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 "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 "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 "preload-index.h"
  20 #include "read-cache-ll.h"
  21 #include "revision.h"
  22 #include "log-tree.h"
  23 #include "setup.h"
  24 #include "submodule.h"
  25 #include "oid-array.h"
  26 #include "tree.h"
  27
  28 #define DIFF_NO_INDEX_EXPLICIT 1
  29 #define DIFF_NO_INDEX_IMPLICIT 2
  30
  31 static const char builtin_diff_usage[] =
  32 "git diff [<options>] [<commit>] [--] [<path>...]\n"
  33 "   or: git diff [<options>] --cached [--merge-base] [<commit>] [--] [<path>...]\n"
  34 "   or: git diff [<options>] [--merge-base] <commit> [<commit>...] <commit> [--] [<path>...]\n"
  35 "   or: git diff [<options>] <commit>...<commit> [--] [<path>...]\n"
  36 "   or: git diff [<options>] <blob> <blob>\n"
  37 "   or: git diff [<options>] --no-index [--] <path> <path>"
  38 "\n"
  39 COMMON_DIFF_OPTIONS_HELP;
  40
  41 static const char *blob_path(struct object_array_entry *entry)
  42 {
  43         return entry->path ? entry->path : entry->name;
  44 }
  45
  46 static void stuff_change(struct diff_options *opt,
  47                          unsigned old_mode, unsigned new_mode,
  48                          const struct object_id *old_oid,
  49                          const struct object_id *new_oid,
  50                          int old_oid_valid,
  51                          int new_oid_valid,
  52                          const char *old_path,
  53                          const char *new_path)
  54 {
  55         struct diff_filespec *one, *two;
  56
  57         if (!is_null_oid(old_oid) && !is_null_oid(new_oid) &&
  58             oideq(old_oid, new_oid) && (old_mode == new_mode))
  59                 return;
  60
  61         if (opt->flags.reverse_diff) {
  62                 SWAP(old_mode, new_mode);
  63                 SWAP(old_oid, new_oid);
  64                 SWAP(old_path, new_path);
  65         }
  66
  67         if (opt->prefix &&
  68             (strncmp(old_path, opt->prefix, opt->prefix_length) ||
  69              strncmp(new_path, opt->prefix, opt->prefix_length)))
  70                 return;
  71
  72         one = alloc_filespec(old_path);
  73         two = alloc_filespec(new_path);
  74         fill_filespec(one, old_oid, old_oid_valid, old_mode);
  75         fill_filespec(two, new_oid, new_oid_valid, new_mode);
  76
  77         diff_queue(&diff_queued_diff, one, two);
  78 }
  79
  80 static int builtin_diff_b_f(struct rev_info *revs,
  81                             int argc, const char **argv UNUSED,
  82                             struct object_array_entry **blob)
  83 {
  84         /* Blob vs file in the working tree*/
  85         struct stat st;
  86         const char *path;
  87
  88         if (argc > 1)
  89                 usage(builtin_diff_usage);
  90
  91         GUARD_PATHSPEC(&revs->prune_data, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
  92         path = revs->prune_data.items[0].match;
  93
  94         if (lstat(path, &st))
  95                 die_errno(_("failed to stat '%s'"), path);
  96         if (!(S_ISREG(st.st_mode) || S_ISLNK(st.st_mode)))
  97                 die(_("'%s': not a regular file or symlink"), path);
  98
  99         diff_set_mnemonic_prefix(&revs->diffopt, "o/", "w/");
 100
 101         if (blob[0]->mode == S_IFINVALID)
 102                 blob[0]->mode = canon_mode(st.st_mode);
 103
 104         stuff_change(&revs->diffopt,
 105                      blob[0]->mode, canon_mode(st.st_mode),
 106                      &blob[0]->item->oid, null_oid(),
 107                      1, 0,
 108                      blob[0]->path ? blob[0]->path : path,
 109                      path);
 110         diffcore_std(&revs->diffopt);
 111         diff_flush(&revs->diffopt);
 112         return 0;
 113 }
 114
 115 static int builtin_diff_blobs(struct rev_info *revs,
 116                               int argc, const char **argv UNUSED,
 117                               struct object_array_entry **blob)
 118 {
 119         const unsigned mode = canon_mode(S_IFREG | 0644);
 120
 121         if (argc > 1)
 122                 usage(builtin_diff_usage);
 123
 124         if (blob[0]->mode == S_IFINVALID)
 125                 blob[0]->mode = mode;
 126
 127         if (blob[1]->mode == S_IFINVALID)
 128                 blob[1]->mode = mode;
 129
 130         stuff_change(&revs->diffopt,
 131                      blob[0]->mode, blob[1]->mode,
 132                      &blob[0]->item->oid, &blob[1]->item->oid,
 133                      1, 1,
 134                      blob_path(blob[0]), blob_path(blob[1]));
 135         diffcore_std(&revs->diffopt);
 136         diff_flush(&revs->diffopt);
 137         return 0;
 138 }
 139
 140 static int builtin_diff_index(struct rev_info *revs,
 141                               int argc, const char **argv)
 142 {
 143         unsigned int option = 0;
 144         while (1 < argc) {
 145                 const char *arg = argv[1];
 146                 if (!strcmp(arg, "--cached") || !strcmp(arg, "--staged"))
 147                         option |= DIFF_INDEX_CACHED;
 148                 else if (!strcmp(arg, "--merge-base"))
 149                         option |= DIFF_INDEX_MERGE_BASE;
 150                 else
 151                         usage(builtin_diff_usage);
 152                 argv++; argc--;
 153         }
 154         /*
 155          * Make sure there is one revision (i.e. pending object),
 156          * and there is no revision filtering parameters.
 157          */
 158         if (revs->pending.nr != 1 ||
 159             revs->max_count != -1 || revs->min_age != -1 ||
 160             revs->max_age != -1)
 161                 usage(builtin_diff_usage);
 162         if (!(option & DIFF_INDEX_CACHED)) {
 163                 setup_work_tree();
 164                 if (repo_read_index_preload(the_repository,
 165                                             &revs->diffopt.pathspec, 0) < 0) {
 166                         perror("repo_read_index_preload");
 167                         return -1;
 168                 }
 169         } else if (repo_read_index(the_repository) < 0) {
 170                 perror("repo_read_cache");
 171                 return -1;
 172         }
 173         return run_diff_index(revs, option);
 174 }
 175
 176 static int builtin_diff_tree(struct rev_info *revs,
 177                              int argc, const char **argv,
 178                              struct object_array_entry *ent0,
 179                              struct object_array_entry *ent1)
 180 {
 181         const struct object_id *(oid[2]);
 182         struct object_id mb_oid;
 183         int merge_base = 0;
 184
 185         while (1 < argc) {
 186                 const char *arg = argv[1];
 187                 if (!strcmp(arg, "--merge-base"))
 188                         merge_base = 1;
 189                 else
 190                         usage(builtin_diff_usage);
 191                 argv++; argc--;
 192         }
 193
 194         if (merge_base) {
 195                 diff_get_merge_base(revs, &mb_oid);
 196                 oid[0] = &mb_oid;
 197                 oid[1] = &revs->pending.objects[1].item->oid;
 198         } else {
 199                 int swap = 0;
 200
 201                 /*
 202                  * We saw two trees, ent0 and ent1.  If ent1 is uninteresting,
 203                  * swap them.
 204                  */
 205                 if (ent1->item->flags & UNINTERESTING)
 206                         swap = 1;
 207                 oid[swap] = &ent0->item->oid;
 208                 oid[1 - swap] = &ent1->item->oid;
 209         }
 210         diff_tree_oid(oid[0], oid[1], "", &revs->diffopt);
 211         log_tree_diff_flush(revs);
 212         return 0;
 213 }
 214
 215 static int builtin_diff_combined(struct rev_info *revs,
 216                                  int argc, const char **argv UNUSED,
 217                                  struct object_array_entry *ent,
 218                                  int ents, int first_non_parent)
 219 {
 220         struct oid_array parents = OID_ARRAY_INIT;
 221         int i;
 222
 223         if (argc > 1)
 224                 usage(builtin_diff_usage);
 225
 226         if (first_non_parent < 0)
 227                 die(_("no merge given, only parents."));
 228         if (first_non_parent >= ents)
 229                 BUG("first_non_parent out of range: %d", first_non_parent);
 230
 231         diff_merges_set_dense_combined_if_unset(revs);
 232
 233         for (i = 0; i < ents; i++) {
 234                 if (i != first_non_parent)
 235                         oid_array_append(&parents, &ent[i].item->oid);
 236         }
 237         diff_tree_combined(&ent[first_non_parent].item->oid, &parents, revs);
 238         oid_array_clear(&parents);
 239         return 0;
 240 }
 241
 242 static void refresh_index_quietly(void)
 243 {
 244         struct lock_file lock_file = LOCK_INIT;
 245         int fd;
 246
 247         fd = repo_hold_locked_index(the_repository, &lock_file, 0);
 248         if (fd < 0)
 249                 return;
 250         discard_index(&the_index);
 251         repo_read_index(the_repository);
 252         refresh_index(&the_index, REFRESH_QUIET|REFRESH_UNMERGED, NULL, NULL,
 253                       NULL);
 254         repo_update_index_if_able(the_repository, &lock_file);
 255 }
 256
 257 static int builtin_diff_files(struct rev_info *revs, int argc, const char **argv)
 258 {
 259         unsigned int options = 0;
 260
 261         while (1 < argc && argv[1][0] == '-') {
 262                 if (!strcmp(argv[1], "--base"))
 263                         revs->max_count = 1;
 264                 else if (!strcmp(argv[1], "--ours"))
 265                         revs->max_count = 2;
 266                 else if (!strcmp(argv[1], "--theirs"))
 267                         revs->max_count = 3;
 268                 else if (!strcmp(argv[1], "-q"))
 269                         options |= DIFF_SILENT_ON_REMOVED;
 270                 else if (!strcmp(argv[1], "-h"))
 271                         usage(builtin_diff_usage);
 272                 else
 273                         return error(_("invalid option: %s"), argv[1]);
 274                 argv++; argc--;
 275         }
 276
 277         /*
 278          * "diff --base" should not combine merges because it was not
 279          * asked to.  "diff -c" should not densify (if the user wants
 280          * dense one, --cc can be explicitly asked for, or just rely
 281          * on the default).
 282          */
 283         if (revs->max_count == -1 &&
 284             (revs->diffopt.output_format & DIFF_FORMAT_PATCH))
 285                 diff_merges_set_dense_combined_if_unset(revs);
 286
 287         setup_work_tree();
 288         if (repo_read_index_preload(the_repository, &revs->diffopt.pathspec,
 289                                     0) < 0) {
 290                 perror("repo_read_index_preload");
 291                 return -1;
 292         }
 293         return run_diff_files(revs, options);
 294 }
 295
 296 struct symdiff {
 297         struct bitmap *skip;
 298         int warn;
 299         const char *base, *left, *right;
 300 };
 301
 302 /*
 303  * Check for symmetric-difference arguments, and if present, arrange
 304  * everything we need to know to handle them correctly.  As a bonus,
 305  * weed out all bogus range-based revision specifications, e.g.,
 306  * "git diff A..B C..D" or "git diff A..B C" get rejected.
 307  *
 308  * For an actual symmetric diff, *symdiff is set this way:
 309  *
 310  *  - its skip is non-NULL and marks *all* rev->pending.objects[i]
 311  *    indices that the caller should ignore (extra merge bases, of
 312  *    which there might be many, and A in A...B).  Note that the
 313  *    chosen merge base and right side are NOT marked.
 314  *  - warn is set if there are multiple merge bases.
 315  *  - base, left, and right point to the names to use in a
 316  *    warning about multiple merge bases.
 317  *
 318  * If there is no symmetric diff argument, sym->skip is NULL and
 319  * sym->warn is cleared.  The remaining fields are not set.
 320  */
 321 static void symdiff_prepare(struct rev_info *rev, struct symdiff *sym)
 322 {
 323         int i, is_symdiff = 0, basecount = 0, othercount = 0;
 324         int lpos = -1, rpos = -1, basepos = -1;
 325         struct bitmap *map = NULL;
 326
 327         /*
 328          * Use the whence fields to find merge bases and left and
 329          * right parts of symmetric difference, so that we do not
 330          * depend on the order that revisions are parsed.  If there
 331          * are any revs that aren't from these sources, we have a
 332          * "git diff C A...B" or "git diff A...B C" case.  Or we
 333          * could even get "git diff A...B C...E", for instance.
 334          *
 335          * If we don't have just one merge base, we pick one
 336          * at random.
 337          *
 338          * NB: REV_CMD_LEFT, REV_CMD_RIGHT are also used for A..B,
 339          * so we must check for SYMMETRIC_LEFT too.  The two arrays
 340          * rev->pending.objects and rev->cmdline.rev are parallel.
 341          */
 342         for (i = 0; i < rev->cmdline.nr; i++) {
 343                 struct object *obj = rev->pending.objects[i].item;
 344                 switch (rev->cmdline.rev[i].whence) {
 345                 case REV_CMD_MERGE_BASE:
 346                         if (basepos < 0)
 347                                 basepos = i;
 348                         basecount++;
 349                         break;          /* do mark all bases */
 350                 case REV_CMD_LEFT:
 351                         if (lpos >= 0)
 352                                 usage(builtin_diff_usage);
 353                         lpos = i;
 354                         if (obj->flags & SYMMETRIC_LEFT) {
 355                                 is_symdiff = 1;
 356                                 break;  /* do mark A */
 357                         }
 358                         continue;
 359                 case REV_CMD_RIGHT:
 360                         if (rpos >= 0)
 361                                 usage(builtin_diff_usage);
 362                         rpos = i;
 363                         continue;       /* don't mark B */
 364                 case REV_CMD_PARENTS_ONLY:
 365                 case REV_CMD_REF:
 366                 case REV_CMD_REV:
 367                         othercount++;
 368                         continue;
 369                 }
 370                 if (!map)
 371                         map = bitmap_new();
 372                 bitmap_set(map, i);
 373         }
 374
 375         /*
 376          * Forbid any additional revs for both A...B and A..B.
 377          */
 378         if (lpos >= 0 && othercount > 0)
 379                 usage(builtin_diff_usage);
 380
 381         if (!is_symdiff) {
 382                 bitmap_free(map);
 383                 sym->warn = 0;
 384                 sym->skip = NULL;
 385                 return;
 386         }
 387
 388         sym->left = rev->pending.objects[lpos].name;
 389         sym->right = rev->pending.objects[rpos].name;
 390         if (basecount == 0)
 391                 die(_("%s...%s: no merge base"), sym->left, sym->right);
 392         sym->base = rev->pending.objects[basepos].name;
 393         bitmap_unset(map, basepos);     /* unmark the base we want */
 394         sym->warn = basecount > 1;
 395         sym->skip = map;
 396 }
 397
 398 int cmd_diff(int argc, const char **argv, const char *prefix)
 399 {
 400         int i;
 401         struct rev_info rev;
 402         struct object_array ent = OBJECT_ARRAY_INIT;
 403         int first_non_parent = -1;
 404         int blobs = 0, paths = 0;
 405         struct object_array_entry *blob[2];
 406         int nongit = 0, no_index = 0;
 407         int result = 0;
 408         struct symdiff sdiff;
 409
 410         /*
 411          * We could get N tree-ish in the rev.pending_objects list.
 412          * Also there could be M blobs there, and P pathspecs. --cached may
 413          * also be present.
 414          *
 415          * N=0, M=0:
 416          *      cache vs files (diff-files)
 417          *
 418          * N=0, M=0, --cached:
 419          *      HEAD vs cache (diff-index --cached)
 420          *
 421          * N=0, M=2:
 422          *      compare two random blobs.  P must be zero.
 423          *
 424          * N=0, M=1, P=1:
 425          *      compare a blob with a working tree file.
 426          *
 427          * N=1, M=0:
 428          *      tree vs files (diff-index)
 429          *
 430          * N=1, M=0, --cached:
 431          *      tree vs cache (diff-index --cached)
 432          *
 433          * N=2, M=0:
 434          *      tree vs tree (diff-tree)
 435          *
 436          * N=0, M=0, P=2:
 437          *      compare two filesystem entities (aka --no-index).
 438          *
 439          * Other cases are errors.
 440          */
 441
 442         /* Were we asked to do --no-index explicitly? */
 443         for (i = 1; i < argc; i++) {
 444                 if (!strcmp(argv[i], "--")) {
 445                         i++;
 446                         break;
 447                 }
 448                 if (!strcmp(argv[i], "--no-index"))
 449                         no_index = DIFF_NO_INDEX_EXPLICIT;
 450                 if (argv[i][0] != '-')
 451                         break;
 452         }
 453
 454         prefix = setup_git_directory_gently(&nongit);
 455
 456         if (!nongit) {
 457                 prepare_repo_settings(the_repository);
 458                 the_repository->settings.command_requires_full_index = 0;
 459         }
 460
 461         if (!no_index) {
 462                 /*
 463                  * Treat git diff with at least one path outside of the
 464                  * repo the same as if the command would have been executed
 465                  * outside of a git repository.  In this case it behaves
 466                  * the same way as "git diff --no-index <a> <b>", which acts
 467                  * as a colourful "diff" replacement.
 468                  */
 469                 if (nongit || ((argc == i + 2) &&
 470                                (!path_inside_repo(prefix, argv[i]) ||
 471                                 !path_inside_repo(prefix, argv[i + 1]))))
 472                         no_index = DIFF_NO_INDEX_IMPLICIT;
 473         }
 474
 475         init_diff_ui_defaults();
 476         git_config(git_diff_ui_config, NULL);
 477         prefix = precompose_argv_prefix(argc, argv, prefix);
 478
 479         repo_init_revisions(the_repository, &rev, prefix);
 480
 481         /* Set up defaults that will apply to both no-index and regular diffs. */
 482         rev.diffopt.stat_width = -1;
 483         rev.diffopt.stat_graph_width = -1;
 484         rev.diffopt.flags.allow_external = 1;
 485         rev.diffopt.flags.allow_textconv = 1;
 486
 487         /* If this is a no-index diff, just run it and exit there. */
 488         if (no_index)
 489                 exit(diff_no_index(&rev, no_index == DIFF_NO_INDEX_IMPLICIT,
 490                                    argc, argv));
 491
 492
 493         /*
 494          * Otherwise, we are doing the usual "git" diff; set up any
 495          * further defaults that apply to regular diffs.
 496          */
 497         rev.diffopt.skip_stat_unmatch = !!diff_auto_refresh_index;
 498
 499         /*
 500          * Default to intent-to-add entries invisible in the
 501          * index. This makes them show up as new files in diff-files
 502          * and not at all in diff-cached.
 503          */
 504         rev.diffopt.ita_invisible_in_index = 1;
 505
 506         if (nongit)
 507                 die(_("Not a git repository"));
 508         argc = setup_revisions(argc, argv, &rev, NULL);
 509         if (!rev.diffopt.output_format) {
 510                 rev.diffopt.output_format = DIFF_FORMAT_PATCH;
 511                 diff_setup_done(&rev.diffopt);
 512         }
 513
 514         rev.diffopt.flags.recursive = 1;
 515         rev.diffopt.rotate_to_strict = 1;
 516
 517         setup_diff_pager(&rev.diffopt);
 518
 519         /*
 520          * Do we have --cached and not have a pending object, then
 521          * default to HEAD by hand.  Eek.
 522          */
 523         if (!rev.pending.nr) {
 524                 int i;
 525                 for (i = 1; i < argc; i++) {
 526                         const char *arg = argv[i];
 527                         if (!strcmp(arg, "--"))
 528                                 break;
 529                         else if (!strcmp(arg, "--cached") ||
 530                                  !strcmp(arg, "--staged")) {
 531                                 add_head_to_pending(&rev);
 532                                 if (!rev.pending.nr) {
 533                                         struct tree *tree;
 534                                         tree = lookup_tree(the_repository,
 535                                                            the_repository->hash_algo->empty_tree);
 536                                         add_pending_object(&rev, &tree->object, "HEAD");
 537                                 }
 538                                 break;
 539                         }
 540                 }
 541         }
 542
 543         symdiff_prepare(&rev, &sdiff);
 544         for (i = 0; i < rev.pending.nr; i++) {
 545                 struct object_array_entry *entry = &rev.pending.objects[i];
 546                 struct object *obj = entry->item;
 547                 const char *name = entry->name;
 548                 int flags = (obj->flags & UNINTERESTING);
 549                 if (!obj->parsed)
 550                         obj = parse_object(the_repository, &obj->oid);
 551                 obj = deref_tag(the_repository, obj, NULL, 0);
 552                 if (!obj)
 553                         die(_("invalid object '%s' given."), name);
 554                 if (obj->type == OBJ_COMMIT)
 555                         obj = &repo_get_commit_tree(the_repository,
 556                                                     ((struct commit *)obj))->object;
 557
 558                 if (obj->type == OBJ_TREE) {
 559                         if (sdiff.skip && bitmap_get(sdiff.skip, i))
 560                                 continue;
 561                         obj->flags |= flags;
 562                         add_object_array(obj, name, &ent);
 563                         if (first_non_parent < 0 &&
 564                             (i >= rev.cmdline.nr || /* HEAD by hand. */
 565                              rev.cmdline.rev[i].whence != REV_CMD_PARENTS_ONLY))
 566                                 first_non_parent = ent.nr - 1;
 567                 } else if (obj->type == OBJ_BLOB) {
 568                         if (2 <= blobs)
 569                                 die(_("more than two blobs given: '%s'"), name);
 570                         blob[blobs] = entry;
 571                         blobs++;
 572
 573                 } else {
 574                         die(_("unhandled object '%s' given."), name);
 575                 }
 576         }
 577         if (rev.prune_data.nr)
 578                 paths += rev.prune_data.nr;
 579
 580         /*
 581          * Now, do the arguments look reasonable?
 582          */
 583         if (!ent.nr) {
 584                 switch (blobs) {
 585                 case 0:
 586                         result = builtin_diff_files(&rev, argc, argv);
 587                         break;
 588                 case 1:
 589                         if (paths != 1)
 590                                 usage(builtin_diff_usage);
 591                         result = builtin_diff_b_f(&rev, argc, argv, blob);
 592                         break;
 593                 case 2:
 594                         if (paths)
 595                                 usage(builtin_diff_usage);
 596                         result = builtin_diff_blobs(&rev, argc, argv, blob);
 597                         break;
 598                 default:
 599                         usage(builtin_diff_usage);
 600                 }
 601         }
 602         else if (blobs)
 603                 usage(builtin_diff_usage);
 604         else if (ent.nr == 1)
 605                 result = builtin_diff_index(&rev, argc, argv);
 606         else if (ent.nr == 2) {
 607                 if (sdiff.warn)
 608                         warning(_("%s...%s: multiple merge bases, using %s"),
 609                                 sdiff.left, sdiff.right, sdiff.base);
 610                 result = builtin_diff_tree(&rev, argc, argv,
 611                                            &ent.objects[0], &ent.objects[1]);
 612         } else
 613                 result = builtin_diff_combined(&rev, argc, argv,
 614                                                ent.objects, ent.nr,
 615                                                first_non_parent);
 616         result = diff_result_code(&rev.diffopt, result);
 617         if (1 < rev.diffopt.skip_stat_unmatch)
 618                 refresh_index_quietly();
 619         release_revisions(&rev);
 620         object_array_clear(&ent);
 621         UNLEAK(blob);
 622         return result;
 623 }