i; j < argc; j++)
860 verify_filename(prefix, argv[j]);
864 paths = get_pathspec(prefix, argv + i);
866 paths = xcalloc(2, sizeof(const char *));
874 return !grep_cache(&opt, paths, cached, external_grep_allowed);
878 die("both --cached and trees are given.");
880 for (i = 0; i < list.nr; i++) {
881 struct object *real_obj;
882 real_obj = deref_tag(list.objects[i].item, NULL, 0);
883 if (grep_object(&opt, paths, real_obj, list.objects[i].name))
886 free_grep_patterns(&opt);