scripts/xsparse.c

   1 /* xsparse - expands compressed sparse file images extracted from GNU tar
   2    archives.
   3
   4    Copyright (C) 2006, 2007 Free Software Foundation, Inc.
   5
   6    Written by Sergey Poznyakoff
   7
   8    This program is free software; you can redistribute it and/or modify it
   9    under the terms of the GNU General Public License as published by the
  10    Free Software Foundation; either version 3, or (at your option) any later
  11    version.
  12
  13    This program is distributed in the hope that it will be useful, but
  14    WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General
  16    Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License along
  19    with this program; if not, write to the Free Software Foundation, Inc.,
  20    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  21
  22 #include <stdlib.h>
  23 #include <stdio.h>
  24 #include <stdarg.h>
  25 #include <unistd.h>
  26 #include <fcntl.h>
  27 #include <sys/stat.h>
  28 #include <limits.h>
  29 #include <errno.h>
  30
  31 /* Bound on length of the string representing an off_t.
  32    See INT_STRLEN_BOUND in intprops.h for explanation */
  33 #define OFF_T_STRLEN_BOUND ((sizeof (off_t) * CHAR_BIT) * 146 / 485 + 1)
  34 #define OFF_T_STRSIZE_BOUND (OFF_T_STRLEN_BOUND+1)
  35
  36 #define BLOCKSIZE 512
  37
  38 struct sp_array
  39 {
  40   off_t offset;
  41   size_t numbytes;
  42 };
  43
  44 char *progname;
  45 int verbose;
  46
  47 void
  48 die (int code, char *fmt, ...)
  49 {
  50   va_list ap;
  51
  52   fprintf (stderr, "%s: ", progname);
  53   va_start (ap, fmt);
  54   vfprintf (stderr, fmt, ap);
  55   va_end (ap);
  56   fprintf (stderr, "\n");
  57   exit (code);
  58 }
  59
  60 void *
  61 emalloc (size_t size)
  62 {
  63   char *p = malloc (size);
  64   if (!p)
  65     die (1, "not enough memory");
  66   return p;
  67 }
  68
  69 off_t
  70 string_to_off (char *p, char **endp)
  71 {
  72   off_t v = 0;
  73
  74   for (; *p; p++)
  75     {
  76       int digit = *p - '0';
  77       off_t x = v * 10;
  78       if (9 < (unsigned) digit)
  79         {
  80           if (endp)
  81             {
  82               *endp = p;
  83               break;
  84             }
  85           die (1, "number parse error near %s", p);
  86         }
  87       else if (x / 10 != v)
  88         die (1, "number out of allowed range, near %s", p);
  89       v = x + digit;
  90       if (v < 0)
  91         die (1, "negative number");
  92     }
  93   if (endp)
  94     *endp = p;
  95   return v;
  96 }
  97
  98 size_t
  99 string_to_size (char *p, char **endp)
 100 {
 101   off_t v = string_to_off (p, endp);
 102   size_t ret = v;
 103   if (ret != v)
 104     die (1, "number too big");
 105   return ret;
 106 }
 107
 108 size_t sparse_map_size;
 109 struct sp_array *sparse_map;
 110
 111 void
 112 get_line (char *s, int size, FILE *stream)
 113 {
 114   char *p = fgets (s, size, stream);
 115   size_t len;
 116
 117   if (!p)
 118     die (1, "unexpected end of file");
 119   len = strlen (p);
 120   if (s[len - 1] != '\n')
 121     die (1, "buffer overflow");
 122   s[len - 1] = 0;
 123 }
 124
 125 int
 126 get_var (FILE *fp, char **name, char **value)
 127 {
 128   static char *buffer;
 129   static size_t bufsize = OFF_T_STRSIZE_BOUND;
 130   char *p, *q;
 131
 132   buffer = emalloc (bufsize);
 133   do
 134     {
 135       size_t len, s;
 136
 137       if (!fgets (buffer, bufsize, fp))
 138         return 0;
 139       len = strlen (buffer);
 140       if (len == 0)
 141         return 0;
 142
 143       s = string_to_size (buffer, &p);
 144       if (*p != ' ')
 145         die (1, "malformed header: expected space but found %s", p);
 146       if (buffer[len-1] != '\n')
 147         {
 148           if (bufsize < s + 1)
 149             {
 150               bufsize = s + 1;
 151               buffer = realloc (buffer, bufsize);
 152               if (!buffer)
 153                 die (1, "not enough memory");
 154             }
 155           if (!fgets (buffer + len, s - len + 1, fp))
 156             die (1, "unexpected end of file or read error");
 157         }
 158       p++;
 159     }
 160   while (memcmp (p, "GNU.sparse.", 11));
 161
 162   p += 11;
 163   q = strchr (p, '=');
 164   if (!q)
 165     die (1, "malformed header: expected `=' not found");
 166   *q++ = 0;
 167   q[strlen (q) - 1] = 0;
 168   *name = p;
 169   *value = q;
 170   return 1;
 171 }
 172
 173 char *outname;
 174 off_t outsize;
 175 unsigned version_major;
 176 unsigned version_minor;
 177
 178 void
 179 read_xheader (char *name)
 180 {
 181   char *kw, *val;
 182   FILE *fp = fopen (name, "r");
 183   char *expect = NULL;
 184   size_t i = 0;
 185
 186   if (verbose)
 187     printf ("Reading extended header file\n");
 188
 189   while (get_var (fp, &kw, &val))
 190     {
 191       if (verbose)
 192         printf ("Found variable GNU.sparse.%s = %s\n", kw, val);
 193
 194       if (expect && strcmp (kw, expect))
 195         die (1, "bad keyword sequence: expected `%s' but found `%s'",
 196              expect, kw);
 197       expect = NULL;
 198       if (strcmp (kw, "name") == 0)
 199         {
 200           outname = emalloc (strlen (val) + 1);
 201           strcpy (outname, val);
 202         }
 203       else if (strcmp (kw, "major") == 0)
 204         {
 205           version_major = string_to_size (val, NULL);
 206         }
 207       else if (strcmp (kw, "minor") == 0)
 208         {
 209           version_minor = string_to_size (val, NULL);
 210         }
 211       else if (strcmp (kw, "realsize") == 0
 212                || strcmp (kw, "size") == 0)
 213         {
 214           outsize = string_to_off (val, NULL);
 215         }
 216       else if (strcmp (kw, "numblocks") == 0)
 217         {
 218           sparse_map_size = string_to_size (val, NULL);
 219           sparse_map = emalloc (sparse_map_size * sizeof *sparse_map);
 220         }
 221       else if (strcmp (kw, "offset") == 0)
 222         {
 223           sparse_map[i].offset = string_to_off (val, NULL);
 224           expect = "numbytes";
 225         }
 226       else if (strcmp (kw, "numbytes") == 0)
 227         {
 228           sparse_map[i++].numbytes = string_to_size (val, NULL);
 229         }
 230       else if (strcmp (kw, "map") == 0)
 231         {
 232           for (i = 0; i < sparse_map_size; i++)
 233             {
 234               sparse_map[i].offset = string_to_off (val, &val);
 235               if (*val != ',')
 236                 die (1, "bad GNU.sparse.map: expected `,' but found `%c'",
 237                      *val);
 238               sparse_map[i].numbytes = string_to_size (val+1, &val);
 239               if (*val != ',')
 240                 {
 241                   if (!(*val == 0 && i == sparse_map_size-1))
 242                     die (1, "bad GNU.sparse.map: expected `,' but found `%c'",
 243                          *val);
 244                 }
 245               else
 246                 val++;
 247             }
 248           if (*val)
 249             die (1, "bad GNU.sparse.map: garbage at the end");
 250         }
 251     }
 252   if (expect)
 253     die (1, "bad keyword sequence: expected `%s' not found", expect);
 254   if (version_major == 0 && sparse_map_size == 0)
 255     die (1, "size of the sparse map unknown");
 256   if (i != sparse_map_size)
 257     die (1, "not all sparse entries supplied");
 258   fclose (fp);
 259 }
 260
 261 void
 262 read_map (FILE *ifp)
 263 {
 264   size_t i;
 265   char nbuf[OFF_T_STRSIZE_BOUND];
 266
 267   if (verbose)
 268     printf ("Reading v.1.0 sparse map\n");
 269
 270   get_line (nbuf, sizeof nbuf, ifp);
 271   sparse_map_size = string_to_size (nbuf, NULL);
 272   sparse_map = emalloc (sparse_map_size * sizeof *sparse_map);
 273
 274   for (i = 0; i < sparse_map_size; i++)
 275     {
 276       get_line (nbuf, sizeof nbuf, ifp);
 277       sparse_map[i].offset = string_to_off (nbuf, NULL);
 278       get_line (nbuf, sizeof nbuf, ifp);
 279       sparse_map[i].numbytes = string_to_size (nbuf, NULL);
 280     }
 281
 282   fseek (ifp, ((ftell (ifp) + BLOCKSIZE - 1) / BLOCKSIZE) * BLOCKSIZE,
 283          SEEK_SET);
 284 }
 285
 286 void
 287 expand_sparse (FILE *sfp, int ofd)
 288 {
 289   size_t i;
 290   size_t maxbytes = 0;
 291   char *buffer;
 292
 293   for (i = 0; i < sparse_map_size; i++)
 294     if (maxbytes < sparse_map[i].numbytes)
 295       maxbytes = sparse_map[i].numbytes;
 296
 297   for (buffer = malloc (maxbytes); !buffer; maxbytes /= 2)
 298     if (maxbytes == 0)
 299       die (1, "not enough memory");
 300
 301   for (i = 0; i < sparse_map_size; i++)
 302     {
 303       size_t size = sparse_map[i].numbytes;
 304
 305       if (size == 0)
 306         ftruncate (ofd, sparse_map[i].offset);
 307       else
 308         {
 309           lseek (ofd, sparse_map[i].offset, SEEK_SET);
 310           while (size)
 311             {
 312               size_t rdsize = (size < maxbytes) ? size : maxbytes;
 313               if (rdsize != fread (buffer, 1, rdsize, sfp))
 314                 die (1, "read error (%d)", errno);
 315               if (rdsize != write (ofd, buffer, rdsize))
 316                 die (1, "write error (%d)", errno);
 317               size -= rdsize;
 318             }
 319         }
 320     }
 321   free (buffer);
 322 }
 323
 324 void
 325 usage (int code)
 326 {
 327   printf ("Usage: %s [OPTIONS] infile [outfile]\n", progname);
 328   printf ("%s: expand sparse files extracted from GNU archives\n",
 329           progname);
 330   printf ("\nOPTIONS are:\n\n");
 331   printf ("  -h           Display this help list\n");
 332   printf ("  -n           Dry run: do nothing, print what would have been done\n");
 333   printf ("  -v           Increase verbosity level\n");
 334   printf ("  -x FILE      Parse extended header FILE\n\n");
 335
 336   exit (code);
 337 }
 338
 339 void
 340 guess_outname (char *name)
 341 {
 342   char *p;
 343   char *s;
 344
 345   if (name[0] == '.' && name[1] == '/')
 346     name += 2;
 347
 348   p = name + strlen (name) - 1;
 349   s = NULL;
 350
 351   for (; p > name && *p != '/'; p--)
 352     ;
 353   if (*p == '/')
 354     s = p + 1;
 355   if (p != name)
 356     {
 357       for (p--; p > name && *p != '/'; p--)
 358         ;
 359     }
 360
 361   if (*p != '/')
 362     {
 363       if (s)
 364         outname = s;
 365       else
 366         {
 367           outname = emalloc (4 + strlen (name));
 368           strcpy (outname, "../");
 369           strcpy (outname + 3, name);
 370         }
 371     }
 372   else
 373     {
 374       size_t len = p - name + 1;
 375       outname = emalloc (len + strlen (s) + 1);
 376       memcpy (outname, name, len);
 377       strcpy (outname + len, s);
 378     }
 379 }
 380
 381 int
 382 main (int argc, char **argv)
 383 {
 384   int c;
 385   int dry_run = 0;
 386   char *xheader_file = NULL;
 387   char *inname;
 388   FILE *ifp;
 389   struct stat st;
 390   int ofd;
 391
 392   progname = argv[0];
 393   while ((c = getopt (argc, argv, "hnvx:")) != EOF)
 394     {
 395       switch (c)
 396         {
 397         case 'h':
 398           usage (0);
 399           break;
 400
 401         case 'x':
 402           xheader_file = optarg;
 403           break;
 404
 405         case 'n':
 406           dry_run = 1;
 407         case 'v':
 408           verbose++;
 409           break;
 410
 411         default:
 412           exit (1);
 413         }
 414     }
 415
 416   argc -= optind;
 417   argv += optind;
 418
 419   if (argc == 0 || argc > 2)
 420     usage (1);
 421
 422   if (xheader_file)
 423     read_xheader (xheader_file);
 424
 425   inname = argv[0];
 426   if (argv[1])
 427     outname = argv[1];
 428
 429   if (stat (inname, &st))
 430     die (1, "cannot stat %s (%d)", inname, errno);
 431
 432   ifp = fopen (inname, "r");
 433   if (ifp == NULL)
 434     die (1, "cannot open file %s (%d)", inname, errno);
 435
 436   if (!xheader_file || version_major == 1)
 437     read_map (ifp);
 438
 439   if (!outname)
 440     guess_outname (inname);
 441
 442   ofd = open (outname, O_RDWR|O_CREAT|O_TRUNC, st.st_mode);
 443   if (ofd == -1)
 444     die (1, "cannot open file %s (%d)", outname, errno);
 445
 446   if (verbose)
 447     printf ("Expanding file `%s' to `%s'\n", inname, outname);
 448
 449   if (dry_run)
 450     {
 451       printf ("Finished dry run\n");
 452       return 0;
 453     }
 454
 455   expand_sparse (ifp, ofd);
 456
 457   fclose (ifp);
 458   close (ofd);
 459
 460   if (verbose)
 461     printf ("Done\n");
 462
 463   if (outsize)
 464     {
 465       if (stat (outname, &st))
 466         die (1, "cannot stat output file %s (%d)", outname, errno);
 467       if (st.st_size != outsize)
 468         die (1, "expanded file has wrong size");
 469     }
 470
 471   return 0;
 472 }
 473