builtin/diff.c

   1 /*
   2  * Builtin "git diff"
   3  *
   4  * Copyright (c) 2006 Junio C Hamano
   5  */
   6
   7 #include "builtin.h"
   8 #include "config.h"
   9 #include "ewah/ewok.h"
  10 #include "lockfile.h"
  11 #include "color.h"
  12 #include "commit.h"
  13 #include "gettext.h"
  14 #include "tag.h"
  15 #include "diff.h"
  16 #include "diff-merges.h"
  17 #include "diffcore.h"
  18 #include "preload-index.h"
  19 #include "read-cache-ll.h"
  20 #include "revision.h"
  21 #include "log-tree.h"
  22 #include "setup.h"
  23 #include "oid-array.h"
  24 #include "tree.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 void 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 }
 111
 112 static void builtin_diff_blobs(struct rev_info *revs,
 113                                int argc, const char **argv UNUSED,
 114                                struct object_array_entry **blob)
 115 {
 116         const unsigned mode = canon_mode(S_IFREG | 0644);
 117
 118         if (argc > 1)
 119                 usage(builtin_diff_usage);
 120
 121         if (blob[0]->mode == S_IFINVALID)
 122                 blob[0]->mode = mode;
 123
 124         if (blob[1]->mode == S_IFINVALID)
 125                 blob[1]->mode = mode;
 126
 127         stuff_change(&revs->diffopt,
 128                      blob[0]->mode, blob[1]->mode,
 129                      &blob[0]->item->oid, &blob[1]->item->oid,
 130                      1, 1,
 131                      blob_path(blob[0]), blob_path(blob[1]));
 132         diffcore_std(&revs->diffopt);
 133         diff_flush(&revs->diffopt);
 134 }
 135
 136 static void 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                         die_errno("repo_read_index_preload");
 163                 }
 164         } else if (repo_read_index(the_repository) < 0) {
 165                 die_errno("repo_read_cache");
 166         }
 167         run_diff_index(revs, option);
 168 }
 169
 170 static void 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 }
 207
 208 static void builtin_diff_combined(struct rev_info *revs,
 209                                   int argc, const char **argv UNUSED,
 210                                   struct object_array_entry *ent,
 211                                   int ents, int first_non_parent)
 212 {
 213         struct oid_array parents = OID_ARRAY_INIT;
 214         int i;
 215
 216         if (argc > 1)
 217                 usage(builtin_diff_usage);
 218
 219         if (first_non_parent < 0)
 220                 die(_("no merge given, only parents."));
 221         if (first_non_parent >= ents)
 222                 BUG("first_non_parent out of range: %d", first_non_parent);
 223
 224         diff_merges_set_dense_combined_if_unset(revs);
 225
 226         for (i = 0; i < ents; i++) {
 227                 if (i != first_non_parent)
 228                         oid_array_append(&parents, &ent[i].item->oid);
 229         }
 230         diff_tree_combined(&ent[first_non_parent].item->oid, &parents, revs);
 231         oid_array_clear(&parents);
 232 }
 233
 234 static void refresh_index_quietly(void)
 235 {
 236         struct lock_file lock_file = LOCK_INIT;
 237         int fd;
 238
 239         fd = repo_hold_locked_index(the_repository, &lock_file, 0);
 240         if (fd < 0)
 241                 return;
 242         discard_index(the_repository->index);
 243         repo_read_index(the_repository);
 244         refresh_index(the_repository->index, REFRESH_QUIET|REFRESH_UNMERGED, NULL, NULL,
 245                       NULL);
 246         repo_update_index_if_able(the_repository, &lock_file);
 247 }
 248
 249 static void builtin_diff_files(struct rev_info *revs, int argc, const char **argv)
 250 {
 251         unsigned int options = 0;
 252
 253         while (1 < argc && argv[1][0] == '-') {
 254                 if (!strcmp(argv[1], "--base"))
 255                         revs->max_count = 1;
 256                 else if (!strcmp(argv[1], "--ours"))
 257                         revs->max_count = 2;
 258                 else if (!strcmp(argv[1], "--theirs"))
 259                         revs->max_count = 3;
 260                 else if (!strcmp(argv[1], "-q"))
 261                         options |= DIFF_SILENT_ON_REMOVED;
 262                 else if (!strcmp(argv[1], "-h"))
 263                         usage(builtin_diff_usage);
 264                 else {
 265                         error(_("invalid option: %s"), argv[1]);
 266                         usage(builtin_diff_usage);
 267                 }
 268                 argv++; argc--;
 269         }
 270
 271         /*
 272          * "diff --base" should not combine merges because it was not
 273          * asked to.  "diff -c" should not densify (if the user wants
 274          * dense one, --cc can be explicitly asked for, or just rely
 275          * on the default).
 276          */
 277         if (revs->max_count == -1 &&
 278             (revs->diffopt.output_format & DIFF_FORMAT_PATCH))
 279                 diff_merges_set_dense_combined_if_unset(revs);
 280
 281         setup_work_tree();
 282         if (repo_read_index_preload(the_repository, &revs->diffopt.pathspec,
 283                                     0) < 0) {
 284                 die_errno("repo_read_index_preload");
 285         }
 286         run_diff_files(revs, options);
 287 }
 288
 289 struct symdiff {
 290         struct bitmap *skip;
 291         int warn;
 292         const char *base, *left, *right;
 293 };
 294
 295 /*
 296  * Check for symmetric-difference arguments, and if present, arrange
 297  * everything we need to know to handle them correctly.  As a bonus,
 298  * weed out all bogus range-based revision specifications, e.g.,
 299  * "git diff A..B C..D" or "git diff A..B C" get rejected.
 300  *
 301  * For an actual symmetric diff, *symdiff is set this way:
 302  *
 303  *  - its skip is non-NULL and marks *all* rev->pending.objects[i]
 304  *    indices that the caller should ignore (extra merge bases, of
 305  *    which there might be many, and A in A...B).  Note that the
 306  *    chosen merge base and right side are NOT marked.
 307  *  - warn is set if there are multiple merge bases.
 308  *  - base, left, and right point to the names to use in a
 309  *    warning about multiple merge bases.
 310  *
 311  * If there is no symmetric diff argument, sym->skip is NULL and
 312  * sym->warn is cleared.  The remaining fields are not set.
 313  */
 314 static void symdiff_prepare(struct rev_info *rev, struct symdiff *sym)
 315 {
 316         int i, is_symdiff = 0, basecount = 0, othercount = 0;
 317         int lpos = -1, rpos = -1, basepos = -1;
 318         struct bitmap *map = NULL;
 319
 320         /*
 321          * Use the whence fields to find merge bases and left and
 322          * right parts of symmetric difference, so that we do not
 323          * depend on the order that revisions are parsed.  If there
 324          * are any revs that aren't from these sources, we have a
 325          * "git diff C A...B" or "git diff A...B C" case.  Or we
 326          * could even get "git diff A...B C...E", for instance.
 327          *
 328          * If we don't have just one merge base, we pick one
 329          * at random.
 330          *
 331          * NB: REV_CMD_LEFT, REV_CMD_RIGHT are also used for A..B,
 332          * so we must check for SYMMETRIC_LEFT too.  The two arrays
 333          * rev->pending.objects and rev->cmdline.rev are parallel.
 334          */
 335         for (i = 0; i < rev->cmdline.nr; i++) {
 336                 struct object *obj = rev->pending.objects[i].item;
 337                 switch (rev->cmdline.rev[i].whence) {
 338                 case REV_CMD_MERGE_BASE:
 339                         if (basepos < 0)
 340                                 basepos = i;
 341                         basecount++;
 342                         break;          /* do mark all bases */
 343                 case REV_CMD_LEFT:
 344                         if (lpos >= 0)
 345                                 usage(builtin_diff_usage);
 346                         lpos = i;
 347                         if (obj->flags & SYMMETRIC_LEFT) {
 348                                 is_symdiff = 1;
 349                                 break;  /* do mark A */
 350                         }
 351                         continue;
 352                 case REV_CMD_RIGHT:
 353                         if (rpos >= 0)
 354                                 usage(builtin_diff_usage);
 355                         rpos = i;
 356                         continue;       /* don't mark B */
 357                 case REV_CMD_PARENTS_ONLY:
 358                 case REV_CMD_REF:
 359                 case REV_CMD_REV:
 360                         othercount++;
 361                         continue;
 362                 }
 363                 if (!map)
 364                         map = bitmap_new();
 365                 bitmap_set(map, i);
 366         }
 367
 368         /*
 369          * Forbid any additional revs for both A...B and A..B.
 370          */
 371         if (lpos >= 0 && othercount > 0)
 372                 usage(builtin_diff_usage);
 373
 374         if (!is_symdiff) {
 375                 bitmap_free(map);
 376                 sym->warn = 0;
 377                 sym->skip = NULL;
 378                 return;
 379         }
 380
 381         sym->left = rev->pending.objects[lpos].name;
 382         sym->right = rev->pending.objects[rpos].name;
 383         if (basecount == 0)
 384                 die(_("%s...%s: no merge base"), sym->left, sym->right);
 385         sym->base = rev->pending.objects[basepos].name;
 386         bitmap_unset(map, basepos);     /* unmark the base we want */
 387         sym->warn = basecount > 1;
 388         sym->skip = map;
 389 }
 390
 391 int cmd_diff(int argc, const char **argv, const char *prefix)
 392 {
 393         int i;
 394         struct rev_info rev;
 395         struct object_array ent = OBJECT_ARRAY_INIT;
 396         int first_non_parent = -1;
 397         int blobs = 0, paths = 0;
 398         struct object_array_entry *blob[2];
 399         int nongit = 0, no_index = 0;
 400         int result;
 401         struct symdiff sdiff;
 402
 403         /*
 404          * We could get N tree-ish in the rev.pending_objects list.
 405          * Also there could be M blobs there, and P pathspecs. --cached may
 406          * also be present.
 407          *
 408          * N=0, M=0:
 409          *      cache vs files (diff-files)
 410          *
 411          * N=0, M=0, --cached:
 412          *      HEAD vs cache (diff-index --cached)
 413          *
 414          * N=0, M=2:
 415          *      compare two random blobs.  P must be zero.
 416          *
 417          * N=0, M=1, P=1:
 418          *      compare a blob with a working tree file.
 419          *
 420          * N=1, M=0:
 421          *      tree vs files (diff-index)
 422          *
 423          * N=1, M=0, --cached:
 424          *      tree vs cache (diff-index --cached)
 425          *
 426          * N=2, M=0:
 427          *      tree vs tree (diff-tree)
 428          *
 429          * N=0, M=0, P=2:
 430          *      compare two filesystem entities (aka --no-index).
 431          *
 432          * Other cases are errors.
 433          */
 434
 435         /* Were we asked to do --no-index explicitly? */
 436         for (i = 1; i < argc; i++) {
 437                 if (!strcmp(argv[i], "--")) {
 438                         i++;
 439                         break;
 440                 }
 441                 if (!strcmp(argv[i], "--no-index"))
 442                         no_index = DIFF_NO_INDEX_EXPLICIT;
 443                 if (argv[i][0] != '-')
 444                         break;
 445         }
 446
 447         prefix = setup_git_directory_gently(&nongit);
 448
 449         if (!nongit) {
 450                 prepare_repo_settings(the_repository);
 451                 the_repository->settings.command_requires_full_index = 0;
 452         }
 453
 454         if (!no_index) {
 455                 /*
 456                  * Treat git diff with at least one path outside of the
 457                  * repo the same as if the command would have been executed
 458                  * outside of a git repository.  In this case it behaves
 459                  * the same way as "git diff --no-index <a> <b>", which acts
 460                  * as a colourful "diff" replacement.
 461                  */
 462                 if (nongit || ((argc == i + 2) &&
 463                                (!path_inside_repo(prefix, argv[i]) ||
 464                                 !path_inside_repo(prefix, argv[i + 1]))))
 465                         no_index = DIFF_NO_INDEX_IMPLICIT;
 466         }
 467
 468         /*
 469          * When operating outside of a Git repository we need to have a hash
 470          * algorithm at hand so that we can generate the blob hashes. We
 471          * default to SHA1 here, but may eventually want to change this to be
 472          * configurable via a command line option.
 473          */
 474         if (nongit)
 475                 repo_set_hash_algo(the_repository, GIT_HASH_SHA1);
 476
 477         init_diff_ui_defaults();
 478         git_config(git_diff_ui_config, NULL);
 479         prefix = precompose_argv_prefix(argc, argv, prefix);
 480
 481         repo_init_revisions(the_repository, &rev, prefix);
 482
 483         /* Set up defaults that will apply to both no-index and regular diffs. */
 484         init_diffstat_widths(&rev.diffopt);
 485         rev.diffopt.flags.allow_external = 1;
 486         rev.diffopt.flags.allow_textconv = 1;
 487
 488         /* If this is a no-index diff, just run it and exit there. */
 489         if (no_index)
 490                 exit(diff_no_index(&rev, no_index == DIFF_NO_INDEX_IMPLICIT,
 491                                    argc, argv));
 492
 493
 494         /*
 495          * Otherwise, we are doing the usual "git" diff; set up any
 496          * further defaults that apply to regular diffs.
 497          */
 498         rev.diffopt.skip_stat_unmatch = !!diff_auto_refresh_index;
 499
 500         /*
 501          * Default to intent-to-add entries invisible in the
 502          * index. This makes them show up as new files in diff-files
 503          * and not at all in diff-cached.
 504          */
 505         rev.diffopt.ita_invisible_in_index = 1;
 506
 507         if (nongit)
 508                 die(_("Not a git repository"));
 509         argc = setup_revisions(argc, argv, &rev, NULL);
 510         if (!rev.diffopt.output_format) {
 511                 rev.diffopt.output_format = DIFF_FORMAT_PATCH;
 512                 diff_setup_done(&rev.diffopt);
 513         }
 514
 515         rev.diffopt.flags.recursive = 1;
 516         rev.diffopt.rotate_to_strict = 1;
 517
 518         setup_diff_pager(&rev.diffopt);
 519
 520         /*
 521          * Do we have --cached and not have a pending object, then
 522          * default to HEAD by hand.  Eek.
 523          */
 524         if (!rev.pending.nr) {
 525                 int i;
 526                 for (i = 1; i < argc; i++) {
 527                         const char *arg = argv[i];
 528                         if (!strcmp(arg, "--"))
 529                                 break;
 530                         else if (!strcmp(arg, "--cached") ||
 531                                  !strcmp(arg, "--staged")) {
 532                                 add_head_to_pending(&rev);
 533                                 if (!rev.pending.nr) {
 534                                         struct tree *tree;
 535                                         tree = lookup_tree(the_repository,
 536                                                            the_repository->hash_algo->empty_tree);
 537                                         add_pending_object(&rev, &tree->object, "HEAD");
 538                                 }
 539                                 break;
 540                         }
 541                 }
 542         }
 543
 544         symdiff_prepare(&rev, &sdiff);
 545         for (i = 0; i < rev.pending.nr; i++) {
 546                 struct object_array_entry *entry = &rev.pending.objects[i];
 547                 struct object *obj = entry->item;
 548                 const char *name = entry->name;
 549                 int flags = (obj->flags & UNINTERESTING);
 550                 if (!obj->parsed)
 551                         obj = parse_object(the_repository, &obj->oid);
 552                 obj = deref_tag(the_repository, obj, NULL, 0);
 553                 if (!obj)
 554                         die(_("invalid object '%s' given."), name);
 555                 if (obj->type == OBJ_COMMIT)
 556                         obj = &repo_get_commit_tree(the_repository,
 557                                                     ((struct commit *)obj))->object;
 558
 559                 if (obj->type == OBJ_TREE) {
 560                         if (sdiff.skip && bitmap_get(sdiff.skip, i))
 561                                 continue;
 562                         obj->flags |= flags;
 563                         add_object_array(obj, name, &ent);
 564                         if (first_non_parent < 0 &&
 565                             (i >= rev.cmdline.nr || /* HEAD by hand. */
 566                              rev.cmdline.rev[i].whence != REV_CMD_PARENTS_ONLY))
 567                                 first_non_parent = ent.nr - 1;
 568                 } else if (obj->type == OBJ_BLOB) {
 569                         if (2 <= blobs)
 570                                 die(_("more than two blobs given: '%s'"), name);
 571                         blob[blobs] = entry;
 572                         blobs++;
 573
 574                 } else {
 575                         die(_("unhandled object '%s' given."), name);
 576                 }
 577         }
 578         if (rev.prune_data.nr)
 579                 paths += rev.prune_data.nr;
 580
 581         /*
 582          * Now, do the arguments look reasonable?
 583          */
 584         if (!ent.nr) {
 585                 switch (blobs) {
 586                 case 0:
 587                         builtin_diff_files(&rev, argc, argv);
 588                         break;
 589                 case 1:
 590                         if (paths != 1)
 591                                 usage(builtin_diff_usage);
 592                         builtin_diff_b_f(&rev, argc, argv, blob);
 593                         break;
 594                 case 2:
 595                         if (paths)
 596                                 usage(builtin_diff_usage);
 597                         builtin_diff_blobs(&rev, argc, argv, blob);
 598                         break;
 599                 default:
 600                         usage(builtin_diff_usage);
 601                 }
 602         }
 603         else if (blobs)
 604                 usage(builtin_diff_usage);
 605         else if (ent.nr == 1)
 606                 builtin_diff_index(&rev, argc, argv);
 607         else if (ent.nr == 2) {
 608                 if (sdiff.warn)
 609                         warning(_("%s...%s: multiple merge bases, using %s"),
 610                                 sdiff.left, sdiff.right, sdiff.base);
 611                 builtin_diff_tree(&rev, argc, argv,
 612                                   &ent.objects[0], &ent.objects[1]);
 613         } else
 614                 builtin_diff_combined(&rev, argc, argv,
 615                                       ent.objects, ent.nr,
 616                                       first_non_parent);
 617         result = diff_result_code(&rev.diffopt);
 618         if (1 < rev.diffopt.skip_stat_unmatch)
 619                 refresh_index_quietly();
 620         release_revisions(&rev);
 621         object_array_clear(&ent);
 622         UNLEAK(blob);
 623         return result;
 624 }