grep: remove #ifdef NO_PTHREADS
[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 36
89f09dd3 37static pthread_t *threads;
5b594f45 38
5b594f45
FK
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 */
9cba13ca 43struct work_item {
8f24a632 44 struct grep_source source;
5b594f45
FK
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
59static struct work_item todo[TODO_SIZE];
60static int todo_start;
61static int todo_end;
62static int todo_done;
63
64/* Has all work items been added? */
65static int all_work_added;
66
67/* This lock protects all the variables above. */
68static pthread_mutex_t grep_mutex;
69
1487a12b
JH
70static inline void grep_lock(void)
71{
8df4c295
ÆAB
72 assert(num_threads);
73 pthread_mutex_lock(&grep_mutex);
1487a12b
JH
74}
75
76static inline void grep_unlock(void)
77{
8df4c295
ÆAB
78 assert(num_threads);
79 pthread_mutex_unlock(&grep_mutex);
1487a12b
JH
80}
81
5b594f45
FK
82/* Signalled when a new work_item is added to todo. */
83static pthread_cond_t cond_add;
84
85/* Signalled when the result from one work_item is written to
86 * stdout.
87 */
88static pthread_cond_t cond_write;
89
90/* Signalled when we are finished with everything. */
91static pthread_cond_t cond_result;
92
08303c36 93static int skip_first_line;
431d6e7b 94
e2e05d61 95static void add_work(struct grep_opt *opt, const struct grep_source *gs)
5b594f45
FK
96{
97 grep_lock();
98
99 while ((todo_end+1) % ARRAY_SIZE(todo) == todo_done) {
100 pthread_cond_wait(&cond_write, &grep_mutex);
101 }
102
e2e05d61 103 todo[todo_end].source = *gs;
9dd5245c 104 if (opt->binary != GREP_BINARY_TEXT)
acd00ea0
NTND
105 grep_source_load_driver(&todo[todo_end].source,
106 opt->repo->index);
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
4002e87c
NTND
236 if (!HAVE_THREADS)
237 return 0;
238
5b594f45
FK
239 grep_lock();
240 all_work_added = 1;
241
242 /* Wait until all work is done. */
243 while (todo_done != todo_end)
244 pthread_cond_wait(&cond_result, &grep_mutex);
245
246 /* Wake up all the consumer threads so they can see that there
247 * is no more work to do.
248 */
249 pthread_cond_broadcast(&cond_add);
250 grep_unlock();
251
89f09dd3 252 for (i = 0; i < num_threads; i++) {
5b594f45
FK
253 void *h;
254 pthread_join(threads[i], &h);
255 hit |= (int) (intptr_t) h;
256 }
257
89f09dd3
VL
258 free(threads);
259
5b594f45 260 pthread_mutex_destroy(&grep_mutex);
b3aeb285 261 pthread_mutex_destroy(&grep_read_mutex);
0579f91d 262 pthread_mutex_destroy(&grep_attr_mutex);
5b594f45
FK
263 pthread_cond_destroy(&cond_add);
264 pthread_cond_destroy(&cond_write);
265 pthread_cond_destroy(&cond_result);
78db6ea9 266 grep_use_locks = 0;
5b594f45
FK
267
268 return hit;
269}
5b594f45 270
15fabd1b
JH
271static int grep_cmd_config(const char *var, const char *value, void *cb)
272{
273 int st = grep_config(var, value, cb);
33c643bb 274 if (git_color_default_config(var, value, cb) < 0)
15fabd1b 275 st = -1;
89f09dd3
VL
276
277 if (!strcmp(var, "grep.threads")) {
278 num_threads = git_config_int(var, value);
279 if (num_threads < 0)
280 die(_("invalid number of threads specified (%d) for %s"),
281 num_threads, var);
4002e87c 282 else if (!HAVE_THREADS && num_threads && num_threads != 1) {
d1edee4a
ÆAB
283 /*
284 * TRANSLATORS: %s is the configuration
285 * variable for tweaking threads, currently
286 * grep.threads
287 */
288 warning(_("no threads support, ignoring %s"), var);
289 num_threads = 0;
290 }
89f09dd3
VL
291 }
292
9071c078
SB
293 if (!strcmp(var, "submodule.recurse"))
294 recurse_submodules = git_config_bool(var, value);
295
15fabd1b
JH
296 return st;
297}
298
1db11086 299static void *lock_and_read_oid_file(const struct object_id *oid, enum object_type *type, unsigned long *size)
5f02d315
JH
300{
301 void *data;
302
b3aeb285 303 grep_read_lock();
b4f5aca4 304 data = read_object_file(oid, type, size);
b3aeb285 305 grep_read_unlock();
5b594f45
FK
306 return data;
307}
308
1db11086 309static int grep_oid(struct grep_opt *opt, const struct object_id *oid,
55c61688
NTND
310 const char *filename, int tree_name_len,
311 const char *path)
5b594f45
FK
312{
313 struct strbuf pathbuf = STRBUF_INIT;
e2e05d61 314 struct grep_source gs;
5b594f45 315
f9ee2fcd
BW
316 if (opt->relative && opt->prefix_length) {
317 quote_path_relative(filename + tree_name_len, opt->prefix, &pathbuf);
318 strbuf_insert(&pathbuf, 0, filename, tree_name_len);
5b594f45
FK
319 } else {
320 strbuf_addstr(&pathbuf, filename);
321 }
322
e2e05d61 323 grep_source_init(&gs, GREP_SOURCE_OID, pathbuf.buf, path, oid);
38ef24dc 324 strbuf_release(&pathbuf);
e2e05d61 325
4002e87c 326 if (HAVE_THREADS && num_threads) {
e2e05d61
RV
327 /*
328 * add_work() copies gs and thus assumes ownership of
329 * its fields, so do not call grep_source_clear()
330 */
331 add_work(opt, &gs);
5b594f45 332 return 0;
4002e87c 333 } else {
5b594f45 334 int hit;
5010cb5f 335
8f24a632 336 hit = grep_source(opt, &gs);
dc49cd76 337
8f24a632
JK
338 grep_source_clear(&gs);
339 return hit;
5010cb5f 340 }
5b594f45
FK
341}
342
343static int grep_file(struct grep_opt *opt, const char *filename)
344{
345 struct strbuf buf = STRBUF_INIT;
e2e05d61 346 struct grep_source gs;
5b594f45 347
f9ee2fcd
BW
348 if (opt->relative && opt->prefix_length)
349 quote_path_relative(filename, opt->prefix, &buf);
350 else
351 strbuf_addstr(&buf, filename);
5b594f45 352
e2e05d61 353 grep_source_init(&gs, GREP_SOURCE_FILE, buf.buf, filename, filename);
38ef24dc 354 strbuf_release(&buf);
e2e05d61 355
4002e87c 356 if (HAVE_THREADS && num_threads) {
e2e05d61
RV
357 /*
358 * add_work() copies gs and thus assumes ownership of
359 * its fields, so do not call grep_source_clear()
360 */
361 add_work(opt, &gs);
5b594f45 362 return 0;
4002e87c 363 } else {
5b594f45 364 int hit;
5b594f45 365
8f24a632
JK
366 hit = grep_source(opt, &gs);
367
368 grep_source_clear(&gs);
5b594f45
FK
369 return hit;
370 }
5010cb5f
JH
371}
372
678e484b
JS
373static void append_path(struct grep_opt *opt, const void *data, size_t len)
374{
375 struct string_list *path_list = opt->output_priv;
376
377 if (len == 1 && *(const char *)data == '\0')
378 return;
0c72cead 379 string_list_append(path_list, xstrndup(data, len));
678e484b
JS
380}
381
382static void run_pager(struct grep_opt *opt, const char *prefix)
383{
384 struct string_list *path_list = opt->output_priv;
850d2fec 385 struct child_process child = CHILD_PROCESS_INIT;
678e484b
JS
386 int i, status;
387
388 for (i = 0; i < path_list->nr; i++)
850d2fec
JK
389 argv_array_push(&child.args, path_list->items[i].string);
390 child.dir = prefix;
391 child.use_shell = 1;
678e484b 392
850d2fec 393 status = run_command(&child);
678e484b
JS
394 if (status)
395 exit(status);
678e484b
JS
396}
397
f9ee2fcd
BW
398static int grep_cache(struct grep_opt *opt, struct repository *repo,
399 const struct pathspec *pathspec, int cached);
400static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
401 struct tree_desc *tree, struct strbuf *base, int tn_len,
402 int check_attr, struct repository *repo);
0281e487 403
f9ee2fcd
BW
404static int grep_submodule(struct grep_opt *opt, struct repository *superproject,
405 const struct pathspec *pathspec,
406 const struct object_id *oid,
407 const char *filename, const char *path)
0281e487 408{
f9ee2fcd
BW
409 struct repository submodule;
410 int hit;
74ed4371 411
f9ee2fcd
BW
412 if (!is_submodule_active(superproject, path))
413 return 0;
0281e487 414
f9ee2fcd
BW
415 if (repo_submodule_init(&submodule, superproject, path))
416 return 0;
be80a239 417
f9ee2fcd 418 repo_read_gitmodules(&submodule);
0281e487 419
74ed4371 420 /*
f9ee2fcd
BW
421 * NEEDSWORK: This adds the submodule's object directory to the list of
422 * alternates for the single in-memory object store. This has some bad
423 * consequences for memory (processed objects will never be freed) and
424 * performance (this increases the number of pack files git has to pay
425 * attention to, to the sum of the number of pack files in all the
426 * repositories processed so far). This can be removed once the object
427 * store is no longer global and instead is a member of the repository
428 * object.
74ed4371 429 */
9560e624 430 grep_read_lock();
90c62155 431 add_to_alternates_memory(submodule.objects->objectdir);
9560e624 432 grep_read_unlock();
74ed4371 433
f9ee2fcd
BW
434 if (oid) {
435 struct object *object;
436 struct tree_desc tree;
437 void *data;
438 unsigned long size;
439 struct strbuf base = STRBUF_INIT;
74ed4371 440
f9ee2fcd 441 object = parse_object_or_die(oid, oid_to_hex(oid));
0281e487 442
f9ee2fcd 443 grep_read_lock();
02f0547e 444 data = read_object_with_reference(&object->oid, tree_type,
f9ee2fcd
BW
445 &size, NULL);
446 grep_read_unlock();
0281e487 447
f9ee2fcd
BW
448 if (!data)
449 die(_("unable to read tree (%s)"), oid_to_hex(&object->oid));
0281e487 450
f9ee2fcd
BW
451 strbuf_addstr(&base, filename);
452 strbuf_addch(&base, '/');
0281e487 453
f9ee2fcd
BW
454 init_tree_desc(&tree, data, size);
455 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
456 object->type == OBJ_COMMIT, &submodule);
457 strbuf_release(&base);
458 free(data);
459 } else {
460 hit = grep_cache(opt, &submodule, pathspec, 1);
e6fac7f3 461 }
0281e487 462
f9ee2fcd
BW
463 repo_clear(&submodule);
464 return hit;
0281e487
BW
465}
466
f9ee2fcd
BW
467static int grep_cache(struct grep_opt *opt, struct repository *repo,
468 const struct pathspec *pathspec, int cached)
5010cb5f
JH
469{
470 int hit = 0;
471 int nr;
0281e487
BW
472 struct strbuf name = STRBUF_INIT;
473 int name_base_len = 0;
f9ee2fcd
BW
474 if (repo->submodule_prefix) {
475 name_base_len = strlen(repo->submodule_prefix);
476 strbuf_addstr(&name, repo->submodule_prefix);
0281e487
BW
477 }
478
b2aa84c7 479 if (repo_read_index(repo) < 0)
5507067d 480 die(_("index file corrupt"));
5010cb5f 481
f9ee2fcd
BW
482 for (nr = 0; nr < repo->index->cache_nr; nr++) {
483 const struct cache_entry *ce = repo->index->cache[nr];
0281e487
BW
484 strbuf_setlen(&name, name_base_len);
485 strbuf_addstr(&name, ce->name);
486
487 if (S_ISREG(ce->ce_mode) &&
a4009b0b 488 match_pathspec(repo->index, pathspec, name.buf, name.len, 0, NULL,
0281e487
BW
489 S_ISDIR(ce->ce_mode) ||
490 S_ISGITLINK(ce->ce_mode))) {
491 /*
492 * If CE_VALID is on, we assume worktree file and its
493 * cache entry are identical, even if worktree file has
494 * been modified, so use cache version instead
495 */
496 if (cached || (ce->ce_flags & CE_VALID) ||
497 ce_skip_worktree(ce)) {
498 if (ce_stage(ce) || ce_intent_to_add(ce))
499 continue;
f9ee2fcd
BW
500 hit |= grep_oid(opt, &ce->oid, name.buf,
501 0, name.buf);
0281e487 502 } else {
f9ee2fcd 503 hit |= grep_file(opt, name.buf);
0281e487
BW
504 }
505 } else if (recurse_submodules && S_ISGITLINK(ce->ce_mode) &&
a4009b0b 506 submodule_path_match(repo->index, pathspec, name.buf, NULL)) {
f9ee2fcd 507 hit |= grep_submodule(opt, repo, pathspec, NULL, ce->name, ce->name);
0281e487 508 } else {
5010cb5f 509 continue;
36f2587f 510 }
0281e487 511
36f2587f
JH
512 if (ce_stage(ce)) {
513 do {
514 nr++;
f9ee2fcd
BW
515 } while (nr < repo->index->cache_nr &&
516 !strcmp(ce->name, repo->index->cache[nr]->name));
36f2587f
JH
517 nr--; /* compensate for loop control */
518 }
c8610a2e
JH
519 if (hit && opt->status_only)
520 break;
5010cb5f 521 }
0281e487
BW
522
523 strbuf_release(&name);
5010cb5f
JH
524 return hit;
525}
526
f34bbc15 527static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
55c61688 528 struct tree_desc *tree, struct strbuf *base, int tn_len,
f9ee2fcd 529 int check_attr, struct repository *repo)
5010cb5f 530{
d688cf07
NTND
531 int hit = 0;
532 enum interesting match = entry_not_interesting;
4c068a98 533 struct name_entry entry;
e5e062b6 534 int old_baselen = base->len;
74ed4371
BW
535 struct strbuf name = STRBUF_INIT;
536 int name_base_len = 0;
f9ee2fcd
BW
537 if (repo->submodule_prefix) {
538 strbuf_addstr(&name, repo->submodule_prefix);
74ed4371
BW
539 name_base_len = name.len;
540 }
5010cb5f 541
4c068a98 542 while (tree_entry(tree, &entry)) {
0de16337 543 int te_len = tree_entry_len(&entry);
e5e062b6 544
d688cf07 545 if (match != all_entries_interesting) {
74ed4371
BW
546 strbuf_addstr(&name, base->buf + tn_len);
547 match = tree_entry_interesting(&entry, &name,
548 0, pathspec);
549 strbuf_setlen(&name, name_base_len);
550
d688cf07 551 if (match == all_entries_not_interesting)
97d0b74a 552 break;
d688cf07 553 if (match == entry_not_interesting)
1376e507
NTND
554 continue;
555 }
5010cb5f 556
1376e507 557 strbuf_add(base, entry.path, te_len);
e0eb889f 558
1376e507 559 if (S_ISREG(entry.mode)) {
1db11086 560 hit |= grep_oid(opt, entry.oid, base->buf, tn_len,
55c61688 561 check_attr ? base->buf + tn_len : NULL);
74ed4371 562 } else if (S_ISDIR(entry.mode)) {
21666f1a 563 enum object_type type;
5010cb5f
JH
564 struct tree_desc sub;
565 void *data;
6fda5e51
LT
566 unsigned long size;
567
1db11086 568 data = lock_and_read_oid_file(entry.oid, &type, &size);
5010cb5f 569 if (!data)
2fc5f9f1 570 die(_("unable to read tree (%s)"),
7d924c91 571 oid_to_hex(entry.oid));
1376e507
NTND
572
573 strbuf_addch(base, '/');
6fda5e51 574 init_tree_desc(&sub, data, size);
55c61688 575 hit |= grep_tree(opt, pathspec, &sub, base, tn_len,
f9ee2fcd 576 check_attr, repo);
5010cb5f 577 free(data);
74ed4371 578 } else if (recurse_submodules && S_ISGITLINK(entry.mode)) {
f9ee2fcd
BW
579 hit |= grep_submodule(opt, repo, pathspec, entry.oid,
580 base->buf, base->buf + tn_len);
5010cb5f 581 }
74ed4371 582
e5e062b6
NTND
583 strbuf_setlen(base, old_baselen);
584
c8610a2e
JH
585 if (hit && opt->status_only)
586 break;
5010cb5f 587 }
74ed4371
BW
588
589 strbuf_release(&name);
5010cb5f
JH
590 return hit;
591}
592
f34bbc15 593static int grep_object(struct grep_opt *opt, const struct pathspec *pathspec,
6856077a 594 struct object *obj, const char *name, const char *path)
5010cb5f 595{
1974632c 596 if (obj->type == OBJ_BLOB)
1db11086 597 return grep_oid(opt, &obj->oid, name, 0, path);
1974632c 598 if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
5010cb5f
JH
599 struct tree_desc tree;
600 void *data;
6fda5e51 601 unsigned long size;
e5e062b6
NTND
602 struct strbuf base;
603 int hit, len;
604
b3aeb285 605 grep_read_lock();
02f0547e 606 data = read_object_with_reference(&obj->oid, tree_type,
6fda5e51 607 &size, NULL);
b3aeb285 608 grep_read_unlock();
8cb5775b 609
5010cb5f 610 if (!data)
f2fd0760 611 die(_("unable to read tree (%s)"), oid_to_hex(&obj->oid));
e5e062b6
NTND
612
613 len = name ? strlen(name) : 0;
614 strbuf_init(&base, PATH_MAX + len + 1);
615 if (len) {
616 strbuf_add(&base, name, len);
617 strbuf_addch(&base, ':');
618 }
6fda5e51 619 init_tree_desc(&tree, data, size);
55c61688 620 hit = grep_tree(opt, pathspec, &tree, &base, base.len,
6856077a 621 obj->type == OBJ_COMMIT, the_repository);
e5e062b6 622 strbuf_release(&base);
5010cb5f
JH
623 free(data);
624 return hit;
625 }
debca9d2 626 die(_("unable to grep from object of type %s"), type_name(obj->type));
5010cb5f
JH
627}
628
f34bbc15 629static int grep_objects(struct grep_opt *opt, const struct pathspec *pathspec,
30d00c39
JN
630 const struct object_array *list)
631{
632 unsigned int i;
633 int hit = 0;
634 const unsigned int nr = list->nr;
635
636 for (i = 0; i < nr; i++) {
637 struct object *real_obj;
a74093da
SB
638 real_obj = deref_tag(the_repository, list->objects[i].item,
639 NULL, 0);
74ed4371
BW
640
641 /* load the gitmodules file for this rev */
642 if (recurse_submodules) {
f793b895 643 submodule_free(the_repository);
cd73de47 644 gitmodules_config_oid(&real_obj->oid);
74ed4371 645 }
6856077a
JT
646 if (grep_object(opt, pathspec, real_obj, list->objects[i].name,
647 list->objects[i].path)) {
30d00c39
JN
648 hit = 1;
649 if (opt->status_only)
650 break;
651 }
652 }
653 return hit;
654}
655
dbfae86a 656static int grep_directory(struct grep_opt *opt, const struct pathspec *pathspec,
85975c0c 657 int exc_std, int use_index)
59332d13
JH
658{
659 struct dir_struct dir;
660 int i, hit = 0;
661
662 memset(&dir, 0, sizeof(dir));
85975c0c
JK
663 if (!use_index)
664 dir.flags |= DIR_NO_GITLINKS;
0a93fb8a
JH
665 if (exc_std)
666 setup_standard_excludes(&dir);
59332d13 667
0d32c183 668 fill_directory(&dir, &the_index, pathspec);
59332d13 669 for (i = 0; i < dir.nr; i++) {
6d2df284 670 if (!dir_path_match(&the_index, dir.entries[i], pathspec, 0, NULL))
9d8b831b 671 continue;
59332d13
JH
672 hit |= grep_file(opt, dir.entries[i]->name);
673 if (hit && opt->status_only)
674 break;
675 }
59332d13
JH
676 return hit;
677}
678
ff3c7f9a
RS
679static int context_callback(const struct option *opt, const char *arg,
680 int unset)
3e230fa1
RS
681{
682 struct grep_opt *grep_opt = opt->value;
683 int value;
684 const char *endp;
685
686 if (unset) {
687 grep_opt->pre_context = grep_opt->post_context = 0;
688 return 0;
689 }
690 value = strtol(arg, (char **)&endp, 10);
691 if (*endp) {
2fc5f9f1 692 return error(_("switch `%c' expects a numerical value"),
3e230fa1
RS
693 opt->short_name);
694 }
695 grep_opt->pre_context = grep_opt->post_context = value;
696 return 0;
697}
698
ff3c7f9a 699static int file_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
700{
701 struct grep_opt *grep_opt = opt->value;
c41dd2fd 702 int from_stdin = !strcmp(arg, "-");
3e230fa1
RS
703 FILE *patterns;
704 int lno = 0;
cfe370c6 705 struct strbuf sb = STRBUF_INIT;
3e230fa1 706
c41dd2fd 707 patterns = from_stdin ? stdin : fopen(arg, "r");
3e230fa1 708 if (!patterns)
2fc5f9f1 709 die_errno(_("cannot open '%s'"), arg);
a5518431 710 while (strbuf_getline(&sb, patterns) == 0) {
3e230fa1
RS
711 /* ignore empty line like grep does */
712 if (sb.len == 0)
713 continue;
ed40a095 714
ec830611
RS
715 append_grep_pat(grep_opt, sb.buf, sb.len, arg, ++lno,
716 GREP_PATTERN);
3e230fa1 717 }
c41dd2fd
RS
718 if (!from_stdin)
719 fclose(patterns);
3e230fa1
RS
720 strbuf_release(&sb);
721 return 0;
722}
723
ff3c7f9a 724static int not_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
725{
726 struct grep_opt *grep_opt = opt->value;
727 append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
728 return 0;
729}
730
ff3c7f9a 731static int and_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
732{
733 struct grep_opt *grep_opt = opt->value;
734 append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
735 return 0;
736}
737
ff3c7f9a 738static int open_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
739{
740 struct grep_opt *grep_opt = opt->value;
741 append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
742 return 0;
743}
744
ff3c7f9a 745static int close_callback(const struct option *opt, const char *arg, int unset)
3e230fa1
RS
746{
747 struct grep_opt *grep_opt = opt->value;
748 append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
749 return 0;
750}
751
ff3c7f9a
RS
752static int pattern_callback(const struct option *opt, const char *arg,
753 int unset)
3e230fa1
RS
754{
755 struct grep_opt *grep_opt = opt->value;
756 append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
757 return 0;
758}
5010cb5f 759
a633fca0 760int cmd_grep(int argc, const char **argv, const char *prefix)
5010cb5f 761{
5010cb5f 762 int hit = 0;
0a93fb8a 763 int cached = 0, untracked = 0, opt_exclude = -1;
5acd64ed 764 int seen_dashdash = 0;
bbc09c22 765 int external_grep_allowed__ignored;
0af88c15 766 const char *show_in_pager = NULL, *default_pager = "dummy";
5010cb5f 767 struct grep_opt opt;
3cd47459 768 struct object_array list = OBJECT_ARRAY_INIT;
f34bbc15 769 struct pathspec pathspec;
183113a5 770 struct string_list path_list = STRING_LIST_INIT_NODUP;
5acd64ed 771 int i;
3e230fa1 772 int dummy;
ff38d1a9 773 int use_index = 1;
84befcd0 774 int pattern_type_arg = GREP_PATTERN_TYPE_UNSPECIFIED;
131f3c96 775 int allow_revs;
cca2c172 776
3e230fa1 777 struct option options[] = {
d5d09d47 778 OPT_BOOL(0, "cached", &cached,
4b407bc5 779 N_("search in index instead of in the work tree")),
cbb08c2e 780 OPT_NEGBIT(0, "no-index", &use_index,
f63cf8c9 781 N_("find in contents not managed by git"), 1),
d5d09d47 782 OPT_BOOL(0, "untracked", &untracked,
4b407bc5 783 N_("search in both tracked and untracked files")),
0a93fb8a 784 OPT_SET_INT(0, "exclude-standard", &opt_exclude,
77fdb8a8 785 N_("ignore files specified via '.gitignore'"), 1),
0281e487 786 OPT_BOOL(0, "recurse-submodules", &recurse_submodules,
4fb1c6aa 787 N_("recursively search in each submodule")),
3e230fa1 788 OPT_GROUP(""),
d5d09d47 789 OPT_BOOL('v', "invert-match", &opt.invert,
4b407bc5 790 N_("show non-matching lines")),
d5d09d47 791 OPT_BOOL('i', "ignore-case", &opt.ignore_case,
4b407bc5 792 N_("case insensitive matching")),
d5d09d47 793 OPT_BOOL('w', "word-regexp", &opt.word_regexp,
4b407bc5 794 N_("match patterns only at word boundaries")),
3e230fa1 795 OPT_SET_INT('a', "text", &opt.binary,
4b407bc5 796 N_("process binary files as text"), GREP_BINARY_TEXT),
3e230fa1 797 OPT_SET_INT('I', NULL, &opt.binary,
4b407bc5 798 N_("don't match patterns in binary files"),
3e230fa1 799 GREP_BINARY_NOMATCH),
335ec3bf
JK
800 OPT_BOOL(0, "textconv", &opt.allow_textconv,
801 N_("process binary files with textconv filters")),
0a09e5ed
RS
802 OPT_SET_INT('r', "recursive", &opt.max_depth,
803 N_("search in subdirectories (default)"), -1),
4b407bc5
NTND
804 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, N_("depth"),
805 N_("descend at most <depth> levels"), PARSE_OPT_NONEG,
a91f453f 806 NULL, 1 },
3e230fa1 807 OPT_GROUP(""),
84befcd0 808 OPT_SET_INT('E', "extended-regexp", &pattern_type_arg,
4b407bc5 809 N_("use extended POSIX regular expressions"),
84befcd0
S
810 GREP_PATTERN_TYPE_ERE),
811 OPT_SET_INT('G', "basic-regexp", &pattern_type_arg,
4b407bc5 812 N_("use basic POSIX regular expressions (default)"),
84befcd0
S
813 GREP_PATTERN_TYPE_BRE),
814 OPT_SET_INT('F', "fixed-strings", &pattern_type_arg,
4b407bc5 815 N_("interpret patterns as fixed strings"),
84befcd0
S
816 GREP_PATTERN_TYPE_FIXED),
817 OPT_SET_INT('P', "perl-regexp", &pattern_type_arg,
4b407bc5 818 N_("use Perl-compatible regular expressions"),
84befcd0 819 GREP_PATTERN_TYPE_PCRE),
3e230fa1 820 OPT_GROUP(""),
d5d09d47 821 OPT_BOOL('n', "line-number", &opt.linenum, N_("show line numbers")),
a449f27f 822 OPT_BOOL(0, "column", &opt.columnnum, N_("show column number of first match")),
4b407bc5
NTND
823 OPT_NEGBIT('h', NULL, &opt.pathname, N_("don't show filenames"), 1),
824 OPT_BIT('H', NULL, &opt.pathname, N_("show filenames"), 1),
3e230fa1 825 OPT_NEGBIT(0, "full-name", &opt.relative,
4b407bc5 826 N_("show filenames relative to top directory"), 1),
d5d09d47 827 OPT_BOOL('l', "files-with-matches", &opt.name_only,
4b407bc5 828 N_("show only filenames instead of matching lines")),
d5d09d47 829 OPT_BOOL(0, "name-only", &opt.name_only,
4b407bc5 830 N_("synonym for --files-with-matches")),
d5d09d47 831 OPT_BOOL('L', "files-without-match",
3e230fa1 832 &opt.unmatch_name_only,
4b407bc5 833 N_("show only the names of files without match")),
caf2de33
NTND
834 OPT_BOOL_F('z', "null", &opt.null_following_name,
835 N_("print NUL after filenames"),
836 PARSE_OPT_NOCOMPLETE),
9d8db06e
TB
837 OPT_BOOL('o', "only-matching", &opt.only_matching,
838 N_("show only matching parts of a line")),
d5d09d47 839 OPT_BOOL('c', "count", &opt.count,
4b407bc5
NTND
840 N_("show the number of matches instead of matching lines")),
841 OPT__COLOR(&opt.color, N_("highlight matches")),
d5d09d47 842 OPT_BOOL(0, "break", &opt.file_break,
4b407bc5 843 N_("print empty line between matches from different files")),
d5d09d47 844 OPT_BOOL(0, "heading", &opt.heading,
4b407bc5 845 N_("show filename only once above matches from same file")),
3e230fa1 846 OPT_GROUP(""),
4b407bc5
NTND
847 OPT_CALLBACK('C', "context", &opt, N_("n"),
848 N_("show <n> context lines before and after matches"),
3e230fa1 849 context_callback),
317f63c2 850 OPT_INTEGER('B', "before-context", &opt.pre_context,
4b407bc5 851 N_("show <n> context lines before matches")),
317f63c2 852 OPT_INTEGER('A', "after-context", &opt.post_context,
4b407bc5 853 N_("show <n> context lines after matches")),
89f09dd3
VL
854 OPT_INTEGER(0, "threads", &num_threads,
855 N_("use <n> worker threads")),
4b407bc5 856 OPT_NUMBER_CALLBACK(&opt, N_("shortcut for -C NUM"),
3e230fa1 857 context_callback),
d5d09d47 858 OPT_BOOL('p', "show-function", &opt.funcname,
4b407bc5 859 N_("show a line with the function name before matches")),
d5d09d47 860 OPT_BOOL('W', "function-context", &opt.funcbody,
4b407bc5 861 N_("show the surrounding function")),
3e230fa1 862 OPT_GROUP(""),
4b407bc5
NTND
863 OPT_CALLBACK('f', NULL, &opt, N_("file"),
864 N_("read patterns from file"), file_callback),
865 { OPTION_CALLBACK, 'e', NULL, &opt, N_("pattern"),
866 N_("match <pattern>"), PARSE_OPT_NONEG, pattern_callback },
3e230fa1 867 { OPTION_CALLBACK, 0, "and", &opt, NULL,
4b407bc5 868 N_("combine patterns specified with -e"),
3e230fa1 869 PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
d5d09d47 870 OPT_BOOL(0, "or", &dummy, ""),
3e230fa1
RS
871 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
872 PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
873 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
874 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
875 open_callback },
876 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
877 PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
878 close_callback },
d52ee6e6 879 OPT__QUIET(&opt.status_only,
4b407bc5 880 N_("indicate hit with exit status without output")),
d5d09d47 881 OPT_BOOL(0, "all-match", &opt.all_match,
4b407bc5 882 N_("show only matches from files that match all patterns")),
3e4a67b4
NTND
883 OPT_SET_INT_F(0, "debug", &opt.debug,
884 N_("show parse tree for grep expression"),
885 1, PARSE_OPT_HIDDEN),
3e230fa1 886 OPT_GROUP(""),
0af88c15 887 { OPTION_STRING, 'O', "open-files-in-pager", &show_in_pager,
4b407bc5 888 N_("pager"), N_("show matching files in the pager"),
caf2de33
NTND
889 PARSE_OPT_OPTARG | PARSE_OPT_NOCOMPLETE,
890 NULL, (intptr_t)default_pager },
891 OPT_BOOL_F(0, "ext-grep", &external_grep_allowed__ignored,
892 N_("allow calling of grep(1) (ignored by this build)"),
893 PARSE_OPT_NOCOMPLETE),
3e230fa1
RS
894 OPT_END()
895 };
5010cb5f 896
38bbc2ea 897 init_grep_defaults(the_repository);
15fabd1b 898 git_config(grep_cmd_config, NULL);
38bbc2ea 899 grep_init(&opt, the_repository, prefix);
7e8f59d5 900
5010cb5f 901 /*
5acd64ed
JH
902 * If there is no -- then the paths must exist in the working
903 * tree. If there is no explicit pattern specified with -e or
904 * -f, we take the first unrecognized non option to be the
905 * pattern, but then what follows it must be zero or more
906 * valid refs up to the -- (if exists), and then existing
907 * paths. If there is an explicit pattern, then the first
82e5a82f 908 * unrecognized non option is the beginning of the refs list
5acd64ed 909 * that continues up to the -- (if exists), and then paths.
5010cb5f 910 */
37782920 911 argc = parse_options(argc, argv, prefix, options, grep_usage,
3e230fa1 912 PARSE_OPT_KEEP_DASHDASH |
44415499 913 PARSE_OPT_STOP_AT_NON_OPTION);
c5c31d33 914 grep_commit_pattern_type(pattern_type_arg, &opt);
3e230fa1 915
ecd9ba61
TG
916 if (use_index && !startup_info->have_repository) {
917 int fallback = 0;
918 git_config_get_bool("grep.fallbacktonoindex", &fallback);
919 if (fallback)
920 use_index = 0;
921 else
922 /* die the same way as if we did it at the beginning */
923 setup_git_directory();
924 }
59332d13 925
1123c67c
JK
926 /*
927 * skip a -- separator; we know it cannot be
928 * separating revisions from pathnames if
929 * we haven't even had any patterns yet
930 */
931 if (argc > 0 && !opt.pattern_list && !strcmp(argv[0], "--")) {
932 argv++;
933 argc--;
934 }
935
3e230fa1
RS
936 /* First unrecognized non-option token */
937 if (argc > 0 && !opt.pattern_list) {
938 append_grep_pattern(&opt, argv[0], "command line", 0,
939 GREP_PATTERN);
940 argv++;
941 argc--;
5010cb5f 942 }
5acd64ed 943
0af88c15
JS
944 if (show_in_pager == default_pager)
945 show_in_pager = git_pager(1);
678e484b 946 if (show_in_pager) {
e7b082a4 947 opt.color = 0;
0af88c15
JS
948 opt.name_only = 1;
949 opt.null_following_name = 1;
950 opt.output_priv = &path_list;
951 opt.output = append_path;
0c72cead 952 string_list_append(&path_list, show_in_pager);
678e484b
JS
953 }
954
f9b9faf6 955 if (!opt.pattern_list)
1a07e59c 956 die(_("no pattern given"));
5b594f45 957
9d8db06e
TB
958 /* --only-matching has no effect with --invert. */
959 if (opt.invert)
960 opt.only_matching = 0;
961
b5b81136
JK
962 /*
963 * We have to find "--" in a separate pass, because its presence
964 * influences how we will parse arguments that come before it.
965 */
966 for (i = 0; i < argc; i++) {
967 if (!strcmp(argv[i], "--")) {
968 seen_dashdash = 1;
969 break;
970 }
971 }
972
973 /*
974 * Resolve any rev arguments. If we have a dashdash, then everything up
975 * to it must resolve as a rev. If not, then we stop at the first
976 * non-rev and assume everything else is a path.
977 */
131f3c96 978 allow_revs = use_index && !untracked;
3e230fa1 979 for (i = 0; i < argc; i++) {
5acd64ed 980 const char *arg = argv[i];
1db11086 981 struct object_id oid;
afa15f3c 982 struct object_context oc;
20d6421c
JK
983 struct object *object;
984
dca3b5f5
JT
985 if (!strcmp(arg, "--")) {
986 i++;
dca3b5f5
JT
987 break;
988 }
20d6421c 989
131f3c96 990 if (!allow_revs) {
d0ffc069 991 if (seen_dashdash)
131f3c96 992 die(_("--no-index or --untracked cannot be used with revs"));
d0ffc069
JK
993 break;
994 }
995
321c89bf 996 if (get_oid_with_context(arg, GET_OID_RECORD_PATH,
e82caf38 997 &oid, &oc)) {
b5b81136
JK
998 if (seen_dashdash)
999 die(_("unable to resolve revision: %s"), arg);
20d6421c 1000 break;
b5b81136 1001 }
20d6421c 1002
c251c83d 1003 object = parse_object_or_die(&oid, arg);
20d6421c
JK
1004 if (!seen_dashdash)
1005 verify_non_filename(prefix, arg);
1006 add_object_array_with_path(object, arg, &list, oc.mode, oc.path);
dc944b65 1007 free(oc.path);
1362671f 1008 }
5acd64ed 1009
b5b81136
JK
1010 /*
1011 * Anything left over is presumed to be a path. But in the non-dashdash
1012 * "do what I mean" case, we verify and complain when that isn't true.
1013 */
a0fe2b0d
JK
1014 if (!seen_dashdash) {
1015 int j;
1016 for (j = i; j < argc; j++)
131f3c96 1017 verify_filename(prefix, argv[j], j == i && allow_revs);
a0fe2b0d
JK
1018 }
1019
1020 parse_pathspec(&pathspec, 0,
1021 PATHSPEC_PREFER_CWD |
1022 (opt.max_depth != -1 ? PATHSPEC_MAXDEPTH_VALID : 0),
1023 prefix, argv + i);
1024 pathspec.max_depth = opt.max_depth;
1025 pathspec.recursive = 1;
eef3df5a 1026 pathspec.recurse_submodules = !!recurse_submodules;
a0fe2b0d 1027
4002e87c
NTND
1028 if (HAVE_THREADS) {
1029 if (list.nr || cached || show_in_pager)
1030 num_threads = 0;
1031 else if (num_threads == 0)
1032 num_threads = GREP_NUM_THREADS_DEFAULT;
1033 else if (num_threads < 0)
1034 die(_("invalid number of threads specified (%d)"), num_threads);
1035 if (num_threads == 1)
1036 num_threads = 0;
1037 } else {
1038 if (num_threads)
1039 warning(_("no threads support, ignoring --threads"));
9ec726a4 1040 num_threads = 0;
4002e87c 1041 }
53b8d931 1042
6d423dd5
ÆAB
1043 if (!num_threads)
1044 /*
1045 * The compiled patterns on the main path are only
1046 * used when not using threading. Otherwise
1047 * start_threads() below calls compile_grep_patterns()
1048 * for each thread.
1049 */
1050 compile_grep_patterns(&opt);
1051
4002e87c 1052 if (HAVE_THREADS && num_threads) {
50dd0f2f
AY
1053 if (!(opt.name_only || opt.unmatch_name_only || opt.count)
1054 && (opt.pre_context || opt.post_context ||
1055 opt.file_break || opt.funcbody))
53b8d931
TR
1056 skip_first_line = 1;
1057 start_threads(&opt);
1058 }
53b8d931 1059
678e484b 1060 if (show_in_pager && (cached || list.nr))
e4fe4ba5 1061 die(_("--open-files-in-pager only works on the worktree"));
678e484b
JS
1062
1063 if (show_in_pager && opt.pattern_list && !opt.pattern_list->next) {
1064 const char *pager = path_list.items[0].string;
1065 int len = strlen(pager);
1066
1067 if (len > 4 && is_dir_sep(pager[len - 5]))
1068 pager += len - 4;
1069
f7febbea
JS
1070 if (opt.ignore_case && !strcmp("less", pager))
1071 string_list_append(&path_list, "-I");
1072
678e484b
JS
1073 if (!strcmp("less", pager) || !strcmp("vi", pager)) {
1074 struct strbuf buf = STRBUF_INIT;
1075 strbuf_addf(&buf, "+/%s%s",
1076 strcmp("less", pager) ? "" : "*",
1077 opt.pattern_list->pattern);
0c72cead 1078 string_list_append(&path_list, buf.buf);
678e484b
JS
1079 strbuf_detach(&buf, NULL);
1080 }
1081 }
1082
74ed4371 1083 if (recurse_submodules && (!use_index || untracked))
1a07e59c 1084 die(_("option not supported with --recurse-submodules"));
0281e487 1085
c2048f0b 1086 if (!show_in_pager && !opt.status_only)
678e484b
JS
1087 setup_pager();
1088
0a93fb8a 1089 if (!use_index && (untracked || cached))
1a07e59c 1090 die(_("--cached or --untracked cannot be used with --no-index"));
678e484b 1091
0a93fb8a 1092 if (!use_index || untracked) {
0a93fb8a 1093 int use_exclude = (opt_exclude < 0) ? use_index : !!opt_exclude;
85975c0c 1094 hit = grep_directory(&opt, &pathspec, use_exclude, use_index);
dbfae86a 1095 } else if (0 <= opt_exclude) {
1a07e59c 1096 die(_("--[no-]exclude-standard cannot be used for tracked contents"));
685359cf 1097 } else if (!list.nr) {
6577f542
NTND
1098 if (!cached)
1099 setup_work_tree();
5b594f45 1100
f9ee2fcd 1101 hit = grep_cache(&opt, the_repository, &pathspec, cached);
685359cf
JS
1102 } else {
1103 if (cached)
1a07e59c 1104 die(_("both --cached and trees are given"));
f9ee2fcd 1105
6856077a 1106 hit = grep_objects(&opt, &pathspec, &list);
5010cb5f 1107 }
5b594f45 1108
89f09dd3 1109 if (num_threads)
5b594f45 1110 hit |= wait_all();
678e484b
JS
1111 if (hit && show_in_pager)
1112 run_pager(&opt, prefix);
7861fa07 1113 clear_pathspec(&pathspec);
b48fb5b6 1114 free_grep_patterns(&opt);
5010cb5f
JH
1115 return !hit;
1116}