unpack-objects.c

   1 #include "cache.h"
   2 #include "object.h"
   3 #include "delta.h"
   4 #include "pack.h"
   5
   6 #include <sys/time.h>
   7
   8 static int dry_run, quiet;
   9 static const char unpack_usage[] = "git-unpack-objects [-q] < pack-file";
  10
  11 /* We always read in 4kB chunks. */
  12 static unsigned char buffer[4096];
  13 static unsigned long offset, len, eof;
  14 static SHA_CTX ctx;
  15
  16 /*
  17  * Make sure at least "min" bytes are available in the buffer, and
  18  * return the pointer to the buffer.
  19  */
  20 static void * fill(int min)
  21 {
  22         if (min <= len)
  23                 return buffer + offset;
  24         if (eof)
  25                 die("unable to fill input");
  26         if (min > sizeof(buffer))
  27                 die("cannot fill %d bytes", min);
  28         if (offset) {
  29                 SHA1_Update(&ctx, buffer, offset);
  30                 memcpy(buffer, buffer + offset, len);
  31                 offset = 0;
  32         }
  33         do {
  34                 int ret = read(0, buffer + len, sizeof(buffer) - len);
  35                 if (ret <= 0) {
  36                         if (!ret)
  37                                 die("early EOF");
  38                         if (errno == EAGAIN || errno == EINTR)
  39                                 continue;
  40                         die("read error on input: %s", strerror(errno));
  41                 }
  42                 len += ret;
  43         } while (len < min);
  44         return buffer;
  45 }
  46
  47 static void use(int bytes)
  48 {
  49         if (bytes > len)
  50                 die("used more bytes than were available");
  51         len -= bytes;
  52         offset += bytes;
  53 }
  54
  55 static void *get_data(unsigned long size)
  56 {
  57         z_stream stream;
  58         void *buf = xmalloc(size);
  59
  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         inflateEnd(&stream);
  79         return buf;
  80 }
  81
  82 struct delta_info {
  83         unsigned char base_sha1[20];
  84         unsigned long size;
  85         void *delta;
  86         struct delta_info *next;
  87 };
  88
  89 static struct delta_info *delta_list;
  90
  91 static void add_delta_to_list(unsigned char *base_sha1, void *delta, unsigned long size)
  92 {
  93         struct delta_info *info = xmalloc(sizeof(*info));
  94
  95         memcpy(info->base_sha1, base_sha1, 20);
  96         info->size = size;
  97         info->delta = delta;
  98         info->next = delta_list;
  99         delta_list = info;
 100 }
 101
 102 static void added_object(unsigned char *sha1, const char *type, void *data, unsigned long size);
 103
 104 static void write_object(void *buf, unsigned long size, const char *type)
 105 {
 106         unsigned char sha1[20];
 107         if (write_sha1_file(buf, size, type, sha1) < 0)
 108                 die("failed to write object");
 109         added_object(sha1, type, buf, size);
 110 }
 111
 112 static int resolve_delta(const char *type,
 113         void *base, unsigned long base_size,
 114         void *delta, unsigned long delta_size)
 115 {
 116         void *result;
 117         unsigned long result_size;
 118
 119         result = patch_delta(base, base_size,
 120                              delta, delta_size,
 121                              &result_size);
 122         if (!result)
 123                 die("failed to apply delta");
 124         free(delta);
 125         write_object(result, result_size, type);
 126         free(result);
 127         return 0;
 128 }
 129
 130 static void added_object(unsigned char *sha1, const char *type, void *data, unsigned long size)
 131 {
 132         struct delta_info **p = &delta_list;
 133         struct delta_info *info;
 134
 135         while ((info = *p) != NULL) {
 136                 if (!memcmp(info->base_sha1, sha1, 20)) {
 137                         *p = info->next;
 138                         p = &delta_list;
 139                         resolve_delta(type, data, size, info->delta, info->size);
 140                         free(info);
 141                         continue;
 142                 }
 143                 p = &info->next;
 144         }
 145 }
 146
 147 static int unpack_non_delta_entry(enum object_type kind, unsigned long size)
 148 {
 149         void *buf = get_data(size);
 150         const char *type;
 151
 152         switch (kind) {
 153         case OBJ_COMMIT: type = "commit"; break;
 154         case OBJ_TREE:   type = "tree"; break;
 155         case OBJ_BLOB:   type = "blob"; break;
 156         case OBJ_TAG:    type = "tag"; break;
 157         default: die("bad type %d", kind);
 158         }
 159         if (!dry_run)
 160                 write_object(buf, size, type);
 161         free(buf);
 162         return 0;
 163 }
 164
 165 static int unpack_delta_entry(unsigned long delta_size)
 166 {
 167         void *delta_data, *base;
 168         unsigned long base_size;
 169         char type[20];
 170         unsigned char base_sha1[20];
 171         int result;
 172
 173         memcpy(base_sha1, fill(20), 20);
 174         use(20);
 175
 176         delta_data = get_data(delta_size);
 177         if (dry_run) {
 178                 free(delta_data);
 179                 return 0;
 180         }
 181
 182         if (!has_sha1_file(base_sha1)) {
 183                 add_delta_to_list(base_sha1, delta_data, delta_size);
 184                 return 0;
 185         }
 186         base = read_sha1_file(base_sha1, type, &base_size);
 187         if (!base)
 188                 die("failed to read delta-pack base object %s", sha1_to_hex(base_sha1));
 189         result = resolve_delta(type, base, base_size, delta_data, delta_size);
 190         free(base);
 191         return result;
 192 }
 193
 194 static void unpack_one(unsigned nr, unsigned total)
 195 {
 196         unsigned shift;
 197         unsigned char *pack, c;
 198         unsigned long size;
 199         enum object_type type;
 200
 201         pack = fill(1);
 202         c = *pack;
 203         use(1);
 204         type = (c >> 4) & 7;
 205         size = (c & 15);
 206         shift = 4;
 207         while (c & 0x80) {
 208                 pack = fill(1);
 209                 c = *pack++;
 210                 use(1);
 211                 size += (c & 0x7f) << shift;
 212                 shift += 7;
 213         }
 214         if (!quiet) {
 215                 static unsigned long last_sec;
 216                 static unsigned last_percent;
 217                 struct timeval now;
 218                 unsigned percentage = (nr * 100) / total;
 219
 220                 gettimeofday(&now, NULL);
 221                 if (percentage != last_percent || now.tv_sec != last_sec) {
 222                         last_sec = now.tv_sec;
 223                         last_percent = percentage;
 224                         fprintf(stderr, "%4u%% (%u/%u) done\r", percentage, nr, total);
 225                 }
 226         }
 227         switch (type) {
 228         case OBJ_COMMIT:
 229         case OBJ_TREE:
 230         case OBJ_BLOB:
 231         case OBJ_TAG:
 232                 unpack_non_delta_entry(type, size);
 233                 return;
 234         case OBJ_DELTA:
 235                 unpack_delta_entry(size);
 236                 return;
 237         default:
 238                 die("bad object type %d", type);
 239         }
 240 }
 241
 242 /*
 243  * We unpack from the end, older files first. Now, usually
 244  * there are deltas etc, so we'll not actually write the
 245  * objects in that order, but we might as well try..
 246  */
 247 static void unpack_all(void)
 248 {
 249         int i;
 250         struct pack_header *hdr = fill(sizeof(struct pack_header));
 251         unsigned version = ntohl(hdr->hdr_version);
 252         unsigned nr_objects = ntohl(hdr->hdr_entries);
 253
 254         if (ntohl(hdr->hdr_signature) != PACK_SIGNATURE)
 255                 die("bad pack file");
 256         if (version != PACK_VERSION)
 257                 die("unable to handle pack file version %d", version);
 258         fprintf(stderr, "Unpacking %d objects\n", nr_objects);
 259
 260         use(sizeof(struct pack_header));
 261         for (i = 0; i < nr_objects; i++)
 262                 unpack_one(i+1, nr_objects);
 263         if (delta_list)
 264                 die("unresolved deltas left after unpacking");
 265 }
 266
 267 int main(int argc, char **argv)
 268 {
 269         int i;
 270         unsigned char sha1[20];
 271
 272         for (i = 1 ; i < argc; i++) {
 273                 const char *arg = argv[i];
 274
 275                 if (*arg == '-') {
 276                         if (!strcmp(arg, "-n")) {
 277                                 dry_run = 1;
 278                                 continue;
 279                         }
 280                         if (!strcmp(arg, "-q")) {
 281                                 quiet = 1;
 282                                 continue;
 283                         }
 284                         usage(unpack_usage);
 285                 }
 286
 287                 /* We don't take any non-flag arguments now.. Maybe some day */
 288                 usage(unpack_usage);
 289         }
 290         SHA1_Init(&ctx);
 291         unpack_all();
 292         SHA1_Update(&ctx, buffer, offset);
 293         SHA1_Final(sha1, &ctx);
 294         if (memcmp(fill(20), sha1, 20))
 295                 die("final sha1 did not match");
 296         use(20);
 297
 298         /* Write the last part of the buffer to stdout */
 299         while (len) {
 300                 int ret = write(1, buffer + offset, len);
 301                 if (!ret)
 302                         break;
 303                 if (ret < 0) {
 304                         if (errno == EAGAIN || errno == EINTR)
 305                                 continue;
 306                         break;
 307                 }
 308                 len -= ret;
 309                 offset += ret;
 310         }
 311
 312         /* All done */
 313         if (!quiet)
 314                 fprintf(stderr, "\n");
 315         return 0;
 316 }