builtin/diff.c

   1 /*
   2  * Builtin "git diff"
   3  *
   4  * Copyright (c) 2006 Junio C Hamano
   5  */
   6 #define USE_THE_INDEX_COMPATIBILITY_MACROS
   7 #include "cache.h"
   8 #include "config.h"
   9 #include "ewah/ewok.h"
  10 #include "lockfile.h"
  11 #include "color.h"
  12 #include "commit.h"
  13 #include "blob.h"
  14 #include "tag.h"
  15 #include "diff.h"
  16 #include "diffcore.h"
  17 #include "revision.h"
  18 #include "log-tree.h"
  19 #include "builtin.h"
  20 #include "submodule.h"
  21 #include "oid-array.h"
  22
  23 #define DIFF_NO_INDEX_EXPLICIT 1
  24 #define DIFF_NO_INDEX_IMPLICIT 2
  25
  26 static const char builtin_diff_usage[] =
  27 "git diff [<options>] [<commit>] [--] [<path>...]\n"
  28 "   or: git diff [<options>] --cached [<commit>] [--] [<path>...]\n"
  29 "   or: git diff [<options>] <commit> [--merge-base] [<commit>...] <commit> [--] [<path>...]\n"
  30 "   or: git diff [<options>] <commit>...<commit>] [--] [<path>...]\n"
  31 "   or: git diff [<options>] <blob> <blob>]\n"
  32 "   or: git diff [<options>] --no-index [--] <path> <path>]\n"
  33 COMMON_DIFF_OPTIONS_HELP;
  34
  35 static const char *blob_path(struct object_array_entry *entry)
  36 {
  37         return entry->path ? entry->path : entry->name;
  38 }
  39
  40 static void stuff_change(struct diff_options *opt,
  41                          unsigned old_mode, unsigned new_mode,
  42                          const struct object_id *old_oid,
  43                          const struct object_id *new_oid,
  44                          int old_oid_valid,
  45                          int new_oid_valid,
  46                          const char *old_path,
  47                          const char *new_path)
  48 {
  49         struct diff_filespec *one, *two;
  50
  51         if (!is_null_oid(old_oid) && !is_null_oid(new_oid) &&
  52             oideq(old_oid, new_oid) && (old_mode == new_mode))
  53                 return;
  54
  55         if (opt->flags.reverse_diff) {
  56                 SWAP(old_mode, new_mode);
  57                 SWAP(old_oid, new_oid);
  58                 SWAP(old_path, new_path);
  59         }
  60
  61         if (opt->prefix &&
  62             (strncmp(old_path, opt->prefix, opt->prefix_length) ||
  63              strncmp(new_path, opt->prefix, opt->prefix_length)))
  64                 return;
  65
  66         one = alloc_filespec(old_path);
  67         two = alloc_filespec(new_path);
  68         fill_filespec(one, old_oid, old_oid_valid, old_mode);
  69         fill_filespec(two, new_oid, new_oid_valid, new_mode);
  70
  71         diff_queue(&diff_queued_diff, one, two);
  72 }
  73
  74 static int builtin_diff_b_f(struct rev_info *revs,
  75                             int argc, const char **argv,
  76                             struct object_array_entry **blob)
  77 {
  78         /* Blob vs file in the working tree*/
  79         struct stat st;
  80         const char *path;
  81
  82         if (argc > 1)
  83                 usage(builtin_diff_usage);
  84
  85         GUARD_PATHSPEC(&revs->prune_data, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
  86         path = revs->prune_data.items[0].match;
  87
  88         if (lstat(path, &st))
  89                 die_errno(_("failed to stat '%s'"), path);
  90         if (!(S_ISREG(st.st_mode) || S_ISLNK(st.st_mode)))
  91                 die(_("'%s': not a regular file or symlink"), path);
  92
  93         diff_set_mnemonic_prefix(&revs->diffopt, "o/", "w/");
  94
  95         if (blob[0]->mode == S_IFINVALID)
  96                 blob[0]->mode = canon_mode(st.st_mode);
  97
  98         stuff_change(&revs->diffopt,
  99                      blob[0]->mode, canon_mode(st.st_mode),
 100                      &blob[0]->item->oid, &null_oid,
 101                      1, 0,
 102                      blob[0]->path ? blob[0]->path : path,
 103                      path);
 104         diffcore_std(&revs->diffopt);
 105         diff_flush(&revs->diffopt);
 106         return 0;
 107 }
 108
 109 static int builtin_diff_blobs(struct rev_info *revs,
 110                               int argc, const char **argv,
 111                               struct object_array_entry **blob)
 112 {
 113         const unsigned mode = canon_mode(S_IFREG | 0644);
 114
 115         if (argc > 1)
 116                 usage(builtin_diff_usage);
 117
 118         if (blob[0]->mode == S_IFINVALID)
 119                 blob[0]->mode = mode;
 120
 121         if (blob[1]->mode == S_IFINVALID)
 122                 blob[1]->mode = mode;
 123
 124         stuff_change(&revs->diffopt,
 125                      blob[0]->mode, blob[1]->mode,
 126                      &blob[0]->item->oid, &blob[1]->item->oid,
 127                      1, 1,
 128                      blob_path(blob[0]), blob_path(blob[1]));
 129         diffcore_std(&revs->diffopt);
 130         diff_flush(&revs->diffopt);
 131         return 0;
 132 }
 133
 134 static int builtin_diff_index(struct rev_info *revs,
 135                               int argc, const char **argv)
 136 {
 137         unsigned int option = 0;
 138         while (1 < argc) {
 139                 const char *arg = argv[1];
 140                 if (!strcmp(arg, "--cached") || !strcmp(arg, "--staged"))
 141                         option |= DIFF_INDEX_CACHED;
 142                 else if (!strcmp(arg, "--merge-base"))
 143                         option |= DIFF_INDEX_MERGE_BASE;
 144                 else
 145                         usage(builtin_diff_usage);
 146                 argv++; argc--;
 147         }
 148         /*
 149          * Make sure there is one revision (i.e. pending object),
 150          * and there is no revision filtering parameters.
 151          */
 152         if (revs->pending.nr != 1 ||
 153             revs->max_count != -1 || revs->min_age != -1 ||
 154             revs->max_age != -1)
 155                 usage(builtin_diff_usage);
 156         if (!(option & DIFF_INDEX_CACHED)) {
 157                 setup_work_tree();
 158                 if (read_cache_preload(&revs->diffopt.pathspec) < 0) {
 159                         perror("read_cache_preload");
 160                         return -1;
 161                 }
 162         } else if (read_cache() < 0) {
 163                 perror("read_cache");
 164                 return -1;
 165         }
 166         return run_diff_index(revs, option);
 167 }
 168
 169 static int builtin_diff_tree(struct rev_info *revs,
 170                              int argc, const char **argv,
 171                              struct object_array_entry *ent0,
 172                              struct object_array_entry *ent1)
 173 {
 174         const struct object_id *(oid[2]);
 175         struct object_id mb_oid;
 176         int merge_base = 0;
 177
 178         while (1 < argc) {
 179                 const char *arg = argv[1];
 180                 if (!strcmp(arg, "--merge-base"))
 181                         merge_base = 1;
 182                 else
 183                         usage(builtin_diff_usage);
 184                 argv++; argc--;
 185         }
 186
 187         if (merge_base) {
 188                 diff_get_merge_base(revs, &mb_oid);
 189                 oid[0] = &mb_oid;
 190                 oid[1] = &revs->pending.objects[1].item->oid;
 191         } else {
 192                 int swap = 0;
 193
 194                 /*
 195                  * We saw two trees, ent0 and ent1.  If ent1 is uninteresting,
 196                  * swap them.
 197                  */
 198                 if (ent1->item->flags & UNINTERESTING)
 199                         swap = 1;
 200                 oid[swap] = &ent0->item->oid;
 201                 oid[1 - swap] = &ent1->item->oid;
 202         }
 203         diff_tree_oid(oid[0], oid[1], "", &revs->diffopt);
 204         log_tree_diff_flush(revs);
 205         return 0;
 206 }
 207
 208 static int builtin_diff_combined(struct rev_info *revs,
 209                                  int argc, const char **argv,
 210                                  struct object_array_entry *ent,
 211                                  int ents)
 212 {
 213         struct oid_array parents = OID_ARRAY_INIT;
 214         int i;
 215
 216         if (argc > 1)
 217                 usage(builtin_diff_usage);
 218
 219         if (!revs->dense_combined_merges && !revs->combine_merges)
 220                 revs->dense_combined_merges = revs->combine_merges = 1;
 221         for (i = 1; i < ents; i++)
 222                 oid_array_append(&parents, &ent[i].item->oid);
 223         diff_tree_combined(&ent[0].item->oid, &parents, revs);
 224         oid_array_clear(&parents);
 225         return 0;
 226 }
 227
 228 static void refresh_index_quietly(void)
 229 {
 230         struct lock_file lock_file = LOCK_INIT;
 231         int fd;
 232
 233         fd = hold_locked_index(&lock_file, 0);
 234         if (fd < 0)
 235                 return;
 236         discard_cache();
 237         read_cache();
 238         refresh_cache(REFRESH_QUIET|REFRESH_UNMERGED);
 239         repo_update_index_if_able(the_repository, &lock_file);
 240 }
 241
 242 static int builtin_diff_files(struct rev_info *revs, int argc, const char **argv)
 243 {
 244         unsigned int options = 0;
 245
 246         while (1 < argc && argv[1][0] == '-') {
 247                 if (!strcmp(argv[1], "--base"))
 248                         revs->max_count = 1;
 249                 else if (!strcmp(argv[1], "--ours"))
 250                         revs->max_count = 2;
 251                 else if (!strcmp(argv[1], "--theirs"))
 252                         revs->max_count = 3;
 253                 else if (!strcmp(argv[1], "-q"))
 254                         options |= DIFF_SILENT_ON_REMOVED;
 255                 else if (!strcmp(argv[1], "-h"))
 256                         usage(builtin_diff_usage);
 257                 else
 258                         return error(_("invalid option: %s"), argv[1]);
 259                 argv++; argc--;
 260         }
 261
 262         /*
 263          * "diff --base" should not combine merges because it was not
 264          * asked to.  "diff -c" should not densify (if the user wants
 265          * dense one, --cc can be explicitly asked for, or just rely
 266          * on the default).
 267          */
 268         if (revs->max_count == -1 && !revs->combine_merges &&
 269             (revs->diffopt.output_format & DIFF_FORMAT_PATCH))
 270                 revs->combine_merges = revs->dense_combined_merges = 1;
 271
 272         setup_work_tree();
 273         if (read_cache_preload(&revs->diffopt.pathspec) < 0) {
 274                 perror("read_cache_preload");
 275                 return -1;
 276         }
 277         return run_diff_files(revs, options);
 278 }
 279
 280 struct symdiff {
 281         struct bitmap *skip;
 282         int warn;
 283         const char *base, *left, *right;
 284 };
 285
 286 /*
 287  * Check for symmetric-difference arguments, and if present, arrange
 288  * everything we need to know to handle them correctly.  As a bonus,
 289  * weed out all bogus range-based revision specifications, e.g.,
 290  * "git diff A..B C..D" or "git diff A..B C" get rejected.
 291  *
 292  * For an actual symmetric diff, *symdiff is set this way:
 293  *
 294  *  - its skip is non-NULL and marks *all* rev->pending.objects[i]
 295  *    indices that the caller should ignore (extra merge bases, of
 296  *    which there might be many, and A in A...B).  Note that the
 297  *    chosen merge base and right side are NOT marked.
 298  *  - warn is set if there are multiple merge bases.
 299  *  - base, left, and right point to the names to use in a
 300  *    warning about multiple merge bases.
 301  *
 302  * If there is no symmetric diff argument, sym->skip is NULL and
 303  * sym->warn is cleared.  The remaining fields are not set.
 304  */
 305 static void symdiff_prepare(struct rev_info *rev, struct symdiff *sym)
 306 {
 307         int i, is_symdiff = 0, basecount = 0, othercount = 0;
 308         int lpos = -1, rpos = -1, basepos = -1;
 309         struct bitmap *map = NULL;
 310
 311         /*
 312          * Use the whence fields to find merge bases and left and
 313          * right parts of symmetric difference, so that we do not
 314          * depend on the order that revisions are parsed.  If there
 315          * are any revs that aren't from these sources, we have a
 316          * "git diff C A...B" or "git diff A...B C" case.  Or we
 317          * could even get "git diff A...B C...E", for instance.
 318          *
 319          * If we don't have just one merge base, we pick one
 320          * at random.
 321          *
 322          * NB: REV_CMD_LEFT, REV_CMD_RIGHT are also used for A..B,
 323          * so we must check for SYMMETRIC_LEFT too.  The two arrays
 324          * rev->pending.objects and rev->cmdline.rev are parallel.
 325          */
 326         for (i = 0; i < rev->cmdline.nr; i++) {
 327                 struct object *obj = rev->pending.objects[i].item;
 328                 switch (rev->cmdline.rev[i].whence) {
 329                 case REV_CMD_MERGE_BASE:
 330                         if (basepos < 0)
 331                                 basepos = i;
 332                         basecount++;
 333                         break;          /* do mark all bases */
 334                 case REV_CMD_LEFT:
 335                         if (lpos >= 0)
 336                                 usage(builtin_diff_usage);
 337                         lpos = i;
 338                         if (obj->flags & SYMMETRIC_LEFT) {
 339                                 is_symdiff = 1;
 340                                 break;  /* do mark A */
 341                         }
 342                         continue;
 343                 case REV_CMD_RIGHT:
 344                         if (rpos >= 0)
 345                                 usage(builtin_diff_usage);
 346                         rpos = i;
 347                         continue;       /* don't mark B */
 348                 case REV_CMD_PARENTS_ONLY:
 349                 case REV_CMD_REF:
 350                 case REV_CMD_REV:
 351                         othercount++;
 352                         continue;
 353                 }
 354                 if (map == NULL)
 355                         map = bitmap_new();
 356                 bitmap_set(map, i);
 357         }
 358
 359         /*
 360          * Forbid any additional revs for both A...B and A..B.
 361          */
 362         if (lpos >= 0 && othercount > 0)
 363                 usage(builtin_diff_usage);
 364
 365         if (!is_symdiff) {
 366                 bitmap_free(map);
 367                 sym->warn = 0;
 368                 sym->skip = NULL;
 369                 return;
 370         }
 371
 372         sym->left = rev->pending.objects[lpos].name;
 373         sym->right = rev->pending.objects[rpos].name;
 374         if (basecount == 0)
 375                 die(_("%s...%s: no merge base"), sym->left, sym->right);
 376         sym->base = rev->pending.objects[basepos].name;
 377         bitmap_unset(map, basepos);     /* unmark the base we want */
 378         sym->warn = basecount > 1;
 379         sym->skip = map;
 380 }
 381
 382 int cmd_diff(int argc, const char **argv, const char *prefix)
 383 {
 384         int i;
 385         struct rev_info rev;
 386         struct object_array ent = OBJECT_ARRAY_INIT;
 387         int blobs = 0, paths = 0;
 388         struct object_array_entry *blob[2];
 389         int nongit = 0, no_index = 0;
 390         int result = 0;
 391         struct symdiff sdiff;
 392
 393         /*
 394          * We could get N tree-ish in the rev.pending_objects list.
 395          * Also there could be M blobs there, and P pathspecs. --cached may
 396          * also be present.
 397          *
 398          * N=0, M=0:
 399          *      cache vs files (diff-files)
 400          *
 401          * N=0, M=0, --cached:
 402          *      HEAD vs cache (diff-index --cached)
 403          *
 404          * N=0, M=2:
 405          *      compare two random blobs.  P must be zero.
 406          *
 407          * N=0, M=1, P=1:
 408          *      compare a blob with a working tree file.
 409          *
 410          * N=1, M=0:
 411          *      tree vs files (diff-index)
 412          *
 413          * N=1, M=0, --cached:
 414          *      tree vs cache (diff-index --cached)
 415          *
 416          * N=2, M=0:
 417          *      tree vs tree (diff-tree)
 418          *
 419          * N=0, M=0, P=2:
 420          *      compare two filesystem entities (aka --no-index).
 421          *
 422          * Other cases are errors.
 423          */
 424
 425         /* Were we asked to do --no-index explicitly? */
 426         for (i = 1; i < argc; i++) {
 427                 if (!strcmp(argv[i], "--")) {
 428                         i++;
 429                         break;
 430                 }
 431                 if (!strcmp(argv[i], "--no-index"))
 432                         no_index = DIFF_NO_INDEX_EXPLICIT;
 433                 if (argv[i][0] != '-')
 434                         break;
 435         }
 436
 437         prefix = setup_git_directory_gently(&nongit);
 438
 439         if (!no_index) {
 440                 /*
 441                  * Treat git diff with at least one path outside of the
 442                  * repo the same as if the command would have been executed
 443                  * outside of a git repository.  In this case it behaves
 444                  * the same way as "git diff --no-index <a> <b>", which acts
 445                  * as a colourful "diff" replacement.
 446                  */
 447                 if (nongit || ((argc == i + 2) &&
 448                                (!path_inside_repo(prefix, argv[i]) ||
 449                                 !path_inside_repo(prefix, argv[i + 1]))))
 450                         no_index = DIFF_NO_INDEX_IMPLICIT;
 451         }
 452
 453         init_diff_ui_defaults();
 454         git_config(git_diff_ui_config, NULL);
 455         precompose_argv(argc, argv);
 456
 457         repo_init_revisions(the_repository, &rev, prefix);
 458
 459         /* Set up defaults that will apply to both no-index and regular diffs. */
 460         rev.diffopt.stat_width = -1;
 461         rev.diffopt.stat_graph_width = -1;
 462         rev.diffopt.flags.allow_external = 1;
 463         rev.diffopt.flags.allow_textconv = 1;
 464
 465         /* If this is a no-index diff, just run it and exit there. */
 466         if (no_index)
 467                 exit(diff_no_index(&rev, no_index == DIFF_NO_INDEX_IMPLICIT,
 468                                    argc, argv));
 469
 470
 471         /*
 472          * Otherwise, we are doing the usual "git" diff; set up any
 473          * further defaults that apply to regular diffs.
 474          */
 475         rev.diffopt.skip_stat_unmatch = !!diff_auto_refresh_index;
 476
 477         /*
 478          * Default to intent-to-add entries invisible in the
 479          * index. This makes them show up as new files in diff-files
 480          * and not at all in diff-cached.
 481          */
 482         rev.diffopt.ita_invisible_in_index = 1;
 483
 484         if (nongit)
 485                 die(_("Not a git repository"));
 486         argc = setup_revisions(argc, argv, &rev, NULL);
 487         if (!rev.diffopt.output_format) {
 488                 rev.diffopt.output_format = DIFF_FORMAT_PATCH;
 489                 diff_setup_done(&rev.diffopt);
 490         }
 491
 492         rev.diffopt.flags.recursive = 1;
 493
 494         setup_diff_pager(&rev.diffopt);
 495
 496         /*
 497          * Do we have --cached and not have a pending object, then
 498          * default to HEAD by hand.  Eek.
 499          */
 500         if (!rev.pending.nr) {
 501                 int i;
 502                 for (i = 1; i < argc; i++) {
 503                         const char *arg = argv[i];
 504                         if (!strcmp(arg, "--"))
 505                                 break;
 506                         else if (!strcmp(arg, "--cached") ||
 507                                  !strcmp(arg, "--staged")) {
 508                                 add_head_to_pending(&rev);
 509                                 if (!rev.pending.nr) {
 510                                         struct tree *tree;
 511                                         tree = lookup_tree(the_repository,
 512                                                            the_repository->hash_algo->empty_tree);
 513                                         add_pending_object(&rev, &tree->object, "HEAD");
 514                                 }
 515                                 break;
 516                         }
 517                 }
 518         }
 519
 520         symdiff_prepare(&rev, &sdiff);
 521         for (i = 0; i < rev.pending.nr; i++) {
 522                 struct object_array_entry *entry = &rev.pending.objects[i];
 523                 struct object *obj = entry->item;
 524                 const char *name = entry->name;
 525                 int flags = (obj->flags & UNINTERESTING);
 526                 if (!obj->parsed)
 527                         obj = parse_object(the_repository, &obj->oid);
 528                 obj = deref_tag(the_repository, obj, NULL, 0);
 529                 if (!obj)
 530                         die(_("invalid object '%s' given."), name);
 531                 if (obj->type == OBJ_COMMIT)
 532                         obj = &get_commit_tree(((struct commit *)obj))->object;
 533
 534                 if (obj->type == OBJ_TREE) {
 535                         if (sdiff.skip && bitmap_get(sdiff.skip, i))
 536                                 continue;
 537                         obj->flags |= flags;
 538                         add_object_array(obj, name, &ent);
 539                 } else if (obj->type == OBJ_BLOB) {
 540                         if (2 <= blobs)
 541                                 die(_("more than two blobs given: '%s'"), name);
 542                         blob[blobs] = entry;
 543                         blobs++;
 544
 545                 } else {
 546                         die(_("unhandled object '%s' given."), name);
 547                 }
 548         }
 549         if (rev.prune_data.nr)
 550                 paths += rev.prune_data.nr;
 551
 552         /*
 553          * Now, do the arguments look reasonable?
 554          */
 555         if (!ent.nr) {
 556                 switch (blobs) {
 557                 case 0:
 558                         result = builtin_diff_files(&rev, argc, argv);
 559                         break;
 560                 case 1:
 561                         if (paths != 1)
 562                                 usage(builtin_diff_usage);
 563                         result = builtin_diff_b_f(&rev, argc, argv, blob);
 564                         break;
 565                 case 2:
 566                         if (paths)
 567                                 usage(builtin_diff_usage);
 568                         result = builtin_diff_blobs(&rev, argc, argv, blob);
 569                         break;
 570                 default:
 571                         usage(builtin_diff_usage);
 572                 }
 573         }
 574         else if (blobs)
 575                 usage(builtin_diff_usage);
 576         else if (ent.nr == 1)
 577                 result = builtin_diff_index(&rev, argc, argv);
 578         else if (ent.nr == 2) {
 579                 if (sdiff.warn)
 580                         warning(_("%s...%s: multiple merge bases, using %s"),
 581                                 sdiff.left, sdiff.right, sdiff.base);
 582                 result = builtin_diff_tree(&rev, argc, argv,
 583                                            &ent.objects[0], &ent.objects[1]);
 584         } else
 585                 result = builtin_diff_combined(&rev, argc, argv,
 586                                                ent.objects, ent.nr);
 587         result = diff_result_code(&rev.diffopt, result);
 588         if (1 < rev.diffopt.skip_stat_unmatch)
 589                 refresh_index_quietly();
 590         UNLEAK(rev);
 591         UNLEAK(ent);
 592         UNLEAK(blob);
 593         return result;
 594 }