builtin/diff.c

   1 /*
   2  * Builtin "git diff"
   3  *
   4  * Copyright (c) 2006 Junio C Hamano
   5  */
   6 #define USE_THE_INDEX_VARIABLE
   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 "gettext.h"
  15 #include "tag.h"
  16 #include "diff.h"
  17 #include "diff-merges.h"
  18 #include "diffcore.h"
  19 #include "revision.h"
  20 #include "log-tree.h"
  21 #include "builtin.h"
  22 #include "setup.h"
  23 #include "submodule.h"
  24 #include "oid-array.h"
  25 #include "tree.h"
  26
  27 #define DIFF_NO_INDEX_EXPLICIT 1
  28 #define DIFF_NO_INDEX_IMPLICIT 2
  29
  30 static const char builtin_diff_usage[] =
  31 "git diff [<options>] [<commit>] [--] [<path>...]\n"
  32 "   or: git diff [<options>] --cached [--merge-base] [<commit>] [--] [<path>...]\n"
  33 "   or: git diff [<options>] [--merge-base] <commit> [<commit>...] <commit> [--] [<path>...]\n"
  34 "   or: git diff [<options>] <commit>...<commit> [--] [<path>...]\n"
  35 "   or: git diff [<options>] <blob> <blob>\n"
  36 "   or: git diff [<options>] --no-index [--] <path> <path>"
  37 "\n"
  38 COMMON_DIFF_OPTIONS_HELP;
  39
  40 static const char *blob_path(struct object_array_entry *entry)
  41 {
  42         return entry->path ? entry->path : entry->name;
  43 }
  44
  45 static void stuff_change(struct diff_options *opt,
  46                          unsigned old_mode, unsigned new_mode,
  47                          const struct object_id *old_oid,
  48                          const struct object_id *new_oid,
  49                          int old_oid_valid,
  50                          int new_oid_valid,
  51                          const char *old_path,
  52                          const char *new_path)
  53 {
  54         struct diff_filespec *one, *two;
  55
  56         if (!is_null_oid(old_oid) && !is_null_oid(new_oid) &&
  57             oideq(old_oid, new_oid) && (old_mode == new_mode))
  58                 return;
  59
  60         if (opt->flags.reverse_diff) {
  61                 SWAP(old_mode, new_mode);
  62                 SWAP(old_oid, new_oid);
  63                 SWAP(old_path, new_path);
  64         }
  65
  66         if (opt->prefix &&
  67             (strncmp(old_path, opt->prefix, opt->prefix_length) ||
  68              strncmp(new_path, opt->prefix, opt->prefix_length)))
  69                 return;
  70
  71         one = alloc_filespec(old_path);
  72         two = alloc_filespec(new_path);
  73         fill_filespec(one, old_oid, old_oid_valid, old_mode);
  74         fill_filespec(two, new_oid, new_oid_valid, new_mode);
  75
  76         diff_queue(&diff_queued_diff, one, two);
  77 }
  78
  79 static int builtin_diff_b_f(struct rev_info *revs,
  80                             int argc, const char **argv UNUSED,
  81                             struct object_array_entry **blob)
  82 {
  83         /* Blob vs file in the working tree*/
  84         struct stat st;
  85         const char *path;
  86
  87         if (argc > 1)
  88                 usage(builtin_diff_usage);
  89
  90         GUARD_PATHSPEC(&revs->prune_data, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
  91         path = revs->prune_data.items[0].match;
  92
  93         if (lstat(path, &st))
  94                 die_errno(_("failed to stat '%s'"), path);
  95         if (!(S_ISREG(st.st_mode) || S_ISLNK(st.st_mode)))
  96                 die(_("'%s': not a regular file or symlink"), path);
  97
  98         diff_set_mnemonic_prefix(&revs->diffopt, "o/", "w/");
  99
 100         if (blob[0]->mode == S_IFINVALID)
 101                 blob[0]->mode = canon_mode(st.st_mode);
 102
 103         stuff_change(&revs->diffopt,
 104                      blob[0]->mode, canon_mode(st.st_mode),
 105                      &blob[0]->item->oid, null_oid(),
 106                      1, 0,
 107                      blob[0]->path ? blob[0]->path : path,
 108                      path);
 109         diffcore_std(&revs->diffopt);
 110         diff_flush(&revs->diffopt);
 111         return 0;
 112 }
 113
 114 static int builtin_diff_blobs(struct rev_info *revs,
 115                               int argc, const char **argv UNUSED,
 116                               struct object_array_entry **blob)
 117 {
 118         const unsigned mode = canon_mode(S_IFREG | 0644);
 119
 120         if (argc > 1)
 121                 usage(builtin_diff_usage);
 122
 123         if (blob[0]->mode == S_IFINVALID)
 124                 blob[0]->mode = mode;
 125
 126         if (blob[1]->mode == S_IFINVALID)
 127                 blob[1]->mode = mode;
 128
 129         stuff_change(&revs->diffopt,
 130                      blob[0]->mode, blob[1]->mode,
 131                      &blob[0]->item->oid, &blob[1]->item->oid,
 132                      1, 1,
 133                      blob_path(blob[0]), blob_path(blob[1]));
 134         diffcore_std(&revs->diffopt);
 135         diff_flush(&revs->diffopt);
 136         return 0;
 137 }
 138
 139 static int builtin_diff_index(struct rev_info *revs,
 140                               int argc, const char **argv)
 141 {
 142         unsigned int option = 0;
 143         while (1 < argc) {
 144                 const char *arg = argv[1];
 145                 if (!strcmp(arg, "--cached") || !strcmp(arg, "--staged"))
 146                         option |= DIFF_INDEX_CACHED;
 147                 else if (!strcmp(arg, "--merge-base"))
 148                         option |= DIFF_INDEX_MERGE_BASE;
 149                 else
 150                         usage(builtin_diff_usage);
 151                 argv++; argc--;
 152         }
 153         /*
 154          * Make sure there is one revision (i.e. pending object),
 155          * and there is no revision filtering parameters.
 156          */
 157         if (revs->pending.nr != 1 ||
 158             revs->max_count != -1 || revs->min_age != -1 ||
 159             revs->max_age != -1)
 160                 usage(builtin_diff_usage);
 161         if (!(option & DIFF_INDEX_CACHED)) {
 162                 setup_work_tree();
 163                 if (repo_read_index_preload(the_repository,
 164                                             &revs->diffopt.pathspec, 0) < 0) {
 165                         perror("repo_read_index_preload");
 166                         return -1;
 167                 }
 168         } else if (repo_read_index(the_repository) < 0) {
 169                 perror("repo_read_cache");
 170                 return -1;
 171         }
 172         return run_diff_index(revs, option);
 173 }
 174
 175 static int builtin_diff_tree(struct rev_info *revs,
 176                              int argc, const char **argv,
 177                              struct object_array_entry *ent0,
 178                              struct object_array_entry *ent1)
 179 {
 180         const struct object_id *(oid[2]);
 181         struct object_id mb_oid;
 182         int merge_base = 0;
 183
 184         while (1 < argc) {
 185                 const char *arg = argv[1];
 186                 if (!strcmp(arg, "--merge-base"))
 187                         merge_base = 1;
 188                 else
 189                         usage(builtin_diff_usage);
 190                 argv++; argc--;
 191         }
 192
 193         if (merge_base) {
 194                 diff_get_merge_base(revs, &mb_oid);
 195                 oid[0] = &mb_oid;
 196                 oid[1] = &revs->pending.objects[1].item->oid;
 197         } else {
 198                 int swap = 0;
 199
 200                 /*
 201                  * We saw two trees, ent0 and ent1.  If ent1 is uninteresting,
 202                  * swap them.
 203                  */
 204                 if (ent1->item->flags & UNINTERESTING)
 205                         swap = 1;
 206                 oid[swap] = &ent0->item->oid;
 207                 oid[1 - swap] = &ent1->item->oid;
 208         }
 209         diff_tree_oid(oid[0], oid[1], "", &revs->diffopt);
 210         log_tree_diff_flush(revs);
 211         return 0;
 212 }
 213
 214 static int builtin_diff_combined(struct rev_info *revs,
 215                                  int argc, const char **argv UNUSED,
 216                                  struct object_array_entry *ent,
 217                                  int ents, int first_non_parent)
 218 {
 219         struct oid_array parents = OID_ARRAY_INIT;
 220         int i;
 221
 222         if (argc > 1)
 223                 usage(builtin_diff_usage);
 224
 225         if (first_non_parent < 0)
 226                 die(_("no merge given, only parents."));
 227         if (first_non_parent >= ents)
 228                 BUG("first_non_parent out of range: %d", first_non_parent);
 229
 230         diff_merges_set_dense_combined_if_unset(revs);
 231
 232         for (i = 0; i < ents; i++) {
 233                 if (i != first_non_parent)
 234                         oid_array_append(&parents, &ent[i].item->oid);
 235         }
 236         diff_tree_combined(&ent[first_non_parent].item->oid, &parents, revs);
 237         oid_array_clear(&parents);
 238         return 0;
 239 }
 240
 241 static void refresh_index_quietly(void)
 242 {
 243         struct lock_file lock_file = LOCK_INIT;
 244         int fd;
 245
 246         fd = repo_hold_locked_index(the_repository, &lock_file, 0);
 247         if (fd < 0)
 248                 return;
 249         discard_index(&the_index);
 250         repo_read_index(the_repository);
 251         refresh_index(&the_index, REFRESH_QUIET|REFRESH_UNMERGED, NULL, NULL,
 252                       NULL);
 253         repo_update_index_if_able(the_repository, &lock_file);
 254 }
 255
 256 static int builtin_diff_files(struct rev_info *revs, int argc, const char **argv)
 257 {
 258         unsigned int options = 0;
 259
 260         while (1 < argc && argv[1][0] == '-') {
 261                 if (!strcmp(argv[1], "--base"))
 262                         revs->max_count = 1;
 263                 else if (!strcmp(argv[1], "--ours"))
 264                         revs->max_count = 2;
 265                 else if (!strcmp(argv[1], "--theirs"))
 266                         revs->max_count = 3;
 267                 else if (!strcmp(argv[1], "-q"))
 268                         options |= DIFF_SILENT_ON_REMOVED;
 269                 else if (!strcmp(argv[1], "-h"))
 270                         usage(builtin_diff_usage);
 271                 else
 272                         return error(_("invalid option: %s"), argv[1]);
 273                 argv++; argc--;
 274         }
 275
 276         /*
 277          * "diff --base" should not combine merges because it was not
 278          * asked to.  "diff -c" should not densify (if the user wants
 279          * dense one, --cc can be explicitly asked for, or just rely
 280          * on the default).
 281          */
 282         if (revs->max_count == -1 &&
 283             (revs->diffopt.output_format & DIFF_FORMAT_PATCH))
 284                 diff_merges_set_dense_combined_if_unset(revs);
 285
 286         setup_work_tree();
 287         if (repo_read_index_preload(the_repository, &revs->diffopt.pathspec,
 288                                     0) < 0) {
 289                 perror("repo_read_index_preload");
 290                 return -1;
 291         }
 292         return run_diff_files(revs, options);
 293 }
 294
 295 struct symdiff {
 296         struct bitmap *skip;
 297         int warn;
 298         const char *base, *left, *right;
 299 };
 300
 301 /*
 302  * Check for symmetric-difference arguments, and if present, arrange
 303  * everything we need to know to handle them correctly.  As a bonus,
 304  * weed out all bogus range-based revision specifications, e.g.,
 305  * "git diff A..B C..D" or "git diff A..B C" get rejected.
 306  *
 307  * For an actual symmetric diff, *symdiff is set this way:
 308  *
 309  *  - its skip is non-NULL and marks *all* rev->pending.objects[i]
 310  *    indices that the caller should ignore (extra merge bases, of
 311  *    which there might be many, and A in A...B).  Note that the
 312  *    chosen merge base and right side are NOT marked.
 313  *  - warn is set if there are multiple merge bases.
 314  *  - base, left, and right point to the names to use in a
 315  *    warning about multiple merge bases.
 316  *
 317  * If there is no symmetric diff argument, sym->skip is NULL and
 318  * sym->warn is cleared.  The remaining fields are not set.
 319  */
 320 static void symdiff_prepare(struct rev_info *rev, struct symdiff *sym)
 321 {
 322         int i, is_symdiff = 0, basecount = 0, othercount = 0;
 323         int lpos = -1, rpos = -1, basepos = -1;
 324         struct bitmap *map = NULL;
 325
 326         /*
 327          * Use the whence fields to find merge bases and left and
 328          * right parts of symmetric difference, so that we do not
 329          * depend on the order that revisions are parsed.  If there
 330          * are any revs that aren't from these sources, we have a
 331          * "git diff C A...B" or "git diff A...B C" case.  Or we
 332          * could even get "git diff A...B C...E", for instance.
 333          *
 334          * If we don't have just one merge base, we pick one
 335          * at random.
 336          *
 337          * NB: REV_CMD_LEFT, REV_CMD_RIGHT are also used for A..B,
 338          * so we must check for SYMMETRIC_LEFT too.  The two arrays
 339          * rev->pending.objects and rev->cmdline.rev are parallel.
 340          */
 341         for (i = 0; i < rev->cmdline.nr; i++) {
 342                 struct object *obj = rev->pending.objects[i].item;
 343                 switch (rev->cmdline.rev[i].whence) {
 344                 case REV_CMD_MERGE_BASE:
 345                         if (basepos < 0)
 346                                 basepos = i;
 347                         basecount++;
 348                         break;          /* do mark all bases */
 349                 case REV_CMD_LEFT:
 350                         if (lpos >= 0)
 351                                 usage(builtin_diff_usage);
 352                         lpos = i;
 353                         if (obj->flags & SYMMETRIC_LEFT) {
 354                                 is_symdiff = 1;
 355                                 break;  /* do mark A */
 356                         }
 357                         continue;
 358                 case REV_CMD_RIGHT:
 359                         if (rpos >= 0)
 360                                 usage(builtin_diff_usage);
 361                         rpos = i;
 362                         continue;       /* don't mark B */
 363                 case REV_CMD_PARENTS_ONLY:
 364                 case REV_CMD_REF:
 365                 case REV_CMD_REV:
 366                         othercount++;
 367                         continue;
 368                 }
 369                 if (!map)
 370                         map = bitmap_new();
 371                 bitmap_set(map, i);
 372         }
 373
 374         /*
 375          * Forbid any additional revs for both A...B and A..B.
 376          */
 377         if (lpos >= 0 && othercount > 0)
 378                 usage(builtin_diff_usage);
 379
 380         if (!is_symdiff) {
 381                 bitmap_free(map);
 382                 sym->warn = 0;
 383                 sym->skip = NULL;
 384                 return;
 385         }
 386
 387         sym->left = rev->pending.objects[lpos].name;
 388         sym->right = rev->pending.objects[rpos].name;
 389         if (basecount == 0)
 390                 die(_("%s...%s: no merge base"), sym->left, sym->right);
 391         sym->base = rev->pending.objects[basepos].name;
 392         bitmap_unset(map, basepos);     /* unmark the base we want */
 393         sym->warn = basecount > 1;
 394         sym->skip = map;
 395 }
 396
 397 int cmd_diff(int argc, const char **argv, const char *prefix)
 398 {
 399         int i;
 400         struct rev_info rev;
 401         struct object_array ent = OBJECT_ARRAY_INIT;
 402         int first_non_parent = -1;
 403         int blobs = 0, paths = 0;
 404         struct object_array_entry *blob[2];
 405         int nongit = 0, no_index = 0;
 406         int result = 0;
 407         struct symdiff sdiff;
 408
 409         /*
 410          * We could get N tree-ish in the rev.pending_objects list.
 411          * Also there could be M blobs there, and P pathspecs. --cached may
 412          * also be present.
 413          *
 414          * N=0, M=0:
 415          *      cache vs files (diff-files)
 416          *
 417          * N=0, M=0, --cached:
 418          *      HEAD vs cache (diff-index --cached)
 419          *
 420          * N=0, M=2:
 421          *      compare two random blobs.  P must be zero.
 422          *
 423          * N=0, M=1, P=1:
 424          *      compare a blob with a working tree file.
 425          *
 426          * N=1, M=0:
 427          *      tree vs files (diff-index)
 428          *
 429          * N=1, M=0, --cached:
 430          *      tree vs cache (diff-index --cached)
 431          *
 432          * N=2, M=0:
 433          *      tree vs tree (diff-tree)
 434          *
 435          * N=0, M=0, P=2:
 436          *      compare two filesystem entities (aka --no-index).
 437          *
 438          * Other cases are errors.
 439          */
 440
 441         /* Were we asked to do --no-index explicitly? */
 442         for (i = 1; i < argc; i++) {
 443                 if (!strcmp(argv[i], "--")) {
 444                         i++;
 445                         break;
 446                 }
 447                 if (!strcmp(argv[i], "--no-index"))
 448                         no_index = DIFF_NO_INDEX_EXPLICIT;
 449                 if (argv[i][0] != '-')
 450                         break;
 451         }
 452
 453         prefix = setup_git_directory_gently(&nongit);
 454
 455         if (!nongit) {
 456                 prepare_repo_settings(the_repository);
 457                 the_repository->settings.command_requires_full_index = 0;
 458         }
 459
 460         if (!no_index) {
 461                 /*
 462                  * Treat git diff with at least one path outside of the
 463                  * repo the same as if the command would have been executed
 464                  * outside of a git repository.  In this case it behaves
 465                  * the same way as "git diff --no-index <a> <b>", which acts
 466                  * as a colourful "diff" replacement.
 467                  */
 468                 if (nongit || ((argc == i + 2) &&
 469                                (!path_inside_repo(prefix, argv[i]) ||
 470                                 !path_inside_repo(prefix, argv[i + 1]))))
 471                         no_index = DIFF_NO_INDEX_IMPLICIT;
 472         }
 473
 474         init_diff_ui_defaults();
 475         git_config(git_diff_ui_config, NULL);
 476         prefix = precompose_argv_prefix(argc, argv, prefix);
 477
 478         repo_init_revisions(the_repository, &rev, prefix);
 479
 480         /* Set up defaults that will apply to both no-index and regular diffs. */
 481         rev.diffopt.stat_width = -1;
 482         rev.diffopt.stat_graph_width = -1;
 483         rev.diffopt.flags.allow_external = 1;
 484         rev.diffopt.flags.allow_textconv = 1;
 485
 486         /* If this is a no-index diff, just run it and exit there. */
 487         if (no_index)
 488                 exit(diff_no_index(&rev, no_index == DIFF_NO_INDEX_IMPLICIT,
 489                                    argc, argv));
 490
 491
 492         /*
 493          * Otherwise, we are doing the usual "git" diff; set up any
 494          * further defaults that apply to regular diffs.
 495          */
 496         rev.diffopt.skip_stat_unmatch = !!diff_auto_refresh_index;
 497
 498         /*
 499          * Default to intent-to-add entries invisible in the
 500          * index. This makes them show up as new files in diff-files
 501          * and not at all in diff-cached.
 502          */
 503         rev.diffopt.ita_invisible_in_index = 1;
 504
 505         if (nongit)
 506                 die(_("Not a git repository"));
 507         argc = setup_revisions(argc, argv, &rev, NULL);
 508         if (!rev.diffopt.output_format) {
 509                 rev.diffopt.output_format = DIFF_FORMAT_PATCH;
 510                 diff_setup_done(&rev.diffopt);
 511         }
 512
 513         rev.diffopt.flags.recursive = 1;
 514         rev.diffopt.rotate_to_strict = 1;
 515
 516         setup_diff_pager(&rev.diffopt);
 517
 518         /*
 519          * Do we have --cached and not have a pending object, then
 520          * default to HEAD by hand.  Eek.
 521          */
 522         if (!rev.pending.nr) {
 523                 int i;
 524                 for (i = 1; i < argc; i++) {
 525                         const char *arg = argv[i];
 526                         if (!strcmp(arg, "--"))
 527                                 break;
 528                         else if (!strcmp(arg, "--cached") ||
 529                                  !strcmp(arg, "--staged")) {
 530                                 add_head_to_pending(&rev);
 531                                 if (!rev.pending.nr) {
 532                                         struct tree *tree;
 533                                         tree = lookup_tree(the_repository,
 534                                                            the_repository->hash_algo->empty_tree);
 535                                         add_pending_object(&rev, &tree->object, "HEAD");
 536                                 }
 537                                 break;
 538                         }
 539                 }
 540         }
 541
 542         symdiff_prepare(&rev, &sdiff);
 543         for (i = 0; i < rev.pending.nr; i++) {
 544                 struct object_array_entry *entry = &rev.pending.objects[i];
 545                 struct object *obj = entry->item;
 546                 const char *name = entry->name;
 547                 int flags = (obj->flags & UNINTERESTING);
 548                 if (!obj->parsed)
 549                         obj = parse_object(the_repository, &obj->oid);
 550                 obj = deref_tag(the_repository, obj, NULL, 0);
 551                 if (!obj)
 552                         die(_("invalid object '%s' given."), name);
 553                 if (obj->type == OBJ_COMMIT)
 554                         obj = &repo_get_commit_tree(the_repository,
 555                                                     ((struct commit *)obj))->object;
 556
 557                 if (obj->type == OBJ_TREE) {
 558                         if (sdiff.skip && bitmap_get(sdiff.skip, i))
 559                                 continue;
 560                         obj->flags |= flags;
 561                         add_object_array(obj, name, &ent);
 562                         if (first_non_parent < 0 &&
 563                             (i >= rev.cmdline.nr || /* HEAD by hand. */
 564                              rev.cmdline.rev[i].whence != REV_CMD_PARENTS_ONLY))
 565                                 first_non_parent = ent.nr - 1;
 566                 } else if (obj->type == OBJ_BLOB) {
 567                         if (2 <= blobs)
 568                                 die(_("more than two blobs given: '%s'"), name);
 569                         blob[blobs] = entry;
 570                         blobs++;
 571
 572                 } else {
 573                         die(_("unhandled object '%s' given."), name);
 574                 }
 575         }
 576         if (rev.prune_data.nr)
 577                 paths += rev.prune_data.nr;
 578
 579         /*
 580          * Now, do the arguments look reasonable?
 581          */
 582         if (!ent.nr) {
 583                 switch (blobs) {
 584                 case 0:
 585                         result = builtin_diff_files(&rev, argc, argv);
 586                         break;
 587                 case 1:
 588                         if (paths != 1)
 589                                 usage(builtin_diff_usage);
 590                         result = builtin_diff_b_f(&rev, argc, argv, blob);
 591                         break;
 592                 case 2:
 593                         if (paths)
 594                                 usage(builtin_diff_usage);
 595                         result = builtin_diff_blobs(&rev, argc, argv, blob);
 596                         break;
 597                 default:
 598                         usage(builtin_diff_usage);
 599                 }
 600         }
 601         else if (blobs)
 602                 usage(builtin_diff_usage);
 603         else if (ent.nr == 1)
 604                 result = builtin_diff_index(&rev, argc, argv);
 605         else if (ent.nr == 2) {
 606                 if (sdiff.warn)
 607                         warning(_("%s...%s: multiple merge bases, using %s"),
 608                                 sdiff.left, sdiff.right, sdiff.base);
 609                 result = builtin_diff_tree(&rev, argc, argv,
 610                                            &ent.objects[0], &ent.objects[1]);
 611         } else
 612                 result = builtin_diff_combined(&rev, argc, argv,
 613                                                ent.objects, ent.nr,
 614                                                first_non_parent);
 615         result = diff_result_code(&rev.diffopt, result);
 616         if (1 < rev.diffopt.skip_stat_unmatch)
 617                 refresh_index_quietly();
 618         release_revisions(&rev);
 619         object_array_clear(&ent);
 620         UNLEAK(blob);
 621         return result;
 622 }