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         if (!dry_run)
 159                 write_object(buf, size, type);
 160         free(buf);
 161         return 0;
 162 }
 163
 164 static int unpack_delta_entry(unsigned long delta_size)
 165 {
 166         void *delta_data, *base;
 167         unsigned long base_size;
 168         char type[20];
 169         unsigned char base_sha1[20];
 170
 171         memcpy(base_sha1, fill(20), 20);
 172         use(20);
 173
 174         delta_data = get_data(delta_size);
 175         if (dry_run) {
 176                 free(delta_data);
 177                 return 0;
 178         }
 179
 180         if (!has_sha1_file(base_sha1)) {
 181                 add_delta_to_list(base_sha1, delta_data, delta_size);
 182                 return 0;
 183         }
 184         base = read_sha1_file(base_sha1, type, &base_size);
 185         if (!base)
 186                 die("failed to read delta-pack base object %s", sha1_to_hex(base_sha1));
 187         return resolve_delta(type, base, base_size, delta_data, delta_size);
 188 }
 189
 190 static void unpack_one(void)
 191 {
 192         unsigned shift;
 193         unsigned char *pack, c;
 194         unsigned long size;
 195         enum object_type type;
 196
 197         pack = fill(1);
 198         c = *pack;
 199         use(1);
 200         type = (c >> 4) & 7;
 201         size = (c & 15);
 202         shift = 4;
 203         while (c & 0x80) {
 204                 pack = fill(1);
 205                 c = *pack++;
 206                 use(1);
 207                 size += (c & 0x7f) << shift;
 208                 shift += 7;
 209         }
 210         switch (type) {
 211         case OBJ_COMMIT:
 212         case OBJ_TREE:
 213         case OBJ_BLOB:
 214         case OBJ_TAG:
 215                 unpack_non_delta_entry(type, size);
 216                 return;
 217         case OBJ_DELTA:
 218                 unpack_delta_entry(size);
 219                 return;
 220         default:
 221                 die("bad object type %d", type);
 222         }
 223 }
 224
 225 /*
 226  * We unpack from the end, older files first. Now, usually
 227  * there are deltas etc, so we'll not actually write the
 228  * objects in that order, but we might as well try..
 229  */
 230 static void unpack_all(void)
 231 {
 232         int i;
 233         struct pack_header *hdr = fill(sizeof(struct pack_header));
 234         unsigned version = ntohl(hdr->hdr_version);
 235         unsigned nr_objects = ntohl(hdr->hdr_entries);
 236
 237         if (ntohl(hdr->hdr_signature) != PACK_SIGNATURE)
 238                 die("bad pack file");
 239         if (version != PACK_VERSION)
 240                 die("unable to handle pack file version %d", version);
 241         fprintf(stderr, "Unpacking %d objects\n", nr_objects);
 242
 243         use(sizeof(struct pack_header));
 244         for (i = 0; i < nr_objects; i++)
 245                 unpack_one();
 246         if (delta_list)
 247                 die("unresolved deltas left after unpacking");
 248 }
 249
 250 int main(int argc, char **argv)
 251 {
 252         int i;
 253         unsigned char sha1[20];
 254
 255         for (i = 1 ; i < argc; i++) {
 256                 const char *arg = argv[i];
 257
 258                 if (*arg == '-') {
 259                         if (!strcmp(arg, "-n")) {
 260                                 dry_run = 1;
 261                                 continue;
 262                         }
 263                         usage(unpack_usage);
 264                 }
 265
 266                 /* We don't take any non-flag arguments now.. Maybe some day */
 267                 usage(unpack_usage);
 268         }
 269         SHA1_Init(&ctx);
 270         unpack_all();
 271         SHA1_Update(&ctx, buffer, offset);
 272         SHA1_Final(sha1, &ctx);
 273         if (memcmp(fill(20), sha1, 20))
 274                 die("final sha1 did not match");
 275         use(20);
 276
 277         /* Write the last part of the buffer to stdout */
 278         while (len) {
 279                 int ret = write(1, buffer + offset, len);
 280                 if (!ret)
 281                         break;
 282                 if (ret < 0) {
 283                         if (errno == EAGAIN || errno == EINTR)
 284                                 continue;
 285                         break;
 286                 }
 287                 len -= ret;
 288                 offset += ret;
 289         }
 290
 291         /* All done */
 292         return 0;
 293 }