builtin / grep.con commit odb_load_loose_cache: fix strbuf leak (7317aa7)
   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
  27static char const * const grep_usage[] = {
  28        N_("git grep [<options>] [-e] <pattern> [<rev>...] [[--] <path>...]"),
  29        NULL
  30};
  31
  32static int recurse_submodules;
  33
  34#define GREP_NUM_THREADS_DEFAULT 8
  35static int num_threads;
  36
  37#ifndef NO_PTHREADS
  38static pthread_t *threads;
  39
  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 */
  44struct work_item {
  45        struct grep_source source;
  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
  71static inline void grep_lock(void)
  72{
  73        assert(num_threads);
  74        pthread_mutex_lock(&grep_mutex);
  75}
  76
  77static inline void grep_unlock(void)
  78{
  79        assert(num_threads);
  80        pthread_mutex_unlock(&grep_mutex);
  81}
  82
  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
  94static int skip_first_line;
  95
  96static void add_work(struct grep_opt *opt, const struct grep_source *gs)
  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
 104        todo[todo_end].source = *gs;
 105        if (opt->binary != GREP_BINARY_TEXT)
 106                grep_source_load_driver(&todo[todo_end].source,
 107                                        opt->repo->index);
 108        todo[todo_end].done = 0;
 109        strbuf_reset(&todo[todo_end].out);
 110        todo_end = (todo_end + 1) % ARRAY_SIZE(todo);
 111
 112        pthread_cond_signal(&cond_add);
 113        grep_unlock();
 114}
 115
 116static struct work_item *get_work(void)
 117{
 118        struct work_item *ret;
 119
 120        grep_lock();
 121        while (todo_start == todo_end && !all_work_added) {
 122                pthread_cond_wait(&cond_add, &grep_mutex);
 123        }
 124
 125        if (todo_start == todo_end && all_work_added) {
 126                ret = NULL;
 127        } else {
 128                ret = &todo[todo_start];
 129                todo_start = (todo_start + 1) % ARRAY_SIZE(todo);
 130        }
 131        grep_unlock();
 132        return ret;
 133}
 134
 135static void work_done(struct work_item *w)
 136{
 137        int old_done;
 138
 139        grep_lock();
 140        w->done = 1;
 141        old_done = todo_done;
 142        for(; todo[todo_done].done && todo_done != todo_start;
 143            todo_done = (todo_done+1) % ARRAY_SIZE(todo)) {
 144                w = &todo[todo_done];
 145                if (w->out.len) {
 146                        const char *p = w->out.buf;
 147                        size_t len = w->out.len;
 148
 149                        /* Skip the leading hunk mark of the first file. */
 150                        if (skip_first_line) {
 151                                while (len) {
 152                                        len--;
 153                                        if (*p++ == '\n')
 154                                                break;
 155                                }
 156                                skip_first_line = 0;
 157                        }
 158
 159                        write_or_die(1, p, len);
 160                }
 161                grep_source_clear(&w->source);
 162        }
 163
 164        if (old_done != todo_done)
 165                pthread_cond_signal(&cond_write);
 166
 167        if (all_work_added && todo_done == todo_end)
 168                pthread_cond_signal(&cond_result);
 169
 170        grep_unlock();
 171}
 172
 173static void *run(void *arg)
 174{
 175        int hit = 0;
 176        struct grep_opt *opt = arg;
 177
 178        while (1) {
 179                struct work_item *w = get_work();
 180                if (!w)
 181                        break;
 182
 183                opt->output_priv = w;
 184                hit |= grep_source(opt, &w->source);
 185                grep_source_clear_data(&w->source);
 186                work_done(w);
 187        }
 188        free_grep_patterns(arg);
 189        free(arg);
 190
 191        return (void*) (intptr_t) hit;
 192}
 193
 194static void strbuf_out(struct grep_opt *opt, const void *buf, size_t size)
 195{
 196        struct work_item *w = opt->output_priv;
 197        strbuf_add(&w->out, buf, size);
 198}
 199
 200static void start_threads(struct grep_opt *opt)
 201{
 202        int i;
 203
 204        pthread_mutex_init(&grep_mutex, NULL);
 205        pthread_mutex_init(&grep_read_mutex, NULL);
 206        pthread_mutex_init(&grep_attr_mutex, NULL);
 207        pthread_cond_init(&cond_add, NULL);
 208        pthread_cond_init(&cond_write, NULL);
 209        pthread_cond_init(&cond_result, NULL);
 210        grep_use_locks = 1;
 211
 212        for (i = 0; i < ARRAY_SIZE(todo); i++) {
 213                strbuf_init(&todo[i].out, 0);
 214        }
 215
 216        threads = xcalloc(num_threads, sizeof(*threads));
 217        for (i = 0; i < num_threads; i++) {
 218                int err;
 219                struct grep_opt *o = grep_opt_dup(opt);
 220                o->output = strbuf_out;
 221                if (i)
 222                        o->debug = 0;
 223                compile_grep_patterns(o);
 224                err = pthread_create(&threads[i], NULL, run, o);
 225
 226                if (err)
 227                        die(_("grep: failed to create thread: %s"),
 228                            strerror(err));
 229        }
 230}
 231
 232static int wait_all(void)
 233{
 234        int hit = 0;
 235        int i;
 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#else /* !NO_PTHREADS */
 269
 270static int wait_all(void)
 271{
 272        return 0;
 273}
 274#endif
 275
 276static int grep_cmd_config(const char *var, const char *value, void *cb)
 277{
 278        int st = grep_config(var, value, cb);
 279        if (git_color_default_config(var, value, cb) < 0)
 280                st = -1;
 281
 282        if (!strcmp(var, "grep.threads")) {
 283                num_threads = git_config_int(var, value);
 284                if (num_threads < 0)
 285                        die(_("invalid number of threads specified (%d) for %s"),
 286                            num_threads, var);
 287#ifdef NO_PTHREADS
 288                else if (num_threads && num_threads != 1) {
 289                        /*
 290                         * TRANSLATORS: %s is the configuration
 291                         * variable for tweaking threads, currently
 292                         * grep.threads
 293                         */
 294                        warning(_("no threads support, ignoring %s"), var);
 295                        num_threads = 0;
 296                }
 297#endif
 298        }
 299
 300        if (!strcmp(var, "submodule.recurse"))
 301                recurse_submodules = git_config_bool(var, value);
 302
 303        return st;
 304}
 305
 306static void *lock_and_read_oid_file(const struct object_id *oid, enum object_type *type, unsigned long *size)
 307{
 308        void *data;
 309
 310        grep_read_lock();
 311        data = read_object_file(oid, type, size);
 312        grep_read_unlock();
 313        return data;
 314}
 315
 316static int grep_oid(struct grep_opt *opt, const struct object_id *oid,
 317                     const char *filename, int tree_name_len,
 318                     const char *path)
 319{
 320        struct strbuf pathbuf = STRBUF_INIT;
 321        struct grep_source gs;
 322
 323        if (opt->relative && opt->prefix_length) {
 324                quote_path_relative(filename + tree_name_len, opt->prefix, &pathbuf);
 325                strbuf_insert(&pathbuf, 0, filename, tree_name_len);
 326        } else {
 327                strbuf_addstr(&pathbuf, filename);
 328        }
 329
 330        grep_source_init(&gs, GREP_SOURCE_OID, pathbuf.buf, path, oid);
 331        strbuf_release(&pathbuf);
 332
 333#ifndef NO_PTHREADS
 334        if (num_threads) {
 335                /*
 336                 * add_work() copies gs and thus assumes ownership of
 337                 * its fields, so do not call grep_source_clear()
 338                 */
 339                add_work(opt, &gs);
 340                return 0;
 341        } else
 342#endif
 343        {
 344                int hit;
 345
 346                hit = grep_source(opt, &gs);
 347
 348                grep_source_clear(&gs);
 349                return hit;
 350        }
 351}
 352
 353static int grep_file(struct grep_opt *opt, const char *filename)
 354{
 355        struct strbuf buf = STRBUF_INIT;
 356        struct grep_source gs;
 357
 358        if (opt->relative && opt->prefix_length)
 359                quote_path_relative(filename, opt->prefix, &buf);
 360        else
 361                strbuf_addstr(&buf, filename);
 362
 363        grep_source_init(&gs, GREP_SOURCE_FILE, buf.buf, filename, filename);
 364        strbuf_release(&buf);
 365
 366#ifndef NO_PTHREADS
 367        if (num_threads) {
 368                /*
 369                 * add_work() copies gs and thus assumes ownership of
 370                 * its fields, so do not call grep_source_clear()
 371                 */
 372                add_work(opt, &gs);
 373                return 0;
 374        } else
 375#endif
 376        {
 377                int hit;
 378
 379                hit = grep_source(opt, &gs);
 380
 381                grep_source_clear(&gs);
 382                return hit;
 383        }
 384}
 385
 386static void append_path(struct grep_opt *opt, const void *data, size_t len)
 387{
 388        struct string_list *path_list = opt->output_priv;
 389
 390        if (len == 1 && *(const char *)data == '\0')
 391                return;
 392        string_list_append(path_list, xstrndup(data, len));
 393}
 394
 395static void run_pager(struct grep_opt *opt, const char *prefix)
 396{
 397        struct string_list *path_list = opt->output_priv;
 398        struct child_process child = CHILD_PROCESS_INIT;
 399        int i, status;
 400
 401        for (i = 0; i < path_list->nr; i++)
 402                argv_array_push(&child.args, path_list->items[i].string);
 403        child.dir = prefix;
 404        child.use_shell = 1;
 405
 406        status = run_command(&child);
 407        if (status)
 408                exit(status);
 409}
 410
 411static int grep_cache(struct grep_opt *opt, struct repository *repo,
 412                      const struct pathspec *pathspec, int cached);
 413static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
 414                     struct tree_desc *tree, struct strbuf *base, int tn_len,
 415                     int check_attr, struct repository *repo);
 416
 417static int grep_submodule(struct grep_opt *opt, struct repository *superproject,
 418                          const struct pathspec *pathspec,
 419                          const struct object_id *oid,
 420                          const char *filename, const char *path)
 421{
 422        struct repository submodule;
 423        int hit;
 424
 425        if (!is_submodule_active(superproject, path))
 426                return 0;
 427
 428        if (repo_submodule_init(&submodule, superproject, path))
 429                return 0;
 430
 431        repo_read_gitmodules(&submodule);
 432
 433        /*
 434         * NEEDSWORK: This adds the submodule's object directory to the list of
 435         * alternates for the single in-memory object store.  This has some bad
 436         * consequences for memory (processed objects will never be freed) and
 437         * performance (this increases the number of pack files git has to pay
 438         * attention to, to the sum of the number of pack files in all the
 439         * repositories processed so far).  This can be removed once the object
 440         * store is no longer global and instead is a member of the repository
 441         * object.
 442         */
 443        grep_read_lock();
 444        add_to_alternates_memory(submodule.objects->odb->path);
 445        grep_read_unlock();
 446
 447        if (oid) {
 448                struct object *object;
 449                struct tree_desc tree;
 450                void *data;
 451                unsigned long size;
 452                struct strbuf base = STRBUF_INIT;
 453
 454                object = parse_object_or_die(oid, oid_to_hex(oid));
 455
 456                grep_read_lock();
 457                data = read_object_with_reference(&object->oid, tree_type,
 458                                                  &size, NULL);
 459                grep_read_unlock();
 460
 461                if (!data)
 462                        die(_("unable to read tree (%s)"), oid_to_hex(&object->oid));
 463
 464                strbuf_addstr(&base, filename);
 465                strbuf_addch(&base, '/');
 466
 467                init_tree_desc(&tree, data, size);
 468                hit = grep_tree(opt, pathspec, &tree, &base, base.len,
 469                                object->type == OBJ_COMMIT, &submodule);
 470                strbuf_release(&base);
 471                free(data);
 472        } else {
 473                hit = grep_cache(opt, &submodule, pathspec, 1);
 474        }
 475
 476        repo_clear(&submodule);
 477        return hit;
 478}
 479
 480static int grep_cache(struct grep_opt *opt, struct repository *repo,
 481                      const struct pathspec *pathspec, int cached)
 482{
 483        int hit = 0;
 484        int nr;
 485        struct strbuf name = STRBUF_INIT;
 486        int name_base_len = 0;
 487        if (repo->submodule_prefix) {
 488                name_base_len = strlen(repo->submodule_prefix);
 489                strbuf_addstr(&name, repo->submodule_prefix);
 490        }
 491
 492        if (repo_read_index(repo) < 0)
 493                die(_("index file corrupt"));
 494
 495        for (nr = 0; nr < repo->index->cache_nr; nr++) {
 496                const struct cache_entry *ce = repo->index->cache[nr];
 497                strbuf_setlen(&name, name_base_len);
 498                strbuf_addstr(&name, ce->name);
 499
 500                if (S_ISREG(ce->ce_mode) &&
 501                    match_pathspec(repo->index, pathspec, name.buf, name.len, 0, NULL,
 502                                   S_ISDIR(ce->ce_mode) ||
 503                                   S_ISGITLINK(ce->ce_mode))) {
 504                        /*
 505                         * If CE_VALID is on, we assume worktree file and its
 506                         * cache entry are identical, even if worktree file has
 507                         * been modified, so use cache version instead
 508                         */
 509                        if (cached || (ce->ce_flags & CE_VALID) ||
 510                            ce_skip_worktree(ce)) {
 511                                if (ce_stage(ce) || ce_intent_to_add(ce))
 512                                        continue;
 513                                hit |= grep_oid(opt, &ce->oid, name.buf,
 514                                                 0, name.buf);
 515                        } else {
 516                                hit |= grep_file(opt, name.buf);
 517                        }
 518                } else if (recurse_submodules && S_ISGITLINK(ce->ce_mode) &&
 519                           submodule_path_match(repo->index, pathspec, name.buf, NULL)) {
 520                        hit |= grep_submodule(opt, repo, pathspec, NULL, ce->name, ce->name);
 521                } else {
 522                        continue;
 523                }
 524
 525                if (ce_stage(ce)) {
 526                        do {
 527                                nr++;
 528                        } while (nr < repo->index->cache_nr &&
 529                                 !strcmp(ce->name, repo->index->cache[nr]->name));
 530                        nr--; /* compensate for loop control */
 531                }
 532                if (hit && opt->status_only)
 533                        break;
 534        }
 535
 536        strbuf_release(&name);
 537        return hit;
 538}
 539
 540static int grep_tree(struct grep_opt *opt, const struct pathspec *pathspec,
 541                     struct tree_desc *tree, struct strbuf *base, int tn_len,
 542                     int check_attr, struct repository *repo)
 543{
 544        int hit = 0;
 545        enum interesting match = entry_not_interesting;
 546        struct name_entry entry;
 547        int old_baselen = base->len;
 548        struct strbuf name = STRBUF_INIT;
 549        int name_base_len = 0;
 550        if (repo->submodule_prefix) {
 551                strbuf_addstr(&name, repo->submodule_prefix);
 552                name_base_len = name.len;
 553        }
 554
 555        while (tree_entry(tree, &entry)) {
 556                int te_len = tree_entry_len(&entry);
 557
 558                if (match != all_entries_interesting) {
 559                        strbuf_addstr(&name, base->buf + tn_len);
 560                        match = tree_entry_interesting(&entry, &name,
 561                                                       0, pathspec);
 562                        strbuf_setlen(&name, name_base_len);
 563
 564                        if (match == all_entries_not_interesting)
 565                                break;
 566                        if (match == entry_not_interesting)
 567                                continue;
 568                }
 569
 570                strbuf_add(base, entry.path, te_len);
 571
 572                if (S_ISREG(entry.mode)) {
 573                        hit |= grep_oid(opt, entry.oid, base->buf, tn_len,
 574                                         check_attr ? base->buf + tn_len : NULL);
 575                } else if (S_ISDIR(entry.mode)) {
 576                        enum object_type type;
 577                        struct tree_desc sub;
 578                        void *data;
 579                        unsigned long size;
 580
 581                        data = lock_and_read_oid_file(entry.oid, &type, &size);
 582                        if (!data)
 583                                die(_("unable to read tree (%s)"),
 584                                    oid_to_hex(entry.oid));
 585
 586                        strbuf_addch(base, '/');
 587                        init_tree_desc(&sub, data, size);
 588                        hit |= grep_tree(opt, pathspec, &sub, base, tn_len,
 589                                         check_attr, repo);
 590                        free(data);
 591                } else if (recurse_submodules && S_ISGITLINK(entry.mode)) {
 592                        hit |= grep_submodule(opt, repo, pathspec, entry.oid,
 593                                              base->buf, base->buf + tn_len);
 594                }
 595
 596                strbuf_setlen(base, old_baselen);
 597
 598                if (hit && opt->status_only)
 599                        break;
 600        }
 601
 602        strbuf_release(&name);
 603        return hit;
 604}
 605
 606static int grep_object(struct grep_opt *opt, const struct pathspec *pathspec,
 607                       struct object *obj, const char *name, const char *path)
 608{
 609        if (obj->type == OBJ_BLOB)
 610                return grep_oid(opt, &obj->oid, name, 0, path);
 611        if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
 612                struct tree_desc tree;
 613                void *data;
 614                unsigned long size;
 615                struct strbuf base;
 616                int hit, len;
 617
 618                grep_read_lock();
 619                data = read_object_with_reference(&obj->oid, tree_type,
 620                                                  &size, NULL);
 621                grep_read_unlock();
 622
 623                if (!data)
 624                        die(_("unable to read tree (%s)"), oid_to_hex(&obj->oid));
 625
 626                len = name ? strlen(name) : 0;
 627                strbuf_init(&base, PATH_MAX + len + 1);
 628                if (len) {
 629                        strbuf_add(&base, name, len);
 630                        strbuf_addch(&base, ':');
 631                }
 632                init_tree_desc(&tree, data, size);
 633                hit = grep_tree(opt, pathspec, &tree, &base, base.len,
 634                                obj->type == OBJ_COMMIT, the_repository);
 635                strbuf_release(&base);
 636                free(data);
 637                return hit;
 638        }
 639        die(_("unable to grep from object of type %s"), type_name(obj->type));
 640}
 641
 642static int grep_objects(struct grep_opt *opt, const struct pathspec *pathspec,
 643                        const struct object_array *list)
 644{
 645        unsigned int i;
 646        int hit = 0;
 647        const unsigned int nr = list->nr;
 648
 649        for (i = 0; i < nr; i++) {
 650                struct object *real_obj;
 651                real_obj = deref_tag(the_repository, list->objects[i].item,
 652                                     NULL, 0);
 653
 654                /* load the gitmodules file for this rev */
 655                if (recurse_submodules) {
 656                        submodule_free(the_repository);
 657                        gitmodules_config_oid(&real_obj->oid);
 658                }
 659                if (grep_object(opt, pathspec, real_obj, list->objects[i].name,
 660                                list->objects[i].path)) {
 661                        hit = 1;
 662                        if (opt->status_only)
 663                                break;
 664                }
 665        }
 666        return hit;
 667}
 668
 669static int grep_directory(struct grep_opt *opt, const struct pathspec *pathspec,
 670                          int exc_std, int use_index)
 671{
 672        struct dir_struct dir;
 673        int i, hit = 0;
 674
 675        memset(&dir, 0, sizeof(dir));
 676        if (!use_index)
 677                dir.flags |= DIR_NO_GITLINKS;
 678        if (exc_std)
 679                setup_standard_excludes(&dir);
 680
 681        fill_directory(&dir, &the_index, pathspec);
 682        for (i = 0; i < dir.nr; i++) {
 683                if (!dir_path_match(&the_index, dir.entries[i], pathspec, 0, NULL))
 684                        continue;
 685                hit |= grep_file(opt, dir.entries[i]->name);
 686                if (hit && opt->status_only)
 687                        break;
 688        }
 689        return hit;
 690}
 691
 692static int context_callback(const struct option *opt, const char *arg,
 693                            int unset)
 694{
 695        struct grep_opt *grep_opt = opt->value;
 696        int value;
 697        const char *endp;
 698
 699        if (unset) {
 700                grep_opt->pre_context = grep_opt->post_context = 0;
 701                return 0;
 702        }
 703        value = strtol(arg, (char **)&endp, 10);
 704        if (*endp) {
 705                return error(_("switch `%c' expects a numerical value"),
 706                             opt->short_name);
 707        }
 708        grep_opt->pre_context = grep_opt->post_context = value;
 709        return 0;
 710}
 711
 712static int file_callback(const struct option *opt, const char *arg, int unset)
 713{
 714        struct grep_opt *grep_opt = opt->value;
 715        int from_stdin = !strcmp(arg, "-");
 716        FILE *patterns;
 717        int lno = 0;
 718        struct strbuf sb = STRBUF_INIT;
 719
 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
 737static int not_callback(const struct option *opt, const char *arg, int unset)
 738{
 739        struct grep_opt *grep_opt = opt->value;
 740        append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
 741        return 0;
 742}
 743
 744static int and_callback(const struct option *opt, const char *arg, int unset)
 745{
 746        struct grep_opt *grep_opt = opt->value;
 747        append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
 748        return 0;
 749}
 750
 751static int open_callback(const struct option *opt, const char *arg, int unset)
 752{
 753        struct grep_opt *grep_opt = opt->value;
 754        append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
 755        return 0;
 756}
 757
 758static int close_callback(const struct option *opt, const char *arg, int unset)
 759{
 760        struct grep_opt *grep_opt = opt->value;
 761        append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
 762        return 0;
 763}
 764
 765static int pattern_callback(const struct option *opt, const char *arg,
 766                            int unset)
 767{
 768        struct grep_opt *grep_opt = opt->value;
 769        append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
 770        return 0;
 771}
 772
 773int cmd_grep(int argc, const char **argv, const char *prefix)
 774{
 775        int hit = 0;
 776        int cached = 0, untracked = 0, opt_exclude = -1;
 777        int seen_dashdash = 0;
 778        int external_grep_allowed__ignored;
 779        const char *show_in_pager = NULL, *default_pager = "dummy";
 780        struct grep_opt opt;
 781        struct object_array list = OBJECT_ARRAY_INIT;
 782        struct pathspec pathspec;
 783        struct string_list path_list = STRING_LIST_INIT_NODUP;
 784        int i;
 785        int dummy;
 786        int use_index = 1;
 787        int pattern_type_arg = GREP_PATTERN_TYPE_UNSPECIFIED;
 788        int allow_revs;
 789
 790        struct option options[] = {
 791                OPT_BOOL(0, "cached", &cached,
 792                        N_("search in index instead of in the work tree")),
 793                OPT_NEGBIT(0, "no-index", &use_index,
 794                         N_("find in contents not managed by git"), 1),
 795                OPT_BOOL(0, "untracked", &untracked,
 796                        N_("search in both tracked and untracked files")),
 797                OPT_SET_INT(0, "exclude-standard", &opt_exclude,
 798                            N_("ignore files specified via '.gitignore'"), 1),
 799                OPT_BOOL(0, "recurse-submodules", &recurse_submodules,
 800                         N_("recursively search in each submodule")),
 801                OPT_GROUP(""),
 802                OPT_BOOL('v', "invert-match", &opt.invert,
 803                        N_("show non-matching lines")),
 804                OPT_BOOL('i', "ignore-case", &opt.ignore_case,
 805                        N_("case insensitive matching")),
 806                OPT_BOOL('w', "word-regexp", &opt.word_regexp,
 807                        N_("match patterns only at word boundaries")),
 808                OPT_SET_INT('a', "text", &opt.binary,
 809                        N_("process binary files as text"), GREP_BINARY_TEXT),
 810                OPT_SET_INT('I', NULL, &opt.binary,
 811                        N_("don't match patterns in binary files"),
 812                        GREP_BINARY_NOMATCH),
 813                OPT_BOOL(0, "textconv", &opt.allow_textconv,
 814                         N_("process binary files with textconv filters")),
 815                OPT_SET_INT('r', "recursive", &opt.max_depth,
 816                            N_("search in subdirectories (default)"), -1),
 817                { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, N_("depth"),
 818                        N_("descend at most <depth> levels"), PARSE_OPT_NONEG,
 819                        NULL, 1 },
 820                OPT_GROUP(""),
 821                OPT_SET_INT('E', "extended-regexp", &pattern_type_arg,
 822                            N_("use extended POSIX regular expressions"),
 823                            GREP_PATTERN_TYPE_ERE),
 824                OPT_SET_INT('G', "basic-regexp", &pattern_type_arg,
 825                            N_("use basic POSIX regular expressions (default)"),
 826                            GREP_PATTERN_TYPE_BRE),
 827                OPT_SET_INT('F', "fixed-strings", &pattern_type_arg,
 828                            N_("interpret patterns as fixed strings"),
 829                            GREP_PATTERN_TYPE_FIXED),
 830                OPT_SET_INT('P', "perl-regexp", &pattern_type_arg,
 831                            N_("use Perl-compatible regular expressions"),
 832                            GREP_PATTERN_TYPE_PCRE),
 833                OPT_GROUP(""),
 834                OPT_BOOL('n', "line-number", &opt.linenum, N_("show line numbers")),
 835                OPT_BOOL(0, "column", &opt.columnnum, N_("show column number of first match")),
 836                OPT_NEGBIT('h', NULL, &opt.pathname, N_("don't show filenames"), 1),
 837                OPT_BIT('H', NULL, &opt.pathname, N_("show filenames"), 1),
 838                OPT_NEGBIT(0, "full-name", &opt.relative,
 839                        N_("show filenames relative to top directory"), 1),
 840                OPT_BOOL('l', "files-with-matches", &opt.name_only,
 841                        N_("show only filenames instead of matching lines")),
 842                OPT_BOOL(0, "name-only", &opt.name_only,
 843                        N_("synonym for --files-with-matches")),
 844                OPT_BOOL('L', "files-without-match",
 845                        &opt.unmatch_name_only,
 846                        N_("show only the names of files without match")),
 847                OPT_BOOL_F('z', "null", &opt.null_following_name,
 848                           N_("print NUL after filenames"),
 849                           PARSE_OPT_NOCOMPLETE),
 850                OPT_BOOL('o', "only-matching", &opt.only_matching,
 851                        N_("show only matching parts of a line")),
 852                OPT_BOOL('c', "count", &opt.count,
 853                        N_("show the number of matches instead of matching lines")),
 854                OPT__COLOR(&opt.color, N_("highlight matches")),
 855                OPT_BOOL(0, "break", &opt.file_break,
 856                        N_("print empty line between matches from different files")),
 857                OPT_BOOL(0, "heading", &opt.heading,
 858                        N_("show filename only once above matches from same file")),
 859                OPT_GROUP(""),
 860                OPT_CALLBACK('C', "context", &opt, N_("n"),
 861                        N_("show <n> context lines before and after matches"),
 862                        context_callback),
 863                OPT_INTEGER('B', "before-context", &opt.pre_context,
 864                        N_("show <n> context lines before matches")),
 865                OPT_INTEGER('A', "after-context", &opt.post_context,
 866                        N_("show <n> context lines after matches")),
 867                OPT_INTEGER(0, "threads", &num_threads,
 868                        N_("use <n> worker threads")),
 869                OPT_NUMBER_CALLBACK(&opt, N_("shortcut for -C NUM"),
 870                        context_callback),
 871                OPT_BOOL('p', "show-function", &opt.funcname,
 872                        N_("show a line with the function name before matches")),
 873                OPT_BOOL('W', "function-context", &opt.funcbody,
 874                        N_("show the surrounding function")),
 875                OPT_GROUP(""),
 876                OPT_CALLBACK('f', NULL, &opt, N_("file"),
 877                        N_("read patterns from file"), file_callback),
 878                { OPTION_CALLBACK, 'e', NULL, &opt, N_("pattern"),
 879                        N_("match <pattern>"), PARSE_OPT_NONEG, pattern_callback },
 880                { OPTION_CALLBACK, 0, "and", &opt, NULL,
 881                  N_("combine patterns specified with -e"),
 882                  PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
 883                OPT_BOOL(0, "or", &dummy, ""),
 884                { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
 885                  PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
 886                { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
 887                  PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
 888                  open_callback },
 889                { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
 890                  PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
 891                  close_callback },
 892                OPT__QUIET(&opt.status_only,
 893                           N_("indicate hit with exit status without output")),
 894                OPT_BOOL(0, "all-match", &opt.all_match,
 895                        N_("show only matches from files that match all patterns")),
 896                OPT_SET_INT_F(0, "debug", &opt.debug,
 897                              N_("show parse tree for grep expression"),
 898                              1, PARSE_OPT_HIDDEN),
 899                OPT_GROUP(""),
 900                { OPTION_STRING, 'O', "open-files-in-pager", &show_in_pager,
 901                        N_("pager"), N_("show matching files in the pager"),
 902                        PARSE_OPT_OPTARG | PARSE_OPT_NOCOMPLETE,
 903                        NULL, (intptr_t)default_pager },
 904                OPT_BOOL_F(0, "ext-grep", &external_grep_allowed__ignored,
 905                           N_("allow calling of grep(1) (ignored by this build)"),
 906                           PARSE_OPT_NOCOMPLETE),
 907                OPT_END()
 908        };
 909
 910        init_grep_defaults(the_repository);
 911        git_config(grep_cmd_config, NULL);
 912        grep_init(&opt, the_repository, prefix);
 913
 914        /*
 915         * If there is no -- then the paths must exist in the working
 916         * tree.  If there is no explicit pattern specified with -e or
 917         * -f, we take the first unrecognized non option to be the
 918         * pattern, but then what follows it must be zero or more
 919         * valid refs up to the -- (if exists), and then existing
 920         * paths.  If there is an explicit pattern, then the first
 921         * unrecognized non option is the beginning of the refs list
 922         * that continues up to the -- (if exists), and then paths.
 923         */
 924        argc = parse_options(argc, argv, prefix, options, grep_usage,
 925                             PARSE_OPT_KEEP_DASHDASH |
 926                             PARSE_OPT_STOP_AT_NON_OPTION);
 927        grep_commit_pattern_type(pattern_type_arg, &opt);
 928
 929        if (use_index && !startup_info->have_repository) {
 930                int fallback = 0;
 931                git_config_get_bool("grep.fallbacktonoindex", &fallback);
 932                if (fallback)
 933                        use_index = 0;
 934                else
 935                        /* die the same way as if we did it at the beginning */
 936                        setup_git_directory();
 937        }
 938
 939        /*
 940         * skip a -- separator; we know it cannot be
 941         * separating revisions from pathnames if
 942         * we haven't even had any patterns yet
 943         */
 944        if (argc > 0 && !opt.pattern_list && !strcmp(argv[0], "--")) {
 945                argv++;
 946                argc--;
 947        }
 948
 949        /* First unrecognized non-option token */
 950        if (argc > 0 && !opt.pattern_list) {
 951                append_grep_pattern(&opt, argv[0], "command line", 0,
 952                                    GREP_PATTERN);
 953                argv++;
 954                argc--;
 955        }
 956
 957        if (show_in_pager == default_pager)
 958                show_in_pager = git_pager(1);
 959        if (show_in_pager) {
 960                opt.color = 0;
 961                opt.name_only = 1;
 962                opt.null_following_name = 1;
 963                opt.output_priv = &path_list;
 964                opt.output = append_path;
 965                string_list_append(&path_list, show_in_pager);
 966        }
 967
 968        if (!opt.pattern_list)
 969                die(_("no pattern given"));
 970
 971        /* --only-matching has no effect with --invert. */
 972        if (opt.invert)
 973                opt.only_matching = 0;
 974
 975        /*
 976         * We have to find "--" in a separate pass, because its presence
 977         * influences how we will parse arguments that come before it.
 978         */
 979        for (i = 0; i < argc; i++) {
 980                if (!strcmp(argv[i], "--")) {
 981                        seen_dashdash = 1;
 982                        break;
 983                }
 984        }
 985
 986        /*
 987         * Resolve any rev arguments. If we have a dashdash, then everything up
 988         * to it must resolve as a rev. If not, then we stop at the first
 989         * non-rev and assume everything else is a path.
 990         */
 991        allow_revs = use_index && !untracked;
 992        for (i = 0; i < argc; i++) {
 993                const char *arg = argv[i];
 994                struct object_id oid;
 995                struct object_context oc;
 996                struct object *object;
 997
 998                if (!strcmp(arg, "--")) {
 999                        i++;
1000                        break;
1001                }
1002
1003                if (!allow_revs) {
1004                        if (seen_dashdash)
1005                                die(_("--no-index or --untracked cannot be used with revs"));
1006                        break;
1007                }
1008
1009                if (get_oid_with_context(arg, GET_OID_RECORD_PATH,
1010                                         &oid, &oc)) {
1011                        if (seen_dashdash)
1012                                die(_("unable to resolve revision: %s"), arg);
1013                        break;
1014                }
1015
1016                object = parse_object_or_die(&oid, arg);
1017                if (!seen_dashdash)
1018                        verify_non_filename(prefix, arg);
1019                add_object_array_with_path(object, arg, &list, oc.mode, oc.path);
1020                free(oc.path);
1021        }
1022
1023        /*
1024         * Anything left over is presumed to be a path. But in the non-dashdash
1025         * "do what I mean" case, we verify and complain when that isn't true.
1026         */
1027        if (!seen_dashdash) {
1028                int j;
1029                for (j = i; j < argc; j++)
1030                        verify_filename(prefix, argv[j], j == i && allow_revs);
1031        }
1032
1033        parse_pathspec(&pathspec, 0,
1034                       PATHSPEC_PREFER_CWD |
1035                       (opt.max_depth != -1 ? PATHSPEC_MAXDEPTH_VALID : 0),
1036                       prefix, argv + i);
1037        pathspec.max_depth = opt.max_depth;
1038        pathspec.recursive = 1;
1039        pathspec.recurse_submodules = !!recurse_submodules;
1040
1041#ifndef NO_PTHREADS
1042        if (list.nr || cached || show_in_pager)
1043                num_threads = 0;
1044        else if (num_threads == 0)
1045                num_threads = GREP_NUM_THREADS_DEFAULT;
1046        else if (num_threads < 0)
1047                die(_("invalid number of threads specified (%d)"), num_threads);
1048        if (num_threads == 1)
1049                num_threads = 0;
1050#else
1051        if (num_threads)
1052                warning(_("no threads support, ignoring --threads"));
1053        num_threads = 0;
1054#endif
1055
1056        if (!num_threads)
1057                /*
1058                 * The compiled patterns on the main path are only
1059                 * used when not using threading. Otherwise
1060                 * start_threads() below calls compile_grep_patterns()
1061                 * for each thread.
1062                 */
1063                compile_grep_patterns(&opt);
1064
1065#ifndef NO_PTHREADS
1066        if (num_threads) {
1067                if (!(opt.name_only || opt.unmatch_name_only || opt.count)
1068                    && (opt.pre_context || opt.post_context ||
1069                        opt.file_break || opt.funcbody))
1070                        skip_first_line = 1;
1071                start_threads(&opt);
1072        }
1073#endif
1074
1075        if (show_in_pager && (cached || list.nr))
1076                die(_("--open-files-in-pager only works on the worktree"));
1077
1078        if (show_in_pager && opt.pattern_list && !opt.pattern_list->next) {
1079                const char *pager = path_list.items[0].string;
1080                int len = strlen(pager);
1081
1082                if (len > 4 && is_dir_sep(pager[len - 5]))
1083                        pager += len - 4;
1084
1085                if (opt.ignore_case && !strcmp("less", pager))
1086                        string_list_append(&path_list, "-I");
1087
1088                if (!strcmp("less", pager) || !strcmp("vi", pager)) {
1089                        struct strbuf buf = STRBUF_INIT;
1090                        strbuf_addf(&buf, "+/%s%s",
1091                                        strcmp("less", pager) ? "" : "*",
1092                                        opt.pattern_list->pattern);
1093                        string_list_append(&path_list, buf.buf);
1094                        strbuf_detach(&buf, NULL);
1095                }
1096        }
1097
1098        if (recurse_submodules && (!use_index || untracked))
1099                die(_("option not supported with --recurse-submodules"));
1100
1101        if (!show_in_pager && !opt.status_only)
1102                setup_pager();
1103
1104        if (!use_index && (untracked || cached))
1105                die(_("--cached or --untracked cannot be used with --no-index"));
1106
1107        if (!use_index || untracked) {
1108                int use_exclude = (opt_exclude < 0) ? use_index : !!opt_exclude;
1109                hit = grep_directory(&opt, &pathspec, use_exclude, use_index);
1110        } else if (0 <= opt_exclude) {
1111                die(_("--[no-]exclude-standard cannot be used for tracked contents"));
1112        } else if (!list.nr) {
1113                if (!cached)
1114                        setup_work_tree();
1115
1116                hit = grep_cache(&opt, the_repository, &pathspec, cached);
1117        } else {
1118                if (cached)
1119                        die(_("both --cached and trees are given"));
1120
1121                hit = grep_objects(&opt, &pathspec, &list);
1122        }
1123
1124        if (num_threads)
1125                hit |= wait_all();
1126        if (hit && show_in_pager)
1127                run_pager(&opt, prefix);
1128        clear_pathspec(&pathspec);
1129        free_grep_patterns(&opt);
1130        return !hit;
1131}