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
  15 /*
  16  * diff-files
  17  */
  18
  19 /*
  20  * Has the work tree entity been removed?
  21  *
  22  * Return 1 if it was removed from the work tree, 0 if an entity to be
  23  * compared with the cache entry ce still exists (the latter includes
  24  * the case where a directory that is not a submodule repository
  25  * exists for ce that is a submodule -- it is a submodule that is not
  26  * checked out).  Return negative for an error.
  27  */
  28 static int check_removed(const struct cache_entry *ce, struct stat *st)
  29 {
  30         if (lstat(ce->name, st) < 0) {
  31                 if (errno != ENOENT && errno != ENOTDIR)
  32                         return -1;
  33                 return 1;
  34         }
  35         if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
  36                 return 1;
  37         if (S_ISDIR(st->st_mode)) {
  38                 unsigned char sub[20];
  39
  40                 /*
  41                  * If ce is already a gitlink, we can have a plain
  42                  * directory (i.e. the submodule is not checked out),
  43                  * or a checked out submodule.  Either case this is not
  44                  * a case where something was removed from the work tree,
  45                  * so we will return 0.
  46                  *
  47                  * Otherwise, if the directory is not a submodule
  48                  * repository, that means ce which was a blob turned into
  49                  * a directory --- the blob was removed!
  50                  */
  51                 if (!S_ISGITLINK(ce->ce_mode) &&
  52                     resolve_gitlink_ref(ce->name, "HEAD", sub))
  53                         return 1;
  54         }
  55         return 0;
  56 }
  57
  58 int run_diff_files(struct rev_info *revs, unsigned int option)
  59 {
  60         int entries, i;
  61         int diff_unmerged_stage = revs->max_count;
  62         int silent_on_removed = option & DIFF_SILENT_ON_REMOVED;
  63         unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
  64                               ? CE_MATCH_RACY_IS_DIRTY : 0);
  65
  66         diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
  67
  68         if (diff_unmerged_stage < 0)
  69                 diff_unmerged_stage = 2;
  70         entries = active_nr;
  71         for (i = 0; i < entries; i++) {
  72                 struct stat st;
  73                 unsigned int oldmode, newmode;
  74                 struct cache_entry *ce = active_cache[i];
  75                 int changed;
  76                 unsigned dirty_submodule = 0;
  77
  78                 if (DIFF_OPT_TST(&revs->diffopt, QUICK) &&
  79                         DIFF_OPT_TST(&revs->diffopt, HAS_CHANGES))
  80                         break;
  81
  82                 if (!ce_path_match(ce, revs->prune_data))
  83                         continue;
  84
  85                 if (ce_stage(ce)) {
  86                         struct combine_diff_path *dpath;
  87                         int num_compare_stages = 0;
  88                         size_t path_len;
  89
  90                         path_len = ce_namelen(ce);
  91
  92                         dpath = xmalloc(combine_diff_path_size(5, path_len));
  93                         dpath->path = (char *) &(dpath->parent[5]);
  94
  95                         dpath->next = NULL;
  96                         dpath->len = path_len;
  97                         memcpy(dpath->path, ce->name, path_len);
  98                         dpath->path[path_len] = '\0';
  99                         hashclr(dpath->sha1);
 100                         memset(&(dpath->parent[0]), 0,
 101                                sizeof(struct combine_diff_parent)*5);
 102
 103                         changed = check_removed(ce, &st);
 104                         if (!changed)
 105                                 dpath->mode = ce_mode_from_stat(ce, st.st_mode);
 106                         else {
 107                                 if (changed < 0) {
 108                                         perror(ce->name);
 109                                         continue;
 110                                 }
 111                                 if (silent_on_removed)
 112                                         continue;
 113                         }
 114
 115                         while (i < entries) {
 116                                 struct cache_entry *nce = active_cache[i];
 117                                 int stage;
 118
 119                                 if (strcmp(ce->name, nce->name))
 120                                         break;
 121
 122                                 /* Stage #2 (ours) is the first parent,
 123                                  * stage #3 (theirs) is the second.
 124                                  */
 125                                 stage = ce_stage(nce);
 126                                 if (2 <= stage) {
 127                                         int mode = nce->ce_mode;
 128                                         num_compare_stages++;
 129                                         hashcpy(dpath->parent[stage-2].sha1, nce->sha1);
 130                                         dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 131                                         dpath->parent[stage-2].status =
 132                                                 DIFF_STATUS_MODIFIED;
 133                                 }
 134
 135                                 /* diff against the proper unmerged stage */
 136                                 if (stage == diff_unmerged_stage)
 137                                         ce = nce;
 138                                 i++;
 139                         }
 140                         /*
 141                          * Compensate for loop update
 142                          */
 143                         i--;
 144
 145                         if (revs->combine_merges && num_compare_stages == 2) {
 146                                 show_combined_diff(dpath, 2,
 147                                                    revs->dense_combined_merges,
 148                                                    revs);
 149                                 free(dpath);
 150                                 continue;
 151                         }
 152                         free(dpath);
 153                         dpath = NULL;
 154
 155                         /*
 156                          * Show the diff for the 'ce' if we found the one
 157                          * from the desired stage.
 158                          */
 159                         diff_unmerge(&revs->diffopt, ce->name, 0, null_sha1);
 160                         if (ce_stage(ce) != diff_unmerged_stage)
 161                                 continue;
 162                 }
 163
 164                 if ((ce_uptodate(ce) && !S_ISGITLINK(ce->ce_mode)) || ce_skip_worktree(ce))
 165                         continue;
 166
 167                 /* If CE_VALID is set, don't look at workdir for file removal */
 168                 changed = (ce->ce_flags & CE_VALID) ? 0 : check_removed(ce, &st);
 169                 if (changed) {
 170                         if (changed < 0) {
 171                                 perror(ce->name);
 172                                 continue;
 173                         }
 174                         if (silent_on_removed)
 175                                 continue;
 176                         diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 177                                        ce->sha1, ce->name, 0);
 178                         continue;
 179                 }
 180                 changed = ce_match_stat(ce, &st, ce_option);
 181                 if (S_ISGITLINK(ce->ce_mode)
 182                     && (!changed || (revs->diffopt.output_format & DIFF_FORMAT_PATCH))
 183                     && is_submodule_modified(ce->name)) {
 184                         changed = 1;
 185                         dirty_submodule = 1;
 186                 }
 187                 if (!changed) {
 188                         ce_mark_uptodate(ce);
 189                         if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 190                                 continue;
 191                 }
 192                 oldmode = ce->ce_mode;
 193                 newmode = ce_mode_from_stat(ce, st.st_mode);
 194                 diff_change(&revs->diffopt, oldmode, newmode,
 195                             ce->sha1, (changed ? null_sha1 : ce->sha1),
 196                             ce->name, 0, dirty_submodule);
 197
 198         }
 199         diffcore_std(&revs->diffopt);
 200         diff_flush(&revs->diffopt);
 201         return 0;
 202 }
 203
 204 /*
 205  * diff-index
 206  */
 207
 208 /* A file entry went away or appeared */
 209 static void diff_index_show_file(struct rev_info *revs,
 210                                  const char *prefix,
 211                                  struct cache_entry *ce,
 212                                  const unsigned char *sha1, unsigned int mode,
 213                                  unsigned dirty_submodule)
 214 {
 215         diff_addremove(&revs->diffopt, prefix[0], mode,
 216                        sha1, ce->name, dirty_submodule);
 217 }
 218
 219 static int get_stat_data(struct cache_entry *ce,
 220                          const unsigned char **sha1p,
 221                          unsigned int *modep,
 222                          int cached, int match_missing,
 223                          unsigned *dirty_submodule, int output_format)
 224 {
 225         const unsigned char *sha1 = ce->sha1;
 226         unsigned int mode = ce->ce_mode;
 227
 228         if (!cached && !ce_uptodate(ce)) {
 229                 int changed;
 230                 struct stat st;
 231                 changed = check_removed(ce, &st);
 232                 if (changed < 0)
 233                         return -1;
 234                 else if (changed) {
 235                         if (match_missing) {
 236                                 *sha1p = sha1;
 237                                 *modep = mode;
 238                                 return 0;
 239                         }
 240                         return -1;
 241                 }
 242                 changed = ce_match_stat(ce, &st, 0);
 243                 if (S_ISGITLINK(ce->ce_mode)
 244                     && (!changed || (output_format & DIFF_FORMAT_PATCH))
 245                     && is_submodule_modified(ce->name)) {
 246                         changed = 1;
 247                         *dirty_submodule = 1;
 248                 }
 249                 if (changed) {
 250                         mode = ce_mode_from_stat(ce, st.st_mode);
 251                         sha1 = null_sha1;
 252                 }
 253         }
 254
 255         *sha1p = sha1;
 256         *modep = mode;
 257         return 0;
 258 }
 259
 260 static void show_new_file(struct rev_info *revs,
 261                           struct cache_entry *new,
 262                           int cached, int match_missing)
 263 {
 264         const unsigned char *sha1;
 265         unsigned int mode;
 266         unsigned dirty_submodule = 0;
 267
 268         /*
 269          * New file in the index: it might actually be different in
 270          * the working copy.
 271          */
 272         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 273             &dirty_submodule, revs->diffopt.output_format) < 0)
 274                 return;
 275
 276         diff_index_show_file(revs, "+", new, sha1, mode, dirty_submodule);
 277 }
 278
 279 static int show_modified(struct rev_info *revs,
 280                          struct cache_entry *old,
 281                          struct cache_entry *new,
 282                          int report_missing,
 283                          int cached, int match_missing)
 284 {
 285         unsigned int mode, oldmode;
 286         const unsigned char *sha1;
 287         unsigned dirty_submodule = 0;
 288
 289         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 290                           &dirty_submodule, revs->diffopt.output_format) < 0) {
 291                 if (report_missing)
 292                         diff_index_show_file(revs, "-", old,
 293                                              old->sha1, old->ce_mode, 0);
 294                 return -1;
 295         }
 296
 297         if (revs->combine_merges && !cached &&
 298             (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
 299                 struct combine_diff_path *p;
 300                 int pathlen = ce_namelen(new);
 301
 302                 p = xmalloc(combine_diff_path_size(2, pathlen));
 303                 p->path = (char *) &p->parent[2];
 304                 p->next = NULL;
 305                 p->len = pathlen;
 306                 memcpy(p->path, new->name, pathlen);
 307                 p->path[pathlen] = 0;
 308                 p->mode = mode;
 309                 hashclr(p->sha1);
 310                 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 311                 p->parent[0].status = DIFF_STATUS_MODIFIED;
 312                 p->parent[0].mode = new->ce_mode;
 313                 hashcpy(p->parent[0].sha1, new->sha1);
 314                 p->parent[1].status = DIFF_STATUS_MODIFIED;
 315                 p->parent[1].mode = old->ce_mode;
 316                 hashcpy(p->parent[1].sha1, old->sha1);
 317                 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 318                 free(p);
 319                 return 0;
 320         }
 321
 322         oldmode = old->ce_mode;
 323         if (mode == oldmode && !hashcmp(sha1, old->sha1) &&
 324             !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 325                 return 0;
 326
 327         diff_change(&revs->diffopt, oldmode, mode,
 328                     old->sha1, sha1, old->name, 0, dirty_submodule);
 329         return 0;
 330 }
 331
 332 /*
 333  * This gets a mix of an existing index and a tree, one pathname entry
 334  * at a time. The index entry may be a single stage-0 one, but it could
 335  * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 336  * give you the position and number of entries in the index).
 337  */
 338 static void do_oneway_diff(struct unpack_trees_options *o,
 339         struct cache_entry *idx,
 340         struct cache_entry *tree)
 341 {
 342         struct rev_info *revs = o->unpack_data;
 343         int match_missing, cached;
 344
 345         /* if the entry is not checked out, don't examine work tree */
 346         cached = o->index_only ||
 347                 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
 348         /*
 349          * Backward compatibility wart - "diff-index -m" does
 350          * not mean "do not ignore merges", but "match_missing".
 351          *
 352          * But with the revision flag parsing, that's found in
 353          * "!revs->ignore_merges".
 354          */
 355         match_missing = !revs->ignore_merges;
 356
 357         if (cached && idx && ce_stage(idx)) {
 358                 diff_unmerge(&revs->diffopt, idx->name, idx->ce_mode,
 359                              idx->sha1);
 360                 return;
 361         }
 362
 363         /*
 364          * Something added to the tree?
 365          */
 366         if (!tree) {
 367                 show_new_file(revs, idx, cached, match_missing);
 368                 return;
 369         }
 370
 371         /*
 372          * Something removed from the tree?
 373          */
 374         if (!idx) {
 375                 diff_index_show_file(revs, "-", tree, tree->sha1, tree->ce_mode, 0);
 376                 return;
 377         }
 378
 379         /* Show difference between old and new */
 380         show_modified(revs, tree, idx, 1, cached, match_missing);
 381 }
 382
 383 static inline void skip_same_name(struct cache_entry *ce, struct unpack_trees_options *o)
 384 {
 385         int len = ce_namelen(ce);
 386         const struct index_state *index = o->src_index;
 387
 388         while (o->pos < index->cache_nr) {
 389                 struct cache_entry *next = index->cache[o->pos];
 390                 if (len != ce_namelen(next))
 391                         break;
 392                 if (memcmp(ce->name, next->name, len))
 393                         break;
 394                 o->pos++;
 395         }
 396 }
 397
 398 /*
 399  * The unpack_trees() interface is designed for merging, so
 400  * the different source entries are designed primarily for
 401  * the source trees, with the old index being really mainly
 402  * used for being replaced by the result.
 403  *
 404  * For diffing, the index is more important, and we only have a
 405  * single tree.
 406  *
 407  * We're supposed to advance o->pos to skip what we have already processed.
 408  *
 409  * This wrapper makes it all more readable, and takes care of all
 410  * the fairly complex unpack_trees() semantic requirements, including
 411  * the skipping, the path matching, the type conflict cases etc.
 412  */
 413 static int oneway_diff(struct cache_entry **src, struct unpack_trees_options *o)
 414 {
 415         struct cache_entry *idx = src[0];
 416         struct cache_entry *tree = src[1];
 417         struct rev_info *revs = o->unpack_data;
 418
 419         if (idx && ce_stage(idx))
 420                 skip_same_name(idx, o);
 421
 422         /*
 423          * Unpack-trees generates a DF/conflict entry if
 424          * there was a directory in the index and a tree
 425          * in the tree. From a diff standpoint, that's a
 426          * delete of the tree and a create of the file.
 427          */
 428         if (tree == o->df_conflict_entry)
 429                 tree = NULL;
 430
 431         if (ce_path_match(idx ? idx : tree, revs->prune_data))
 432                 do_oneway_diff(o, idx, tree);
 433
 434         return 0;
 435 }
 436
 437 int run_diff_index(struct rev_info *revs, int cached)
 438 {
 439         struct object *ent;
 440         struct tree *tree;
 441         const char *tree_name;
 442         struct unpack_trees_options opts;
 443         struct tree_desc t;
 444
 445         ent = revs->pending.objects[0].item;
 446         tree_name = revs->pending.objects[0].name;
 447         tree = parse_tree_indirect(ent->sha1);
 448         if (!tree)
 449                 return error("bad tree object %s", tree_name);
 450
 451         memset(&opts, 0, sizeof(opts));
 452         opts.head_idx = 1;
 453         opts.index_only = cached;
 454         opts.diff_index_cached = (cached &&
 455                                   !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 456         opts.merge = 1;
 457         opts.fn = oneway_diff;
 458         opts.unpack_data = revs;
 459         opts.src_index = &the_index;
 460         opts.dst_index = NULL;
 461
 462         init_tree_desc(&t, tree->buffer, tree->size);
 463         if (unpack_trees(1, &t, &opts))
 464                 exit(128);
 465
 466         diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 467         diffcore_std(&revs->diffopt);
 468         diff_flush(&revs->diffopt);
 469         return 0;
 470 }
 471
 472 int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
 473 {
 474         struct tree *tree;
 475         struct rev_info revs;
 476         int i;
 477         struct cache_entry **dst;
 478         struct cache_entry *last = NULL;
 479         struct unpack_trees_options opts;
 480         struct tree_desc t;
 481
 482         /*
 483          * This is used by git-blame to run diff-cache internally;
 484          * it potentially needs to repeatedly run this, so we will
 485          * start by removing the higher order entries the last round
 486          * left behind.
 487          */
 488         dst = active_cache;
 489         for (i = 0; i < active_nr; i++) {
 490                 struct cache_entry *ce = active_cache[i];
 491                 if (ce_stage(ce)) {
 492                         if (last && !strcmp(ce->name, last->name))
 493                                 continue;
 494                         cache_tree_invalidate_path(active_cache_tree,
 495                                                    ce->name);
 496                         last = ce;
 497                         ce->ce_flags |= CE_REMOVE;
 498                 }
 499                 *dst++ = ce;
 500         }
 501         active_nr = dst - active_cache;
 502
 503         init_revisions(&revs, NULL);
 504         revs.prune_data = opt->paths;
 505         tree = parse_tree_indirect(tree_sha1);
 506         if (!tree)
 507                 die("bad tree object %s", sha1_to_hex(tree_sha1));
 508
 509         memset(&opts, 0, sizeof(opts));
 510         opts.head_idx = 1;
 511         opts.index_only = 1;
 512         opts.diff_index_cached = !DIFF_OPT_TST(opt, FIND_COPIES_HARDER);
 513         opts.merge = 1;
 514         opts.fn = oneway_diff;
 515         opts.unpack_data = &revs;
 516         opts.src_index = &the_index;
 517         opts.dst_index = &the_index;
 518
 519         init_tree_desc(&t, tree->buffer, tree->size);
 520         if (unpack_trees(1, &t, &opts))
 521                 exit(128);
 522         return 0;
 523 }
 524
 525 int index_differs_from(const char *def, int diff_flags)
 526 {
 527         struct rev_info rev;
 528
 529         init_revisions(&rev, NULL);
 530         setup_revisions(0, NULL, &rev, def);
 531         DIFF_OPT_SET(&rev.diffopt, QUICK);
 532         DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 533         rev.diffopt.flags |= diff_flags;
 534         run_diff_index(&rev, 1);
 535         if (rev.pending.alloc)
 536                 free(rev.pending.objects);
 537         return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 538 }