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