sha1_file.c

   1 /*
   2  * GIT - The information manager from hell
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  *
   6  * This handles basic git sha1 object files - packing, unpacking,
   7  * creation etc.
   8  */
   9 #include <sys/types.h>
  10 #include <dirent.h>
  11 #include "cache.h"
  12 #include "delta.h"
  13 #include "pack.h"
  14
  15 #ifndef O_NOATIME
  16 #if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
  17 #define O_NOATIME 01000000
  18 #else
  19 #define O_NOATIME 0
  20 #endif
  21 #endif
  22
  23 static unsigned int sha1_file_open_flag = O_NOATIME;
  24
  25 static unsigned hexval(char c)
  26 {
  27         if (c >= '0' && c <= '9')
  28                 return c - '0';
  29         if (c >= 'a' && c <= 'f')
  30                 return c - 'a' + 10;
  31         if (c >= 'A' && c <= 'F')
  32                 return c - 'A' + 10;
  33         return ~0;
  34 }
  35
  36 int get_sha1_hex(const char *hex, unsigned char *sha1)
  37 {
  38         int i;
  39         for (i = 0; i < 20; i++) {
  40                 unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]);
  41                 if (val & ~0xff)
  42                         return -1;
  43                 *sha1++ = val;
  44                 hex += 2;
  45         }
  46         return 0;
  47 }
  48
  49 static int get_sha1_file(const char *path, unsigned char *result)
  50 {
  51         char buffer[60];
  52         int fd = open(path, O_RDONLY);
  53         int len;
  54
  55         if (fd < 0)
  56                 return -1;
  57         len = read(fd, buffer, sizeof(buffer));
  58         close(fd);
  59         if (len < 40)
  60                 return -1;
  61         return get_sha1_hex(buffer, result);
  62 }
  63
  64 static char *git_dir, *git_object_dir, *git_index_file, *git_refs_dir;
  65 static void setup_git_env(void)
  66 {
  67         git_dir = gitenv(GIT_DIR_ENVIRONMENT);
  68         if (!git_dir)
  69                 git_dir = DEFAULT_GIT_DIR_ENVIRONMENT;
  70         git_object_dir = gitenv(DB_ENVIRONMENT);
  71         if (!git_object_dir) {
  72                 git_object_dir = xmalloc(strlen(git_dir) + 9);
  73                 sprintf(git_object_dir, "%s/objects", git_dir);
  74         }
  75         git_refs_dir = xmalloc(strlen(git_dir) + 6);
  76         sprintf(git_refs_dir, "%s/refs", git_dir);
  77         git_index_file = gitenv(INDEX_ENVIRONMENT);
  78         if (!git_index_file) {
  79                 git_index_file = xmalloc(strlen(git_dir) + 7);
  80                 sprintf(git_index_file, "%s/index", git_dir);
  81         }
  82 }
  83
  84 char *get_object_directory(void)
  85 {
  86         if (!git_object_dir)
  87                 setup_git_env();
  88         return git_object_dir;
  89 }
  90
  91 char *get_refs_directory(void)
  92 {
  93         if (!git_refs_dir)
  94                 setup_git_env();
  95         return git_refs_dir;
  96 }
  97
  98 char *get_index_file(void)
  99 {
 100         if (!git_index_file)
 101                 setup_git_env();
 102         return git_index_file;
 103 }
 104
 105 int get_sha1(const char *str, unsigned char *sha1)
 106 {
 107         static char pathname[PATH_MAX];
 108         static const char *prefix[] = {
 109                 "",
 110                 "refs",
 111                 "refs/tags",
 112                 "refs/heads",
 113                 "refs/snap",
 114                 NULL
 115         };
 116         const char **p;
 117
 118         if (!get_sha1_hex(str, sha1))
 119                 return 0;
 120
 121         if (!git_dir)
 122                 setup_git_env();
 123         for (p = prefix; *p; p++) {
 124                 snprintf(pathname, sizeof(pathname), "%s/%s/%s",
 125                          git_dir, *p, str);
 126                 if (!get_sha1_file(pathname, sha1))
 127                         return 0;
 128         }
 129
 130         return -1;
 131 }
 132
 133 char * sha1_to_hex(const unsigned char *sha1)
 134 {
 135         static char buffer[50];
 136         static const char hex[] = "0123456789abcdef";
 137         char *buf = buffer;
 138         int i;
 139
 140         for (i = 0; i < 20; i++) {
 141                 unsigned int val = *sha1++;
 142                 *buf++ = hex[val >> 4];
 143                 *buf++ = hex[val & 0xf];
 144         }
 145         return buffer;
 146 }
 147
 148 static void fill_sha1_path(char *pathbuf, const unsigned char *sha1)
 149 {
 150         int i;
 151         for (i = 0; i < 20; i++) {
 152                 static char hex[] = "0123456789abcdef";
 153                 unsigned int val = sha1[i];
 154                 char *pos = pathbuf + i*2 + (i > 0);
 155                 *pos++ = hex[val >> 4];
 156                 *pos = hex[val & 0xf];
 157         }
 158 }
 159
 160 /*
 161  * NOTE! This returns a statically allocated buffer, so you have to be
 162  * careful about using it. Do a "strdup()" if you need to save the
 163  * filename.
 164  *
 165  * Also note that this returns the location for creating.  Reading
 166  * SHA1 file can happen from any alternate directory listed in the
 167  * DB_ENVIRONMENT environment variable if it is not found in
 168  * the primary object database.
 169  */
 170 char *sha1_file_name(const unsigned char *sha1)
 171 {
 172         static char *name, *base;
 173
 174         if (!base) {
 175                 const char *sha1_file_directory = get_object_directory();
 176                 int len = strlen(sha1_file_directory);
 177                 base = xmalloc(len + 60);
 178                 memcpy(base, sha1_file_directory, len);
 179                 memset(base+len, 0, 60);
 180                 base[len] = '/';
 181                 base[len+3] = '/';
 182                 name = base + len + 1;
 183         }
 184         fill_sha1_path(name, sha1);
 185         return base;
 186 }
 187
 188 struct alternate_object_database *alt_odb;
 189
 190 /*
 191  * Prepare alternate object database registry.
 192  * alt_odb points at an array of struct alternate_object_database.
 193  * This array is terminated with an element that has both its base
 194  * and name set to NULL.  alt_odb[n] comes from n'th non-empty
 195  * element from colon separated ALTERNATE_DB_ENVIRONMENT environment
 196  * variable, and its base points at a statically allocated buffer
 197  * that contains "/the/directory/corresponding/to/.git/objects/...",
 198  * while its name points just after the slash at the end of
 199  * ".git/objects/" in the example above, and has enough space to hold
 200  * 40-byte hex SHA1, an extra slash for the first level indirection,
 201  * and the terminating NUL.
 202  * This function allocates the alt_odb array and all the strings
 203  * pointed by base fields of the array elements with one xmalloc();
 204  * the string pool immediately follows the array.
 205  */
 206 void prepare_alt_odb(void)
 207 {
 208         int pass, totlen, i;
 209         const char *cp, *last;
 210         char *op = NULL;
 211         const char *alt = gitenv(ALTERNATE_DB_ENVIRONMENT) ? : "";
 212
 213         if (alt_odb)
 214                 return;
 215         /* The first pass counts how large an area to allocate to
 216          * hold the entire alt_odb structure, including array of
 217          * structs and path buffers for them.  The second pass fills
 218          * the structure and prepares the path buffers for use by
 219          * fill_sha1_path().
 220          */
 221         for (totlen = pass = 0; pass < 2; pass++) {
 222                 last = alt;
 223                 i = 0;
 224                 do {
 225                         cp = strchr(last, ':') ? : last + strlen(last);
 226                         if (last != cp) {
 227                                 /* 43 = 40-byte + 2 '/' + terminating NUL */
 228                                 int pfxlen = cp - last;
 229                                 int entlen = pfxlen + 43;
 230                                 if (pass == 0)
 231                                         totlen += entlen;
 232                                 else {
 233                                         alt_odb[i].base = op;
 234                                         alt_odb[i].name = op + pfxlen + 1;
 235                                         memcpy(op, last, pfxlen);
 236                                         op[pfxlen] = op[pfxlen + 3] = '/';
 237                                         op[entlen-1] = 0;
 238                                         op += entlen;
 239                                 }
 240                                 i++;
 241                         }
 242                         while (*cp && *cp == ':')
 243                                 cp++;
 244                         last = cp;
 245                 } while (*cp);
 246                 if (pass)
 247                         break;
 248                 alt_odb = xmalloc(sizeof(*alt_odb) * (i + 1) + totlen);
 249                 alt_odb[i].base = alt_odb[i].name = NULL;
 250                 op = (char*)(&alt_odb[i+1]);
 251         }
 252 }
 253
 254 static char *find_sha1_file(const unsigned char *sha1, struct stat *st)
 255 {
 256         int i;
 257         char *name = sha1_file_name(sha1);
 258
 259         if (!stat(name, st))
 260                 return name;
 261         prepare_alt_odb();
 262         for (i = 0; (name = alt_odb[i].name) != NULL; i++) {
 263                 fill_sha1_path(name, sha1);
 264                 if (!stat(alt_odb[i].base, st))
 265                         return alt_odb[i].base;
 266         }
 267         return NULL;
 268 }
 269
 270 #define PACK_MAX_SZ (1<<26)
 271 static int pack_used_ctr;
 272 static unsigned long pack_mapped;
 273 struct packed_git *packed_git;
 274
 275 struct pack_entry {
 276         unsigned int offset;
 277         unsigned char sha1[20];
 278         struct packed_git *p;
 279 };
 280
 281 static int check_packed_git_idx(const char *path, unsigned long *idx_size_,
 282                                 void **idx_map_)
 283 {
 284         void *idx_map;
 285         unsigned int *index;
 286         unsigned long idx_size;
 287         int nr, i;
 288         int fd = open(path, O_RDONLY);
 289         struct stat st;
 290         if (fd < 0)
 291                 return -1;
 292         if (fstat(fd, &st)) {
 293                 close(fd);
 294                 return -1;
 295         }
 296         idx_size = st.st_size;
 297         idx_map = mmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
 298         close(fd);
 299         if (idx_map == MAP_FAILED)
 300                 return -1;
 301
 302         index = idx_map;
 303
 304         /* check index map */
 305         if (idx_size < 4*256 + 20)
 306                 return error("index file too small");
 307         nr = 0;
 308         for (i = 0; i < 256; i++) {
 309                 unsigned int n = ntohl(index[i]);
 310                 if (n < nr)
 311                         return error("non-monotonic index");
 312                 nr = n;
 313         }
 314
 315         /*
 316          * Total size:
 317          *  - 256 index entries 4 bytes each
 318          *  - 24-byte entries * nr (20-byte sha1 + 4-byte offset)
 319          *  - 20-byte SHA1 of the packfile
 320          *  - 20-byte SHA1 file checksum
 321          */
 322         if (idx_size != 4*256 + nr * 24 + 20 + 20)
 323                 return error("wrong index file size");
 324
 325         *idx_map_ = idx_map;
 326         *idx_size_ = idx_size;
 327         return 0;
 328 }
 329
 330 static void unuse_one_packed_git(void)
 331 {
 332         /* NOTYET */
 333 }
 334
 335 static int use_packed_git(struct packed_git *p)
 336 {
 337         if (!p->pack_base) {
 338                 int fd;
 339                 struct stat st;
 340                 void *map;
 341
 342                 pack_mapped += p->pack_size;
 343                 while (PACK_MAX_SZ < pack_mapped)
 344                         unuse_one_packed_git();
 345                 fd = open(p->pack_name, O_RDONLY);
 346                 if (fd < 0)
 347                         return -1;
 348                 if (fstat(fd, &st)) {
 349                         close(fd);
 350                         return -1;
 351                 }
 352                 if (st.st_size != p->pack_size)
 353                         return -1;
 354                 map = mmap(NULL, p->pack_size, PROT_READ, MAP_PRIVATE, fd, 0);
 355                 close(fd);
 356                 if (map == MAP_FAILED)
 357                         return -1;
 358                 p->pack_base = map;
 359         }
 360         p->pack_last_used = pack_used_ctr++;
 361         return 0;
 362 }
 363
 364 static struct packed_git *add_packed_git(char *path, int path_len)
 365 {
 366         struct stat st;
 367         struct packed_git *p;
 368         unsigned long idx_size;
 369         void *idx_map;
 370
 371         if (check_packed_git_idx(path, &idx_size, &idx_map))
 372                 return NULL;
 373
 374         /* do we have a corresponding .pack file? */
 375         strcpy(path + path_len - 4, ".pack");
 376         if (stat(path, &st) || !S_ISREG(st.st_mode)) {
 377                 munmap(idx_map, idx_size);
 378                 return NULL;
 379         }
 380         /* ok, it looks sane as far as we can check without
 381          * actually mapping the pack file.
 382          */
 383         p = xmalloc(sizeof(*p) + path_len + 2);
 384         strcpy(p->pack_name, path);
 385         p->index_size = idx_size;
 386         p->pack_size = st.st_size;
 387         p->index_base = idx_map;
 388         p->next = NULL;
 389         p->pack_base = NULL;
 390         p->pack_last_used = 0;
 391         return p;
 392 }
 393
 394 static void prepare_packed_git_one(char *objdir)
 395 {
 396         char path[PATH_MAX];
 397         int len;
 398         DIR *dir;
 399         struct dirent *de;
 400
 401         sprintf(path, "%s/pack", objdir);
 402         len = strlen(path);
 403         dir = opendir(path);
 404         if (!dir)
 405                 return;
 406         path[len++] = '/';
 407         while ((de = readdir(dir)) != NULL) {
 408                 int namelen = strlen(de->d_name);
 409                 struct packed_git *p;
 410
 411                 if (strcmp(de->d_name + namelen - 4, ".idx"))
 412                         continue;
 413
 414                 /* we have .idx.  Is it a file we can map? */
 415                 strcpy(path + len, de->d_name);
 416                 p = add_packed_git(path, len + namelen);
 417                 if (!p)
 418                         continue;
 419                 p->next = packed_git;
 420                 packed_git = p;
 421         }
 422 }
 423
 424 void prepare_packed_git(void)
 425 {
 426         int i;
 427         static int run_once = 0;
 428
 429         if (run_once++)
 430                 return;
 431
 432         prepare_packed_git_one(get_object_directory());
 433         prepare_alt_odb();
 434         for (i = 0; alt_odb[i].base != NULL; i++) {
 435                 alt_odb[i].name[0] = 0;
 436                 prepare_packed_git_one(alt_odb[i].base);
 437         }
 438 }
 439
 440 int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type)
 441 {
 442         char header[100];
 443         unsigned char real_sha1[20];
 444         SHA_CTX c;
 445
 446         SHA1_Init(&c);
 447         SHA1_Update(&c, header, 1+sprintf(header, "%s %lu", type, size));
 448         SHA1_Update(&c, map, size);
 449         SHA1_Final(real_sha1, &c);
 450         return memcmp(sha1, real_sha1, 20) ? -1 : 0;
 451 }
 452
 453 static void *map_sha1_file_internal(const unsigned char *sha1,
 454                                     unsigned long *size,
 455                                     int say_error)
 456 {
 457         struct stat st;
 458         void *map;
 459         int fd;
 460         char *filename = find_sha1_file(sha1, &st);
 461
 462         if (!filename) {
 463                 if (say_error)
 464                         error("cannot map sha1 file %s", sha1_to_hex(sha1));
 465                 return NULL;
 466         }
 467
 468         fd = open(filename, O_RDONLY | sha1_file_open_flag);
 469         if (fd < 0) {
 470                 /* See if it works without O_NOATIME */
 471                 switch (sha1_file_open_flag) {
 472                 default:
 473                         fd = open(filename, O_RDONLY);
 474                         if (fd >= 0)
 475                                 break;
 476                 /* Fallthrough */
 477                 case 0:
 478                         if (say_error)
 479                                 perror(filename);
 480                         return NULL;
 481                 }
 482
 483                 /* If it failed once, it will probably fail again.
 484                  * Stop using O_NOATIME
 485                  */
 486                 sha1_file_open_flag = 0;
 487         }
 488         map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
 489         close(fd);
 490         if (-1 == (int)(long)map)
 491                 return NULL;
 492         *size = st.st_size;
 493         return map;
 494 }
 495
 496 void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
 497 {
 498         return map_sha1_file_internal(sha1, size, 1);
 499 }
 500
 501 int unpack_sha1_header(z_stream *stream, void *map, unsigned long mapsize, void *buffer, unsigned long size)
 502 {
 503         /* Get the data stream */
 504         memset(stream, 0, sizeof(*stream));
 505         stream->next_in = map;
 506         stream->avail_in = mapsize;
 507         stream->next_out = buffer;
 508         stream->avail_out = size;
 509
 510         inflateInit(stream);
 511         return inflate(stream, 0);
 512 }
 513
 514 void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size)
 515 {
 516         int bytes = strlen(buffer) + 1;
 517         unsigned char *buf = xmalloc(1+size);
 518
 519         memcpy(buf, buffer + bytes, stream->total_out - bytes);
 520         bytes = stream->total_out - bytes;
 521         if (bytes < size) {
 522                 stream->next_out = buf + bytes;
 523                 stream->avail_out = size - bytes;
 524                 while (inflate(stream, Z_FINISH) == Z_OK)
 525                         /* nothing */;
 526         }
 527         buf[size] = 0;
 528         inflateEnd(stream);
 529         return buf;
 530 }
 531
 532 /*
 533  * We used to just use "sscanf()", but that's actually way
 534  * too permissive for what we want to check. So do an anal
 535  * object header parse by hand.
 536  */
 537 int parse_sha1_header(char *hdr, char *type, unsigned long *sizep)
 538 {
 539         int i;
 540         unsigned long size;
 541
 542         /*
 543          * The type can be at most ten bytes (including the
 544          * terminating '\0' that we add), and is followed by
 545          * a space.
 546          */
 547         i = 10;
 548         for (;;) {
 549                 char c = *hdr++;
 550                 if (c == ' ')
 551                         break;
 552                 if (!--i)
 553                         return -1;
 554                 *type++ = c;
 555         }
 556         *type = 0;
 557
 558         /*
 559          * The length must follow immediately, and be in canonical
 560          * decimal format (ie "010" is not valid).
 561          */
 562         size = *hdr++ - '0';
 563         if (size > 9)
 564                 return -1;
 565         if (size) {
 566                 for (;;) {
 567                         unsigned long c = *hdr - '0';
 568                         if (c > 9)
 569                                 break;
 570                         hdr++;
 571                         size = size * 10 + c;
 572                 }
 573         }
 574         *sizep = size;
 575
 576         /*
 577          * The length must be followed by a zero byte
 578          */
 579         return *hdr ? -1 : 0;
 580 }
 581
 582 void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
 583 {
 584         int ret;
 585         z_stream stream;
 586         char hdr[8192];
 587
 588         ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr));
 589         if (ret < Z_OK || parse_sha1_header(hdr, type, size) < 0)
 590                 return NULL;
 591
 592         return unpack_sha1_rest(&stream, hdr, *size);
 593 }
 594
 595 static int packed_delta_info(unsigned char *base_sha1,
 596                              unsigned long delta_size,
 597                              unsigned long left,
 598                              char *type,
 599                              unsigned long *sizep)
 600 {
 601         unsigned char *data;
 602         unsigned char delta_head[64];
 603         int i;
 604         unsigned char cmd;
 605         unsigned long data_size, result_size, base_size, verify_base_size;
 606         z_stream stream;
 607         int st;
 608
 609         if (left < 20)
 610                 die("truncated pack file");
 611         if (sha1_object_info(base_sha1, type, &base_size))
 612                 die("cannot get info for delta-pack base");
 613
 614         data = base_sha1 + 20;
 615         data_size = left - 20;
 616
 617         memset(&stream, 0, sizeof(stream));
 618
 619         stream.next_in = data;
 620         stream.avail_in = data_size;
 621         stream.next_out = delta_head;
 622         stream.avail_out = sizeof(delta_head);
 623
 624         inflateInit(&stream);
 625         st = inflate(&stream, Z_FINISH);
 626         inflateEnd(&stream);
 627         if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head))
 628                 die("delta data unpack-initial failed");
 629
 630         /* Examine the initial part of the delta to figure out
 631          * the result size.  Verify the base size while we are at it.
 632          */
 633         data = delta_head;
 634         verify_base_size = i = 0;
 635         cmd = *data++;
 636         while (cmd) {
 637                 if (cmd & 1)
 638                         verify_base_size |= *data++ << i;
 639                 i += 8;
 640                 cmd >>= 1;
 641         }
 642
 643         /* Read the result size */
 644         result_size = i = 0;
 645         cmd = *data++;
 646         while (cmd) {
 647                 if (cmd & 1)
 648                         result_size |= *data++ << i;
 649                 i += 8;
 650                 cmd >>= 1;
 651         }
 652         if (verify_base_size != base_size)
 653                 die("delta base size mismatch");
 654
 655         *sizep = result_size;
 656         return 0;
 657 }
 658
 659 static unsigned long unpack_object_header(struct packed_git *p, unsigned long offset,
 660         enum object_type *type, unsigned long *sizep)
 661 {
 662         unsigned char *pack, c;
 663         unsigned long size;
 664
 665         if (offset >= p->pack_size)
 666                 die("object offset outside of pack file");
 667
 668         pack =  p->pack_base + offset;
 669         c = *pack++;
 670         offset++;
 671         *type = (c >> 4) & 7;
 672         size = c & 15;
 673         while (c & 0x80) {
 674                 if (offset >= p->pack_size)
 675                         die("object offset outside of pack file");
 676                 c = *pack++;
 677                 offset++;
 678                 size = (size << 7) | (c & 0x7f);
 679         }
 680         *sizep = size;
 681         return offset;
 682 }
 683
 684 static int packed_object_info(struct pack_entry *entry,
 685                               char *type, unsigned long *sizep)
 686 {
 687         struct packed_git *p = entry->p;
 688         unsigned long offset, size, left;
 689         unsigned char *pack;
 690         enum object_type kind;
 691
 692         if (use_packed_git(p))
 693                 die("cannot map packed file");
 694
 695         offset = unpack_object_header(p, entry->offset, &kind, &size);
 696         pack = p->pack_base + offset;
 697         left = p->pack_size - offset;
 698
 699         switch (kind) {
 700         case OBJ_DELTA:
 701                 return packed_delta_info(pack, size, left, type, sizep);
 702                 break;
 703         case OBJ_COMMIT:
 704                 strcpy(type, "commit");
 705                 break;
 706         case OBJ_TREE:
 707                 strcpy(type, "tree");
 708                 break;
 709         case OBJ_BLOB:
 710                 strcpy(type, "blob");
 711                 break;
 712         case OBJ_TAG:
 713                 strcpy(type, "tag");
 714                 break;
 715         default:
 716                 die("corrupted pack file");
 717         }
 718         *sizep = size;
 719         return 0;
 720 }
 721
 722 /* forward declaration for a mutually recursive function */
 723 static void *unpack_entry(struct pack_entry *, char *, unsigned long *);
 724
 725 static void *unpack_delta_entry(unsigned char *base_sha1,
 726                                 unsigned long delta_size,
 727                                 unsigned long left,
 728                                 char *type,
 729                                 unsigned long *sizep)
 730 {
 731         void *data, *delta_data, *result, *base;
 732         unsigned long data_size, result_size, base_size;
 733         z_stream stream;
 734         int st;
 735
 736         if (left < 20)
 737                 die("truncated pack file");
 738         data = base_sha1 + 20;
 739         data_size = left - 20;
 740         delta_data = xmalloc(delta_size);
 741
 742         memset(&stream, 0, sizeof(stream));
 743
 744         stream.next_in = data;
 745         stream.avail_in = data_size;
 746         stream.next_out = delta_data;
 747         stream.avail_out = delta_size;
 748
 749         inflateInit(&stream);
 750         st = inflate(&stream, Z_FINISH);
 751         inflateEnd(&stream);
 752         if ((st != Z_STREAM_END) || stream.total_out != delta_size)
 753                 die("delta data unpack failed");
 754
 755         /* This may recursively unpack the base, which is what we want */
 756         base = read_sha1_file(base_sha1, type, &base_size);
 757         if (!base)
 758                 die("failed to read delta-pack base object %s",
 759                     sha1_to_hex(base_sha1));
 760         result = patch_delta(base, base_size,
 761                              delta_data, delta_size,
 762                              &result_size);
 763         if (!result)
 764                 die("failed to apply delta");
 765         free(delta_data);
 766         free(base);
 767         *sizep = result_size;
 768         return result;
 769 }
 770
 771 static void *unpack_non_delta_entry(unsigned char *data,
 772                                     unsigned long size,
 773                                     unsigned long left)
 774 {
 775         int st;
 776         z_stream stream;
 777         char *buffer;
 778
 779         buffer = xmalloc(size + 1);
 780         buffer[size] = 0;
 781         memset(&stream, 0, sizeof(stream));
 782         stream.next_in = data;
 783         stream.avail_in = left;
 784         stream.next_out = buffer;
 785         stream.avail_out = size;
 786
 787         inflateInit(&stream);
 788         st = inflate(&stream, Z_FINISH);
 789         inflateEnd(&stream);
 790         if ((st != Z_STREAM_END) || stream.total_out != size) {
 791                 free(buffer);
 792                 return NULL;
 793         }
 794
 795         return buffer;
 796 }
 797
 798 static void *unpack_entry(struct pack_entry *entry,
 799                           char *type, unsigned long *sizep)
 800 {
 801         struct packed_git *p = entry->p;
 802         unsigned long offset, size, left;
 803         unsigned char *pack;
 804         enum object_type kind;
 805
 806         if (use_packed_git(p))
 807                 die("cannot map packed file");
 808
 809         offset = unpack_object_header(p, entry->offset, &kind, &size);
 810         pack = p->pack_base + offset;
 811         left = p->pack_size - offset;
 812         switch (kind) {
 813         case OBJ_DELTA:
 814                 return unpack_delta_entry(pack, size, left, type, sizep);
 815         case OBJ_COMMIT:
 816                 strcpy(type, "commit");
 817                 break;
 818         case OBJ_TREE:
 819                 strcpy(type, "tree");
 820                 break;
 821         case OBJ_BLOB:
 822                 strcpy(type, "blob");
 823                 break;
 824         case OBJ_TAG:
 825                 strcpy(type, "tag");
 826                 break;
 827         default:
 828                 die("corrupted pack file");
 829         }
 830         *sizep = size;
 831         return unpack_non_delta_entry(pack, size, left);
 832 }
 833
 834 int num_packed_objects(const struct packed_git *p)
 835 {
 836         /* See check_packed_git_idx and pack-objects.c */
 837         return (p->index_size - 20 - 20 - 4*256) / 24;
 838 }
 839
 840 int nth_packed_object_sha1(const struct packed_git *p, int n,
 841                            unsigned char* sha1)
 842 {
 843         void *index = p->index_base + 256;
 844         if (n < 0 || num_packed_objects(p) <= n)
 845                 return -1;
 846         memcpy(sha1, (index + 24 * n + 4), 20);
 847         return 0;
 848 }
 849
 850 static int find_pack_entry_1(const unsigned char *sha1,
 851                              struct pack_entry *e, struct packed_git *p)
 852 {
 853         int *level1_ofs = p->index_base;
 854         int hi = ntohl(level1_ofs[*sha1]);
 855         int lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1]));
 856         void *index = p->index_base + 256;
 857
 858         do {
 859                 int mi = (lo + hi) / 2;
 860                 int cmp = memcmp(index + 24 * mi + 4, sha1, 20);
 861                 if (!cmp) {
 862                         e->offset = ntohl(*((int*)(index + 24 * mi)));
 863                         memcpy(e->sha1, sha1, 20);
 864                         e->p = p;
 865                         return 1;
 866                 }
 867                 if (cmp > 0)
 868                         hi = mi;
 869                 else
 870                         lo = mi+1;
 871         } while (lo < hi);
 872         return 0;
 873 }
 874
 875 static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e)
 876 {
 877         struct packed_git *p;
 878         prepare_packed_git();
 879
 880         for (p = packed_git; p; p = p->next) {
 881                 if (find_pack_entry_1(sha1, e, p))
 882                         return 1;
 883         }
 884         return 0;
 885 }
 886
 887 int sha1_object_info(const unsigned char *sha1, char *type, unsigned long *sizep)
 888 {
 889         int status;
 890         unsigned long mapsize, size;
 891         void *map;
 892         z_stream stream;
 893         char hdr[128];
 894
 895         map = map_sha1_file_internal(sha1, &mapsize, 0);
 896         if (!map) {
 897                 struct pack_entry e;
 898
 899                 if (!find_pack_entry(sha1, &e))
 900                         return error("unable to find %s", sha1_to_hex(sha1));
 901                 if (!packed_object_info(&e, type, sizep))
 902                         return 0;
 903                 /* sheesh */
 904                 map = unpack_entry(&e, type, sizep);
 905                 free(map);
 906                 return (map == NULL) ? 0 : -1;
 907         }
 908         if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
 909                 status = error("unable to unpack %s header",
 910                                sha1_to_hex(sha1));
 911         if (parse_sha1_header(hdr, type, &size) < 0)
 912                 status = error("unable to parse %s header", sha1_to_hex(sha1));
 913         else {
 914                 status = 0;
 915                 *sizep = size;
 916         }
 917         inflateEnd(&stream);
 918         munmap(map, mapsize);
 919         return status;
 920 }
 921
 922 static void *read_packed_sha1(const unsigned char *sha1, char *type, unsigned long *size)
 923 {
 924         struct pack_entry e;
 925
 926         if (!find_pack_entry(sha1, &e)) {
 927                 error("cannot read sha1_file for %s", sha1_to_hex(sha1));
 928                 return NULL;
 929         }
 930         return unpack_entry(&e, type, size);
 931 }
 932
 933 void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
 934 {
 935         unsigned long mapsize;
 936         void *map, *buf;
 937
 938         map = map_sha1_file_internal(sha1, &mapsize, 0);
 939         if (map) {
 940                 buf = unpack_sha1_file(map, mapsize, type, size);
 941                 munmap(map, mapsize);
 942                 return buf;
 943         }
 944         return read_packed_sha1(sha1, type, size);
 945 }
 946
 947 void *read_object_with_reference(const unsigned char *sha1,
 948                                  const char *required_type,
 949                                  unsigned long *size,
 950                                  unsigned char *actual_sha1_return)
 951 {
 952         char type[20];
 953         void *buffer;
 954         unsigned long isize;
 955         unsigned char actual_sha1[20];
 956
 957         memcpy(actual_sha1, sha1, 20);
 958         while (1) {
 959                 int ref_length = -1;
 960                 const char *ref_type = NULL;
 961
 962                 buffer = read_sha1_file(actual_sha1, type, &isize);
 963                 if (!buffer)
 964                         return NULL;
 965                 if (!strcmp(type, required_type)) {
 966                         *size = isize;
 967                         if (actual_sha1_return)
 968                                 memcpy(actual_sha1_return, actual_sha1, 20);
 969                         return buffer;
 970                 }
 971                 /* Handle references */
 972                 else if (!strcmp(type, "commit"))
 973                         ref_type = "tree ";
 974                 else if (!strcmp(type, "tag"))
 975                         ref_type = "object ";
 976                 else {
 977                         free(buffer);
 978                         return NULL;
 979                 }
 980                 ref_length = strlen(ref_type);
 981
 982                 if (memcmp(buffer, ref_type, ref_length) ||
 983                     get_sha1_hex(buffer + ref_length, actual_sha1)) {
 984                         free(buffer);
 985                         return NULL;
 986                 }
 987                 /* Now we have the ID of the referred-to object in
 988                  * actual_sha1.  Check again. */
 989         }
 990 }
 991
 992 static char *write_sha1_file_prepare(void *buf,
 993                                      unsigned long len,
 994                                      const char *type,
 995                                      unsigned char *sha1,
 996                                      unsigned char *hdr,
 997                                      int *hdrlen)
 998 {
 999         SHA_CTX c;
1000
1001         /* Generate the header */
1002         *hdrlen = sprintf((char *)hdr, "%s %lu", type, len)+1;
1003
1004         /* Sha1.. */
1005         SHA1_Init(&c);
1006         SHA1_Update(&c, hdr, *hdrlen);
1007         SHA1_Update(&c, buf, len);
1008         SHA1_Final(sha1, &c);
1009
1010         return sha1_file_name(sha1);
1011 }
1012
1013 int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1)
1014 {
1015         int size;
1016         unsigned char *compressed;
1017         z_stream stream;
1018         unsigned char sha1[20];
1019         char *filename;
1020         static char tmpfile[PATH_MAX];
1021         unsigned char hdr[50];
1022         int fd, hdrlen, ret;
1023
1024         /* Normally if we have it in the pack then we do not bother writing
1025          * it out into .git/objects/??/?{38} file.
1026          */
1027         filename = write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
1028         if (returnsha1)
1029                 memcpy(returnsha1, sha1, 20);
1030         if (has_sha1_file(sha1))
1031                 return 0;
1032         fd = open(filename, O_RDONLY);
1033         if (fd >= 0) {
1034                 /*
1035                  * FIXME!!! We might do collision checking here, but we'd
1036                  * need to uncompress the old file and check it. Later.
1037                  */
1038                 close(fd);
1039                 return 0;
1040         }
1041
1042         if (errno != ENOENT) {
1043                 fprintf(stderr, "sha1 file %s: %s", filename, strerror(errno));
1044                 return -1;
1045         }
1046
1047         snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
1048
1049         fd = mkstemp(tmpfile);
1050         if (fd < 0) {
1051                 fprintf(stderr, "unable to create temporary sha1 filename %s: %s", tmpfile, strerror(errno));
1052                 return -1;
1053         }
1054
1055         /* Set it up */
1056         memset(&stream, 0, sizeof(stream));
1057         deflateInit(&stream, Z_BEST_COMPRESSION);
1058         size = deflateBound(&stream, len+hdrlen);
1059         compressed = xmalloc(size);
1060
1061         /* Compress it */
1062         stream.next_out = compressed;
1063         stream.avail_out = size;
1064
1065         /* First header.. */
1066         stream.next_in = hdr;
1067         stream.avail_in = hdrlen;
1068         while (deflate(&stream, 0) == Z_OK)
1069                 /* nothing */;
1070
1071         /* Then the data itself.. */
1072         stream.next_in = buf;
1073         stream.avail_in = len;
1074         while (deflate(&stream, Z_FINISH) == Z_OK)
1075                 /* nothing */;
1076         deflateEnd(&stream);
1077         size = stream.total_out;
1078
1079         if (write(fd, compressed, size) != size)
1080                 die("unable to write file");
1081         fchmod(fd, 0444);
1082         close(fd);
1083         free(compressed);
1084
1085         ret = link(tmpfile, filename);
1086         if (ret < 0) {
1087                 ret = errno;
1088
1089                 /*
1090                  * Coda hack - coda doesn't like cross-directory links,
1091                  * so we fall back to a rename, which will mean that it
1092                  * won't be able to check collisions, but that's not a
1093                  * big deal.
1094                  *
1095                  * When this succeeds, we just return 0. We have nothing
1096                  * left to unlink.
1097                  */
1098                 if (ret == EXDEV && !rename(tmpfile, filename))
1099                         return 0;
1100         }
1101         unlink(tmpfile);
1102         if (ret) {
1103                 if (ret != EEXIST) {
1104                         fprintf(stderr, "unable to write sha1 filename %s: %s", filename, strerror(ret));
1105                         return -1;
1106                 }
1107                 /* FIXME!!! Collision check here ? */
1108         }
1109
1110         return 0;
1111 }
1112
1113 int write_sha1_from_fd(const unsigned char *sha1, int fd)
1114 {
1115         char *filename = sha1_file_name(sha1);
1116
1117         int local;
1118         z_stream stream;
1119         unsigned char real_sha1[20];
1120         unsigned char buf[4096];
1121         unsigned char discard[4096];
1122         int ret;
1123         SHA_CTX c;
1124
1125         local = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
1126
1127         if (local < 0)
1128                 return error("Couldn't open %s\n", filename);
1129
1130         memset(&stream, 0, sizeof(stream));
1131
1132         inflateInit(&stream);
1133
1134         SHA1_Init(&c);
1135
1136         do {
1137                 ssize_t size;
1138                 size = read(fd, buf, 4096);
1139                 if (size <= 0) {
1140                         close(local);
1141                         unlink(filename);
1142                         if (!size)
1143                                 return error("Connection closed?");
1144                         perror("Reading from connection");
1145                         return -1;
1146                 }
1147                 write(local, buf, size);
1148                 stream.avail_in = size;
1149                 stream.next_in = buf;
1150                 do {
1151                         stream.next_out = discard;
1152                         stream.avail_out = sizeof(discard);
1153                         ret = inflate(&stream, Z_SYNC_FLUSH);
1154                         SHA1_Update(&c, discard, sizeof(discard) -
1155                                     stream.avail_out);
1156                 } while (stream.avail_in && ret == Z_OK);
1157
1158         } while (ret == Z_OK);
1159         inflateEnd(&stream);
1160
1161         close(local);
1162         SHA1_Final(real_sha1, &c);
1163         if (ret != Z_STREAM_END) {
1164                 unlink(filename);
1165                 return error("File %s corrupted", sha1_to_hex(sha1));
1166         }
1167         if (memcmp(sha1, real_sha1, 20)) {
1168                 unlink(filename);
1169                 return error("File %s has bad hash\n", sha1_to_hex(sha1));
1170         }
1171
1172         return 0;
1173 }
1174
1175 int has_sha1_file(const unsigned char *sha1)
1176 {
1177         struct stat st;
1178         struct pack_entry e;
1179
1180         if (find_sha1_file(sha1, &st))
1181                 return 1;
1182         return find_pack_entry(sha1, &e);
1183 }
1184
1185 int index_fd(unsigned char *sha1, int fd, struct stat *st)
1186 {
1187         unsigned long size = st->st_size;
1188         void *buf;
1189         int ret;
1190
1191         buf = "";
1192         if (size)
1193                 buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
1194         close(fd);
1195         if ((int)(long)buf == -1)
1196                 return -1;
1197
1198         ret = write_sha1_file(buf, size, "blob", sha1);
1199         if (size)
1200                 munmap(buf, size);
1201         return ret;
1202 }