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