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_OPT_TST(&revs->diffopt, QUICK) &&
  77                     !revs->diffopt.filter &&
  78                     DIFF_OPT_TST(&revs->diffopt, HAS_CHANGES))
  79                         break;
  80
  81                 if (!ce_path_match(ce, revs->prune_data))
  82                         continue;
  83
  84                 if (ce_stage(ce)) {
  85                         struct combine_diff_path *dpath;
  86                         int num_compare_stages = 0;
  87                         size_t path_len;
  88
  89                         path_len = ce_namelen(ce);
  90
  91                         dpath = xmalloc(combine_diff_path_size(5, path_len));
  92                         dpath->path = (char *) &(dpath->parent[5]);
  93
  94                         dpath->next = NULL;
  95                         dpath->len = path_len;
  96                         memcpy(dpath->path, ce->name, path_len);
  97                         dpath->path[path_len] = '\0';
  98                         hashclr(dpath->sha1);
  99                         memset(&(dpath->parent[0]), 0,
 100                                sizeof(struct combine_diff_parent)*5);
 101
 102                         changed = check_removed(ce, &st);
 103                         if (!changed)
 104                                 dpath->mode = ce_mode_from_stat(ce, st.st_mode);
 105                         else {
 106                                 if (changed < 0) {
 107                                         perror(ce->name);
 108                                         continue;
 109                                 }
 110                                 if (silent_on_removed)
 111                                         continue;
 112                         }
 113
 114                         while (i < entries) {
 115                                 struct cache_entry *nce = active_cache[i];
 116                                 int stage;
 117
 118                                 if (strcmp(ce->name, nce->name))
 119                                         break;
 120
 121                                 /* Stage #2 (ours) is the first parent,
 122                                  * stage #3 (theirs) is the second.
 123                                  */
 124                                 stage = ce_stage(nce);
 125                                 if (2 <= stage) {
 126                                         int mode = nce->ce_mode;
 127                                         num_compare_stages++;
 128                                         hashcpy(dpath->parent[stage-2].sha1, nce->sha1);
 129                                         dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 130                                         dpath->parent[stage-2].status =
 131                                                 DIFF_STATUS_MODIFIED;
 132                                 }
 133
 134                                 /* diff against the proper unmerged stage */
 135                                 if (stage == diff_unmerged_stage)
 136                                         ce = nce;
 137                                 i++;
 138                         }
 139                         /*
 140                          * Compensate for loop update
 141                          */
 142                         i--;
 143
 144                         if (revs->combine_merges && num_compare_stages == 2) {
 145                                 show_combined_diff(dpath, 2,
 146                                                    revs->dense_combined_merges,
 147                                                    revs);
 148                                 free(dpath);
 149                                 continue;
 150                         }
 151                         free(dpath);
 152                         dpath = NULL;
 153
 154                         /*
 155                          * Show the diff for the 'ce' if we found the one
 156                          * from the desired stage.
 157                          */
 158                         diff_unmerge(&revs->diffopt, ce->name, 0, null_sha1);
 159                         if (ce_stage(ce) != diff_unmerged_stage)
 160                                 continue;
 161                 }
 162
 163                 if (ce_uptodate(ce))
 164                         continue;
 165
 166                 changed = check_removed(ce, &st);
 167                 if (changed) {
 168                         if (changed < 0) {
 169                                 perror(ce->name);
 170                                 continue;
 171                         }
 172                         if (silent_on_removed)
 173                                 continue;
 174                         diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 175                                        ce->sha1, ce->name);
 176                         continue;
 177                 }
 178                 changed = ce_match_stat(ce, &st, ce_option);
 179                 if (!changed) {
 180                         ce_mark_uptodate(ce);
 181                         if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 182                                 continue;
 183                 }
 184                 oldmode = ce->ce_mode;
 185                 newmode = ce_mode_from_stat(ce, st.st_mode);
 186                 diff_change(&revs->diffopt, oldmode, newmode,
 187                             ce->sha1, (changed ? null_sha1 : ce->sha1),
 188                             ce->name);
 189
 190         }
 191         diffcore_std(&revs->diffopt);
 192         diff_flush(&revs->diffopt);
 193         return 0;
 194 }
 195
 196 /*
 197  * diff-index
 198  */
 199
 200 /* A file entry went away or appeared */
 201 static void diff_index_show_file(struct rev_info *revs,
 202                                  const char *prefix,
 203                                  struct cache_entry *ce,
 204                                  const unsigned char *sha1, unsigned int mode)
 205 {
 206         diff_addremove(&revs->diffopt, prefix[0], mode,
 207                        sha1, ce->name);
 208 }
 209
 210 static int get_stat_data(struct cache_entry *ce,
 211                          const unsigned char **sha1p,
 212                          unsigned int *modep,
 213                          int cached, int match_missing)
 214 {
 215         const unsigned char *sha1 = ce->sha1;
 216         unsigned int mode = ce->ce_mode;
 217
 218         if (!cached && !ce_uptodate(ce)) {
 219                 int changed;
 220                 struct stat st;
 221                 changed = check_removed(ce, &st);
 222                 if (changed < 0)
 223                         return -1;
 224                 else if (changed) {
 225                         if (match_missing) {
 226                                 *sha1p = sha1;
 227                                 *modep = mode;
 228                                 return 0;
 229                         }
 230                         return -1;
 231                 }
 232                 changed = ce_match_stat(ce, &st, 0);
 233                 if (changed) {
 234                         mode = ce_mode_from_stat(ce, st.st_mode);
 235                         sha1 = null_sha1;
 236                 }
 237         }
 238
 239         *sha1p = sha1;
 240         *modep = mode;
 241         return 0;
 242 }
 243
 244 static void show_new_file(struct rev_info *revs,
 245                           struct cache_entry *new,
 246                           int cached, int match_missing)
 247 {
 248         const unsigned char *sha1;
 249         unsigned int mode;
 250
 251         /*
 252          * New file in the index: it might actually be different in
 253          * the working copy.
 254          */
 255         if (get_stat_data(new, &sha1, &mode, cached, match_missing) < 0)
 256                 return;
 257
 258         diff_index_show_file(revs, "+", new, sha1, mode);
 259 }
 260
 261 static int show_modified(struct rev_info *revs,
 262                          struct cache_entry *old,
 263                          struct cache_entry *new,
 264                          int report_missing,
 265                          int cached, int match_missing)
 266 {
 267         unsigned int mode, oldmode;
 268         const unsigned char *sha1;
 269
 270         if (get_stat_data(new, &sha1, &mode, cached, match_missing) < 0) {
 271                 if (report_missing)
 272                         diff_index_show_file(revs, "-", old,
 273                                              old->sha1, old->ce_mode);
 274                 return -1;
 275         }
 276
 277         if (revs->combine_merges && !cached &&
 278             (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
 279                 struct combine_diff_path *p;
 280                 int pathlen = ce_namelen(new);
 281
 282                 p = xmalloc(combine_diff_path_size(2, pathlen));
 283                 p->path = (char *) &p->parent[2];
 284                 p->next = NULL;
 285                 p->len = pathlen;
 286                 memcpy(p->path, new->name, pathlen);
 287                 p->path[pathlen] = 0;
 288                 p->mode = mode;
 289                 hashclr(p->sha1);
 290                 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 291                 p->parent[0].status = DIFF_STATUS_MODIFIED;
 292                 p->parent[0].mode = new->ce_mode;
 293                 hashcpy(p->parent[0].sha1, new->sha1);
 294                 p->parent[1].status = DIFF_STATUS_MODIFIED;
 295                 p->parent[1].mode = old->ce_mode;
 296                 hashcpy(p->parent[1].sha1, old->sha1);
 297                 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 298                 free(p);
 299                 return 0;
 300         }
 301
 302         oldmode = old->ce_mode;
 303         if (mode == oldmode && !hashcmp(sha1, old->sha1) &&
 304             !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 305                 return 0;
 306
 307         diff_change(&revs->diffopt, oldmode, mode,
 308                     old->sha1, sha1, old->name);
 309         return 0;
 310 }
 311
 312 /*
 313  * This turns all merge entries into "stage 3". That guarantees that
 314  * when we read in the new tree (into "stage 1"), we won't lose sight
 315  * of the fact that we had unmerged entries.
 316  */
 317 static void mark_merge_entries(void)
 318 {
 319         int i;
 320         for (i = 0; i < active_nr; i++) {
 321                 struct cache_entry *ce = active_cache[i];
 322                 if (!ce_stage(ce))
 323                         continue;
 324                 ce->ce_flags |= CE_STAGEMASK;
 325         }
 326 }
 327
 328 /*
 329  * This gets a mix of an existing index and a tree, one pathname entry
 330  * at a time. The index entry may be a single stage-0 one, but it could
 331  * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 332  * give you the position and number of entries in the index).
 333  */
 334 static void do_oneway_diff(struct unpack_trees_options *o,
 335         struct cache_entry *idx,
 336         struct cache_entry *tree)
 337 {
 338         struct rev_info *revs = o->unpack_data;
 339         int match_missing, cached;
 340
 341         /*
 342          * Backward compatibility wart - "diff-index -m" does
 343          * not mean "do not ignore merges", but "match_missing".
 344          *
 345          * But with the revision flag parsing, that's found in
 346          * "!revs->ignore_merges".
 347          */
 348         cached = o->index_only;
 349         match_missing = !revs->ignore_merges;
 350
 351         if (cached && idx && ce_stage(idx)) {
 352                 if (tree)
 353                         diff_unmerge(&revs->diffopt, idx->name, idx->ce_mode, idx->sha1);
 354                 return;
 355         }
 356
 357         /*
 358          * Something added to the tree?
 359          */
 360         if (!tree) {
 361                 show_new_file(revs, idx, cached, match_missing);
 362                 return;
 363         }
 364
 365         /*
 366          * Something removed from the tree?
 367          */
 368         if (!idx) {
 369                 diff_index_show_file(revs, "-", tree, tree->sha1, tree->ce_mode);
 370                 return;
 371         }
 372
 373         /* Show difference between old and new */
 374         show_modified(revs, tree, idx, 1, cached, match_missing);
 375 }
 376
 377 static inline void skip_same_name(struct cache_entry *ce, struct unpack_trees_options *o)
 378 {
 379         int len = ce_namelen(ce);
 380         const struct index_state *index = o->src_index;
 381
 382         while (o->pos < index->cache_nr) {
 383                 struct cache_entry *next = index->cache[o->pos];
 384                 if (len != ce_namelen(next))
 385                         break;
 386                 if (memcmp(ce->name, next->name, len))
 387                         break;
 388                 o->pos++;
 389         }
 390 }
 391
 392 /*
 393  * The unpack_trees() interface is designed for merging, so
 394  * the different source entries are designed primarily for
 395  * the source trees, with the old index being really mainly
 396  * used for being replaced by the result.
 397  *
 398  * For diffing, the index is more important, and we only have a
 399  * single tree.
 400  *
 401  * We're supposed to return how many index entries we want to skip.
 402  *
 403  * This wrapper makes it all more readable, and takes care of all
 404  * the fairly complex unpack_trees() semantic requirements, including
 405  * the skipping, the path matching, the type conflict cases etc.
 406  */
 407 static int oneway_diff(struct cache_entry **src, struct unpack_trees_options *o)
 408 {
 409         struct cache_entry *idx = src[0];
 410         struct cache_entry *tree = src[1];
 411         struct rev_info *revs = o->unpack_data;
 412
 413         if (idx && ce_stage(idx))
 414                 skip_same_name(idx, o);
 415
 416         /*
 417          * Unpack-trees generates a DF/conflict entry if
 418          * there was a directory in the index and a tree
 419          * in the tree. From a diff standpoint, that's a
 420          * delete of the tree and a create of the file.
 421          */
 422         if (tree == o->df_conflict_entry)
 423                 tree = NULL;
 424
 425         if (ce_path_match(idx ? idx : tree, revs->prune_data))
 426                 do_oneway_diff(o, idx, tree);
 427
 428         return 0;
 429 }
 430
 431 int run_diff_index(struct rev_info *revs, int cached)
 432 {
 433         struct object *ent;
 434         struct tree *tree;
 435         const char *tree_name;
 436         struct unpack_trees_options opts;
 437         struct tree_desc t;
 438
 439         mark_merge_entries();
 440
 441         ent = revs->pending.objects[0].item;
 442         tree_name = revs->pending.objects[0].name;
 443         tree = parse_tree_indirect(ent->sha1);
 444         if (!tree)
 445                 return error("bad tree object %s", tree_name);
 446
 447         memset(&opts, 0, sizeof(opts));
 448         opts.head_idx = 1;
 449         opts.index_only = cached;
 450         opts.diff_index_cached = (cached &&
 451                                   !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 452         opts.merge = 1;
 453         opts.fn = oneway_diff;
 454         opts.unpack_data = revs;
 455         opts.src_index = &the_index;
 456         opts.dst_index = NULL;
 457
 458         init_tree_desc(&t, tree->buffer, tree->size);
 459         if (unpack_trees(1, &t, &opts))
 460                 exit(128);
 461
 462         diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 463         diffcore_std(&revs->diffopt);
 464         diff_flush(&revs->diffopt);
 465         return 0;
 466 }
 467
 468 int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
 469 {
 470         struct tree *tree;
 471         struct rev_info revs;
 472         int i;
 473         struct cache_entry **dst;
 474         struct cache_entry *last = NULL;
 475         struct unpack_trees_options opts;
 476         struct tree_desc t;
 477
 478         /*
 479          * This is used by git-blame to run diff-cache internally;
 480          * it potentially needs to repeatedly run this, so we will
 481          * start by removing the higher order entries the last round
 482          * left behind.
 483          */
 484         dst = active_cache;
 485         for (i = 0; i < active_nr; i++) {
 486                 struct cache_entry *ce = active_cache[i];
 487                 if (ce_stage(ce)) {
 488                         if (last && !strcmp(ce->name, last->name))
 489                                 continue;
 490                         cache_tree_invalidate_path(active_cache_tree,
 491                                                    ce->name);
 492                         last = ce;
 493                         ce->ce_flags |= CE_REMOVE;
 494                 }
 495                 *dst++ = ce;
 496         }
 497         active_nr = dst - active_cache;
 498
 499         init_revisions(&revs, NULL);
 500         revs.prune_data = opt->paths;
 501         tree = parse_tree_indirect(tree_sha1);
 502         if (!tree)
 503                 die("bad tree object %s", sha1_to_hex(tree_sha1));
 504
 505         memset(&opts, 0, sizeof(opts));
 506         opts.head_idx = 1;
 507         opts.index_only = 1;
 508         opts.diff_index_cached = !DIFF_OPT_TST(opt, FIND_COPIES_HARDER);
 509         opts.merge = 1;
 510         opts.fn = oneway_diff;
 511         opts.unpack_data = &revs;
 512         opts.src_index = &the_index;
 513         opts.dst_index = &the_index;
 514
 515         init_tree_desc(&t, tree->buffer, tree->size);
 516         if (unpack_trees(1, &t, &opts))
 517                 exit(128);
 518         return 0;
 519 }
 520
 521 int index_differs_from(const char *def, int diff_flags)
 522 {
 523         struct rev_info rev;
 524
 525         init_revisions(&rev, NULL);
 526         setup_revisions(0, NULL, &rev, def);
 527         DIFF_OPT_SET(&rev.diffopt, QUICK);
 528         DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 529         rev.diffopt.flags |= diff_flags;
 530         run_diff_index(&rev, 1);
 531         if (rev.pending.alloc)
 532                 free(rev.pending.objects);
 533         return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 534 }