builtin-grep.c

   1 /*
   2  * Builtin "git grep"
   3  *
   4  * Copyright (c) 2006 Junio C Hamano
   5  */
   6 #include "cache.h"
   7 #include "blob.h"
   8 #include "tree.h"
   9 #include "commit.h"
  10 #include "tag.h"
  11 #include "tree-walk.h"
  12 #include "builtin.h"
  13 #include "parse-options.h"
  14 #include "userdiff.h"
  15 #include "grep.h"
  16
  17 #ifndef NO_EXTERNAL_GREP
  18 #ifdef __unix__
  19 #define NO_EXTERNAL_GREP 0
  20 #else
  21 #define NO_EXTERNAL_GREP 1
  22 #endif
  23 #endif
  24
  25 static char const * const grep_usage[] = {
  26         "git grep [options] [-e] <pattern> [<rev>...] [[--] path...]",
  27         NULL
  28 };
  29
  30 static int grep_config(const char *var, const char *value, void *cb)
  31 {
  32         struct grep_opt *opt = cb;
  33
  34         switch (userdiff_config(var, value)) {
  35         case 0: break;
  36         case -1: return -1;
  37         default: return 0;
  38         }
  39
  40         if (!strcmp(var, "color.grep")) {
  41                 opt->color = git_config_colorbool(var, value, -1);
  42                 return 0;
  43         }
  44         if (!strcmp(var, "color.grep.external"))
  45                 return git_config_string(&(opt->color_external), var, value);
  46         if (!strcmp(var, "color.grep.match")) {
  47                 if (!value)
  48                         return config_error_nonbool(var);
  49                 color_parse(value, var, opt->color_match);
  50                 return 0;
  51         }
  52         return git_color_default_config(var, value, cb);
  53 }
  54
  55 /*
  56  * Return non-zero if max_depth is negative or path has no more then max_depth
  57  * slashes.
  58  */
  59 static int accept_subdir(const char *path, int max_depth)
  60 {
  61         if (max_depth < 0)
  62                 return 1;
  63
  64         while ((path = strchr(path, '/')) != NULL) {
  65                 max_depth--;
  66                 if (max_depth < 0)
  67                         return 0;
  68                 path++;
  69         }
  70         return 1;
  71 }
  72
  73 /*
  74  * Return non-zero if name is a subdirectory of match and is not too deep.
  75  */
  76 static int is_subdir(const char *name, int namelen,
  77                 const char *match, int matchlen, int max_depth)
  78 {
  79         if (matchlen > namelen || strncmp(name, match, matchlen))
  80                 return 0;
  81
  82         if (name[matchlen] == '\0') /* exact match */
  83                 return 1;
  84
  85         if (!matchlen || match[matchlen-1] == '/' || name[matchlen] == '/')
  86                 return accept_subdir(name + matchlen + 1, max_depth);
  87
  88         return 0;
  89 }
  90
  91 /*
  92  * git grep pathspecs are somewhat different from diff-tree pathspecs;
  93  * pathname wildcards are allowed.
  94  */
  95 static int pathspec_matches(const char **paths, const char *name, int max_depth)
  96 {
  97         int namelen, i;
  98         if (!paths || !*paths)
  99                 return accept_subdir(name, max_depth);
 100         namelen = strlen(name);
 101         for (i = 0; paths[i]; i++) {
 102                 const char *match = paths[i];
 103                 int matchlen = strlen(match);
 104                 const char *cp, *meta;
 105
 106                 if (is_subdir(name, namelen, match, matchlen, max_depth))
 107                         return 1;
 108                 if (!fnmatch(match, name, 0))
 109                         return 1;
 110                 if (name[namelen-1] != '/')
 111                         continue;
 112
 113                 /* We are being asked if the directory ("name") is worth
 114                  * descending into.
 115                  *
 116                  * Find the longest leading directory name that does
 117                  * not have metacharacter in the pathspec; the name
 118                  * we are looking at must overlap with that directory.
 119                  */
 120                 for (cp = match, meta = NULL; cp - match < matchlen; cp++) {
 121                         char ch = *cp;
 122                         if (ch == '*' || ch == '[' || ch == '?') {
 123                                 meta = cp;
 124                                 break;
 125                         }
 126                 }
 127                 if (!meta)
 128                         meta = cp; /* fully literal */
 129
 130                 if (namelen <= meta - match) {
 131                         /* Looking at "Documentation/" and
 132                          * the pattern says "Documentation/howto/", or
 133                          * "Documentation/diff*.txt".  The name we
 134                          * have should match prefix.
 135                          */
 136                         if (!memcmp(match, name, namelen))
 137                                 return 1;
 138                         continue;
 139                 }
 140
 141                 if (meta - match < namelen) {
 142                         /* Looking at "Documentation/howto/" and
 143                          * the pattern says "Documentation/h*";
 144                          * match up to "Do.../h"; this avoids descending
 145                          * into "Documentation/technical/".
 146                          */
 147                         if (!memcmp(match, name, meta - match))
 148                                 return 1;
 149                         continue;
 150                 }
 151         }
 152         return 0;
 153 }
 154
 155 static int grep_sha1(struct grep_opt *opt, const unsigned char *sha1, const char *name, int tree_name_len)
 156 {
 157         unsigned long size;
 158         char *data;
 159         enum object_type type;
 160         char *to_free = NULL;
 161         int hit;
 162
 163         data = read_sha1_file(sha1, &type, &size);
 164         if (!data) {
 165                 error("'%s': unable to read %s", name, sha1_to_hex(sha1));
 166                 return 0;
 167         }
 168         if (opt->relative && opt->prefix_length) {
 169                 static char name_buf[PATH_MAX];
 170                 char *cp;
 171                 int name_len = strlen(name) - opt->prefix_length + 1;
 172
 173                 if (!tree_name_len)
 174                         name += opt->prefix_length;
 175                 else {
 176                         if (ARRAY_SIZE(name_buf) <= name_len)
 177                                 cp = to_free = xmalloc(name_len);
 178                         else
 179                                 cp = name_buf;
 180                         memcpy(cp, name, tree_name_len);
 181                         strcpy(cp + tree_name_len,
 182                                name + tree_name_len + opt->prefix_length);
 183                         name = cp;
 184                 }
 185         }
 186         hit = grep_buffer(opt, name, data, size);
 187         free(data);
 188         free(to_free);
 189         return hit;
 190 }
 191
 192 static int grep_file(struct grep_opt *opt, const char *filename)
 193 {
 194         struct stat st;
 195         int i;
 196         char *data;
 197         size_t sz;
 198
 199         if (lstat(filename, &st) < 0) {
 200         err_ret:
 201                 if (errno != ENOENT)
 202                         error("'%s': %s", filename, strerror(errno));
 203                 return 0;
 204         }
 205         if (!st.st_size)
 206                 return 0; /* empty file -- no grep hit */
 207         if (!S_ISREG(st.st_mode))
 208                 return 0;
 209         sz = xsize_t(st.st_size);
 210         i = open(filename, O_RDONLY);
 211         if (i < 0)
 212                 goto err_ret;
 213         data = xmalloc(sz + 1);
 214         if (st.st_size != read_in_full(i, data, sz)) {
 215                 error("'%s': short read %s", filename, strerror(errno));
 216                 close(i);
 217                 free(data);
 218                 return 0;
 219         }
 220         close(i);
 221         if (opt->relative && opt->prefix_length)
 222                 filename += opt->prefix_length;
 223         i = grep_buffer(opt, filename, data, sz);
 224         free(data);
 225         return i;
 226 }
 227
 228 #if !NO_EXTERNAL_GREP
 229 static int exec_grep(int argc, const char **argv)
 230 {
 231         pid_t pid;
 232         int status;
 233
 234         argv[argc] = NULL;
 235         pid = fork();
 236         if (pid < 0)
 237                 return pid;
 238         if (!pid) {
 239                 execvp("grep", (char **) argv);
 240                 exit(255);
 241         }
 242         while (waitpid(pid, &status, 0) < 0) {
 243                 if (errno == EINTR)
 244                         continue;
 245                 return -1;
 246         }
 247         if (WIFEXITED(status)) {
 248                 if (!WEXITSTATUS(status))
 249                         return 1;
 250                 return 0;
 251         }
 252         return -1;
 253 }
 254
 255 #define MAXARGS 1000
 256 #define ARGBUF 4096
 257 #define push_arg(a) do { \
 258         if (nr < MAXARGS) argv[nr++] = (a); \
 259         else die("maximum number of args exceeded"); \
 260         } while (0)
 261
 262 /*
 263  * If you send a singleton filename to grep, it does not give
 264  * the name of the file.  GNU grep has "-H" but we would want
 265  * that behaviour in a portable way.
 266  *
 267  * So we keep two pathnames in argv buffer unsent to grep in
 268  * the main loop if we need to do more than one grep.
 269  */
 270 static int flush_grep(struct grep_opt *opt,
 271                       int argc, int arg0, const char **argv, int *kept)
 272 {
 273         int status;
 274         int count = argc - arg0;
 275         const char *kept_0 = NULL;
 276
 277         if (count <= 2) {
 278                 /*
 279                  * Because we keep at least 2 paths in the call from
 280                  * the main loop (i.e. kept != NULL), and MAXARGS is
 281                  * far greater than 2, this usually is a call to
 282                  * conclude the grep.  However, the user could attempt
 283                  * to overflow the argv buffer by giving too many
 284                  * options to leave very small number of real
 285                  * arguments even for the call in the main loop.
 286                  */
 287                 if (kept)
 288                         die("insanely many options to grep");
 289
 290                 /*
 291                  * If we have two or more paths, we do not have to do
 292                  * anything special, but we need to push /dev/null to
 293                  * get "-H" behaviour of GNU grep portably but when we
 294                  * are not doing "-l" nor "-L" nor "-c".
 295                  */
 296                 if (count == 1 &&
 297                     !opt->name_only &&
 298                     !opt->unmatch_name_only &&
 299                     !opt->count) {
 300                         argv[argc++] = "/dev/null";
 301                         argv[argc] = NULL;
 302                 }
 303         }
 304
 305         else if (kept) {
 306                 /*
 307                  * Called because we found many paths and haven't finished
 308                  * iterating over the cache yet.  We keep two paths
 309                  * for the concluding call.  argv[argc-2] and argv[argc-1]
 310                  * has the last two paths, so save the first one away,
 311                  * replace it with NULL while sending the list to grep,
 312                  * and recover them after we are done.
 313                  */
 314                 *kept = 2;
 315                 kept_0 = argv[argc-2];
 316                 argv[argc-2] = NULL;
 317                 argc -= 2;
 318         }
 319
 320         if (opt->pre_context || opt->post_context) {
 321                 /*
 322                  * grep handles hunk marks between files, but we need to
 323                  * do that ourselves between multiple calls.
 324                  */
 325                 if (opt->show_hunk_mark)
 326                         write_or_die(1, "--\n", 3);
 327                 else
 328                         opt->show_hunk_mark = 1;
 329         }
 330
 331         status = exec_grep(argc, argv);
 332
 333         if (kept_0) {
 334                 /*
 335                  * Then recover them.  Now the last arg is beyond the
 336                  * terminating NULL which is at argc, and the second
 337                  * from the last is what we saved away in kept_0
 338                  */
 339                 argv[arg0++] = kept_0;
 340                 argv[arg0] = argv[argc+1];
 341         }
 342         return status;
 343 }
 344
 345 static void grep_add_color(struct strbuf *sb, const char *escape_seq)
 346 {
 347         size_t orig_len = sb->len;
 348
 349         while (*escape_seq) {
 350                 if (*escape_seq == 'm')
 351                         strbuf_addch(sb, ';');
 352                 else if (*escape_seq != '\033' && *escape_seq  != '[')
 353                         strbuf_addch(sb, *escape_seq);
 354                 escape_seq++;
 355         }
 356         if (sb->len > orig_len && sb->buf[sb->len - 1] == ';')
 357                 strbuf_setlen(sb, sb->len - 1);
 358 }
 359
 360 static int has_skip_worktree_entry(struct grep_opt *opt, const char **paths)
 361 {
 362         int nr;
 363         for (nr = 0; nr < active_nr; nr++) {
 364                 struct cache_entry *ce = active_cache[nr];
 365                 if (!S_ISREG(ce->ce_mode))
 366                         continue;
 367                 if (!pathspec_matches(paths, ce->name, opt->max_depth))
 368                         continue;
 369                 if (ce_skip_worktree(ce))
 370                         return 1;
 371         }
 372         return 0;
 373 }
 374
 375 static int external_grep(struct grep_opt *opt, const char **paths, int cached)
 376 {
 377         int i, nr, argc, hit, len, status;
 378         const char *argv[MAXARGS+1];
 379         char randarg[ARGBUF];
 380         char *argptr = randarg;
 381         struct grep_pat *p;
 382
 383         if (opt->extended || (opt->relative && opt->prefix_length)
 384             || has_skip_worktree_entry(opt, paths))
 385                 return -1;
 386         len = nr = 0;
 387         push_arg("grep");
 388         if (opt->fixed)
 389                 push_arg("-F");
 390         if (opt->linenum)
 391                 push_arg("-n");
 392         if (!opt->pathname)
 393                 push_arg("-h");
 394         if (opt->regflags & REG_EXTENDED)
 395                 push_arg("-E");
 396         if (opt->regflags & REG_ICASE)
 397                 push_arg("-i");
 398         if (opt->binary == GREP_BINARY_NOMATCH)
 399                 push_arg("-I");
 400         if (opt->word_regexp)
 401                 push_arg("-w");
 402         if (opt->name_only)
 403                 push_arg("-l");
 404         if (opt->unmatch_name_only)
 405                 push_arg("-L");
 406         if (opt->null_following_name)
 407                 /* in GNU grep git's "-z" translates to "-Z" */
 408                 push_arg("-Z");
 409         if (opt->count)
 410                 push_arg("-c");
 411         if (opt->post_context || opt->pre_context) {
 412                 if (opt->post_context != opt->pre_context) {
 413                         if (opt->pre_context) {
 414                                 push_arg("-B");
 415                                 len += snprintf(argptr, sizeof(randarg)-len,
 416                                                 "%u", opt->pre_context) + 1;
 417                                 if (sizeof(randarg) <= len)
 418                                         die("maximum length of args exceeded");
 419                                 push_arg(argptr);
 420                                 argptr += len;
 421                         }
 422                         if (opt->post_context) {
 423                                 push_arg("-A");
 424                                 len += snprintf(argptr, sizeof(randarg)-len,
 425                                                 "%u", opt->post_context) + 1;
 426                                 if (sizeof(randarg) <= len)
 427                                         die("maximum length of args exceeded");
 428                                 push_arg(argptr);
 429                                 argptr += len;
 430                         }
 431                 }
 432                 else {
 433                         push_arg("-C");
 434                         len += snprintf(argptr, sizeof(randarg)-len,
 435                                         "%u", opt->post_context) + 1;
 436                         if (sizeof(randarg) <= len)
 437                                 die("maximum length of args exceeded");
 438                         push_arg(argptr);
 439                         argptr += len;
 440                 }
 441         }
 442         for (p = opt->pattern_list; p; p = p->next) {
 443                 push_arg("-e");
 444                 push_arg(p->pattern);
 445         }
 446         if (opt->color) {
 447                 struct strbuf sb = STRBUF_INIT;
 448
 449                 grep_add_color(&sb, opt->color_match);
 450                 setenv("GREP_COLOR", sb.buf, 1);
 451
 452                 strbuf_reset(&sb);
 453                 strbuf_addstr(&sb, "mt=");
 454                 grep_add_color(&sb, opt->color_match);
 455                 strbuf_addstr(&sb, ":sl=:cx=:fn=:ln=:bn=:se=");
 456                 setenv("GREP_COLORS", sb.buf, 1);
 457
 458                 strbuf_release(&sb);
 459
 460                 if (opt->color_external && strlen(opt->color_external) > 0)
 461                         push_arg(opt->color_external);
 462         }
 463
 464         hit = 0;
 465         argc = nr;
 466         for (i = 0; i < active_nr; i++) {
 467                 struct cache_entry *ce = active_cache[i];
 468                 char *name;
 469                 int kept;
 470                 if (!S_ISREG(ce->ce_mode))
 471                         continue;
 472                 if (!pathspec_matches(paths, ce->name, opt->max_depth))
 473                         continue;
 474                 name = ce->name;
 475                 if (name[0] == '-') {
 476                         int len = ce_namelen(ce);
 477                         name = xmalloc(len + 3);
 478                         memcpy(name, "./", 2);
 479                         memcpy(name + 2, ce->name, len + 1);
 480                 }
 481                 argv[argc++] = name;
 482                 if (MAXARGS <= argc) {
 483                         status = flush_grep(opt, argc, nr, argv, &kept);
 484                         if (0 < status)
 485                                 hit = 1;
 486                         argc = nr + kept;
 487                 }
 488                 if (ce_stage(ce)) {
 489                         do {
 490                                 i++;
 491                         } while (i < active_nr &&
 492                                  !strcmp(ce->name, active_cache[i]->name));
 493                         i--; /* compensate for loop control */
 494                 }
 495         }
 496         if (argc > nr) {
 497                 status = flush_grep(opt, argc, nr, argv, NULL);
 498                 if (0 < status)
 499                         hit = 1;
 500         }
 501         return hit;
 502 }
 503 #endif
 504
 505 static int grep_cache(struct grep_opt *opt, const char **paths, int cached,
 506                       int external_grep_allowed)
 507 {
 508         int hit = 0;
 509         int nr;
 510         read_cache();
 511
 512 #if !NO_EXTERNAL_GREP
 513         /*
 514          * Use the external "grep" command for the case where
 515          * we grep through the checked-out files. It tends to
 516          * be a lot more optimized
 517          */
 518         if (!cached && external_grep_allowed) {
 519                 hit = external_grep(opt, paths, cached);
 520                 if (hit >= 0)
 521                         return hit;
 522         }
 523 #endif
 524
 525         for (nr = 0; nr < active_nr; nr++) {
 526                 struct cache_entry *ce = active_cache[nr];
 527                 if (!S_ISREG(ce->ce_mode))
 528                         continue;
 529                 if (!pathspec_matches(paths, ce->name, opt->max_depth))
 530                         continue;
 531                 /*
 532                  * If CE_VALID is on, we assume worktree file and its cache entry
 533                  * are identical, even if worktree file has been modified, so use
 534                  * cache version instead
 535                  */
 536                 if (cached || (ce->ce_flags & CE_VALID) || ce_skip_worktree(ce)) {
 537                         if (ce_stage(ce))
 538                                 continue;
 539                         hit |= grep_sha1(opt, ce->sha1, ce->name, 0);
 540                 }
 541                 else
 542                         hit |= grep_file(opt, ce->name);
 543                 if (ce_stage(ce)) {
 544                         do {
 545                                 nr++;
 546                         } while (nr < active_nr &&
 547                                  !strcmp(ce->name, active_cache[nr]->name));
 548                         nr--; /* compensate for loop control */
 549                 }
 550         }
 551         free_grep_patterns(opt);
 552         return hit;
 553 }
 554
 555 static int grep_tree(struct grep_opt *opt, const char **paths,
 556                      struct tree_desc *tree,
 557                      const char *tree_name, const char *base)
 558 {
 559         int len;
 560         int hit = 0;
 561         struct name_entry entry;
 562         char *down;
 563         int tn_len = strlen(tree_name);
 564         struct strbuf pathbuf;
 565
 566         strbuf_init(&pathbuf, PATH_MAX + tn_len);
 567
 568         if (tn_len) {
 569                 strbuf_add(&pathbuf, tree_name, tn_len);
 570                 strbuf_addch(&pathbuf, ':');
 571                 tn_len = pathbuf.len;
 572         }
 573         strbuf_addstr(&pathbuf, base);
 574         len = pathbuf.len;
 575
 576         while (tree_entry(tree, &entry)) {
 577                 int te_len = tree_entry_len(entry.path, entry.sha1);
 578                 pathbuf.len = len;
 579                 strbuf_add(&pathbuf, entry.path, te_len);
 580
 581                 if (S_ISDIR(entry.mode))
 582                         /* Match "abc/" against pathspec to
 583                          * decide if we want to descend into "abc"
 584                          * directory.
 585                          */
 586                         strbuf_addch(&pathbuf, '/');
 587
 588                 down = pathbuf.buf + tn_len;
 589                 if (!pathspec_matches(paths, down, opt->max_depth))
 590                         ;
 591                 else if (S_ISREG(entry.mode))
 592                         hit |= grep_sha1(opt, entry.sha1, pathbuf.buf, tn_len);
 593                 else if (S_ISDIR(entry.mode)) {
 594                         enum object_type type;
 595                         struct tree_desc sub;
 596                         void *data;
 597                         unsigned long size;
 598
 599                         data = read_sha1_file(entry.sha1, &type, &size);
 600                         if (!data)
 601                                 die("unable to read tree (%s)",
 602                                     sha1_to_hex(entry.sha1));
 603                         init_tree_desc(&sub, data, size);
 604                         hit |= grep_tree(opt, paths, &sub, tree_name, down);
 605                         free(data);
 606                 }
 607         }
 608         strbuf_release(&pathbuf);
 609         return hit;
 610 }
 611
 612 static int grep_object(struct grep_opt *opt, const char **paths,
 613                        struct object *obj, const char *name)
 614 {
 615         if (obj->type == OBJ_BLOB)
 616                 return grep_sha1(opt, obj->sha1, name, 0);
 617         if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
 618                 struct tree_desc tree;
 619                 void *data;
 620                 unsigned long size;
 621                 int hit;
 622                 data = read_object_with_reference(obj->sha1, tree_type,
 623                                                   &size, NULL);
 624                 if (!data)
 625                         die("unable to read tree (%s)", sha1_to_hex(obj->sha1));
 626                 init_tree_desc(&tree, data, size);
 627                 hit = grep_tree(opt, paths, &tree, name, "");
 628                 free(data);
 629                 return hit;
 630         }
 631         die("unable to grep from object of type %s", typename(obj->type));
 632 }
 633
 634 static int context_callback(const struct option *opt, const char *arg,
 635                             int unset)
 636 {
 637         struct grep_opt *grep_opt = opt->value;
 638         int value;
 639         const char *endp;
 640
 641         if (unset) {
 642                 grep_opt->pre_context = grep_opt->post_context = 0;
 643                 return 0;
 644         }
 645         value = strtol(arg, (char **)&endp, 10);
 646         if (*endp) {
 647                 return error("switch `%c' expects a numerical value",
 648                              opt->short_name);
 649         }
 650         grep_opt->pre_context = grep_opt->post_context = value;
 651         return 0;
 652 }
 653
 654 static int file_callback(const struct option *opt, const char *arg, int unset)
 655 {
 656         struct grep_opt *grep_opt = opt->value;
 657         FILE *patterns;
 658         int lno = 0;
 659         struct strbuf sb;
 660
 661         patterns = fopen(arg, "r");
 662         if (!patterns)
 663                 die_errno("cannot open '%s'", arg);
 664         while (strbuf_getline(&sb, patterns, '\n') == 0) {
 665                 /* ignore empty line like grep does */
 666                 if (sb.len == 0)
 667                         continue;
 668                 append_grep_pattern(grep_opt, strbuf_detach(&sb, NULL), arg,
 669                                     ++lno, GREP_PATTERN);
 670         }
 671         fclose(patterns);
 672         strbuf_release(&sb);
 673         return 0;
 674 }
 675
 676 static int not_callback(const struct option *opt, const char *arg, int unset)
 677 {
 678         struct grep_opt *grep_opt = opt->value;
 679         append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
 680         return 0;
 681 }
 682
 683 static int and_callback(const struct option *opt, const char *arg, int unset)
 684 {
 685         struct grep_opt *grep_opt = opt->value;
 686         append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
 687         return 0;
 688 }
 689
 690 static int open_callback(const struct option *opt, const char *arg, int unset)
 691 {
 692         struct grep_opt *grep_opt = opt->value;
 693         append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
 694         return 0;
 695 }
 696
 697 static int close_callback(const struct option *opt, const char *arg, int unset)
 698 {
 699         struct grep_opt *grep_opt = opt->value;
 700         append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
 701         return 0;
 702 }
 703
 704 static int pattern_callback(const struct option *opt, const char *arg,
 705                             int unset)
 706 {
 707         struct grep_opt *grep_opt = opt->value;
 708         append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
 709         return 0;
 710 }
 711
 712 static int help_callback(const struct option *opt, const char *arg, int unset)
 713 {
 714         return -1;
 715 }
 716
 717 int cmd_grep(int argc, const char **argv, const char *prefix)
 718 {
 719         int hit = 0;
 720         int cached = 0;
 721         int external_grep_allowed = 1;
 722         int seen_dashdash = 0;
 723         struct grep_opt opt;
 724         struct object_array list = { 0, 0, NULL };
 725         const char **paths = NULL;
 726         int i;
 727         int dummy;
 728         struct option options[] = {
 729                 OPT_BOOLEAN(0, "cached", &cached,
 730                         "search in index instead of in the work tree"),
 731                 OPT_GROUP(""),
 732                 OPT_BOOLEAN('v', "invert-match", &opt.invert,
 733                         "show non-matching lines"),
 734                 OPT_BIT('i', "ignore-case", &opt.regflags,
 735                         "case insensitive matching", REG_ICASE),
 736                 OPT_BOOLEAN('w', "word-regexp", &opt.word_regexp,
 737                         "match patterns only at word boundaries"),
 738                 OPT_SET_INT('a', "text", &opt.binary,
 739                         "process binary files as text", GREP_BINARY_TEXT),
 740                 OPT_SET_INT('I', NULL, &opt.binary,
 741                         "don't match patterns in binary files",
 742                         GREP_BINARY_NOMATCH),
 743                 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, "depth",
 744                         "descend at most <depth> levels", PARSE_OPT_NONEG,
 745                         NULL, 1 },
 746                 OPT_GROUP(""),
 747                 OPT_BIT('E', "extended-regexp", &opt.regflags,
 748                         "use extended POSIX regular expressions", REG_EXTENDED),
 749                 OPT_NEGBIT('G', "basic-regexp", &opt.regflags,
 750                         "use basic POSIX regular expressions (default)",
 751                         REG_EXTENDED),
 752                 OPT_BOOLEAN('F', "fixed-strings", &opt.fixed,
 753                         "interpret patterns as fixed strings"),
 754                 OPT_GROUP(""),
 755                 OPT_BOOLEAN('n', NULL, &opt.linenum, "show line numbers"),
 756                 OPT_NEGBIT('h', NULL, &opt.pathname, "don't show filenames", 1),
 757                 OPT_BIT('H', NULL, &opt.pathname, "show filenames", 1),
 758                 OPT_NEGBIT(0, "full-name", &opt.relative,
 759                         "show filenames relative to top directory", 1),
 760                 OPT_BOOLEAN('l', "files-with-matches", &opt.name_only,
 761                         "show only filenames instead of matching lines"),
 762                 OPT_BOOLEAN(0, "name-only", &opt.name_only,
 763                         "synonym for --files-with-matches"),
 764                 OPT_BOOLEAN('L', "files-without-match",
 765                         &opt.unmatch_name_only,
 766                         "show only the names of files without match"),
 767                 OPT_BOOLEAN('z', "null", &opt.null_following_name,
 768                         "print NUL after filenames"),
 769                 OPT_BOOLEAN('c', "count", &opt.count,
 770                         "show the number of matches instead of matching lines"),
 771                 OPT_SET_INT(0, "color", &opt.color, "highlight matches", 1),
 772                 OPT_GROUP(""),
 773                 OPT_CALLBACK('C', NULL, &opt, "n",
 774                         "show <n> context lines before and after matches",
 775                         context_callback),
 776                 OPT_INTEGER('B', NULL, &opt.pre_context,
 777                         "show <n> context lines before matches"),
 778                 OPT_INTEGER('A', NULL, &opt.post_context,
 779                         "show <n> context lines after matches"),
 780                 OPT_NUMBER_CALLBACK(&opt, "shortcut for -C NUM",
 781                         context_callback),
 782                 OPT_BOOLEAN('p', "show-function", &opt.funcname,
 783                         "show a line with the function name before matches"),
 784                 OPT_GROUP(""),
 785                 OPT_CALLBACK('f', NULL, &opt, "file",
 786                         "read patterns from file", file_callback),
 787                 { OPTION_CALLBACK, 'e', NULL, &opt, "pattern",
 788                         "match <pattern>", PARSE_OPT_NONEG, pattern_callback },
 789                 { OPTION_CALLBACK, 0, "and", &opt, NULL,
 790                   "combine patterns specified with -e",
 791                   PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
 792                 OPT_BOOLEAN(0, "or", &dummy, ""),
 793                 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
 794                   PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
 795                 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
 796                   PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
 797                   open_callback },
 798                 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
 799                   PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
 800                   close_callback },
 801                 OPT_BOOLEAN(0, "all-match", &opt.all_match,
 802                         "show only matches from files that match all patterns"),
 803                 OPT_GROUP(""),
 804 #if NO_EXTERNAL_GREP
 805                 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
 806                         "allow calling of grep(1) (ignored by this build)"),
 807 #else
 808                 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
 809                         "allow calling of grep(1) (default)"),
 810 #endif
 811                 { OPTION_CALLBACK, 0, "help-all", &options, NULL, "show usage",
 812                   PARSE_OPT_HIDDEN | PARSE_OPT_NOARG, help_callback },
 813                 OPT_END()
 814         };
 815
 816         memset(&opt, 0, sizeof(opt));
 817         opt.prefix_length = (prefix && *prefix) ? strlen(prefix) : 0;
 818         opt.relative = 1;
 819         opt.pathname = 1;
 820         opt.pattern_tail = &opt.pattern_list;
 821         opt.regflags = REG_NEWLINE;
 822         opt.max_depth = -1;
 823
 824         strcpy(opt.color_match, GIT_COLOR_RED GIT_COLOR_BOLD);
 825         opt.color = -1;
 826         git_config(grep_config, &opt);
 827         if (opt.color == -1)
 828                 opt.color = git_use_color_default;
 829
 830         /*
 831          * If there is no -- then the paths must exist in the working
 832          * tree.  If there is no explicit pattern specified with -e or
 833          * -f, we take the first unrecognized non option to be the
 834          * pattern, but then what follows it must be zero or more
 835          * valid refs up to the -- (if exists), and then existing
 836          * paths.  If there is an explicit pattern, then the first
 837          * unrecognized non option is the beginning of the refs list
 838          * that continues up to the -- (if exists), and then paths.
 839          */
 840         argc = parse_options(argc, argv, prefix, options, grep_usage,
 841                              PARSE_OPT_KEEP_DASHDASH |
 842                              PARSE_OPT_STOP_AT_NON_OPTION |
 843                              PARSE_OPT_NO_INTERNAL_HELP);
 844
 845         /* First unrecognized non-option token */
 846         if (argc > 0 && !opt.pattern_list) {
 847                 append_grep_pattern(&opt, argv[0], "command line", 0,
 848                                     GREP_PATTERN);
 849                 argv++;
 850                 argc--;
 851         }
 852
 853         if ((opt.color && !opt.color_external) || opt.funcname)
 854                 external_grep_allowed = 0;
 855         if (!opt.pattern_list)
 856                 die("no pattern given.");
 857         if ((opt.regflags != REG_NEWLINE) && opt.fixed)
 858                 die("cannot mix --fixed-strings and regexp");
 859         compile_grep_patterns(&opt);
 860
 861         /* Check revs and then paths */
 862         for (i = 0; i < argc; i++) {
 863                 const char *arg = argv[i];
 864                 unsigned char sha1[20];
 865                 /* Is it a rev? */
 866                 if (!get_sha1(arg, sha1)) {
 867                         struct object *object = parse_object(sha1);
 868                         if (!object)
 869                                 die("bad object %s", arg);
 870                         add_object_array(object, arg, &list);
 871                         continue;
 872                 }
 873                 if (!strcmp(arg, "--")) {
 874                         i++;
 875                         seen_dashdash = 1;
 876                 }
 877                 break;
 878         }
 879
 880         /* The rest are paths */
 881         if (!seen_dashdash) {
 882                 int j;
 883                 for (j = i; j < argc; j++)
 884                         verify_filename(prefix, argv[j]);
 885         }
 886
 887         if (i < argc) {
 888                 paths = get_pathspec(prefix, argv + i);
 889                 if (opt.prefix_length && opt.relative) {
 890                         /* Make sure we do not get outside of paths */
 891                         for (i = 0; paths[i]; i++)
 892                                 if (strncmp(prefix, paths[i], opt.prefix_length))
 893                                         die("git grep: cannot generate relative filenames containing '..'");
 894                 }
 895         }
 896         else if (prefix) {
 897                 paths = xcalloc(2, sizeof(const char *));
 898                 paths[0] = prefix;
 899                 paths[1] = NULL;
 900         }
 901
 902         if (!list.nr) {
 903                 if (!cached)
 904                         setup_work_tree();
 905                 return !grep_cache(&opt, paths, cached, external_grep_allowed);
 906         }
 907
 908         if (cached)
 909                 die("both --cached and trees are given.");
 910
 911         for (i = 0; i < list.nr; i++) {
 912                 struct object *real_obj;
 913                 real_obj = deref_tag(list.objects[i].item, NULL, 0);
 914                 if (grep_object(&opt, paths, real_obj, list.objects[i].name))
 915                         hit = 1;
 916         }
 917         free_grep_patterns(&opt);
 918         return !hit;
 919 }