The sixth batch
[git.git] / builtin / grep.c
blob7d2f8e5adb69c428847bc23b81bf822c916ccdea
1 /*
2 * Builtin "git grep"
4 * Copyright (c) 2006 Junio C Hamano
5 */
6 #define USE_THE_INDEX_COMPATIBILITY_MACROS
7 #include "cache.h"
8 #include "repository.h"
9 #include "config.h"
10 #include "blob.h"
11 #include "tree.h"
12 #include "commit.h"
13 #include "tag.h"
14 #include "tree-walk.h"
15 #include "builtin.h"
16 #include "parse-options.h"
17 #include "string-list.h"
18 #include "run-command.h"
19 #include "userdiff.h"
20 #include "grep.h"
21 #include "quote.h"
22 #include "dir.h"
23 #include "pathspec.h"
24 #include "submodule.h"
25 #include "submodule-config.h"
26 #include "object-store.h"
27 #include "packfile.h"
29 static char const * const grep_usage[] = {
30 N_("git grep [<options>] [-e] <pattern> [<rev>...] [[--] <path>...]"),
31 NULL
34 static int recurse_submodules;
36 static int num_threads;
38 static pthread_t *threads;
40 /* We use one producer thread and THREADS consumer
41 * threads. The producer adds struct work_items to 'todo' and the
42 * consumers pick work items from the same array.
44 struct work_item {
45 struct grep_source source;
46 char done;
47 struct strbuf out;
50 /* In the range [todo_done, todo_start) in 'todo' we have work_items
51 * that have been or are processed by a consumer thread. We haven't
52 * written the result for these to stdout yet.
54 * The work_items in [todo_start, todo_end) are waiting to be picked
55 * up by a consumer thread.
57 * The ranges are modulo TODO_SIZE.
59 #define TODO_SIZE 128
60 static struct work_item todo[TODO_SIZE];
61 static int todo_start;
62 static int todo_end;
63 static int todo_done;
65 /* Has all work items been added? */
66 static int all_work_added;
68 /* This lock protects all the variables above. */
69 static pthread_mutex_t grep_mutex;
71 static inline void grep_lock(void)
73 pthread_mutex_lock(&grep_mutex);
76 static inline void grep_unlock(void)
78 pthread_mutex_unlock(&grep_mutex);
81 /* Signalled when a new work_item is added to todo. */
82 static pthread_cond_t cond_add;
84 /* Signalled when the result from one work_item is written to
85 * stdout.
87 static pthread_cond_t cond_write;
89 /* Signalled when we are finished with everything. */
90 static pthread_cond_t cond_result;
92 static int skip_first_line;
94 static void add_work(struct grep_opt *opt, struct grep_source *gs)
96 if (opt->binary != GREP_BINARY_TEXT)
97 grep_source_load_driver(gs, opt->repo->index);
99 grep_lock();
101 while ((todo_end+1) % ARRAY_SIZE(todo) == todo_done) {
102 pthread_cond_wait(&cond_write, &grep_mutex);
105 todo[todo_end].source = *gs;
106 todo[todo_end].done = 0;
107 strbuf_reset(&todo[todo_end].out);
108 todo_end = (todo_end + 1) % ARRAY_SIZE(todo);
110 pthread_cond_signal(&cond_add);
111 grep_unlock();
114 static struct work_item *get_work(void)
116 struct work_item *ret;
118 grep_lock();
119 while (todo_start == todo_end && !all_work_added) {
120 pthread_cond_wait(&cond_add, &grep_mutex);
123 if (todo_start == todo_end && all_work_added) {
124 ret = NULL;
125 } else {
126 ret = &todo[todo_start];
127 todo_start = (todo_start + 1) % ARRAY_SIZE(todo);
129 grep_unlock();
130 return ret;
133 static void work_done(struct work_item *w)
135 int old_done;
137 grep_lock();
138 w->done = 1;
139 old_done = todo_done;
140 for(; todo[todo_done].done && todo_done != todo_start;
141 todo_done = (todo_done+1) % ARRAY_SIZE(todo)) {
142 w = &todo[todo_done];
143 if (w->out.len) {
144 const char *p = w->out.buf;
145 size_t len = w->out.len;
147 /* Skip the leading hunk mark of the first file. */
148 if (skip_first_line) {
149 while (len) {
150 len--;
151 if (*p++ == '\n')
152 break;
154 skip_first_line = 0;
157 write_or_die(1, p, len);
159 grep_source_clear(&w->source);
162 if (old_done != todo_done)
163 pthread_cond_signal(&cond_write);
165 if (all_work_added && todo_done == todo_end)
166 pthread_cond_signal(&cond_result);
168 grep_unlock();
171 static void *run(void *arg)
173 int hit = 0;
174 struct grep_opt *opt = arg;
176 while (1) {
177 struct work_item *w = get_work();
178 if (!w)
179 break;
181 opt->output_priv = w;
182 hit |= grep_source(opt, &w->source);
183 grep_source_clear_data(&w->source);
184 work_done(w);
186 free_grep_patterns(arg);
187 free(arg);
189 return (void*) (intptr_t) hit;
192 static void strbuf_out(struct grep_opt *opt, const void *buf, size_t size)
194 struct work_item *w = opt->output_priv;
195 strbuf_add(&w->out, buf, size);
198 static void start_threads(struct grep_opt *opt)
200 int i;
202 pthread_mutex_init(&grep_mutex, NULL);
203 pthread_mutex_init(&grep_attr_mutex, NULL);
204 pthread_cond_init(&cond_add, NULL);
205 pthread_cond_init(&cond_write, NULL);
206 pthread_cond_init(&cond_result, NULL);
207 grep_use_locks = 1;
208 enable_obj_read_lock();
210 for (i = 0; i < ARRAY_SIZE(todo); i++) {
211 strbuf_init(&todo[i].out, 0);
214 CALLOC_ARRAY(threads, num_threads);
215 for (i = 0; i < num_threads; i++) {
216 int err;
217 struct grep_opt *o = grep_opt_dup(opt);
218 o->output = strbuf_out;
219 compile_grep_patterns(o);
220 err = pthread_create(&threads[i], NULL, run, o);
222 if (err)
223 die(_("grep: failed to create thread: %s"),
224 strerror(err));
228 static int wait_all(void)
230 int hit = 0;
231 int i;
233 if (!HAVE_THREADS)
234 BUG("Never call this function unless you have started threads");
236 grep_lock();
237 all_work_added = 1;
239 /* Wait until all work is done. */
240 while (todo_done != todo_end)
241 pthread_cond_wait(&cond_result, &grep_mutex);
243 /* Wake up all the consumer threads so they can see that there
244 * is no more work to do.
246 pthread_cond_broadcast(&cond_add);
247 grep_unlock();
249 for (i = 0; i < num_threads; i++) {
250 void *h;
251 pthread_join(threads[i], &h);
252 hit |= (int) (intptr_t) h;
255 free(threads);
257 pthread_mutex_destroy(&grep_mutex);
258 pthread_mutex_destroy(&grep_attr_mutex);
259 pthread_cond_destroy(&cond_add);
260 pthread_cond_destroy(&cond_write);
261 pthread_cond_destroy(&cond_result);
262 grep_use_locks = 0;
263 disable_obj_read_lock();
265 return hit;
268 static int grep_cmd_config(const char *var, const char *value, void *cb)
270 int st = grep_config(var, value, cb);
271 if (git_color_default_config(var, value, cb) < 0)
272 st = -1;
274 if (!strcmp(var, "grep.threads")) {
275 num_threads = git_config_int(var, value);
276 if (num_threads < 0)
277 die(_("invalid number of threads specified (%d) for %s"),
278 num_threads, var);
279 else if (!HAVE_THREADS && num_threads > 1) {
281 * TRANSLATORS: %s is the configuration
282 * variable for tweaking threads, currently
283 * grep.threads
285 warning(_("no threads support, ignoring %s"), var);
286 num_threads = 1;
290 if (!strcmp(var, "submodule.recurse"))
291 recurse_submodules = git_config_bool(var, value);
293 return st;
296 static void grep_source_name(struct grep_opt *opt, const char *filename,
297 int tree_name_len, struct strbuf *out)
299 strbuf_reset(out);
301 if (opt->null_following_name) {
302 if (opt->relative && opt->prefix_length) {
303 struct strbuf rel_buf = STRBUF_INIT;
304 const char *rel_name =
305 relative_path(filename + tree_name_len,
306 opt->prefix, &rel_buf);
308 if (tree_name_len)
309 strbuf_add(out, filename, tree_name_len);
311 strbuf_addstr(out, rel_name);
312 strbuf_release(&rel_buf);
313 } else {
314 strbuf_addstr(out, filename);
316 return;
319 if (opt->relative && opt->prefix_length)
320 quote_path(filename + tree_name_len, opt->prefix, out, 0);
321 else
322 quote_c_style(filename + tree_name_len, out, NULL, 0);
324 if (tree_name_len)
325 strbuf_insert(out, 0, filename, tree_name_len);
328 static int grep_oid(struct grep_opt *opt, const struct object_id *oid,
329 const char *filename, int tree_name_len,
330 const char *path)
332 struct strbuf pathbuf = STRBUF_INIT;
333 struct grep_source gs;
335 grep_source_name(opt, filename, tree_name_len, &pathbuf);
336 grep_source_init(&gs, GREP_SOURCE_OID, pathbuf.buf, path, oid);
337 strbuf_release(&pathbuf);
339 if (num_threads > 1) {
341 * add_work() copies gs and thus assumes ownership of
342 * its fields, so do not call grep_source_clear()
344 add_work(opt, &gs);
345 return 0;
346 } else {
347 int hit;
349 hit = grep_source(opt, &gs);
351 grep_source_clear(&gs);
352 return hit;
356 static int grep_file(struct grep_opt *opt, const char *filename)
358 struct strbuf buf = STRBUF_INIT;
359 struct grep_source gs;
361 grep_source_name(opt, filename, 0, &buf);
362 grep_source_init(&gs, GREP_SOURCE_FILE, buf.buf, filename, filename);
363 strbuf_release(&buf);
365 if (num_threads > 1) {
367 * add_work() copies gs and thus assumes ownership of
368 * its fields, so do not call grep_source_clear()
370 add_work(opt, &gs);
371 return 0;
372 } else {
373 int hit;
375 hit = grep_source(opt, &gs);
377 grep_source_clear(&gs);
378 return hit;
382 static void append_path(struct grep_opt *opt, const void *data, size_t len)
384 struct string_list *path_list = opt->output_priv;
386 if (len == 1 && *(const char *)data == '\0')
387 return;
388 string_list_append(path_list, xstrndup(data, len));
391 static void run_pager(struct grep_opt *opt, const char *prefix)
393 struct string_list *path_list = opt->output_priv;
394 struct child_process child = CHILD_PROCESS_INIT;
395 int i, status;
397 for (i = 0; i < path_list->nr; i++)
398 strvec_push(&child.args, path_list->items[i].string);
399 child.dir = prefix;
400 child.use_shell = 1;
402 status = run_command(&child);
403 if (status)
404 exit(status);
407 static int grep_cache(struct grep_opt *opt,
408 const struct pathspec *pathspec, int cached);
409 static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
410 struct tree_desc *tree, struct strbuf *base, int tn_len,
411 int check_attr);
413 static int grep_submodule(struct grep_opt *opt,
414 const struct pathspec *pathspec,
415 const struct object_id *oid,
416 const char *filename, const char *path, int cached)
418 struct repository subrepo;
419 struct repository *superproject = opt->repo;
420 const struct submodule *sub;
421 struct grep_opt subopt;
422 int hit;
424 sub = submodule_from_path(superproject, null_oid(), path);
426 if (!is_submodule_active(superproject, path))
427 return 0;
429 if (repo_submodule_init(&subrepo, superproject, sub))
430 return 0;
433 * NEEDSWORK: repo_read_gitmodules() might call
434 * add_to_alternates_memory() via config_from_gitmodules(). This
435 * operation causes a race condition with concurrent object readings
436 * performed by the worker threads. That's why we need obj_read_lock()
437 * here. It should be removed once it's no longer necessary to add the
438 * subrepo's odbs to the in-memory alternates list.
440 obj_read_lock();
441 repo_read_gitmodules(&subrepo, 0);
444 * NEEDSWORK: This adds the submodule's object directory to the list of
445 * alternates for the single in-memory object store. This has some bad
446 * consequences for memory (processed objects will never be freed) and
447 * performance (this increases the number of pack files git has to pay
448 * attention to, to the sum of the number of pack files in all the
449 * repositories processed so far). This can be removed once the object
450 * store is no longer global and instead is a member of the repository
451 * object.
453 add_to_alternates_memory(subrepo.objects->odb->path);
454 obj_read_unlock();
456 memcpy(&subopt, opt, sizeof(subopt));
457 subopt.repo = &subrepo;
459 if (oid) {
460 struct object *object;
461 struct tree_desc tree;
462 void *data;
463 unsigned long size;
464 struct strbuf base = STRBUF_INIT;
466 obj_read_lock();
467 object = parse_object_or_die(oid, NULL);
468 obj_read_unlock();
469 data = read_object_with_reference(&subrepo,
470 &object->oid, tree_type,
471 &size, NULL);
472 if (!data)
473 die(_("unable to read tree (%s)"), oid_to_hex(&object->oid));
475 strbuf_addstr(&base, filename);
476 strbuf_addch(&base, '/');
478 init_tree_desc(&tree, data, size);
479 hit = grep_tree(&subopt, pathspec, &tree, &base, base.len,
480 object->type == OBJ_COMMIT);
481 strbuf_release(&base);
482 free(data);
483 } else {
484 hit = grep_cache(&subopt, pathspec, cached);
487 repo_clear(&subrepo);
488 return hit;
491 static int grep_cache(struct grep_opt *opt,
492 const struct pathspec *pathspec, int cached)
494 struct repository *repo = opt->repo;
495 int hit = 0;
496 int nr;
497 struct strbuf name = STRBUF_INIT;
498 int name_base_len = 0;
499 if (repo->submodule_prefix) {
500 name_base_len = strlen(repo->submodule_prefix);
501 strbuf_addstr(&name, repo->submodule_prefix);
504 if (repo_read_index(repo) < 0)
505 die(_("index file corrupt"));
507 /* TODO: audit for interaction with sparse-index. */
508 ensure_full_index(repo->index);
509 for (nr = 0; nr < repo->index->cache_nr; nr++) {
510 const struct cache_entry *ce = repo->index->cache[nr];
512 if (!cached && ce_skip_worktree(ce))
513 continue;
515 strbuf_setlen(&name, name_base_len);
516 strbuf_addstr(&name, ce->name);
518 if (S_ISREG(ce->ce_mode) &&
519 match_pathspec(repo->index, pathspec, name.buf, name.len, 0, NULL,
520 S_ISDIR(ce->ce_mode) ||
521 S_ISGITLINK(ce->ce_mode))) {
523 * If CE_VALID is on, we assume worktree file and its
524 * cache entry are identical, even if worktree file has
525 * been modified, so use cache version instead
527 if (cached || (ce->ce_flags & CE_VALID)) {
528 if (ce_stage(ce) || ce_intent_to_add(ce))
529 continue;
530 hit |= grep_oid(opt, &ce->oid, name.buf,
531 0, name.buf);
532 } else {
533 hit |= grep_file(opt, name.buf);
535 } else if (recurse_submodules && S_ISGITLINK(ce->ce_mode) &&
536 submodule_path_match(repo->index, pathspec, name.buf, NULL)) {
537 hit |= grep_submodule(opt, pathspec, NULL, ce->name,
538 ce->name, cached);
539 } else {
540 continue;
543 if (ce_stage(ce)) {
544 do {
545 nr++;
546 } while (nr < repo->index->cache_nr &&
547 !strcmp(ce->name, repo->index->cache[nr]->name));
548 nr--; /* compensate for loop control */
550 if (hit && opt->status_only)
551 break;
554 strbuf_release(&name);
555 return hit;
558 static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
559 struct tree_desc *tree, struct strbuf *base, int tn_len,
560 int check_attr)
562 struct repository *repo = opt->repo;
563 int hit = 0;
564 enum interesting match = entry_not_interesting;
565 struct name_entry entry;
566 int old_baselen = base->len;
567 struct strbuf name = STRBUF_INIT;
568 int name_base_len = 0;
569 if (repo->submodule_prefix) {
570 strbuf_addstr(&name, repo->submodule_prefix);
571 name_base_len = name.len;
574 while (tree_entry(tree, &entry)) {
575 int te_len = tree_entry_len(&entry);
577 if (match != all_entries_interesting) {
578 strbuf_addstr(&name, base->buf + tn_len);
579 match = tree_entry_interesting(repo->index,
580 &entry, &name,
581 0, pathspec);
582 strbuf_setlen(&name, name_base_len);
584 if (match == all_entries_not_interesting)
585 break;
586 if (match == entry_not_interesting)
587 continue;
590 strbuf_add(base, entry.path, te_len);
592 if (S_ISREG(entry.mode)) {
593 hit |= grep_oid(opt, &entry.oid, base->buf, tn_len,
594 check_attr ? base->buf + tn_len : NULL);
595 } else if (S_ISDIR(entry.mode)) {
596 enum object_type type;
597 struct tree_desc sub;
598 void *data;
599 unsigned long size;
601 data = read_object_file(&entry.oid, &type, &size);
602 if (!data)
603 die(_("unable to read tree (%s)"),
604 oid_to_hex(&entry.oid));
606 strbuf_addch(base, '/');
607 init_tree_desc(&sub, data, size);
608 hit |= grep_tree(opt, pathspec, &sub, base, tn_len,
609 check_attr);
610 free(data);
611 } else if (recurse_submodules && S_ISGITLINK(entry.mode)) {
612 hit |= grep_submodule(opt, pathspec, &entry.oid,
613 base->buf, base->buf + tn_len,
614 1); /* ignored */
617 strbuf_setlen(base, old_baselen);
619 if (hit && opt->status_only)
620 break;
623 strbuf_release(&name);
624 return hit;
627 static int grep_object(struct grep_opt *opt, const struct pathspec *pathspec,
628 struct object *obj, const char *name, const char *path)
630 if (obj->type == OBJ_BLOB)
631 return grep_oid(opt, &obj->oid, name, 0, path);
632 if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
633 struct tree_desc tree;
634 void *data;
635 unsigned long size;
636 struct strbuf base;
637 int hit, len;
639 data = read_object_with_reference(opt->repo,
640 &obj->oid, tree_type,
641 &size, NULL);
642 if (!data)
643 die(_("unable to read tree (%s)"), oid_to_hex(&obj->oid));
645 len = name ? strlen(name) : 0;
646 strbuf_init(&base, PATH_MAX + len + 1);
647 if (len) {
648 strbuf_add(&base, name, len);
649 strbuf_addch(&base, ':');
651 init_tree_desc(&tree, data, size);
652 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
653 obj->type == OBJ_COMMIT);
654 strbuf_release(&base);
655 free(data);
656 return hit;
658 die(_("unable to grep from object of type %s"), type_name(obj->type));
661 static int grep_objects(struct grep_opt *opt, const struct pathspec *pathspec,
662 const struct object_array *list)
664 unsigned int i;
665 int hit = 0;
666 const unsigned int nr = list->nr;
668 for (i = 0; i < nr; i++) {
669 struct object *real_obj;
671 obj_read_lock();
672 real_obj = deref_tag(opt->repo, list->objects[i].item,
673 NULL, 0);
674 obj_read_unlock();
676 if (!real_obj) {
677 char hex[GIT_MAX_HEXSZ + 1];
678 const char *name = list->objects[i].name;
680 if (!name) {
681 oid_to_hex_r(hex, &list->objects[i].item->oid);
682 name = hex;
684 die(_("invalid object '%s' given."), name);
687 /* load the gitmodules file for this rev */
688 if (recurse_submodules) {
689 submodule_free(opt->repo);
690 obj_read_lock();
691 gitmodules_config_oid(&real_obj->oid);
692 obj_read_unlock();
694 if (grep_object(opt, pathspec, real_obj, list->objects[i].name,
695 list->objects[i].path)) {
696 hit = 1;
697 if (opt->status_only)
698 break;
701 return hit;
704 static int grep_directory(struct grep_opt *opt, const struct pathspec *pathspec,
705 int exc_std, int use_index)
707 struct dir_struct dir = DIR_INIT;
708 int i, hit = 0;
710 if (!use_index)
711 dir.flags |= DIR_NO_GITLINKS;
712 if (exc_std)
713 setup_standard_excludes(&dir);
715 fill_directory(&dir, opt->repo->index, pathspec);
716 for (i = 0; i < dir.nr; i++) {
717 hit |= grep_file(opt, dir.entries[i]->name);
718 if (hit && opt->status_only)
719 break;
721 dir_clear(&dir);
722 return hit;
725 static int context_callback(const struct option *opt, const char *arg,
726 int unset)
728 struct grep_opt *grep_opt = opt->value;
729 int value;
730 const char *endp;
732 if (unset) {
733 grep_opt->pre_context = grep_opt->post_context = 0;
734 return 0;
736 value = strtol(arg, (char **)&endp, 10);
737 if (*endp) {
738 return error(_("switch `%c' expects a numerical value"),
739 opt->short_name);
741 grep_opt->pre_context = grep_opt->post_context = value;
742 return 0;
745 static int file_callback(const struct option *opt, const char *arg, int unset)
747 struct grep_opt *grep_opt = opt->value;
748 int from_stdin;
749 FILE *patterns;
750 int lno = 0;
751 struct strbuf sb = STRBUF_INIT;
753 BUG_ON_OPT_NEG(unset);
755 from_stdin = !strcmp(arg, "-");
756 patterns = from_stdin ? stdin : fopen(arg, "r");
757 if (!patterns)
758 die_errno(_("cannot open '%s'"), arg);
759 while (strbuf_getline(&sb, patterns) == 0) {
760 /* ignore empty line like grep does */
761 if (sb.len == 0)
762 continue;
764 append_grep_pat(grep_opt, sb.buf, sb.len, arg, ++lno,
765 GREP_PATTERN);
767 if (!from_stdin)
768 fclose(patterns);
769 strbuf_release(&sb);
770 return 0;
773 static int not_callback(const struct option *opt, const char *arg, int unset)
775 struct grep_opt *grep_opt = opt->value;
776 BUG_ON_OPT_NEG(unset);
777 BUG_ON_OPT_ARG(arg);
778 append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
779 return 0;
782 static int and_callback(const struct option *opt, const char *arg, int unset)
784 struct grep_opt *grep_opt = opt->value;
785 BUG_ON_OPT_NEG(unset);
786 BUG_ON_OPT_ARG(arg);
787 append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
788 return 0;
791 static int open_callback(const struct option *opt, const char *arg, int unset)
793 struct grep_opt *grep_opt = opt->value;
794 BUG_ON_OPT_NEG(unset);
795 BUG_ON_OPT_ARG(arg);
796 append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
797 return 0;
800 static int close_callback(const struct option *opt, const char *arg, int unset)
802 struct grep_opt *grep_opt = opt->value;
803 BUG_ON_OPT_NEG(unset);
804 BUG_ON_OPT_ARG(arg);
805 append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
806 return 0;
809 static int pattern_callback(const struct option *opt, const char *arg,
810 int unset)
812 struct grep_opt *grep_opt = opt->value;
813 BUG_ON_OPT_NEG(unset);
814 append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
815 return 0;
818 int cmd_grep(int argc, const char **argv, const char *prefix)
820 int hit = 0;
821 int cached = 0, untracked = 0, opt_exclude = -1;
822 int seen_dashdash = 0;
823 int external_grep_allowed__ignored;
824 const char *show_in_pager = NULL, *default_pager = "dummy";
825 struct grep_opt opt;
826 struct object_array list = OBJECT_ARRAY_INIT;
827 struct pathspec pathspec;
828 struct string_list path_list = STRING_LIST_INIT_NODUP;
829 int i;
830 int dummy;
831 int use_index = 1;
832 int pattern_type_arg = GREP_PATTERN_TYPE_UNSPECIFIED;
833 int allow_revs;
835 struct option options[] = {
836 OPT_BOOL(0, "cached", &cached,
837 N_("search in index instead of in the work tree")),
838 OPT_NEGBIT(0, "no-index", &use_index,
839 N_("find in contents not managed by git"), 1),
840 OPT_BOOL(0, "untracked", &untracked,
841 N_("search in both tracked and untracked files")),
842 OPT_SET_INT(0, "exclude-standard", &opt_exclude,
843 N_("ignore files specified via '.gitignore'"), 1),
844 OPT_BOOL(0, "recurse-submodules", &recurse_submodules,
845 N_("recursively search in each submodule")),
846 OPT_GROUP(""),
847 OPT_BOOL('v', "invert-match", &opt.invert,
848 N_("show non-matching lines")),
849 OPT_BOOL('i', "ignore-case", &opt.ignore_case,
850 N_("case insensitive matching")),
851 OPT_BOOL('w', "word-regexp", &opt.word_regexp,
852 N_("match patterns only at word boundaries")),
853 OPT_SET_INT('a', "text", &opt.binary,
854 N_("process binary files as text"), GREP_BINARY_TEXT),
855 OPT_SET_INT('I', NULL, &opt.binary,
856 N_("don't match patterns in binary files"),
857 GREP_BINARY_NOMATCH),
858 OPT_BOOL(0, "textconv", &opt.allow_textconv,
859 N_("process binary files with textconv filters")),
860 OPT_SET_INT('r', "recursive", &opt.max_depth,
861 N_("search in subdirectories (default)"), -1),
862 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, N_("depth"),
863 N_("descend at most <depth> levels"), PARSE_OPT_NONEG,
864 NULL, 1 },
865 OPT_GROUP(""),
866 OPT_SET_INT('E', "extended-regexp", &pattern_type_arg,
867 N_("use extended POSIX regular expressions"),
868 GREP_PATTERN_TYPE_ERE),
869 OPT_SET_INT('G', "basic-regexp", &pattern_type_arg,
870 N_("use basic POSIX regular expressions (default)"),
871 GREP_PATTERN_TYPE_BRE),
872 OPT_SET_INT('F', "fixed-strings", &pattern_type_arg,
873 N_("interpret patterns as fixed strings"),
874 GREP_PATTERN_TYPE_FIXED),
875 OPT_SET_INT('P', "perl-regexp", &pattern_type_arg,
876 N_("use Perl-compatible regular expressions"),
877 GREP_PATTERN_TYPE_PCRE),
878 OPT_GROUP(""),
879 OPT_BOOL('n', "line-number", &opt.linenum, N_("show line numbers")),
880 OPT_BOOL(0, "column", &opt.columnnum, N_("show column number of first match")),
881 OPT_NEGBIT('h', NULL, &opt.pathname, N_("don't show filenames"), 1),
882 OPT_BIT('H', NULL, &opt.pathname, N_("show filenames"), 1),
883 OPT_NEGBIT(0, "full-name", &opt.relative,
884 N_("show filenames relative to top directory"), 1),
885 OPT_BOOL('l', "files-with-matches", &opt.name_only,
886 N_("show only filenames instead of matching lines")),
887 OPT_BOOL(0, "name-only", &opt.name_only,
888 N_("synonym for --files-with-matches")),
889 OPT_BOOL('L', "files-without-match",
890 &opt.unmatch_name_only,
891 N_("show only the names of files without match")),
892 OPT_BOOL_F('z', "null", &opt.null_following_name,
893 N_("print NUL after filenames"),
894 PARSE_OPT_NOCOMPLETE),
895 OPT_BOOL('o', "only-matching", &opt.only_matching,
896 N_("show only matching parts of a line")),
897 OPT_BOOL('c', "count", &opt.count,
898 N_("show the number of matches instead of matching lines")),
899 OPT__COLOR(&opt.color, N_("highlight matches")),
900 OPT_BOOL(0, "break", &opt.file_break,
901 N_("print empty line between matches from different files")),
902 OPT_BOOL(0, "heading", &opt.heading,
903 N_("show filename only once above matches from same file")),
904 OPT_GROUP(""),
905 OPT_CALLBACK('C', "context", &opt, N_("n"),
906 N_("show <n> context lines before and after matches"),
907 context_callback),
908 OPT_INTEGER('B', "before-context", &opt.pre_context,
909 N_("show <n> context lines before matches")),
910 OPT_INTEGER('A', "after-context", &opt.post_context,
911 N_("show <n> context lines after matches")),
912 OPT_INTEGER(0, "threads", &num_threads,
913 N_("use <n> worker threads")),
914 OPT_NUMBER_CALLBACK(&opt, N_("shortcut for -C NUM"),
915 context_callback),
916 OPT_BOOL('p', "show-function", &opt.funcname,
917 N_("show a line with the function name before matches")),
918 OPT_BOOL('W', "function-context", &opt.funcbody,
919 N_("show the surrounding function")),
920 OPT_GROUP(""),
921 OPT_CALLBACK('f', NULL, &opt, N_("file"),
922 N_("read patterns from file"), file_callback),
923 OPT_CALLBACK_F('e', NULL, &opt, N_("pattern"),
924 N_("match <pattern>"), PARSE_OPT_NONEG, pattern_callback),
925 OPT_CALLBACK_F(0, "and", &opt, NULL,
926 N_("combine patterns specified with -e"),
927 PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback),
928 OPT_BOOL(0, "or", &dummy, ""),
929 OPT_CALLBACK_F(0, "not", &opt, NULL, "",
930 PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback),
931 OPT_CALLBACK_F('(', NULL, &opt, NULL, "",
932 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
933 open_callback),
934 OPT_CALLBACK_F(')', NULL, &opt, NULL, "",
935 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
936 close_callback),
937 OPT__QUIET(&opt.status_only,
938 N_("indicate hit with exit status without output")),
939 OPT_BOOL(0, "all-match", &opt.all_match,
940 N_("show only matches from files that match all patterns")),
941 OPT_GROUP(""),
942 { OPTION_STRING, 'O', "open-files-in-pager", &show_in_pager,
943 N_("pager"), N_("show matching files in the pager"),
944 PARSE_OPT_OPTARG | PARSE_OPT_NOCOMPLETE,
945 NULL, (intptr_t)default_pager },
946 OPT_BOOL_F(0, "ext-grep", &external_grep_allowed__ignored,
947 N_("allow calling of grep(1) (ignored by this build)"),
948 PARSE_OPT_NOCOMPLETE),
949 OPT_END()
952 git_config(grep_cmd_config, NULL);
953 grep_init(&opt, the_repository, prefix);
956 * If there is no -- then the paths must exist in the working
957 * tree. If there is no explicit pattern specified with -e or
958 * -f, we take the first unrecognized non option to be the
959 * pattern, but then what follows it must be zero or more
960 * valid refs up to the -- (if exists), and then existing
961 * paths. If there is an explicit pattern, then the first
962 * unrecognized non option is the beginning of the refs list
963 * that continues up to the -- (if exists), and then paths.
965 argc = parse_options(argc, argv, prefix, options, grep_usage,
966 PARSE_OPT_KEEP_DASHDASH |
967 PARSE_OPT_STOP_AT_NON_OPTION);
968 grep_commit_pattern_type(pattern_type_arg, &opt);
970 if (use_index && !startup_info->have_repository) {
971 int fallback = 0;
972 git_config_get_bool("grep.fallbacktonoindex", &fallback);
973 if (fallback)
974 use_index = 0;
975 else
976 /* die the same way as if we did it at the beginning */
977 setup_git_directory();
979 /* Ignore --recurse-submodules if --no-index is given or implied */
980 if (!use_index)
981 recurse_submodules = 0;
984 * skip a -- separator; we know it cannot be
985 * separating revisions from pathnames if
986 * we haven't even had any patterns yet
988 if (argc > 0 && !opt.pattern_list && !strcmp(argv[0], "--")) {
989 argv++;
990 argc--;
993 /* First unrecognized non-option token */
994 if (argc > 0 && !opt.pattern_list) {
995 append_grep_pattern(&opt, argv[0], "command line", 0,
996 GREP_PATTERN);
997 argv++;
998 argc--;
1001 if (show_in_pager == default_pager)
1002 show_in_pager = git_pager(1);
1003 if (show_in_pager) {
1004 opt.color = 0;
1005 opt.name_only = 1;
1006 opt.null_following_name = 1;
1007 opt.output_priv = &path_list;
1008 opt.output = append_path;
1009 string_list_append(&path_list, show_in_pager);
1012 if (!opt.pattern_list)
1013 die(_("no pattern given"));
1015 /* --only-matching has no effect with --invert. */
1016 if (opt.invert)
1017 opt.only_matching = 0;
1020 * We have to find "--" in a separate pass, because its presence
1021 * influences how we will parse arguments that come before it.
1023 for (i = 0; i < argc; i++) {
1024 if (!strcmp(argv[i], "--")) {
1025 seen_dashdash = 1;
1026 break;
1031 * Resolve any rev arguments. If we have a dashdash, then everything up
1032 * to it must resolve as a rev. If not, then we stop at the first
1033 * non-rev and assume everything else is a path.
1035 allow_revs = use_index && !untracked;
1036 for (i = 0; i < argc; i++) {
1037 const char *arg = argv[i];
1038 struct object_id oid;
1039 struct object_context oc;
1040 struct object *object;
1042 if (!strcmp(arg, "--")) {
1043 i++;
1044 break;
1047 if (!allow_revs) {
1048 if (seen_dashdash)
1049 die(_("--no-index or --untracked cannot be used with revs"));
1050 break;
1053 if (get_oid_with_context(the_repository, arg,
1054 GET_OID_RECORD_PATH,
1055 &oid, &oc)) {
1056 if (seen_dashdash)
1057 die(_("unable to resolve revision: %s"), arg);
1058 break;
1061 object = parse_object_or_die(&oid, arg);
1062 if (!seen_dashdash)
1063 verify_non_filename(prefix, arg);
1064 add_object_array_with_path(object, arg, &list, oc.mode, oc.path);
1065 free(oc.path);
1069 * Anything left over is presumed to be a path. But in the non-dashdash
1070 * "do what I mean" case, we verify and complain when that isn't true.
1072 if (!seen_dashdash) {
1073 int j;
1074 for (j = i; j < argc; j++)
1075 verify_filename(prefix, argv[j], j == i && allow_revs);
1078 parse_pathspec(&pathspec, 0,
1079 PATHSPEC_PREFER_CWD |
1080 (opt.max_depth != -1 ? PATHSPEC_MAXDEPTH_VALID : 0),
1081 prefix, argv + i);
1082 pathspec.max_depth = opt.max_depth;
1083 pathspec.recursive = 1;
1084 pathspec.recurse_submodules = !!recurse_submodules;
1086 if (recurse_submodules && untracked)
1087 die(_("--untracked not supported with --recurse-submodules"));
1089 if (show_in_pager) {
1090 if (num_threads > 1)
1091 warning(_("invalid option combination, ignoring --threads"));
1092 num_threads = 1;
1093 } else if (!HAVE_THREADS && num_threads > 1) {
1094 warning(_("no threads support, ignoring --threads"));
1095 num_threads = 1;
1096 } else if (num_threads < 0)
1097 die(_("invalid number of threads specified (%d)"), num_threads);
1098 else if (num_threads == 0)
1099 num_threads = HAVE_THREADS ? online_cpus() : 1;
1101 if (num_threads > 1) {
1102 if (!HAVE_THREADS)
1103 BUG("Somebody got num_threads calculation wrong!");
1104 if (!(opt.name_only || opt.unmatch_name_only || opt.count)
1105 && (opt.pre_context || opt.post_context ||
1106 opt.file_break || opt.funcbody))
1107 skip_first_line = 1;
1110 * Pre-read gitmodules (if not read already) and force eager
1111 * initialization of packed_git to prevent racy lazy
1112 * reading/initialization once worker threads are started.
1114 if (recurse_submodules)
1115 repo_read_gitmodules(the_repository, 1);
1116 if (startup_info->have_repository)
1117 (void)get_packed_git(the_repository);
1119 start_threads(&opt);
1120 } else {
1122 * The compiled patterns on the main path are only
1123 * used when not using threading. Otherwise
1124 * start_threads() above calls compile_grep_patterns()
1125 * for each thread.
1127 compile_grep_patterns(&opt);
1130 if (show_in_pager && (cached || list.nr))
1131 die(_("--open-files-in-pager only works on the worktree"));
1133 if (show_in_pager && opt.pattern_list && !opt.pattern_list->next) {
1134 const char *pager = path_list.items[0].string;
1135 int len = strlen(pager);
1137 if (len > 4 && is_dir_sep(pager[len - 5]))
1138 pager += len - 4;
1140 if (opt.ignore_case && !strcmp("less", pager))
1141 string_list_append(&path_list, "-I");
1143 if (!strcmp("less", pager) || !strcmp("vi", pager)) {
1144 struct strbuf buf = STRBUF_INIT;
1145 strbuf_addf(&buf, "+/%s%s",
1146 strcmp("less", pager) ? "" : "*",
1147 opt.pattern_list->pattern);
1148 string_list_append(&path_list,
1149 strbuf_detach(&buf, NULL));
1153 if (!show_in_pager && !opt.status_only)
1154 setup_pager();
1156 if (!use_index && (untracked || cached))
1157 die(_("--cached or --untracked cannot be used with --no-index"));
1159 if (untracked && cached)
1160 die(_("--untracked cannot be used with --cached"));
1162 if (!use_index || untracked) {
1163 int use_exclude = (opt_exclude < 0) ? use_index : !!opt_exclude;
1164 hit = grep_directory(&opt, &pathspec, use_exclude, use_index);
1165 } else if (0 <= opt_exclude) {
1166 die(_("--[no-]exclude-standard cannot be used for tracked contents"));
1167 } else if (!list.nr) {
1168 if (!cached)
1169 setup_work_tree();
1171 hit = grep_cache(&opt, &pathspec, cached);
1172 } else {
1173 if (cached)
1174 die(_("both --cached and trees are given"));
1176 hit = grep_objects(&opt, &pathspec, &list);
1179 if (num_threads > 1)
1180 hit |= wait_all();
1181 if (hit && show_in_pager)
1182 run_pager(&opt, prefix);
1183 clear_pathspec(&pathspec);
1184 free_grep_patterns(&opt);
1185 return !hit;