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         default:
 153                 die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
 154         }
 155         if (ce->ce_mtime.sec != htonl(st->st_mtime))
 156                 changed |= MTIME_CHANGED;
 157         if (ce->ce_ctime.sec != htonl(st->st_ctime))
 158                 changed |= CTIME_CHANGED;
 159
 160 #ifdef USE_NSEC
 161         /*
 162          * nsec seems unreliable - not all filesystems support it, so
 163          * as long as it is in the inode cache you get right nsec
 164          * but after it gets flushed, you get zero nsec.
 165          */
 166         if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
 167                 changed |= MTIME_CHANGED;
 168         if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
 169                 changed |= CTIME_CHANGED;
 170 #endif
 171
 172         if (ce->ce_uid != htonl(st->st_uid) ||
 173             ce->ce_gid != htonl(st->st_gid))
 174                 changed |= OWNER_CHANGED;
 175         if (ce->ce_ino != htonl(st->st_ino))
 176                 changed |= INODE_CHANGED;
 177
 178 #ifdef USE_STDEV
 179         /*
 180          * st_dev breaks on network filesystems where different
 181          * clients will have different views of what "device"
 182          * the filesystem is on
 183          */
 184         if (ce->ce_dev != htonl(st->st_dev))
 185                 changed |= INODE_CHANGED;
 186 #endif
 187
 188         if (ce->ce_size != htonl(st->st_size))
 189                 changed |= DATA_CHANGED;
 190
 191         return changed;
 192 }
 193
 194 int ie_match_stat(struct index_state *istate,
 195                   struct cache_entry *ce, struct stat *st, int options)
 196 {
 197         unsigned int changed;
 198         int ignore_valid = options & 01;
 199         int assume_racy_is_modified = options & 02;
 200
 201         /*
 202          * If it's marked as always valid in the index, it's
 203          * valid whatever the checked-out copy says.
 204          */
 205         if (!ignore_valid && (ce->ce_flags & htons(CE_VALID)))
 206                 return 0;
 207
 208         changed = ce_match_stat_basic(ce, st);
 209
 210         /*
 211          * Within 1 second of this sequence:
 212          *      echo xyzzy >file && git-update-index --add file
 213          * running this command:
 214          *      echo frotz >file
 215          * would give a falsely clean cache entry.  The mtime and
 216          * length match the cache, and other stat fields do not change.
 217          *
 218          * We could detect this at update-index time (the cache entry
 219          * being registered/updated records the same time as "now")
 220          * and delay the return from git-update-index, but that would
 221          * effectively mean we can make at most one commit per second,
 222          * which is not acceptable.  Instead, we check cache entries
 223          * whose mtime are the same as the index file timestamp more
 224          * carefully than others.
 225          */
 226         if (!changed &&
 227             istate->timestamp &&
 228             istate->timestamp <= ntohl(ce->ce_mtime.sec)) {
 229                 if (assume_racy_is_modified)
 230                         changed |= DATA_CHANGED;
 231                 else
 232                         changed |= ce_modified_check_fs(ce, st);
 233         }
 234
 235         return changed;
 236 }
 237
 238 int ie_modified(struct index_state *istate,
 239                 struct cache_entry *ce, struct stat *st, int really)
 240 {
 241         int changed, changed_fs;
 242         changed = ie_match_stat(istate, ce, st, really);
 243         if (!changed)
 244                 return 0;
 245         /*
 246          * If the mode or type has changed, there's no point in trying
 247          * to refresh the entry - it's not going to match
 248          */
 249         if (changed & (MODE_CHANGED | TYPE_CHANGED))
 250                 return changed;
 251
 252         /* Immediately after read-tree or update-index --cacheinfo,
 253          * the length field is zero.  For other cases the ce_size
 254          * should match the SHA1 recorded in the index entry.
 255          */
 256         if ((changed & DATA_CHANGED) && ce->ce_size != htonl(0))
 257                 return changed;
 258
 259         changed_fs = ce_modified_check_fs(ce, st);
 260         if (changed_fs)
 261                 return changed | changed_fs;
 262         return 0;
 263 }
 264
 265 int base_name_compare(const char *name1, int len1, int mode1,
 266                       const char *name2, int len2, int mode2)
 267 {
 268         unsigned char c1, c2;
 269         int len = len1 < len2 ? len1 : len2;
 270         int cmp;
 271
 272         cmp = memcmp(name1, name2, len);
 273         if (cmp)
 274                 return cmp;
 275         c1 = name1[len];
 276         c2 = name2[len];
 277         if (!c1 && S_ISDIR(mode1))
 278                 c1 = '/';
 279         if (!c2 && S_ISDIR(mode2))
 280                 c2 = '/';
 281         return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
 282 }
 283
 284 int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
 285 {
 286         int len1 = flags1 & CE_NAMEMASK;
 287         int len2 = flags2 & CE_NAMEMASK;
 288         int len = len1 < len2 ? len1 : len2;
 289         int cmp;
 290
 291         cmp = memcmp(name1, name2, len);
 292         if (cmp)
 293                 return cmp;
 294         if (len1 < len2)
 295                 return -1;
 296         if (len1 > len2)
 297                 return 1;
 298
 299         /* Compare stages  */
 300         flags1 &= CE_STAGEMASK;
 301         flags2 &= CE_STAGEMASK;
 302
 303         if (flags1 < flags2)
 304                 return -1;
 305         if (flags1 > flags2)
 306                 return 1;
 307         return 0;
 308 }
 309
 310 int index_name_pos(struct index_state *istate, const char *name, int namelen)
 311 {
 312         int first, last;
 313
 314         first = 0;
 315         last = istate->cache_nr;
 316         while (last > first) {
 317                 int next = (last + first) >> 1;
 318                 struct cache_entry *ce = istate->cache[next];
 319                 int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
 320                 if (!cmp)
 321                         return next;
 322                 if (cmp < 0) {
 323                         last = next;
 324                         continue;
 325                 }
 326                 first = next+1;
 327         }
 328         return -first-1;
 329 }
 330
 331 /* Remove entry, return true if there are more entries to go.. */
 332 int remove_index_entry_at(struct index_state *istate, int pos)
 333 {
 334         istate->cache_changed = 1;
 335         istate->cache_nr--;
 336         if (pos >= istate->cache_nr)
 337                 return 0;
 338         memmove(istate->cache + pos,
 339                 istate->cache + pos + 1,
 340                 (istate->cache_nr - pos) * sizeof(struct cache_entry *));
 341         return 1;
 342 }
 343
 344 int remove_file_from_index(struct index_state *istate, const char *path)
 345 {
 346         int pos = index_name_pos(istate, path, strlen(path));
 347         if (pos < 0)
 348                 pos = -pos-1;
 349         while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
 350                 remove_index_entry_at(istate, pos);
 351         return 0;
 352 }
 353
 354 static int compare_name(struct cache_entry *ce, const char *path, int namelen)
 355 {
 356         return namelen != ce_namelen(ce) || memcmp(path, ce->name, namelen);
 357 }
 358
 359 static int index_name_pos_also_unmerged(struct index_state *istate,
 360         const char *path, int namelen)
 361 {
 362         int pos = index_name_pos(istate, path, namelen);
 363         struct cache_entry *ce;
 364
 365         if (pos >= 0)
 366                 return pos;
 367
 368         /* maybe unmerged? */
 369         pos = -1 - pos;
 370         if (pos >= istate->cache_nr ||
 371                         compare_name((ce = istate->cache[pos]), path, namelen))
 372                 return -1;
 373
 374         /* order of preference: stage 2, 1, 3 */
 375         if (ce_stage(ce) == 1 && pos + 1 < istate->cache_nr &&
 376                         ce_stage((ce = istate->cache[pos + 1])) == 2 &&
 377                         !compare_name(ce, path, namelen))
 378                 pos++;
 379         return pos;
 380 }
 381
 382 int add_file_to_index(struct index_state *istate, const char *path, int verbose)
 383 {
 384         int size, namelen, pos;
 385         struct stat st;
 386         struct cache_entry *ce;
 387
 388         if (lstat(path, &st))
 389                 die("%s: unable to stat (%s)", path, strerror(errno));
 390
 391         if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode) && !S_ISDIR(st.st_mode))
 392                 die("%s: can only add regular files, symbolic links or git-directories", path);
 393
 394         namelen = strlen(path);
 395         if (S_ISDIR(st.st_mode)) {
 396                 while (namelen && path[namelen-1] == '/')
 397                         namelen--;
 398         }
 399         size = cache_entry_size(namelen);
 400         ce = xcalloc(1, size);
 401         memcpy(ce->name, path, namelen);
 402         ce->ce_flags = htons(namelen);
 403         fill_stat_cache_info(ce, &st);
 404
 405         if (trust_executable_bit && has_symlinks)
 406                 ce->ce_mode = create_ce_mode(st.st_mode);
 407         else {
 408                 /* If there is an existing entry, pick the mode bits and type
 409                  * from it, otherwise assume unexecutable regular file.
 410                  */
 411                 struct cache_entry *ent;
 412                 int pos = index_name_pos_also_unmerged(istate, path, namelen);
 413
 414                 ent = (0 <= pos) ? istate->cache[pos] : NULL;
 415                 ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
 416         }
 417
 418         pos = index_name_pos(istate, ce->name, namelen);
 419         if (0 <= pos &&
 420             !ce_stage(istate->cache[pos]) &&
 421             !ie_modified(istate, istate->cache[pos], &st, 1)) {
 422                 /* Nothing changed, really */
 423                 free(ce);
 424                 return 0;
 425         }
 426
 427         if (index_path(ce->sha1, path, &st, 1))
 428                 die("unable to index file %s", path);
 429         if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
 430                 die("unable to add %s to index",path);
 431         if (verbose)
 432                 printf("add '%s'\n", path);
 433         cache_tree_invalidate_path(istate->cache_tree, path);
 434         return 0;
 435 }
 436
 437 int ce_same_name(struct cache_entry *a, struct cache_entry *b)
 438 {
 439         int len = ce_namelen(a);
 440         return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
 441 }
 442
 443 int ce_path_match(const struct cache_entry *ce, const char **pathspec)
 444 {
 445         const char *match, *name;
 446         int len;
 447
 448         if (!pathspec)
 449                 return 1;
 450
 451         len = ce_namelen(ce);
 452         name = ce->name;
 453         while ((match = *pathspec++) != NULL) {
 454                 int matchlen = strlen(match);
 455                 if (matchlen > len)
 456                         continue;
 457                 if (memcmp(name, match, matchlen))
 458                         continue;
 459                 if (matchlen && name[matchlen-1] == '/')
 460                         return 1;
 461                 if (name[matchlen] == '/' || !name[matchlen])
 462                         return 1;
 463                 if (!matchlen)
 464                         return 1;
 465         }
 466         return 0;
 467 }
 468
 469 /*
 470  * We fundamentally don't like some paths: we don't want
 471  * dot or dot-dot anywhere, and for obvious reasons don't
 472  * want to recurse into ".git" either.
 473  *
 474  * Also, we don't want double slashes or slashes at the
 475  * end that can make pathnames ambiguous.
 476  */
 477 static int verify_dotfile(const char *rest)
 478 {
 479         /*
 480          * The first character was '.', but that
 481          * has already been discarded, we now test
 482          * the rest.
 483          */
 484         switch (*rest) {
 485         /* "." is not allowed */
 486         case '\0': case '/':
 487                 return 0;
 488
 489         /*
 490          * ".git" followed by  NUL or slash is bad. This
 491          * shares the path end test with the ".." case.
 492          */
 493         case 'g':
 494                 if (rest[1] != 'i')
 495                         break;
 496                 if (rest[2] != 't')
 497                         break;
 498                 rest += 2;
 499         /* fallthrough */
 500         case '.':
 501                 if (rest[1] == '\0' || rest[1] == '/')
 502                         return 0;
 503         }
 504         return 1;
 505 }
 506
 507 int verify_path(const char *path)
 508 {
 509         char c;
 510
 511         goto inside;
 512         for (;;) {
 513                 if (!c)
 514                         return 1;
 515                 if (c == '/') {
 516 inside:
 517                         c = *path++;
 518                         switch (c) {
 519                         default:
 520                                 continue;
 521                         case '/': case '\0':
 522                                 break;
 523                         case '.':
 524                                 if (verify_dotfile(path))
 525                                         continue;
 526                         }
 527                         return 0;
 528                 }
 529                 c = *path++;
 530         }
 531 }
 532
 533 /*
 534  * Do we have another file that has the beginning components being a
 535  * proper superset of the name we're trying to add?
 536  */
 537 static int has_file_name(struct index_state *istate,
 538                          const struct cache_entry *ce, int pos, int ok_to_replace)
 539 {
 540         int retval = 0;
 541         int len = ce_namelen(ce);
 542         int stage = ce_stage(ce);
 543         const char *name = ce->name;
 544
 545         while (pos < istate->cache_nr) {
 546                 struct cache_entry *p = istate->cache[pos++];
 547
 548                 if (len >= ce_namelen(p))
 549                         break;
 550                 if (memcmp(name, p->name, len))
 551                         break;
 552                 if (ce_stage(p) != stage)
 553                         continue;
 554                 if (p->name[len] != '/')
 555                         continue;
 556                 if (!ce_stage(p) && !p->ce_mode)
 557                         continue;
 558                 retval = -1;
 559                 if (!ok_to_replace)
 560                         break;
 561                 remove_index_entry_at(istate, --pos);
 562         }
 563         return retval;
 564 }
 565
 566 /*
 567  * Do we have another file with a pathname that is a proper
 568  * subset of the name we're trying to add?
 569  */
 570 static int has_dir_name(struct index_state *istate,
 571                         const struct cache_entry *ce, int pos, int ok_to_replace)
 572 {
 573         int retval = 0;
 574         int stage = ce_stage(ce);
 575         const char *name = ce->name;
 576         const char *slash = name + ce_namelen(ce);
 577
 578         for (;;) {
 579                 int len;
 580
 581                 for (;;) {
 582                         if (*--slash == '/')
 583                                 break;
 584                         if (slash <= ce->name)
 585                                 return retval;
 586                 }
 587                 len = slash - name;
 588
 589                 pos = index_name_pos(istate, name, ntohs(create_ce_flags(len, stage)));
 590                 if (pos >= 0) {
 591                         /*
 592                          * Found one, but not so fast.  This could
 593                          * be a marker that says "I was here, but
 594                          * I am being removed".  Such an entry is
 595                          * not a part of the resulting tree, and
 596                          * it is Ok to have a directory at the same
 597                          * path.
 598                          */
 599                         if (stage || istate->cache[pos]->ce_mode) {
 600                                 retval = -1;
 601                                 if (!ok_to_replace)
 602                                         break;
 603                                 remove_index_entry_at(istate, pos);
 604                                 continue;
 605                         }
 606                 }
 607                 else
 608                         pos = -pos-1;
 609
 610                 /*
 611                  * Trivial optimization: if we find an entry that
 612                  * already matches the sub-directory, then we know
 613                  * we're ok, and we can exit.
 614                  */
 615                 while (pos < istate->cache_nr) {
 616                         struct cache_entry *p = istate->cache[pos];
 617                         if ((ce_namelen(p) <= len) ||
 618                             (p->name[len] != '/') ||
 619                             memcmp(p->name, name, len))
 620                                 break; /* not our subdirectory */
 621                         if (ce_stage(p) == stage && (stage || p->ce_mode))
 622                                 /* p is at the same stage as our entry, and
 623                                  * is a subdirectory of what we are looking
 624                                  * at, so we cannot have conflicts at our
 625                                  * level or anything shorter.
 626                                  */
 627                                 return retval;
 628                         pos++;
 629                 }
 630         }
 631         return retval;
 632 }
 633
 634 /* We may be in a situation where we already have path/file and path
 635  * is being added, or we already have path and path/file is being
 636  * added.  Either one would result in a nonsense tree that has path
 637  * twice when git-write-tree tries to write it out.  Prevent it.
 638  *
 639  * If ok-to-replace is specified, we remove the conflicting entries
 640  * from the cache so the caller should recompute the insert position.
 641  * When this happens, we return non-zero.
 642  */
 643 static int check_file_directory_conflict(struct index_state *istate,
 644                                          const struct cache_entry *ce,
 645                                          int pos, int ok_to_replace)
 646 {
 647         int retval;
 648
 649         /*
 650          * When ce is an "I am going away" entry, we allow it to be added
 651          */
 652         if (!ce_stage(ce) && !ce->ce_mode)
 653                 return 0;
 654
 655         /*
 656          * We check if the path is a sub-path of a subsequent pathname
 657          * first, since removing those will not change the position
 658          * in the array.
 659          */
 660         retval = has_file_name(istate, ce, pos, ok_to_replace);
 661
 662         /*
 663          * Then check if the path might have a clashing sub-directory
 664          * before it.
 665          */
 666         return retval + has_dir_name(istate, ce, pos, ok_to_replace);
 667 }
 668
 669 static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
 670 {
 671         int pos;
 672         int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
 673         int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
 674         int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
 675
 676         pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 677
 678         /* existing match? Just replace it. */
 679         if (pos >= 0) {
 680                 istate->cache_changed = 1;
 681                 istate->cache[pos] = ce;
 682                 return 0;
 683         }
 684         pos = -pos-1;
 685
 686         /*
 687          * Inserting a merged entry ("stage 0") into the index
 688          * will always replace all non-merged entries..
 689          */
 690         if (pos < istate->cache_nr && ce_stage(ce) == 0) {
 691                 while (ce_same_name(istate->cache[pos], ce)) {
 692                         ok_to_add = 1;
 693                         if (!remove_index_entry_at(istate, pos))
 694                                 break;
 695                 }
 696         }
 697
 698         if (!ok_to_add)
 699                 return -1;
 700         if (!verify_path(ce->name))
 701                 return -1;
 702
 703         if (!skip_df_check &&
 704             check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
 705                 if (!ok_to_replace)
 706                         return error("'%s' appears as both a file and as a directory",
 707                                      ce->name);
 708                 pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 709                 pos = -pos-1;
 710         }
 711         return pos + 1;
 712 }
 713
 714 int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
 715 {
 716         int pos;
 717
 718         if (option & ADD_CACHE_JUST_APPEND)
 719                 pos = istate->cache_nr;
 720         else {
 721                 int ret;
 722                 ret = add_index_entry_with_check(istate, ce, option);
 723                 if (ret <= 0)
 724                         return ret;
 725                 pos = ret - 1;
 726         }
 727
 728         /* Make sure the array is big enough .. */
 729         if (istate->cache_nr == istate->cache_alloc) {
 730                 istate->cache_alloc = alloc_nr(istate->cache_alloc);
 731                 istate->cache = xrealloc(istate->cache,
 732                                         istate->cache_alloc * sizeof(struct cache_entry *));
 733         }
 734
 735         /* Add it in.. */
 736         istate->cache_nr++;
 737         if (istate->cache_nr > pos + 1)
 738                 memmove(istate->cache + pos + 1,
 739                         istate->cache + pos,
 740                         (istate->cache_nr - pos - 1) * sizeof(ce));
 741         istate->cache[pos] = ce;
 742         istate->cache_changed = 1;
 743         return 0;
 744 }
 745
 746 /*
 747  * "refresh" does not calculate a new sha1 file or bring the
 748  * cache up-to-date for mode/content changes. But what it
 749  * _does_ do is to "re-match" the stat information of a file
 750  * with the cache, so that you can refresh the cache for a
 751  * file that hasn't been changed but where the stat entry is
 752  * out of date.
 753  *
 754  * For example, you'd want to do this after doing a "git-read-tree",
 755  * to link up the stat cache details with the proper files.
 756  */
 757 static struct cache_entry *refresh_cache_ent(struct index_state *istate,
 758                                              struct cache_entry *ce, int really, int *err)
 759 {
 760         struct stat st;
 761         struct cache_entry *updated;
 762         int changed, size;
 763
 764         if (lstat(ce->name, &st) < 0) {
 765                 if (err)
 766                         *err = errno;
 767                 return NULL;
 768         }
 769
 770         changed = ie_match_stat(istate, ce, &st, really);
 771         if (!changed) {
 772                 if (really && assume_unchanged &&
 773                     !(ce->ce_flags & htons(CE_VALID)))
 774                         ; /* mark this one VALID again */
 775                 else
 776                         return ce;
 777         }
 778
 779         if (ie_modified(istate, ce, &st, really)) {
 780                 if (err)
 781                         *err = EINVAL;
 782                 return NULL;
 783         }
 784
 785         size = ce_size(ce);
 786         updated = xmalloc(size);
 787         memcpy(updated, ce, size);
 788         fill_stat_cache_info(updated, &st);
 789
 790         /* In this case, if really is not set, we should leave
 791          * CE_VALID bit alone.  Otherwise, paths marked with
 792          * --no-assume-unchanged (i.e. things to be edited) will
 793          * reacquire CE_VALID bit automatically, which is not
 794          * really what we want.
 795          */
 796         if (!really && assume_unchanged && !(ce->ce_flags & htons(CE_VALID)))
 797                 updated->ce_flags &= ~htons(CE_VALID);
 798
 799         return updated;
 800 }
 801
 802 int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen)
 803 {
 804         int i;
 805         int has_errors = 0;
 806         int really = (flags & REFRESH_REALLY) != 0;
 807         int allow_unmerged = (flags & REFRESH_UNMERGED) != 0;
 808         int quiet = (flags & REFRESH_QUIET) != 0;
 809         int not_new = (flags & REFRESH_IGNORE_MISSING) != 0;
 810
 811         for (i = 0; i < istate->cache_nr; i++) {
 812                 struct cache_entry *ce, *new;
 813                 int cache_errno = 0;
 814
 815                 ce = istate->cache[i];
 816                 if (ce_stage(ce)) {
 817                         while ((i < istate->cache_nr) &&
 818                                ! strcmp(istate->cache[i]->name, ce->name))
 819                                 i++;
 820                         i--;
 821                         if (allow_unmerged)
 822                                 continue;
 823                         printf("%s: needs merge\n", ce->name);
 824                         has_errors = 1;
 825                         continue;
 826                 }
 827
 828                 if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
 829                         continue;
 830
 831                 new = refresh_cache_ent(istate, ce, really, &cache_errno);
 832                 if (new == ce)
 833                         continue;
 834                 if (!new) {
 835                         if (not_new && cache_errno == ENOENT)
 836                                 continue;
 837                         if (really && cache_errno == EINVAL) {
 838                                 /* If we are doing --really-refresh that
 839                                  * means the index is not valid anymore.
 840                                  */
 841                                 ce->ce_flags &= ~htons(CE_VALID);
 842                                 istate->cache_changed = 1;
 843                         }
 844                         if (quiet)
 845                                 continue;
 846                         printf("%s: needs update\n", ce->name);
 847                         has_errors = 1;
 848                         continue;
 849                 }
 850                 istate->cache_changed = 1;
 851                 /* You can NOT just free istate->cache[i] here, since it
 852                  * might not be necessarily malloc()ed but can also come
 853                  * from mmap(). */
 854                 istate->cache[i] = new;
 855         }
 856         return has_errors;
 857 }
 858
 859 struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
 860 {
 861         return refresh_cache_ent(&the_index, ce, really, NULL);
 862 }
 863
 864 static int verify_hdr(struct cache_header *hdr, unsigned long size)
 865 {
 866         SHA_CTX c;
 867         unsigned char sha1[20];
 868
 869         if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
 870                 return error("bad signature");
 871         if (hdr->hdr_version != htonl(2))
 872                 return error("bad index version");
 873         SHA1_Init(&c);
 874         SHA1_Update(&c, hdr, size - 20);
 875         SHA1_Final(sha1, &c);
 876         if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
 877                 return error("bad index file sha1 signature");
 878         return 0;
 879 }
 880
 881 static int read_index_extension(struct index_state *istate,
 882                                 const char *ext, void *data, unsigned long sz)
 883 {
 884         switch (CACHE_EXT(ext)) {
 885         case CACHE_EXT_TREE:
 886                 istate->cache_tree = cache_tree_read(data, sz);
 887                 break;
 888         default:
 889                 if (*ext < 'A' || 'Z' < *ext)
 890                         return error("index uses %.4s extension, which we do not understand",
 891                                      ext);
 892                 fprintf(stderr, "ignoring %.4s extension\n", ext);
 893                 break;
 894         }
 895         return 0;
 896 }
 897
 898 int read_index(struct index_state *istate)
 899 {
 900         return read_index_from(istate, get_index_file());
 901 }
 902
 903 /* remember to discard_cache() before reading a different cache! */
 904 int read_index_from(struct index_state *istate, const char *path)
 905 {
 906         int fd, i;
 907         struct stat st;
 908         unsigned long offset;
 909         struct cache_header *hdr;
 910
 911         errno = EBUSY;
 912         if (istate->mmap)
 913                 return istate->cache_nr;
 914
 915         errno = ENOENT;
 916         istate->timestamp = 0;
 917         fd = open(path, O_RDONLY);
 918         if (fd < 0) {
 919                 if (errno == ENOENT)
 920                         return 0;
 921                 die("index file open failed (%s)", strerror(errno));
 922         }
 923
 924         if (fstat(fd, &st))
 925                 die("cannot stat the open index (%s)", strerror(errno));
 926
 927         errno = EINVAL;
 928         istate->mmap_size = xsize_t(st.st_size);
 929         if (istate->mmap_size < sizeof(struct cache_header) + 20)
 930                 die("index file smaller than expected");
 931
 932         istate->mmap = xmmap(NULL, istate->mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
 933         close(fd);
 934
 935         hdr = istate->mmap;
 936         if (verify_hdr(hdr, istate->mmap_size) < 0)
 937                 goto unmap;
 938
 939         istate->cache_nr = ntohl(hdr->hdr_entries);
 940         istate->cache_alloc = alloc_nr(istate->cache_nr);
 941         istate->cache = xcalloc(istate->cache_alloc, sizeof(struct cache_entry *));
 942
 943         offset = sizeof(*hdr);
 944         for (i = 0; i < istate->cache_nr; i++) {
 945                 struct cache_entry *ce;
 946
 947                 ce = (struct cache_entry *)((char *)(istate->mmap) + offset);
 948                 offset = offset + ce_size(ce);
 949                 istate->cache[i] = ce;
 950         }
 951         istate->timestamp = st.st_mtime;
 952         while (offset <= istate->mmap_size - 20 - 8) {
 953                 /* After an array of active_nr index entries,
 954                  * there can be arbitrary number of extended
 955                  * sections, each of which is prefixed with
 956                  * extension name (4-byte) and section length
 957                  * in 4-byte network byte order.
 958                  */
 959                 unsigned long extsize;
 960                 memcpy(&extsize, (char *)(istate->mmap) + offset + 4, 4);
 961                 extsize = ntohl(extsize);
 962                 if (read_index_extension(istate,
 963                                          ((const char *) (istate->mmap)) + offset,
 964                                          (char *) (istate->mmap) + offset + 8,
 965                                          extsize) < 0)
 966                         goto unmap;
 967                 offset += 8;
 968                 offset += extsize;
 969         }
 970         return istate->cache_nr;
 971
 972 unmap:
 973         munmap(istate->mmap, istate->mmap_size);
 974         errno = EINVAL;
 975         die("index file corrupt");
 976 }
 977
 978 int discard_index(struct index_state *istate)
 979 {
 980         int ret;
 981
 982         istate->cache_nr = 0;
 983         istate->cache_changed = 0;
 984         istate->timestamp = 0;
 985         cache_tree_free(&(istate->cache_tree));
 986         if (istate->mmap == NULL)
 987                 return 0;
 988         ret = munmap(istate->mmap, istate->mmap_size);
 989         istate->mmap = NULL;
 990         istate->mmap_size = 0;
 991
 992         /* no need to throw away allocated active_cache */
 993         return ret;
 994 }
 995
 996 #define WRITE_BUFFER_SIZE 8192
 997 static unsigned char write_buffer[WRITE_BUFFER_SIZE];
 998 static unsigned long write_buffer_len;
 999
1000 static int ce_write_flush(SHA_CTX *context, int fd)
1001 {
1002         unsigned int buffered = write_buffer_len;
1003         if (buffered) {
1004                 SHA1_Update(context, write_buffer, buffered);
1005                 if (write_in_full(fd, write_buffer, buffered) != buffered)
1006                         return -1;
1007                 write_buffer_len = 0;
1008         }
1009         return 0;
1010 }
1011
1012 static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
1013 {
1014         while (len) {
1015                 unsigned int buffered = write_buffer_len;
1016                 unsigned int partial = WRITE_BUFFER_SIZE - buffered;
1017                 if (partial > len)
1018                         partial = len;
1019                 memcpy(write_buffer + buffered, data, partial);
1020                 buffered += partial;
1021                 if (buffered == WRITE_BUFFER_SIZE) {
1022                         write_buffer_len = buffered;
1023                         if (ce_write_flush(context, fd))
1024                                 return -1;
1025                         buffered = 0;
1026                 }
1027                 write_buffer_len = buffered;
1028                 len -= partial;
1029                 data = (char *) data + partial;
1030         }
1031         return 0;
1032 }
1033
1034 static int write_index_ext_header(SHA_CTX *context, int fd,
1035                                   unsigned int ext, unsigned int sz)
1036 {
1037         ext = htonl(ext);
1038         sz = htonl(sz);
1039         return ((ce_write(context, fd, &ext, 4) < 0) ||
1040                 (ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1041 }
1042
1043 static int ce_flush(SHA_CTX *context, int fd)
1044 {
1045         unsigned int left = write_buffer_len;
1046
1047         if (left) {
1048                 write_buffer_len = 0;
1049                 SHA1_Update(context, write_buffer, left);
1050         }
1051
1052         /* Flush first if not enough space for SHA1 signature */
1053         if (left + 20 > WRITE_BUFFER_SIZE) {
1054                 if (write_in_full(fd, write_buffer, left) != left)
1055                         return -1;
1056                 left = 0;
1057         }
1058
1059         /* Append the SHA1 signature at the end */
1060         SHA1_Final(write_buffer + left, context);
1061         left += 20;
1062         return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1063 }
1064
1065 static void ce_smudge_racily_clean_entry(struct cache_entry *ce)
1066 {
1067         /*
1068          * The only thing we care about in this function is to smudge the
1069          * falsely clean entry due to touch-update-touch race, so we leave
1070          * everything else as they are.  We are called for entries whose
1071          * ce_mtime match the index file mtime.
1072          */
1073         struct stat st;
1074
1075         if (lstat(ce->name, &st) < 0)
1076                 return;
1077         if (ce_match_stat_basic(ce, &st))
1078                 return;
1079         if (ce_modified_check_fs(ce, &st)) {
1080                 /* This is "racily clean"; smudge it.  Note that this
1081                  * is a tricky code.  At first glance, it may appear
1082                  * that it can break with this sequence:
1083                  *
1084                  * $ echo xyzzy >frotz
1085                  * $ git-update-index --add frotz
1086                  * $ : >frotz
1087                  * $ sleep 3
1088                  * $ echo filfre >nitfol
1089                  * $ git-update-index --add nitfol
1090                  *
1091                  * but it does not.  When the second update-index runs,
1092                  * it notices that the entry "frotz" has the same timestamp
1093                  * as index, and if we were to smudge it by resetting its
1094                  * size to zero here, then the object name recorded
1095                  * in index is the 6-byte file but the cached stat information
1096                  * becomes zero --- which would then match what we would
1097                  * obtain from the filesystem next time we stat("frotz").
1098                  *
1099                  * However, the second update-index, before calling
1100                  * this function, notices that the cached size is 6
1101                  * bytes and what is on the filesystem is an empty
1102                  * file, and never calls us, so the cached size information
1103                  * for "frotz" stays 6 which does not match the filesystem.
1104                  */
1105                 ce->ce_size = htonl(0);
1106         }
1107 }
1108
1109 int write_index(struct index_state *istate, int newfd)
1110 {
1111         SHA_CTX c;
1112         struct cache_header hdr;
1113         int i, removed;
1114         struct cache_entry **cache = istate->cache;
1115         int entries = istate->cache_nr;
1116
1117         for (i = removed = 0; i < entries; i++)
1118                 if (!cache[i]->ce_mode)
1119                         removed++;
1120
1121         hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1122         hdr.hdr_version = htonl(2);
1123         hdr.hdr_entries = htonl(entries - removed);
1124
1125         SHA1_Init(&c);
1126         if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1127                 return -1;
1128
1129         for (i = 0; i < entries; i++) {
1130                 struct cache_entry *ce = cache[i];
1131                 if (!ce->ce_mode)
1132                         continue;
1133                 if (istate->timestamp &&
1134                     istate->timestamp <= ntohl(ce->ce_mtime.sec))
1135                         ce_smudge_racily_clean_entry(ce);
1136                 if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1137                         return -1;
1138         }
1139
1140         /* Write extension data here */
1141         if (istate->cache_tree) {
1142                 unsigned long sz;
1143                 void *data = cache_tree_write(istate->cache_tree, &sz);
1144                 if (data &&
1145                     !write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sz) &&
1146                     !ce_write(&c, newfd, data, sz))
1147                         free(data);
1148                 else {
1149                         free(data);
1150                         return -1;
1151                 }
1152         }
1153         return ce_flush(&c, newfd);
1154 }