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