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
  14 /*
  15  * diff-files
  16  */
  17
  18 /*
  19  * Has the work tree entity been removed?
  20  *
  21  * Return 1 if it was removed from the work tree, 0 if an entity to be
  22  * compared with the cache entry ce still exists (the latter includes
  23  * the case where a directory that is not a submodule repository
  24  * exists for ce that is a submodule -- it is a submodule that is not
  25  * checked out).  Return negative for an error.
  26  */
  27 static int check_removed(const struct cache_entry *ce, struct stat *st)
  28 {
  29         if (lstat(ce->name, st) < 0) {
  30                 if (errno != ENOENT && errno != ENOTDIR)
  31                         return -1;
  32                 return 1;
  33         }
  34         if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
  35                 return 1;
  36         if (S_ISDIR(st->st_mode)) {
  37                 unsigned char sub[20];
  38
  39                 /*
  40                  * If ce is already a gitlink, we can have a plain
  41                  * directory (i.e. the submodule is not checked out),
  42                  * or a checked out submodule.  Either case this is not
  43                  * a case where something was removed from the work tree,
  44                  * so we will return 0.
  45                  *
  46                  * Otherwise, if the directory is not a submodule
  47                  * repository, that means ce which was a blob turned into
  48                  * a directory --- the blob was removed!
  49                  */
  50                 if (!S_ISGITLINK(ce->ce_mode) &&
  51                     resolve_gitlink_ref(ce->name, "HEAD", sub))
  52                         return 1;
  53         }
  54         return 0;
  55 }
  56
  57 int run_diff_files(struct rev_info *revs, unsigned int option)
  58 {
  59         int entries, i;
  60         int diff_unmerged_stage = revs->max_count;
  61         int silent_on_removed = option & DIFF_SILENT_ON_REMOVED;
  62         unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
  63                               ? CE_MATCH_RACY_IS_DIRTY : 0);
  64
  65         diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
  66
  67         if (diff_unmerged_stage < 0)
  68                 diff_unmerged_stage = 2;
  69         entries = active_nr;
  70         for (i = 0; i < entries; i++) {
  71                 struct stat st;
  72                 unsigned int oldmode, newmode;
  73                 struct cache_entry *ce = active_cache[i];
  74                 int changed;
  75
  76                 if (diff_can_quit_early(&revs->diffopt))
  77                         break;
  78
  79                 if (!ce_path_match(ce, revs->prune_data))
  80                         continue;
  81
  82                 if (ce_stage(ce)) {
  83                         struct combine_diff_path *dpath;
  84                         int num_compare_stages = 0;
  85                         size_t path_len;
  86
  87                         path_len = ce_namelen(ce);
  88
  89                         dpath = xmalloc(combine_diff_path_size(5, path_len));
  90                         dpath->path = (char *) &(dpath->parent[5]);
  91
  92                         dpath->next = NULL;
  93                         dpath->len = path_len;
  94                         memcpy(dpath->path, ce->name, path_len);
  95                         dpath->path[path_len] = '\0';
  96                         hashclr(dpath->sha1);
  97                         memset(&(dpath->parent[0]), 0,
  98                                sizeof(struct combine_diff_parent)*5);
  99
 100                         changed = check_removed(ce, &st);
 101                         if (!changed)
 102                                 dpath->mode = ce_mode_from_stat(ce, st.st_mode);
 103                         else {
 104                                 if (changed < 0) {
 105                                         perror(ce->name);
 106                                         continue;
 107                                 }
 108                                 if (silent_on_removed)
 109                                         continue;
 110                         }
 111
 112                         while (i < entries) {
 113                                 struct cache_entry *nce = active_cache[i];
 114                                 int stage;
 115
 116                                 if (strcmp(ce->name, nce->name))
 117                                         break;
 118
 119                                 /* Stage #2 (ours) is the first parent,
 120                                  * stage #3 (theirs) is the second.
 121                                  */
 122                                 stage = ce_stage(nce);
 123                                 if (2 <= stage) {
 124                                         int mode = nce->ce_mode;
 125                                         num_compare_stages++;
 126                                         hashcpy(dpath->parent[stage-2].sha1, nce->sha1);
 127                                         dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 128                                         dpath->parent[stage-2].status =
 129                                                 DIFF_STATUS_MODIFIED;
 130                                 }
 131
 132                                 /* diff against the proper unmerged stage */
 133                                 if (stage == diff_unmerged_stage)
 134                                         ce = nce;
 135                                 i++;
 136                         }
 137                         /*
 138                          * Compensate for loop update
 139                          */
 140                         i--;
 141
 142                         if (revs->combine_merges && num_compare_stages == 2) {
 143                                 show_combined_diff(dpath, 2,
 144                                                    revs->dense_combined_merges,
 145                                                    revs);
 146                                 free(dpath);
 147                                 continue;
 148                         }
 149                         free(dpath);
 150                         dpath = NULL;
 151
 152                         /*
 153                          * Show the diff for the 'ce' if we found the one
 154                          * from the desired stage.
 155                          */
 156                         diff_unmerge(&revs->diffopt, ce->name, 0, null_sha1);
 157                         if (ce_stage(ce) != diff_unmerged_stage)
 158                                 continue;
 159                 }
 160
 161                 if (ce_uptodate(ce))
 162                         continue;
 163
 164                 changed = check_removed(ce, &st);
 165                 if (changed) {
 166                         if (changed < 0) {
 167                                 perror(ce->name);
 168                                 continue;
 169                         }
 170                         if (silent_on_removed)
 171                                 continue;
 172                         diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 173                                        ce->sha1, ce->name);
 174                         continue;
 175                 }
 176                 changed = ce_match_stat(ce, &st, ce_option);
 177                 if (!changed) {
 178                         ce_mark_uptodate(ce);
 179                         if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 180                                 continue;
 181                 }
 182                 oldmode = ce->ce_mode;
 183                 newmode = ce_mode_from_stat(ce, st.st_mode);
 184                 diff_change(&revs->diffopt, oldmode, newmode,
 185                             ce->sha1, (changed ? null_sha1 : ce->sha1),
 186                             ce->name);
 187
 188         }
 189         diffcore_std(&revs->diffopt);
 190         diff_flush(&revs->diffopt);
 191         return 0;
 192 }
 193
 194 /*
 195  * diff-index
 196  */
 197
 198 /* A file entry went away or appeared */
 199 static void diff_index_show_file(struct rev_info *revs,
 200                                  const char *prefix,
 201                                  struct cache_entry *ce,
 202                                  const unsigned char *sha1, unsigned int mode)
 203 {
 204         diff_addremove(&revs->diffopt, prefix[0], mode,
 205                        sha1, ce->name);
 206 }
 207
 208 static int get_stat_data(struct cache_entry *ce,
 209                          const unsigned char **sha1p,
 210                          unsigned int *modep,
 211                          int cached, int match_missing)
 212 {
 213         const unsigned char *sha1 = ce->sha1;
 214         unsigned int mode = ce->ce_mode;
 215
 216         if (!cached && !ce_uptodate(ce)) {
 217                 int changed;
 218                 struct stat st;
 219                 changed = check_removed(ce, &st);
 220                 if (changed < 0)
 221                         return -1;
 222                 else if (changed) {
 223                         if (match_missing) {
 224                                 *sha1p = sha1;
 225                                 *modep = mode;
 226                                 return 0;
 227                         }
 228                         return -1;
 229                 }
 230                 changed = ce_match_stat(ce, &st, 0);
 231                 if (changed) {
 232                         mode = ce_mode_from_stat(ce, st.st_mode);
 233                         sha1 = null_sha1;
 234                 }
 235         }
 236
 237         *sha1p = sha1;
 238         *modep = mode;
 239         return 0;
 240 }
 241
 242 static void show_new_file(struct rev_info *revs,
 243                           struct cache_entry *new,
 244                           int cached, int match_missing)
 245 {
 246         const unsigned char *sha1;
 247         unsigned int mode;
 248
 249         /*
 250          * New file in the index: it might actually be different in
 251          * the working copy.
 252          */
 253         if (get_stat_data(new, &sha1, &mode, cached, match_missing) < 0)
 254                 return;
 255
 256         diff_index_show_file(revs, "+", new, sha1, mode);
 257 }
 258
 259 static int show_modified(struct rev_info *revs,
 260                          struct cache_entry *old,
 261                          struct cache_entry *new,
 262                          int report_missing,
 263                          int cached, int match_missing)
 264 {
 265         unsigned int mode, oldmode;
 266         const unsigned char *sha1;
 267
 268         if (get_stat_data(new, &sha1, &mode, cached, match_missing) < 0) {
 269                 if (report_missing)
 270                         diff_index_show_file(revs, "-", old,
 271                                              old->sha1, old->ce_mode);
 272                 return -1;
 273         }
 274
 275         if (revs->combine_merges && !cached &&
 276             (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
 277                 struct combine_diff_path *p;
 278                 int pathlen = ce_namelen(new);
 279
 280                 p = xmalloc(combine_diff_path_size(2, pathlen));
 281                 p->path = (char *) &p->parent[2];
 282                 p->next = NULL;
 283                 p->len = pathlen;
 284                 memcpy(p->path, new->name, pathlen);
 285                 p->path[pathlen] = 0;
 286                 p->mode = mode;
 287                 hashclr(p->sha1);
 288                 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 289                 p->parent[0].status = DIFF_STATUS_MODIFIED;
 290                 p->parent[0].mode = new->ce_mode;
 291                 hashcpy(p->parent[0].sha1, new->sha1);
 292                 p->parent[1].status = DIFF_STATUS_MODIFIED;
 293                 p->parent[1].mode = old->ce_mode;
 294                 hashcpy(p->parent[1].sha1, old->sha1);
 295                 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 296                 free(p);
 297                 return 0;
 298         }
 299
 300         oldmode = old->ce_mode;
 301         if (mode == oldmode && !hashcmp(sha1, old->sha1) &&
 302             !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 303                 return 0;
 304
 305         diff_change(&revs->diffopt, oldmode, mode,
 306                     old->sha1, sha1, old->name);
 307         return 0;
 308 }
 309
 310 /*
 311  * This turns all merge entries into "stage 3". That guarantees that
 312  * when we read in the new tree (into "stage 1"), we won't lose sight
 313  * of the fact that we had unmerged entries.
 314  */
 315 static void mark_merge_entries(void)
 316 {
 317         int i;
 318         for (i = 0; i < active_nr; i++) {
 319                 struct cache_entry *ce = active_cache[i];
 320                 if (!ce_stage(ce))
 321                         continue;
 322                 ce->ce_flags |= CE_STAGEMASK;
 323         }
 324 }
 325
 326 /*
 327  * This gets a mix of an existing index and a tree, one pathname entry
 328  * at a time. The index entry may be a single stage-0 one, but it could
 329  * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 330  * give you the position and number of entries in the index).
 331  */
 332 static void do_oneway_diff(struct unpack_trees_options *o,
 333         struct cache_entry *idx,
 334         struct cache_entry *tree)
 335 {
 336         struct rev_info *revs = o->unpack_data;
 337         int match_missing, cached;
 338
 339         /*
 340          * Backward compatibility wart - "diff-index -m" does
 341          * not mean "do not ignore merges", but "match_missing".
 342          *
 343          * But with the revision flag parsing, that's found in
 344          * "!revs->ignore_merges".
 345          */
 346         cached = o->index_only;
 347         match_missing = !revs->ignore_merges;
 348
 349         if (cached && idx && ce_stage(idx)) {
 350                 if (tree)
 351                         diff_unmerge(&revs->diffopt, idx->name, idx->ce_mode, idx->sha1);
 352                 return;
 353         }
 354
 355         /*
 356          * Something added to the tree?
 357          */
 358         if (!tree) {
 359                 show_new_file(revs, idx, cached, match_missing);
 360                 return;
 361         }
 362
 363         /*
 364          * Something removed from the tree?
 365          */
 366         if (!idx) {
 367                 diff_index_show_file(revs, "-", tree, tree->sha1, tree->ce_mode);
 368                 return;
 369         }
 370
 371         /* Show difference between old and new */
 372         show_modified(revs, tree, idx, 1, cached, match_missing);
 373 }
 374
 375 static inline void skip_same_name(struct cache_entry *ce, struct unpack_trees_options *o)
 376 {
 377         int len = ce_namelen(ce);
 378         const struct index_state *index = o->src_index;
 379
 380         while (o->pos < index->cache_nr) {
 381                 struct cache_entry *next = index->cache[o->pos];
 382                 if (len != ce_namelen(next))
 383                         break;
 384                 if (memcmp(ce->name, next->name, len))
 385                         break;
 386                 o->pos++;
 387         }
 388 }
 389
 390 /*
 391  * The unpack_trees() interface is designed for merging, so
 392  * the different source entries are designed primarily for
 393  * the source trees, with the old index being really mainly
 394  * used for being replaced by the result.
 395  *
 396  * For diffing, the index is more important, and we only have a
 397  * single tree.
 398  *
 399  * We're supposed to return how many index entries we want to skip.
 400  *
 401  * This wrapper makes it all more readable, and takes care of all
 402  * the fairly complex unpack_trees() semantic requirements, including
 403  * the skipping, the path matching, the type conflict cases etc.
 404  */
 405 static int oneway_diff(struct cache_entry **src, struct unpack_trees_options *o)
 406 {
 407         struct cache_entry *idx = src[0];
 408         struct cache_entry *tree = src[1];
 409         struct rev_info *revs = o->unpack_data;
 410
 411         if (idx && ce_stage(idx))
 412                 skip_same_name(idx, o);
 413
 414         /*
 415          * Unpack-trees generates a DF/conflict entry if
 416          * there was a directory in the index and a tree
 417          * in the tree. From a diff standpoint, that's a
 418          * delete of the tree and a create of the file.
 419          */
 420         if (tree == o->df_conflict_entry)
 421                 tree = NULL;
 422
 423         if (ce_path_match(idx ? idx : tree, revs->prune_data))
 424                 do_oneway_diff(o, idx, tree);
 425
 426         return 0;
 427 }
 428
 429 int run_diff_index(struct rev_info *revs, int cached)
 430 {
 431         struct object *ent;
 432         struct tree *tree;
 433         const char *tree_name;
 434         struct unpack_trees_options opts;
 435         struct tree_desc t;
 436
 437         mark_merge_entries();
 438
 439         ent = revs->pending.objects[0].item;
 440         tree_name = revs->pending.objects[0].name;
 441         tree = parse_tree_indirect(ent->sha1);
 442         if (!tree)
 443                 return error("bad tree object %s", tree_name);
 444
 445         memset(&opts, 0, sizeof(opts));
 446         opts.head_idx = 1;
 447         opts.index_only = cached;
 448         opts.diff_index_cached = (cached &&
 449                                   !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 450         opts.merge = 1;
 451         opts.fn = oneway_diff;
 452         opts.unpack_data = revs;
 453         opts.src_index = &the_index;
 454         opts.dst_index = NULL;
 455
 456         init_tree_desc(&t, tree->buffer, tree->size);
 457         if (unpack_trees(1, &t, &opts))
 458                 exit(128);
 459
 460         diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 461         diffcore_std(&revs->diffopt);
 462         diff_flush(&revs->diffopt);
 463         return 0;
 464 }
 465
 466 int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
 467 {
 468         struct tree *tree;
 469         struct rev_info revs;
 470         int i;
 471         struct cache_entry **dst;
 472         struct cache_entry *last = NULL;
 473         struct unpack_trees_options opts;
 474         struct tree_desc t;
 475
 476         /*
 477          * This is used by git-blame to run diff-cache internally;
 478          * it potentially needs to repeatedly run this, so we will
 479          * start by removing the higher order entries the last round
 480          * left behind.
 481          */
 482         dst = active_cache;
 483         for (i = 0; i < active_nr; i++) {
 484                 struct cache_entry *ce = active_cache[i];
 485                 if (ce_stage(ce)) {
 486                         if (last && !strcmp(ce->name, last->name))
 487                                 continue;
 488                         cache_tree_invalidate_path(active_cache_tree,
 489                                                    ce->name);
 490                         last = ce;
 491                         ce->ce_flags |= CE_REMOVE;
 492                 }
 493                 *dst++ = ce;
 494         }
 495         active_nr = dst - active_cache;
 496
 497         init_revisions(&revs, NULL);
 498         revs.prune_data = opt->paths;
 499         tree = parse_tree_indirect(tree_sha1);
 500         if (!tree)
 501                 die("bad tree object %s", sha1_to_hex(tree_sha1));
 502
 503         memset(&opts, 0, sizeof(opts));
 504         opts.head_idx = 1;
 505         opts.index_only = 1;
 506         opts.diff_index_cached = !DIFF_OPT_TST(opt, FIND_COPIES_HARDER);
 507         opts.merge = 1;
 508         opts.fn = oneway_diff;
 509         opts.unpack_data = &revs;
 510         opts.src_index = &the_index;
 511         opts.dst_index = &the_index;
 512
 513         init_tree_desc(&t, tree->buffer, tree->size);
 514         if (unpack_trees(1, &t, &opts))
 515                 exit(128);
 516         return 0;
 517 }
 518
 519 int index_differs_from(const char *def, int diff_flags)
 520 {
 521         struct rev_info rev;
 522
 523         init_revisions(&rev, NULL);
 524         setup_revisions(0, NULL, &rev, def);
 525         DIFF_OPT_SET(&rev.diffopt, QUICK);
 526         DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 527         rev.diffopt.flags |= diff_flags;
 528         run_diff_index(&rev, 1);
 529         if (rev.pending.alloc)
 530                 free(rev.pending.objects);
 531         return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 532 }