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