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