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));
211 if (opt->relative && opt->prefix_length)
212 filename = quote_path_relative(filename, -1, &buf, opt->prefix);
213 i = grep_buffer(opt, filename, data, sz);
214 strbuf_release(&buf);
219 #if !NO_EXTERNAL_GREP
220 static int exec_grep(int argc, const char **argv)
230 execvp("grep", (char **) argv);
233 while (waitpid(pid, &status, 0) < 0) {
238 if (WIFEXITED(status)) {
239 if (!WEXITSTATUS(status))
248 #define push_arg(a) do { \
249 if (nr < MAXARGS) argv[nr++] = (a); \
250 else die("maximum number of args exceeded"); \
254 * If you send a singleton filename to grep, it does not give
255 * the name of the file. GNU grep has "-H" but we would want
256 * that behaviour in a portable way.
258 * So we keep two pathnames in argv buffer unsent to grep in
259 * the main loop if we need to do more than one grep.
261 static int flush_grep(struct grep_opt *opt,
262 int argc, int arg0, const char **argv, int *kept)
265 int count = argc - arg0;
266 const char *kept_0 = NULL;
270 * Because we keep at least 2 paths in the call from
271 * the main loop (i.e. kept != NULL), and MAXARGS is
272 * far greater than 2, this usually is a call to
273 * conclude the grep. However, the user could attempt
274 * to overflow the argv buffer by giving too many
275 * options to leave very small number of real
276 * arguments even for the call in the main loop.
279 die("insanely many options to grep");
282 * If we have two or more paths, we do not have to do
283 * anything special, but we need to push /dev/null to
284 * get "-H" behaviour of GNU grep portably but when we
285 * are not doing "-l" nor "-L" nor "-c".
289 !opt->unmatch_name_only &&
291 argv[argc++] = "/dev/null";
298 * Called because we found many paths and haven't finished
299 * iterating over the cache yet. We keep two paths
300 * for the concluding call. argv[argc-2] and argv[argc-1]
301 * has the last two paths, so save the first one away,
302 * replace it with NULL while sending the list to grep,
303 * and recover them after we are done.
306 kept_0 = argv[argc-2];
311 if (opt->pre_context || opt->post_context) {
313 * grep handles hunk marks between files, but we need to
314 * do that ourselves between multiple calls.
316 if (opt->show_hunk_mark)
317 write_or_die(1, "--\n", 3);
319 opt->show_hunk_mark = 1;
322 status = exec_grep(argc, argv);
326 * Then recover them. Now the last arg is beyond the
327 * terminating NULL which is at argc, and the second
328 * from the last is what we saved away in kept_0
330 argv[arg0++] = kept_0;
331 argv[arg0] = argv[argc+1];
336 static void grep_add_color(struct strbuf *sb, const char *escape_seq)
338 size_t orig_len = sb->len;
340 while (*escape_seq) {
341 if (*escape_seq == 'm')
342 strbuf_addch(sb, ';');
343 else if (*escape_seq != '\033' && *escape_seq != '[')
344 strbuf_addch(sb, *escape_seq);
347 if (sb->len > orig_len && sb->buf[sb->len - 1] == ';')
348 strbuf_setlen(sb, sb->len - 1);
351 static int external_grep(struct grep_opt *opt, const char **paths, int cached)
353 int i, nr, argc, hit, len, status;
354 const char *argv[MAXARGS+1];
355 char randarg[ARGBUF];
356 char *argptr = randarg;
359 if (opt->extended || (opt->relative && opt->prefix_length))
369 if (opt->regflags & REG_EXTENDED)
371 if (opt->ignore_case)
373 if (opt->binary == GREP_BINARY_NOMATCH)
375 if (opt->word_regexp)
379 if (opt->unmatch_name_only)
381 if (opt->null_following_name)
382 /* in GNU grep git's "-z" translates to "-Z" */
386 if (opt->post_context || opt->pre_context) {
387 if (opt->post_context != opt->pre_context) {
388 if (opt->pre_context) {
390 len += snprintf(argptr, sizeof(randarg)-len,
391 "%u", opt->pre_context) + 1;
392 if (sizeof(randarg) <= len)
393 die("maximum length of args exceeded");
397 if (opt->post_context) {
399 len += snprintf(argptr, sizeof(randarg)-len,
400 "%u", opt->post_context) + 1;
401 if (sizeof(randarg) <= len)
402 die("maximum length of args exceeded");
409 len += snprintf(argptr, sizeof(randarg)-len,
410 "%u", opt->post_context) + 1;
411 if (sizeof(randarg) <= len)
412 die("maximum length of args exceeded");
417 for (p = opt->pattern_list; p; p = p->next) {
419 push_arg(p->pattern);
422 struct strbuf sb = STRBUF_INIT;
424 grep_add_color(&sb, opt->color_match);
425 setenv("GREP_COLOR", sb.buf, 1);
428 strbuf_addstr(&sb, "mt=");
429 grep_add_color(&sb, opt->color_match);
430 strbuf_addstr(&sb, ":sl=:cx=:fn=:ln=:bn=:se=");
431 setenv("GREP_COLORS", sb.buf, 1);
435 if (opt->color_external && strlen(opt->color_external) > 0)
436 push_arg(opt->color_external);
441 for (i = 0; i < active_nr; i++) {
442 struct cache_entry *ce = active_cache[i];
445 if (!S_ISREG(ce->ce_mode))
447 if (!pathspec_matches(paths, ce->name, opt->max_depth))
450 if (name[0] == '-') {
451 int len = ce_namelen(ce);
452 name = xmalloc(len + 3);
453 memcpy(name, "./", 2);
454 memcpy(name + 2, ce->name, len + 1);
457 if (MAXARGS <= argc) {
458 status = flush_grep(opt, argc, nr, argv, &kept);
466 } while (i < active_nr &&
467 !strcmp(ce->name, active_cache[i]->name));
468 i--; /* compensate for loop control */
472 status = flush_grep(opt, argc, nr, argv, NULL);
480 static int grep_cache(struct grep_opt *opt, const char **paths, int cached,
481 int external_grep_allowed)
487 #if !NO_EXTERNAL_GREP
489 * Use the external "grep" command for the case where
490 * we grep through the checked-out files. It tends to
491 * be a lot more optimized
493 if (!cached && external_grep_allowed) {
494 hit = external_grep(opt, paths, cached);
501 for (nr = 0; nr < active_nr; nr++) {
502 struct cache_entry *ce = active_cache[nr];
503 if (!S_ISREG(ce->ce_mode))
505 if (!pathspec_matches(paths, ce->name, opt->max_depth))
508 * If CE_VALID is on, we assume worktree file and its cache entry
509 * are identical, even if worktree file has been modified, so use
510 * cache version instead
512 if (cached || (ce->ce_flags & CE_VALID)) {
515 hit |= grep_sha1(opt, ce->sha1, ce->name, 0);
518 hit |= grep_file(opt, ce->name);
522 } while (nr < active_nr &&
523 !strcmp(ce->name, active_cache[nr]->name));
524 nr--; /* compensate for loop control */
527 free_grep_patterns(opt);
531 static int grep_tree(struct grep_opt *opt, const char **paths,
532 struct tree_desc *tree,
533 const char *tree_name, const char *base)
537 struct name_entry entry;
539 int tn_len = strlen(tree_name);
540 struct strbuf pathbuf;
542 strbuf_init(&pathbuf, PATH_MAX + tn_len);
545 strbuf_add(&pathbuf, tree_name, tn_len);
546 strbuf_addch(&pathbuf, ':');
547 tn_len = pathbuf.len;
549 strbuf_addstr(&pathbuf, base);
552 while (tree_entry(tree, &entry)) {
553 int te_len = tree_entry_len(entry.path, entry.sha1);
555 strbuf_add(&pathbuf, entry.path, te_len);
557 if (S_ISDIR(entry.mode))
558 /* Match "abc/" against pathspec to
559 * decide if we want to descend into "abc"
562 strbuf_addch(&pathbuf, '/');
564 down = pathbuf.buf + tn_len;
565 if (!pathspec_matches(paths, down, opt->max_depth))
567 else if (S_ISREG(entry.mode))
568 hit |= grep_sha1(opt, entry.sha1, pathbuf.buf, tn_len);
569 else if (S_ISDIR(entry.mode)) {
570 enum object_type type;
571 struct tree_desc sub;
575 data = read_sha1_file(entry.sha1, &type, &size);
577 die("unable to read tree (%s)",
578 sha1_to_hex(entry.sha1));
579 init_tree_desc(&sub, data, size);
580 hit |= grep_tree(opt, paths, &sub, tree_name, down);
584 strbuf_release(&pathbuf);
588 static int grep_object(struct grep_opt *opt, const char **paths,
589 struct object *obj, const char *name)
591 if (obj->type == OBJ_BLOB)
592 return grep_sha1(opt, obj->sha1, name, 0);
593 if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
594 struct tree_desc tree;
598 data = read_object_with_reference(obj->sha1, tree_type,
601 die("unable to read tree (%s)", sha1_to_hex(obj->sha1));
602 init_tree_desc(&tree, data, size);
603 hit = grep_tree(opt, paths, &tree, name, "");
607 die("unable to grep from object of type %s", typename(obj->type));
610 static int context_callback(const struct option *opt, const char *arg,
613 struct grep_opt *grep_opt = opt->value;
618 grep_opt->pre_context = grep_opt->post_context = 0;
621 value = strtol(arg, (char **)&endp, 10);
623 return error("switch `%c' expects a numerical value",
626 grep_opt->pre_context = grep_opt->post_context = value;
630 static int file_callback(const struct option *opt, const char *arg, int unset)
632 struct grep_opt *grep_opt = opt->value;
635 struct strbuf sb = STRBUF_INIT;
637 patterns = fopen(arg, "r");
639 die_errno("cannot open '%s'", arg);
640 while (strbuf_getline(&sb, patterns, '\n') == 0) {
641 /* ignore empty line like grep does */
644 append_grep_pattern(grep_opt, strbuf_detach(&sb, NULL), arg,
645 ++lno, GREP_PATTERN);
652 static int not_callback(const struct option *opt, const char *arg, int unset)
654 struct grep_opt *grep_opt = opt->value;
655 append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
659 static int and_callback(const struct option *opt, const char *arg, int unset)
661 struct grep_opt *grep_opt = opt->value;
662 append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
666 static int open_callback(const struct option *opt, const char *arg, int unset)
668 struct grep_opt *grep_opt = opt->value;
669 append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
673 static int close_callback(const struct option *opt, const char *arg, int unset)
675 struct grep_opt *grep_opt = opt->value;
676 append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
680 static int pattern_callback(const struct option *opt, const char *arg,
683 struct grep_opt *grep_opt = opt->value;
684 append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
688 static int help_callback(const struct option *opt, const char *arg, int unset)
693 int cmd_grep(int argc, const char **argv, const char *prefix)
697 int external_grep_allowed = 1;
698 int seen_dashdash = 0;
700 struct object_array list = { 0, 0, NULL };
701 const char **paths = NULL;
704 struct option options[] = {
705 OPT_BOOLEAN(0, "cached", &cached,
706 "search in index instead of in the work tree"),
708 OPT_BOOLEAN('v', "invert-match", &opt.invert,
709 "show non-matching lines"),
710 OPT_BOOLEAN('i', "ignore-case", &opt.ignore_case,
711 "case insensitive matching"),
712 OPT_BOOLEAN('w', "word-regexp", &opt.word_regexp,
713 "match patterns only at word boundaries"),
714 OPT_SET_INT('a', "text", &opt.binary,
715 "process binary files as text", GREP_BINARY_TEXT),
716 OPT_SET_INT('I', NULL, &opt.binary,
717 "don't match patterns in binary files",
718 GREP_BINARY_NOMATCH),
719 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, "depth",
720 "descend at most <depth> levels", PARSE_OPT_NONEG,
723 OPT_BIT('E', "extended-regexp", &opt.regflags,
724 "use extended POSIX regular expressions", REG_EXTENDED),
725 OPT_NEGBIT('G', "basic-regexp", &opt.regflags,
726 "use basic POSIX regular expressions (default)",
728 OPT_BOOLEAN('F', "fixed-strings", &opt.fixed,
729 "interpret patterns as fixed strings"),
731 OPT_BOOLEAN('n', NULL, &opt.linenum, "show line numbers"),
732 OPT_NEGBIT('h', NULL, &opt.pathname, "don't show filenames", 1),
733 OPT_BIT('H', NULL, &opt.pathname, "show filenames", 1),
734 OPT_NEGBIT(0, "full-name", &opt.relative,
735 "show filenames relative to top directory", 1),
736 OPT_BOOLEAN('l', "files-with-matches", &opt.name_only,
737 "show only filenames instead of matching lines"),
738 OPT_BOOLEAN(0, "name-only", &opt.name_only,
739 "synonym for --files-with-matches"),
740 OPT_BOOLEAN('L', "files-without-match",
741 &opt.unmatch_name_only,
742 "show only the names of files without match"),
743 OPT_BOOLEAN('z', "null", &opt.null_following_name,
744 "print NUL after filenames"),
745 OPT_BOOLEAN('c', "count", &opt.count,
746 "show the number of matches instead of matching lines"),
747 OPT_SET_INT(0, "color", &opt.color, "highlight matches", 1),
749 OPT_CALLBACK('C', NULL, &opt, "n",
750 "show <n> context lines before and after matches",
752 OPT_INTEGER('B', NULL, &opt.pre_context,
753 "show <n> context lines before matches"),
754 OPT_INTEGER('A', NULL, &opt.post_context,
755 "show <n> context lines after matches"),
756 OPT_NUMBER_CALLBACK(&opt, "shortcut for -C NUM",
758 OPT_BOOLEAN('p', "show-function", &opt.funcname,
759 "show a line with the function name before matches"),
761 OPT_CALLBACK('f', NULL, &opt, "file",
762 "read patterns from file", file_callback),
763 { OPTION_CALLBACK, 'e', NULL, &opt, "pattern",
764 "match <pattern>", PARSE_OPT_NONEG, pattern_callback },
765 { OPTION_CALLBACK, 0, "and", &opt, NULL,
766 "combine patterns specified with -e",
767 PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
768 OPT_BOOLEAN(0, "or", &dummy, ""),
769 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
770 PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
771 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
772 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
774 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
775 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
777 OPT_BOOLEAN(0, "all-match", &opt.all_match,
778 "show only matches from files that match all patterns"),
781 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
782 "allow calling of grep(1) (ignored by this build)"),
784 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
785 "allow calling of grep(1) (default)"),
787 { OPTION_CALLBACK, 0, "help-all", &options, NULL, "show usage",
788 PARSE_OPT_HIDDEN | PARSE_OPT_NOARG, help_callback },
792 memset(&opt, 0, sizeof(opt));
794 opt.prefix_length = (prefix && *prefix) ? strlen(prefix) : 0;
797 opt.pattern_tail = &opt.pattern_list;
798 opt.regflags = REG_NEWLINE;
801 strcpy(opt.color_match, GIT_COLOR_RED GIT_COLOR_BOLD);
803 git_config(grep_config, &opt);
805 opt.color = git_use_color_default;
808 * If there is no -- then the paths must exist in the working
809 * tree. If there is no explicit pattern specified with -e or
810 * -f, we take the first unrecognized non option to be the
811 * pattern, but then what follows it must be zero or more
812 * valid refs up to the -- (if exists), and then existing
813 * paths. If there is an explicit pattern, then the first
814 * unrecognized non option is the beginning of the refs list
815 * that continues up to the -- (if exists), and then paths.
817 argc = parse_options(argc, argv, prefix, options, grep_usage,
818 PARSE_OPT_KEEP_DASHDASH |
819 PARSE_OPT_STOP_AT_NON_OPTION |
820 PARSE_OPT_NO_INTERNAL_HELP);
822 /* First unrecognized non-option token */
823 if (argc > 0 && !opt.pattern_list) {
824 append_grep_pattern(&opt, argv[0], "command line", 0,
830 if ((opt.color && !opt.color_external) || opt.funcname)
831 external_grep_allowed = 0;
832 if (!opt.pattern_list)
833 die("no pattern given.");
834 if (!opt.fixed && opt.ignore_case)
835 opt.regflags |= REG_ICASE;
836 if ((opt.regflags != REG_NEWLINE) && opt.fixed)
837 die("cannot mix --fixed-strings and regexp");
838 compile_grep_patterns(&opt);
840 /* Check revs and then paths */
841 for (i = 0; i < argc; i++) {
842 const char *arg = argv[i];
843 unsigned char sha1[20];
845 if (!get_sha1(arg, sha1)) {
846 struct object *object = parse_object(sha1);
848 die("bad object %s", arg);
849 add_object_array(object, arg, &list);
852 if (!strcmp(arg, "--")) {
859 /* The rest are paths */
860 if (!seen_dashdash) {
862 for (j = i; j < argc; j++)
863 verify_filename(prefix, argv[j]);
867 paths = get_pathspec(prefix, argv + i);
869 paths = xcalloc(2, sizeof(const char *));
877 return !grep_cache(&opt, paths, cached, external_grep_allowed);
881 die("both --cached and trees are given.");
883 for (i = 0; i < list.nr; i++) {
884 struct object *real_obj;
885 real_obj = deref_tag(list.objects[i].item, NULL, 0);
886 if (grep_object(&opt, paths, real_obj, list.objects[i].name))
889 free_grep_patterns(&opt);