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 /*
  59  * Has a file changed or has a submodule new commits or a dirty work tree?
  60  *
  61  * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
  62  * option is set, the caller does not only want to know if a submodule is
  63  * modified at all but wants to know all the conditions that are met (new
  64  * commits, untracked content and/or modified content).
  65  */
  66 static int match_stat_with_submodule(struct diff_options *diffopt,
  67                                       struct cache_entry *ce, struct stat *st,
  68                                       unsigned ce_option, unsigned *dirty_submodule)
  69 {
  70         int changed = ce_match_stat(ce, st, ce_option);
  71         if (S_ISGITLINK(ce->ce_mode)) {
  72                 unsigned orig_flags = diffopt->flags;
  73                 if (!DIFF_OPT_TST(diffopt, OVERRIDE_SUBMODULE_CONFIG))
  74                         set_diffopt_flags_from_submodule_config(diffopt, ce->name);
  75                 if (DIFF_OPT_TST(diffopt, IGNORE_SUBMODULES))
  76                         changed = 0;
  77                 else if (!DIFF_OPT_TST(diffopt, IGNORE_DIRTY_SUBMODULES)
  78                     && (!changed || DIFF_OPT_TST(diffopt, DIRTY_SUBMODULES)))
  79                         *dirty_submodule = is_submodule_modified(ce->name, DIFF_OPT_TST(diffopt, IGNORE_UNTRACKED_IN_SUBMODULES));
  80                 diffopt->flags = orig_flags;
  81         }
  82         return changed;
  83 }
  84
  85 int run_diff_files(struct rev_info *revs, unsigned int option)
  86 {
  87         int entries, i;
  88         int diff_unmerged_stage = revs->max_count;
  89         int silent_on_removed = option & DIFF_SILENT_ON_REMOVED;
  90         unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
  91                               ? CE_MATCH_RACY_IS_DIRTY : 0);
  92
  93         diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
  94
  95         if (diff_unmerged_stage < 0)
  96                 diff_unmerged_stage = 2;
  97         entries = active_nr;
  98         for (i = 0; i < entries; i++) {
  99                 struct stat st;
 100                 unsigned int oldmode, newmode;
 101                 struct cache_entry *ce = active_cache[i];
 102                 int changed;
 103                 unsigned dirty_submodule = 0;
 104
 105                 if (DIFF_OPT_TST(&revs->diffopt, QUICK) &&
 106                         DIFF_OPT_TST(&revs->diffopt, HAS_CHANGES))
 107                         break;
 108
 109                 if (!ce_path_match(ce, revs->prune_data))
 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
 119                         path_len = ce_namelen(ce);
 120
 121                         dpath = xmalloc(combine_diff_path_size(5, path_len));
 122                         dpath->path = (char *) &(dpath->parent[5]);
 123
 124                         dpath->next = NULL;
 125                         dpath->len = path_len;
 126                         memcpy(dpath->path, ce->name, path_len);
 127                         dpath->path[path_len] = '\0';
 128                         hashclr(dpath->sha1);
 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                                 if (silent_on_removed)
 141                                         continue;
 142                                 wt_mode = 0;
 143                         }
 144                         dpath->mode = wt_mode;
 145
 146                         while (i < entries) {
 147                                 struct cache_entry *nce = active_cache[i];
 148                                 int stage;
 149
 150                                 if (strcmp(ce->name, nce->name))
 151                                         break;
 152
 153                                 /* Stage #2 (ours) is the first parent,
 154                                  * stage #3 (theirs) is the second.
 155                                  */
 156                                 stage = ce_stage(nce);
 157                                 if (2 <= stage) {
 158                                         int mode = nce->ce_mode;
 159                                         num_compare_stages++;
 160                                         hashcpy(dpath->parent[stage-2].sha1, nce->sha1);
 161                                         dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 162                                         dpath->parent[stage-2].status =
 163                                                 DIFF_STATUS_MODIFIED;
 164                                 }
 165
 166                                 /* diff against the proper unmerged stage */
 167                                 if (stage == diff_unmerged_stage)
 168                                         ce = nce;
 169                                 i++;
 170                         }
 171                         /*
 172                          * Compensate for loop update
 173                          */
 174                         i--;
 175
 176                         if (revs->combine_merges && num_compare_stages == 2) {
 177                                 show_combined_diff(dpath, 2,
 178                                                    revs->dense_combined_merges,
 179                                                    revs);
 180                                 free(dpath);
 181                                 continue;
 182                         }
 183                         free(dpath);
 184                         dpath = NULL;
 185
 186                         /*
 187                          * Show the diff for the 'ce' if we found the one
 188                          * from the desired stage.
 189                          */
 190                         pair = diff_unmerge(&revs->diffopt, ce->name);
 191                         if (wt_mode)
 192                                 pair->two->mode = wt_mode;
 193                         if (ce_stage(ce) != diff_unmerged_stage)
 194                                 continue;
 195                 }
 196
 197                 if (ce_uptodate(ce) || ce_skip_worktree(ce))
 198                         continue;
 199
 200                 /* If CE_VALID is set, don't look at workdir for file removal */
 201                 changed = (ce->ce_flags & CE_VALID) ? 0 : check_removed(ce, &st);
 202                 if (changed) {
 203                         if (changed < 0) {
 204                                 perror(ce->name);
 205                                 continue;
 206                         }
 207                         if (silent_on_removed)
 208                                 continue;
 209                         diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 210                                        ce->sha1, ce->name, 0);
 211                         continue;
 212                 }
 213                 changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
 214                                                     ce_option, &dirty_submodule);
 215                 if (!changed && !dirty_submodule) {
 216                         ce_mark_uptodate(ce);
 217                         if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 218                                 continue;
 219                 }
 220                 oldmode = ce->ce_mode;
 221                 newmode = ce_mode_from_stat(ce, st.st_mode);
 222                 diff_change(&revs->diffopt, oldmode, newmode,
 223                             ce->sha1, (changed ? null_sha1 : ce->sha1),
 224                             ce->name, 0, dirty_submodule);
 225
 226         }
 227         diffcore_std(&revs->diffopt);
 228         diff_flush(&revs->diffopt);
 229         return 0;
 230 }
 231
 232 /*
 233  * diff-index
 234  */
 235
 236 /* A file entry went away or appeared */
 237 static void diff_index_show_file(struct rev_info *revs,
 238                                  const char *prefix,
 239                                  struct cache_entry *ce,
 240                                  const unsigned char *sha1, unsigned int mode,
 241                                  unsigned dirty_submodule)
 242 {
 243         diff_addremove(&revs->diffopt, prefix[0], mode,
 244                        sha1, ce->name, dirty_submodule);
 245 }
 246
 247 static int get_stat_data(struct cache_entry *ce,
 248                          const unsigned char **sha1p,
 249                          unsigned int *modep,
 250                          int cached, int match_missing,
 251                          unsigned *dirty_submodule, struct diff_options *diffopt)
 252 {
 253         const unsigned char *sha1 = ce->sha1;
 254         unsigned int mode = ce->ce_mode;
 255
 256         if (!cached && !ce_uptodate(ce)) {
 257                 int changed;
 258                 struct stat st;
 259                 changed = check_removed(ce, &st);
 260                 if (changed < 0)
 261                         return -1;
 262                 else if (changed) {
 263                         if (match_missing) {
 264                                 *sha1p = sha1;
 265                                 *modep = mode;
 266                                 return 0;
 267                         }
 268                         return -1;
 269                 }
 270                 changed = match_stat_with_submodule(diffopt, ce, &st,
 271                                                     0, dirty_submodule);
 272                 if (changed) {
 273                         mode = ce_mode_from_stat(ce, st.st_mode);
 274                         sha1 = null_sha1;
 275                 }
 276         }
 277
 278         *sha1p = sha1;
 279         *modep = mode;
 280         return 0;
 281 }
 282
 283 static void show_new_file(struct rev_info *revs,
 284                           struct cache_entry *new,
 285                           int cached, int match_missing)
 286 {
 287         const unsigned char *sha1;
 288         unsigned int mode;
 289         unsigned dirty_submodule = 0;
 290
 291         /*
 292          * New file in the index: it might actually be different in
 293          * the working copy.
 294          */
 295         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 296             &dirty_submodule, &revs->diffopt) < 0)
 297                 return;
 298
 299         diff_index_show_file(revs, "+", new, sha1, mode, dirty_submodule);
 300 }
 301
 302 static int show_modified(struct rev_info *revs,
 303                          struct cache_entry *old,
 304                          struct cache_entry *new,
 305                          int report_missing,
 306                          int cached, int match_missing)
 307 {
 308         unsigned int mode, oldmode;
 309         const unsigned char *sha1;
 310         unsigned dirty_submodule = 0;
 311
 312         if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 313                           &dirty_submodule, &revs->diffopt) < 0) {
 314                 if (report_missing)
 315                         diff_index_show_file(revs, "-", old,
 316                                              old->sha1, old->ce_mode, 0);
 317                 return -1;
 318         }
 319
 320         if (revs->combine_merges && !cached &&
 321             (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
 322                 struct combine_diff_path *p;
 323                 int pathlen = ce_namelen(new);
 324
 325                 p = xmalloc(combine_diff_path_size(2, pathlen));
 326                 p->path = (char *) &p->parent[2];
 327                 p->next = NULL;
 328                 p->len = pathlen;
 329                 memcpy(p->path, new->name, pathlen);
 330                 p->path[pathlen] = 0;
 331                 p->mode = mode;
 332                 hashclr(p->sha1);
 333                 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 334                 p->parent[0].status = DIFF_STATUS_MODIFIED;
 335                 p->parent[0].mode = new->ce_mode;
 336                 hashcpy(p->parent[0].sha1, new->sha1);
 337                 p->parent[1].status = DIFF_STATUS_MODIFIED;
 338                 p->parent[1].mode = old->ce_mode;
 339                 hashcpy(p->parent[1].sha1, old->sha1);
 340                 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 341                 free(p);
 342                 return 0;
 343         }
 344
 345         oldmode = old->ce_mode;
 346         if (mode == oldmode && !hashcmp(sha1, old->sha1) && !dirty_submodule &&
 347             !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 348                 return 0;
 349
 350         diff_change(&revs->diffopt, oldmode, mode,
 351                     old->sha1, sha1, old->name, 0, dirty_submodule);
 352         return 0;
 353 }
 354
 355 /*
 356  * This gets a mix of an existing index and a tree, one pathname entry
 357  * at a time. The index entry may be a single stage-0 one, but it could
 358  * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 359  * give you the position and number of entries in the index).
 360  */
 361 static void do_oneway_diff(struct unpack_trees_options *o,
 362         struct cache_entry *idx,
 363         struct cache_entry *tree)
 364 {
 365         struct rev_info *revs = o->unpack_data;
 366         int match_missing, cached;
 367
 368         /* if the entry is not checked out, don't examine work tree */
 369         cached = o->index_only ||
 370                 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
 371         /*
 372          * Backward compatibility wart - "diff-index -m" does
 373          * not mean "do not ignore merges", but "match_missing".
 374          *
 375          * But with the revision flag parsing, that's found in
 376          * "!revs->ignore_merges".
 377          */
 378         match_missing = !revs->ignore_merges;
 379
 380         if (cached && idx && ce_stage(idx)) {
 381                 struct diff_filepair *pair;
 382                 pair = diff_unmerge(&revs->diffopt, idx->name);
 383                 fill_filespec(pair->one, idx->sha1, idx->ce_mode);
 384                 return;
 385         }
 386
 387         /*
 388          * Something added to the tree?
 389          */
 390         if (!tree) {
 391                 show_new_file(revs, idx, cached, match_missing);
 392                 return;
 393         }
 394
 395         /*
 396          * Something removed from the tree?
 397          */
 398         if (!idx) {
 399                 diff_index_show_file(revs, "-", tree, tree->sha1, tree->ce_mode, 0);
 400                 return;
 401         }
 402
 403         /* Show difference between old and new */
 404         show_modified(revs, tree, idx, 1, cached, match_missing);
 405 }
 406
 407 /*
 408  * The unpack_trees() interface is designed for merging, so
 409  * the different source entries are designed primarily for
 410  * the source trees, with the old index being really mainly
 411  * used for being replaced by the result.
 412  *
 413  * For diffing, the index is more important, and we only have a
 414  * single tree.
 415  *
 416  * We're supposed to advance o->pos to skip what we have already processed.
 417  *
 418  * This wrapper makes it all more readable, and takes care of all
 419  * the fairly complex unpack_trees() semantic requirements, including
 420  * the skipping, the path matching, the type conflict cases etc.
 421  */
 422 static int oneway_diff(struct cache_entry **src, struct unpack_trees_options *o)
 423 {
 424         struct cache_entry *idx = src[0];
 425         struct cache_entry *tree = src[1];
 426         struct rev_info *revs = o->unpack_data;
 427
 428         /*
 429          * Unpack-trees generates a DF/conflict entry if
 430          * there was a directory in the index and a tree
 431          * in the tree. From a diff standpoint, that's a
 432          * delete of the tree and a create of the file.
 433          */
 434         if (tree == o->df_conflict_entry)
 435                 tree = NULL;
 436
 437         if (ce_path_match(idx ? idx : tree, revs->prune_data))
 438                 do_oneway_diff(o, idx, tree);
 439
 440         return 0;
 441 }
 442
 443 int run_diff_index(struct rev_info *revs, int cached)
 444 {
 445         struct object *ent;
 446         struct tree *tree;
 447         const char *tree_name;
 448         struct unpack_trees_options opts;
 449         struct tree_desc t;
 450
 451         ent = revs->pending.objects[0].item;
 452         tree_name = revs->pending.objects[0].name;
 453         tree = parse_tree_indirect(ent->sha1);
 454         if (!tree)
 455                 return error("bad tree object %s", tree_name);
 456
 457         memset(&opts, 0, sizeof(opts));
 458         opts.head_idx = 1;
 459         opts.index_only = cached;
 460         opts.diff_index_cached = (cached &&
 461                                   !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 462         opts.merge = 1;
 463         opts.fn = oneway_diff;
 464         opts.unpack_data = revs;
 465         opts.src_index = &the_index;
 466         opts.dst_index = NULL;
 467
 468         init_tree_desc(&t, tree->buffer, tree->size);
 469         if (unpack_trees(1, &t, &opts))
 470                 exit(128);
 471
 472         diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 473         diffcore_fix_diff_index(&revs->diffopt);
 474         diffcore_std(&revs->diffopt);
 475         diff_flush(&revs->diffopt);
 476         return 0;
 477 }
 478
 479 int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
 480 {
 481         struct tree *tree;
 482         struct rev_info revs;
 483         int i;
 484         struct cache_entry **dst;
 485         struct cache_entry *last = NULL;
 486         struct unpack_trees_options opts;
 487         struct tree_desc t;
 488
 489         /*
 490          * This is used by git-blame to run diff-cache internally;
 491          * it potentially needs to repeatedly run this, so we will
 492          * start by removing the higher order entries the last round
 493          * left behind.
 494          */
 495         dst = active_cache;
 496         for (i = 0; i < active_nr; i++) {
 497                 struct cache_entry *ce = active_cache[i];
 498                 if (ce_stage(ce)) {
 499                         if (last && !strcmp(ce->name, last->name))
 500                                 continue;
 501                         cache_tree_invalidate_path(active_cache_tree,
 502                                                    ce->name);
 503                         last = ce;
 504                         ce->ce_flags |= CE_REMOVE;
 505                 }
 506                 *dst++ = ce;
 507         }
 508         active_nr = dst - active_cache;
 509
 510         init_revisions(&revs, NULL);
 511         revs.prune_data = opt->paths;
 512         tree = parse_tree_indirect(tree_sha1);
 513         if (!tree)
 514                 die("bad tree object %s", sha1_to_hex(tree_sha1));
 515
 516         memset(&opts, 0, sizeof(opts));
 517         opts.head_idx = 1;
 518         opts.index_only = 1;
 519         opts.diff_index_cached = !DIFF_OPT_TST(opt, FIND_COPIES_HARDER);
 520         opts.merge = 1;
 521         opts.fn = oneway_diff;
 522         opts.unpack_data = &revs;
 523         opts.src_index = &the_index;
 524         opts.dst_index = &the_index;
 525
 526         init_tree_desc(&t, tree->buffer, tree->size);
 527         if (unpack_trees(1, &t, &opts))
 528                 exit(128);
 529         return 0;
 530 }
 531
 532 int index_differs_from(const char *def, int diff_flags)
 533 {
 534         struct rev_info rev;
 535         struct setup_revision_opt opt;
 536
 537         init_revisions(&rev, NULL);
 538         memset(&opt, 0, sizeof(opt));
 539         opt.def = def;
 540         setup_revisions(0, NULL, &rev, &opt);
 541         DIFF_OPT_SET(&rev.diffopt, QUICK);
 542         DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 543         rev.diffopt.flags |= diff_flags;
 544         run_diff_index(&rev, 1);
 545         if (rev.pending.alloc)
 546                 free(rev.pending.objects);
 547         return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 548 }