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