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 char *git_path(const char *fmt, ...)
 106 {
 107         static char pathname[PATH_MAX], *ret;
 108         va_list args;
 109         int len;
 110
 111         if (!git_dir)
 112                 setup_git_env();
 113         len = strlen(git_dir);
 114         if (len > PATH_MAX-100)
 115                 return "pad-path";
 116         memcpy(pathname, git_dir, len);
 117         if (len && git_dir[len-1] != '/')
 118                 pathname[len++] = '/';
 119         va_start(args, fmt);
 120         vsnprintf(pathname + len, sizeof(pathname) - len, fmt, args);
 121         va_end(args);
 122         ret = pathname;
 123
 124         /* Clean it up */
 125         if (!memcmp(pathname, "./", 2)) {
 126                 ret += 2;
 127                 while (*ret == '/')
 128                         ret++;
 129         }
 130         return ret;
 131 }
 132
 133 int safe_create_leading_directories(char *path)
 134 {
 135         char *pos = path;
 136
 137         while (pos) {
 138                 pos = strchr(pos, '/');
 139                 if (!pos)
 140                         break;
 141                 *pos = 0;
 142                 if (mkdir(path, 0777) < 0)
 143                         if (errno != EEXIST) {
 144                                 *pos = '/';
 145                                 return -1;
 146                         }
 147                 *pos++ = '/';
 148         }
 149         return 0;
 150 }
 151
 152 int get_sha1(const char *str, unsigned char *sha1)
 153 {
 154         static const char *prefix[] = {
 155                 "",
 156                 "refs",
 157                 "refs/tags",
 158                 "refs/heads",
 159                 "refs/snap",
 160                 NULL
 161         };
 162         const char **p;
 163
 164         if (!get_sha1_hex(str, sha1))
 165                 return 0;
 166
 167         for (p = prefix; *p; p++) {
 168                 char * pathname = git_path("%s/%s", *p, str);
 169                 if (!get_sha1_file(pathname, sha1))
 170                         return 0;
 171         }
 172
 173         return -1;
 174 }
 175
 176 char * sha1_to_hex(const unsigned char *sha1)
 177 {
 178         static char buffer[50];
 179         static const char hex[] = "0123456789abcdef";
 180         char *buf = buffer;
 181         int i;
 182
 183         for (i = 0; i < 20; i++) {
 184                 unsigned int val = *sha1++;
 185                 *buf++ = hex[val >> 4];
 186                 *buf++ = hex[val & 0xf];
 187         }
 188         return buffer;
 189 }
 190
 191 static void fill_sha1_path(char *pathbuf, const unsigned char *sha1)
 192 {
 193         int i;
 194         for (i = 0; i < 20; i++) {
 195                 static char hex[] = "0123456789abcdef";
 196                 unsigned int val = sha1[i];
 197                 char *pos = pathbuf + i*2 + (i > 0);
 198                 *pos++ = hex[val >> 4];
 199                 *pos = hex[val & 0xf];
 200         }
 201 }
 202
 203 /*
 204  * NOTE! This returns a statically allocated buffer, so you have to be
 205  * careful about using it. Do a "strdup()" if you need to save the
 206  * filename.
 207  *
 208  * Also note that this returns the location for creating.  Reading
 209  * SHA1 file can happen from any alternate directory listed in the
 210  * DB_ENVIRONMENT environment variable if it is not found in
 211  * the primary object database.
 212  */
 213 char *sha1_file_name(const unsigned char *sha1)
 214 {
 215         static char *name, *base;
 216
 217         if (!base) {
 218                 const char *sha1_file_directory = get_object_directory();
 219                 int len = strlen(sha1_file_directory);
 220                 base = xmalloc(len + 60);
 221                 memcpy(base, sha1_file_directory, len);
 222                 memset(base+len, 0, 60);
 223                 base[len] = '/';
 224                 base[len+3] = '/';
 225                 name = base + len + 1;
 226         }
 227         fill_sha1_path(name, sha1);
 228         return base;
 229 }
 230
 231 struct alternate_object_database *alt_odb;
 232
 233 /*
 234  * Prepare alternate object database registry.
 235  * alt_odb points at an array of struct alternate_object_database.
 236  * This array is terminated with an element that has both its base
 237  * and name set to NULL.  alt_odb[n] comes from n'th non-empty
 238  * element from colon separated ALTERNATE_DB_ENVIRONMENT environment
 239  * variable, and its base points at a statically allocated buffer
 240  * that contains "/the/directory/corresponding/to/.git/objects/...",
 241  * while its name points just after the slash at the end of
 242  * ".git/objects/" in the example above, and has enough space to hold
 243  * 40-byte hex SHA1, an extra slash for the first level indirection,
 244  * and the terminating NUL.
 245  * This function allocates the alt_odb array and all the strings
 246  * pointed by base fields of the array elements with one xmalloc();
 247  * the string pool immediately follows the array.
 248  */
 249 void prepare_alt_odb(void)
 250 {
 251         int pass, totlen, i;
 252         const char *cp, *last;
 253         char *op = NULL;
 254         const char *alt = gitenv(ALTERNATE_DB_ENVIRONMENT) ? : "";
 255
 256         if (alt_odb)
 257                 return;
 258         /* The first pass counts how large an area to allocate to
 259          * hold the entire alt_odb structure, including array of
 260          * structs and path buffers for them.  The second pass fills
 261          * the structure and prepares the path buffers for use by
 262          * fill_sha1_path().
 263          */
 264         for (totlen = pass = 0; pass < 2; pass++) {
 265                 last = alt;
 266                 i = 0;
 267                 do {
 268                         cp = strchr(last, ':') ? : last + strlen(last);
 269                         if (last != cp) {
 270                                 /* 43 = 40-byte + 2 '/' + terminating NUL */
 271                                 int pfxlen = cp - last;
 272                                 int entlen = pfxlen + 43;
 273                                 if (pass == 0)
 274                                         totlen += entlen;
 275                                 else {
 276                                         alt_odb[i].base = op;
 277                                         alt_odb[i].name = op + pfxlen + 1;
 278                                         memcpy(op, last, pfxlen);
 279                                         op[pfxlen] = op[pfxlen + 3] = '/';
 280                                         op[entlen-1] = 0;
 281                                         op += entlen;
 282                                 }
 283                                 i++;
 284                         }
 285                         while (*cp && *cp == ':')
 286                                 cp++;
 287                         last = cp;
 288                 } while (*cp);
 289                 if (pass)
 290                         break;
 291                 alt_odb = xmalloc(sizeof(*alt_odb) * (i + 1) + totlen);
 292                 alt_odb[i].base = alt_odb[i].name = NULL;
 293                 op = (char*)(&alt_odb[i+1]);
 294         }
 295 }
 296
 297 static char *find_sha1_file(const unsigned char *sha1, struct stat *st)
 298 {
 299         int i;
 300         char *name = sha1_file_name(sha1);
 301
 302         if (!stat(name, st))
 303                 return name;
 304         prepare_alt_odb();
 305         for (i = 0; (name = alt_odb[i].name) != NULL; i++) {
 306                 fill_sha1_path(name, sha1);
 307                 if (!stat(alt_odb[i].base, st))
 308                         return alt_odb[i].base;
 309         }
 310         return NULL;
 311 }
 312
 313 #define PACK_MAX_SZ (1<<26)
 314 static int pack_used_ctr;
 315 static unsigned long pack_mapped;
 316 struct packed_git *packed_git;
 317
 318 static int check_packed_git_idx(const char *path, unsigned long *idx_size_,
 319                                 void **idx_map_)
 320 {
 321         void *idx_map;
 322         unsigned int *index;
 323         unsigned long idx_size;
 324         int nr, i;
 325         int fd = open(path, O_RDONLY);
 326         struct stat st;
 327         if (fd < 0)
 328                 return -1;
 329         if (fstat(fd, &st)) {
 330                 close(fd);
 331                 return -1;
 332         }
 333         idx_size = st.st_size;
 334         idx_map = mmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
 335         close(fd);
 336         if (idx_map == MAP_FAILED)
 337                 return -1;
 338
 339         index = idx_map;
 340         *idx_map_ = idx_map;
 341         *idx_size_ = idx_size;
 342
 343         /* check index map */
 344         if (idx_size < 4*256 + 20 + 20)
 345                 return error("index file too small");
 346         nr = 0;
 347         for (i = 0; i < 256; i++) {
 348                 unsigned int n = ntohl(index[i]);
 349                 if (n < nr)
 350                         return error("non-monotonic index");
 351                 nr = n;
 352         }
 353
 354         /*
 355          * Total size:
 356          *  - 256 index entries 4 bytes each
 357          *  - 24-byte entries * nr (20-byte sha1 + 4-byte offset)
 358          *  - 20-byte SHA1 of the packfile
 359          *  - 20-byte SHA1 file checksum
 360          */
 361         if (idx_size != 4*256 + nr * 24 + 20 + 20)
 362                 return error("wrong index file size");
 363
 364         return 0;
 365 }
 366
 367 static int unuse_one_packed_git(void)
 368 {
 369         struct packed_git *p, *lru = NULL;
 370
 371         for (p = packed_git; p; p = p->next) {
 372                 if (p->pack_use_cnt || !p->pack_base)
 373                         continue;
 374                 if (!lru || p->pack_last_used < lru->pack_last_used)
 375                         lru = p;
 376         }
 377         if (!lru)
 378                 return 0;
 379         munmap(lru->pack_base, lru->pack_size);
 380         lru->pack_base = NULL;
 381         return 1;
 382 }
 383
 384 void unuse_packed_git(struct packed_git *p)
 385 {
 386         p->pack_use_cnt--;
 387 }
 388
 389 int use_packed_git(struct packed_git *p)
 390 {
 391         if (!p->pack_base) {
 392                 int fd;
 393                 struct stat st;
 394                 void *map;
 395
 396                 pack_mapped += p->pack_size;
 397                 while (PACK_MAX_SZ < pack_mapped && unuse_one_packed_git())
 398                         ; /* nothing */
 399                 fd = open(p->pack_name, O_RDONLY);
 400                 if (fd < 0)
 401                         die("packfile %s cannot be opened", p->pack_name);
 402                 if (fstat(fd, &st)) {
 403                         close(fd);
 404                         die("packfile %s cannot be opened", p->pack_name);
 405                 }
 406                 if (st.st_size != p->pack_size)
 407                         die("packfile %s size mismatch.", p->pack_name);
 408                 map = mmap(NULL, p->pack_size, PROT_READ, MAP_PRIVATE, fd, 0);
 409                 close(fd);
 410                 if (map == MAP_FAILED)
 411                         die("packfile %s cannot be mapped.", p->pack_name);
 412                 p->pack_base = map;
 413
 414                 /* Check if the pack file matches with the index file.
 415                  * this is cheap.
 416                  */
 417                 if (memcmp((char*)(p->index_base) + p->index_size - 40,
 418                            p->pack_base + p->pack_size - 20, 20))
 419                         die("packfile %s does not match index.", p->pack_name);
 420         }
 421         p->pack_last_used = pack_used_ctr++;
 422         p->pack_use_cnt++;
 423         return 0;
 424 }
 425
 426 struct packed_git *add_packed_git(char *path, int path_len)
 427 {
 428         struct stat st;
 429         struct packed_git *p;
 430         unsigned long idx_size;
 431         void *idx_map;
 432
 433         if (check_packed_git_idx(path, &idx_size, &idx_map))
 434                 return NULL;
 435
 436         /* do we have a corresponding .pack file? */
 437         strcpy(path + path_len - 4, ".pack");
 438         if (stat(path, &st) || !S_ISREG(st.st_mode)) {
 439                 munmap(idx_map, idx_size);
 440                 return NULL;
 441         }
 442         /* ok, it looks sane as far as we can check without
 443          * actually mapping the pack file.
 444          */
 445         p = xmalloc(sizeof(*p) + path_len + 2);
 446         strcpy(p->pack_name, path);
 447         p->index_size = idx_size;
 448         p->pack_size = st.st_size;
 449         p->index_base = idx_map;
 450         p->next = NULL;
 451         p->pack_base = NULL;
 452         p->pack_last_used = 0;
 453         p->pack_use_cnt = 0;
 454         return p;
 455 }
 456
 457 static void prepare_packed_git_one(char *objdir)
 458 {
 459         char path[PATH_MAX];
 460         int len;
 461         DIR *dir;
 462         struct dirent *de;
 463
 464         sprintf(path, "%s/pack", objdir);
 465         len = strlen(path);
 466         dir = opendir(path);
 467         if (!dir)
 468                 return;
 469         path[len++] = '/';
 470         while ((de = readdir(dir)) != NULL) {
 471                 int namelen = strlen(de->d_name);
 472                 struct packed_git *p;
 473
 474                 if (strcmp(de->d_name + namelen - 4, ".idx"))
 475                         continue;
 476
 477                 /* we have .idx.  Is it a file we can map? */
 478                 strcpy(path + len, de->d_name);
 479                 p = add_packed_git(path, len + namelen);
 480                 if (!p)
 481                         continue;
 482                 p->next = packed_git;
 483                 packed_git = p;
 484         }
 485         closedir(dir);
 486 }
 487
 488 void prepare_packed_git(void)
 489 {
 490         int i;
 491         static int run_once = 0;
 492
 493         if (run_once++)
 494                 return;
 495
 496         prepare_packed_git_one(get_object_directory());
 497         prepare_alt_odb();
 498         for (i = 0; alt_odb[i].base != NULL; i++) {
 499                 alt_odb[i].name[0] = 0;
 500                 prepare_packed_git_one(alt_odb[i].base);
 501         }
 502 }
 503
 504 int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type)
 505 {
 506         char header[100];
 507         unsigned char real_sha1[20];
 508         SHA_CTX c;
 509
 510         SHA1_Init(&c);
 511         SHA1_Update(&c, header, 1+sprintf(header, "%s %lu", type, size));
 512         SHA1_Update(&c, map, size);
 513         SHA1_Final(real_sha1, &c);
 514         return memcmp(sha1, real_sha1, 20) ? -1 : 0;
 515 }
 516
 517 static void *map_sha1_file_internal(const unsigned char *sha1,
 518                                     unsigned long *size,
 519                                     int say_error)
 520 {
 521         struct stat st;
 522         void *map;
 523         int fd;
 524         char *filename = find_sha1_file(sha1, &st);
 525
 526         if (!filename) {
 527                 if (say_error)
 528                         error("cannot map sha1 file %s", sha1_to_hex(sha1));
 529                 return NULL;
 530         }
 531
 532         fd = open(filename, O_RDONLY | sha1_file_open_flag);
 533         if (fd < 0) {
 534                 /* See if it works without O_NOATIME */
 535                 switch (sha1_file_open_flag) {
 536                 default:
 537                         fd = open(filename, O_RDONLY);
 538                         if (fd >= 0)
 539                                 break;
 540                 /* Fallthrough */
 541                 case 0:
 542                         if (say_error)
 543                                 perror(filename);
 544                         return NULL;
 545                 }
 546
 547                 /* If it failed once, it will probably fail again.
 548                  * Stop using O_NOATIME
 549                  */
 550                 sha1_file_open_flag = 0;
 551         }
 552         map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
 553         close(fd);
 554         if (-1 == (int)(long)map)
 555                 return NULL;
 556         *size = st.st_size;
 557         return map;
 558 }
 559
 560 void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
 561 {
 562         return map_sha1_file_internal(sha1, size, 1);
 563 }
 564
 565 int unpack_sha1_header(z_stream *stream, void *map, unsigned long mapsize, void *buffer, unsigned long size)
 566 {
 567         /* Get the data stream */
 568         memset(stream, 0, sizeof(*stream));
 569         stream->next_in = map;
 570         stream->avail_in = mapsize;
 571         stream->next_out = buffer;
 572         stream->avail_out = size;
 573
 574         inflateInit(stream);
 575         return inflate(stream, 0);
 576 }
 577
 578 static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size)
 579 {
 580         int bytes = strlen(buffer) + 1;
 581         unsigned char *buf = xmalloc(1+size);
 582
 583         memcpy(buf, buffer + bytes, stream->total_out - bytes);
 584         bytes = stream->total_out - bytes;
 585         if (bytes < size) {
 586                 stream->next_out = buf + bytes;
 587                 stream->avail_out = size - bytes;
 588                 while (inflate(stream, Z_FINISH) == Z_OK)
 589                         /* nothing */;
 590         }
 591         buf[size] = 0;
 592         inflateEnd(stream);
 593         return buf;
 594 }
 595
 596 /*
 597  * We used to just use "sscanf()", but that's actually way
 598  * too permissive for what we want to check. So do an anal
 599  * object header parse by hand.
 600  */
 601 int parse_sha1_header(char *hdr, char *type, unsigned long *sizep)
 602 {
 603         int i;
 604         unsigned long size;
 605
 606         /*
 607          * The type can be at most ten bytes (including the
 608          * terminating '\0' that we add), and is followed by
 609          * a space.
 610          */
 611         i = 10;
 612         for (;;) {
 613                 char c = *hdr++;
 614                 if (c == ' ')
 615                         break;
 616                 if (!--i)
 617                         return -1;
 618                 *type++ = c;
 619         }
 620         *type = 0;
 621
 622         /*
 623          * The length must follow immediately, and be in canonical
 624          * decimal format (ie "010" is not valid).
 625          */
 626         size = *hdr++ - '0';
 627         if (size > 9)
 628                 return -1;
 629         if (size) {
 630                 for (;;) {
 631                         unsigned long c = *hdr - '0';
 632                         if (c > 9)
 633                                 break;
 634                         hdr++;
 635                         size = size * 10 + c;
 636                 }
 637         }
 638         *sizep = size;
 639
 640         /*
 641          * The length must be followed by a zero byte
 642          */
 643         return *hdr ? -1 : 0;
 644 }
 645
 646 void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
 647 {
 648         int ret;
 649         z_stream stream;
 650         char hdr[8192];
 651
 652         ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr));
 653         if (ret < Z_OK || parse_sha1_header(hdr, type, size) < 0)
 654                 return NULL;
 655
 656         return unpack_sha1_rest(&stream, hdr, *size);
 657 }
 658
 659 /* forward declaration for a mutually recursive function */
 660 static int packed_object_info(struct pack_entry *entry,
 661                               char *type, unsigned long *sizep);
 662
 663 static int packed_delta_info(unsigned char *base_sha1,
 664                              unsigned long delta_size,
 665                              unsigned long left,
 666                              char *type,
 667                              unsigned long *sizep,
 668                              struct packed_git *p)
 669 {
 670         struct pack_entry base_ent;
 671
 672         if (left < 20)
 673                 die("truncated pack file");
 674
 675         /* The base entry _must_ be in the same pack */
 676         if (!find_pack_entry_one(base_sha1, &base_ent, p))
 677                 die("failed to find delta-pack base object %s",
 678                     sha1_to_hex(base_sha1));
 679
 680         /* We choose to only get the type of the base object and
 681          * ignore potentially corrupt pack file that expects the delta
 682          * based on a base with a wrong size.  This saves tons of
 683          * inflate() calls.
 684          */
 685
 686         if (packed_object_info(&base_ent, type, NULL))
 687                 die("cannot get info for delta-pack base");
 688
 689         if (sizep) {
 690                 const unsigned char *data;
 691                 unsigned char delta_head[64];
 692                 unsigned long result_size;
 693                 z_stream stream;
 694                 int st;
 695
 696                 memset(&stream, 0, sizeof(stream));
 697
 698                 data = stream.next_in = base_sha1 + 20;
 699                 stream.avail_in = left - 20;
 700                 stream.next_out = delta_head;
 701                 stream.avail_out = sizeof(delta_head);
 702
 703                 inflateInit(&stream);
 704                 st = inflate(&stream, Z_FINISH);
 705                 inflateEnd(&stream);
 706                 if ((st != Z_STREAM_END) &&
 707                     stream.total_out != sizeof(delta_head))
 708                         die("delta data unpack-initial failed");
 709
 710                 /* Examine the initial part of the delta to figure out
 711                  * the result size.
 712                  */
 713                 data = delta_head;
 714                 get_delta_hdr_size(&data); /* ignore base size */
 715
 716                 /* Read the result size */
 717                 result_size = get_delta_hdr_size(&data);
 718                 *sizep = result_size;
 719         }
 720         return 0;
 721 }
 722
 723 static unsigned long unpack_object_header(struct packed_git *p, unsigned long offset,
 724         enum object_type *type, unsigned long *sizep)
 725 {
 726         unsigned shift;
 727         unsigned char *pack, c;
 728         unsigned long size;
 729
 730         if (offset >= p->pack_size)
 731                 die("object offset outside of pack file");
 732
 733         pack =  p->pack_base + offset;
 734         c = *pack++;
 735         offset++;
 736         *type = (c >> 4) & 7;
 737         size = c & 15;
 738         shift = 4;
 739         while (c & 0x80) {
 740                 if (offset >= p->pack_size)
 741                         die("object offset outside of pack file");
 742                 c = *pack++;
 743                 offset++;
 744                 size += (c & 0x7f) << shift;
 745                 shift += 7;
 746         }
 747         *sizep = size;
 748         return offset;
 749 }
 750
 751 void packed_object_info_detail(struct pack_entry *e,
 752                                char *type,
 753                                unsigned long *size,
 754                                unsigned long *store_size,
 755                                int *delta_chain_length,
 756                                unsigned char *base_sha1)
 757 {
 758         struct packed_git *p = e->p;
 759         unsigned long offset, left;
 760         unsigned char *pack;
 761         enum object_type kind;
 762
 763         offset = unpack_object_header(p, e->offset, &kind, size);
 764         pack = p->pack_base + offset;
 765         left = p->pack_size - offset;
 766         if (kind != OBJ_DELTA)
 767                 *delta_chain_length = 0;
 768         else {
 769                 int chain_length = 0;
 770                 memcpy(base_sha1, pack, 20);
 771                 do {
 772                         struct pack_entry base_ent;
 773                         unsigned long junk;
 774
 775                         find_pack_entry_one(pack, &base_ent, p);
 776                         offset = unpack_object_header(p, base_ent.offset,
 777                                                       &kind, &junk);
 778                         pack = p->pack_base + offset;
 779                         chain_length++;
 780                 } while (kind == OBJ_DELTA);
 781                 *delta_chain_length = chain_length;
 782         }
 783         switch (kind) {
 784         case OBJ_COMMIT:
 785                 strcpy(type, "commit");
 786                 break;
 787         case OBJ_TREE:
 788                 strcpy(type, "tree");
 789                 break;
 790         case OBJ_BLOB:
 791                 strcpy(type, "blob");
 792                 break;
 793         case OBJ_TAG:
 794                 strcpy(type, "tag");
 795                 break;
 796         default:
 797                 die("corrupted pack file");
 798         }
 799         *store_size = 0; /* notyet */
 800 }
 801
 802 static int packed_object_info(struct pack_entry *entry,
 803                               char *type, unsigned long *sizep)
 804 {
 805         struct packed_git *p = entry->p;
 806         unsigned long offset, size, left;
 807         unsigned char *pack;
 808         enum object_type kind;
 809         int retval;
 810
 811         if (use_packed_git(p))
 812                 die("cannot map packed file");
 813
 814         offset = unpack_object_header(p, entry->offset, &kind, &size);
 815         pack = p->pack_base + offset;
 816         left = p->pack_size - offset;
 817
 818         switch (kind) {
 819         case OBJ_DELTA:
 820                 retval = packed_delta_info(pack, size, left, type, sizep, p);
 821                 unuse_packed_git(p);
 822                 return retval;
 823         case OBJ_COMMIT:
 824                 strcpy(type, "commit");
 825                 break;
 826         case OBJ_TREE:
 827                 strcpy(type, "tree");
 828                 break;
 829         case OBJ_BLOB:
 830                 strcpy(type, "blob");
 831                 break;
 832         case OBJ_TAG:
 833                 strcpy(type, "tag");
 834                 break;
 835         default:
 836                 die("corrupted pack file");
 837         }
 838         if (sizep)
 839                 *sizep = size;
 840         unuse_packed_git(p);
 841         return 0;
 842 }
 843
 844 /* forward declaration for a mutually recursive function */
 845 static void *unpack_entry(struct pack_entry *, char *, unsigned long *);
 846
 847 static void *unpack_delta_entry(unsigned char *base_sha1,
 848                                 unsigned long delta_size,
 849                                 unsigned long left,
 850                                 char *type,
 851                                 unsigned long *sizep,
 852                                 struct packed_git *p)
 853 {
 854         struct pack_entry base_ent;
 855         void *data, *delta_data, *result, *base;
 856         unsigned long data_size, result_size, base_size;
 857         z_stream stream;
 858         int st;
 859
 860         if (left < 20)
 861                 die("truncated pack file");
 862         data = base_sha1 + 20;
 863         data_size = left - 20;
 864         delta_data = xmalloc(delta_size);
 865
 866         memset(&stream, 0, sizeof(stream));
 867
 868         stream.next_in = data;
 869         stream.avail_in = data_size;
 870         stream.next_out = delta_data;
 871         stream.avail_out = delta_size;
 872
 873         inflateInit(&stream);
 874         st = inflate(&stream, Z_FINISH);
 875         inflateEnd(&stream);
 876         if ((st != Z_STREAM_END) || stream.total_out != delta_size)
 877                 die("delta data unpack failed");
 878
 879         /* The base entry _must_ be in the same pack */
 880         if (!find_pack_entry_one(base_sha1, &base_ent, p))
 881                 die("failed to find delta-pack base object %s",
 882                     sha1_to_hex(base_sha1));
 883         base = unpack_entry_gently(&base_ent, type, &base_size);
 884         if (!base)
 885                 die("failed to read delta-pack base object %s",
 886                     sha1_to_hex(base_sha1));
 887         result = patch_delta(base, base_size,
 888                              delta_data, delta_size,
 889                              &result_size);
 890         if (!result)
 891                 die("failed to apply delta");
 892         free(delta_data);
 893         free(base);
 894         *sizep = result_size;
 895         return result;
 896 }
 897
 898 static void *unpack_non_delta_entry(unsigned char *data,
 899                                     unsigned long size,
 900                                     unsigned long left)
 901 {
 902         int st;
 903         z_stream stream;
 904         unsigned char *buffer;
 905
 906         buffer = xmalloc(size + 1);
 907         buffer[size] = 0;
 908         memset(&stream, 0, sizeof(stream));
 909         stream.next_in = data;
 910         stream.avail_in = left;
 911         stream.next_out = buffer;
 912         stream.avail_out = size;
 913
 914         inflateInit(&stream);
 915         st = inflate(&stream, Z_FINISH);
 916         inflateEnd(&stream);
 917         if ((st != Z_STREAM_END) || stream.total_out != size) {
 918                 free(buffer);
 919                 return NULL;
 920         }
 921
 922         return buffer;
 923 }
 924
 925 static void *unpack_entry(struct pack_entry *entry,
 926                           char *type, unsigned long *sizep)
 927 {
 928         struct packed_git *p = entry->p;
 929         void *retval;
 930
 931         if (use_packed_git(p))
 932                 die("cannot map packed file");
 933         retval = unpack_entry_gently(entry, type, sizep);
 934         unuse_packed_git(p);
 935         if (!retval)
 936                 die("corrupted pack file");
 937         return retval;
 938 }
 939
 940 /* The caller is responsible for use_packed_git()/unuse_packed_git() pair */
 941 void *unpack_entry_gently(struct pack_entry *entry,
 942                           char *type, unsigned long *sizep)
 943 {
 944         struct packed_git *p = entry->p;
 945         unsigned long offset, size, left;
 946         unsigned char *pack;
 947         enum object_type kind;
 948         void *retval;
 949
 950         offset = unpack_object_header(p, entry->offset, &kind, &size);
 951         pack = p->pack_base + offset;
 952         left = p->pack_size - offset;
 953         switch (kind) {
 954         case OBJ_DELTA:
 955                 retval = unpack_delta_entry(pack, size, left, type, sizep, p);
 956                 return retval;
 957         case OBJ_COMMIT:
 958                 strcpy(type, "commit");
 959                 break;
 960         case OBJ_TREE:
 961                 strcpy(type, "tree");
 962                 break;
 963         case OBJ_BLOB:
 964                 strcpy(type, "blob");
 965                 break;
 966         case OBJ_TAG:
 967                 strcpy(type, "tag");
 968                 break;
 969         default:
 970                 return NULL;
 971         }
 972         *sizep = size;
 973         retval = unpack_non_delta_entry(pack, size, left);
 974         return retval;
 975 }
 976
 977 int num_packed_objects(const struct packed_git *p)
 978 {
 979         /* See check_packed_git_idx() */
 980         return (p->index_size - 20 - 20 - 4*256) / 24;
 981 }
 982
 983 int nth_packed_object_sha1(const struct packed_git *p, int n,
 984                            unsigned char* sha1)
 985 {
 986         void *index = p->index_base + 256;
 987         if (n < 0 || num_packed_objects(p) <= n)
 988                 return -1;
 989         memcpy(sha1, (index + 24 * n + 4), 20);
 990         return 0;
 991 }
 992
 993 int find_pack_entry_one(const unsigned char *sha1,
 994                         struct pack_entry *e, struct packed_git *p)
 995 {
 996         unsigned int *level1_ofs = p->index_base;
 997         int hi = ntohl(level1_ofs[*sha1]);
 998         int lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1]));
 999         void *index = p->index_base + 256;
1000
1001         do {
1002                 int mi = (lo + hi) / 2;
1003                 int cmp = memcmp(index + 24 * mi + 4, sha1, 20);
1004                 if (!cmp) {
1005                         e->offset = ntohl(*((int*)(index + 24 * mi)));
1006                         memcpy(e->sha1, sha1, 20);
1007                         e->p = p;
1008                         return 1;
1009                 }
1010                 if (cmp > 0)
1011                         hi = mi;
1012                 else
1013                         lo = mi+1;
1014         } while (lo < hi);
1015         return 0;
1016 }
1017
1018 static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e)
1019 {
1020         struct packed_git *p;
1021         prepare_packed_git();
1022
1023         for (p = packed_git; p; p = p->next) {
1024                 if (find_pack_entry_one(sha1, e, p))
1025                         return 1;
1026         }
1027         return 0;
1028 }
1029
1030 int sha1_object_info(const unsigned char *sha1, char *type, unsigned long *sizep)
1031 {
1032         int status;
1033         unsigned long mapsize, size;
1034         void *map;
1035         z_stream stream;
1036         char hdr[128];
1037
1038         map = map_sha1_file_internal(sha1, &mapsize, 0);
1039         if (!map) {
1040                 struct pack_entry e;
1041
1042                 if (!find_pack_entry(sha1, &e))
1043                         return error("unable to find %s", sha1_to_hex(sha1));
1044                 return packed_object_info(&e, type, sizep);
1045         }
1046         if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
1047                 status = error("unable to unpack %s header",
1048                                sha1_to_hex(sha1));
1049         if (parse_sha1_header(hdr, type, &size) < 0)
1050                 status = error("unable to parse %s header", sha1_to_hex(sha1));
1051         else {
1052                 status = 0;
1053                 if (sizep)
1054                         *sizep = size;
1055         }
1056         inflateEnd(&stream);
1057         munmap(map, mapsize);
1058         return status;
1059 }
1060
1061 static void *read_packed_sha1(const unsigned char *sha1, char *type, unsigned long *size)
1062 {
1063         struct pack_entry e;
1064
1065         if (!find_pack_entry(sha1, &e)) {
1066                 error("cannot read sha1_file for %s", sha1_to_hex(sha1));
1067                 return NULL;
1068         }
1069         return unpack_entry(&e, type, size);
1070 }
1071
1072 void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
1073 {
1074         unsigned long mapsize;
1075         void *map, *buf;
1076
1077         map = map_sha1_file_internal(sha1, &mapsize, 0);
1078         if (map) {
1079                 buf = unpack_sha1_file(map, mapsize, type, size);
1080                 munmap(map, mapsize);
1081                 return buf;
1082         }
1083         return read_packed_sha1(sha1, type, size);
1084 }
1085
1086 void *read_object_with_reference(const unsigned char *sha1,
1087                                  const char *required_type,
1088                                  unsigned long *size,
1089                                  unsigned char *actual_sha1_return)
1090 {
1091         char type[20];
1092         void *buffer;
1093         unsigned long isize;
1094         unsigned char actual_sha1[20];
1095
1096         memcpy(actual_sha1, sha1, 20);
1097         while (1) {
1098                 int ref_length = -1;
1099                 const char *ref_type = NULL;
1100
1101                 buffer = read_sha1_file(actual_sha1, type, &isize);
1102                 if (!buffer)
1103                         return NULL;
1104                 if (!strcmp(type, required_type)) {
1105                         *size = isize;
1106                         if (actual_sha1_return)
1107                                 memcpy(actual_sha1_return, actual_sha1, 20);
1108                         return buffer;
1109                 }
1110                 /* Handle references */
1111                 else if (!strcmp(type, "commit"))
1112                         ref_type = "tree ";
1113                 else if (!strcmp(type, "tag"))
1114                         ref_type = "object ";
1115                 else {
1116                         free(buffer);
1117                         return NULL;
1118                 }
1119                 ref_length = strlen(ref_type);
1120
1121                 if (memcmp(buffer, ref_type, ref_length) ||
1122                     get_sha1_hex(buffer + ref_length, actual_sha1)) {
1123                         free(buffer);
1124                         return NULL;
1125                 }
1126                 /* Now we have the ID of the referred-to object in
1127                  * actual_sha1.  Check again. */
1128         }
1129 }
1130
1131 static char *write_sha1_file_prepare(void *buf,
1132                                      unsigned long len,
1133                                      const char *type,
1134                                      unsigned char *sha1,
1135                                      unsigned char *hdr,
1136                                      int *hdrlen)
1137 {
1138         SHA_CTX c;
1139
1140         /* Generate the header */
1141         *hdrlen = sprintf((char *)hdr, "%s %lu", type, len)+1;
1142
1143         /* Sha1.. */
1144         SHA1_Init(&c);
1145         SHA1_Update(&c, hdr, *hdrlen);
1146         SHA1_Update(&c, buf, len);
1147         SHA1_Final(sha1, &c);
1148
1149         return sha1_file_name(sha1);
1150 }
1151
1152 int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1)
1153 {
1154         int size;
1155         unsigned char *compressed;
1156         z_stream stream;
1157         unsigned char sha1[20];
1158         char *filename;
1159         static char tmpfile[PATH_MAX];
1160         unsigned char hdr[50];
1161         int fd, hdrlen, ret;
1162
1163         /* Normally if we have it in the pack then we do not bother writing
1164          * it out into .git/objects/??/?{38} file.
1165          */
1166         filename = write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
1167         if (returnsha1)
1168                 memcpy(returnsha1, sha1, 20);
1169         if (has_sha1_file(sha1))
1170                 return 0;
1171         fd = open(filename, O_RDONLY);
1172         if (fd >= 0) {
1173                 /*
1174                  * FIXME!!! We might do collision checking here, but we'd
1175                  * need to uncompress the old file and check it. Later.
1176                  */
1177                 close(fd);
1178                 return 0;
1179         }
1180
1181         if (errno != ENOENT) {
1182                 fprintf(stderr, "sha1 file %s: %s", filename, strerror(errno));
1183                 return -1;
1184         }
1185
1186         snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
1187
1188         fd = mkstemp(tmpfile);
1189         if (fd < 0) {
1190                 fprintf(stderr, "unable to create temporary sha1 filename %s: %s", tmpfile, strerror(errno));
1191                 return -1;
1192         }
1193
1194         /* Set it up */
1195         memset(&stream, 0, sizeof(stream));
1196         deflateInit(&stream, Z_BEST_COMPRESSION);
1197         size = deflateBound(&stream, len+hdrlen);
1198         compressed = xmalloc(size);
1199
1200         /* Compress it */
1201         stream.next_out = compressed;
1202         stream.avail_out = size;
1203
1204         /* First header.. */
1205         stream.next_in = hdr;
1206         stream.avail_in = hdrlen;
1207         while (deflate(&stream, 0) == Z_OK)
1208                 /* nothing */;
1209
1210         /* Then the data itself.. */
1211         stream.next_in = buf;
1212         stream.avail_in = len;
1213         while (deflate(&stream, Z_FINISH) == Z_OK)
1214                 /* nothing */;
1215         deflateEnd(&stream);
1216         size = stream.total_out;
1217
1218         if (write(fd, compressed, size) != size)
1219                 die("unable to write file");
1220         fchmod(fd, 0444);
1221         close(fd);
1222         free(compressed);
1223
1224         ret = link(tmpfile, filename);
1225         if (ret < 0) {
1226                 ret = errno;
1227
1228                 /*
1229                  * Coda hack - coda doesn't like cross-directory links,
1230                  * so we fall back to a rename, which will mean that it
1231                  * won't be able to check collisions, but that's not a
1232                  * big deal.
1233                  *
1234                  * When this succeeds, we just return 0. We have nothing
1235                  * left to unlink.
1236                  */
1237                 if (ret == EXDEV && !rename(tmpfile, filename))
1238                         return 0;
1239         }
1240         unlink(tmpfile);
1241         if (ret) {
1242                 if (ret != EEXIST) {
1243                         fprintf(stderr, "unable to write sha1 filename %s: %s", filename, strerror(ret));
1244                         return -1;
1245                 }
1246                 /* FIXME!!! Collision check here ? */
1247         }
1248
1249         return 0;
1250 }
1251
1252 int write_sha1_from_fd(const unsigned char *sha1, int fd)
1253 {
1254         char *filename = sha1_file_name(sha1);
1255
1256         int local;
1257         z_stream stream;
1258         unsigned char real_sha1[20];
1259         unsigned char buf[4096];
1260         unsigned char discard[4096];
1261         int ret;
1262         SHA_CTX c;
1263
1264         local = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
1265
1266         if (local < 0)
1267                 return error("Couldn't open %s\n", filename);
1268
1269         memset(&stream, 0, sizeof(stream));
1270
1271         inflateInit(&stream);
1272
1273         SHA1_Init(&c);
1274
1275         do {
1276                 ssize_t size;
1277                 size = read(fd, buf, 4096);
1278                 if (size <= 0) {
1279                         close(local);
1280                         unlink(filename);
1281                         if (!size)
1282                                 return error("Connection closed?");
1283                         perror("Reading from connection");
1284                         return -1;
1285                 }
1286                 write(local, buf, size);
1287                 stream.avail_in = size;
1288                 stream.next_in = buf;
1289                 do {
1290                         stream.next_out = discard;
1291                         stream.avail_out = sizeof(discard);
1292                         ret = inflate(&stream, Z_SYNC_FLUSH);
1293                         SHA1_Update(&c, discard, sizeof(discard) -
1294                                     stream.avail_out);
1295                 } while (stream.avail_in && ret == Z_OK);
1296
1297         } while (ret == Z_OK);
1298         inflateEnd(&stream);
1299
1300         close(local);
1301         SHA1_Final(real_sha1, &c);
1302         if (ret != Z_STREAM_END) {
1303                 unlink(filename);
1304                 return error("File %s corrupted", sha1_to_hex(sha1));
1305         }
1306         if (memcmp(sha1, real_sha1, 20)) {
1307                 unlink(filename);
1308                 return error("File %s has bad hash\n", sha1_to_hex(sha1));
1309         }
1310
1311         return 0;
1312 }
1313
1314 int has_sha1_pack(const unsigned char *sha1)
1315 {
1316         struct pack_entry e;
1317         return find_pack_entry(sha1, &e);
1318 }
1319
1320 int has_sha1_file(const unsigned char *sha1)
1321 {
1322         struct stat st;
1323         struct pack_entry e;
1324
1325         if (find_sha1_file(sha1, &st))
1326                 return 1;
1327         return find_pack_entry(sha1, &e);
1328 }
1329
1330 int index_fd(unsigned char *sha1, int fd, struct stat *st)
1331 {
1332         unsigned long size = st->st_size;
1333         void *buf;
1334         int ret;
1335
1336         buf = "";
1337         if (size)
1338                 buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
1339         close(fd);
1340         if ((int)(long)buf == -1)
1341                 return -1;
1342
1343         ret = write_sha1_file(buf, size, "blob", sha1);
1344         if (size)
1345                 munmap(buf, size);
1346         return ret;
1347 }