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         cache_tree_invalidate_path(istate->cache_tree, path);
 350         while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
 351                 remove_index_entry_at(istate, pos);
 352         return 0;
 353 }
 354
 355 static int compare_name(struct cache_entry *ce, const char *path, int namelen)
 356 {
 357         return namelen != ce_namelen(ce) || memcmp(path, ce->name, namelen);
 358 }
 359
 360 static int index_name_pos_also_unmerged(struct index_state *istate,
 361         const char *path, int namelen)
 362 {
 363         int pos = index_name_pos(istate, path, namelen);
 364         struct cache_entry *ce;
 365
 366         if (pos >= 0)
 367                 return pos;
 368
 369         /* maybe unmerged? */
 370         pos = -1 - pos;
 371         if (pos >= istate->cache_nr ||
 372                         compare_name((ce = istate->cache[pos]), path, namelen))
 373                 return -1;
 374
 375         /* order of preference: stage 2, 1, 3 */
 376         if (ce_stage(ce) == 1 && pos + 1 < istate->cache_nr &&
 377                         ce_stage((ce = istate->cache[pos + 1])) == 2 &&
 378                         !compare_name(ce, path, namelen))
 379                 pos++;
 380         return pos;
 381 }
 382
 383 int add_file_to_index(struct index_state *istate, const char *path, int verbose)
 384 {
 385         int size, namelen, pos;
 386         struct stat st;
 387         struct cache_entry *ce;
 388
 389         if (lstat(path, &st))
 390                 die("%s: unable to stat (%s)", path, strerror(errno));
 391
 392         if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode) && !S_ISDIR(st.st_mode))
 393                 die("%s: can only add regular files, symbolic links or git-directories", path);
 394
 395         namelen = strlen(path);
 396         if (S_ISDIR(st.st_mode)) {
 397                 while (namelen && path[namelen-1] == '/')
 398                         namelen--;
 399         }
 400         size = cache_entry_size(namelen);
 401         ce = xcalloc(1, size);
 402         memcpy(ce->name, path, namelen);
 403         ce->ce_flags = htons(namelen);
 404         fill_stat_cache_info(ce, &st);
 405
 406         if (trust_executable_bit && has_symlinks)
 407                 ce->ce_mode = create_ce_mode(st.st_mode);
 408         else {
 409                 /* If there is an existing entry, pick the mode bits and type
 410                  * from it, otherwise assume unexecutable regular file.
 411                  */
 412                 struct cache_entry *ent;
 413                 int pos = index_name_pos_also_unmerged(istate, path, namelen);
 414
 415                 ent = (0 <= pos) ? istate->cache[pos] : NULL;
 416                 ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
 417         }
 418
 419         pos = index_name_pos(istate, ce->name, namelen);
 420         if (0 <= pos &&
 421             !ce_stage(istate->cache[pos]) &&
 422             !ie_modified(istate, istate->cache[pos], &st, 1)) {
 423                 /* Nothing changed, really */
 424                 free(ce);
 425                 return 0;
 426         }
 427
 428         if (index_path(ce->sha1, path, &st, 1))
 429                 die("unable to index file %s", path);
 430         if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
 431                 die("unable to add %s to index",path);
 432         if (verbose)
 433                 printf("add '%s'\n", path);
 434         return 0;
 435 }
 436
 437 struct cache_entry *make_cache_entry(unsigned int mode,
 438                 const unsigned char *sha1, const char *path, int stage,
 439                 int refresh)
 440 {
 441         int size, len;
 442         struct cache_entry *ce;
 443
 444         if (!verify_path(path))
 445                 return NULL;
 446
 447         len = strlen(path);
 448         size = cache_entry_size(len);
 449         ce = xcalloc(1, size);
 450
 451         hashcpy(ce->sha1, sha1);
 452         memcpy(ce->name, path, len);
 453         ce->ce_flags = create_ce_flags(len, stage);
 454         ce->ce_mode = create_ce_mode(mode);
 455
 456         if (refresh)
 457                 return refresh_cache_entry(ce, 0);
 458
 459         return ce;
 460 }
 461
 462 int ce_same_name(struct cache_entry *a, struct cache_entry *b)
 463 {
 464         int len = ce_namelen(a);
 465         return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
 466 }
 467
 468 int ce_path_match(const struct cache_entry *ce, const char **pathspec)
 469 {
 470         const char *match, *name;
 471         int len;
 472
 473         if (!pathspec)
 474                 return 1;
 475
 476         len = ce_namelen(ce);
 477         name = ce->name;
 478         while ((match = *pathspec++) != NULL) {
 479                 int matchlen = strlen(match);
 480                 if (matchlen > len)
 481                         continue;
 482                 if (memcmp(name, match, matchlen))
 483                         continue;
 484                 if (matchlen && name[matchlen-1] == '/')
 485                         return 1;
 486                 if (name[matchlen] == '/' || !name[matchlen])
 487                         return 1;
 488                 if (!matchlen)
 489                         return 1;
 490         }
 491         return 0;
 492 }
 493
 494 /*
 495  * We fundamentally don't like some paths: we don't want
 496  * dot or dot-dot anywhere, and for obvious reasons don't
 497  * want to recurse into ".git" either.
 498  *
 499  * Also, we don't want double slashes or slashes at the
 500  * end that can make pathnames ambiguous.
 501  */
 502 static int verify_dotfile(const char *rest)
 503 {
 504         /*
 505          * The first character was '.', but that
 506          * has already been discarded, we now test
 507          * the rest.
 508          */
 509         switch (*rest) {
 510         /* "." is not allowed */
 511         case '\0': case '/':
 512                 return 0;
 513
 514         /*
 515          * ".git" followed by  NUL or slash is bad. This
 516          * shares the path end test with the ".." case.
 517          */
 518         case 'g':
 519                 if (rest[1] != 'i')
 520                         break;
 521                 if (rest[2] != 't')
 522                         break;
 523                 rest += 2;
 524         /* fallthrough */
 525         case '.':
 526                 if (rest[1] == '\0' || rest[1] == '/')
 527                         return 0;
 528         }
 529         return 1;
 530 }
 531
 532 int verify_path(const char *path)
 533 {
 534         char c;
 535
 536         goto inside;
 537         for (;;) {
 538                 if (!c)
 539                         return 1;
 540                 if (c == '/') {
 541 inside:
 542                         c = *path++;
 543                         switch (c) {
 544                         default:
 545                                 continue;
 546                         case '/': case '\0':
 547                                 break;
 548                         case '.':
 549                                 if (verify_dotfile(path))
 550                                         continue;
 551                         }
 552                         return 0;
 553                 }
 554                 c = *path++;
 555         }
 556 }
 557
 558 /*
 559  * Do we have another file that has the beginning components being a
 560  * proper superset of the name we're trying to add?
 561  */
 562 static int has_file_name(struct index_state *istate,
 563                          const struct cache_entry *ce, int pos, int ok_to_replace)
 564 {
 565         int retval = 0;
 566         int len = ce_namelen(ce);
 567         int stage = ce_stage(ce);
 568         const char *name = ce->name;
 569
 570         while (pos < istate->cache_nr) {
 571                 struct cache_entry *p = istate->cache[pos++];
 572
 573                 if (len >= ce_namelen(p))
 574                         break;
 575                 if (memcmp(name, p->name, len))
 576                         break;
 577                 if (ce_stage(p) != stage)
 578                         continue;
 579                 if (p->name[len] != '/')
 580                         continue;
 581                 if (!ce_stage(p) && !p->ce_mode)
 582                         continue;
 583                 retval = -1;
 584                 if (!ok_to_replace)
 585                         break;
 586                 remove_index_entry_at(istate, --pos);
 587         }
 588         return retval;
 589 }
 590
 591 /*
 592  * Do we have another file with a pathname that is a proper
 593  * subset of the name we're trying to add?
 594  */
 595 static int has_dir_name(struct index_state *istate,
 596                         const struct cache_entry *ce, int pos, int ok_to_replace)
 597 {
 598         int retval = 0;
 599         int stage = ce_stage(ce);
 600         const char *name = ce->name;
 601         const char *slash = name + ce_namelen(ce);
 602
 603         for (;;) {
 604                 int len;
 605
 606                 for (;;) {
 607                         if (*--slash == '/')
 608                                 break;
 609                         if (slash <= ce->name)
 610                                 return retval;
 611                 }
 612                 len = slash - name;
 613
 614                 pos = index_name_pos(istate, name, ntohs(create_ce_flags(len, stage)));
 615                 if (pos >= 0) {
 616                         /*
 617                          * Found one, but not so fast.  This could
 618                          * be a marker that says "I was here, but
 619                          * I am being removed".  Such an entry is
 620                          * not a part of the resulting tree, and
 621                          * it is Ok to have a directory at the same
 622                          * path.
 623                          */
 624                         if (stage || istate->cache[pos]->ce_mode) {
 625                                 retval = -1;
 626                                 if (!ok_to_replace)
 627                                         break;
 628                                 remove_index_entry_at(istate, pos);
 629                                 continue;
 630                         }
 631                 }
 632                 else
 633                         pos = -pos-1;
 634
 635                 /*
 636                  * Trivial optimization: if we find an entry that
 637                  * already matches the sub-directory, then we know
 638                  * we're ok, and we can exit.
 639                  */
 640                 while (pos < istate->cache_nr) {
 641                         struct cache_entry *p = istate->cache[pos];
 642                         if ((ce_namelen(p) <= len) ||
 643                             (p->name[len] != '/') ||
 644                             memcmp(p->name, name, len))
 645                                 break; /* not our subdirectory */
 646                         if (ce_stage(p) == stage && (stage || p->ce_mode))
 647                                 /* p is at the same stage as our entry, and
 648                                  * is a subdirectory of what we are looking
 649                                  * at, so we cannot have conflicts at our
 650                                  * level or anything shorter.
 651                                  */
 652                                 return retval;
 653                         pos++;
 654                 }
 655         }
 656         return retval;
 657 }
 658
 659 /* We may be in a situation where we already have path/file and path
 660  * is being added, or we already have path and path/file is being
 661  * added.  Either one would result in a nonsense tree that has path
 662  * twice when git-write-tree tries to write it out.  Prevent it.
 663  *
 664  * If ok-to-replace is specified, we remove the conflicting entries
 665  * from the cache so the caller should recompute the insert position.
 666  * When this happens, we return non-zero.
 667  */
 668 static int check_file_directory_conflict(struct index_state *istate,
 669                                          const struct cache_entry *ce,
 670                                          int pos, int ok_to_replace)
 671 {
 672         int retval;
 673
 674         /*
 675          * When ce is an "I am going away" entry, we allow it to be added
 676          */
 677         if (!ce_stage(ce) && !ce->ce_mode)
 678                 return 0;
 679
 680         /*
 681          * We check if the path is a sub-path of a subsequent pathname
 682          * first, since removing those will not change the position
 683          * in the array.
 684          */
 685         retval = has_file_name(istate, ce, pos, ok_to_replace);
 686
 687         /*
 688          * Then check if the path might have a clashing sub-directory
 689          * before it.
 690          */
 691         return retval + has_dir_name(istate, ce, pos, ok_to_replace);
 692 }
 693
 694 static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
 695 {
 696         int pos;
 697         int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
 698         int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
 699         int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
 700
 701         cache_tree_invalidate_path(istate->cache_tree, ce->name);
 702         pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 703
 704         /* existing match? Just replace it. */
 705         if (pos >= 0) {
 706                 istate->cache_changed = 1;
 707                 istate->cache[pos] = ce;
 708                 return 0;
 709         }
 710         pos = -pos-1;
 711
 712         /*
 713          * Inserting a merged entry ("stage 0") into the index
 714          * will always replace all non-merged entries..
 715          */
 716         if (pos < istate->cache_nr && ce_stage(ce) == 0) {
 717                 while (ce_same_name(istate->cache[pos], ce)) {
 718                         ok_to_add = 1;
 719                         if (!remove_index_entry_at(istate, pos))
 720                                 break;
 721                 }
 722         }
 723
 724         if (!ok_to_add)
 725                 return -1;
 726         if (!verify_path(ce->name))
 727                 return -1;
 728
 729         if (!skip_df_check &&
 730             check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
 731                 if (!ok_to_replace)
 732                         return error("'%s' appears as both a file and as a directory",
 733                                      ce->name);
 734                 pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 735                 pos = -pos-1;
 736         }
 737         return pos + 1;
 738 }
 739
 740 int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
 741 {
 742         int pos;
 743
 744         if (option & ADD_CACHE_JUST_APPEND)
 745                 pos = istate->cache_nr;
 746         else {
 747                 int ret;
 748                 ret = add_index_entry_with_check(istate, ce, option);
 749                 if (ret <= 0)
 750                         return ret;
 751                 pos = ret - 1;
 752         }
 753
 754         /* Make sure the array is big enough .. */
 755         if (istate->cache_nr == istate->cache_alloc) {
 756                 istate->cache_alloc = alloc_nr(istate->cache_alloc);
 757                 istate->cache = xrealloc(istate->cache,
 758                                         istate->cache_alloc * sizeof(struct cache_entry *));
 759         }
 760
 761         /* Add it in.. */
 762         istate->cache_nr++;
 763         if (istate->cache_nr > pos + 1)
 764                 memmove(istate->cache + pos + 1,
 765                         istate->cache + pos,
 766                         (istate->cache_nr - pos - 1) * sizeof(ce));
 767         istate->cache[pos] = ce;
 768         istate->cache_changed = 1;
 769         return 0;
 770 }
 771
 772 /*
 773  * "refresh" does not calculate a new sha1 file or bring the
 774  * cache up-to-date for mode/content changes. But what it
 775  * _does_ do is to "re-match" the stat information of a file
 776  * with the cache, so that you can refresh the cache for a
 777  * file that hasn't been changed but where the stat entry is
 778  * out of date.
 779  *
 780  * For example, you'd want to do this after doing a "git-read-tree",
 781  * to link up the stat cache details with the proper files.
 782  */
 783 static struct cache_entry *refresh_cache_ent(struct index_state *istate,
 784                                              struct cache_entry *ce, int really, int *err)
 785 {
 786         struct stat st;
 787         struct cache_entry *updated;
 788         int changed, size;
 789
 790         if (lstat(ce->name, &st) < 0) {
 791                 if (err)
 792                         *err = errno;
 793                 return NULL;
 794         }
 795
 796         changed = ie_match_stat(istate, ce, &st, really);
 797         if (!changed) {
 798                 if (really && assume_unchanged &&
 799                     !(ce->ce_flags & htons(CE_VALID)))
 800                         ; /* mark this one VALID again */
 801                 else
 802                         return ce;
 803         }
 804
 805         if (ie_modified(istate, ce, &st, really)) {
 806                 if (err)
 807                         *err = EINVAL;
 808                 return NULL;
 809         }
 810
 811         size = ce_size(ce);
 812         updated = xmalloc(size);
 813         memcpy(updated, ce, size);
 814         fill_stat_cache_info(updated, &st);
 815
 816         /* In this case, if really is not set, we should leave
 817          * CE_VALID bit alone.  Otherwise, paths marked with
 818          * --no-assume-unchanged (i.e. things to be edited) will
 819          * reacquire CE_VALID bit automatically, which is not
 820          * really what we want.
 821          */
 822         if (!really && assume_unchanged && !(ce->ce_flags & htons(CE_VALID)))
 823                 updated->ce_flags &= ~htons(CE_VALID);
 824
 825         return updated;
 826 }
 827
 828 int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen)
 829 {
 830         int i;
 831         int has_errors = 0;
 832         int really = (flags & REFRESH_REALLY) != 0;
 833         int allow_unmerged = (flags & REFRESH_UNMERGED) != 0;
 834         int quiet = (flags & REFRESH_QUIET) != 0;
 835         int not_new = (flags & REFRESH_IGNORE_MISSING) != 0;
 836
 837         for (i = 0; i < istate->cache_nr; i++) {
 838                 struct cache_entry *ce, *new;
 839                 int cache_errno = 0;
 840
 841                 ce = istate->cache[i];
 842                 if (ce_stage(ce)) {
 843                         while ((i < istate->cache_nr) &&
 844                                ! strcmp(istate->cache[i]->name, ce->name))
 845                                 i++;
 846                         i--;
 847                         if (allow_unmerged)
 848                                 continue;
 849                         printf("%s: needs merge\n", ce->name);
 850                         has_errors = 1;
 851                         continue;
 852                 }
 853
 854                 if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
 855                         continue;
 856
 857                 new = refresh_cache_ent(istate, ce, really, &cache_errno);
 858                 if (new == ce)
 859                         continue;
 860                 if (!new) {
 861                         if (not_new && cache_errno == ENOENT)
 862                                 continue;
 863                         if (really && cache_errno == EINVAL) {
 864                                 /* If we are doing --really-refresh that
 865                                  * means the index is not valid anymore.
 866                                  */
 867                                 ce->ce_flags &= ~htons(CE_VALID);
 868                                 istate->cache_changed = 1;
 869                         }
 870                         if (quiet)
 871                                 continue;
 872                         printf("%s: needs update\n", ce->name);
 873                         has_errors = 1;
 874                         continue;
 875                 }
 876                 istate->cache_changed = 1;
 877                 /* You can NOT just free istate->cache[i] here, since it
 878                  * might not be necessarily malloc()ed but can also come
 879                  * from mmap(). */
 880                 istate->cache[i] = new;
 881         }
 882         return has_errors;
 883 }
 884
 885 struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
 886 {
 887         return refresh_cache_ent(&the_index, ce, really, NULL);
 888 }
 889
 890 static int verify_hdr(struct cache_header *hdr, unsigned long size)
 891 {
 892         SHA_CTX c;
 893         unsigned char sha1[20];
 894
 895         if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
 896                 return error("bad signature");
 897         if (hdr->hdr_version != htonl(2))
 898                 return error("bad index version");
 899         SHA1_Init(&c);
 900         SHA1_Update(&c, hdr, size - 20);
 901         SHA1_Final(sha1, &c);
 902         if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
 903                 return error("bad index file sha1 signature");
 904         return 0;
 905 }
 906
 907 static int read_index_extension(struct index_state *istate,
 908                                 const char *ext, void *data, unsigned long sz)
 909 {
 910         switch (CACHE_EXT(ext)) {
 911         case CACHE_EXT_TREE:
 912                 istate->cache_tree = cache_tree_read(data, sz);
 913                 break;
 914         default:
 915                 if (*ext < 'A' || 'Z' < *ext)
 916                         return error("index uses %.4s extension, which we do not understand",
 917                                      ext);
 918                 fprintf(stderr, "ignoring %.4s extension\n", ext);
 919                 break;
 920         }
 921         return 0;
 922 }
 923
 924 int read_index(struct index_state *istate)
 925 {
 926         return read_index_from(istate, get_index_file());
 927 }
 928
 929 /* remember to discard_cache() before reading a different cache! */
 930 int read_index_from(struct index_state *istate, const char *path)
 931 {
 932         int fd, i;
 933         struct stat st;
 934         unsigned long offset;
 935         struct cache_header *hdr;
 936
 937         errno = EBUSY;
 938         if (istate->mmap)
 939                 return istate->cache_nr;
 940
 941         errno = ENOENT;
 942         istate->timestamp = 0;
 943         fd = open(path, O_RDONLY);
 944         if (fd < 0) {
 945                 if (errno == ENOENT)
 946                         return 0;
 947                 die("index file open failed (%s)", strerror(errno));
 948         }
 949
 950         if (fstat(fd, &st))
 951                 die("cannot stat the open index (%s)", strerror(errno));
 952
 953         errno = EINVAL;
 954         istate->mmap_size = xsize_t(st.st_size);
 955         if (istate->mmap_size < sizeof(struct cache_header) + 20)
 956                 die("index file smaller than expected");
 957
 958         istate->mmap = xmmap(NULL, istate->mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
 959         close(fd);
 960
 961         hdr = istate->mmap;
 962         if (verify_hdr(hdr, istate->mmap_size) < 0)
 963                 goto unmap;
 964
 965         istate->cache_nr = ntohl(hdr->hdr_entries);
 966         istate->cache_alloc = alloc_nr(istate->cache_nr);
 967         istate->cache = xcalloc(istate->cache_alloc, sizeof(struct cache_entry *));
 968
 969         offset = sizeof(*hdr);
 970         for (i = 0; i < istate->cache_nr; i++) {
 971                 struct cache_entry *ce;
 972
 973                 ce = (struct cache_entry *)((char *)(istate->mmap) + offset);
 974                 offset = offset + ce_size(ce);
 975                 istate->cache[i] = ce;
 976         }
 977         istate->timestamp = st.st_mtime;
 978         while (offset <= istate->mmap_size - 20 - 8) {
 979                 /* After an array of active_nr index entries,
 980                  * there can be arbitrary number of extended
 981                  * sections, each of which is prefixed with
 982                  * extension name (4-byte) and section length
 983                  * in 4-byte network byte order.
 984                  */
 985                 unsigned long extsize;
 986                 memcpy(&extsize, (char *)(istate->mmap) + offset + 4, 4);
 987                 extsize = ntohl(extsize);
 988                 if (read_index_extension(istate,
 989                                          ((const char *) (istate->mmap)) + offset,
 990                                          (char *) (istate->mmap) + offset + 8,
 991                                          extsize) < 0)
 992                         goto unmap;
 993                 offset += 8;
 994                 offset += extsize;
 995         }
 996         return istate->cache_nr;
 997
 998 unmap:
 999         munmap(istate->mmap, istate->mmap_size);
1000         errno = EINVAL;
1001         die("index file corrupt");
1002 }
1003
1004 int discard_index(struct index_state *istate)
1005 {
1006         int ret;
1007
1008         istate->cache_nr = 0;
1009         istate->cache_changed = 0;
1010         istate->timestamp = 0;
1011         cache_tree_free(&(istate->cache_tree));
1012         if (istate->mmap == NULL)
1013                 return 0;
1014         ret = munmap(istate->mmap, istate->mmap_size);
1015         istate->mmap = NULL;
1016         istate->mmap_size = 0;
1017
1018         /* no need to throw away allocated active_cache */
1019         return ret;
1020 }
1021
1022 #define WRITE_BUFFER_SIZE 8192
1023 static unsigned char write_buffer[WRITE_BUFFER_SIZE];
1024 static unsigned long write_buffer_len;
1025
1026 static int ce_write_flush(SHA_CTX *context, int fd)
1027 {
1028         unsigned int buffered = write_buffer_len;
1029         if (buffered) {
1030                 SHA1_Update(context, write_buffer, buffered);
1031                 if (write_in_full(fd, write_buffer, buffered) != buffered)
1032                         return -1;
1033                 write_buffer_len = 0;
1034         }
1035         return 0;
1036 }
1037
1038 static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
1039 {
1040         while (len) {
1041                 unsigned int buffered = write_buffer_len;
1042                 unsigned int partial = WRITE_BUFFER_SIZE - buffered;
1043                 if (partial > len)
1044                         partial = len;
1045                 memcpy(write_buffer + buffered, data, partial);
1046                 buffered += partial;
1047                 if (buffered == WRITE_BUFFER_SIZE) {
1048                         write_buffer_len = buffered;
1049                         if (ce_write_flush(context, fd))
1050                                 return -1;
1051                         buffered = 0;
1052                 }
1053                 write_buffer_len = buffered;
1054                 len -= partial;
1055                 data = (char *) data + partial;
1056         }
1057         return 0;
1058 }
1059
1060 static int write_index_ext_header(SHA_CTX *context, int fd,
1061                                   unsigned int ext, unsigned int sz)
1062 {
1063         ext = htonl(ext);
1064         sz = htonl(sz);
1065         return ((ce_write(context, fd, &ext, 4) < 0) ||
1066                 (ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1067 }
1068
1069 static int ce_flush(SHA_CTX *context, int fd)
1070 {
1071         unsigned int left = write_buffer_len;
1072
1073         if (left) {
1074                 write_buffer_len = 0;
1075                 SHA1_Update(context, write_buffer, left);
1076         }
1077
1078         /* Flush first if not enough space for SHA1 signature */
1079         if (left + 20 > WRITE_BUFFER_SIZE) {
1080                 if (write_in_full(fd, write_buffer, left) != left)
1081                         return -1;
1082                 left = 0;
1083         }
1084
1085         /* Append the SHA1 signature at the end */
1086         SHA1_Final(write_buffer + left, context);
1087         left += 20;
1088         return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1089 }
1090
1091 static void ce_smudge_racily_clean_entry(struct cache_entry *ce)
1092 {
1093         /*
1094          * The only thing we care about in this function is to smudge the
1095          * falsely clean entry due to touch-update-touch race, so we leave
1096          * everything else as they are.  We are called for entries whose
1097          * ce_mtime match the index file mtime.
1098          */
1099         struct stat st;
1100
1101         if (lstat(ce->name, &st) < 0)
1102                 return;
1103         if (ce_match_stat_basic(ce, &st))
1104                 return;
1105         if (ce_modified_check_fs(ce, &st)) {
1106                 /* This is "racily clean"; smudge it.  Note that this
1107                  * is a tricky code.  At first glance, it may appear
1108                  * that it can break with this sequence:
1109                  *
1110                  * $ echo xyzzy >frotz
1111                  * $ git-update-index --add frotz
1112                  * $ : >frotz
1113                  * $ sleep 3
1114                  * $ echo filfre >nitfol
1115                  * $ git-update-index --add nitfol
1116                  *
1117                  * but it does not.  When the second update-index runs,
1118                  * it notices that the entry "frotz" has the same timestamp
1119                  * as index, and if we were to smudge it by resetting its
1120                  * size to zero here, then the object name recorded
1121                  * in index is the 6-byte file but the cached stat information
1122                  * becomes zero --- which would then match what we would
1123                  * obtain from the filesystem next time we stat("frotz").
1124                  *
1125                  * However, the second update-index, before calling
1126                  * this function, notices that the cached size is 6
1127                  * bytes and what is on the filesystem is an empty
1128                  * file, and never calls us, so the cached size information
1129                  * for "frotz" stays 6 which does not match the filesystem.
1130                  */
1131                 ce->ce_size = htonl(0);
1132         }
1133 }
1134
1135 int write_index(struct index_state *istate, int newfd)
1136 {
1137         SHA_CTX c;
1138         struct cache_header hdr;
1139         int i, err, removed;
1140         struct cache_entry **cache = istate->cache;
1141         int entries = istate->cache_nr;
1142
1143         for (i = removed = 0; i < entries; i++)
1144                 if (!cache[i]->ce_mode)
1145                         removed++;
1146
1147         hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1148         hdr.hdr_version = htonl(2);
1149         hdr.hdr_entries = htonl(entries - removed);
1150
1151         SHA1_Init(&c);
1152         if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1153                 return -1;
1154
1155         for (i = 0; i < entries; i++) {
1156                 struct cache_entry *ce = cache[i];
1157                 if (!ce->ce_mode)
1158                         continue;
1159                 if (istate->timestamp &&
1160                     istate->timestamp <= ntohl(ce->ce_mtime.sec))
1161                         ce_smudge_racily_clean_entry(ce);
1162                 if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1163                         return -1;
1164         }
1165
1166         /* Write extension data here */
1167         if (istate->cache_tree) {
1168                 struct strbuf sb;
1169
1170                 strbuf_init(&sb, 0);
1171                 cache_tree_write(&sb, istate->cache_tree);
1172                 err = write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sb.len) < 0
1173                         || ce_write(&c, newfd, sb.buf, sb.len) < 0;
1174                 strbuf_release(&sb);
1175                 if (err)
1176                         return -1;
1177         }
1178         return ce_flush(&c, newfd);
1179 }