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