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 [-n] [-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 = xread(0, buffer + len, sizeof(buffer) - len);
  35                 if (ret <= 0) {
  36                         if (!ret)
  37                                 die("early EOF");
  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         memset(&stream, 0, sizeof(stream));
  59
  60         stream.next_out = buf;
  61         stream.avail_out = size;
  62         stream.next_in = fill(1);
  63         stream.avail_in = len;
  64         inflateInit(&stream);
  65
  66         for (;;) {
  67                 int ret = inflate(&stream, 0);
  68                 use(len - stream.avail_in);
  69                 if (stream.total_out == size && ret == Z_STREAM_END)
  70                         break;
  71                 if (ret != Z_OK)
  72                         die("inflate returned %d\n", ret);
  73                 stream.next_in = fill(1);
  74                 stream.avail_in = len;
  75         }
  76         inflateEnd(&stream);
  77         return buf;
  78 }
  79
  80 struct delta_info {
  81         unsigned char base_sha1[20];
  82         unsigned long size;
  83         void *delta;
  84         struct delta_info *next;
  85 };
  86
  87 static struct delta_info *delta_list;
  88
  89 static void add_delta_to_list(unsigned char *base_sha1, void *delta, unsigned long size)
  90 {
  91         struct delta_info *info = xmalloc(sizeof(*info));
  92
  93         memcpy(info->base_sha1, base_sha1, 20);
  94         info->size = size;
  95         info->delta = delta;
  96         info->next = delta_list;
  97         delta_list = info;
  98 }
  99
 100 static void added_object(unsigned char *sha1, const char *type, void *data, unsigned long size);
 101
 102 static void write_object(void *buf, unsigned long size, const char *type)
 103 {
 104         unsigned char sha1[20];
 105         if (write_sha1_file(buf, size, type, sha1) < 0)
 106                 die("failed to write object");
 107         added_object(sha1, type, buf, size);
 108 }
 109
 110 static int resolve_delta(const char *type,
 111         void *base, unsigned long base_size,
 112         void *delta, unsigned long delta_size)
 113 {
 114         void *result;
 115         unsigned long result_size;
 116
 117         result = patch_delta(base, base_size,
 118                              delta, delta_size,
 119                              &result_size);
 120         if (!result)
 121                 die("failed to apply delta");
 122         free(delta);
 123         write_object(result, result_size, type);
 124         free(result);
 125         return 0;
 126 }
 127
 128 static void added_object(unsigned char *sha1, const char *type, void *data, unsigned long size)
 129 {
 130         struct delta_info **p = &delta_list;
 131         struct delta_info *info;
 132
 133         while ((info = *p) != NULL) {
 134                 if (!memcmp(info->base_sha1, sha1, 20)) {
 135                         *p = info->next;
 136                         p = &delta_list;
 137                         resolve_delta(type, data, size, info->delta, info->size);
 138                         free(info);
 139                         continue;
 140                 }
 141                 p = &info->next;
 142         }
 143 }
 144
 145 static int unpack_non_delta_entry(enum object_type kind, unsigned long size)
 146 {
 147         void *buf = get_data(size);
 148         const char *type;
 149
 150         switch (kind) {
 151         case OBJ_COMMIT: type = "commit"; break;
 152         case OBJ_TREE:   type = "tree"; break;
 153         case OBJ_BLOB:   type = "blob"; break;
 154         case OBJ_TAG:    type = "tag"; break;
 155         default: die("bad type %d", kind);
 156         }
 157         if (!dry_run)
 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         int result;
 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         result = resolve_delta(type, base, base_size, delta_data, delta_size);
 188         free(base);
 189         return result;
 190 }
 191
 192 static void unpack_one(unsigned nr, unsigned total)
 193 {
 194         unsigned shift;
 195         unsigned char *pack, c;
 196         unsigned long size;
 197         enum object_type type;
 198
 199         pack = fill(1);
 200         c = *pack;
 201         use(1);
 202         type = (c >> 4) & 7;
 203         size = (c & 15);
 204         shift = 4;
 205         while (c & 0x80) {
 206                 pack = fill(1);
 207                 c = *pack++;
 208                 use(1);
 209                 size += (c & 0x7f) << shift;
 210                 shift += 7;
 211         }
 212         if (!quiet) {
 213                 static unsigned long last_sec;
 214                 static unsigned last_percent;
 215                 struct timeval now;
 216                 unsigned percentage = (nr * 100) / total;
 217
 218                 gettimeofday(&now, NULL);
 219                 if (percentage != last_percent || now.tv_sec != last_sec) {
 220                         last_sec = now.tv_sec;
 221                         last_percent = percentage;
 222                         fprintf(stderr, "%4u%% (%u/%u) done\r", percentage, nr, total);
 223                 }
 224         }
 225         switch (type) {
 226         case OBJ_COMMIT:
 227         case OBJ_TREE:
 228         case OBJ_BLOB:
 229         case OBJ_TAG:
 230                 unpack_non_delta_entry(type, size);
 231                 return;
 232         case OBJ_DELTA:
 233                 unpack_delta_entry(size);
 234                 return;
 235         default:
 236                 die("bad object type %d", type);
 237         }
 238 }
 239
 240 /*
 241  * We unpack from the end, older files first. Now, usually
 242  * there are deltas etc, so we'll not actually write the
 243  * objects in that order, but we might as well try..
 244  */
 245 static void unpack_all(void)
 246 {
 247         int i;
 248         struct pack_header *hdr = fill(sizeof(struct pack_header));
 249         unsigned version = ntohl(hdr->hdr_version);
 250         unsigned nr_objects = ntohl(hdr->hdr_entries);
 251
 252         if (ntohl(hdr->hdr_signature) != PACK_SIGNATURE)
 253                 die("bad pack file");
 254         if (version != PACK_VERSION)
 255                 die("unable to handle pack file version %d", version);
 256         fprintf(stderr, "Unpacking %d objects\n", nr_objects);
 257
 258         use(sizeof(struct pack_header));
 259         for (i = 0; i < nr_objects; i++)
 260                 unpack_one(i+1, nr_objects);
 261         if (delta_list)
 262                 die("unresolved deltas left after unpacking");
 263 }
 264
 265 int main(int argc, char **argv)
 266 {
 267         int i;
 268         unsigned char sha1[20];
 269
 270         setup_git_directory();
 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 = xwrite(1, buffer + offset, len);
 301                 if (ret <= 0)
 302                         break;
 303                 len -= ret;
 304                 offset += ret;
 305         }
 306
 307         /* All done */
 308         if (!quiet)
 309                 fprintf(stderr, "\n");
 310         return 0;
 311 }