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         trace_argv_printf(argv, "trace: grep:");
 236         pid = fork();
 237         if (pid < 0)
 238                 return pid;
 239         if (!pid) {
 240                 execvp("grep", (char **) argv);
 241                 exit(255);
 242         }
 243         while (waitpid(pid, &status, 0) < 0) {
 244                 if (errno == EINTR)
 245                         continue;
 246                 return -1;
 247         }
 248         if (WIFEXITED(status)) {
 249                 if (!WEXITSTATUS(status))
 250                         return 1;
 251                 return 0;
 252         }
 253         return -1;
 254 }
 255
 256 #define MAXARGS 1000
 257 #define ARGBUF 4096
 258 #define push_arg(a) do { \
 259         if (nr < MAXARGS) argv[nr++] = (a); \
 260         else die("maximum number of args exceeded"); \
 261         } while (0)
 262
 263 /*
 264  * If you send a singleton filename to grep, it does not give
 265  * the name of the file.  GNU grep has "-H" but we would want
 266  * that behaviour in a portable way.
 267  *
 268  * So we keep two pathnames in argv buffer unsent to grep in
 269  * the main loop if we need to do more than one grep.
 270  */
 271 static int flush_grep(struct grep_opt *opt,
 272                       int argc, int arg0, const char **argv, int *kept)
 273 {
 274         int status;
 275         int count = argc - arg0;
 276         const char *kept_0 = NULL;
 277
 278         if (count <= 2) {
 279                 /*
 280                  * Because we keep at least 2 paths in the call from
 281                  * the main loop (i.e. kept != NULL), and MAXARGS is
 282                  * far greater than 2, this usually is a call to
 283                  * conclude the grep.  However, the user could attempt
 284                  * to overflow the argv buffer by giving too many
 285                  * options to leave very small number of real
 286                  * arguments even for the call in the main loop.
 287                  */
 288                 if (kept)
 289                         die("insanely many options to grep");
 290
 291                 /*
 292                  * If we have two or more paths, we do not have to do
 293                  * anything special, but we need to push /dev/null to
 294                  * get "-H" behaviour of GNU grep portably but when we
 295                  * are not doing "-l" nor "-L" nor "-c".
 296                  */
 297                 if (count == 1 &&
 298                     !opt->name_only &&
 299                     !opt->unmatch_name_only &&
 300                     !opt->count) {
 301                         argv[argc++] = "/dev/null";
 302                         argv[argc] = NULL;
 303                 }
 304         }
 305
 306         else if (kept) {
 307                 /*
 308                  * Called because we found many paths and haven't finished
 309                  * iterating over the cache yet.  We keep two paths
 310                  * for the concluding call.  argv[argc-2] and argv[argc-1]
 311                  * has the last two paths, so save the first one away,
 312                  * replace it with NULL while sending the list to grep,
 313                  * and recover them after we are done.
 314                  */
 315                 *kept = 2;
 316                 kept_0 = argv[argc-2];
 317                 argv[argc-2] = NULL;
 318                 argc -= 2;
 319         }
 320
 321         if (opt->pre_context || opt->post_context) {
 322                 /*
 323                  * grep handles hunk marks between files, but we need to
 324                  * do that ourselves between multiple calls.
 325                  */
 326                 if (opt->show_hunk_mark)
 327                         write_or_die(1, "--\n", 3);
 328                 else
 329                         opt->show_hunk_mark = 1;
 330         }
 331
 332         status = exec_grep(argc, argv);
 333
 334         if (kept_0) {
 335                 /*
 336                  * Then recover them.  Now the last arg is beyond the
 337                  * terminating NULL which is at argc, and the second
 338                  * from the last is what we saved away in kept_0
 339                  */
 340                 argv[arg0++] = kept_0;
 341                 argv[arg0] = argv[argc+1];
 342         }
 343         return status;
 344 }
 345
 346 static void grep_add_color(struct strbuf *sb, const char *escape_seq)
 347 {
 348         size_t orig_len = sb->len;
 349
 350         while (*escape_seq) {
 351                 if (*escape_seq == 'm')
 352                         strbuf_addch(sb, ';');
 353                 else if (*escape_seq != '\033' && *escape_seq  != '[')
 354                         strbuf_addch(sb, *escape_seq);
 355                 escape_seq++;
 356         }
 357         if (sb->len > orig_len && sb->buf[sb->len - 1] == ';')
 358                 strbuf_setlen(sb, sb->len - 1);
 359 }
 360
 361 static int has_skip_worktree_entry(struct grep_opt *opt, const char **paths)
 362 {
 363         int nr;
 364         for (nr = 0; nr < active_nr; nr++) {
 365                 struct cache_entry *ce = active_cache[nr];
 366                 if (!S_ISREG(ce->ce_mode))
 367                         continue;
 368                 if (!pathspec_matches(paths, ce->name, opt->max_depth))
 369                         continue;
 370                 if (ce_skip_worktree(ce))
 371                         return 1;
 372         }
 373         return 0;
 374 }
 375
 376 static int external_grep(struct grep_opt *opt, const char **paths, int cached)
 377 {
 378         int i, nr, argc, hit, len, status;
 379         const char *argv[MAXARGS+1];
 380         char randarg[ARGBUF];
 381         char *argptr = randarg;
 382         struct grep_pat *p;
 383
 384         if (opt->extended || (opt->relative && opt->prefix_length)
 385             || has_skip_worktree_entry(opt, paths))
 386                 return -1;
 387         len = nr = 0;
 388         push_arg("grep");
 389         if (opt->fixed)
 390                 push_arg("-F");
 391         if (opt->linenum)
 392                 push_arg("-n");
 393         if (!opt->pathname)
 394                 push_arg("-h");
 395         if (opt->regflags & REG_EXTENDED)
 396                 push_arg("-E");
 397         if (opt->regflags & REG_ICASE)
 398                 push_arg("-i");
 399         if (opt->binary == GREP_BINARY_NOMATCH)
 400                 push_arg("-I");
 401         if (opt->word_regexp)
 402                 push_arg("-w");
 403         if (opt->name_only)
 404                 push_arg("-l");
 405         if (opt->unmatch_name_only)
 406                 push_arg("-L");
 407         if (opt->null_following_name)
 408                 /* in GNU grep git's "-z" translates to "-Z" */
 409                 push_arg("-Z");
 410         if (opt->count)
 411                 push_arg("-c");
 412         if (opt->post_context || opt->pre_context) {
 413                 if (opt->post_context != opt->pre_context) {
 414                         if (opt->pre_context) {
 415                                 push_arg("-B");
 416                                 len += snprintf(argptr, sizeof(randarg)-len,
 417                                                 "%u", opt->pre_context) + 1;
 418                                 if (sizeof(randarg) <= len)
 419                                         die("maximum length of args exceeded");
 420                                 push_arg(argptr);
 421                                 argptr += len;
 422                         }
 423                         if (opt->post_context) {
 424                                 push_arg("-A");
 425                                 len += snprintf(argptr, sizeof(randarg)-len,
 426                                                 "%u", opt->post_context) + 1;
 427                                 if (sizeof(randarg) <= len)
 428                                         die("maximum length of args exceeded");
 429                                 push_arg(argptr);
 430                                 argptr += len;
 431                         }
 432                 }
 433                 else {
 434                         push_arg("-C");
 435                         len += snprintf(argptr, sizeof(randarg)-len,
 436                                         "%u", opt->post_context) + 1;
 437                         if (sizeof(randarg) <= len)
 438                                 die("maximum length of args exceeded");
 439                         push_arg(argptr);
 440                         argptr += len;
 441                 }
 442         }
 443         for (p = opt->pattern_list; p; p = p->next) {
 444                 push_arg("-e");
 445                 push_arg(p->pattern);
 446         }
 447         if (opt->color) {
 448                 struct strbuf sb = STRBUF_INIT;
 449
 450                 grep_add_color(&sb, opt->color_match);
 451                 setenv("GREP_COLOR", sb.buf, 1);
 452
 453                 strbuf_reset(&sb);
 454                 strbuf_addstr(&sb, "mt=");
 455                 grep_add_color(&sb, opt->color_match);
 456                 strbuf_addstr(&sb, ":sl=:cx=:fn=:ln=:bn=:se=");
 457                 setenv("GREP_COLORS", sb.buf, 1);
 458
 459                 strbuf_release(&sb);
 460
 461                 if (opt->color_external && strlen(opt->color_external) > 0)
 462                         push_arg(opt->color_external);
 463         }
 464
 465         hit = 0;
 466         argc = nr;
 467         for (i = 0; i < active_nr; i++) {
 468                 struct cache_entry *ce = active_cache[i];
 469                 char *name;
 470                 int kept;
 471                 if (!S_ISREG(ce->ce_mode))
 472                         continue;
 473                 if (!pathspec_matches(paths, ce->name, opt->max_depth))
 474                         continue;
 475                 name = ce->name;
 476                 if (name[0] == '-') {
 477                         int len = ce_namelen(ce);
 478                         name = xmalloc(len + 3);
 479                         memcpy(name, "./", 2);
 480                         memcpy(name + 2, ce->name, len + 1);
 481                 }
 482                 argv[argc++] = name;
 483                 if (MAXARGS <= argc) {
 484                         status = flush_grep(opt, argc, nr, argv, &kept);
 485                         if (0 < status)
 486                                 hit = 1;
 487                         argc = nr + kept;
 488                 }
 489                 if (ce_stage(ce)) {
 490                         do {
 491                                 i++;
 492                         } while (i < active_nr &&
 493                                  !strcmp(ce->name, active_cache[i]->name));
 494                         i--; /* compensate for loop control */
 495                 }
 496         }
 497         if (argc > nr) {
 498                 status = flush_grep(opt, argc, nr, argv, NULL);
 499                 if (0 < status)
 500                         hit = 1;
 501         }
 502         return hit;
 503 }
 504 #endif
 505
 506 static int grep_cache(struct grep_opt *opt, const char **paths, int cached,
 507                       int external_grep_allowed)
 508 {
 509         int hit = 0;
 510         int nr;
 511         read_cache();
 512
 513 #if !NO_EXTERNAL_GREP
 514         /*
 515          * Use the external "grep" command for the case where
 516          * we grep through the checked-out files. It tends to
 517          * be a lot more optimized
 518          */
 519         if (!cached && external_grep_allowed) {
 520                 hit = external_grep(opt, paths, cached);
 521                 if (hit >= 0)
 522                         return hit;
 523         }
 524 #endif
 525
 526         for (nr = 0; nr < active_nr; nr++) {
 527                 struct cache_entry *ce = active_cache[nr];
 528                 if (!S_ISREG(ce->ce_mode))
 529                         continue;
 530                 if (!pathspec_matches(paths, ce->name, opt->max_depth))
 531                         continue;
 532                 /*
 533                  * If CE_VALID is on, we assume worktree file and its cache entry
 534                  * are identical, even if worktree file has been modified, so use
 535                  * cache version instead
 536                  */
 537                 if (cached || (ce->ce_flags & CE_VALID) || ce_skip_worktree(ce)) {
 538                         if (ce_stage(ce))
 539                                 continue;
 540                         hit |= grep_sha1(opt, ce->sha1, ce->name, 0);
 541                 }
 542                 else
 543                         hit |= grep_file(opt, ce->name);
 544                 if (ce_stage(ce)) {
 545                         do {
 546                                 nr++;
 547                         } while (nr < active_nr &&
 548                                  !strcmp(ce->name, active_cache[nr]->name));
 549                         nr--; /* compensate for loop control */
 550                 }
 551         }
 552         free_grep_patterns(opt);
 553         return hit;
 554 }
 555
 556 static int grep_tree(struct grep_opt *opt, const char **paths,
 557                      struct tree_desc *tree,
 558                      const char *tree_name, const char *base)
 559 {
 560         int len;
 561         int hit = 0;
 562         struct name_entry entry;
 563         char *down;
 564         int tn_len = strlen(tree_name);
 565         struct strbuf pathbuf;
 566
 567         strbuf_init(&pathbuf, PATH_MAX + tn_len);
 568
 569         if (tn_len) {
 570                 strbuf_add(&pathbuf, tree_name, tn_len);
 571                 strbuf_addch(&pathbuf, ':');
 572                 tn_len = pathbuf.len;
 573         }
 574         strbuf_addstr(&pathbuf, base);
 575         len = pathbuf.len;
 576
 577         while (tree_entry(tree, &entry)) {
 578                 int te_len = tree_entry_len(entry.path, entry.sha1);
 579                 pathbuf.len = len;
 580                 strbuf_add(&pathbuf, entry.path, te_len);
 581
 582                 if (S_ISDIR(entry.mode))
 583                         /* Match "abc/" against pathspec to
 584                          * decide if we want to descend into "abc"
 585                          * directory.
 586                          */
 587                         strbuf_addch(&pathbuf, '/');
 588
 589                 down = pathbuf.buf + tn_len;
 590                 if (!pathspec_matches(paths, down, opt->max_depth))
 591                         ;
 592                 else if (S_ISREG(entry.mode))
 593                         hit |= grep_sha1(opt, entry.sha1, pathbuf.buf, tn_len);
 594                 else if (S_ISDIR(entry.mode)) {
 595                         enum object_type type;
 596                         struct tree_desc sub;
 597                         void *data;
 598                         unsigned long size;
 599
 600                         data = read_sha1_file(entry.sha1, &type, &size);
 601                         if (!data)
 602                                 die("unable to read tree (%s)",
 603                                     sha1_to_hex(entry.sha1));
 604                         init_tree_desc(&sub, data, size);
 605                         hit |= grep_tree(opt, paths, &sub, tree_name, down);
 606                         free(data);
 607                 }
 608         }
 609         strbuf_release(&pathbuf);
 610         return hit;
 611 }
 612
 613 static int grep_object(struct grep_opt *opt, const char **paths,
 614                        struct object *obj, const char *name)
 615 {
 616         if (obj->type == OBJ_BLOB)
 617                 return grep_sha1(opt, obj->sha1, name, 0);
 618         if (obj->type == OBJ_COMMIT || obj->type == OBJ_TREE) {
 619                 struct tree_desc tree;
 620                 void *data;
 621                 unsigned long size;
 622                 int hit;
 623                 data = read_object_with_reference(obj->sha1, tree_type,
 624                                                   &size, NULL);
 625                 if (!data)
 626                         die("unable to read tree (%s)", sha1_to_hex(obj->sha1));
 627                 init_tree_desc(&tree, data, size);
 628                 hit = grep_tree(opt, paths, &tree, name, "");
 629                 free(data);
 630                 return hit;
 631         }
 632         die("unable to grep from object of type %s", typename(obj->type));
 633 }
 634
 635 static int context_callback(const struct option *opt, const char *arg,
 636                             int unset)
 637 {
 638         struct grep_opt *grep_opt = opt->value;
 639         int value;
 640         const char *endp;
 641
 642         if (unset) {
 643                 grep_opt->pre_context = grep_opt->post_context = 0;
 644                 return 0;
 645         }
 646         value = strtol(arg, (char **)&endp, 10);
 647         if (*endp) {
 648                 return error("switch `%c' expects a numerical value",
 649                              opt->short_name);
 650         }
 651         grep_opt->pre_context = grep_opt->post_context = value;
 652         return 0;
 653 }
 654
 655 static int file_callback(const struct option *opt, const char *arg, int unset)
 656 {
 657         struct grep_opt *grep_opt = opt->value;
 658         FILE *patterns;
 659         int lno = 0;
 660         struct strbuf sb;
 661
 662         patterns = fopen(arg, "r");
 663         if (!patterns)
 664                 die_errno("cannot open '%s'", arg);
 665         while (strbuf_getline(&sb, patterns, '\n') == 0) {
 666                 /* ignore empty line like grep does */
 667                 if (sb.len == 0)
 668                         continue;
 669                 append_grep_pattern(grep_opt, strbuf_detach(&sb, NULL), arg,
 670                                     ++lno, GREP_PATTERN);
 671         }
 672         fclose(patterns);
 673         strbuf_release(&sb);
 674         return 0;
 675 }
 676
 677 static int not_callback(const struct option *opt, const char *arg, int unset)
 678 {
 679         struct grep_opt *grep_opt = opt->value;
 680         append_grep_pattern(grep_opt, "--not", "command line", 0, GREP_NOT);
 681         return 0;
 682 }
 683
 684 static int and_callback(const struct option *opt, const char *arg, int unset)
 685 {
 686         struct grep_opt *grep_opt = opt->value;
 687         append_grep_pattern(grep_opt, "--and", "command line", 0, GREP_AND);
 688         return 0;
 689 }
 690
 691 static int open_callback(const struct option *opt, const char *arg, int unset)
 692 {
 693         struct grep_opt *grep_opt = opt->value;
 694         append_grep_pattern(grep_opt, "(", "command line", 0, GREP_OPEN_PAREN);
 695         return 0;
 696 }
 697
 698 static int close_callback(const struct option *opt, const char *arg, int unset)
 699 {
 700         struct grep_opt *grep_opt = opt->value;
 701         append_grep_pattern(grep_opt, ")", "command line", 0, GREP_CLOSE_PAREN);
 702         return 0;
 703 }
 704
 705 static int pattern_callback(const struct option *opt, const char *arg,
 706                             int unset)
 707 {
 708         struct grep_opt *grep_opt = opt->value;
 709         append_grep_pattern(grep_opt, arg, "-e option", 0, GREP_PATTERN);
 710         return 0;
 711 }
 712
 713 static int help_callback(const struct option *opt, const char *arg, int unset)
 714 {
 715         return -1;
 716 }
 717
 718 int cmd_grep(int argc, const char **argv, const char *prefix)
 719 {
 720         int hit = 0;
 721         int cached = 0;
 722         int external_grep_allowed = 1;
 723         int seen_dashdash = 0;
 724         struct grep_opt opt;
 725         struct object_array list = { 0, 0, NULL };
 726         const char **paths = NULL;
 727         int i;
 728         int dummy;
 729         struct option options[] = {
 730                 OPT_BOOLEAN(0, "cached", &cached,
 731                         "search in index instead of in the work tree"),
 732                 OPT_GROUP(""),
 733                 OPT_BOOLEAN('v', "invert-match", &opt.invert,
 734                         "show non-matching lines"),
 735                 OPT_BIT('i', "ignore-case", &opt.regflags,
 736                         "case insensitive matching", REG_ICASE),
 737                 OPT_BOOLEAN('w', "word-regexp", &opt.word_regexp,
 738                         "match patterns only at word boundaries"),
 739                 OPT_SET_INT('a', "text", &opt.binary,
 740                         "process binary files as text", GREP_BINARY_TEXT),
 741                 OPT_SET_INT('I', NULL, &opt.binary,
 742                         "don't match patterns in binary files",
 743                         GREP_BINARY_NOMATCH),
 744                 { OPTION_INTEGER, 0, "max-depth", &opt.max_depth, "depth",
 745                         "descend at most <depth> levels", PARSE_OPT_NONEG,
 746                         NULL, 1 },
 747                 OPT_GROUP(""),
 748                 OPT_BIT('E', "extended-regexp", &opt.regflags,
 749                         "use extended POSIX regular expressions", REG_EXTENDED),
 750                 OPT_NEGBIT('G', "basic-regexp", &opt.regflags,
 751                         "use basic POSIX regular expressions (default)",
 752                         REG_EXTENDED),
 753                 OPT_BOOLEAN('F', "fixed-strings", &opt.fixed,
 754                         "interpret patterns as fixed strings"),
 755                 OPT_GROUP(""),
 756                 OPT_BOOLEAN('n', NULL, &opt.linenum, "show line numbers"),
 757                 OPT_NEGBIT('h', NULL, &opt.pathname, "don't show filenames", 1),
 758                 OPT_BIT('H', NULL, &opt.pathname, "show filenames", 1),
 759                 OPT_NEGBIT(0, "full-name", &opt.relative,
 760                         "show filenames relative to top directory", 1),
 761                 OPT_BOOLEAN('l', "files-with-matches", &opt.name_only,
 762                         "show only filenames instead of matching lines"),
 763                 OPT_BOOLEAN(0, "name-only", &opt.name_only,
 764                         "synonym for --files-with-matches"),
 765                 OPT_BOOLEAN('L', "files-without-match",
 766                         &opt.unmatch_name_only,
 767                         "show only the names of files without match"),
 768                 OPT_BOOLEAN('z', "null", &opt.null_following_name,
 769                         "print NUL after filenames"),
 770                 OPT_BOOLEAN('c', "count", &opt.count,
 771                         "show the number of matches instead of matching lines"),
 772                 OPT_SET_INT(0, "color", &opt.color, "highlight matches", 1),
 773                 OPT_GROUP(""),
 774                 OPT_CALLBACK('C', NULL, &opt, "n",
 775                         "show <n> context lines before and after matches",
 776                         context_callback),
 777                 OPT_INTEGER('B', NULL, &opt.pre_context,
 778                         "show <n> context lines before matches"),
 779                 OPT_INTEGER('A', NULL, &opt.post_context,
 780                         "show <n> context lines after matches"),
 781                 OPT_NUMBER_CALLBACK(&opt, "shortcut for -C NUM",
 782                         context_callback),
 783                 OPT_BOOLEAN('p', "show-function", &opt.funcname,
 784                         "show a line with the function name before matches"),
 785                 OPT_GROUP(""),
 786                 OPT_CALLBACK('f', NULL, &opt, "file",
 787                         "read patterns from file", file_callback),
 788                 { OPTION_CALLBACK, 'e', NULL, &opt, "pattern",
 789                         "match <pattern>", PARSE_OPT_NONEG, pattern_callback },
 790                 { OPTION_CALLBACK, 0, "and", &opt, NULL,
 791                   "combine patterns specified with -e",
 792                   PARSE_OPT_NOARG | PARSE_OPT_NONEG, and_callback },
 793                 OPT_BOOLEAN(0, "or", &dummy, ""),
 794                 { OPTION_CALLBACK, 0, "not", &opt, NULL, "",
 795                   PARSE_OPT_NOARG | PARSE_OPT_NONEG, not_callback },
 796                 { OPTION_CALLBACK, '(', NULL, &opt, NULL, "",
 797                   PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
 798                   open_callback },
 799                 { OPTION_CALLBACK, ')', NULL, &opt, NULL, "",
 800                   PARSE_OPT_NOARG | PARSE_OPT_NONEG | PARSE_OPT_NODASH,
 801                   close_callback },
 802                 OPT_BOOLEAN(0, "all-match", &opt.all_match,
 803                         "show only matches from files that match all patterns"),
 804                 OPT_GROUP(""),
 805 #if NO_EXTERNAL_GREP
 806                 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
 807                         "allow calling of grep(1) (ignored by this build)"),
 808 #else
 809                 OPT_BOOLEAN(0, "ext-grep", &external_grep_allowed,
 810                         "allow calling of grep(1) (default)"),
 811 #endif
 812                 { OPTION_CALLBACK, 0, "help-all", &options, NULL, "show usage",
 813                   PARSE_OPT_HIDDEN | PARSE_OPT_NOARG, help_callback },
 814                 OPT_END()
 815         };
 816
 817         memset(&opt, 0, sizeof(opt));
 818         opt.prefix_length = (prefix && *prefix) ? strlen(prefix) : 0;
 819         opt.relative = 1;
 820         opt.pathname = 1;
 821         opt.pattern_tail = &opt.pattern_list;
 822         opt.regflags = REG_NEWLINE;
 823         opt.max_depth = -1;
 824
 825         strcpy(opt.color_match, GIT_COLOR_RED GIT_COLOR_BOLD);
 826         opt.color = -1;
 827         git_config(grep_config, &opt);
 828         if (opt.color == -1)
 829                 opt.color = git_use_color_default;
 830
 831         /*
 832          * If there is no -- then the paths must exist in the working
 833          * tree.  If there is no explicit pattern specified with -e or
 834          * -f, we take the first unrecognized non option to be the
 835          * pattern, but then what follows it must be zero or more
 836          * valid refs up to the -- (if exists), and then existing
 837          * paths.  If there is an explicit pattern, then the first
 838          * unrecognized non option is the beginning of the refs list
 839          * that continues up to the -- (if exists), and then paths.
 840          */
 841         argc = parse_options(argc, argv, prefix, options, grep_usage,
 842                              PARSE_OPT_KEEP_DASHDASH |
 843                              PARSE_OPT_STOP_AT_NON_OPTION |
 844                              PARSE_OPT_NO_INTERNAL_HELP);
 845
 846         /* First unrecognized non-option token */
 847         if (argc > 0 && !opt.pattern_list) {
 848                 append_grep_pattern(&opt, argv[0], "command line", 0,
 849                                     GREP_PATTERN);
 850                 argv++;
 851                 argc--;
 852         }
 853
 854         if ((opt.color && !opt.color_external) || opt.funcname)
 855                 external_grep_allowed = 0;
 856         if (!opt.pattern_list)
 857                 die("no pattern given.");
 858         if ((opt.regflags != REG_NEWLINE) && opt.fixed)
 859                 die("cannot mix --fixed-strings and regexp");
 860         compile_grep_patterns(&opt);
 861
 862         /* Check revs and then paths */
 863         for (i = 0; i < argc; i++) {
 864                 const char *arg = argv[i];
 865                 unsigned char sha1[20];
 866                 /* Is it a rev? */
 867                 if (!get_sha1(arg, sha1)) {
 868                         struct object *object = parse_object(sha1);
 869                         if (!object)
 870                                 die("bad object %s", arg);
 871                         add_object_array(object, arg, &list);
 872                         continue;
 873                 }
 874                 if (!strcmp(arg, "--")) {
 875                         i++;
 876                         seen_dashdash = 1;
 877                 }
 878                 break;
 879         }
 880
 881         /* The rest are paths */
 882         if (!seen_dashdash) {
 883                 int j;
 884                 for (j = i; j < argc; j++)
 885                         verify_filename(prefix, argv[j]);
 886         }
 887
 888         if (i < argc) {
 889                 paths = get_pathspec(prefix, argv + i);
 890                 if (opt.prefix_length && opt.relative) {
 891                         /* Make sure we do not get outside of paths */
 892                         for (i = 0; paths[i]; i++)
 893                                 if (strncmp(prefix, paths[i], opt.prefix_length))
 894                                         die("git grep: cannot generate relative filenames containing '..'");
 895                 }
 896         }
 897         else if (prefix) {
 898                 paths = xcalloc(2, sizeof(const char *));
 899                 paths[0] = prefix;
 900                 paths[1] = NULL;
 901         }
 902
 903         if (!list.nr) {
 904                 if (!cached)
 905                         setup_work_tree();
 906                 return !grep_cache(&opt, paths, cached, external_grep_allowed);
 907         }
 908
 909         if (cached)
 910                 die("both --cached and trees are given.");
 911
 912         for (i = 0; i < list.nr; i++) {
 913                 struct object *real_obj;
 914                 real_obj = deref_tag(list.objects[i].item, NULL, 0);
 915                 if (grep_object(&opt, paths, real_obj, list.objects[i].name))
 916                         hit = 1;
 917         }
 918         free_grep_patterns(&opt);
 919         return !hit;
 920 }