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                                         hashcpy(dpath->parent[stage-2].oid.hash, nce->sha1);
 159                                         dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 160                                         dpath->parent[stage-2].status =
 161                                                 DIFF_STATUS_MODIFIED;
 162                                 }
 163
 164                                 /* diff against the proper unmerged stage */
 165                                 if (stage == diff_unmerged_stage)
 166                                         ce = nce;
 167                                 i++;
 168                         }
 169                         /*
 170                          * Compensate for loop update
 171                          */
 172                         i--;
 173
 174                         if (revs->combine_merges && num_compare_stages == 2) {
 175                                 show_combined_diff(dpath, 2,
 176                                                    revs->dense_combined_merges,
 177                                                    revs);
 178                                 free(dpath);
 179                                 continue;
 180                         }
 181                         free(dpath);
 182                         dpath = NULL;
 183
 184                         /*
 185                          * Show the diff for the 'ce' if we found the one
 186                          * from the desired stage.
 187                          */
 188                         pair = diff_unmerge(&revs->diffopt, ce->name);
 189                         if (wt_mode)
 190                                 pair->two->mode = wt_mode;
 191                         if (ce_stage(ce) != diff_unmerged_stage)
 192                                 continue;
 193                 }
 194
 195                 if (ce_uptodate(ce) || ce_skip_worktree(ce))
 196                         continue;
 197
 198                 /* If CE_VALID is set, don't look at workdir for file removal */
 199                 if (ce->ce_flags & CE_VALID) {
 200                         changed = 0;
 201                         newmode = ce->ce_mode;
 202                 } else {
 203                         struct stat st;
 204
 205                         changed = check_removed(ce, &st);
 206                         if (changed) {
 207                                 if (changed < 0) {
 208                                         perror(ce->name);
 209                                         continue;
 210                                 }
 211                                 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 212                                                ce->sha1, !is_null_sha1(ce->sha1),
 213                                                ce->name, 0);
 214                                 continue;
 215                         } else if (ce->ce_flags & CE_INTENT_TO_ADD) {
 216                                 diff_addremove(&revs->diffopt, '+', ce->ce_mode,
 217                                                EMPTY_BLOB_SHA1_BIN, 0,
 218                                                ce->name, 0);
 219                                 continue;
 220                         }
 221
 222                         changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
 223                                                             ce_option, &dirty_submodule);
 224                         newmode = ce_mode_from_stat(ce, st.st_mode);
 225                 }
 226
 227                 if (!changed && !dirty_submodule) {
 228                         ce_mark_uptodate(ce);
 229                         if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 230                                 continue;
 231                 }
 232                 oldmode = ce->ce_mode;
 233                 old_sha1 = ce->sha1;
 234                 new_sha1 = changed ? null_sha1 : ce->sha1;
 235                 diff_change(&revs->diffopt, oldmode, newmode,
 236                             old_sha1, new_sha1,
 237                             !is_null_sha1(old_sha1),
 238                             !is_null_sha1(new_sha1),
 239                             ce->name, 0, dirty_submodule);
 240
 241         }
 242         diffcore_std(&revs->diffopt);
 243         diff_flush(&revs->diffopt);
 244         return 0;
 245 }
 246
 247 /*
 248  * diff-index
 249  */
 250
 251 /* A file entry went away or appeared */
 252 static void diff_index_show_file(struct rev_info *revs,
 253                                  const char *prefix,
 254                                  const struct cache_entry *ce,
 255                                  const unsigned char *sha1, int sha1_valid,
 256                                  unsigned int mode,
 257                                  unsigned dirty_submodule)
 258 {
 259         diff_addremove(&revs->diffopt, prefix[0], mode,
 260                        sha1, sha1_valid, ce->name, dirty_submodule);
 261 }
 262
 263 static int get_stat_data(const struct cache_entry *ce,
 264                          const unsigned char **sha1p,
 265                          unsigned int *modep,
 266                          int cached, int match_missing,
 267                          unsigned *dirty_submodule, struct diff_options *diffopt)
 268 {
 269         const unsigned char *sha1 = ce->sha1;
 270         unsigned int mode = ce->ce_mode;
 271
 272         if (!cached && !ce_uptodate(ce)) {
 273                 int changed;
 274                 struct stat st;
 275                 changed = check_removed(ce, &st);
 276                 if (changed < 0)
 277                         return -1;
 278                 else if (changed) {
 279                         if (match_missing) {
 280                                 *sha1p = sha1;
 281                                 *modep = mode;
 282                                 return 0;
 283                         }
 284                         return -1;
 285                 }
 286                 changed = match_stat_with_submodule(diffopt, ce, &st,
 287                                                     0, dirty_submodule);
 288                 if (changed) {
 289                         mode = ce_mode_from_stat(ce, st.st_mode);
 290                         sha1 = null_sha1;
 291                 }
 292         }
 293
 294         *sha1p = sha1;
 295         *modep = mode;
 296         return 0;
 297 }
 298
 299 static void show_new_file(struct rev_info *revs,
 300                           const struct cache_entry *new,
 301                           int cached, int match_missing)
 302 {
 303         const unsigned char *sha1;
 304         unsigned int mode;
 305         unsigned dirty_submodule = 0;
 306
 307         /*
 308          * New file in the index: it might actually be different in
 309          * the working tree.
 310          */
 311         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 312             &dirty_submodule, &revs->diffopt) < 0)
 313                 return;
 314
 315         diff_index_show_file(revs, "+", new, sha1, !is_null_sha1(sha1), mode, dirty_submodule);
 316 }
 317
 318 static int show_modified(struct rev_info *revs,
 319                          const struct cache_entry *old,
 320                          const struct cache_entry *new,
 321                          int report_missing,
 322                          int cached, int match_missing)
 323 {
 324         unsigned int mode, oldmode;
 325         const unsigned char *sha1;
 326         unsigned dirty_submodule = 0;
 327
 328         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 329                           &dirty_submodule, &revs->diffopt) < 0) {
 330                 if (report_missing)
 331                         diff_index_show_file(revs, "-", old,
 332                                              old->sha1, 1, old->ce_mode, 0);
 333                 return -1;
 334         }
 335
 336         if (revs->combine_merges && !cached &&
 337             (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
 338                 struct combine_diff_path *p;
 339                 int pathlen = ce_namelen(new);
 340
 341                 p = xmalloc(combine_diff_path_size(2, pathlen));
 342                 p->path = (char *) &p->parent[2];
 343                 p->next = NULL;
 344                 memcpy(p->path, new->name, pathlen);
 345                 p->path[pathlen] = 0;
 346                 p->mode = mode;
 347                 oidclr(&p->oid);
 348                 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 349                 p->parent[0].status = DIFF_STATUS_MODIFIED;
 350                 p->parent[0].mode = new->ce_mode;
 351                 hashcpy(p->parent[0].oid.hash, new->sha1);
 352                 p->parent[1].status = DIFF_STATUS_MODIFIED;
 353                 p->parent[1].mode = old->ce_mode;
 354                 hashcpy(p->parent[1].oid.hash, old->sha1);
 355                 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 356                 free(p);
 357                 return 0;
 358         }
 359
 360         oldmode = old->ce_mode;
 361         if (mode == oldmode && !hashcmp(sha1, old->sha1) && !dirty_submodule &&
 362             !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 363                 return 0;
 364
 365         diff_change(&revs->diffopt, oldmode, mode,
 366                     old->sha1, sha1, 1, !is_null_sha1(sha1),
 367                     old->name, 0, dirty_submodule);
 368         return 0;
 369 }
 370
 371 /*
 372  * This gets a mix of an existing index and a tree, one pathname entry
 373  * at a time. The index entry may be a single stage-0 one, but it could
 374  * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 375  * give you the position and number of entries in the index).
 376  */
 377 static void do_oneway_diff(struct unpack_trees_options *o,
 378                            const struct cache_entry *idx,
 379                            const struct cache_entry *tree)
 380 {
 381         struct rev_info *revs = o->unpack_data;
 382         int match_missing, cached;
 383
 384         /* i-t-a entries do not actually exist in the index */
 385         if (idx && (idx->ce_flags & CE_INTENT_TO_ADD)) {
 386                 idx = NULL;
 387                 if (!tree)
 388                         return; /* nothing to diff.. */
 389         }
 390
 391         /* if the entry is not checked out, don't examine work tree */
 392         cached = o->index_only ||
 393                 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
 394         /*
 395          * Backward compatibility wart - "diff-index -m" does
 396          * not mean "do not ignore merges", but "match_missing".
 397          *
 398          * But with the revision flag parsing, that's found in
 399          * "!revs->ignore_merges".
 400          */
 401         match_missing = !revs->ignore_merges;
 402
 403         if (cached && idx && ce_stage(idx)) {
 404                 struct diff_filepair *pair;
 405                 pair = diff_unmerge(&revs->diffopt, idx->name);
 406                 if (tree)
 407                         fill_filespec(pair->one, tree->sha1, 1, tree->ce_mode);
 408                 return;
 409         }
 410
 411         /*
 412          * Something added to the tree?
 413          */
 414         if (!tree) {
 415                 show_new_file(revs, idx, cached, match_missing);
 416                 return;
 417         }
 418
 419         /*
 420          * Something removed from the tree?
 421          */
 422         if (!idx) {
 423                 diff_index_show_file(revs, "-", tree, tree->sha1, 1, tree->ce_mode, 0);
 424                 return;
 425         }
 426
 427         /* Show difference between old and new */
 428         show_modified(revs, tree, idx, 1, cached, match_missing);
 429 }
 430
 431 /*
 432  * The unpack_trees() interface is designed for merging, so
 433  * the different source entries are designed primarily for
 434  * the source trees, with the old index being really mainly
 435  * used for being replaced by the result.
 436  *
 437  * For diffing, the index is more important, and we only have a
 438  * single tree.
 439  *
 440  * We're supposed to advance o->pos to skip what we have already processed.
 441  *
 442  * This wrapper makes it all more readable, and takes care of all
 443  * the fairly complex unpack_trees() semantic requirements, including
 444  * the skipping, the path matching, the type conflict cases etc.
 445  */
 446 static int oneway_diff(const struct cache_entry * const *src,
 447                        struct unpack_trees_options *o)
 448 {
 449         const struct cache_entry *idx = src[0];
 450         const struct cache_entry *tree = src[1];
 451         struct rev_info *revs = o->unpack_data;
 452
 453         /*
 454          * Unpack-trees generates a DF/conflict entry if
 455          * there was a directory in the index and a tree
 456          * in the tree. From a diff standpoint, that's a
 457          * delete of the tree and a create of the file.
 458          */
 459         if (tree == o->df_conflict_entry)
 460                 tree = NULL;
 461
 462         if (ce_path_match(idx ? idx : tree, &revs->prune_data, NULL)) {
 463                 do_oneway_diff(o, idx, tree);
 464                 if (diff_can_quit_early(&revs->diffopt)) {
 465                         o->exiting_early = 1;
 466                         return -1;
 467                 }
 468         }
 469
 470         return 0;
 471 }
 472
 473 static int diff_cache(struct rev_info *revs,
 474                       const unsigned char *tree_sha1,
 475                       const char *tree_name,
 476                       int cached)
 477 {
 478         struct tree *tree;
 479         struct tree_desc t;
 480         struct unpack_trees_options opts;
 481
 482         tree = parse_tree_indirect(tree_sha1);
 483         if (!tree)
 484                 return error("bad tree object %s",
 485                              tree_name ? tree_name : sha1_to_hex(tree_sha1));
 486         memset(&opts, 0, sizeof(opts));
 487         opts.head_idx = 1;
 488         opts.index_only = cached;
 489         opts.diff_index_cached = (cached &&
 490                                   !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 491         opts.merge = 1;
 492         opts.fn = oneway_diff;
 493         opts.unpack_data = revs;
 494         opts.src_index = &the_index;
 495         opts.dst_index = NULL;
 496         opts.pathspec = &revs->diffopt.pathspec;
 497         opts.pathspec->recursive = 1;
 498
 499         init_tree_desc(&t, tree->buffer, tree->size);
 500         return unpack_trees(1, &t, &opts);
 501 }
 502
 503 int run_diff_index(struct rev_info *revs, int cached)
 504 {
 505         struct object_array_entry *ent;
 506
 507         ent = revs->pending.objects;
 508         if (diff_cache(revs, ent->item->sha1, ent->name, cached))
 509                 exit(128);
 510
 511         diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 512         diffcore_fix_diff_index(&revs->diffopt);
 513         diffcore_std(&revs->diffopt);
 514         diff_flush(&revs->diffopt);
 515         return 0;
 516 }
 517
 518 int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
 519 {
 520         struct rev_info revs;
 521
 522         init_revisions(&revs, NULL);
 523         copy_pathspec(&revs.prune_data, &opt->pathspec);
 524         revs.diffopt = *opt;
 525
 526         if (diff_cache(&revs, tree_sha1, NULL, 1))
 527                 exit(128);
 528         return 0;
 529 }
 530
 531 int index_differs_from(const char *def, int diff_flags)
 532 {
 533         struct rev_info rev;
 534         struct setup_revision_opt opt;
 535
 536         init_revisions(&rev, NULL);
 537         memset(&opt, 0, sizeof(opt));
 538         opt.def = def;
 539         setup_revisions(0, NULL, &rev, &opt);
 540         DIFF_OPT_SET(&rev.diffopt, QUICK);
 541         DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 542         rev.diffopt.flags |= diff_flags;
 543         run_diff_index(&rev, 1);
 544         if (rev.pending.alloc)
 545                 free(rev.pending.objects);
 546         return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 547 }