cache-tree.c

   1 #include "cache.h"
   2 #include "tree.h"
   3 #include "tree-walk.h"
   4 #include "cache-tree.h"
   5
   6 #ifndef DEBUG
   7 #define DEBUG 0
   8 #endif
   9
  10 struct cache_tree *cache_tree(void)
  11 {
  12         struct cache_tree *it = xcalloc(1, sizeof(struct cache_tree));
  13         it->entry_count = -1;
  14         return it;
  15 }
  16
  17 void cache_tree_free(struct cache_tree **it_p)
  18 {
  19         int i;
  20         struct cache_tree *it = *it_p;
  21
  22         if (!it)
  23                 return;
  24         for (i = 0; i < it->subtree_nr; i++)
  25                 if (it->down[i]) {
  26                         cache_tree_free(&it->down[i]->cache_tree);
  27                         free(it->down[i]);
  28                 }
  29         free(it->down);
  30         free(it);
  31         *it_p = NULL;
  32 }
  33
  34 static int subtree_name_cmp(const char *one, int onelen,
  35                             const char *two, int twolen)
  36 {
  37         if (onelen < twolen)
  38                 return -1;
  39         if (twolen < onelen)
  40                 return 1;
  41         return memcmp(one, two, onelen);
  42 }
  43
  44 static int subtree_pos(struct cache_tree *it, const char *path, int pathlen)
  45 {
  46         struct cache_tree_sub **down = it->down;
  47         int lo, hi;
  48         lo = 0;
  49         hi = it->subtree_nr;
  50         while (lo < hi) {
  51                 int mi = (lo + hi) / 2;
  52                 struct cache_tree_sub *mdl = down[mi];
  53                 int cmp = subtree_name_cmp(path, pathlen,
  54                                            mdl->name, mdl->namelen);
  55                 if (!cmp)
  56                         return mi;
  57                 if (cmp < 0)
  58                         hi = mi;
  59                 else
  60                         lo = mi + 1;
  61         }
  62         return -lo-1;
  63 }
  64
  65 static struct cache_tree_sub *find_subtree(struct cache_tree *it,
  66                                            const char *path,
  67                                            int pathlen,
  68                                            int create)
  69 {
  70         struct cache_tree_sub *down;
  71         int pos = subtree_pos(it, path, pathlen);
  72         if (0 <= pos)
  73                 return it->down[pos];
  74         if (!create)
  75                 return NULL;
  76
  77         pos = -pos-1;
  78         ALLOC_GROW(it->down, it->subtree_nr + 1, it->subtree_alloc);
  79         it->subtree_nr++;
  80
  81         down = xmalloc(sizeof(*down) + pathlen + 1);
  82         down->cache_tree = NULL;
  83         down->namelen = pathlen;
  84         memcpy(down->name, path, pathlen);
  85         down->name[pathlen] = 0;
  86
  87         if (pos < it->subtree_nr)
  88                 memmove(it->down + pos + 1,
  89                         it->down + pos,
  90                         sizeof(down) * (it->subtree_nr - pos - 1));
  91         it->down[pos] = down;
  92         return down;
  93 }
  94
  95 struct cache_tree_sub *cache_tree_sub(struct cache_tree *it, const char *path)
  96 {
  97         int pathlen = strlen(path);
  98         return find_subtree(it, path, pathlen, 1);
  99 }
 100
 101 void cache_tree_invalidate_path(struct cache_tree *it, const char *path)
 102 {
 103         /* a/b/c
 104          * ==> invalidate self
 105          * ==> find "a", have it invalidate "b/c"
 106          * a
 107          * ==> invalidate self
 108          * ==> if "a" exists as a subtree, remove it.
 109          */
 110         const char *slash;
 111         int namelen;
 112         struct cache_tree_sub *down;
 113
 114 #if DEBUG
 115         fprintf(stderr, "cache-tree invalidate <%s>\n", path);
 116 #endif
 117
 118         if (!it)
 119                 return;
 120         slash = strchr(path, '/');
 121         it->entry_count = -1;
 122         if (!slash) {
 123                 int pos;
 124                 namelen = strlen(path);
 125                 pos = subtree_pos(it, path, namelen);
 126                 if (0 <= pos) {
 127                         cache_tree_free(&it->down[pos]->cache_tree);
 128                         free(it->down[pos]);
 129                         /* 0 1 2 3 4 5
 130                          *       ^     ^subtree_nr = 6
 131                          *       pos
 132                          * move 4 and 5 up one place (2 entries)
 133                          * 2 = 6 - 3 - 1 = subtree_nr - pos - 1
 134                          */
 135                         memmove(it->down+pos, it->down+pos+1,
 136                                 sizeof(struct cache_tree_sub *) *
 137                                 (it->subtree_nr - pos - 1));
 138                         it->subtree_nr--;
 139                 }
 140                 return;
 141         }
 142         namelen = slash - path;
 143         down = find_subtree(it, path, namelen, 0);
 144         if (down)
 145                 cache_tree_invalidate_path(down->cache_tree, slash + 1);
 146 }
 147
 148 static int verify_cache(const struct cache_entry * const *cache,
 149                         int entries, int flags)
 150 {
 151         int i, funny;
 152         int silent = flags & WRITE_TREE_SILENT;
 153
 154         /* Verify that the tree is merged */
 155         funny = 0;
 156         for (i = 0; i < entries; i++) {
 157                 const struct cache_entry *ce = cache[i];
 158                 if (ce_stage(ce)) {
 159                         if (silent)
 160                                 return -1;
 161                         if (10 < ++funny) {
 162                                 fprintf(stderr, "...\n");
 163                                 break;
 164                         }
 165                         fprintf(stderr, "%s: unmerged (%s)\n",
 166                                 ce->name, sha1_to_hex(ce->sha1));
 167                 }
 168         }
 169         if (funny)
 170                 return -1;
 171
 172         /* Also verify that the cache does not have path and path/file
 173          * at the same time.  At this point we know the cache has only
 174          * stage 0 entries.
 175          */
 176         funny = 0;
 177         for (i = 0; i < entries - 1; i++) {
 178                 /* path/file always comes after path because of the way
 179                  * the cache is sorted.  Also path can appear only once,
 180                  * which means conflicting one would immediately follow.
 181                  */
 182                 const char *this_name = cache[i]->name;
 183                 const char *next_name = cache[i+1]->name;
 184                 int this_len = strlen(this_name);
 185                 if (this_len < strlen(next_name) &&
 186                     strncmp(this_name, next_name, this_len) == 0 &&
 187                     next_name[this_len] == '/') {
 188                         if (10 < ++funny) {
 189                                 fprintf(stderr, "...\n");
 190                                 break;
 191                         }
 192                         fprintf(stderr, "You have both %s and %s\n",
 193                                 this_name, next_name);
 194                 }
 195         }
 196         if (funny)
 197                 return -1;
 198         return 0;
 199 }
 200
 201 static void discard_unused_subtrees(struct cache_tree *it)
 202 {
 203         struct cache_tree_sub **down = it->down;
 204         int nr = it->subtree_nr;
 205         int dst, src;
 206         for (dst = src = 0; src < nr; src++) {
 207                 struct cache_tree_sub *s = down[src];
 208                 if (s->used)
 209                         down[dst++] = s;
 210                 else {
 211                         cache_tree_free(&s->cache_tree);
 212                         free(s);
 213                         it->subtree_nr--;
 214                 }
 215         }
 216 }
 217
 218 int cache_tree_fully_valid(struct cache_tree *it)
 219 {
 220         int i;
 221         if (!it)
 222                 return 0;
 223         if (it->entry_count < 0 || !has_sha1_file(it->sha1))
 224                 return 0;
 225         for (i = 0; i < it->subtree_nr; i++) {
 226                 if (!cache_tree_fully_valid(it->down[i]->cache_tree))
 227                         return 0;
 228         }
 229         return 1;
 230 }
 231
 232 static int update_one(struct cache_tree *it,
 233                       const struct cache_entry * const *cache,
 234                       int entries,
 235                       const char *base,
 236                       int baselen,
 237                       int *skip_count,
 238                       int flags)
 239 {
 240         struct strbuf buffer;
 241         int missing_ok = flags & WRITE_TREE_MISSING_OK;
 242         int dryrun = flags & WRITE_TREE_DRY_RUN;
 243         int to_invalidate = 0;
 244         int i;
 245
 246         *skip_count = 0;
 247
 248         if (0 <= it->entry_count && has_sha1_file(it->sha1))
 249                 return it->entry_count;
 250
 251         /*
 252          * We first scan for subtrees and update them; we start by
 253          * marking existing subtrees -- the ones that are unmarked
 254          * should not be in the result.
 255          */
 256         for (i = 0; i < it->subtree_nr; i++)
 257                 it->down[i]->used = 0;
 258
 259         /*
 260          * Find the subtrees and update them.
 261          */
 262         i = 0;
 263         while (i < entries) {
 264                 const struct cache_entry *ce = cache[i];
 265                 struct cache_tree_sub *sub;
 266                 const char *path, *slash;
 267                 int pathlen, sublen, subcnt, subskip;
 268
 269                 path = ce->name;
 270                 pathlen = ce_namelen(ce);
 271                 if (pathlen <= baselen || memcmp(base, path, baselen))
 272                         break; /* at the end of this level */
 273
 274                 slash = strchr(path + baselen, '/');
 275                 if (!slash) {
 276                         i++;
 277                         continue;
 278                 }
 279                 /*
 280                  * a/bbb/c (base = a/, slash = /c)
 281                  * ==>
 282                  * path+baselen = bbb/c, sublen = 3
 283                  */
 284                 sublen = slash - (path + baselen);
 285                 sub = find_subtree(it, path + baselen, sublen, 1);
 286                 if (!sub->cache_tree)
 287                         sub->cache_tree = cache_tree();
 288                 subcnt = update_one(sub->cache_tree,
 289                                     cache + i, entries - i,
 290                                     path,
 291                                     baselen + sublen + 1,
 292                                     &subskip,
 293                                     flags);
 294                 if (subcnt < 0)
 295                         return subcnt;
 296                 i += subcnt;
 297                 sub->count = subcnt; /* to be used in the next loop */
 298                 *skip_count += subskip;
 299                 sub->used = 1;
 300         }
 301
 302         discard_unused_subtrees(it);
 303
 304         /*
 305          * Then write out the tree object for this level.
 306          */
 307         strbuf_init(&buffer, 8192);
 308
 309         i = 0;
 310         while (i < entries) {
 311                 const struct cache_entry *ce = cache[i];
 312                 struct cache_tree_sub *sub;
 313                 const char *path, *slash;
 314                 int pathlen, entlen;
 315                 const unsigned char *sha1;
 316                 unsigned mode;
 317
 318                 path = ce->name;
 319                 pathlen = ce_namelen(ce);
 320                 if (pathlen <= baselen || memcmp(base, path, baselen))
 321                         break; /* at the end of this level */
 322
 323                 slash = strchr(path + baselen, '/');
 324                 if (slash) {
 325                         entlen = slash - (path + baselen);
 326                         sub = find_subtree(it, path + baselen, entlen, 0);
 327                         if (!sub)
 328                                 die("cache-tree.c: '%.*s' in '%s' not found",
 329                                     entlen, path + baselen, path);
 330                         i += sub->count;
 331                         sha1 = sub->cache_tree->sha1;
 332                         mode = S_IFDIR;
 333                         if (sub->cache_tree->entry_count < 0)
 334                                 to_invalidate = 1;
 335                 }
 336                 else {
 337                         sha1 = ce->sha1;
 338                         mode = ce->ce_mode;
 339                         entlen = pathlen - baselen;
 340                         i++;
 341                 }
 342                 if (mode != S_IFGITLINK && !missing_ok && !has_sha1_file(sha1)) {
 343                         strbuf_release(&buffer);
 344                         return error("invalid object %06o %s for '%.*s'",
 345                                 mode, sha1_to_hex(sha1), entlen+baselen, path);
 346                 }
 347
 348                 /*
 349                  * CE_REMOVE entries are removed before the index is
 350                  * written to disk. Skip them to remain consistent
 351                  * with the future on-disk index.
 352                  */
 353                 if (ce->ce_flags & CE_REMOVE) {
 354                         *skip_count = *skip_count + 1;
 355                         continue;
 356                 }
 357
 358                 /*
 359                  * CE_INTENT_TO_ADD entries exist on on-disk index but
 360                  * they are not part of generated trees. Invalidate up
 361                  * to root to force cache-tree users to read elsewhere.
 362                  */
 363                 if (ce->ce_flags & CE_INTENT_TO_ADD) {
 364                         to_invalidate = 1;
 365                         continue;
 366                 }
 367
 368                 strbuf_grow(&buffer, entlen + 100);
 369                 strbuf_addf(&buffer, "%o %.*s%c", mode, entlen, path + baselen, '\0');
 370                 strbuf_add(&buffer, sha1, 20);
 371
 372 #if DEBUG
 373                 fprintf(stderr, "cache-tree update-one %o %.*s\n",
 374                         mode, entlen, path + baselen);
 375 #endif
 376         }
 377
 378         if (dryrun)
 379                 hash_sha1_file(buffer.buf, buffer.len, tree_type, it->sha1);
 380         else if (write_sha1_file(buffer.buf, buffer.len, tree_type, it->sha1)) {
 381                 strbuf_release(&buffer);
 382                 return -1;
 383         }
 384
 385         strbuf_release(&buffer);
 386         it->entry_count = to_invalidate ? -1 : i - *skip_count;
 387 #if DEBUG
 388         fprintf(stderr, "cache-tree update-one (%d ent, %d subtree) %s\n",
 389                 it->entry_count, it->subtree_nr,
 390                 sha1_to_hex(it->sha1));
 391 #endif
 392         return i;
 393 }
 394
 395 int cache_tree_update(struct cache_tree *it,
 396                       const struct cache_entry * const *cache,
 397                       int entries,
 398                       int flags)
 399 {
 400         int i, skip;
 401         i = verify_cache(cache, entries, flags);
 402         if (i)
 403                 return i;
 404         i = update_one(it, cache, entries, "", 0, &skip, flags);
 405         if (i < 0)
 406                 return i;
 407         return 0;
 408 }
 409
 410 static void write_one(struct strbuf *buffer, struct cache_tree *it,
 411                       const char *path, int pathlen)
 412 {
 413         int i;
 414
 415         /* One "cache-tree" entry consists of the following:
 416          * path (NUL terminated)
 417          * entry_count, subtree_nr ("%d %d\n")
 418          * tree-sha1 (missing if invalid)
 419          * subtree_nr "cache-tree" entries for subtrees.
 420          */
 421         strbuf_grow(buffer, pathlen + 100);
 422         strbuf_add(buffer, path, pathlen);
 423         strbuf_addf(buffer, "%c%d %d\n", 0, it->entry_count, it->subtree_nr);
 424
 425 #if DEBUG
 426         if (0 <= it->entry_count)
 427                 fprintf(stderr, "cache-tree <%.*s> (%d ent, %d subtree) %s\n",
 428                         pathlen, path, it->entry_count, it->subtree_nr,
 429                         sha1_to_hex(it->sha1));
 430         else
 431                 fprintf(stderr, "cache-tree <%.*s> (%d subtree) invalid\n",
 432                         pathlen, path, it->subtree_nr);
 433 #endif
 434
 435         if (0 <= it->entry_count) {
 436                 strbuf_add(buffer, it->sha1, 20);
 437         }
 438         for (i = 0; i < it->subtree_nr; i++) {
 439                 struct cache_tree_sub *down = it->down[i];
 440                 if (i) {
 441                         struct cache_tree_sub *prev = it->down[i-1];
 442                         if (subtree_name_cmp(down->name, down->namelen,
 443                                              prev->name, prev->namelen) <= 0)
 444                                 die("fatal - unsorted cache subtree");
 445                 }
 446                 write_one(buffer, down->cache_tree, down->name, down->namelen);
 447         }
 448 }
 449
 450 void cache_tree_write(struct strbuf *sb, struct cache_tree *root)
 451 {
 452         write_one(sb, root, "", 0);
 453 }
 454
 455 static struct cache_tree *read_one(const char **buffer, unsigned long *size_p)
 456 {
 457         const char *buf = *buffer;
 458         unsigned long size = *size_p;
 459         const char *cp;
 460         char *ep;
 461         struct cache_tree *it;
 462         int i, subtree_nr;
 463
 464         it = NULL;
 465         /* skip name, but make sure name exists */
 466         while (size && *buf) {
 467                 size--;
 468                 buf++;
 469         }
 470         if (!size)
 471                 goto free_return;
 472         buf++; size--;
 473         it = cache_tree();
 474
 475         cp = buf;
 476         it->entry_count = strtol(cp, &ep, 10);
 477         if (cp == ep)
 478                 goto free_return;
 479         cp = ep;
 480         subtree_nr = strtol(cp, &ep, 10);
 481         if (cp == ep)
 482                 goto free_return;
 483         while (size && *buf && *buf != '\n') {
 484                 size--;
 485                 buf++;
 486         }
 487         if (!size)
 488                 goto free_return;
 489         buf++; size--;
 490         if (0 <= it->entry_count) {
 491                 if (size < 20)
 492                         goto free_return;
 493                 hashcpy(it->sha1, (const unsigned char*)buf);
 494                 buf += 20;
 495                 size -= 20;
 496         }
 497
 498 #if DEBUG
 499         if (0 <= it->entry_count)
 500                 fprintf(stderr, "cache-tree <%s> (%d ent, %d subtree) %s\n",
 501                         *buffer, it->entry_count, subtree_nr,
 502                         sha1_to_hex(it->sha1));
 503         else
 504                 fprintf(stderr, "cache-tree <%s> (%d subtrees) invalid\n",
 505                         *buffer, subtree_nr);
 506 #endif
 507
 508         /*
 509          * Just a heuristic -- we do not add directories that often but
 510          * we do not want to have to extend it immediately when we do,
 511          * hence +2.
 512          */
 513         it->subtree_alloc = subtree_nr + 2;
 514         it->down = xcalloc(it->subtree_alloc, sizeof(struct cache_tree_sub *));
 515         for (i = 0; i < subtree_nr; i++) {
 516                 /* read each subtree */
 517                 struct cache_tree *sub;
 518                 struct cache_tree_sub *subtree;
 519                 const char *name = buf;
 520
 521                 sub = read_one(&buf, &size);
 522                 if (!sub)
 523                         goto free_return;
 524                 subtree = cache_tree_sub(it, name);
 525                 subtree->cache_tree = sub;
 526         }
 527         if (subtree_nr != it->subtree_nr)
 528                 die("cache-tree: internal error");
 529         *buffer = buf;
 530         *size_p = size;
 531         return it;
 532
 533  free_return:
 534         cache_tree_free(&it);
 535         return NULL;
 536 }
 537
 538 struct cache_tree *cache_tree_read(const char *buffer, unsigned long size)
 539 {
 540         if (buffer[0])
 541                 return NULL; /* not the whole tree */
 542         return read_one(&buffer, &size);
 543 }
 544
 545 static struct cache_tree *cache_tree_find(struct cache_tree *it, const char *path)
 546 {
 547         if (!it)
 548                 return NULL;
 549         while (*path) {
 550                 const char *slash;
 551                 struct cache_tree_sub *sub;
 552
 553                 slash = strchrnul(path, '/');
 554                 /*
 555                  * Between path and slash is the name of the subtree
 556                  * to look for.
 557                  */
 558                 sub = find_subtree(it, path, slash - path, 0);
 559                 if (!sub)
 560                         return NULL;
 561                 it = sub->cache_tree;
 562
 563                 path = slash;
 564                 while (*path == '/')
 565                         path++;
 566         }
 567         return it;
 568 }
 569
 570 int write_cache_as_tree(unsigned char *sha1, int flags, const char *prefix)
 571 {
 572         int entries, was_valid, newfd;
 573         struct lock_file *lock_file;
 574
 575         /*
 576          * We can't free this memory, it becomes part of a linked list
 577          * parsed atexit()
 578          */
 579         lock_file = xcalloc(1, sizeof(struct lock_file));
 580
 581         newfd = hold_locked_index(lock_file, 1);
 582
 583         entries = read_cache();
 584         if (entries < 0)
 585                 return WRITE_TREE_UNREADABLE_INDEX;
 586         if (flags & WRITE_TREE_IGNORE_CACHE_TREE)
 587                 cache_tree_free(&(active_cache_tree));
 588
 589         if (!active_cache_tree)
 590                 active_cache_tree = cache_tree();
 591
 592         was_valid = cache_tree_fully_valid(active_cache_tree);
 593         if (!was_valid) {
 594                 if (cache_tree_update(active_cache_tree,
 595                                       (const struct cache_entry * const *)active_cache,
 596                                       active_nr, flags) < 0)
 597                         return WRITE_TREE_UNMERGED_INDEX;
 598                 if (0 <= newfd) {
 599                         if (!write_cache(newfd, active_cache, active_nr) &&
 600                             !commit_lock_file(lock_file))
 601                                 newfd = -1;
 602                 }
 603                 /* Not being able to write is fine -- we are only interested
 604                  * in updating the cache-tree part, and if the next caller
 605                  * ends up using the old index with unupdated cache-tree part
 606                  * it misses the work we did here, but that is just a
 607                  * performance penalty and not a big deal.
 608                  */
 609         }
 610
 611         if (prefix) {
 612                 struct cache_tree *subtree =
 613                         cache_tree_find(active_cache_tree, prefix);
 614                 if (!subtree)
 615                         return WRITE_TREE_PREFIX_ERROR;
 616                 hashcpy(sha1, subtree->sha1);
 617         }
 618         else
 619                 hashcpy(sha1, active_cache_tree->sha1);
 620
 621         if (0 <= newfd)
 622                 rollback_lock_file(lock_file);
 623
 624         return 0;
 625 }
 626
 627 static void prime_cache_tree_rec(struct cache_tree *it, struct tree *tree)
 628 {
 629         struct tree_desc desc;
 630         struct name_entry entry;
 631         int cnt;
 632
 633         hashcpy(it->sha1, tree->object.sha1);
 634         init_tree_desc(&desc, tree->buffer, tree->size);
 635         cnt = 0;
 636         while (tree_entry(&desc, &entry)) {
 637                 if (!S_ISDIR(entry.mode))
 638                         cnt++;
 639                 else {
 640                         struct cache_tree_sub *sub;
 641                         struct tree *subtree = lookup_tree(entry.sha1);
 642                         if (!subtree->object.parsed)
 643                                 parse_tree(subtree);
 644                         sub = cache_tree_sub(it, entry.path);
 645                         sub->cache_tree = cache_tree();
 646                         prime_cache_tree_rec(sub->cache_tree, subtree);
 647                         cnt += sub->cache_tree->entry_count;
 648                 }
 649         }
 650         it->entry_count = cnt;
 651 }
 652
 653 void prime_cache_tree(struct cache_tree **it, struct tree *tree)
 654 {
 655         cache_tree_free(it);
 656         *it = cache_tree();
 657         prime_cache_tree_rec(*it, tree);
 658 }
 659
 660 /*
 661  * find the cache_tree that corresponds to the current level without
 662  * exploding the full path into textual form.  The root of the
 663  * cache tree is given as "root", and our current level is "info".
 664  * (1) When at root level, info->prev is NULL, so it is "root" itself.
 665  * (2) Otherwise, find the cache_tree that corresponds to one level
 666  *     above us, and find ourselves in there.
 667  */
 668 static struct cache_tree *find_cache_tree_from_traversal(struct cache_tree *root,
 669                                                          struct traverse_info *info)
 670 {
 671         struct cache_tree *our_parent;
 672
 673         if (!info->prev)
 674                 return root;
 675         our_parent = find_cache_tree_from_traversal(root, info->prev);
 676         return cache_tree_find(our_parent, info->name.path);
 677 }
 678
 679 int cache_tree_matches_traversal(struct cache_tree *root,
 680                                  struct name_entry *ent,
 681                                  struct traverse_info *info)
 682 {
 683         struct cache_tree *it;
 684
 685         it = find_cache_tree_from_traversal(root, info);
 686         it = cache_tree_find(it, ent->path);
 687         if (it && it->entry_count > 0 && !hashcmp(ent->sha1, it->sha1))
 688                 return it->entry_count;
 689         return 0;
 690 }
 691
 692 int update_main_cache_tree(int flags)
 693 {
 694         if (!the_index.cache_tree)
 695                 the_index.cache_tree = cache_tree();
 696         return cache_tree_update(the_index.cache_tree,
 697                                  (const struct cache_entry * const *)the_index.cache,
 698                                  the_index.cache_nr, flags);
 699 }