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 static int do_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 0;
 120         slash = strchrnul(path, '/');
 121         namelen = slash - path;
 122         it->entry_count = -1;
 123         if (!*slash) {
 124                 int pos;
 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 1;
 141         }
 142         down = find_subtree(it, path, namelen, 0);
 143         if (down)
 144                 do_invalidate_path(down->cache_tree, slash + 1);
 145         return 1;
 146 }
 147
 148 void cache_tree_invalidate_path(struct index_state *istate, const char *path)
 149 {
 150         if (do_invalidate_path(istate->cache_tree, path))
 151                 istate->cache_changed |= CACHE_TREE_CHANGED;
 152 }
 153
 154 static int verify_cache(struct cache_entry **cache,
 155                         int entries, int flags)
 156 {
 157         int i, funny;
 158         int silent = flags & WRITE_TREE_SILENT;
 159
 160         /* Verify that the tree is merged */
 161         funny = 0;
 162         for (i = 0; i < entries; i++) {
 163                 const struct cache_entry *ce = cache[i];
 164                 if (ce_stage(ce)) {
 165                         if (silent)
 166                                 return -1;
 167                         if (10 < ++funny) {
 168                                 fprintf(stderr, "...\n");
 169                                 break;
 170                         }
 171                         fprintf(stderr, "%s: unmerged (%s)\n",
 172                                 ce->name, sha1_to_hex(ce->sha1));
 173                 }
 174         }
 175         if (funny)
 176                 return -1;
 177
 178         /* Also verify that the cache does not have path and path/file
 179          * at the same time.  At this point we know the cache has only
 180          * stage 0 entries.
 181          */
 182         funny = 0;
 183         for (i = 0; i < entries - 1; i++) {
 184                 /* path/file always comes after path because of the way
 185                  * the cache is sorted.  Also path can appear only once,
 186                  * which means conflicting one would immediately follow.
 187                  */
 188                 const char *this_name = cache[i]->name;
 189                 const char *next_name = cache[i+1]->name;
 190                 int this_len = strlen(this_name);
 191                 if (this_len < strlen(next_name) &&
 192                     strncmp(this_name, next_name, this_len) == 0 &&
 193                     next_name[this_len] == '/') {
 194                         if (10 < ++funny) {
 195                                 fprintf(stderr, "...\n");
 196                                 break;
 197                         }
 198                         fprintf(stderr, "You have both %s and %s\n",
 199                                 this_name, next_name);
 200                 }
 201         }
 202         if (funny)
 203                 return -1;
 204         return 0;
 205 }
 206
 207 static void discard_unused_subtrees(struct cache_tree *it)
 208 {
 209         struct cache_tree_sub **down = it->down;
 210         int nr = it->subtree_nr;
 211         int dst, src;
 212         for (dst = src = 0; src < nr; src++) {
 213                 struct cache_tree_sub *s = down[src];
 214                 if (s->used)
 215                         down[dst++] = s;
 216                 else {
 217                         cache_tree_free(&s->cache_tree);
 218                         free(s);
 219                         it->subtree_nr--;
 220                 }
 221         }
 222 }
 223
 224 int cache_tree_fully_valid(struct cache_tree *it)
 225 {
 226         int i;
 227         if (!it)
 228                 return 0;
 229         if (it->entry_count < 0 || !has_sha1_file(it->sha1))
 230                 return 0;
 231         for (i = 0; i < it->subtree_nr; i++) {
 232                 if (!cache_tree_fully_valid(it->down[i]->cache_tree))
 233                         return 0;
 234         }
 235         return 1;
 236 }
 237
 238 static int update_one(struct cache_tree *it,
 239                       struct cache_entry **cache,
 240                       int entries,
 241                       const char *base,
 242                       int baselen,
 243                       int *skip_count,
 244                       int flags)
 245 {
 246         struct strbuf buffer;
 247         int missing_ok = flags & WRITE_TREE_MISSING_OK;
 248         int dryrun = flags & WRITE_TREE_DRY_RUN;
 249         int to_invalidate = 0;
 250         int i;
 251
 252         *skip_count = 0;
 253
 254         if (0 <= it->entry_count && has_sha1_file(it->sha1))
 255                 return it->entry_count;
 256
 257         /*
 258          * We first scan for subtrees and update them; we start by
 259          * marking existing subtrees -- the ones that are unmarked
 260          * should not be in the result.
 261          */
 262         for (i = 0; i < it->subtree_nr; i++)
 263                 it->down[i]->used = 0;
 264
 265         /*
 266          * Find the subtrees and update them.
 267          */
 268         i = 0;
 269         while (i < entries) {
 270                 const struct cache_entry *ce = cache[i];
 271                 struct cache_tree_sub *sub;
 272                 const char *path, *slash;
 273                 int pathlen, sublen, subcnt, subskip;
 274
 275                 path = ce->name;
 276                 pathlen = ce_namelen(ce);
 277                 if (pathlen <= baselen || memcmp(base, path, baselen))
 278                         break; /* at the end of this level */
 279
 280                 slash = strchr(path + baselen, '/');
 281                 if (!slash) {
 282                         i++;
 283                         continue;
 284                 }
 285                 /*
 286                  * a/bbb/c (base = a/, slash = /c)
 287                  * ==>
 288                  * path+baselen = bbb/c, sublen = 3
 289                  */
 290                 sublen = slash - (path + baselen);
 291                 sub = find_subtree(it, path + baselen, sublen, 1);
 292                 if (!sub->cache_tree)
 293                         sub->cache_tree = cache_tree();
 294                 subcnt = update_one(sub->cache_tree,
 295                                     cache + i, entries - i,
 296                                     path,
 297                                     baselen + sublen + 1,
 298                                     &subskip,
 299                                     flags);
 300                 if (subcnt < 0)
 301                         return subcnt;
 302                 if (!subcnt)
 303                         die("index cache-tree records empty sub-tree");
 304                 i += subcnt;
 305                 sub->count = subcnt; /* to be used in the next loop */
 306                 *skip_count += subskip;
 307                 sub->used = 1;
 308         }
 309
 310         discard_unused_subtrees(it);
 311
 312         /*
 313          * Then write out the tree object for this level.
 314          */
 315         strbuf_init(&buffer, 8192);
 316
 317         i = 0;
 318         while (i < entries) {
 319                 const struct cache_entry *ce = cache[i];
 320                 struct cache_tree_sub *sub;
 321                 const char *path, *slash;
 322                 int pathlen, entlen;
 323                 const unsigned char *sha1;
 324                 unsigned mode;
 325
 326                 path = ce->name;
 327                 pathlen = ce_namelen(ce);
 328                 if (pathlen <= baselen || memcmp(base, path, baselen))
 329                         break; /* at the end of this level */
 330
 331                 slash = strchr(path + baselen, '/');
 332                 if (slash) {
 333                         entlen = slash - (path + baselen);
 334                         sub = find_subtree(it, path + baselen, entlen, 0);
 335                         if (!sub)
 336                                 die("cache-tree.c: '%.*s' in '%s' not found",
 337                                     entlen, path + baselen, path);
 338                         i += sub->count;
 339                         sha1 = sub->cache_tree->sha1;
 340                         mode = S_IFDIR;
 341                         if (sub->cache_tree->entry_count < 0)
 342                                 to_invalidate = 1;
 343                 }
 344                 else {
 345                         sha1 = ce->sha1;
 346                         mode = ce->ce_mode;
 347                         entlen = pathlen - baselen;
 348                         i++;
 349                 }
 350                 if (mode != S_IFGITLINK && !missing_ok && !has_sha1_file(sha1)) {
 351                         strbuf_release(&buffer);
 352                         return error("invalid object %06o %s for '%.*s'",
 353                                 mode, sha1_to_hex(sha1), entlen+baselen, path);
 354                 }
 355
 356                 /*
 357                  * CE_REMOVE entries are removed before the index is
 358                  * written to disk. Skip them to remain consistent
 359                  * with the future on-disk index.
 360                  */
 361                 if (ce->ce_flags & CE_REMOVE) {
 362                         *skip_count = *skip_count + 1;
 363                         continue;
 364                 }
 365
 366                 /*
 367                  * CE_INTENT_TO_ADD entries exist on on-disk index but
 368                  * they are not part of generated trees. Invalidate up
 369                  * to root to force cache-tree users to read elsewhere.
 370                  */
 371                 if (ce->ce_flags & CE_INTENT_TO_ADD) {
 372                         to_invalidate = 1;
 373                         continue;
 374                 }
 375
 376                 strbuf_grow(&buffer, entlen + 100);
 377                 strbuf_addf(&buffer, "%o %.*s%c", mode, entlen, path + baselen, '\0');
 378                 strbuf_add(&buffer, sha1, 20);
 379
 380 #if DEBUG
 381                 fprintf(stderr, "cache-tree update-one %o %.*s\n",
 382                         mode, entlen, path + baselen);
 383 #endif
 384         }
 385
 386         if (dryrun)
 387                 hash_sha1_file(buffer.buf, buffer.len, tree_type, it->sha1);
 388         else if (write_sha1_file(buffer.buf, buffer.len, tree_type, it->sha1)) {
 389                 strbuf_release(&buffer);
 390                 return -1;
 391         }
 392
 393         strbuf_release(&buffer);
 394         it->entry_count = to_invalidate ? -1 : i - *skip_count;
 395 #if DEBUG
 396         fprintf(stderr, "cache-tree update-one (%d ent, %d subtree) %s\n",
 397                 it->entry_count, it->subtree_nr,
 398                 sha1_to_hex(it->sha1));
 399 #endif
 400         return i;
 401 }
 402
 403 int cache_tree_update(struct index_state *istate, int flags)
 404 {
 405         struct cache_tree *it = istate->cache_tree;
 406         struct cache_entry **cache = istate->cache;
 407         int entries = istate->cache_nr;
 408         int skip, i = verify_cache(cache, entries, flags);
 409
 410         if (i)
 411                 return i;
 412         i = update_one(it, cache, entries, "", 0, &skip, flags);
 413         if (i < 0)
 414                 return i;
 415         istate->cache_changed |= CACHE_TREE_CHANGED;
 416         return 0;
 417 }
 418
 419 static void write_one(struct strbuf *buffer, struct cache_tree *it,
 420                       const char *path, int pathlen)
 421 {
 422         int i;
 423
 424         /* One "cache-tree" entry consists of the following:
 425          * path (NUL terminated)
 426          * entry_count, subtree_nr ("%d %d\n")
 427          * tree-sha1 (missing if invalid)
 428          * subtree_nr "cache-tree" entries for subtrees.
 429          */
 430         strbuf_grow(buffer, pathlen + 100);
 431         strbuf_add(buffer, path, pathlen);
 432         strbuf_addf(buffer, "%c%d %d\n", 0, it->entry_count, it->subtree_nr);
 433
 434 #if DEBUG
 435         if (0 <= it->entry_count)
 436                 fprintf(stderr, "cache-tree <%.*s> (%d ent, %d subtree) %s\n",
 437                         pathlen, path, it->entry_count, it->subtree_nr,
 438                         sha1_to_hex(it->sha1));
 439         else
 440                 fprintf(stderr, "cache-tree <%.*s> (%d subtree) invalid\n",
 441                         pathlen, path, it->subtree_nr);
 442 #endif
 443
 444         if (0 <= it->entry_count) {
 445                 strbuf_add(buffer, it->sha1, 20);
 446         }
 447         for (i = 0; i < it->subtree_nr; i++) {
 448                 struct cache_tree_sub *down = it->down[i];
 449                 if (i) {
 450                         struct cache_tree_sub *prev = it->down[i-1];
 451                         if (subtree_name_cmp(down->name, down->namelen,
 452                                              prev->name, prev->namelen) <= 0)
 453                                 die("fatal - unsorted cache subtree");
 454                 }
 455                 write_one(buffer, down->cache_tree, down->name, down->namelen);
 456         }
 457 }
 458
 459 void cache_tree_write(struct strbuf *sb, struct cache_tree *root)
 460 {
 461         write_one(sb, root, "", 0);
 462 }
 463
 464 static struct cache_tree *read_one(const char **buffer, unsigned long *size_p)
 465 {
 466         const char *buf = *buffer;
 467         unsigned long size = *size_p;
 468         const char *cp;
 469         char *ep;
 470         struct cache_tree *it;
 471         int i, subtree_nr;
 472
 473         it = NULL;
 474         /* skip name, but make sure name exists */
 475         while (size && *buf) {
 476                 size--;
 477                 buf++;
 478         }
 479         if (!size)
 480                 goto free_return;
 481         buf++; size--;
 482         it = cache_tree();
 483
 484         cp = buf;
 485         it->entry_count = strtol(cp, &ep, 10);
 486         if (cp == ep)
 487                 goto free_return;
 488         cp = ep;
 489         subtree_nr = strtol(cp, &ep, 10);
 490         if (cp == ep)
 491                 goto free_return;
 492         while (size && *buf && *buf != '\n') {
 493                 size--;
 494                 buf++;
 495         }
 496         if (!size)
 497                 goto free_return;
 498         buf++; size--;
 499         if (0 <= it->entry_count) {
 500                 if (size < 20)
 501                         goto free_return;
 502                 hashcpy(it->sha1, (const unsigned char*)buf);
 503                 buf += 20;
 504                 size -= 20;
 505         }
 506
 507 #if DEBUG
 508         if (0 <= it->entry_count)
 509                 fprintf(stderr, "cache-tree <%s> (%d ent, %d subtree) %s\n",
 510                         *buffer, it->entry_count, subtree_nr,
 511                         sha1_to_hex(it->sha1));
 512         else
 513                 fprintf(stderr, "cache-tree <%s> (%d subtrees) invalid\n",
 514                         *buffer, subtree_nr);
 515 #endif
 516
 517         /*
 518          * Just a heuristic -- we do not add directories that often but
 519          * we do not want to have to extend it immediately when we do,
 520          * hence +2.
 521          */
 522         it->subtree_alloc = subtree_nr + 2;
 523         it->down = xcalloc(it->subtree_alloc, sizeof(struct cache_tree_sub *));
 524         for (i = 0; i < subtree_nr; i++) {
 525                 /* read each subtree */
 526                 struct cache_tree *sub;
 527                 struct cache_tree_sub *subtree;
 528                 const char *name = buf;
 529
 530                 sub = read_one(&buf, &size);
 531                 if (!sub)
 532                         goto free_return;
 533                 subtree = cache_tree_sub(it, name);
 534                 subtree->cache_tree = sub;
 535         }
 536         if (subtree_nr != it->subtree_nr)
 537                 die("cache-tree: internal error");
 538         *buffer = buf;
 539         *size_p = size;
 540         return it;
 541
 542  free_return:
 543         cache_tree_free(&it);
 544         return NULL;
 545 }
 546
 547 struct cache_tree *cache_tree_read(const char *buffer, unsigned long size)
 548 {
 549         if (buffer[0])
 550                 return NULL; /* not the whole tree */
 551         return read_one(&buffer, &size);
 552 }
 553
 554 static struct cache_tree *cache_tree_find(struct cache_tree *it, const char *path)
 555 {
 556         if (!it)
 557                 return NULL;
 558         while (*path) {
 559                 const char *slash;
 560                 struct cache_tree_sub *sub;
 561
 562                 slash = strchrnul(path, '/');
 563                 /*
 564                  * Between path and slash is the name of the subtree
 565                  * to look for.
 566                  */
 567                 sub = find_subtree(it, path, slash - path, 0);
 568                 if (!sub)
 569                         return NULL;
 570                 it = sub->cache_tree;
 571
 572                 path = slash;
 573                 while (*path == '/')
 574                         path++;
 575         }
 576         return it;
 577 }
 578
 579 int write_cache_as_tree(unsigned char *sha1, int flags, const char *prefix)
 580 {
 581         int entries, was_valid, newfd;
 582         struct lock_file *lock_file;
 583
 584         /*
 585          * We can't free this memory, it becomes part of a linked list
 586          * parsed atexit()
 587          */
 588         lock_file = xcalloc(1, sizeof(struct lock_file));
 589
 590         newfd = hold_locked_index(lock_file, 1);
 591
 592         entries = read_cache();
 593         if (entries < 0)
 594                 return WRITE_TREE_UNREADABLE_INDEX;
 595         if (flags & WRITE_TREE_IGNORE_CACHE_TREE)
 596                 cache_tree_free(&(active_cache_tree));
 597
 598         if (!active_cache_tree)
 599                 active_cache_tree = cache_tree();
 600
 601         was_valid = cache_tree_fully_valid(active_cache_tree);
 602         if (!was_valid) {
 603                 if (cache_tree_update(&the_index, flags) < 0)
 604                         return WRITE_TREE_UNMERGED_INDEX;
 605                 if (0 <= newfd) {
 606                         if (!write_locked_index(&the_index, lock_file, COMMIT_LOCK))
 607                                 newfd = -1;
 608                 }
 609                 /* Not being able to write is fine -- we are only interested
 610                  * in updating the cache-tree part, and if the next caller
 611                  * ends up using the old index with unupdated cache-tree part
 612                  * it misses the work we did here, but that is just a
 613                  * performance penalty and not a big deal.
 614                  */
 615         }
 616
 617         if (prefix) {
 618                 struct cache_tree *subtree =
 619                         cache_tree_find(active_cache_tree, prefix);
 620                 if (!subtree)
 621                         return WRITE_TREE_PREFIX_ERROR;
 622                 hashcpy(sha1, subtree->sha1);
 623         }
 624         else
 625                 hashcpy(sha1, active_cache_tree->sha1);
 626
 627         if (0 <= newfd)
 628                 rollback_lock_file(lock_file);
 629
 630         return 0;
 631 }
 632
 633 static void prime_cache_tree_rec(struct cache_tree *it, struct tree *tree)
 634 {
 635         struct tree_desc desc;
 636         struct name_entry entry;
 637         int cnt;
 638
 639         hashcpy(it->sha1, tree->object.sha1);
 640         init_tree_desc(&desc, tree->buffer, tree->size);
 641         cnt = 0;
 642         while (tree_entry(&desc, &entry)) {
 643                 if (!S_ISDIR(entry.mode))
 644                         cnt++;
 645                 else {
 646                         struct cache_tree_sub *sub;
 647                         struct tree *subtree = lookup_tree(entry.sha1);
 648                         if (!subtree->object.parsed)
 649                                 parse_tree(subtree);
 650                         sub = cache_tree_sub(it, entry.path);
 651                         sub->cache_tree = cache_tree();
 652                         prime_cache_tree_rec(sub->cache_tree, subtree);
 653                         cnt += sub->cache_tree->entry_count;
 654                 }
 655         }
 656         it->entry_count = cnt;
 657 }
 658
 659 void prime_cache_tree(struct index_state *istate, struct tree *tree)
 660 {
 661         cache_tree_free(&istate->cache_tree);
 662         istate->cache_tree = cache_tree();
 663         prime_cache_tree_rec(istate->cache_tree, tree);
 664         istate->cache_changed |= CACHE_TREE_CHANGED;
 665 }
 666
 667 /*
 668  * find the cache_tree that corresponds to the current level without
 669  * exploding the full path into textual form.  The root of the
 670  * cache tree is given as "root", and our current level is "info".
 671  * (1) When at root level, info->prev is NULL, so it is "root" itself.
 672  * (2) Otherwise, find the cache_tree that corresponds to one level
 673  *     above us, and find ourselves in there.
 674  */
 675 static struct cache_tree *find_cache_tree_from_traversal(struct cache_tree *root,
 676                                                          struct traverse_info *info)
 677 {
 678         struct cache_tree *our_parent;
 679
 680         if (!info->prev)
 681                 return root;
 682         our_parent = find_cache_tree_from_traversal(root, info->prev);
 683         return cache_tree_find(our_parent, info->name.path);
 684 }
 685
 686 int cache_tree_matches_traversal(struct cache_tree *root,
 687                                  struct name_entry *ent,
 688                                  struct traverse_info *info)
 689 {
 690         struct cache_tree *it;
 691
 692         it = find_cache_tree_from_traversal(root, info);
 693         it = cache_tree_find(it, ent->path);
 694         if (it && it->entry_count > 0 && !hashcmp(ent->sha1, it->sha1))
 695                 return it->entry_count;
 696         return 0;
 697 }
 698
 699 int update_main_cache_tree(int flags)
 700 {
 701         if (!the_index.cache_tree)
 702                 the_index.cache_tree = cache_tree();
 703         return cache_tree_update(&the_index, flags);
 704 }