diff-lib.c

   1 /*
   2  * Copyright (C) 2005 Junio C Hamano
   3  */
   4 #include "cache.h"
   5 #include "quote.h"
   6 #include "commit.h"
   7 #include "diff.h"
   8 #include "diffcore.h"
   9 #include "revision.h"
  10 #include "cache-tree.h"
  11 #include "unpack-trees.h"
  12 #include "refs.h"
  13 #include "submodule.h"
  14 #include "dir.h"
  15
  16 /*
  17  * diff-files
  18  */
  19
  20 /*
  21  * Has the work tree entity been removed?
  22  *
  23  * Return 1 if it was removed from the work tree, 0 if an entity to be
  24  * compared with the cache entry ce still exists (the latter includes
  25  * the case where a directory that is not a submodule repository
  26  * exists for ce that is a submodule -- it is a submodule that is not
  27  * checked out).  Return negative for an error.
  28  */
  29 static int check_removed(const struct cache_entry *ce, struct stat *st)
  30 {
  31         if (lstat(ce->name, st) < 0) {
  32                 if (errno != ENOENT && errno != ENOTDIR)
  33                         return -1;
  34                 return 1;
  35         }
  36         if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
  37                 return 1;
  38         if (S_ISDIR(st->st_mode)) {
  39                 unsigned char sub[20];
  40
  41                 /*
  42                  * If ce is already a gitlink, we can have a plain
  43                  * directory (i.e. the submodule is not checked out),
  44                  * or a checked out submodule.  Either case this is not
  45                  * a case where something was removed from the work tree,
  46                  * so we will return 0.
  47                  *
  48                  * Otherwise, if the directory is not a submodule
  49                  * repository, that means ce which was a blob turned into
  50                  * a directory --- the blob was removed!
  51                  */
  52                 if (!S_ISGITLINK(ce->ce_mode) &&
  53                     resolve_gitlink_ref(ce->name, "HEAD", sub))
  54                         return 1;
  55         }
  56         return 0;
  57 }
  58
  59 /*
  60  * Has a file changed or has a submodule new commits or a dirty work tree?
  61  *
  62  * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
  63  * option is set, the caller does not only want to know if a submodule is
  64  * modified at all but wants to know all the conditions that are met (new
  65  * commits, untracked content and/or modified content).
  66  */
  67 static int match_stat_with_submodule(struct diff_options *diffopt,
  68                                      const struct cache_entry *ce,
  69                                      struct stat *st, unsigned ce_option,
  70                                      unsigned *dirty_submodule)
  71 {
  72         int changed = ce_match_stat(ce, st, ce_option);
  73         if (S_ISGITLINK(ce->ce_mode)) {
  74                 unsigned orig_flags = diffopt->flags;
  75                 if (!DIFF_OPT_TST(diffopt, OVERRIDE_SUBMODULE_CONFIG))
  76                         set_diffopt_flags_from_submodule_config(diffopt, ce->name);
  77                 if (DIFF_OPT_TST(diffopt, IGNORE_SUBMODULES))
  78                         changed = 0;
  79                 else if (!DIFF_OPT_TST(diffopt, IGNORE_DIRTY_SUBMODULES)
  80                     && (!changed || DIFF_OPT_TST(diffopt, DIRTY_SUBMODULES)))
  81                         *dirty_submodule = is_submodule_modified(ce->name, DIFF_OPT_TST(diffopt, IGNORE_UNTRACKED_IN_SUBMODULES));
  82                 diffopt->flags = orig_flags;
  83         }
  84         return changed;
  85 }
  86
  87 int run_diff_files(struct rev_info *revs, unsigned int option)
  88 {
  89         int entries, i;
  90         int diff_unmerged_stage = revs->max_count;
  91         unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
  92                               ? CE_MATCH_RACY_IS_DIRTY : 0);
  93
  94         diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
  95
  96         if (diff_unmerged_stage < 0)
  97                 diff_unmerged_stage = 2;
  98         entries = active_nr;
  99         for (i = 0; i < entries; i++) {
 100                 unsigned int oldmode, newmode;
 101                 struct cache_entry *ce = active_cache[i];
 102                 int changed;
 103                 unsigned dirty_submodule = 0;
 104                 const unsigned char *old_sha1, *new_sha1;
 105
 106                 if (diff_can_quit_early(&revs->diffopt))
 107                         break;
 108
 109                 if (!ce_path_match(ce, &revs->prune_data, NULL))
 110                         continue;
 111
 112                 if (ce_stage(ce)) {
 113                         struct combine_diff_path *dpath;
 114                         struct diff_filepair *pair;
 115                         unsigned int wt_mode = 0;
 116                         int num_compare_stages = 0;
 117                         size_t path_len;
 118                         struct stat st;
 119
 120                         path_len = ce_namelen(ce);
 121
 122                         dpath = xmalloc(combine_diff_path_size(5, path_len));
 123                         dpath->path = (char *) &(dpath->parent[5]);
 124
 125                         dpath->next = NULL;
 126                         memcpy(dpath->path, ce->name, path_len);
 127                         dpath->path[path_len] = '\0';
 128                         oidclr(&dpath->oid);
 129                         memset(&(dpath->parent[0]), 0,
 130                                sizeof(struct combine_diff_parent)*5);
 131
 132                         changed = check_removed(ce, &st);
 133                         if (!changed)
 134                                 wt_mode = ce_mode_from_stat(ce, st.st_mode);
 135                         else {
 136                                 if (changed < 0) {
 137                                         perror(ce->name);
 138                                         continue;
 139                                 }
 140                                 wt_mode = 0;
 141                         }
 142                         dpath->mode = wt_mode;
 143
 144                         while (i < entries) {
 145                                 struct cache_entry *nce = active_cache[i];
 146                                 int stage;
 147
 148                                 if (strcmp(ce->name, nce->name))
 149                                         break;
 150
 151                                 /* Stage #2 (ours) is the first parent,
 152                                  * stage #3 (theirs) is the second.
 153                                  */
 154                                 stage = ce_stage(nce);
 155                                 if (2 <= stage) {
 156                                         int mode = nce->ce_mode;
 157                                         num_compare_stages++;
 158                                         oidcpy(&dpath->parent[stage - 2].oid,
 159                                                &nce->oid);
 160                                         dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 161                                         dpath->parent[stage-2].status =
 162                                                 DIFF_STATUS_MODIFIED;
 163                                 }
 164
 165                                 /* diff against the proper unmerged stage */
 166                                 if (stage == diff_unmerged_stage)
 167                                         ce = nce;
 168                                 i++;
 169                         }
 170                         /*
 171                          * Compensate for loop update
 172                          */
 173                         i--;
 174
 175                         if (revs->combine_merges && num_compare_stages == 2) {
 176                                 show_combined_diff(dpath, 2,
 177                                                    revs->dense_combined_merges,
 178                                                    revs);
 179                                 free(dpath);
 180                                 continue;
 181                         }
 182                         free(dpath);
 183                         dpath = NULL;
 184
 185                         /*
 186                          * Show the diff for the 'ce' if we found the one
 187                          * from the desired stage.
 188                          */
 189                         pair = diff_unmerge(&revs->diffopt, ce->name);
 190                         if (wt_mode)
 191                                 pair->two->mode = wt_mode;
 192                         if (ce_stage(ce) != diff_unmerged_stage)
 193                                 continue;
 194                 }
 195
 196                 if (ce_uptodate(ce) || ce_skip_worktree(ce))
 197                         continue;
 198
 199                 /* If CE_VALID is set, don't look at workdir for file removal */
 200                 if (ce->ce_flags & CE_VALID) {
 201                         changed = 0;
 202                         newmode = ce->ce_mode;
 203                 } else {
 204                         struct stat st;
 205
 206                         changed = check_removed(ce, &st);
 207                         if (changed) {
 208                                 if (changed < 0) {
 209                                         perror(ce->name);
 210                                         continue;
 211                                 }
 212                                 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 213                                                ce->oid.hash,
 214                                                !is_null_oid(&ce->oid),
 215                                                ce->name, 0);
 216                                 continue;
 217                         } else if (revs->diffopt.ita_invisible_in_index &&
 218                                    ce_intent_to_add(ce)) {
 219                                 diff_addremove(&revs->diffopt, '+', ce->ce_mode,
 220                                                EMPTY_BLOB_SHA1_BIN, 0,
 221                                                ce->name, 0);
 222                                 continue;
 223                         }
 224
 225                         changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
 226                                                             ce_option, &dirty_submodule);
 227                         newmode = ce_mode_from_stat(ce, st.st_mode);
 228                 }
 229
 230                 if (!changed && !dirty_submodule) {
 231                         ce_mark_uptodate(ce);
 232                         if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 233                                 continue;
 234                 }
 235                 oldmode = ce->ce_mode;
 236                 old_sha1 = ce->oid.hash;
 237                 new_sha1 = changed ? null_sha1 : ce->oid.hash;
 238                 diff_change(&revs->diffopt, oldmode, newmode,
 239                             old_sha1, new_sha1,
 240                             !is_null_sha1(old_sha1),
 241                             !is_null_sha1(new_sha1),
 242                             ce->name, 0, dirty_submodule);
 243
 244         }
 245         diffcore_std(&revs->diffopt);
 246         diff_flush(&revs->diffopt);
 247         return 0;
 248 }
 249
 250 /*
 251  * diff-index
 252  */
 253
 254 /* A file entry went away or appeared */
 255 static void diff_index_show_file(struct rev_info *revs,
 256                                  const char *prefix,
 257                                  const struct cache_entry *ce,
 258                                  const unsigned char *sha1, int sha1_valid,
 259                                  unsigned int mode,
 260                                  unsigned dirty_submodule)
 261 {
 262         diff_addremove(&revs->diffopt, prefix[0], mode,
 263                        sha1, sha1_valid, ce->name, dirty_submodule);
 264 }
 265
 266 static int get_stat_data(const struct cache_entry *ce,
 267                          const unsigned char **sha1p,
 268                          unsigned int *modep,
 269                          int cached, int match_missing,
 270                          unsigned *dirty_submodule, struct diff_options *diffopt)
 271 {
 272         const unsigned char *sha1 = ce->oid.hash;
 273         unsigned int mode = ce->ce_mode;
 274
 275         if (!cached && !ce_uptodate(ce)) {
 276                 int changed;
 277                 struct stat st;
 278                 changed = check_removed(ce, &st);
 279                 if (changed < 0)
 280                         return -1;
 281                 else if (changed) {
 282                         if (match_missing) {
 283                                 *sha1p = sha1;
 284                                 *modep = mode;
 285                                 return 0;
 286                         }
 287                         return -1;
 288                 }
 289                 changed = match_stat_with_submodule(diffopt, ce, &st,
 290                                                     0, dirty_submodule);
 291                 if (changed) {
 292                         mode = ce_mode_from_stat(ce, st.st_mode);
 293                         sha1 = null_sha1;
 294                 }
 295         }
 296
 297         *sha1p = sha1;
 298         *modep = mode;
 299         return 0;
 300 }
 301
 302 static void show_new_file(struct rev_info *revs,
 303                           const struct cache_entry *new,
 304                           int cached, int match_missing)
 305 {
 306         const unsigned char *sha1;
 307         unsigned int mode;
 308         unsigned dirty_submodule = 0;
 309
 310         /*
 311          * New file in the index: it might actually be different in
 312          * the working tree.
 313          */
 314         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 315             &dirty_submodule, &revs->diffopt) < 0)
 316                 return;
 317
 318         diff_index_show_file(revs, "+", new, sha1, !is_null_sha1(sha1), mode, dirty_submodule);
 319 }
 320
 321 static int show_modified(struct rev_info *revs,
 322                          const struct cache_entry *old,
 323                          const struct cache_entry *new,
 324                          int report_missing,
 325                          int cached, int match_missing)
 326 {
 327         unsigned int mode, oldmode;
 328         const unsigned char *sha1;
 329         unsigned dirty_submodule = 0;
 330
 331         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 332                           &dirty_submodule, &revs->diffopt) < 0) {
 333                 if (report_missing)
 334                         diff_index_show_file(revs, "-", old,
 335                                              old->oid.hash, 1, old->ce_mode,
 336                                              0);
 337                 return -1;
 338         }
 339
 340         if (revs->combine_merges && !cached &&
 341             (hashcmp(sha1, old->oid.hash) || oidcmp(&old->oid, &new->oid))) {
 342                 struct combine_diff_path *p;
 343                 int pathlen = ce_namelen(new);
 344
 345                 p = xmalloc(combine_diff_path_size(2, pathlen));
 346                 p->path = (char *) &p->parent[2];
 347                 p->next = NULL;
 348                 memcpy(p->path, new->name, pathlen);
 349                 p->path[pathlen] = 0;
 350                 p->mode = mode;
 351                 oidclr(&p->oid);
 352                 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 353                 p->parent[0].status = DIFF_STATUS_MODIFIED;
 354                 p->parent[0].mode = new->ce_mode;
 355                 oidcpy(&p->parent[0].oid, &new->oid);
 356                 p->parent[1].status = DIFF_STATUS_MODIFIED;
 357                 p->parent[1].mode = old->ce_mode;
 358                 oidcpy(&p->parent[1].oid, &old->oid);
 359                 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 360                 free(p);
 361                 return 0;
 362         }
 363
 364         oldmode = old->ce_mode;
 365         if (mode == oldmode && !hashcmp(sha1, old->oid.hash) && !dirty_submodule &&
 366             !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 367                 return 0;
 368
 369         diff_change(&revs->diffopt, oldmode, mode,
 370                     old->oid.hash, sha1, 1, !is_null_sha1(sha1),
 371                     old->name, 0, dirty_submodule);
 372         return 0;
 373 }
 374
 375 /*
 376  * This gets a mix of an existing index and a tree, one pathname entry
 377  * at a time. The index entry may be a single stage-0 one, but it could
 378  * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 379  * give you the position and number of entries in the index).
 380  */
 381 static void do_oneway_diff(struct unpack_trees_options *o,
 382                            const struct cache_entry *idx,
 383                            const struct cache_entry *tree)
 384 {
 385         struct rev_info *revs = o->unpack_data;
 386         int match_missing, cached;
 387
 388         /* i-t-a entries do not actually exist in the index */
 389         if (revs->diffopt.ita_invisible_in_index &&
 390             idx && ce_intent_to_add(idx)) {
 391                 idx = NULL;
 392                 if (!tree)
 393                         return; /* nothing to diff.. */
 394         }
 395
 396         /* if the entry is not checked out, don't examine work tree */
 397         cached = o->index_only ||
 398                 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
 399         /*
 400          * Backward compatibility wart - "diff-index -m" does
 401          * not mean "do not ignore merges", but "match_missing".
 402          *
 403          * But with the revision flag parsing, that's found in
 404          * "!revs->ignore_merges".
 405          */
 406         match_missing = !revs->ignore_merges;
 407
 408         if (cached && idx && ce_stage(idx)) {
 409                 struct diff_filepair *pair;
 410                 pair = diff_unmerge(&revs->diffopt, idx->name);
 411                 if (tree)
 412                         fill_filespec(pair->one, tree->oid.hash, 1,
 413                                       tree->ce_mode);
 414                 return;
 415         }
 416
 417         /*
 418          * Something added to the tree?
 419          */
 420         if (!tree) {
 421                 show_new_file(revs, idx, cached, match_missing);
 422                 return;
 423         }
 424
 425         /*
 426          * Something removed from the tree?
 427          */
 428         if (!idx) {
 429                 diff_index_show_file(revs, "-", tree, tree->oid.hash, 1,
 430                                      tree->ce_mode, 0);
 431                 return;
 432         }
 433
 434         /* Show difference between old and new */
 435         show_modified(revs, tree, idx, 1, cached, match_missing);
 436 }
 437
 438 /*
 439  * The unpack_trees() interface is designed for merging, so
 440  * the different source entries are designed primarily for
 441  * the source trees, with the old index being really mainly
 442  * used for being replaced by the result.
 443  *
 444  * For diffing, the index is more important, and we only have a
 445  * single tree.
 446  *
 447  * We're supposed to advance o->pos to skip what we have already processed.
 448  *
 449  * This wrapper makes it all more readable, and takes care of all
 450  * the fairly complex unpack_trees() semantic requirements, including
 451  * the skipping, the path matching, the type conflict cases etc.
 452  */
 453 static int oneway_diff(const struct cache_entry * const *src,
 454                        struct unpack_trees_options *o)
 455 {
 456         const struct cache_entry *idx = src[0];
 457         const struct cache_entry *tree = src[1];
 458         struct rev_info *revs = o->unpack_data;
 459
 460         /*
 461          * Unpack-trees generates a DF/conflict entry if
 462          * there was a directory in the index and a tree
 463          * in the tree. From a diff standpoint, that's a
 464          * delete of the tree and a create of the file.
 465          */
 466         if (tree == o->df_conflict_entry)
 467                 tree = NULL;
 468
 469         if (ce_path_match(idx ? idx : tree, &revs->prune_data, NULL)) {
 470                 do_oneway_diff(o, idx, tree);
 471                 if (diff_can_quit_early(&revs->diffopt)) {
 472                         o->exiting_early = 1;
 473                         return -1;
 474                 }
 475         }
 476
 477         return 0;
 478 }
 479
 480 static int diff_cache(struct rev_info *revs,
 481                       const struct object_id *tree_oid,
 482                       const char *tree_name,
 483                       int cached)
 484 {
 485         struct tree *tree;
 486         struct tree_desc t;
 487         struct unpack_trees_options opts;
 488
 489         tree = parse_tree_indirect(tree_oid);
 490         if (!tree)
 491                 return error("bad tree object %s",
 492                              tree_name ? tree_name : oid_to_hex(tree_oid));
 493         memset(&opts, 0, sizeof(opts));
 494         opts.head_idx = 1;
 495         opts.index_only = cached;
 496         opts.diff_index_cached = (cached &&
 497                                   !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 498         opts.merge = 1;
 499         opts.fn = oneway_diff;
 500         opts.unpack_data = revs;
 501         opts.src_index = &the_index;
 502         opts.dst_index = NULL;
 503         opts.pathspec = &revs->diffopt.pathspec;
 504         opts.pathspec->recursive = 1;
 505
 506         init_tree_desc(&t, tree->buffer, tree->size);
 507         return unpack_trees(1, &t, &opts);
 508 }
 509
 510 int run_diff_index(struct rev_info *revs, int cached)
 511 {
 512         struct object_array_entry *ent;
 513
 514         ent = revs->pending.objects;
 515         if (diff_cache(revs, &ent->item->oid, ent->name, cached))
 516                 exit(128);
 517
 518         diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 519         diffcore_fix_diff_index(&revs->diffopt);
 520         diffcore_std(&revs->diffopt);
 521         diff_flush(&revs->diffopt);
 522         return 0;
 523 }
 524
 525 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
 526 {
 527         struct rev_info revs;
 528
 529         init_revisions(&revs, NULL);
 530         copy_pathspec(&revs.prune_data, &opt->pathspec);
 531         revs.diffopt = *opt;
 532
 533         if (diff_cache(&revs, tree_oid, NULL, 1))
 534                 exit(128);
 535         return 0;
 536 }
 537
 538 int index_differs_from(const char *def, int diff_flags,
 539                        int ita_invisible_in_index)
 540 {
 541         struct rev_info rev;
 542         struct setup_revision_opt opt;
 543
 544         init_revisions(&rev, NULL);
 545         memset(&opt, 0, sizeof(opt));
 546         opt.def = def;
 547         setup_revisions(0, NULL, &rev, &opt);
 548         DIFF_OPT_SET(&rev.diffopt, QUICK);
 549         DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 550         rev.diffopt.flags |= diff_flags;
 551         rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
 552         run_diff_index(&rev, 1);
 553         if (rev.pending.alloc)
 554                 free(rev.pending.objects);
 555         return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 556 }