submodule: add a helper to check if it is safe to write to .gitmodules
[git/git.git] / builtin / grep.c
CommitLineData
5010cb5f
JH
1/*
2 * Builtin "git grep"
3 *
4 * Copyright (c) 2006 Junio C Hamano
5 */
6#include "cache.h"
627d9342 7#include "repository.h"
b2141fc1 8#include "config.h"
5010cb5f
JH
9#include "blob.h"
10#include "tree.h"
11#include "commit.h"
12#include "tag.h"
1362671f 13#include "tree-walk.h"
5010cb5f 14#include "builtin.h"
3e230fa1 15#include "parse-options.h"
678e484b
JS
16#include "string-list.h"
17#include "run-command.h"
60ecac98 18#include "userdiff.h"
83b5d2f5 19#include "grep.h"
493b7a08 20#include "quote.h"
59332d13 21#include "dir.h"
64acde94 22#include "pathspec.h"
0281e487 23#include "submodule.h"
74ed4371 24#include "submodule-config.h"
90c62155 25#include "object-store.h"
5b594f45 26
3e230fa1 27static char const * const grep_usage[] = {
9c9b4f2f 28 N_("git grep [<options>] [-e] <pattern> [<rev>...] [[--] <path>...]"),
3e230fa1
RS
29 NULL
30};
31
0281e487 32static int recurse_submodules;
0281e487 33
89f09dd3
VL
34#define GREP_NUM_THREADS_DEFAULT 8
35static int num_threads;
5b594f45
FK
36
37#ifndef NO_PTHREADS
89f09dd3 38static pthread_t *threads;
5b594f45 39
5b594f45
FK
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.
43 */
9cba13ca 44struct work_item {
8f24a632 45 struct grep_source source;
5b594f45
FK
46 char done;
47 struct strbuf out;
48};
49
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.
53 *
54 * The work_items in [todo_start, todo_end) are waiting to be picked
55 * up by a consumer thread.
56 *
57 * The ranges are modulo TODO_SIZE.
58 */
59#define TODO_SIZE 128
60static struct work_item todo[TODO_SIZE];
61static int todo_start;
62static int todo_end;
63static int todo_done;
64
65/* Has all work items been added? */
66static int all_work_added;
67
68/* This lock protects all the variables above. */
69static pthread_mutex_t grep_mutex;
70
1487a12b
JH
71static inline void grep_lock(void)
72{
8df4c295
ÆAB
73 assert(num_threads);
74 pthread_mutex_lock(&grep_mutex);
1487a12b
JH
75}
76
77static inline void grep_unlock(void)
78{
8df4c295
ÆAB
79 assert(num_threads);
80 pthread_mutex_unlock(&grep_mutex);
1487a12b
JH
81}
82
5b594f45
FK
83/* Signalled when a new work_item is added to todo. */
84static pthread_cond_t cond_add;
85
86/* Signalled when the result from one work_item is written to
87 * stdout.
88 */
89static pthread_cond_t cond_write;
90
91/* Signalled when we are finished with everything. */
92static pthread_cond_t cond_result;
93
08303c36 94static int skip_first_line;
431d6e7b 95
e2e05d61 96static void add_work(struct grep_opt *opt, const struct grep_source *gs)
5b594f45
FK
97{
98 grep_lock();
99
100 while ((todo_end+1) % ARRAY_SIZE(todo) == todo_done) {
101 pthread_cond_wait(&cond_write, &grep_mutex);
102 }
103
e2e05d61 104 todo[todo_end].source = *gs;
9dd5245c
JK
105 if (opt->binary != GREP_BINARY_TEXT)
106 grep_source_load_driver(&todo[todo_end].source);
5b594f45
FK
107 todo[todo_end].done = 0;
108 strbuf_reset(&todo[todo_end].out);
109 todo_end = (todo_end + 1) % ARRAY_SIZE(todo);
110
111 pthread_cond_signal(&cond_add);
112 grep_unlock();
113}
114
115static struct work_item *get_work(void)
116{
117 struct work_item *ret;
118
119 grep_lock();
120 while (todo_start == todo_end && !all_work_added) {
121 pthread_cond_wait(&cond_add, &grep_mutex);
122 }
123
124 if (todo_start == todo_end && all_work_added) {
125 ret = NULL;
126 } else {
127 ret = &todo[todo_start];
128 todo_start = (todo_start + 1) % ARRAY_SIZE(todo);
129 }
130 grep_unlock();
131 return ret;
132}
133
5b594f45
FK
134static void work_done(struct work_item *w)
135{
136 int old_done;
137
138 grep_lock();
139 w->done = 1;
140 old_done = todo_done;
141 for(; todo[todo_done].done && todo_done != todo_start;
142 todo_done = (todo_done+1) % ARRAY_SIZE(todo)) {
143 w = &todo[todo_done];
431d6e7b 144 if (w->out.len) {
08303c36
RS
145 const char *p = w->out.buf;
146 size_t len = w->out.len;
147
148 /* Skip the leading hunk mark of the first file. */
149 if (skip_first_line) {
150 while (len) {
151 len--;
152 if (*p++ == '\n')
153 break;
154 }
155 skip_first_line = 0;
156 }
157
158 write_or_die(1, p, len);
431d6e7b 159 }
8f24a632 160 grep_source_clear(&w->source);
5b594f45
FK
161 }
162
163 if (old_done != todo_done)
164 pthread_cond_signal(&cond_write);
165
166 if (all_work_added && todo_done == todo_end)
167 pthread_cond_signal(&cond_result);
168
169 grep_unlock();
170}
171
172static void *run(void *arg)
173{
174 int hit = 0;
175 struct grep_opt *opt = arg;
176
177 while (1) {
178 struct work_item *w = get_work();
179 if (!w)
180 break;
181
182 opt->output_priv = w;
f9ee2fcd 183 hit |= grep_source(opt, &w->source);
8f24a632 184 grep_source_clear_data(&w->source);
5b594f45
FK
185 work_done(w);
186 }
bfac23d9
DM
187 free_grep_patterns(arg);
188 free(arg);
5b594f45
FK
189
190 return (void*) (intptr_t) hit;
191}
192
193static void strbuf_out(struct grep_opt *opt, const void *buf, size_t size)
194{
195 struct work_item *w = opt->output_priv;
196 strbuf_add(&w->out, buf, size);
197}
198
199static void start_threads(struct grep_opt *opt)
200{
201 int i;
202
203 pthread_mutex_init(&grep_mutex, NULL);
b3aeb285 204 pthread_mutex_init(&grep_read_mutex, NULL);
0579f91d 205 pthread_mutex_init(&grep_attr_mutex, NULL);
5b594f45
FK
206 pthread_cond_init(&cond_add, NULL);
207 pthread_cond_init(&cond_write, NULL);
208 pthread_cond_init(&cond_result, NULL);
78db6ea9 209 grep_use_locks = 1;
5b594f45
FK
210
211 for (i = 0; i < ARRAY_SIZE(todo); i++) {
212 strbuf_init(&todo[i].out, 0);
213 }
214
89f09dd3
VL
215 threads = xcalloc(num_threads, sizeof(*threads));
216 for (i = 0; i < num_threads; i++) {
5b594f45
FK
217 int err;
218 struct grep_opt *o = grep_opt_dup(opt);
219 o->output = strbuf_out;
6d423dd5
ÆAB
220 if (i)
221 o->debug = 0;
5b594f45
FK
222 compile_grep_patterns(o);
223 err = pthread_create(&threads[i], NULL, run, o);
224
225 if (err)
2fc5f9f1 226 die(_("grep: failed to create thread: %s"),
5b594f45
FK
227 strerror(err));
228 }
229}
230
231static int wait_all(void)
232{
233 int hit = 0;
234 int i;
235
236 grep_lock();
237 all_work_added = 1;
238
239 /* Wait until all work is done. */
240 while (todo_done != todo_end)
241 pthread_cond_wait(&cond_result, &grep_mutex);
242
243 /* Wake up all the consumer threads so they can see that there
244 * is no more work to do.
245 */
246 pthread_cond_broadcast(&cond_add);
247 grep_unlock();
248
89f09dd3 249 for (i = 0; i < num_threads; i++) {
5b594f45
FK
250 void *h;
251 pthread_join(threads[i], &h);
252 hit |= (int) (intptr_t) h;
253 }
254
89f09dd3
VL
255 free(threads);
256
5b594f45 257 pthread_mutex_destroy(&grep_mutex);
b3aeb285 258 pthread_mutex_destroy(&grep_read_mutex);
0579f91d 259 pthread_mutex_destroy(&grep_attr_mutex);
5b594f45
FK
260 pthread_cond_destroy(&cond_add);
261 pthread_cond_destroy(&cond_write);
262 pthread_cond_destroy(&cond_result);
78db6ea9 263 grep_use_locks = 0;
5b594f45
FK
264
265 return hit;
266}
267#else /* !NO_PTHREADS */
5b594f45
FK
268
269static int wait_all(void)
270{
271 return 0;
272}
273#endif
274
15fabd1b
JH
275static int grep_cmd_config(const char *var, const char *value, void *cb)
276{
277 int st = grep_config(var, value, cb);
33c643bb 278 if (git_color_default_config(var, value, cb) < 0)
15fabd1b 279 st = -1;
89f09dd3
VL
280
281 if (!strcmp(var, "grep.threads")) {
282 num_threads = git_config_int(var, value);
283 if (num_threads < 0)
284 die(_("invalid number of threads specified (%d) for %s"),
285 num_threads, var);
d1edee4a
ÆAB
286#ifdef NO_PTHREADS
287 else if (num_threads && num_threads != 1) {
288 /*
289 * TRANSLATORS: %s is the configuration
290 * variable for tweaking threads, currently
291 * grep.threads
292 */
293 warning(_("no threads support, ignoring %s"), var);
294 num_threads = 0;
295 }
296#endif
89f09dd3
VL
297 }
298
9071c078
SB
299 if (!strcmp(var, "submodule.recurse"))
300 recurse_submodules = git_config_bool(var, value);
301
15fabd1b
JH
302 return st;
303}
304
1db11086 305static void *lock_and_read_oid_file(const struct object_id *oid, enum object_type *type, unsigned long *size)
5f02d315
JH
306{
307 void *data;
308
b3aeb285 309 grep_read_lock();
b4f5aca4 310 data = read_object_file(oid, type, size);
b3aeb285 311 grep_read_unlock();
5b594f45
FK
312 return data;
313}
314
1db11086 315static int grep_oid(struct grep_opt *opt, const struct object_id *oid,
55c61688
NTND
316 const char *filename, int tree_name_len,
317 const char *path)
5b594f45
FK
318{
319 struct strbuf pathbuf = STRBUF_INIT;
e2e05d61 320 struct grep_source gs;
5b594f45 321
f9ee2fcd
BW
322 if (opt->relative && opt->prefix_length) {
323 quote_path_relative(filename + tree_name_len, opt->prefix, &pathbuf);
324 strbuf_insert(&pathbuf, 0, filename, tree_name_len);
5b594f45
FK
325 } else {
326 strbuf_addstr(&pathbuf, filename);
327 }
328
e2e05d61 329 grep_source_init(&gs, GREP_SOURCE_OID, pathbuf.buf, path, oid);
38ef24dc 330 strbuf_release(&pathbuf);
e2e05d61 331
5b594f45 332#ifndef NO_PTHREADS
89f09dd3 333 if (num_threads) {
e2e05d61
RV
334 /*
335 * add_work() copies gs and thus assumes ownership of
336 * its fields, so do not call grep_source_clear()
337 */
338 add_work(opt, &gs);
5b594f45
FK
339 return 0;
340 } else
341#endif
342 {
343 int hit;
5010cb5f 344
8f24a632 345 hit = grep_source(opt, &gs);
dc49cd76 346
8f24a632
JK
347 grep_source_clear(&gs);
348 return hit;
5010cb5f 349 }
5b594f45
FK
350}
351
352static int grep_file(struct grep_opt *opt, const char *filename)
353{
354 struct strbuf buf = STRBUF_INIT;
e2e05d61 355 struct grep_source gs;
5b594f45 356
f9ee2fcd
BW
357 if (opt->relative && opt->prefix_length)
358 quote_path_relative(filename, opt->prefix, &buf);
359 else
360 strbuf_addstr(&buf, filename);
5b594f45 361
e2e05d61 362 grep_source_init(&gs, GREP_SOURCE_FILE, buf.buf, filename, filename);
38ef24dc 363 strbuf_release(&buf);
e2e05d61 364
5b594f45 365#ifndef NO_PTHREADS
89f09dd3 366 if (num_threads) {
e2e05d61
RV
367 /*
368 * add_work() copies gs and thus assumes ownership of
369 * its fields, so do not call grep_source_clear()
370 */
371 add_work(opt, &gs);
5b594f45
FK
372 return 0;
373 } else
374#endif
375 {
376 int hit;
5b594f45 377
8f24a632
JK
378 hit = grep_source(opt, &gs);
379
380 grep_source_clear(&gs);
5b594f45
FK
381 return hit;
382 }
5010cb5f
JH
383}
384
678e484b
JS
385static void append_path(struct grep_opt *opt, const void *data, size_t len)
386{
387 struct string_list *path_list = opt->output_priv;
388
389 if (len == 1 && *(const char *)data == '\0')
390 return;
0c72cead 391 string_list_append(path_list, xstrndup(data, len));
678e484b
JS
392}
393
394static void run_pager(struct grep_opt *opt, const char *prefix)
395{
396 struct string_list *path_list = opt->output_priv;
850d2fec 397 struct child_process child = CHILD_PROCESS_INIT;
678e484b
JS
398 int i, status;
399
400 for (i = 0; i < path_list->nr; i++)
850d2fec
JK
401 argv_array_push(&child.args, path_list->items[i].string);
402 child.dir = prefix;
403 child.use_shell = 1;
678e484b 404
850d2fec 405 status = run_command(&child);
678e484b
JS
406 if (status)
407 exit(status);
678e484b
JS
408}
409
f9ee2fcd
BW
410static int grep_cache(struct grep_opt *opt, struct repository *repo,
411 const struct pathspec *pathspec, int cached);
412static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
413 struct tree_desc *tree, struct strbuf *base, int tn_len,
414 int check_attr, struct repository *repo);
0281e487 415
f9ee2fcd
BW
416static int grep_submodule(struct grep_opt *opt, struct repository *superproject,
417 const struct pathspec *pathspec,
418 const struct object_id *oid,
419 const char *filename, const char *path)
0281e487 420{
f9ee2fcd
BW
421 struct repository submodule;
422 int hit;
74ed4371 423
f9ee2fcd
BW
424 if (!is_submodule_active(superproject, path))
425 return 0;
0281e487 426
f9ee2fcd
BW
427 if (repo_submodule_init(&submodule, superproject, path))
428 return 0;
be80a239 429
f9ee2fcd 430 repo_read_gitmodules(&submodule);
0281e487 431
74ed4371 432 /*
f9ee2fcd
BW
433 * NEEDSWORK: This adds the submodule's object directory to the list of
434 * alternates for the single in-memory object store. This has some bad
435 * consequences for memory (processed objects will never be freed) and
436 * performance (this increases the number of pack files git has to pay
437 * attention to, to the sum of the number of pack files in all the
438 * repositories processed so far). This can be removed once the object
439 * store is no longer global and instead is a member of the repository
440 * object.
74ed4371 441 */
9560e624 442 grep_read_lock();
90c62155 443 add_to_alternates_memory(submodule.objects->objectdir);
9560e624 444 grep_read_unlock();
74ed4371 445
f9ee2fcd
BW
446 if (oid) {
447 struct object *object;
448 struct tree_desc tree;
449 void *data;
450 unsigned long size;
451 struct strbuf base = STRBUF_INIT;
74ed4371 452
f9ee2fcd 453 object = parse_object_or_die(oid, oid_to_hex(oid));
0281e487 454
f9ee2fcd 455 grep_read_lock();
02f0547e 456 data = read_object_with_reference(&object->oid, tree_type,
f9ee2fcd
BW
457 &size, NULL);
458 grep_read_unlock();
0281e487 459
f9ee2fcd
BW
460 if (!data)
461 die(_("unable to read tree (%s)"), oid_to_hex(&object->oid));
0281e487 462
f9ee2fcd
BW
463 strbuf_addstr(&base, filename);
464 strbuf_addch(&base, '/');
0281e487 465
f9ee2fcd
BW
466 init_tree_desc(&tree, data, size);
467 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
468 object->type == OBJ_COMMIT, &submodule);
469 strbuf_release(&base);
470 free(data);
471 } else {
472 hit = grep_cache(opt, &submodule, pathspec, 1);
e6fac7f3 473 }
0281e487 474
f9ee2fcd
BW
475 repo_clear(&submodule);
476 return hit;
0281e487
BW
477}
478
f9ee2fcd
BW
479static int grep_cache(struct grep_opt *opt, struct repository *repo,
480 const struct pathspec *pathspec, int cached)
5010cb5f
JH
481{
482 int hit = 0;
483 int nr;
0281e487
BW
484 struct strbuf name = STRBUF_INIT;
485 int name_base_len = 0;
f9ee2fcd
BW
486 if (repo->submodule_prefix) {
487 name_base_len = strlen(repo->submodule_prefix);
488 strbuf_addstr(&name, repo->submodule_prefix);
0281e487
BW
489 }
490
b2aa84c7 491 if (repo_read_index(repo) < 0)
5507067d 492 die(_("index file corrupt"));
5010cb5f 493
f9ee2fcd
BW
494 for (nr = 0; nr < repo->index->cache_nr; nr++) {
495 const struct cache_entry *ce = repo->index->cache[nr];
0281e487
BW
496 strbuf_setlen(&name, name_base_len);
497 strbuf_addstr(&name, ce->name);
498
499 if (S_ISREG(ce->ce_mode) &&
a4009b0b 500 match_pathspec(repo->index, pathspec, name.buf, name.len, 0, NULL,
0281e487
BW
501 S_ISDIR(ce->ce_mode) ||
502 S_ISGITLINK(ce->ce_mode))) {
503 /*
504 * If CE_VALID is on, we assume worktree file and its
505 * cache entry are identical, even if worktree file has
506 * been modified, so use cache version instead
507 */
508 if (cached || (ce->ce_flags & CE_VALID) ||
509 ce_skip_worktree(ce)) {
510 if (ce_stage(ce) || ce_intent_to_add(ce))
511 continue;
f9ee2fcd
BW
512 hit |= grep_oid(opt, &ce->oid, name.buf,
513 0, name.buf);
0281e487 514 } else {
f9ee2fcd 515 hit |= grep_file(opt, name.buf);
0281e487
BW
516 }
517 } else if (recurse_submodules && S_ISGITLINK(ce->ce_mode) &&
a4009b0b 518 submodule_path_match(repo->index, pathspec, name.buf, NULL)) {
f9ee2fcd 519 hit |= grep_submodule(opt, repo, pathspec, NULL, ce->name, ce->name);
0281e487 520 } else {
5010cb5f 521 continue;
36f2587f 522 }
0281e487 523
36f2587f
JH
524 if (ce_stage(ce)) {
525 do {
526 nr++;
f9ee2fcd
BW
527 } while (nr < repo->index->cache_nr &&
528 !strcmp(ce->name, repo->index->cache[nr]->name));
36f2587f
JH
529 nr--; /* compensate for loop control */
530 }
c8610a2e
JH
531 if (hit && opt->status_only)
532 break;
5010cb5f 533 }
0281e487
BW
534
535 strbuf_release(&name);
5010cb5f
JH
536 return hit;
537}
538
f34bbc15 539static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
55c61688 540 struct tree_desc *tree, struct strbuf *base, int tn_len,
f9ee2fcd 541 int check_attr, struct repository *repo)
5010cb5f 542{
d688cf07
NTND
543 int hit = 0;
544 enum interesting match = entry_not_interesting;
4c068a98 545 struct name_entry entry;
e5e062b6 546 int old_baselen = base->len;
74ed4371
BW
547 struct strbuf name = STRBUF_INIT;
548 int name_base_len = 0;
f9ee2fcd
BW
549 if (repo->submodule_prefix) {
550 strbuf_addstr(&name, repo->submodule_prefix);
74ed4371
BW
551 name_base_len = name.len;
552 }
5010cb5f 553
4c068a98 554 while (tree_entry(tree, &entry)) {
0de16337 555 int te_len = tree_entry_len(&entry);
e5e062b6 556
d688cf07 557 if (match != all_entries_interesting) {
74ed4371
BW
558 strbuf_addstr(&name, base->buf + tn_len);
559 match = tree_entry_interesting(&entry, &name,
560 0, pathspec);
561 strbuf_setlen(&name, name_base_len);
562
d688cf07 563 if (match == all_entries_not_interesting)
97d0b74a 564 break;
d688cf07 565 if (match == entry_not_interesting)
1376e507
NTND
566 continue;
567 }
5010cb5f 568
1376e507 569 strbuf_add(base, entry.path, te_len);
e0eb889f 570
1376e507 571 if (S_ISREG(entry.mode)) {
1db11086 572 hit |= grep_oid(opt, entry.oid, base->buf, tn_len,
55c61688 573 check_attr ? base->buf + tn_len : NULL);
74ed4371 574 } else if (S_ISDIR(entry.mode)) {
21666f1a 575 enum object_type type;
5010cb5f
JH
576 struct tree_desc sub;
577 void *data;
6fda5e51
LT
578 unsigned long size;
579
1db11086 580 data = lock_and_read_oid_file(entry.oid, &type, &size);
5010cb5f 581 if (!data)
2fc5f9f1 582 die(_("unable to read tree (%s)"),
7d924c91 583 oid_to_hex(entry.oid));
1376e507
NTND
584
585 strbuf_addch(base, '/');
6fda5e51 586 init_tree_desc(&sub, data, size);
55c61688 587 hit |= grep_tree(opt, pathspec, &sub, base, tn_len,
f9ee2fcd 588 check_attr, repo);
5010cb5f 589 free(data);
74ed4371 590 } else if (recurse_submodules && S_ISGITLINK(entry.mode)) {
f9ee2fcd
BW
591 hit |= grep_submodule(opt, repo, pathspec, entry.oid,
592 base->buf, base->buf + tn_len);
5010cb5f 593 }
74ed4371 594
e5e062b6
NTND
595 strbuf_setlen(base, old_baselen);
596
c8610a2e
JH
597 if (hit && opt->status_only)
598 break;
5010cb5f 599 }
74ed4371
BW
600
601 strbuf_release(&name);
5010cb5f
JH
602 return hit;
603}
604
f34bbc15 605static int grep_object(struct grep_opt *opt, const struct pathspec *pathspec,
6856077a 606 struct object *obj, const char *name, const char *path)
5010cb5f 607{
1974632c 608 if (obj->type == OBJ_BLOB)
1db11086 609 return grep_oid(opt, &obj->oid, name, 0, path);
1974632c 610 if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
5010cb5f
JH
611 struct tree_desc tree;
612 void *data;
6fda5e51 613 unsigned long size;
e5e062b6
NTND
614 struct strbuf base;
615 int hit, len;
616
b3aeb285 617 grep_read_lock();
02f0547e 618 data = read_object_with_reference(&obj->oid, tree_type,
6fda5e51 619 &size, NULL);
b3aeb285 620 grep_read_unlock();
8cb5775b 621
5010cb5f 622 if (!data)
f2fd0760 623 die(_("unable to read tree (%s)"), oid_to_hex(&obj->oid));
e5e062b6
NTND
624
625 len = name ? strlen(name) : 0;
626 strbuf_init(&base, PATH_MAX + len + 1);
627 if (len) {
628 strbuf_add(&base, name, len);
629 strbuf_addch(&base, ':');
630 }
6fda5e51 631 init_tree_desc(&tree, data, size);
55c61688 632 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
6856077a 633 obj->type == OBJ_COMMIT, the_repository);
e5e062b6 634 strbuf_release(&base);
5010cb5f
JH
635 free(data);
636 return hit;
637 }
debca9d2 638 die(_("unable to grep from object of type %s"), type_name(obj->type));
5010cb5f
JH
639}
640
f34bbc15 641static int grep_objects(struct grep_opt *opt, const struct pathspec *pathspec,
30d00c39
JN
642 const struct object_array *list)
643{
644 unsigned int i;
645 int hit = 0;
646 const unsigned int nr = list->nr;
647
648 for (i = 0; i < nr; i++) {
649 struct object *real_obj;
a74093da
SB
650 real_obj = deref_tag(the_repository, list->objects[i].item,
651 NULL, 0);
74ed4371
BW
652
653 /* load the gitmodules file for this rev */
654 if (recurse_submodules) {
f793b895 655 submodule_free(the_repository);
cd73de47 656 gitmodules_config_oid(&real_obj->oid);
74ed4371 657 }
6856077a
JT
658 if (grep_object(opt, pathspec, real_obj, list->objects[i].name,
659 list->objects[i].path)) {
30d00c39
JN
660 hit = 1;
661 if (opt->status_only)
662 break;
663 }
664 }
665 return hit;
666}
667
dbfae86a 668static int grep_directory(struct grep_opt *opt, const struct pathspec *pathspec,
85975c0c 669 int exc_std, int use_index)
59332d13
JH
670{
671 struct dir_struct dir;
672 int i, hit = 0;
673
674 memset(&dir, 0, sizeof(dir));
85975c0c
JK
675 if (!use_index)
676 dir.flags |= DIR_NO_GITLINKS;
0a93fb8a
JH
677 if (exc_std)
678 setup_standard_excludes(&dir);
59332d13 679
0d32c183 680 fill_directory(&dir, &the_index, pathspec);
59332d13 681 for (i = 0; i < dir.nr; i++) {
6d2df284 682 if (!dir_path_match(&the_index, dir.entries[i], pathspec, 0, NULL))
9d8b831b 683 continue;
59332d13
JH
684 hit |= grep_file(opt, dir.entries[i]->name);
685 if (hit && opt->status_only)
686 break;
687 }
59332d13
JH
688 return hit;
689}
690
ff3c7f9a
RS
691static int context_callback(const struct option *opt, const char *arg,
692 int unset)
3e230fa1
RS
693{
694 struct grep_opt *grep_opt = opt->value;
695 int value;
696 const char *endp;
697
698 if (unset) {
699 grep_opt->pre_context = grep_opt->post_context = 0;
700 return 0;
701 }
702 value = strtol(arg, (char **)&endp, 10);
703 if (*endp) {
2fc5f9f1 704 return error(_("switch `%c' expects a numerical value"),
3e230fa1
RS
705 opt->short_name);
706 }
707 grep_opt->pre_context = grep_opt->post_context = value;
708 return 0;
709}
710
ff3c7f9a 711static int file_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
712{
713 struct grep_opt *grep_opt = opt->value;
c41dd2fd 714 int from_stdin = !strcmp(arg, "-");
3e230fa1
RS
715 FILE *patterns;
716 int lno = 0;
cfe370c6 717 struct strbuf sb = STRBUF_INIT;
3e230fa1 718
c41dd2fd 719 patterns = from_stdin ? stdin : fopen(arg, "r");
3e230fa1 720 if (!patterns)
2fc5f9f1 721 die_errno(_("cannot open '%s'"), arg);
a5518431 722 while (strbuf_getline(&sb, patterns) == 0) {
3e230fa1
RS
723 /* ignore empty line like grep does */
724 if (sb.len == 0)
725 continue;
ed40a095 726
ec830611
RS
727 append_grep_pat(grep_opt, sb.buf, sb.len, arg, ++lno,
728 GREP_PATTERN);
3e230fa1 729 }
c41dd2fd
RS
730 if (!from_stdin)
731 fclose(patterns);
3e230fa1
RS
732 strbuf_release(&sb);
733 return 0;
734}
735
ff3c7f9a 736static int not_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
737{
738 struct grep_opt *grep_opt = opt->value;
739 append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
740 return 0;
741}
742
ff3c7f9a 743static int and_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
744{
745 struct grep_opt *grep_opt = opt->value;
746 append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
747 return 0;
748}
749
ff3c7f9a 750static int open_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
751{
752 struct grep_opt *grep_opt = opt->value;
753 append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
754 return 0;
755}
756
ff3c7f9a 757static int close_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
758{
759 struct grep_opt *grep_opt = opt->value;
760 append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
761 return 0;
762}
763
ff3c7f9a
RS
764static int pattern_callback(const struct option *opt, const char *arg,
765 int unset)
3e230fa1
RS
766{
767 struct grep_opt *grep_opt = opt->value;
768 append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
769 return 0;
770}
5010cb5f 771
a633fca0 772int cmd_grep(int argc, const char **argv, const char *prefix)
5010cb5f 773{
5010cb5f 774 int hit = 0;
0a93fb8a 775 int cached = 0, untracked = 0, opt_exclude = -1;
5acd64ed 776 int seen_dashdash = 0;
bbc09c22 777 int external_grep_allowed__ignored;
0af88c15 778 const char *show_in_pager = NULL, *default_pager = "dummy";
5010cb5f 779 struct grep_opt opt;
3cd47459 780 struct object_array list = OBJECT_ARRAY_INIT;
f34bbc15 781 struct pathspec pathspec;
183113a5 782 struct string_list path_list = STRING_LIST_INIT_NODUP;
5acd64ed 783 int i;
3e230fa1 784 int dummy;
ff38d1a9 785 int use_index = 1;
84befcd0 786 int pattern_type_arg = GREP_PATTERN_TYPE_UNSPECIFIED;
131f3c96 787 int allow_revs;
cca2c172 788
3e230fa1 789 struct option options[] = {
d5d09d47 790 OPT_BOOL(0, "cached", &cached,
4b407bc5 791 N_("search in index instead of in the work tree")),
cbb08c2e 792 OPT_NEGBIT(0, "no-index", &use_index,
f63cf8c9 793 N_("find in contents not managed by git"), 1),
d5d09d47 794 OPT_BOOL(0, "untracked", &untracked,
4b407bc5 795 N_("search in both tracked and untracked files")),
0a93fb8a 796 OPT_SET_INT(0, "exclude-standard", &opt_exclude,
77fdb8a8 797 N_("ignore files specified via '.gitignore'"), 1),
0281e487 798 OPT_BOOL(0, "recurse-submodules", &recurse_submodules,
4fb1c6aa 799 N_("recursively search in each submodule")),
3e230fa1 800 OPT_GROUP(""),
d5d09d47 801 OPT_BOOL('v', "invert-match", &opt.invert,
4b407bc5 802 N_("show non-matching lines")),
d5d09d47 803 OPT_BOOL('i', "ignore-case", &opt.ignore_case,
4b407bc5 804 N_("case insensitive matching")),
d5d09d47 805 OPT_BOOL('w', "word-regexp", &opt.word_regexp,
4b407bc5 806 N_("match patterns only at word boundaries")),
3e230fa1 807 OPT_SET_INT('a', "text", &opt.binary,
4b407bc5 808 N_("process binary files as text"), GREP_BINARY_TEXT),
3e230fa1 809 OPT_SET_INT('I', NULL, &opt.binary,
4b407bc5 810 N_("don't match patterns in binary files"),
3e230fa1 811 GREP_BINARY_NOMATCH),
335ec3bf
JK
812 OPT_BOOL(0, "textconv", &opt.allow_textconv,
813 N_("process binary files with textconv filters")),
4b407bc5
NTND
814 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, N_("depth"),
815 N_("descend at most <depth> levels"), PARSE_OPT_NONEG,
a91f453f 816 NULL, 1 },
3e230fa1 817 OPT_GROUP(""),
84befcd0 818 OPT_SET_INT('E', "extended-regexp", &pattern_type_arg,
4b407bc5 819 N_("use extended POSIX regular expressions"),
84befcd0
S
820 GREP_PATTERN_TYPE_ERE),
821 OPT_SET_INT('G', "basic-regexp", &pattern_type_arg,
4b407bc5 822 N_("use basic POSIX regular expressions (default)"),
84befcd0
S
823 GREP_PATTERN_TYPE_BRE),
824 OPT_SET_INT('F', "fixed-strings", &pattern_type_arg,
4b407bc5 825 N_("interpret patterns as fixed strings"),
84befcd0
S
826 GREP_PATTERN_TYPE_FIXED),
827 OPT_SET_INT('P', "perl-regexp", &pattern_type_arg,
4b407bc5 828 N_("use Perl-compatible regular expressions"),
84befcd0 829 GREP_PATTERN_TYPE_PCRE),
3e230fa1 830 OPT_GROUP(""),
d5d09d47 831 OPT_BOOL('n', "line-number", &opt.linenum, N_("show line numbers")),
a449f27f 832 OPT_BOOL(0, "column", &opt.columnnum, N_("show column number of first match")),
4b407bc5
NTND
833 OPT_NEGBIT('h', NULL, &opt.pathname, N_("don't show filenames"), 1),
834 OPT_BIT('H', NULL, &opt.pathname, N_("show filenames"), 1),
3e230fa1 835 OPT_NEGBIT(0, "full-name", &opt.relative,
4b407bc5 836 N_("show filenames relative to top directory"), 1),
d5d09d47 837 OPT_BOOL('l', "files-with-matches", &opt.name_only,
4b407bc5 838 N_("show only filenames instead of matching lines")),
d5d09d47 839 OPT_BOOL(0, "name-only", &opt.name_only,
4b407bc5 840 N_("synonym for --files-with-matches")),
d5d09d47 841 OPT_BOOL('L', "files-without-match",
3e230fa1 842 &opt.unmatch_name_only,
4b407bc5 843 N_("show only the names of files without match")),
caf2de33
NTND
844 OPT_BOOL_F('z', "null", &opt.null_following_name,
845 N_("print NUL after filenames"),
846 PARSE_OPT_NOCOMPLETE),
9d8db06e
TB
847 OPT_BOOL('o', "only-matching", &opt.only_matching,
848 N_("show only matching parts of a line")),
d5d09d47 849 OPT_BOOL('c', "count", &opt.count,
4b407bc5
NTND
850 N_("show the number of matches instead of matching lines")),
851 OPT__COLOR(&opt.color, N_("highlight matches")),
d5d09d47 852 OPT_BOOL(0, "break", &opt.file_break,
4b407bc5 853 N_("print empty line between matches from different files")),
d5d09d47 854 OPT_BOOL(0, "heading", &opt.heading,
4b407bc5 855 N_("show filename only once above matches from same file")),
3e230fa1 856 OPT_GROUP(""),
4b407bc5
NTND
857 OPT_CALLBACK('C', "context", &opt, N_("n"),
858 N_("show <n> context lines before and after matches"),
3e230fa1 859 context_callback),
317f63c2 860 OPT_INTEGER('B', "before-context", &opt.pre_context,
4b407bc5 861 N_("show <n> context lines before matches")),
317f63c2 862 OPT_INTEGER('A', "after-context", &opt.post_context,
4b407bc5 863 N_("show <n> context lines after matches")),
89f09dd3
VL
864 OPT_INTEGER(0, "threads", &num_threads,
865 N_("use <n> worker threads")),
4b407bc5 866 OPT_NUMBER_CALLBACK(&opt, N_("shortcut for -C NUM"),
3e230fa1 867 context_callback),
d5d09d47 868 OPT_BOOL('p', "show-function", &opt.funcname,
4b407bc5 869 N_("show a line with the function name before matches")),
d5d09d47 870 OPT_BOOL('W', "function-context", &opt.funcbody,
4b407bc5 871 N_("show the surrounding function")),
3e230fa1 872 OPT_GROUP(""),
4b407bc5
NTND
873 OPT_CALLBACK('f', NULL, &opt, N_("file"),
874 N_("read patterns from file"), file_callback),
875 { OPTION_CALLBACK, 'e', NULL, &opt, N_("pattern"),
876 N_("match <pattern>"), PARSE_OPT_NONEG, pattern_callback },
3e230fa1 877 { OPTION_CALLBACK, 0, "and", &opt, NULL,
4b407bc5 878 N_("combine patterns specified with -e"),
3e230fa1 879 PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
d5d09d47 880 OPT_BOOL(0, "or", &dummy, ""),
3e230fa1
RS
881 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
882 PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
883 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
884 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
885 open_callback },
886 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
887 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
888 close_callback },
d52ee6e6 889 OPT__QUIET(&opt.status_only,
4b407bc5 890 N_("indicate hit with exit status without output")),
d5d09d47 891 OPT_BOOL(0, "all-match", &opt.all_match,
4b407bc5 892 N_("show only matches from files that match all patterns")),
3e4a67b4
NTND
893 OPT_SET_INT_F(0, "debug", &opt.debug,
894 N_("show parse tree for grep expression"),
895 1, PARSE_OPT_HIDDEN),
3e230fa1 896 OPT_GROUP(""),
0af88c15 897 { OPTION_STRING, 'O', "open-files-in-pager", &show_in_pager,
4b407bc5 898 N_("pager"), N_("show matching files in the pager"),
caf2de33
NTND
899 PARSE_OPT_OPTARG | PARSE_OPT_NOCOMPLETE,
900 NULL, (intptr_t)default_pager },
901 OPT_BOOL_F(0, "ext-grep", &external_grep_allowed__ignored,
902 N_("allow calling of grep(1) (ignored by this build)"),
903 PARSE_OPT_NOCOMPLETE),
3e230fa1
RS
904 OPT_END()
905 };
5010cb5f 906
15fabd1b
JH
907 init_grep_defaults();
908 git_config(grep_cmd_config, NULL);
909 grep_init(&opt, prefix);
7e8f59d5 910
5010cb5f 911 /*
5acd64ed
JH
912 * If there is no -- then the paths must exist in the working
913 * tree. If there is no explicit pattern specified with -e or
914 * -f, we take the first unrecognized non option to be the
915 * pattern, but then what follows it must be zero or more
916 * valid refs up to the -- (if exists), and then existing
917 * paths. If there is an explicit pattern, then the first
82e5a82f 918 * unrecognized non option is the beginning of the refs list
5acd64ed 919 * that continues up to the -- (if exists), and then paths.
5010cb5f 920 */
37782920 921 argc = parse_options(argc, argv, prefix, options, grep_usage,
3e230fa1 922 PARSE_OPT_KEEP_DASHDASH |
44415499 923 PARSE_OPT_STOP_AT_NON_OPTION);
c5c31d33 924 grep_commit_pattern_type(pattern_type_arg, &opt);
3e230fa1 925
ecd9ba61
TG
926 if (use_index && !startup_info->have_repository) {
927 int fallback = 0;
928 git_config_get_bool("grep.fallbacktonoindex", &fallback);
929 if (fallback)
930 use_index = 0;
931 else
932 /* die the same way as if we did it at the beginning */
933 setup_git_directory();
934 }
59332d13 935
1123c67c
JK
936 /*
937 * skip a -- separator; we know it cannot be
938 * separating revisions from pathnames if
939 * we haven't even had any patterns yet
940 */
941 if (argc > 0 && !opt.pattern_list && !strcmp(argv[0], "--")) {
942 argv++;
943 argc--;
944 }
945
3e230fa1
RS
946 /* First unrecognized non-option token */
947 if (argc > 0 && !opt.pattern_list) {
948 append_grep_pattern(&opt, argv[0], "command line", 0,
949 GREP_PATTERN);
950 argv++;
951 argc--;
5010cb5f 952 }
5acd64ed 953
0af88c15
JS
954 if (show_in_pager == default_pager)
955 show_in_pager = git_pager(1);
678e484b 956 if (show_in_pager) {
e7b082a4 957 opt.color = 0;
0af88c15
JS
958 opt.name_only = 1;
959 opt.null_following_name = 1;
960 opt.output_priv = &path_list;
961 opt.output = append_path;
0c72cead 962 string_list_append(&path_list, show_in_pager);
678e484b
JS
963 }
964
f9b9faf6 965 if (!opt.pattern_list)
1a07e59c 966 die(_("no pattern given"));
5b594f45 967
9d8db06e
TB
968 /* --only-matching has no effect with --invert. */
969 if (opt.invert)
970 opt.only_matching = 0;
971
b5b81136
JK
972 /*
973 * We have to find "--" in a separate pass, because its presence
974 * influences how we will parse arguments that come before it.
975 */
976 for (i = 0; i < argc; i++) {
977 if (!strcmp(argv[i], "--")) {
978 seen_dashdash = 1;
979 break;
980 }
981 }
982
983 /*
984 * Resolve any rev arguments. If we have a dashdash, then everything up
985 * to it must resolve as a rev. If not, then we stop at the first
986 * non-rev and assume everything else is a path.
987 */
131f3c96 988 allow_revs = use_index && !untracked;
3e230fa1 989 for (i = 0; i < argc; i++) {
5acd64ed 990 const char *arg = argv[i];
1db11086 991 struct object_id oid;
afa15f3c 992 struct object_context oc;
20d6421c
JK
993 struct object *object;
994
dca3b5f5
JT
995 if (!strcmp(arg, "--")) {
996 i++;
dca3b5f5
JT
997 break;
998 }
20d6421c 999
131f3c96 1000 if (!allow_revs) {
d0ffc069 1001 if (seen_dashdash)
131f3c96 1002 die(_("--no-index or --untracked cannot be used with revs"));
d0ffc069
JK
1003 break;
1004 }
1005
321c89bf 1006 if (get_oid_with_context(arg, GET_OID_RECORD_PATH,
e82caf38 1007 &oid, &oc)) {
b5b81136
JK
1008 if (seen_dashdash)
1009 die(_("unable to resolve revision: %s"), arg);
20d6421c 1010 break;
b5b81136 1011 }
20d6421c 1012
c251c83d 1013 object = parse_object_or_die(&oid, arg);
20d6421c
JK
1014 if (!seen_dashdash)
1015 verify_non_filename(prefix, arg);
1016 add_object_array_with_path(object, arg, &list, oc.mode, oc.path);
dc944b65 1017 free(oc.path);
1362671f 1018 }
5acd64ed 1019
b5b81136
JK
1020 /*
1021 * Anything left over is presumed to be a path. But in the non-dashdash
1022 * "do what I mean" case, we verify and complain when that isn't true.
1023 */
a0fe2b0d
JK
1024 if (!seen_dashdash) {
1025 int j;
1026 for (j = i; j < argc; j++)
131f3c96 1027 verify_filename(prefix, argv[j], j == i && allow_revs);
a0fe2b0d
JK
1028 }
1029
1030 parse_pathspec(&pathspec, 0,
1031 PATHSPEC_PREFER_CWD |
1032 (opt.max_depth != -1 ? PATHSPEC_MAXDEPTH_VALID : 0),
1033 prefix, argv + i);
1034 pathspec.max_depth = opt.max_depth;
1035 pathspec.recursive = 1;
eef3df5a 1036 pathspec.recurse_submodules = !!recurse_submodules;
a0fe2b0d 1037
53b8d931 1038#ifndef NO_PTHREADS
044b1f3c 1039 if (list.nr || cached || show_in_pager)
89f09dd3
VL
1040 num_threads = 0;
1041 else if (num_threads == 0)
1042 num_threads = GREP_NUM_THREADS_DEFAULT;
1043 else if (num_threads < 0)
1044 die(_("invalid number of threads specified (%d)"), num_threads);
9ec726a4
ÆAB
1045 if (num_threads == 1)
1046 num_threads = 0;
53b8d931 1047#else
d1edee4a
ÆAB
1048 if (num_threads)
1049 warning(_("no threads support, ignoring --threads"));
89f09dd3 1050 num_threads = 0;
53b8d931
TR
1051#endif
1052
6d423dd5
ÆAB
1053 if (!num_threads)
1054 /*
1055 * The compiled patterns on the main path are only
1056 * used when not using threading. Otherwise
1057 * start_threads() below calls compile_grep_patterns()
1058 * for each thread.
1059 */
1060 compile_grep_patterns(&opt);
1061
53b8d931 1062#ifndef NO_PTHREADS
89f09dd3 1063 if (num_threads) {
50dd0f2f
AY
1064 if (!(opt.name_only || opt.unmatch_name_only || opt.count)
1065 && (opt.pre_context || opt.post_context ||
1066 opt.file_break || opt.funcbody))
53b8d931
TR
1067 skip_first_line = 1;
1068 start_threads(&opt);
1069 }
1070#endif
1071
678e484b 1072 if (show_in_pager && (cached || list.nr))
e4fe4ba5 1073 die(_("--open-files-in-pager only works on the worktree"));
678e484b
JS
1074
1075 if (show_in_pager && opt.pattern_list && !opt.pattern_list->next) {
1076 const char *pager = path_list.items[0].string;
1077 int len = strlen(pager);
1078
1079 if (len > 4 && is_dir_sep(pager[len - 5]))
1080 pager += len - 4;
1081
f7febbea
JS
1082 if (opt.ignore_case && !strcmp("less", pager))
1083 string_list_append(&path_list, "-I");
1084
678e484b
JS
1085 if (!strcmp("less", pager) || !strcmp("vi", pager)) {
1086 struct strbuf buf = STRBUF_INIT;
1087 strbuf_addf(&buf, "+/%s%s",
1088 strcmp("less", pager) ? "" : "*",
1089 opt.pattern_list->pattern);
0c72cead 1090 string_list_append(&path_list, buf.buf);
678e484b
JS
1091 strbuf_detach(&buf, NULL);
1092 }
1093 }
1094
74ed4371 1095 if (recurse_submodules && (!use_index || untracked))
1a07e59c 1096 die(_("option not supported with --recurse-submodules"));
0281e487 1097
c2048f0b 1098 if (!show_in_pager && !opt.status_only)
678e484b
JS
1099 setup_pager();
1100
0a93fb8a 1101 if (!use_index && (untracked || cached))
1a07e59c 1102 die(_("--cached or --untracked cannot be used with --no-index"));
678e484b 1103
0a93fb8a 1104 if (!use_index || untracked) {
0a93fb8a 1105 int use_exclude = (opt_exclude < 0) ? use_index : !!opt_exclude;
85975c0c 1106 hit = grep_directory(&opt, &pathspec, use_exclude, use_index);
dbfae86a 1107 } else if (0 <= opt_exclude) {
1a07e59c 1108 die(_("--[no-]exclude-standard cannot be used for tracked contents"));
685359cf 1109 } else if (!list.nr) {
6577f542
NTND
1110 if (!cached)
1111 setup_work_tree();
5b594f45 1112
f9ee2fcd 1113 hit = grep_cache(&opt, the_repository, &pathspec, cached);
685359cf
JS
1114 } else {
1115 if (cached)
1a07e59c 1116 die(_("both --cached and trees are given"));
f9ee2fcd 1117
6856077a 1118 hit = grep_objects(&opt, &pathspec, &list);
5010cb5f 1119 }
5b594f45 1120
89f09dd3 1121 if (num_threads)
5b594f45 1122 hit |= wait_all();
678e484b
JS
1123 if (hit && show_in_pager)
1124 run_pager(&opt, prefix);
7861fa07 1125 clear_pathspec(&pathspec);
b48fb5b6 1126 free_grep_patterns(&opt);
5010cb5f
JH
1127 return !hit;
1128}