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