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