read-cache.c

   1 /*
   2  * GIT - The information manager from hell
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  */
   6 #define NO_THE_INDEX_COMPATIBILITY_MACROS
   7 #include "cache.h"
   8 #include "cache-tree.h"
   9 #include "refs.h"
  10 #include "dir.h"
  11
  12 /* Index extensions.
  13  *
  14  * The first letter should be 'A'..'Z' for extensions that are not
  15  * necessary for a correct operation (i.e. optimization data).
  16  * When new extensions are added that _needs_ to be understood in
  17  * order to correctly interpret the index file, pick character that
  18  * is outside the range, to cause the reader to abort.
  19  */
  20
  21 #define CACHE_EXT(s) ( (s[0]<<24)|(s[1]<<16)|(s[2]<<8)|(s[3]) )
  22 #define CACHE_EXT_TREE 0x54524545       /* "TREE" */
  23
  24 struct index_state the_index;
  25
  26 /*
  27  * This only updates the "non-critical" parts of the directory
  28  * cache, ie the parts that aren't tracked by GIT, and only used
  29  * to validate the cache.
  30  */
  31 void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
  32 {
  33         ce->ce_ctime.sec = htonl(st->st_ctime);
  34         ce->ce_mtime.sec = htonl(st->st_mtime);
  35 #ifdef USE_NSEC
  36         ce->ce_ctime.nsec = htonl(st->st_ctim.tv_nsec);
  37         ce->ce_mtime.nsec = htonl(st->st_mtim.tv_nsec);
  38 #endif
  39         ce->ce_dev = htonl(st->st_dev);
  40         ce->ce_ino = htonl(st->st_ino);
  41         ce->ce_uid = htonl(st->st_uid);
  42         ce->ce_gid = htonl(st->st_gid);
  43         ce->ce_size = htonl(st->st_size);
  44
  45         if (assume_unchanged)
  46                 ce->ce_flags |= htons(CE_VALID);
  47 }
  48
  49 static int ce_compare_data(struct cache_entry *ce, struct stat *st)
  50 {
  51         int match = -1;
  52         int fd = open(ce->name, O_RDONLY);
  53
  54         if (fd >= 0) {
  55                 unsigned char sha1[20];
  56                 if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name))
  57                         match = hashcmp(sha1, ce->sha1);
  58                 /* index_fd() closed the file descriptor already */
  59         }
  60         return match;
  61 }
  62
  63 static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
  64 {
  65         int match = -1;
  66         char *target;
  67         void *buffer;
  68         unsigned long size;
  69         enum object_type type;
  70         int len;
  71
  72         target = xmalloc(expected_size);
  73         len = readlink(ce->name, target, expected_size);
  74         if (len != expected_size) {
  75                 free(target);
  76                 return -1;
  77         }
  78         buffer = read_sha1_file(ce->sha1, &type, &size);
  79         if (!buffer) {
  80                 free(target);
  81                 return -1;
  82         }
  83         if (size == expected_size)
  84                 match = memcmp(buffer, target, size);
  85         free(buffer);
  86         free(target);
  87         return match;
  88 }
  89
  90 static int ce_compare_gitlink(struct cache_entry *ce)
  91 {
  92         unsigned char sha1[20];
  93
  94         /*
  95          * We don't actually require that the .git directory
  96          * under GITLINK directory be a valid git directory. It
  97          * might even be missing (in case nobody populated that
  98          * sub-project).
  99          *
 100          * If so, we consider it always to match.
 101          */
 102         if (resolve_gitlink_ref(ce->name, "HEAD", sha1) < 0)
 103                 return 0;
 104         return hashcmp(sha1, ce->sha1);
 105 }
 106
 107 static int ce_modified_check_fs(struct cache_entry *ce, struct stat *st)
 108 {
 109         switch (st->st_mode & S_IFMT) {
 110         case S_IFREG:
 111                 if (ce_compare_data(ce, st))
 112                         return DATA_CHANGED;
 113                 break;
 114         case S_IFLNK:
 115                 if (ce_compare_link(ce, xsize_t(st->st_size)))
 116                         return DATA_CHANGED;
 117                 break;
 118         case S_IFDIR:
 119                 if (S_ISGITLINK(ntohl(ce->ce_mode)))
 120                         return 0;
 121         default:
 122                 return TYPE_CHANGED;
 123         }
 124         return 0;
 125 }
 126
 127 static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
 128 {
 129         unsigned int changed = 0;
 130
 131         switch (ntohl(ce->ce_mode) & S_IFMT) {
 132         case S_IFREG:
 133                 changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
 134                 /* We consider only the owner x bit to be relevant for
 135                  * "mode changes"
 136                  */
 137                 if (trust_executable_bit &&
 138                     (0100 & (ntohl(ce->ce_mode) ^ st->st_mode)))
 139                         changed |= MODE_CHANGED;
 140                 break;
 141         case S_IFLNK:
 142                 if (!S_ISLNK(st->st_mode) &&
 143                     (has_symlinks || !S_ISREG(st->st_mode)))
 144                         changed |= TYPE_CHANGED;
 145                 break;
 146         case S_IFGITLINK:
 147                 if (!S_ISDIR(st->st_mode))
 148                         changed |= TYPE_CHANGED;
 149                 else if (ce_compare_gitlink(ce))
 150                         changed |= DATA_CHANGED;
 151                 return changed;
 152         case 0: /* Special case: unmerged file in index */
 153                 return MODE_CHANGED | DATA_CHANGED | TYPE_CHANGED;
 154         default:
 155                 die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
 156         }
 157         if (ce->ce_mtime.sec != htonl(st->st_mtime))
 158                 changed |= MTIME_CHANGED;
 159         if (ce->ce_ctime.sec != htonl(st->st_ctime))
 160                 changed |= CTIME_CHANGED;
 161
 162 #ifdef USE_NSEC
 163         /*
 164          * nsec seems unreliable - not all filesystems support it, so
 165          * as long as it is in the inode cache you get right nsec
 166          * but after it gets flushed, you get zero nsec.
 167          */
 168         if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
 169                 changed |= MTIME_CHANGED;
 170         if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
 171                 changed |= CTIME_CHANGED;
 172 #endif
 173
 174         if (ce->ce_uid != htonl(st->st_uid) ||
 175             ce->ce_gid != htonl(st->st_gid))
 176                 changed |= OWNER_CHANGED;
 177         if (ce->ce_ino != htonl(st->st_ino))
 178                 changed |= INODE_CHANGED;
 179
 180 #ifdef USE_STDEV
 181         /*
 182          * st_dev breaks on network filesystems where different
 183          * clients will have different views of what "device"
 184          * the filesystem is on
 185          */
 186         if (ce->ce_dev != htonl(st->st_dev))
 187                 changed |= INODE_CHANGED;
 188 #endif
 189
 190         if (ce->ce_size != htonl(st->st_size))
 191                 changed |= DATA_CHANGED;
 192
 193         return changed;
 194 }
 195
 196 int ie_match_stat(struct index_state *istate,
 197                   struct cache_entry *ce, struct stat *st,
 198                   unsigned int options)
 199 {
 200         unsigned int changed;
 201         int ignore_valid = options & CE_MATCH_IGNORE_VALID;
 202         int assume_racy_is_modified = options & CE_MATCH_RACY_IS_DIRTY;
 203
 204         /*
 205          * If it's marked as always valid in the index, it's
 206          * valid whatever the checked-out copy says.
 207          */
 208         if (!ignore_valid && (ce->ce_flags & htons(CE_VALID)))
 209                 return 0;
 210
 211         changed = ce_match_stat_basic(ce, st);
 212
 213         /*
 214          * Within 1 second of this sequence:
 215          *      echo xyzzy >file && git-update-index --add file
 216          * running this command:
 217          *      echo frotz >file
 218          * would give a falsely clean cache entry.  The mtime and
 219          * length match the cache, and other stat fields do not change.
 220          *
 221          * We could detect this at update-index time (the cache entry
 222          * being registered/updated records the same time as "now")
 223          * and delay the return from git-update-index, but that would
 224          * effectively mean we can make at most one commit per second,
 225          * which is not acceptable.  Instead, we check cache entries
 226          * whose mtime are the same as the index file timestamp more
 227          * carefully than others.
 228          */
 229         if (!changed &&
 230             istate->timestamp &&
 231             istate->timestamp <= ntohl(ce->ce_mtime.sec)) {
 232                 if (assume_racy_is_modified)
 233                         changed |= DATA_CHANGED;
 234                 else
 235                         changed |= ce_modified_check_fs(ce, st);
 236         }
 237
 238         return changed;
 239 }
 240
 241 int ie_modified(struct index_state *istate,
 242                 struct cache_entry *ce, struct stat *st, unsigned int options)
 243 {
 244         int changed, changed_fs;
 245
 246         changed = ie_match_stat(istate, ce, st, options);
 247         if (!changed)
 248                 return 0;
 249         /*
 250          * If the mode or type has changed, there's no point in trying
 251          * to refresh the entry - it's not going to match
 252          */
 253         if (changed & (MODE_CHANGED | TYPE_CHANGED))
 254                 return changed;
 255
 256         /* Immediately after read-tree or update-index --cacheinfo,
 257          * the length field is zero.  For other cases the ce_size
 258          * should match the SHA1 recorded in the index entry.
 259          */
 260         if ((changed & DATA_CHANGED) && ce->ce_size != htonl(0))
 261                 return changed;
 262
 263         changed_fs = ce_modified_check_fs(ce, st);
 264         if (changed_fs)
 265                 return changed | changed_fs;
 266         return 0;
 267 }
 268
 269 int base_name_compare(const char *name1, int len1, int mode1,
 270                       const char *name2, int len2, int mode2)
 271 {
 272         unsigned char c1, c2;
 273         int len = len1 < len2 ? len1 : len2;
 274         int cmp;
 275
 276         cmp = memcmp(name1, name2, len);
 277         if (cmp)
 278                 return cmp;
 279         c1 = name1[len];
 280         c2 = name2[len];
 281         if (!c1 && S_ISDIR(mode1))
 282                 c1 = '/';
 283         if (!c2 && S_ISDIR(mode2))
 284                 c2 = '/';
 285         return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
 286 }
 287
 288 int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
 289 {
 290         int len1 = flags1 & CE_NAMEMASK;
 291         int len2 = flags2 & CE_NAMEMASK;
 292         int len = len1 < len2 ? len1 : len2;
 293         int cmp;
 294
 295         cmp = memcmp(name1, name2, len);
 296         if (cmp)
 297                 return cmp;
 298         if (len1 < len2)
 299                 return -1;
 300         if (len1 > len2)
 301                 return 1;
 302
 303         /* Compare stages  */
 304         flags1 &= CE_STAGEMASK;
 305         flags2 &= CE_STAGEMASK;
 306
 307         if (flags1 < flags2)
 308                 return -1;
 309         if (flags1 > flags2)
 310                 return 1;
 311         return 0;
 312 }
 313
 314 int index_name_pos(struct index_state *istate, const char *name, int namelen)
 315 {
 316         int first, last;
 317
 318         first = 0;
 319         last = istate->cache_nr;
 320         while (last > first) {
 321                 int next = (last + first) >> 1;
 322                 struct cache_entry *ce = istate->cache[next];
 323                 int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
 324                 if (!cmp)
 325                         return next;
 326                 if (cmp < 0) {
 327                         last = next;
 328                         continue;
 329                 }
 330                 first = next+1;
 331         }
 332         return -first-1;
 333 }
 334
 335 /* Remove entry, return true if there are more entries to go.. */
 336 int remove_index_entry_at(struct index_state *istate, int pos)
 337 {
 338         istate->cache_changed = 1;
 339         istate->cache_nr--;
 340         if (pos >= istate->cache_nr)
 341                 return 0;
 342         memmove(istate->cache + pos,
 343                 istate->cache + pos + 1,
 344                 (istate->cache_nr - pos) * sizeof(struct cache_entry *));
 345         return 1;
 346 }
 347
 348 int remove_file_from_index(struct index_state *istate, const char *path)
 349 {
 350         int pos = index_name_pos(istate, path, strlen(path));
 351         if (pos < 0)
 352                 pos = -pos-1;
 353         cache_tree_invalidate_path(istate->cache_tree, path);
 354         while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
 355                 remove_index_entry_at(istate, pos);
 356         return 0;
 357 }
 358
 359 static int compare_name(struct cache_entry *ce, const char *path, int namelen)
 360 {
 361         return namelen != ce_namelen(ce) || memcmp(path, ce->name, namelen);
 362 }
 363
 364 static int index_name_pos_also_unmerged(struct index_state *istate,
 365         const char *path, int namelen)
 366 {
 367         int pos = index_name_pos(istate, path, namelen);
 368         struct cache_entry *ce;
 369
 370         if (pos >= 0)
 371                 return pos;
 372
 373         /* maybe unmerged? */
 374         pos = -1 - pos;
 375         if (pos >= istate->cache_nr ||
 376                         compare_name((ce = istate->cache[pos]), path, namelen))
 377                 return -1;
 378
 379         /* order of preference: stage 2, 1, 3 */
 380         if (ce_stage(ce) == 1 && pos + 1 < istate->cache_nr &&
 381                         ce_stage((ce = istate->cache[pos + 1])) == 2 &&
 382                         !compare_name(ce, path, namelen))
 383                 pos++;
 384         return pos;
 385 }
 386
 387 int add_file_to_index(struct index_state *istate, const char *path, int verbose)
 388 {
 389         int size, namelen, pos;
 390         struct stat st;
 391         struct cache_entry *ce;
 392         unsigned ce_option = CE_MATCH_IGNORE_VALID|CE_MATCH_RACY_IS_DIRTY;
 393
 394         if (lstat(path, &st))
 395                 die("%s: unable to stat (%s)", path, strerror(errno));
 396
 397         if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode) && !S_ISDIR(st.st_mode))
 398                 die("%s: can only add regular files, symbolic links or git-directories", path);
 399
 400         namelen = strlen(path);
 401         if (S_ISDIR(st.st_mode)) {
 402                 while (namelen && path[namelen-1] == '/')
 403                         namelen--;
 404         }
 405         size = cache_entry_size(namelen);
 406         ce = xcalloc(1, size);
 407         memcpy(ce->name, path, namelen);
 408         ce->ce_flags = htons(namelen);
 409         fill_stat_cache_info(ce, &st);
 410
 411         if (trust_executable_bit && has_symlinks)
 412                 ce->ce_mode = create_ce_mode(st.st_mode);
 413         else {
 414                 /* If there is an existing entry, pick the mode bits and type
 415                  * from it, otherwise assume unexecutable regular file.
 416                  */
 417                 struct cache_entry *ent;
 418                 int pos = index_name_pos_also_unmerged(istate, path, namelen);
 419
 420                 ent = (0 <= pos) ? istate->cache[pos] : NULL;
 421                 ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
 422         }
 423
 424         pos = index_name_pos(istate, ce->name, namelen);
 425         if (0 <= pos &&
 426             !ce_stage(istate->cache[pos]) &&
 427             !ie_match_stat(istate, istate->cache[pos], &st, ce_option)) {
 428                 /* Nothing changed, really */
 429                 free(ce);
 430                 return 0;
 431         }
 432
 433         if (index_path(ce->sha1, path, &st, 1))
 434                 die("unable to index file %s", path);
 435         if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
 436                 die("unable to add %s to index",path);
 437         if (verbose)
 438                 printf("add '%s'\n", path);
 439         return 0;
 440 }
 441
 442 struct cache_entry *make_cache_entry(unsigned int mode,
 443                 const unsigned char *sha1, const char *path, int stage,
 444                 int refresh)
 445 {
 446         int size, len;
 447         struct cache_entry *ce;
 448
 449         if (!verify_path(path))
 450                 return NULL;
 451
 452         len = strlen(path);
 453         size = cache_entry_size(len);
 454         ce = xcalloc(1, size);
 455
 456         hashcpy(ce->sha1, sha1);
 457         memcpy(ce->name, path, len);
 458         ce->ce_flags = create_ce_flags(len, stage);
 459         ce->ce_mode = create_ce_mode(mode);
 460
 461         if (refresh)
 462                 return refresh_cache_entry(ce, 0);
 463
 464         return ce;
 465 }
 466
 467 int ce_same_name(struct cache_entry *a, struct cache_entry *b)
 468 {
 469         int len = ce_namelen(a);
 470         return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
 471 }
 472
 473 int ce_path_match(const struct cache_entry *ce, const char **pathspec)
 474 {
 475         const char *match, *name;
 476         int len;
 477
 478         if (!pathspec)
 479                 return 1;
 480
 481         len = ce_namelen(ce);
 482         name = ce->name;
 483         while ((match = *pathspec++) != NULL) {
 484                 int matchlen = strlen(match);
 485                 if (matchlen > len)
 486                         continue;
 487                 if (memcmp(name, match, matchlen))
 488                         continue;
 489                 if (matchlen && name[matchlen-1] == '/')
 490                         return 1;
 491                 if (name[matchlen] == '/' || !name[matchlen])
 492                         return 1;
 493                 if (!matchlen)
 494                         return 1;
 495         }
 496         return 0;
 497 }
 498
 499 /*
 500  * We fundamentally don't like some paths: we don't want
 501  * dot or dot-dot anywhere, and for obvious reasons don't
 502  * want to recurse into ".git" either.
 503  *
 504  * Also, we don't want double slashes or slashes at the
 505  * end that can make pathnames ambiguous.
 506  */
 507 static int verify_dotfile(const char *rest)
 508 {
 509         /*
 510          * The first character was '.', but that
 511          * has already been discarded, we now test
 512          * the rest.
 513          */
 514         switch (*rest) {
 515         /* "." is not allowed */
 516         case '\0': case '/':
 517                 return 0;
 518
 519         /*
 520          * ".git" followed by  NUL or slash is bad. This
 521          * shares the path end test with the ".." case.
 522          */
 523         case 'g':
 524                 if (rest[1] != 'i')
 525                         break;
 526                 if (rest[2] != 't')
 527                         break;
 528                 rest += 2;
 529         /* fallthrough */
 530         case '.':
 531                 if (rest[1] == '\0' || rest[1] == '/')
 532                         return 0;
 533         }
 534         return 1;
 535 }
 536
 537 int verify_path(const char *path)
 538 {
 539         char c;
 540
 541         goto inside;
 542         for (;;) {
 543                 if (!c)
 544                         return 1;
 545                 if (c == '/') {
 546 inside:
 547                         c = *path++;
 548                         switch (c) {
 549                         default:
 550                                 continue;
 551                         case '/': case '\0':
 552                                 break;
 553                         case '.':
 554                                 if (verify_dotfile(path))
 555                                         continue;
 556                         }
 557                         return 0;
 558                 }
 559                 c = *path++;
 560         }
 561 }
 562
 563 /*
 564  * Do we have another file that has the beginning components being a
 565  * proper superset of the name we're trying to add?
 566  */
 567 static int has_file_name(struct index_state *istate,
 568                          const struct cache_entry *ce, int pos, int ok_to_replace)
 569 {
 570         int retval = 0;
 571         int len = ce_namelen(ce);
 572         int stage = ce_stage(ce);
 573         const char *name = ce->name;
 574
 575         while (pos < istate->cache_nr) {
 576                 struct cache_entry *p = istate->cache[pos++];
 577
 578                 if (len >= ce_namelen(p))
 579                         break;
 580                 if (memcmp(name, p->name, len))
 581                         break;
 582                 if (ce_stage(p) != stage)
 583                         continue;
 584                 if (p->name[len] != '/')
 585                         continue;
 586                 if (!ce_stage(p) && !p->ce_mode)
 587                         continue;
 588                 retval = -1;
 589                 if (!ok_to_replace)
 590                         break;
 591                 remove_index_entry_at(istate, --pos);
 592         }
 593         return retval;
 594 }
 595
 596 /*
 597  * Do we have another file with a pathname that is a proper
 598  * subset of the name we're trying to add?
 599  */
 600 static int has_dir_name(struct index_state *istate,
 601                         const struct cache_entry *ce, int pos, int ok_to_replace)
 602 {
 603         int retval = 0;
 604         int stage = ce_stage(ce);
 605         const char *name = ce->name;
 606         const char *slash = name + ce_namelen(ce);
 607
 608         for (;;) {
 609                 int len;
 610
 611                 for (;;) {
 612                         if (*--slash == '/')
 613                                 break;
 614                         if (slash <= ce->name)
 615                                 return retval;
 616                 }
 617                 len = slash - name;
 618
 619                 pos = index_name_pos(istate, name, ntohs(create_ce_flags(len, stage)));
 620                 if (pos >= 0) {
 621                         /*
 622                          * Found one, but not so fast.  This could
 623                          * be a marker that says "I was here, but
 624                          * I am being removed".  Such an entry is
 625                          * not a part of the resulting tree, and
 626                          * it is Ok to have a directory at the same
 627                          * path.
 628                          */
 629                         if (stage || istate->cache[pos]->ce_mode) {
 630                                 retval = -1;
 631                                 if (!ok_to_replace)
 632                                         break;
 633                                 remove_index_entry_at(istate, pos);
 634                                 continue;
 635                         }
 636                 }
 637                 else
 638                         pos = -pos-1;
 639
 640                 /*
 641                  * Trivial optimization: if we find an entry that
 642                  * already matches the sub-directory, then we know
 643                  * we're ok, and we can exit.
 644                  */
 645                 while (pos < istate->cache_nr) {
 646                         struct cache_entry *p = istate->cache[pos];
 647                         if ((ce_namelen(p) <= len) ||
 648                             (p->name[len] != '/') ||
 649                             memcmp(p->name, name, len))
 650                                 break; /* not our subdirectory */
 651                         if (ce_stage(p) == stage && (stage || p->ce_mode))
 652                                 /* p is at the same stage as our entry, and
 653                                  * is a subdirectory of what we are looking
 654                                  * at, so we cannot have conflicts at our
 655                                  * level or anything shorter.
 656                                  */
 657                                 return retval;
 658                         pos++;
 659                 }
 660         }
 661         return retval;
 662 }
 663
 664 /* We may be in a situation where we already have path/file and path
 665  * is being added, or we already have path and path/file is being
 666  * added.  Either one would result in a nonsense tree that has path
 667  * twice when git-write-tree tries to write it out.  Prevent it.
 668  *
 669  * If ok-to-replace is specified, we remove the conflicting entries
 670  * from the cache so the caller should recompute the insert position.
 671  * When this happens, we return non-zero.
 672  */
 673 static int check_file_directory_conflict(struct index_state *istate,
 674                                          const struct cache_entry *ce,
 675                                          int pos, int ok_to_replace)
 676 {
 677         int retval;
 678
 679         /*
 680          * When ce is an "I am going away" entry, we allow it to be added
 681          */
 682         if (!ce_stage(ce) && !ce->ce_mode)
 683                 return 0;
 684
 685         /*
 686          * We check if the path is a sub-path of a subsequent pathname
 687          * first, since removing those will not change the position
 688          * in the array.
 689          */
 690         retval = has_file_name(istate, ce, pos, ok_to_replace);
 691
 692         /*
 693          * Then check if the path might have a clashing sub-directory
 694          * before it.
 695          */
 696         return retval + has_dir_name(istate, ce, pos, ok_to_replace);
 697 }
 698
 699 static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
 700 {
 701         int pos;
 702         int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
 703         int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
 704         int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
 705
 706         cache_tree_invalidate_path(istate->cache_tree, ce->name);
 707         pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 708
 709         /* existing match? Just replace it. */
 710         if (pos >= 0) {
 711                 istate->cache_changed = 1;
 712                 istate->cache[pos] = ce;
 713                 return 0;
 714         }
 715         pos = -pos-1;
 716
 717         /*
 718          * Inserting a merged entry ("stage 0") into the index
 719          * will always replace all non-merged entries..
 720          */
 721         if (pos < istate->cache_nr && ce_stage(ce) == 0) {
 722                 while (ce_same_name(istate->cache[pos], ce)) {
 723                         ok_to_add = 1;
 724                         if (!remove_index_entry_at(istate, pos))
 725                                 break;
 726                 }
 727         }
 728
 729         if (!ok_to_add)
 730                 return -1;
 731         if (!verify_path(ce->name))
 732                 return -1;
 733
 734         if (!skip_df_check &&
 735             check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
 736                 if (!ok_to_replace)
 737                         return error("'%s' appears as both a file and as a directory",
 738                                      ce->name);
 739                 pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 740                 pos = -pos-1;
 741         }
 742         return pos + 1;
 743 }
 744
 745 int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
 746 {
 747         int pos;
 748
 749         if (option & ADD_CACHE_JUST_APPEND)
 750                 pos = istate->cache_nr;
 751         else {
 752                 int ret;
 753                 ret = add_index_entry_with_check(istate, ce, option);
 754                 if (ret <= 0)
 755                         return ret;
 756                 pos = ret - 1;
 757         }
 758
 759         /* Make sure the array is big enough .. */
 760         if (istate->cache_nr == istate->cache_alloc) {
 761                 istate->cache_alloc = alloc_nr(istate->cache_alloc);
 762                 istate->cache = xrealloc(istate->cache,
 763                                         istate->cache_alloc * sizeof(struct cache_entry *));
 764         }
 765
 766         /* Add it in.. */
 767         istate->cache_nr++;
 768         if (istate->cache_nr > pos + 1)
 769                 memmove(istate->cache + pos + 1,
 770                         istate->cache + pos,
 771                         (istate->cache_nr - pos - 1) * sizeof(ce));
 772         istate->cache[pos] = ce;
 773         istate->cache_changed = 1;
 774         return 0;
 775 }
 776
 777 /*
 778  * "refresh" does not calculate a new sha1 file or bring the
 779  * cache up-to-date for mode/content changes. But what it
 780  * _does_ do is to "re-match" the stat information of a file
 781  * with the cache, so that you can refresh the cache for a
 782  * file that hasn't been changed but where the stat entry is
 783  * out of date.
 784  *
 785  * For example, you'd want to do this after doing a "git-read-tree",
 786  * to link up the stat cache details with the proper files.
 787  */
 788 static struct cache_entry *refresh_cache_ent(struct index_state *istate,
 789                                              struct cache_entry *ce,
 790                                              unsigned int options, int *err)
 791 {
 792         struct stat st;
 793         struct cache_entry *updated;
 794         int changed, size;
 795         int ignore_valid = options & CE_MATCH_IGNORE_VALID;
 796
 797         if (lstat(ce->name, &st) < 0) {
 798                 if (err)
 799                         *err = errno;
 800                 return NULL;
 801         }
 802
 803         changed = ie_match_stat(istate, ce, &st, options);
 804         if (!changed) {
 805                 /*
 806                  * The path is unchanged.  If we were told to ignore
 807                  * valid bit, then we did the actual stat check and
 808                  * found that the entry is unmodified.  If the entry
 809                  * is not marked VALID, this is the place to mark it
 810                  * valid again, under "assume unchanged" mode.
 811                  */
 812                 if (ignore_valid && assume_unchanged &&
 813                     !(ce->ce_flags & htons(CE_VALID)))
 814                         ; /* mark this one VALID again */
 815                 else
 816                         return ce;
 817         }
 818
 819         if (ie_modified(istate, ce, &st, options)) {
 820                 if (err)
 821                         *err = EINVAL;
 822                 return NULL;
 823         }
 824
 825         size = ce_size(ce);
 826         updated = xmalloc(size);
 827         memcpy(updated, ce, size);
 828         fill_stat_cache_info(updated, &st);
 829
 830         /*
 831          * If ignore_valid is not set, we should leave CE_VALID bit
 832          * alone.  Otherwise, paths marked with --no-assume-unchanged
 833          * (i.e. things to be edited) will reacquire CE_VALID bit
 834          * automatically, which is not really what we want.
 835          */
 836         if (!ignore_valid && assume_unchanged &&
 837             !(ce->ce_flags & htons(CE_VALID)))
 838                 updated->ce_flags &= ~htons(CE_VALID);
 839
 840         return updated;
 841 }
 842
 843 int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen)
 844 {
 845         int i;
 846         int has_errors = 0;
 847         int really = (flags & REFRESH_REALLY) != 0;
 848         int allow_unmerged = (flags & REFRESH_UNMERGED) != 0;
 849         int quiet = (flags & REFRESH_QUIET) != 0;
 850         int not_new = (flags & REFRESH_IGNORE_MISSING) != 0;
 851         unsigned int options = really ? CE_MATCH_IGNORE_VALID : 0;
 852
 853         for (i = 0; i < istate->cache_nr; i++) {
 854                 struct cache_entry *ce, *new;
 855                 int cache_errno = 0;
 856
 857                 ce = istate->cache[i];
 858                 if (ce_stage(ce)) {
 859                         while ((i < istate->cache_nr) &&
 860                                ! strcmp(istate->cache[i]->name, ce->name))
 861                                 i++;
 862                         i--;
 863                         if (allow_unmerged)
 864                                 continue;
 865                         printf("%s: needs merge\n", ce->name);
 866                         has_errors = 1;
 867                         continue;
 868                 }
 869
 870                 if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
 871                         continue;
 872
 873                 new = refresh_cache_ent(istate, ce, options, &cache_errno);
 874                 if (new == ce)
 875                         continue;
 876                 if (!new) {
 877                         if (not_new && cache_errno == ENOENT)
 878                                 continue;
 879                         if (really && cache_errno == EINVAL) {
 880                                 /* If we are doing --really-refresh that
 881                                  * means the index is not valid anymore.
 882                                  */
 883                                 ce->ce_flags &= ~htons(CE_VALID);
 884                                 istate->cache_changed = 1;
 885                         }
 886                         if (quiet)
 887                                 continue;
 888                         printf("%s: needs update\n", ce->name);
 889                         has_errors = 1;
 890                         continue;
 891                 }
 892                 istate->cache_changed = 1;
 893                 /* You can NOT just free istate->cache[i] here, since it
 894                  * might not be necessarily malloc()ed but can also come
 895                  * from mmap(). */
 896                 istate->cache[i] = new;
 897         }
 898         return has_errors;
 899 }
 900
 901 struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
 902 {
 903         return refresh_cache_ent(&the_index, ce, really, NULL);
 904 }
 905
 906 static int verify_hdr(struct cache_header *hdr, unsigned long size)
 907 {
 908         SHA_CTX c;
 909         unsigned char sha1[20];
 910
 911         if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
 912                 return error("bad signature");
 913         if (hdr->hdr_version != htonl(2))
 914                 return error("bad index version");
 915         SHA1_Init(&c);
 916         SHA1_Update(&c, hdr, size - 20);
 917         SHA1_Final(sha1, &c);
 918         if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
 919                 return error("bad index file sha1 signature");
 920         return 0;
 921 }
 922
 923 static int read_index_extension(struct index_state *istate,
 924                                 const char *ext, void *data, unsigned long sz)
 925 {
 926         switch (CACHE_EXT(ext)) {
 927         case CACHE_EXT_TREE:
 928                 istate->cache_tree = cache_tree_read(data, sz);
 929                 break;
 930         default:
 931                 if (*ext < 'A' || 'Z' < *ext)
 932                         return error("index uses %.4s extension, which we do not understand",
 933                                      ext);
 934                 fprintf(stderr, "ignoring %.4s extension\n", ext);
 935                 break;
 936         }
 937         return 0;
 938 }
 939
 940 int read_index(struct index_state *istate)
 941 {
 942         return read_index_from(istate, get_index_file());
 943 }
 944
 945 /* remember to discard_cache() before reading a different cache! */
 946 int read_index_from(struct index_state *istate, const char *path)
 947 {
 948         int fd, i;
 949         struct stat st;
 950         unsigned long offset;
 951         struct cache_header *hdr;
 952
 953         errno = EBUSY;
 954         if (istate->mmap)
 955                 return istate->cache_nr;
 956
 957         errno = ENOENT;
 958         istate->timestamp = 0;
 959         fd = open(path, O_RDONLY);
 960         if (fd < 0) {
 961                 if (errno == ENOENT)
 962                         return 0;
 963                 die("index file open failed (%s)", strerror(errno));
 964         }
 965
 966         if (fstat(fd, &st))
 967                 die("cannot stat the open index (%s)", strerror(errno));
 968
 969         errno = EINVAL;
 970         istate->mmap_size = xsize_t(st.st_size);
 971         if (istate->mmap_size < sizeof(struct cache_header) + 20)
 972                 die("index file smaller than expected");
 973
 974         istate->mmap = xmmap(NULL, istate->mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
 975         close(fd);
 976
 977         hdr = istate->mmap;
 978         if (verify_hdr(hdr, istate->mmap_size) < 0)
 979                 goto unmap;
 980
 981         istate->cache_nr = ntohl(hdr->hdr_entries);
 982         istate->cache_alloc = alloc_nr(istate->cache_nr);
 983         istate->cache = xcalloc(istate->cache_alloc, sizeof(struct cache_entry *));
 984
 985         offset = sizeof(*hdr);
 986         for (i = 0; i < istate->cache_nr; i++) {
 987                 struct cache_entry *ce;
 988
 989                 ce = (struct cache_entry *)((char *)(istate->mmap) + offset);
 990                 offset = offset + ce_size(ce);
 991                 istate->cache[i] = ce;
 992         }
 993         istate->timestamp = st.st_mtime;
 994         while (offset <= istate->mmap_size - 20 - 8) {
 995                 /* After an array of active_nr index entries,
 996                  * there can be arbitrary number of extended
 997                  * sections, each of which is prefixed with
 998                  * extension name (4-byte) and section length
 999                  * in 4-byte network byte order.
1000                  */
1001                 unsigned long extsize;
1002                 memcpy(&extsize, (char *)(istate->mmap) + offset + 4, 4);
1003                 extsize = ntohl(extsize);
1004                 if (read_index_extension(istate,
1005                                          ((const char *) (istate->mmap)) + offset,
1006                                          (char *) (istate->mmap) + offset + 8,
1007                                          extsize) < 0)
1008                         goto unmap;
1009                 offset += 8;
1010                 offset += extsize;
1011         }
1012         return istate->cache_nr;
1013
1014 unmap:
1015         munmap(istate->mmap, istate->mmap_size);
1016         errno = EINVAL;
1017         die("index file corrupt");
1018 }
1019
1020 int discard_index(struct index_state *istate)
1021 {
1022         int ret;
1023
1024         istate->cache_nr = 0;
1025         istate->cache_changed = 0;
1026         istate->timestamp = 0;
1027         cache_tree_free(&(istate->cache_tree));
1028         if (istate->mmap == NULL)
1029                 return 0;
1030         ret = munmap(istate->mmap, istate->mmap_size);
1031         istate->mmap = NULL;
1032         istate->mmap_size = 0;
1033
1034         /* no need to throw away allocated active_cache */
1035         return ret;
1036 }
1037
1038 #define WRITE_BUFFER_SIZE 8192
1039 static unsigned char write_buffer[WRITE_BUFFER_SIZE];
1040 static unsigned long write_buffer_len;
1041
1042 static int ce_write_flush(SHA_CTX *context, int fd)
1043 {
1044         unsigned int buffered = write_buffer_len;
1045         if (buffered) {
1046                 SHA1_Update(context, write_buffer, buffered);
1047                 if (write_in_full(fd, write_buffer, buffered) != buffered)
1048                         return -1;
1049                 write_buffer_len = 0;
1050         }
1051         return 0;
1052 }
1053
1054 static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
1055 {
1056         while (len) {
1057                 unsigned int buffered = write_buffer_len;
1058                 unsigned int partial = WRITE_BUFFER_SIZE - buffered;
1059                 if (partial > len)
1060                         partial = len;
1061                 memcpy(write_buffer + buffered, data, partial);
1062                 buffered += partial;
1063                 if (buffered == WRITE_BUFFER_SIZE) {
1064                         write_buffer_len = buffered;
1065                         if (ce_write_flush(context, fd))
1066                                 return -1;
1067                         buffered = 0;
1068                 }
1069                 write_buffer_len = buffered;
1070                 len -= partial;
1071                 data = (char *) data + partial;
1072         }
1073         return 0;
1074 }
1075
1076 static int write_index_ext_header(SHA_CTX *context, int fd,
1077                                   unsigned int ext, unsigned int sz)
1078 {
1079         ext = htonl(ext);
1080         sz = htonl(sz);
1081         return ((ce_write(context, fd, &ext, 4) < 0) ||
1082                 (ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1083 }
1084
1085 static int ce_flush(SHA_CTX *context, int fd)
1086 {
1087         unsigned int left = write_buffer_len;
1088
1089         if (left) {
1090                 write_buffer_len = 0;
1091                 SHA1_Update(context, write_buffer, left);
1092         }
1093
1094         /* Flush first if not enough space for SHA1 signature */
1095         if (left + 20 > WRITE_BUFFER_SIZE) {
1096                 if (write_in_full(fd, write_buffer, left) != left)
1097                         return -1;
1098                 left = 0;
1099         }
1100
1101         /* Append the SHA1 signature at the end */
1102         SHA1_Final(write_buffer + left, context);
1103         left += 20;
1104         return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1105 }
1106
1107 static void ce_smudge_racily_clean_entry(struct cache_entry *ce)
1108 {
1109         /*
1110          * The only thing we care about in this function is to smudge the
1111          * falsely clean entry due to touch-update-touch race, so we leave
1112          * everything else as they are.  We are called for entries whose
1113          * ce_mtime match the index file mtime.
1114          */
1115         struct stat st;
1116
1117         if (lstat(ce->name, &st) < 0)
1118                 return;
1119         if (ce_match_stat_basic(ce, &st))
1120                 return;
1121         if (ce_modified_check_fs(ce, &st)) {
1122                 /* This is "racily clean"; smudge it.  Note that this
1123                  * is a tricky code.  At first glance, it may appear
1124                  * that it can break with this sequence:
1125                  *
1126                  * $ echo xyzzy >frotz
1127                  * $ git-update-index --add frotz
1128                  * $ : >frotz
1129                  * $ sleep 3
1130                  * $ echo filfre >nitfol
1131                  * $ git-update-index --add nitfol
1132                  *
1133                  * but it does not.  When the second update-index runs,
1134                  * it notices that the entry "frotz" has the same timestamp
1135                  * as index, and if we were to smudge it by resetting its
1136                  * size to zero here, then the object name recorded
1137                  * in index is the 6-byte file but the cached stat information
1138                  * becomes zero --- which would then match what we would
1139                  * obtain from the filesystem next time we stat("frotz").
1140                  *
1141                  * However, the second update-index, before calling
1142                  * this function, notices that the cached size is 6
1143                  * bytes and what is on the filesystem is an empty
1144                  * file, and never calls us, so the cached size information
1145                  * for "frotz" stays 6 which does not match the filesystem.
1146                  */
1147                 ce->ce_size = htonl(0);
1148         }
1149 }
1150
1151 int write_index(struct index_state *istate, int newfd)
1152 {
1153         SHA_CTX c;
1154         struct cache_header hdr;
1155         int i, err, removed;
1156         struct cache_entry **cache = istate->cache;
1157         int entries = istate->cache_nr;
1158
1159         for (i = removed = 0; i < entries; i++)
1160                 if (!cache[i]->ce_mode)
1161                         removed++;
1162
1163         hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1164         hdr.hdr_version = htonl(2);
1165         hdr.hdr_entries = htonl(entries - removed);
1166
1167         SHA1_Init(&c);
1168         if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1169                 return -1;
1170
1171         for (i = 0; i < entries; i++) {
1172                 struct cache_entry *ce = cache[i];
1173                 if (!ce->ce_mode)
1174                         continue;
1175                 if (istate->timestamp &&
1176                     istate->timestamp <= ntohl(ce->ce_mtime.sec))
1177                         ce_smudge_racily_clean_entry(ce);
1178                 if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1179                         return -1;
1180         }
1181
1182         /* Write extension data here */
1183         if (istate->cache_tree) {
1184                 struct strbuf sb;
1185
1186                 strbuf_init(&sb, 0);
1187                 cache_tree_write(&sb, istate->cache_tree);
1188                 err = write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sb.len) < 0
1189                         || ce_write(&c, newfd, sb.buf, sb.len) < 0;
1190                 strbuf_release(&sb);
1191                 if (err)
1192                         return -1;
1193         }
1194         return ce_flush(&c, newfd);
1195 }