cache-tree.c

   1 #define USE_THE_REPOSITORY_VARIABLE
   2
   3 #include "git-compat-util.h"
   4 #include "hex.h"
   5 #include "lockfile.h"
   6 #include "tree.h"
   7 #include "tree-walk.h"
   8 #include "cache-tree.h"
   9 #include "bulk-checkin.h"
  10 #include "object-file.h"
  11 #include "object-store-ll.h"
  12 #include "read-cache-ll.h"
  13 #include "replace-object.h"
  14 #include "repository.h"
  15 #include "promisor-remote.h"
  16 #include "trace.h"
  17 #include "trace2.h"
  18
  19 #ifndef DEBUG_CACHE_TREE
  20 #define DEBUG_CACHE_TREE 0
  21 #endif
  22
  23 struct cache_tree *cache_tree(void)
  24 {
  25         struct cache_tree *it = xcalloc(1, sizeof(struct cache_tree));
  26         it->entry_count = -1;
  27         return it;
  28 }
  29
  30 void cache_tree_free(struct cache_tree **it_p)
  31 {
  32         int i;
  33         struct cache_tree *it = *it_p;
  34
  35         if (!it)
  36                 return;
  37         for (i = 0; i < it->subtree_nr; i++)
  38                 if (it->down[i]) {
  39                         cache_tree_free(&it->down[i]->cache_tree);
  40                         free(it->down[i]);
  41                 }
  42         free(it->down);
  43         free(it);
  44         *it_p = NULL;
  45 }
  46
  47 static int subtree_name_cmp(const char *one, int onelen,
  48                             const char *two, int twolen)
  49 {
  50         if (onelen < twolen)
  51                 return -1;
  52         if (twolen < onelen)
  53                 return 1;
  54         return memcmp(one, two, onelen);
  55 }
  56
  57 int cache_tree_subtree_pos(struct cache_tree *it, const char *path, int pathlen)
  58 {
  59         struct cache_tree_sub **down = it->down;
  60         int lo, hi;
  61         lo = 0;
  62         hi = it->subtree_nr;
  63         while (lo < hi) {
  64                 int mi = lo + (hi - lo) / 2;
  65                 struct cache_tree_sub *mdl = down[mi];
  66                 int cmp = subtree_name_cmp(path, pathlen,
  67                                            mdl->name, mdl->namelen);
  68                 if (!cmp)
  69                         return mi;
  70                 if (cmp < 0)
  71                         hi = mi;
  72                 else
  73                         lo = mi + 1;
  74         }
  75         return -lo-1;
  76 }
  77
  78 static struct cache_tree_sub *find_subtree(struct cache_tree *it,
  79                                            const char *path,
  80                                            int pathlen,
  81                                            int create)
  82 {
  83         struct cache_tree_sub *down;
  84         int pos = cache_tree_subtree_pos(it, path, pathlen);
  85         if (0 <= pos)
  86                 return it->down[pos];
  87         if (!create)
  88                 return NULL;
  89
  90         pos = -pos-1;
  91         ALLOC_GROW(it->down, it->subtree_nr + 1, it->subtree_alloc);
  92         it->subtree_nr++;
  93
  94         FLEX_ALLOC_MEM(down, name, path, pathlen);
  95         down->cache_tree = NULL;
  96         down->namelen = pathlen;
  97
  98         if (pos < it->subtree_nr)
  99                 MOVE_ARRAY(it->down + pos + 1, it->down + pos,
 100                            it->subtree_nr - pos - 1);
 101         it->down[pos] = down;
 102         return down;
 103 }
 104
 105 struct cache_tree_sub *cache_tree_sub(struct cache_tree *it, const char *path)
 106 {
 107         int pathlen = strlen(path);
 108         return find_subtree(it, path, pathlen, 1);
 109 }
 110
 111 static int do_invalidate_path(struct cache_tree *it, const char *path)
 112 {
 113         /* a/b/c
 114          * ==> invalidate self
 115          * ==> find "a", have it invalidate "b/c"
 116          * a
 117          * ==> invalidate self
 118          * ==> if "a" exists as a subtree, remove it.
 119          */
 120         const char *slash;
 121         int namelen;
 122         struct cache_tree_sub *down;
 123
 124 #if DEBUG_CACHE_TREE
 125         fprintf(stderr, "cache-tree invalidate <%s>\n", path);
 126 #endif
 127
 128         if (!it)
 129                 return 0;
 130         slash = strchrnul(path, '/');
 131         namelen = slash - path;
 132         it->entry_count = -1;
 133         if (!*slash) {
 134                 int pos;
 135                 pos = cache_tree_subtree_pos(it, path, namelen);
 136                 if (0 <= pos) {
 137                         cache_tree_free(&it->down[pos]->cache_tree);
 138                         free(it->down[pos]);
 139                         /* 0 1 2 3 4 5
 140                          *       ^     ^subtree_nr = 6
 141                          *       pos
 142                          * move 4 and 5 up one place (2 entries)
 143                          * 2 = 6 - 3 - 1 = subtree_nr - pos - 1
 144                          */
 145                         MOVE_ARRAY(it->down + pos, it->down + pos + 1,
 146                                    it->subtree_nr - pos - 1);
 147                         it->subtree_nr--;
 148                 }
 149                 return 1;
 150         }
 151         down = find_subtree(it, path, namelen, 0);
 152         if (down)
 153                 do_invalidate_path(down->cache_tree, slash + 1);
 154         return 1;
 155 }
 156
 157 void cache_tree_invalidate_path(struct index_state *istate, const char *path)
 158 {
 159         if (do_invalidate_path(istate->cache_tree, path))
 160                 istate->cache_changed |= CACHE_TREE_CHANGED;
 161 }
 162
 163 static int verify_cache(struct index_state *istate, int flags)
 164 {
 165         unsigned i, funny;
 166         int silent = flags & WRITE_TREE_SILENT;
 167
 168         /* Verify that the tree is merged */
 169         funny = 0;
 170         for (i = 0; i < istate->cache_nr; i++) {
 171                 const struct cache_entry *ce = istate->cache[i];
 172                 if (ce_stage(ce)) {
 173                         if (silent)
 174                                 return -1;
 175                         if (10 < ++funny) {
 176                                 fprintf(stderr, "...\n");
 177                                 break;
 178                         }
 179                         fprintf(stderr, "%s: unmerged (%s)\n",
 180                                 ce->name, oid_to_hex(&ce->oid));
 181                 }
 182         }
 183         if (funny)
 184                 return -1;
 185
 186         /* Also verify that the cache does not have path and path/file
 187          * at the same time.  At this point we know the cache has only
 188          * stage 0 entries.
 189          */
 190         funny = 0;
 191         for (i = 0; i + 1 < istate->cache_nr; i++) {
 192                 /* path/file always comes after path because of the way
 193                  * the cache is sorted.  Also path can appear only once,
 194                  * which means conflicting one would immediately follow.
 195                  */
 196                 const struct cache_entry *this_ce = istate->cache[i];
 197                 const struct cache_entry *next_ce = istate->cache[i + 1];
 198                 const char *this_name = this_ce->name;
 199                 const char *next_name = next_ce->name;
 200                 int this_len = ce_namelen(this_ce);
 201                 if (this_len < ce_namelen(next_ce) &&
 202                     next_name[this_len] == '/' &&
 203                     strncmp(this_name, next_name, this_len) == 0) {
 204                         if (10 < ++funny) {
 205                                 fprintf(stderr, "...\n");
 206                                 break;
 207                         }
 208                         fprintf(stderr, "You have both %s and %s\n",
 209                                 this_name, next_name);
 210                 }
 211         }
 212         if (funny)
 213                 return -1;
 214         return 0;
 215 }
 216
 217 static void discard_unused_subtrees(struct cache_tree *it)
 218 {
 219         struct cache_tree_sub **down = it->down;
 220         int nr = it->subtree_nr;
 221         int dst, src;
 222         for (dst = src = 0; src < nr; src++) {
 223                 struct cache_tree_sub *s = down[src];
 224                 if (s->used)
 225                         down[dst++] = s;
 226                 else {
 227                         cache_tree_free(&s->cache_tree);
 228                         free(s);
 229                         it->subtree_nr--;
 230                 }
 231         }
 232 }
 233
 234 int cache_tree_fully_valid(struct cache_tree *it)
 235 {
 236         int i;
 237         if (!it)
 238                 return 0;
 239         if (it->entry_count < 0 || !repo_has_object_file(the_repository, &it->oid))
 240                 return 0;
 241         for (i = 0; i < it->subtree_nr; i++) {
 242                 if (!cache_tree_fully_valid(it->down[i]->cache_tree))
 243                         return 0;
 244         }
 245         return 1;
 246 }
 247
 248 static int must_check_existence(const struct cache_entry *ce)
 249 {
 250         return !(repo_has_promisor_remote(the_repository) && ce_skip_worktree(ce));
 251 }
 252
 253 static int update_one(struct cache_tree *it,
 254                       struct cache_entry **cache,
 255                       int entries,
 256                       const char *base,
 257                       int baselen,
 258                       int *skip_count,
 259                       int flags)
 260 {
 261         struct strbuf buffer;
 262         int missing_ok = flags & WRITE_TREE_MISSING_OK;
 263         int dryrun = flags & WRITE_TREE_DRY_RUN;
 264         int repair = flags & WRITE_TREE_REPAIR;
 265         int to_invalidate = 0;
 266         int i;
 267
 268         assert(!(dryrun && repair));
 269
 270         *skip_count = 0;
 271
 272         /*
 273          * If the first entry of this region is a sparse directory
 274          * entry corresponding exactly to 'base', then this cache_tree
 275          * struct is a "leaf" in the data structure, pointing to the
 276          * tree OID specified in the entry.
 277          */
 278         if (entries > 0) {
 279                 const struct cache_entry *ce = cache[0];
 280
 281                 if (S_ISSPARSEDIR(ce->ce_mode) &&
 282                     ce->ce_namelen == baselen &&
 283                     !strncmp(ce->name, base, baselen)) {
 284                         it->entry_count = 1;
 285                         oidcpy(&it->oid, &ce->oid);
 286                         return 1;
 287                 }
 288         }
 289
 290         if (0 <= it->entry_count && repo_has_object_file(the_repository, &it->oid))
 291                 return it->entry_count;
 292
 293         /*
 294          * We first scan for subtrees and update them; we start by
 295          * marking existing subtrees -- the ones that are unmarked
 296          * should not be in the result.
 297          */
 298         for (i = 0; i < it->subtree_nr; i++)
 299                 it->down[i]->used = 0;
 300
 301         /*
 302          * Find the subtrees and update them.
 303          */
 304         i = 0;
 305         while (i < entries) {
 306                 const struct cache_entry *ce = cache[i];
 307                 struct cache_tree_sub *sub;
 308                 const char *path, *slash;
 309                 int pathlen, sublen, subcnt, subskip;
 310
 311                 path = ce->name;
 312                 pathlen = ce_namelen(ce);
 313                 if (pathlen <= baselen || memcmp(base, path, baselen))
 314                         break; /* at the end of this level */
 315
 316                 slash = strchr(path + baselen, '/');
 317                 if (!slash) {
 318                         i++;
 319                         continue;
 320                 }
 321                 /*
 322                  * a/bbb/c (base = a/, slash = /c)
 323                  * ==>
 324                  * path+baselen = bbb/c, sublen = 3
 325                  */
 326                 sublen = slash - (path + baselen);
 327                 sub = find_subtree(it, path + baselen, sublen, 1);
 328                 if (!sub->cache_tree)
 329                         sub->cache_tree = cache_tree();
 330                 subcnt = update_one(sub->cache_tree,
 331                                     cache + i, entries - i,
 332                                     path,
 333                                     baselen + sublen + 1,
 334                                     &subskip,
 335                                     flags);
 336                 if (subcnt < 0)
 337                         return subcnt;
 338                 if (!subcnt)
 339                         die("index cache-tree records empty sub-tree");
 340                 i += subcnt;
 341                 sub->count = subcnt; /* to be used in the next loop */
 342                 *skip_count += subskip;
 343                 sub->used = 1;
 344         }
 345
 346         discard_unused_subtrees(it);
 347
 348         /*
 349          * Then write out the tree object for this level.
 350          */
 351         strbuf_init(&buffer, 8192);
 352
 353         i = 0;
 354         while (i < entries) {
 355                 const struct cache_entry *ce = cache[i];
 356                 struct cache_tree_sub *sub = NULL;
 357                 const char *path, *slash;
 358                 int pathlen, entlen;
 359                 const struct object_id *oid;
 360                 unsigned mode;
 361                 int expected_missing = 0;
 362                 int contains_ita = 0;
 363                 int ce_missing_ok;
 364
 365                 path = ce->name;
 366                 pathlen = ce_namelen(ce);
 367                 if (pathlen <= baselen || memcmp(base, path, baselen))
 368                         break; /* at the end of this level */
 369
 370                 slash = strchr(path + baselen, '/');
 371                 if (slash) {
 372                         entlen = slash - (path + baselen);
 373                         sub = find_subtree(it, path + baselen, entlen, 0);
 374                         if (!sub)
 375                                 die("cache-tree.c: '%.*s' in '%s' not found",
 376                                     entlen, path + baselen, path);
 377                         i += sub->count;
 378                         oid = &sub->cache_tree->oid;
 379                         mode = S_IFDIR;
 380                         contains_ita = sub->cache_tree->entry_count < 0;
 381                         if (contains_ita) {
 382                                 to_invalidate = 1;
 383                                 expected_missing = 1;
 384                         }
 385                 }
 386                 else {
 387                         oid = &ce->oid;
 388                         mode = ce->ce_mode;
 389                         entlen = pathlen - baselen;
 390                         i++;
 391                 }
 392
 393                 ce_missing_ok = mode == S_IFGITLINK || missing_ok ||
 394                         !must_check_existence(ce);
 395                 if (is_null_oid(oid) ||
 396                     (!ce_missing_ok && !repo_has_object_file(the_repository, oid))) {
 397                         strbuf_release(&buffer);
 398                         if (expected_missing)
 399                                 return -1;
 400                         return error("invalid object %06o %s for '%.*s'",
 401                                 mode, oid_to_hex(oid), entlen+baselen, path);
 402                 }
 403
 404                 /*
 405                  * CE_REMOVE entries are removed before the index is
 406                  * written to disk. Skip them to remain consistent
 407                  * with the future on-disk index.
 408                  */
 409                 if (ce->ce_flags & CE_REMOVE) {
 410                         *skip_count = *skip_count + 1;
 411                         continue;
 412                 }
 413
 414                 /*
 415                  * CE_INTENT_TO_ADD entries exist in on-disk index but
 416                  * they are not part of generated trees. Invalidate up
 417                  * to root to force cache-tree users to read elsewhere.
 418                  */
 419                 if (!sub && ce_intent_to_add(ce)) {
 420                         to_invalidate = 1;
 421                         continue;
 422                 }
 423
 424                 /*
 425                  * "sub" can be an empty tree if all subentries are i-t-a.
 426                  */
 427                 if (contains_ita && is_empty_tree_oid(oid, the_repository->hash_algo))
 428                         continue;
 429
 430                 strbuf_grow(&buffer, entlen + 100);
 431                 strbuf_addf(&buffer, "%o %.*s%c", mode, entlen, path + baselen, '\0');
 432                 strbuf_add(&buffer, oid->hash, the_hash_algo->rawsz);
 433
 434 #if DEBUG_CACHE_TREE
 435                 fprintf(stderr, "cache-tree update-one %o %.*s\n",
 436                         mode, entlen, path + baselen);
 437 #endif
 438         }
 439
 440         if (repair) {
 441                 struct object_id oid;
 442                 hash_object_file(the_hash_algo, buffer.buf, buffer.len,
 443                                  OBJ_TREE, &oid);
 444                 if (repo_has_object_file_with_flags(the_repository, &oid, OBJECT_INFO_SKIP_FETCH_OBJECT))
 445                         oidcpy(&it->oid, &oid);
 446                 else
 447                         to_invalidate = 1;
 448         } else if (dryrun) {
 449                 hash_object_file(the_hash_algo, buffer.buf, buffer.len,
 450                                  OBJ_TREE, &it->oid);
 451         } else if (write_object_file_flags(buffer.buf, buffer.len, OBJ_TREE,
 452                                            &it->oid, NULL, flags & WRITE_TREE_SILENT
 453                                            ? HASH_SILENT : 0)) {
 454                 strbuf_release(&buffer);
 455                 return -1;
 456         }
 457
 458         strbuf_release(&buffer);
 459         it->entry_count = to_invalidate ? -1 : i - *skip_count;
 460 #if DEBUG_CACHE_TREE
 461         fprintf(stderr, "cache-tree update-one (%d ent, %d subtree) %s\n",
 462                 it->entry_count, it->subtree_nr,
 463                 oid_to_hex(&it->oid));
 464 #endif
 465         return i;
 466 }
 467
 468 int cache_tree_update(struct index_state *istate, int flags)
 469 {
 470         int skip, i;
 471
 472         i = verify_cache(istate, flags);
 473
 474         if (i)
 475                 return i;
 476
 477         if (!istate->cache_tree)
 478                 istate->cache_tree = cache_tree();
 479
 480         if (!(flags & WRITE_TREE_MISSING_OK) && repo_has_promisor_remote(the_repository))
 481                 prefetch_cache_entries(istate, must_check_existence);
 482
 483         trace_performance_enter();
 484         trace2_region_enter("cache_tree", "update", the_repository);
 485         begin_odb_transaction();
 486         i = update_one(istate->cache_tree, istate->cache, istate->cache_nr,
 487                        "", 0, &skip, flags);
 488         end_odb_transaction();
 489         trace2_region_leave("cache_tree", "update", the_repository);
 490         trace_performance_leave("cache_tree_update");
 491         if (i < 0)
 492                 return i;
 493         istate->cache_changed |= CACHE_TREE_CHANGED;
 494         return 0;
 495 }
 496
 497 static void write_one(struct strbuf *buffer, struct cache_tree *it,
 498                       const char *path, int pathlen)
 499 {
 500         int i;
 501
 502         /* One "cache-tree" entry consists of the following:
 503          * path (NUL terminated)
 504          * entry_count, subtree_nr ("%d %d\n")
 505          * tree-sha1 (missing if invalid)
 506          * subtree_nr "cache-tree" entries for subtrees.
 507          */
 508         strbuf_grow(buffer, pathlen + 100);
 509         strbuf_add(buffer, path, pathlen);
 510         strbuf_addf(buffer, "%c%d %d\n", 0, it->entry_count, it->subtree_nr);
 511
 512 #if DEBUG_CACHE_TREE
 513         if (0 <= it->entry_count)
 514                 fprintf(stderr, "cache-tree <%.*s> (%d ent, %d subtree) %s\n",
 515                         pathlen, path, it->entry_count, it->subtree_nr,
 516                         oid_to_hex(&it->oid));
 517         else
 518                 fprintf(stderr, "cache-tree <%.*s> (%d subtree) invalid\n",
 519                         pathlen, path, it->subtree_nr);
 520 #endif
 521
 522         if (0 <= it->entry_count) {
 523                 strbuf_add(buffer, it->oid.hash, the_hash_algo->rawsz);
 524         }
 525         for (i = 0; i < it->subtree_nr; i++) {
 526                 struct cache_tree_sub *down = it->down[i];
 527                 if (i) {
 528                         struct cache_tree_sub *prev = it->down[i-1];
 529                         if (subtree_name_cmp(down->name, down->namelen,
 530                                              prev->name, prev->namelen) <= 0)
 531                                 die("fatal - unsorted cache subtree");
 532                 }
 533                 write_one(buffer, down->cache_tree, down->name, down->namelen);
 534         }
 535 }
 536
 537 void cache_tree_write(struct strbuf *sb, struct cache_tree *root)
 538 {
 539         trace2_region_enter("cache_tree", "write", the_repository);
 540         write_one(sb, root, "", 0);
 541         trace2_region_leave("cache_tree", "write", the_repository);
 542 }
 543
 544 static struct cache_tree *read_one(const char **buffer, unsigned long *size_p)
 545 {
 546         const char *buf = *buffer;
 547         unsigned long size = *size_p;
 548         const char *cp;
 549         char *ep;
 550         struct cache_tree *it;
 551         int i, subtree_nr;
 552         const unsigned rawsz = the_hash_algo->rawsz;
 553
 554         it = NULL;
 555         /* skip name, but make sure name exists */
 556         while (size && *buf) {
 557                 size--;
 558                 buf++;
 559         }
 560         if (!size)
 561                 goto free_return;
 562         buf++; size--;
 563         it = cache_tree();
 564
 565         cp = buf;
 566         it->entry_count = strtol(cp, &ep, 10);
 567         if (cp == ep)
 568                 goto free_return;
 569         cp = ep;
 570         subtree_nr = strtol(cp, &ep, 10);
 571         if (cp == ep)
 572                 goto free_return;
 573         while (size && *buf && *buf != '\n') {
 574                 size--;
 575                 buf++;
 576         }
 577         if (!size)
 578                 goto free_return;
 579         buf++; size--;
 580         if (0 <= it->entry_count) {
 581                 if (size < rawsz)
 582                         goto free_return;
 583                 oidread(&it->oid, (const unsigned char *)buf,
 584                         the_repository->hash_algo);
 585                 buf += rawsz;
 586                 size -= rawsz;
 587         }
 588
 589 #if DEBUG_CACHE_TREE
 590         if (0 <= it->entry_count)
 591                 fprintf(stderr, "cache-tree <%s> (%d ent, %d subtree) %s\n",
 592                         *buffer, it->entry_count, subtree_nr,
 593                         oid_to_hex(&it->oid));
 594         else
 595                 fprintf(stderr, "cache-tree <%s> (%d subtrees) invalid\n",
 596                         *buffer, subtree_nr);
 597 #endif
 598
 599         /*
 600          * Just a heuristic -- we do not add directories that often but
 601          * we do not want to have to extend it immediately when we do,
 602          * hence +2.
 603          */
 604         it->subtree_alloc = subtree_nr + 2;
 605         CALLOC_ARRAY(it->down, it->subtree_alloc);
 606         for (i = 0; i < subtree_nr; i++) {
 607                 /* read each subtree */
 608                 struct cache_tree *sub;
 609                 struct cache_tree_sub *subtree;
 610                 const char *name = buf;
 611
 612                 sub = read_one(&buf, &size);
 613                 if (!sub)
 614                         goto free_return;
 615                 subtree = cache_tree_sub(it, name);
 616                 subtree->cache_tree = sub;
 617         }
 618         if (subtree_nr != it->subtree_nr)
 619                 die("cache-tree: internal error");
 620         *buffer = buf;
 621         *size_p = size;
 622         return it;
 623
 624  free_return:
 625         cache_tree_free(&it);
 626         return NULL;
 627 }
 628
 629 struct cache_tree *cache_tree_read(const char *buffer, unsigned long size)
 630 {
 631         struct cache_tree *result;
 632
 633         if (buffer[0])
 634                 return NULL; /* not the whole tree */
 635
 636         trace2_region_enter("cache_tree", "read", the_repository);
 637         result = read_one(&buffer, &size);
 638         trace2_region_leave("cache_tree", "read", the_repository);
 639
 640         return result;
 641 }
 642
 643 static struct cache_tree *cache_tree_find(struct cache_tree *it, const char *path)
 644 {
 645         if (!it)
 646                 return NULL;
 647         while (*path) {
 648                 const char *slash;
 649                 struct cache_tree_sub *sub;
 650
 651                 slash = strchrnul(path, '/');
 652                 /*
 653                  * Between path and slash is the name of the subtree
 654                  * to look for.
 655                  */
 656                 sub = find_subtree(it, path, slash - path, 0);
 657                 if (!sub)
 658                         return NULL;
 659                 it = sub->cache_tree;
 660
 661                 path = slash;
 662                 while (*path == '/')
 663                         path++;
 664         }
 665         return it;
 666 }
 667
 668 static int write_index_as_tree_internal(struct object_id *oid,
 669                                         struct index_state *index_state,
 670                                         int cache_tree_valid,
 671                                         int flags,
 672                                         const char *prefix)
 673 {
 674         if (flags & WRITE_TREE_IGNORE_CACHE_TREE) {
 675                 cache_tree_free(&index_state->cache_tree);
 676                 cache_tree_valid = 0;
 677         }
 678
 679         if (!cache_tree_valid && cache_tree_update(index_state, flags) < 0)
 680                 return WRITE_TREE_UNMERGED_INDEX;
 681
 682         if (prefix) {
 683                 struct cache_tree *subtree;
 684                 subtree = cache_tree_find(index_state->cache_tree, prefix);
 685                 if (!subtree)
 686                         return WRITE_TREE_PREFIX_ERROR;
 687                 oidcpy(oid, &subtree->oid);
 688         }
 689         else
 690                 oidcpy(oid, &index_state->cache_tree->oid);
 691
 692         return 0;
 693 }
 694
 695 struct tree* write_in_core_index_as_tree(struct repository *repo) {
 696         struct object_id o;
 697         int was_valid, ret;
 698
 699         struct index_state *index_state = repo->index;
 700         was_valid = index_state->cache_tree &&
 701                     cache_tree_fully_valid(index_state->cache_tree);
 702
 703         ret = write_index_as_tree_internal(&o, index_state, was_valid, 0, NULL);
 704         if (ret == WRITE_TREE_UNMERGED_INDEX) {
 705                 int i;
 706                 bug("there are unmerged index entries:");
 707                 for (i = 0; i < index_state->cache_nr; i++) {
 708                         const struct cache_entry *ce = index_state->cache[i];
 709                         if (ce_stage(ce))
 710                                 bug("%d %.*s", ce_stage(ce),
 711                                     (int)ce_namelen(ce), ce->name);
 712                 }
 713                 BUG("unmerged index entries when writing in-core index");
 714         }
 715
 716         return lookup_tree(repo, &index_state->cache_tree->oid);
 717 }
 718
 719
 720 int write_index_as_tree(struct object_id *oid, struct index_state *index_state, const char *index_path, int flags, const char *prefix)
 721 {
 722         int entries, was_valid;
 723         struct lock_file lock_file = LOCK_INIT;
 724         int ret;
 725
 726         hold_lock_file_for_update(&lock_file, index_path, LOCK_DIE_ON_ERROR);
 727
 728         entries = read_index_from(index_state, index_path,
 729                                   repo_get_git_dir(the_repository));
 730         if (entries < 0) {
 731                 ret = WRITE_TREE_UNREADABLE_INDEX;
 732                 goto out;
 733         }
 734
 735         was_valid = !(flags & WRITE_TREE_IGNORE_CACHE_TREE) &&
 736                     index_state->cache_tree &&
 737                     cache_tree_fully_valid(index_state->cache_tree);
 738
 739         ret = write_index_as_tree_internal(oid, index_state, was_valid, flags,
 740                                            prefix);
 741         if (!ret && !was_valid) {
 742                 write_locked_index(index_state, &lock_file, COMMIT_LOCK);
 743                 /* Not being able to write is fine -- we are only interested
 744                  * in updating the cache-tree part, and if the next caller
 745                  * ends up using the old index with unupdated cache-tree part
 746                  * it misses the work we did here, but that is just a
 747                  * performance penalty and not a big deal.
 748                  */
 749         }
 750
 751 out:
 752         rollback_lock_file(&lock_file);
 753         return ret;
 754 }
 755
 756 static void prime_cache_tree_sparse_dir(struct cache_tree *it,
 757                                         struct tree *tree)
 758 {
 759
 760         oidcpy(&it->oid, &tree->object.oid);
 761         it->entry_count = 1;
 762 }
 763
 764 static void prime_cache_tree_rec(struct repository *r,
 765                                  struct cache_tree *it,
 766                                  struct tree *tree,
 767                                  struct strbuf *tree_path)
 768 {
 769         struct tree_desc desc;
 770         struct name_entry entry;
 771         int cnt;
 772         size_t base_path_len = tree_path->len;
 773
 774         oidcpy(&it->oid, &tree->object.oid);
 775
 776         init_tree_desc(&desc, &tree->object.oid, tree->buffer, tree->size);
 777         cnt = 0;
 778         while (tree_entry(&desc, &entry)) {
 779                 if (!S_ISDIR(entry.mode))
 780                         cnt++;
 781                 else {
 782                         struct cache_tree_sub *sub;
 783                         struct tree *subtree = lookup_tree(r, &entry.oid);
 784
 785                         if (parse_tree(subtree) < 0)
 786                                 exit(128);
 787                         sub = cache_tree_sub(it, entry.path);
 788                         sub->cache_tree = cache_tree();
 789
 790                         /*
 791                          * Recursively-constructed subtree path is only needed when working
 792                          * in a sparse index (where it's used to determine whether the
 793                          * subtree is a sparse directory in the index).
 794                          */
 795                         if (r->index->sparse_index) {
 796                                 strbuf_setlen(tree_path, base_path_len);
 797                                 strbuf_add(tree_path, entry.path, entry.pathlen);
 798                                 strbuf_addch(tree_path, '/');
 799                         }
 800
 801                         /*
 802                          * If a sparse index is in use, the directory being processed may be
 803                          * sparse. To confirm that, we can check whether an entry with that
 804                          * exact name exists in the index. If it does, the created subtree
 805                          * should be sparse. Otherwise, cache tree expansion should continue
 806                          * as normal.
 807                          */
 808                         if (r->index->sparse_index &&
 809                             index_entry_exists(r->index, tree_path->buf, tree_path->len))
 810                                 prime_cache_tree_sparse_dir(sub->cache_tree, subtree);
 811                         else
 812                                 prime_cache_tree_rec(r, sub->cache_tree, subtree, tree_path);
 813                         cnt += sub->cache_tree->entry_count;
 814                 }
 815         }
 816
 817         it->entry_count = cnt;
 818 }
 819
 820 void prime_cache_tree(struct repository *r,
 821                       struct index_state *istate,
 822                       struct tree *tree)
 823 {
 824         struct strbuf tree_path = STRBUF_INIT;
 825
 826         trace2_region_enter("cache-tree", "prime_cache_tree", r);
 827         cache_tree_free(&istate->cache_tree);
 828         istate->cache_tree = cache_tree();
 829
 830         prime_cache_tree_rec(r, istate->cache_tree, tree, &tree_path);
 831         strbuf_release(&tree_path);
 832         istate->cache_changed |= CACHE_TREE_CHANGED;
 833         trace2_region_leave("cache-tree", "prime_cache_tree", r);
 834 }
 835
 836 /*
 837  * find the cache_tree that corresponds to the current level without
 838  * exploding the full path into textual form.  The root of the
 839  * cache tree is given as "root", and our current level is "info".
 840  * (1) When at root level, info->prev is NULL, so it is "root" itself.
 841  * (2) Otherwise, find the cache_tree that corresponds to one level
 842  *     above us, and find ourselves in there.
 843  */
 844 static struct cache_tree *find_cache_tree_from_traversal(struct cache_tree *root,
 845                                                          struct traverse_info *info)
 846 {
 847         struct cache_tree *our_parent;
 848
 849         if (!info->prev)
 850                 return root;
 851         our_parent = find_cache_tree_from_traversal(root, info->prev);
 852         return cache_tree_find(our_parent, info->name);
 853 }
 854
 855 int cache_tree_matches_traversal(struct cache_tree *root,
 856                                  struct name_entry *ent,
 857                                  struct traverse_info *info)
 858 {
 859         struct cache_tree *it;
 860
 861         it = find_cache_tree_from_traversal(root, info);
 862         it = cache_tree_find(it, ent->path);
 863         if (it && it->entry_count > 0 && oideq(&ent->oid, &it->oid))
 864                 return it->entry_count;
 865         return 0;
 866 }
 867
 868 static void verify_one_sparse(struct index_state *istate,
 869                               struct strbuf *path,
 870                               int pos)
 871 {
 872         struct cache_entry *ce = istate->cache[pos];
 873
 874         if (!S_ISSPARSEDIR(ce->ce_mode))
 875                 BUG("directory '%s' is present in index, but not sparse",
 876                     path->buf);
 877 }
 878
 879 /*
 880  * Returns:
 881  *  0 - Verification completed.
 882  *  1 - Restart verification - a call to ensure_full_index() freed the cache
 883  *      tree that is being verified and verification needs to be restarted from
 884  *      the new toplevel cache tree.
 885  */
 886 static int verify_one(struct repository *r,
 887                       struct index_state *istate,
 888                       struct cache_tree *it,
 889                       struct strbuf *path)
 890 {
 891         int i, pos, len = path->len;
 892         struct strbuf tree_buf = STRBUF_INIT;
 893         struct object_id new_oid;
 894
 895         for (i = 0; i < it->subtree_nr; i++) {
 896                 strbuf_addf(path, "%s/", it->down[i]->name);
 897                 if (verify_one(r, istate, it->down[i]->cache_tree, path))
 898                         return 1;
 899                 strbuf_setlen(path, len);
 900         }
 901
 902         if (it->entry_count < 0 ||
 903             /* no verification on tests (t7003) that replace trees */
 904             lookup_replace_object(r, &it->oid) != &it->oid)
 905                 return 0;
 906
 907         if (path->len) {
 908                 /*
 909                  * If the index is sparse and the cache tree is not
 910                  * index_name_pos() may trigger ensure_full_index() which will
 911                  * free the tree that is being verified.
 912                  */
 913                 int is_sparse = istate->sparse_index;
 914                 pos = index_name_pos(istate, path->buf, path->len);
 915                 if (is_sparse && !istate->sparse_index)
 916                         return 1;
 917
 918                 if (pos >= 0) {
 919                         verify_one_sparse(istate, path, pos);
 920                         return 0;
 921                 }
 922
 923                 pos = -pos - 1;
 924         } else {
 925                 pos = 0;
 926         }
 927
 928         i = 0;
 929         while (i < it->entry_count) {
 930                 struct cache_entry *ce = istate->cache[pos + i];
 931                 const char *slash;
 932                 struct cache_tree_sub *sub = NULL;
 933                 const struct object_id *oid;
 934                 const char *name;
 935                 unsigned mode;
 936                 int entlen;
 937
 938                 if (ce->ce_flags & (CE_STAGEMASK | CE_INTENT_TO_ADD | CE_REMOVE))
 939                         BUG("%s with flags 0x%x should not be in cache-tree",
 940                             ce->name, ce->ce_flags);
 941                 name = ce->name + path->len;
 942                 slash = strchr(name, '/');
 943                 if (slash) {
 944                         entlen = slash - name;
 945                         sub = find_subtree(it, ce->name + path->len, entlen, 0);
 946                         if (!sub || sub->cache_tree->entry_count < 0)
 947                                 BUG("bad subtree '%.*s'", entlen, name);
 948                         oid = &sub->cache_tree->oid;
 949                         mode = S_IFDIR;
 950                         i += sub->cache_tree->entry_count;
 951                 } else {
 952                         oid = &ce->oid;
 953                         mode = ce->ce_mode;
 954                         entlen = ce_namelen(ce) - path->len;
 955                         i++;
 956                 }
 957                 strbuf_addf(&tree_buf, "%o %.*s%c", mode, entlen, name, '\0');
 958                 strbuf_add(&tree_buf, oid->hash, r->hash_algo->rawsz);
 959         }
 960         hash_object_file(r->hash_algo, tree_buf.buf, tree_buf.len, OBJ_TREE,
 961                          &new_oid);
 962         if (!oideq(&new_oid, &it->oid))
 963                 BUG("cache-tree for path %.*s does not match. "
 964                     "Expected %s got %s", len, path->buf,
 965                     oid_to_hex(&new_oid), oid_to_hex(&it->oid));
 966         strbuf_setlen(path, len);
 967         strbuf_release(&tree_buf);
 968         return 0;
 969 }
 970
 971 void cache_tree_verify(struct repository *r, struct index_state *istate)
 972 {
 973         struct strbuf path = STRBUF_INIT;
 974
 975         if (!istate->cache_tree)
 976                 return;
 977         if (verify_one(r, istate, istate->cache_tree, &path)) {
 978                 strbuf_reset(&path);
 979                 if (verify_one(r, istate, istate->cache_tree, &path))
 980                         BUG("ensure_full_index() called twice while verifying cache tree");
 981         }
 982         strbuf_release(&path);
 983 }