ls-refs: filter refs using namespace-stripped name
[git/git.git] / builtin / grep.c
1 /*
2 * Builtin "git grep"
3 *
4 * Copyright (c) 2006 Junio C Hamano
5 */
6 #include "cache.h"
7 #include "repository.h"
8 #include "config.h"
9 #include "blob.h"
10 #include "tree.h"
11 #include "commit.h"
12 #include "tag.h"
13 #include "tree-walk.h"
14 #include "builtin.h"
15 #include "parse-options.h"
16 #include "string-list.h"
17 #include "run-command.h"
18 #include "userdiff.h"
19 #include "grep.h"
20 #include "quote.h"
21 #include "dir.h"
22 #include "pathspec.h"
23 #include "submodule.h"
24 #include "submodule-config.h"
25 #include "object-store.h"
26
27 static char const * const grep_usage[] = {
28 N_("git grep [<options>] [-e] <pattern> [<rev>...] [[--] <path>...]"),
29 NULL
30 };
31
32 static int recurse_submodules;
33
34 #define GREP_NUM_THREADS_DEFAULT 8
35 static int num_threads;
36
37 static pthread_t *threads;
38
39 /* We use one producer thread and THREADS consumer
40 * threads. The producer adds struct work_items to 'todo' and the
41 * consumers pick work items from the same array.
42 */
43 struct work_item {
44 struct grep_source source;
45 char done;
46 struct strbuf out;
47 };
48
49 /* In the range [todo_done, todo_start) in 'todo' we have work_items
50 * that have been or are processed by a consumer thread. We haven't
51 * written the result for these to stdout yet.
52 *
53 * The work_items in [todo_start, todo_end) are waiting to be picked
54 * up by a consumer thread.
55 *
56 * The ranges are modulo TODO_SIZE.
57 */
58 #define TODO_SIZE 128
59 static struct work_item todo[TODO_SIZE];
60 static int todo_start;
61 static int todo_end;
62 static int todo_done;
63
64 /* Has all work items been added? */
65 static int all_work_added;
66
67 /* This lock protects all the variables above. */
68 static pthread_mutex_t grep_mutex;
69
70 static inline void grep_lock(void)
71 {
72 pthread_mutex_lock(&grep_mutex);
73 }
74
75 static inline void grep_unlock(void)
76 {
77 pthread_mutex_unlock(&grep_mutex);
78 }
79
80 /* Signalled when a new work_item is added to todo. */
81 static pthread_cond_t cond_add;
82
83 /* Signalled when the result from one work_item is written to
84 * stdout.
85 */
86 static pthread_cond_t cond_write;
87
88 /* Signalled when we are finished with everything. */
89 static pthread_cond_t cond_result;
90
91 static int skip_first_line;
92
93 static void add_work(struct grep_opt *opt, const struct grep_source *gs)
94 {
95 grep_lock();
96
97 while ((todo_end+1) % ARRAY_SIZE(todo) == todo_done) {
98 pthread_cond_wait(&cond_write, &grep_mutex);
99 }
100
101 todo[todo_end].source = *gs;
102 if (opt->binary != GREP_BINARY_TEXT)
103 grep_source_load_driver(&todo[todo_end].source,
104 opt->repo->index);
105 todo[todo_end].done = 0;
106 strbuf_reset(&todo[todo_end].out);
107 todo_end = (todo_end + 1) % ARRAY_SIZE(todo);
108
109 pthread_cond_signal(&cond_add);
110 grep_unlock();
111 }
112
113 static struct work_item *get_work(void)
114 {
115 struct work_item *ret;
116
117 grep_lock();
118 while (todo_start == todo_end && !all_work_added) {
119 pthread_cond_wait(&cond_add, &grep_mutex);
120 }
121
122 if (todo_start == todo_end && all_work_added) {
123 ret = NULL;
124 } else {
125 ret = &todo[todo_start];
126 todo_start = (todo_start + 1) % ARRAY_SIZE(todo);
127 }
128 grep_unlock();
129 return ret;
130 }
131
132 static void work_done(struct work_item *w)
133 {
134 int old_done;
135
136 grep_lock();
137 w->done = 1;
138 old_done = todo_done;
139 for(; todo[todo_done].done && todo_done != todo_start;
140 todo_done = (todo_done+1) % ARRAY_SIZE(todo)) {
141 w = &todo[todo_done];
142 if (w->out.len) {
143 const char *p = w->out.buf;
144 size_t len = w->out.len;
145
146 /* Skip the leading hunk mark of the first file. */
147 if (skip_first_line) {
148 while (len) {
149 len--;
150 if (*p++ == '\n')
151 break;
152 }
153 skip_first_line = 0;
154 }
155
156 write_or_die(1, p, len);
157 }
158 grep_source_clear(&w->source);
159 }
160
161 if (old_done != todo_done)
162 pthread_cond_signal(&cond_write);
163
164 if (all_work_added && todo_done == todo_end)
165 pthread_cond_signal(&cond_result);
166
167 grep_unlock();
168 }
169
170 static void *run(void *arg)
171 {
172 int hit = 0;
173 struct grep_opt *opt = arg;
174
175 while (1) {
176 struct work_item *w = get_work();
177 if (!w)
178 break;
179
180 opt->output_priv = w;
181 hit |= grep_source(opt, &w->source);
182 grep_source_clear_data(&w->source);
183 work_done(w);
184 }
185 free_grep_patterns(arg);
186 free(arg);
187
188 return (void*) (intptr_t) hit;
189 }
190
191 static void strbuf_out(struct grep_opt *opt, const void *buf, size_t size)
192 {
193 struct work_item *w = opt->output_priv;
194 strbuf_add(&w->out, buf, size);
195 }
196
197 static void start_threads(struct grep_opt *opt)
198 {
199 int i;
200
201 pthread_mutex_init(&grep_mutex, NULL);
202 pthread_mutex_init(&grep_read_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
209 for (i = 0; i < ARRAY_SIZE(todo); i++) {
210 strbuf_init(&todo[i].out, 0);
211 }
212
213 threads = xcalloc(num_threads, sizeof(*threads));
214 for (i = 0; i < num_threads; i++) {
215 int err;
216 struct grep_opt *o = grep_opt_dup(opt);
217 o->output = strbuf_out;
218 if (i)
219 o->debug = 0;
220 compile_grep_patterns(o);
221 err = pthread_create(&threads[i], NULL, run, o);
222
223 if (err)
224 die(_("grep: failed to create thread: %s"),
225 strerror(err));
226 }
227 }
228
229 static int wait_all(void)
230 {
231 int hit = 0;
232 int i;
233
234 if (!HAVE_THREADS)
235 BUG("Never call this function unless you have started threads");
236
237 grep_lock();
238 all_work_added = 1;
239
240 /* Wait until all work is done. */
241 while (todo_done != todo_end)
242 pthread_cond_wait(&cond_result, &grep_mutex);
243
244 /* Wake up all the consumer threads so they can see that there
245 * is no more work to do.
246 */
247 pthread_cond_broadcast(&cond_add);
248 grep_unlock();
249
250 for (i = 0; i < num_threads; i++) {
251 void *h;
252 pthread_join(threads[i], &h);
253 hit |= (int) (intptr_t) h;
254 }
255
256 free(threads);
257
258 pthread_mutex_destroy(&grep_mutex);
259 pthread_mutex_destroy(&grep_read_mutex);
260 pthread_mutex_destroy(&grep_attr_mutex);
261 pthread_cond_destroy(&cond_add);
262 pthread_cond_destroy(&cond_write);
263 pthread_cond_destroy(&cond_result);
264 grep_use_locks = 0;
265
266 return hit;
267 }
268
269 static int grep_cmd_config(const char *var, const char *value, void *cb)
270 {
271 int st = grep_config(var, value, cb);
272 if (git_color_default_config(var, value, cb) < 0)
273 st = -1;
274
275 if (!strcmp(var, "grep.threads")) {
276 num_threads = git_config_int(var, value);
277 if (num_threads < 0)
278 die(_("invalid number of threads specified (%d) for %s"),
279 num_threads, var);
280 else if (!HAVE_THREADS && num_threads > 1) {
281 /*
282 * TRANSLATORS: %s is the configuration
283 * variable for tweaking threads, currently
284 * grep.threads
285 */
286 warning(_("no threads support, ignoring %s"), var);
287 num_threads = 1;
288 }
289 }
290
291 if (!strcmp(var, "submodule.recurse"))
292 recurse_submodules = git_config_bool(var, value);
293
294 return st;
295 }
296
297 static void *lock_and_read_oid_file(const struct object_id *oid, enum object_type *type, unsigned long *size)
298 {
299 void *data;
300
301 grep_read_lock();
302 data = read_object_file(oid, type, size);
303 grep_read_unlock();
304 return data;
305 }
306
307 static int grep_oid(struct grep_opt *opt, const struct object_id *oid,
308 const char *filename, int tree_name_len,
309 const char *path)
310 {
311 struct strbuf pathbuf = STRBUF_INIT;
312 struct grep_source gs;
313
314 if (opt->relative && opt->prefix_length) {
315 quote_path_relative(filename + tree_name_len, opt->prefix, &pathbuf);
316 strbuf_insert(&pathbuf, 0, filename, tree_name_len);
317 } else {
318 strbuf_addstr(&pathbuf, filename);
319 }
320
321 grep_source_init(&gs, GREP_SOURCE_OID, pathbuf.buf, path, oid);
322 strbuf_release(&pathbuf);
323
324 if (num_threads > 1) {
325 /*
326 * add_work() copies gs and thus assumes ownership of
327 * its fields, so do not call grep_source_clear()
328 */
329 add_work(opt, &gs);
330 return 0;
331 } else {
332 int hit;
333
334 hit = grep_source(opt, &gs);
335
336 grep_source_clear(&gs);
337 return hit;
338 }
339 }
340
341 static int grep_file(struct grep_opt *opt, const char *filename)
342 {
343 struct strbuf buf = STRBUF_INIT;
344 struct grep_source gs;
345
346 if (opt->relative && opt->prefix_length)
347 quote_path_relative(filename, opt->prefix, &buf);
348 else
349 strbuf_addstr(&buf, filename);
350
351 grep_source_init(&gs, GREP_SOURCE_FILE, buf.buf, filename, filename);
352 strbuf_release(&buf);
353
354 if (num_threads > 1) {
355 /*
356 * add_work() copies gs and thus assumes ownership of
357 * its fields, so do not call grep_source_clear()
358 */
359 add_work(opt, &gs);
360 return 0;
361 } else {
362 int hit;
363
364 hit = grep_source(opt, &gs);
365
366 grep_source_clear(&gs);
367 return hit;
368 }
369 }
370
371 static void append_path(struct grep_opt *opt, const void *data, size_t len)
372 {
373 struct string_list *path_list = opt->output_priv;
374
375 if (len == 1 && *(const char *)data == '\0')
376 return;
377 string_list_append(path_list, xstrndup(data, len));
378 }
379
380 static void run_pager(struct grep_opt *opt, const char *prefix)
381 {
382 struct string_list *path_list = opt->output_priv;
383 struct child_process child = CHILD_PROCESS_INIT;
384 int i, status;
385
386 for (i = 0; i < path_list->nr; i++)
387 argv_array_push(&child.args, path_list->items[i].string);
388 child.dir = prefix;
389 child.use_shell = 1;
390
391 status = run_command(&child);
392 if (status)
393 exit(status);
394 }
395
396 static int grep_cache(struct grep_opt *opt, struct repository *repo,
397 const struct pathspec *pathspec, int cached);
398 static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
399 struct tree_desc *tree, struct strbuf *base, int tn_len,
400 int check_attr, struct repository *repo);
401
402 static int grep_submodule(struct grep_opt *opt, struct repository *superproject,
403 const struct pathspec *pathspec,
404 const struct object_id *oid,
405 const char *filename, const char *path)
406 {
407 struct repository submodule;
408 int hit;
409
410 /*
411 * NEEDSWORK: submodules functions need to be protected because they
412 * access the object store via config_from_gitmodules(): the latter
413 * uses get_oid() which, for now, relies on the global the_repository
414 * object.
415 */
416 grep_read_lock();
417
418 if (!is_submodule_active(superproject, path)) {
419 grep_read_unlock();
420 return 0;
421 }
422
423 if (repo_submodule_init(&submodule, superproject, path)) {
424 grep_read_unlock();
425 return 0;
426 }
427
428 repo_read_gitmodules(&submodule);
429
430 /*
431 * NEEDSWORK: This adds the submodule's object directory to the list of
432 * alternates for the single in-memory object store. This has some bad
433 * consequences for memory (processed objects will never be freed) and
434 * performance (this increases the number of pack files git has to pay
435 * attention to, to the sum of the number of pack files in all the
436 * repositories processed so far). This can be removed once the object
437 * store is no longer global and instead is a member of the repository
438 * object.
439 */
440 add_to_alternates_memory(submodule.objects->odb->path);
441 grep_read_unlock();
442
443 if (oid) {
444 struct object *object;
445 struct tree_desc tree;
446 void *data;
447 unsigned long size;
448 struct strbuf base = STRBUF_INIT;
449
450 object = parse_object_or_die(oid, oid_to_hex(oid));
451
452 grep_read_lock();
453 data = read_object_with_reference(&object->oid, tree_type,
454 &size, NULL);
455 grep_read_unlock();
456
457 if (!data)
458 die(_("unable to read tree (%s)"), oid_to_hex(&object->oid));
459
460 strbuf_addstr(&base, filename);
461 strbuf_addch(&base, '/');
462
463 init_tree_desc(&tree, data, size);
464 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
465 object->type == OBJ_COMMIT, &submodule);
466 strbuf_release(&base);
467 free(data);
468 } else {
469 hit = grep_cache(opt, &submodule, pathspec, 1);
470 }
471
472 repo_clear(&submodule);
473 return hit;
474 }
475
476 static int grep_cache(struct grep_opt *opt, struct repository *repo,
477 const struct pathspec *pathspec, int cached)
478 {
479 int hit = 0;
480 int nr;
481 struct strbuf name = STRBUF_INIT;
482 int name_base_len = 0;
483 if (repo->submodule_prefix) {
484 name_base_len = strlen(repo->submodule_prefix);
485 strbuf_addstr(&name, repo->submodule_prefix);
486 }
487
488 if (repo_read_index(repo) < 0)
489 die(_("index file corrupt"));
490
491 for (nr = 0; nr < repo->index->cache_nr; nr++) {
492 const struct cache_entry *ce = repo->index->cache[nr];
493 strbuf_setlen(&name, name_base_len);
494 strbuf_addstr(&name, ce->name);
495
496 if (S_ISREG(ce->ce_mode) &&
497 match_pathspec(repo->index, pathspec, name.buf, name.len, 0, NULL,
498 S_ISDIR(ce->ce_mode) ||
499 S_ISGITLINK(ce->ce_mode))) {
500 /*
501 * If CE_VALID is on, we assume worktree file and its
502 * cache entry are identical, even if worktree file has
503 * been modified, so use cache version instead
504 */
505 if (cached || (ce->ce_flags & CE_VALID) ||
506 ce_skip_worktree(ce)) {
507 if (ce_stage(ce) || ce_intent_to_add(ce))
508 continue;
509 hit |= grep_oid(opt, &ce->oid, name.buf,
510 0, name.buf);
511 } else {
512 hit |= grep_file(opt, name.buf);
513 }
514 } else if (recurse_submodules && S_ISGITLINK(ce->ce_mode) &&
515 submodule_path_match(repo->index, pathspec, name.buf, NULL)) {
516 hit |= grep_submodule(opt, repo, pathspec, NULL, ce->name, ce->name);
517 } else {
518 continue;
519 }
520
521 if (ce_stage(ce)) {
522 do {
523 nr++;
524 } while (nr < repo->index->cache_nr &&
525 !strcmp(ce->name, repo->index->cache[nr]->name));
526 nr--; /* compensate for loop control */
527 }
528 if (hit && opt->status_only)
529 break;
530 }
531
532 strbuf_release(&name);
533 return hit;
534 }
535
536 static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
537 struct tree_desc *tree, struct strbuf *base, int tn_len,
538 int check_attr, struct repository *repo)
539 {
540 int hit = 0;
541 enum interesting match = entry_not_interesting;
542 struct name_entry entry;
543 int old_baselen = base->len;
544 struct strbuf name = STRBUF_INIT;
545 int name_base_len = 0;
546 if (repo->submodule_prefix) {
547 strbuf_addstr(&name, repo->submodule_prefix);
548 name_base_len = name.len;
549 }
550
551 while (tree_entry(tree, &entry)) {
552 int te_len = tree_entry_len(&entry);
553
554 if (match != all_entries_interesting) {
555 strbuf_addstr(&name, base->buf + tn_len);
556 match = tree_entry_interesting(repo->index,
557 &entry, &name,
558 0, pathspec);
559 strbuf_setlen(&name, name_base_len);
560
561 if (match == all_entries_not_interesting)
562 break;
563 if (match == entry_not_interesting)
564 continue;
565 }
566
567 strbuf_add(base, entry.path, te_len);
568
569 if (S_ISREG(entry.mode)) {
570 hit |= grep_oid(opt, entry.oid, base->buf, tn_len,
571 check_attr ? base->buf + tn_len : NULL);
572 } else if (S_ISDIR(entry.mode)) {
573 enum object_type type;
574 struct tree_desc sub;
575 void *data;
576 unsigned long size;
577
578 data = lock_and_read_oid_file(entry.oid, &type, &size);
579 if (!data)
580 die(_("unable to read tree (%s)"),
581 oid_to_hex(entry.oid));
582
583 strbuf_addch(base, '/');
584 init_tree_desc(&sub, data, size);
585 hit |= grep_tree(opt, pathspec, &sub, base, tn_len,
586 check_attr, repo);
587 free(data);
588 } else if (recurse_submodules && S_ISGITLINK(entry.mode)) {
589 hit |= grep_submodule(opt, repo, pathspec, entry.oid,
590 base->buf, base->buf + tn_len);
591 }
592
593 strbuf_setlen(base, old_baselen);
594
595 if (hit && opt->status_only)
596 break;
597 }
598
599 strbuf_release(&name);
600 return hit;
601 }
602
603 static int grep_object(struct grep_opt *opt, const struct pathspec *pathspec,
604 struct object *obj, const char *name, const char *path)
605 {
606 if (obj->type == OBJ_BLOB)
607 return grep_oid(opt, &obj->oid, name, 0, path);
608 if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
609 struct tree_desc tree;
610 void *data;
611 unsigned long size;
612 struct strbuf base;
613 int hit, len;
614
615 grep_read_lock();
616 data = read_object_with_reference(&obj->oid, tree_type,
617 &size, NULL);
618 grep_read_unlock();
619
620 if (!data)
621 die(_("unable to read tree (%s)"), oid_to_hex(&obj->oid));
622
623 len = name ? strlen(name) : 0;
624 strbuf_init(&base, PATH_MAX + len + 1);
625 if (len) {
626 strbuf_add(&base, name, len);
627 strbuf_addch(&base, ':');
628 }
629 init_tree_desc(&tree, data, size);
630 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
631 obj->type == OBJ_COMMIT, the_repository);
632 strbuf_release(&base);
633 free(data);
634 return hit;
635 }
636 die(_("unable to grep from object of type %s"), type_name(obj->type));
637 }
638
639 static int grep_objects(struct grep_opt *opt, const struct pathspec *pathspec,
640 const struct object_array *list)
641 {
642 unsigned int i;
643 int hit = 0;
644 const unsigned int nr = list->nr;
645
646 for (i = 0; i < nr; i++) {
647 struct object *real_obj;
648 real_obj = deref_tag(the_repository, list->objects[i].item,
649 NULL, 0);
650
651 /* load the gitmodules file for this rev */
652 if (recurse_submodules) {
653 submodule_free(the_repository);
654 gitmodules_config_oid(&real_obj->oid);
655 }
656 if (grep_object(opt, pathspec, real_obj, list->objects[i].name,
657 list->objects[i].path)) {
658 hit = 1;
659 if (opt->status_only)
660 break;
661 }
662 }
663 return hit;
664 }
665
666 static int grep_directory(struct grep_opt *opt, const struct pathspec *pathspec,
667 int exc_std, int use_index)
668 {
669 struct dir_struct dir;
670 int i, hit = 0;
671
672 memset(&dir, 0, sizeof(dir));
673 if (!use_index)
674 dir.flags |= DIR_NO_GITLINKS;
675 if (exc_std)
676 setup_standard_excludes(&dir);
677
678 fill_directory(&dir, &the_index, pathspec);
679 for (i = 0; i < dir.nr; i++) {
680 if (!dir_path_match(&the_index, dir.entries[i], pathspec, 0, NULL))
681 continue;
682 hit |= grep_file(opt, dir.entries[i]->name);
683 if (hit && opt->status_only)
684 break;
685 }
686 return hit;
687 }
688
689 static int context_callback(const struct option *opt, const char *arg,
690 int unset)
691 {
692 struct grep_opt *grep_opt = opt->value;
693 int value;
694 const char *endp;
695
696 if (unset) {
697 grep_opt->pre_context = grep_opt->post_context = 0;
698 return 0;
699 }
700 value = strtol(arg, (char **)&endp, 10);
701 if (*endp) {
702 return error(_("switch `%c' expects a numerical value"),
703 opt->short_name);
704 }
705 grep_opt->pre_context = grep_opt->post_context = value;
706 return 0;
707 }
708
709 static int file_callback(const struct option *opt, const char *arg, int unset)
710 {
711 struct grep_opt *grep_opt = opt->value;
712 int from_stdin;
713 FILE *patterns;
714 int lno = 0;
715 struct strbuf sb = STRBUF_INIT;
716
717 BUG_ON_OPT_NEG(unset);
718
719 from_stdin = !strcmp(arg, "-");
720 patterns = from_stdin ? stdin : fopen(arg, "r");
721 if (!patterns)
722 die_errno(_("cannot open '%s'"), arg);
723 while (strbuf_getline(&sb, patterns) == 0) {
724 /* ignore empty line like grep does */
725 if (sb.len == 0)
726 continue;
727
728 append_grep_pat(grep_opt, sb.buf, sb.len, arg, ++lno,
729 GREP_PATTERN);
730 }
731 if (!from_stdin)
732 fclose(patterns);
733 strbuf_release(&sb);
734 return 0;
735 }
736
737 static int not_callback(const struct option *opt, const char *arg, int unset)
738 {
739 struct grep_opt *grep_opt = opt->value;
740 BUG_ON_OPT_NEG(unset);
741 BUG_ON_OPT_ARG(arg);
742 append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
743 return 0;
744 }
745
746 static int and_callback(const struct option *opt, const char *arg, int unset)
747 {
748 struct grep_opt *grep_opt = opt->value;
749 BUG_ON_OPT_NEG(unset);
750 BUG_ON_OPT_ARG(arg);
751 append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
752 return 0;
753 }
754
755 static int open_callback(const struct option *opt, const char *arg, int unset)
756 {
757 struct grep_opt *grep_opt = opt->value;
758 BUG_ON_OPT_NEG(unset);
759 BUG_ON_OPT_ARG(arg);
760 append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
761 return 0;
762 }
763
764 static int close_callback(const struct option *opt, const char *arg, int unset)
765 {
766 struct grep_opt *grep_opt = opt->value;
767 BUG_ON_OPT_NEG(unset);
768 BUG_ON_OPT_ARG(arg);
769 append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
770 return 0;
771 }
772
773 static int pattern_callback(const struct option *opt, const char *arg,
774 int unset)
775 {
776 struct grep_opt *grep_opt = opt->value;
777 BUG_ON_OPT_NEG(unset);
778 append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
779 return 0;
780 }
781
782 int cmd_grep(int argc, const char **argv, const char *prefix)
783 {
784 int hit = 0;
785 int cached = 0, untracked = 0, opt_exclude = -1;
786 int seen_dashdash = 0;
787 int external_grep_allowed__ignored;
788 const char *show_in_pager = NULL, *default_pager = "dummy";
789 struct grep_opt opt;
790 struct object_array list = OBJECT_ARRAY_INIT;
791 struct pathspec pathspec;
792 struct string_list path_list = STRING_LIST_INIT_NODUP;
793 int i;
794 int dummy;
795 int use_index = 1;
796 int pattern_type_arg = GREP_PATTERN_TYPE_UNSPECIFIED;
797 int allow_revs;
798
799 struct option options[] = {
800 OPT_BOOL(0, "cached", &cached,
801 N_("search in index instead of in the work tree")),
802 OPT_NEGBIT(0, "no-index", &use_index,
803 N_("find in contents not managed by git"), 1),
804 OPT_BOOL(0, "untracked", &untracked,
805 N_("search in both tracked and untracked files")),
806 OPT_SET_INT(0, "exclude-standard", &opt_exclude,
807 N_("ignore files specified via '.gitignore'"), 1),
808 OPT_BOOL(0, "recurse-submodules", &recurse_submodules,
809 N_("recursively search in each submodule")),
810 OPT_GROUP(""),
811 OPT_BOOL('v', "invert-match", &opt.invert,
812 N_("show non-matching lines")),
813 OPT_BOOL('i', "ignore-case", &opt.ignore_case,
814 N_("case insensitive matching")),
815 OPT_BOOL('w', "word-regexp", &opt.word_regexp,
816 N_("match patterns only at word boundaries")),
817 OPT_SET_INT('a', "text", &opt.binary,
818 N_("process binary files as text"), GREP_BINARY_TEXT),
819 OPT_SET_INT('I', NULL, &opt.binary,
820 N_("don't match patterns in binary files"),
821 GREP_BINARY_NOMATCH),
822 OPT_BOOL(0, "textconv", &opt.allow_textconv,
823 N_("process binary files with textconv filters")),
824 OPT_SET_INT('r', "recursive", &opt.max_depth,
825 N_("search in subdirectories (default)"), -1),
826 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, N_("depth"),
827 N_("descend at most <depth> levels"), PARSE_OPT_NONEG,
828 NULL, 1 },
829 OPT_GROUP(""),
830 OPT_SET_INT('E', "extended-regexp", &pattern_type_arg,
831 N_("use extended POSIX regular expressions"),
832 GREP_PATTERN_TYPE_ERE),
833 OPT_SET_INT('G', "basic-regexp", &pattern_type_arg,
834 N_("use basic POSIX regular expressions (default)"),
835 GREP_PATTERN_TYPE_BRE),
836 OPT_SET_INT('F', "fixed-strings", &pattern_type_arg,
837 N_("interpret patterns as fixed strings"),
838 GREP_PATTERN_TYPE_FIXED),
839 OPT_SET_INT('P', "perl-regexp", &pattern_type_arg,
840 N_("use Perl-compatible regular expressions"),
841 GREP_PATTERN_TYPE_PCRE),
842 OPT_GROUP(""),
843 OPT_BOOL('n', "line-number", &opt.linenum, N_("show line numbers")),
844 OPT_BOOL(0, "column", &opt.columnnum, N_("show column number of first match")),
845 OPT_NEGBIT('h', NULL, &opt.pathname, N_("don't show filenames"), 1),
846 OPT_BIT('H', NULL, &opt.pathname, N_("show filenames"), 1),
847 OPT_NEGBIT(0, "full-name", &opt.relative,
848 N_("show filenames relative to top directory"), 1),
849 OPT_BOOL('l', "files-with-matches", &opt.name_only,
850 N_("show only filenames instead of matching lines")),
851 OPT_BOOL(0, "name-only", &opt.name_only,
852 N_("synonym for --files-with-matches")),
853 OPT_BOOL('L', "files-without-match",
854 &opt.unmatch_name_only,
855 N_("show only the names of files without match")),
856 OPT_BOOL_F('z', "null", &opt.null_following_name,
857 N_("print NUL after filenames"),
858 PARSE_OPT_NOCOMPLETE),
859 OPT_BOOL('o', "only-matching", &opt.only_matching,
860 N_("show only matching parts of a line")),
861 OPT_BOOL('c', "count", &opt.count,
862 N_("show the number of matches instead of matching lines")),
863 OPT__COLOR(&opt.color, N_("highlight matches")),
864 OPT_BOOL(0, "break", &opt.file_break,
865 N_("print empty line between matches from different files")),
866 OPT_BOOL(0, "heading", &opt.heading,
867 N_("show filename only once above matches from same file")),
868 OPT_GROUP(""),
869 OPT_CALLBACK('C', "context", &opt, N_("n"),
870 N_("show <n> context lines before and after matches"),
871 context_callback),
872 OPT_INTEGER('B', "before-context", &opt.pre_context,
873 N_("show <n> context lines before matches")),
874 OPT_INTEGER('A', "after-context", &opt.post_context,
875 N_("show <n> context lines after matches")),
876 OPT_INTEGER(0, "threads", &num_threads,
877 N_("use <n> worker threads")),
878 OPT_NUMBER_CALLBACK(&opt, N_("shortcut for -C NUM"),
879 context_callback),
880 OPT_BOOL('p', "show-function", &opt.funcname,
881 N_("show a line with the function name before matches")),
882 OPT_BOOL('W', "function-context", &opt.funcbody,
883 N_("show the surrounding function")),
884 OPT_GROUP(""),
885 OPT_CALLBACK('f', NULL, &opt, N_("file"),
886 N_("read patterns from file"), file_callback),
887 { OPTION_CALLBACK, 'e', NULL, &opt, N_("pattern"),
888 N_("match <pattern>"), PARSE_OPT_NONEG, pattern_callback },
889 { OPTION_CALLBACK, 0, "and", &opt, NULL,
890 N_("combine patterns specified with -e"),
891 PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
892 OPT_BOOL(0, "or", &dummy, ""),
893 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
894 PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
895 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
896 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
897 open_callback },
898 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
899 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
900 close_callback },
901 OPT__QUIET(&opt.status_only,
902 N_("indicate hit with exit status without output")),
903 OPT_BOOL(0, "all-match", &opt.all_match,
904 N_("show only matches from files that match all patterns")),
905 OPT_SET_INT_F(0, "debug", &opt.debug,
906 N_("show parse tree for grep expression"),
907 1, PARSE_OPT_HIDDEN),
908 OPT_GROUP(""),
909 { OPTION_STRING, 'O', "open-files-in-pager", &show_in_pager,
910 N_("pager"), N_("show matching files in the pager"),
911 PARSE_OPT_OPTARG | PARSE_OPT_NOCOMPLETE,
912 NULL, (intptr_t)default_pager },
913 OPT_BOOL_F(0, "ext-grep", &external_grep_allowed__ignored,
914 N_("allow calling of grep(1) (ignored by this build)"),
915 PARSE_OPT_NOCOMPLETE),
916 OPT_END()
917 };
918
919 init_grep_defaults(the_repository);
920 git_config(grep_cmd_config, NULL);
921 grep_init(&opt, the_repository, prefix);
922
923 /*
924 * If there is no -- then the paths must exist in the working
925 * tree. If there is no explicit pattern specified with -e or
926 * -f, we take the first unrecognized non option to be the
927 * pattern, but then what follows it must be zero or more
928 * valid refs up to the -- (if exists), and then existing
929 * paths. If there is an explicit pattern, then the first
930 * unrecognized non option is the beginning of the refs list
931 * that continues up to the -- (if exists), and then paths.
932 */
933 argc = parse_options(argc, argv, prefix, options, grep_usage,
934 PARSE_OPT_KEEP_DASHDASH |
935 PARSE_OPT_STOP_AT_NON_OPTION);
936 grep_commit_pattern_type(pattern_type_arg, &opt);
937
938 if (use_index && !startup_info->have_repository) {
939 int fallback = 0;
940 git_config_get_bool("grep.fallbacktonoindex", &fallback);
941 if (fallback)
942 use_index = 0;
943 else
944 /* die the same way as if we did it at the beginning */
945 setup_git_directory();
946 }
947
948 /*
949 * skip a -- separator; we know it cannot be
950 * separating revisions from pathnames if
951 * we haven't even had any patterns yet
952 */
953 if (argc > 0 && !opt.pattern_list && !strcmp(argv[0], "--")) {
954 argv++;
955 argc--;
956 }
957
958 /* First unrecognized non-option token */
959 if (argc > 0 && !opt.pattern_list) {
960 append_grep_pattern(&opt, argv[0], "command line", 0,
961 GREP_PATTERN);
962 argv++;
963 argc--;
964 }
965
966 if (show_in_pager == default_pager)
967 show_in_pager = git_pager(1);
968 if (show_in_pager) {
969 opt.color = 0;
970 opt.name_only = 1;
971 opt.null_following_name = 1;
972 opt.output_priv = &path_list;
973 opt.output = append_path;
974 string_list_append(&path_list, show_in_pager);
975 }
976
977 if (!opt.pattern_list)
978 die(_("no pattern given"));
979
980 /* --only-matching has no effect with --invert. */
981 if (opt.invert)
982 opt.only_matching = 0;
983
984 /*
985 * We have to find "--" in a separate pass, because its presence
986 * influences how we will parse arguments that come before it.
987 */
988 for (i = 0; i < argc; i++) {
989 if (!strcmp(argv[i], "--")) {
990 seen_dashdash = 1;
991 break;
992 }
993 }
994
995 /*
996 * Resolve any rev arguments. If we have a dashdash, then everything up
997 * to it must resolve as a rev. If not, then we stop at the first
998 * non-rev and assume everything else is a path.
999 */
1000 allow_revs = use_index && !untracked;
1001 for (i = 0; i < argc; i++) {
1002 const char *arg = argv[i];
1003 struct object_id oid;
1004 struct object_context oc;
1005 struct object *object;
1006
1007 if (!strcmp(arg, "--")) {
1008 i++;
1009 break;
1010 }
1011
1012 if (!allow_revs) {
1013 if (seen_dashdash)
1014 die(_("--no-index or --untracked cannot be used with revs"));
1015 break;
1016 }
1017
1018 if (get_oid_with_context(arg, GET_OID_RECORD_PATH,
1019 &oid, &oc)) {
1020 if (seen_dashdash)
1021 die(_("unable to resolve revision: %s"), arg);
1022 break;
1023 }
1024
1025 object = parse_object_or_die(&oid, arg);
1026 if (!seen_dashdash)
1027 verify_non_filename(prefix, arg);
1028 add_object_array_with_path(object, arg, &list, oc.mode, oc.path);
1029 free(oc.path);
1030 }
1031
1032 /*
1033 * Anything left over is presumed to be a path. But in the non-dashdash
1034 * "do what I mean" case, we verify and complain when that isn't true.
1035 */
1036 if (!seen_dashdash) {
1037 int j;
1038 for (j = i; j < argc; j++)
1039 verify_filename(prefix, argv[j], j == i && allow_revs);
1040 }
1041
1042 parse_pathspec(&pathspec, 0,
1043 PATHSPEC_PREFER_CWD |
1044 (opt.max_depth != -1 ? PATHSPEC_MAXDEPTH_VALID : 0),
1045 prefix, argv + i);
1046 pathspec.max_depth = opt.max_depth;
1047 pathspec.recursive = 1;
1048 pathspec.recurse_submodules = !!recurse_submodules;
1049
1050 if (list.nr || cached || show_in_pager) {
1051 if (num_threads > 1)
1052 warning(_("invalid option combination, ignoring --threads"));
1053 num_threads = 1;
1054 } else if (!HAVE_THREADS && num_threads > 1) {
1055 warning(_("no threads support, ignoring --threads"));
1056 num_threads = 1;
1057 } else if (num_threads < 0)
1058 die(_("invalid number of threads specified (%d)"), num_threads);
1059 else if (num_threads == 0)
1060 num_threads = HAVE_THREADS ? GREP_NUM_THREADS_DEFAULT : 1;
1061
1062 if (num_threads > 1) {
1063 if (!HAVE_THREADS)
1064 BUG("Somebody got num_threads calculation wrong!");
1065 if (!(opt.name_only || opt.unmatch_name_only || opt.count)
1066 && (opt.pre_context || opt.post_context ||
1067 opt.file_break || opt.funcbody))
1068 skip_first_line = 1;
1069 start_threads(&opt);
1070 } else {
1071 /*
1072 * The compiled patterns on the main path are only
1073 * used when not using threading. Otherwise
1074 * start_threads() above calls compile_grep_patterns()
1075 * for each thread.
1076 */
1077 compile_grep_patterns(&opt);
1078 }
1079
1080 if (show_in_pager && (cached || list.nr))
1081 die(_("--open-files-in-pager only works on the worktree"));
1082
1083 if (show_in_pager && opt.pattern_list && !opt.pattern_list->next) {
1084 const char *pager = path_list.items[0].string;
1085 int len = strlen(pager);
1086
1087 if (len > 4 && is_dir_sep(pager[len - 5]))
1088 pager += len - 4;
1089
1090 if (opt.ignore_case && !strcmp("less", pager))
1091 string_list_append(&path_list, "-I");
1092
1093 if (!strcmp("less", pager) || !strcmp("vi", pager)) {
1094 struct strbuf buf = STRBUF_INIT;
1095 strbuf_addf(&buf, "+/%s%s",
1096 strcmp("less", pager) ? "" : "*",
1097 opt.pattern_list->pattern);
1098 string_list_append(&path_list, buf.buf);
1099 strbuf_detach(&buf, NULL);
1100 }
1101 }
1102
1103 if (recurse_submodules && (!use_index || untracked))
1104 die(_("option not supported with --recurse-submodules"));
1105
1106 if (!show_in_pager && !opt.status_only)
1107 setup_pager();
1108
1109 if (!use_index && (untracked || cached))
1110 die(_("--cached or --untracked cannot be used with --no-index"));
1111
1112 if (!use_index || untracked) {
1113 int use_exclude = (opt_exclude < 0) ? use_index : !!opt_exclude;
1114 hit = grep_directory(&opt, &pathspec, use_exclude, use_index);
1115 } else if (0 <= opt_exclude) {
1116 die(_("--[no-]exclude-standard cannot be used for tracked contents"));
1117 } else if (!list.nr) {
1118 if (!cached)
1119 setup_work_tree();
1120
1121 hit = grep_cache(&opt, the_repository, &pathspec, cached);
1122 } else {
1123 if (cached)
1124 die(_("both --cached and trees are given"));
1125
1126 hit = grep_objects(&opt, &pathspec, &list);
1127 }
1128
1129 if (num_threads > 1)
1130 hit |= wait_all();
1131 if (hit && show_in_pager)
1132 run_pager(&opt, prefix);
1133 clear_pathspec(&pathspec);
1134 free_grep_patterns(&opt);
1135 return !hit;
1136 }