unpack-objects.c

   1 #include "cache.h"
   2 #include "object.h"
   3 #include "delta.h"
   4 #include "pack.h"
   5
   6 static int dry_run;
   7 static const char unpack_usage[] = "git-unpack-objects < pack-file";
   8
   9 /* We always read in 4kB chunks. */
  10 static unsigned char buffer[4096];
  11 static unsigned long offset, len, eof;
  12 static SHA_CTX ctx;
  13
  14 /*
  15  * Make sure at least "min" bytes are available in the buffer, and
  16  * return the pointer to the buffer.
  17  */
  18 static void * fill(int min)
  19 {
  20         if (min <= len)
  21                 return buffer + offset;
  22         if (eof)
  23                 die("unable to fill input");
  24         if (min > sizeof(buffer))
  25                 die("cannot fill %d bytes", min);
  26         if (offset) {
  27                 SHA1_Update(&ctx, buffer, offset);
  28                 memcpy(buffer, buffer + offset, len);
  29                 offset = 0;
  30         }
  31         do {
  32                 int ret = read(0, buffer + len, sizeof(buffer) - len);
  33                 if (ret <= 0) {
  34                         if (!ret)
  35                                 die("early EOF");
  36                         if (errno == EAGAIN || errno == EINTR)
  37                                 continue;
  38                         die("read error on input: %s", strerror(errno));
  39                 }
  40                 len += ret;
  41         } while (len < min);
  42         return buffer;
  43 }
  44
  45 static void use(int bytes)
  46 {
  47         if (bytes > len)
  48                 die("used more bytes than were available");
  49         len -= bytes;
  50         offset += bytes;
  51 }
  52
  53 static void *get_data(unsigned long size)
  54 {
  55         z_stream stream;
  56         void *buf = xmalloc(size);
  57
  58         if (!size)
  59                 return buf;
  60         memset(&stream, 0, sizeof(stream));
  61
  62         stream.next_out = buf;
  63         stream.avail_out = size;
  64         stream.next_in = fill(1);
  65         stream.avail_in = len;
  66         inflateInit(&stream);
  67
  68         for (;;) {
  69                 int ret = inflate(&stream, 0);
  70                 use(len - stream.avail_in);
  71                 if (stream.total_out == size && ret == Z_STREAM_END)
  72                         break;
  73                 if (ret != Z_OK)
  74                         die("inflate returned %d\n", ret);
  75                 stream.next_in = fill(1);
  76                 stream.avail_in = len;
  77         }
  78         return buf;
  79 }
  80
  81 struct delta_info {
  82         unsigned char base_sha1[20];
  83         unsigned long size;
  84         void *delta;
  85         struct delta_info *next;
  86 };
  87
  88 static struct delta_info *delta_list;
  89
  90 static void add_delta_to_list(unsigned char *base_sha1, void *delta, unsigned long size)
  91 {
  92         struct delta_info *info = xmalloc(sizeof(*info));
  93
  94         memcpy(info->base_sha1, base_sha1, 20);
  95         info->size = size;
  96         info->delta = delta;
  97         info->next = delta_list;
  98         delta_list = info;
  99 }
 100
 101 static void added_object(unsigned char *sha1, const char *type, void *data, unsigned long size);
 102
 103 static void write_object(void *buf, unsigned long size, const char *type)
 104 {
 105         unsigned char sha1[20];
 106         if (write_sha1_file(buf, size, type, sha1) < 0)
 107                 die("failed to write object");
 108         added_object(sha1, type, buf, size);
 109 }
 110
 111 static int resolve_delta(const char *type,
 112         void *base, unsigned long base_size,
 113         void *delta, unsigned long delta_size)
 114 {
 115         void *result;
 116         unsigned long result_size;
 117
 118         result = patch_delta(base, base_size,
 119                              delta, delta_size,
 120                              &result_size);
 121         if (!result)
 122                 die("failed to apply delta");
 123         free(delta);
 124         write_object(result, result_size, type);
 125         free(result);
 126         return 0;
 127 }
 128
 129 static void added_object(unsigned char *sha1, const char *type, void *data, unsigned long size)
 130 {
 131         struct delta_info **p = &delta_list;
 132         struct delta_info *info;
 133
 134         while ((info = *p) != NULL) {
 135                 if (!memcmp(info->base_sha1, sha1, 20)) {
 136                         *p = info->next;
 137                         p = &delta_list;
 138                         resolve_delta(type, data, size, info->delta, info->size);
 139                         free(info);
 140                         continue;
 141                 }
 142                 p = &info->next;
 143         }
 144 }
 145
 146 static int unpack_non_delta_entry(enum object_type kind, unsigned long size)
 147 {
 148         void *buf = get_data(size);
 149         const char *type;
 150
 151         switch (kind) {
 152         case OBJ_COMMIT: type = "commit"; break;
 153         case OBJ_TREE:   type = "tree"; break;
 154         case OBJ_BLOB:   type = "blob"; break;
 155         case OBJ_TAG:    type = "tag"; break;
 156         default: die("bad type %d", kind);
 157         }
 158         write_object(buf, size, type);
 159         free(buf);
 160         return 0;
 161 }
 162
 163 static int unpack_delta_entry(unsigned long delta_size)
 164 {
 165         void *delta_data, *base;
 166         unsigned long base_size;
 167         char type[20];
 168         unsigned char base_sha1[20];
 169
 170         memcpy(base_sha1, fill(20), 20);
 171         use(20);
 172
 173         delta_data = get_data(delta_size);
 174
 175         if (!has_sha1_file(base_sha1)) {
 176                 add_delta_to_list(base_sha1, delta_data, delta_size);
 177                 return 0;
 178         }
 179         base = read_sha1_file(base_sha1, type, &base_size);
 180         if (!base)
 181                 die("failed to read delta-pack base object %s", sha1_to_hex(base_sha1));
 182         return resolve_delta(type, base, base_size, delta_data, delta_size);
 183 }
 184
 185 static void unpack_one(void)
 186 {
 187         unsigned shift;
 188         unsigned char *pack, c;
 189         unsigned long size;
 190         enum object_type type;
 191
 192         pack = fill(1);
 193         c = *pack;
 194         use(1);
 195         type = (c >> 4) & 7;
 196         size = (c & 15);
 197         shift = 4;
 198         while (c & 0x80) {
 199                 pack = fill(1);
 200                 c = *pack++;
 201                 use(1);
 202                 size += (c & 0x7f) << shift;
 203                 shift += 7;
 204         }
 205         switch (type) {
 206         case OBJ_COMMIT:
 207         case OBJ_TREE:
 208         case OBJ_BLOB:
 209         case OBJ_TAG:
 210                 unpack_non_delta_entry(type, size);
 211                 return;
 212         case OBJ_DELTA:
 213                 unpack_delta_entry(size);
 214                 return;
 215         default:
 216                 die("bad object type %d", type);
 217         }
 218 }
 219
 220 /*
 221  * We unpack from the end, older files first. Now, usually
 222  * there are deltas etc, so we'll not actually write the
 223  * objects in that order, but we might as well try..
 224  */
 225 static void unpack_all(void)
 226 {
 227         int i;
 228         struct pack_header *hdr = fill(sizeof(struct pack_header));
 229         unsigned version = ntohl(hdr->hdr_version);
 230         unsigned nr_objects = ntohl(hdr->hdr_entries);
 231
 232         if (ntohl(hdr->hdr_signature) != PACK_SIGNATURE)
 233                 die("bad pack file");
 234         if (version != PACK_VERSION)
 235                 die("unable to handle pack file version %d", version);
 236         fprintf(stderr, "Unpacking %d objects\n", nr_objects);
 237
 238         use(sizeof(struct pack_header));
 239         for (i = 0; i < nr_objects; i++)
 240                 unpack_one();
 241         if (delta_list)
 242                 die("unresolved deltas left after unpacking");
 243 }
 244
 245 int main(int argc, char **argv)
 246 {
 247         int i;
 248         unsigned char sha1[20];
 249
 250         for (i = 1 ; i < argc; i++) {
 251                 const char *arg = argv[i];
 252
 253                 if (*arg == '-') {
 254                         if (!strcmp(arg, "-n")) {
 255                                 dry_run = 1;
 256                                 continue;
 257                         }
 258                         usage(unpack_usage);
 259                 }
 260
 261                 /* We don't take any non-flag arguments now.. Maybe some day */
 262                 usage(unpack_usage);
 263         }
 264         SHA1_Init(&ctx);
 265         unpack_all();
 266         SHA1_Update(&ctx, buffer, offset);
 267         SHA1_Final(sha1, &ctx);
 268         if (memcmp(fill(20), sha1, 20))
 269                 die("final sha1 did not match");
 270         use(20);
 271
 272         /* Write the last part of the buffer to stdout */
 273         while (len) {
 274                 int ret = write(1, buffer + offset, len);
 275                 if (!ret)
 276                         break;
 277                 if (ret < 0) {
 278                         if (errno == EAGAIN || errno == EINTR)
 279                                 continue;
 280                         break;
 281                 }
 282                 len -= ret;
 283                 offset += ret;
 284         }
 285
 286         /* All done */
 287         return 0;
 288 }