gcc/java/zextract.c

   1 /* Handle a .class file embedded in a .zip archive.
   2    This extracts a member from a .zip file, but does not handle
   3    uncompression (since that is not needed for classes.zip).
   4    Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004, 2005
   5    Free Software Foundation, Inc.
   6
   7 This file is part of GCC.
   8
   9 GCC is free software; you can redistribute it and/or modify
  10 it under the terms of the GNU General Public License as published by
  11 the Free Software Foundation; either version 2, or (at your option)
  12 any later version.
  13
  14 GCC is distributed in the hope that it will be useful,
  15 but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17 GNU General Public License for more details.
  18
  19 You should have received a copy of the GNU General Public License
  20 along with GCC; see the file COPYING.  If not, write to
  21 the Free Software Foundation, 51 Franklin Street, Fifth Floor,
  22 Boston, MA 02110-1301, USA.
  23
  24 Java and all Java-based marks are trademarks or registered trademarks
  25 of Sun Microsystems, Inc. in the United States and other countries.
  26 The Free Software Foundation is independent of Sun Microsystems, Inc.  */
  27
  28 /* Written by Per Bothner <bothner@cygnus.com>, February 1996. */
  29
  30 #include "config.h"
  31 #include "system.h"
  32 #include "coretypes.h"
  33 #include "tm.h"
  34 #include "zipfile.h"
  35
  36 /* This stuff is partly based on the 28 August 1994 public release of the
  37 Info-ZIP group's portable UnZip zipfile-extraction program (and related
  38 utilities). */
  39
  40 /*************/
  41 /*  Defines  */
  42 /*************/
  43
  44 #define UNZIP
  45 #define UNZIP_VERSION     20   /* compatible with PKUNZIP 2.0 */
  46 #define VMS_UNZIP_VERSION 42   /* if OS-needed-to-extract is VMS:  can do */
  47
  48
  49 #define ZSUFX             ".zip"
  50 #define CENTRAL_HDR_SIG   "\113\001\002"   /* the infamous "PK" signature */
  51 #define LOCAL_HDR_SIG     "\113\003\004"   /*  bytes, sans "P" (so unzip */
  52 #define END_CENTRAL_SIG   "\113\005\006"   /*  executable not mistaken for */
  53 #define EXTD_LOCAL_SIG    "\113\007\010"   /*  zipfile itself) */
  54
  55 #define STORED            0    /* compression methods */
  56 #define SHRUNK            1
  57 #define REDUCED1          2
  58 #define REDUCED2          3
  59 #define REDUCED3          4
  60 #define REDUCED4          5
  61 #define IMPLODED          6
  62 #define TOKENIZED         7
  63 #define DEFLATED          8
  64 #define NUM_METHODS       9    /* index of last method + 1 */
  65 /* don't forget to update list_files() appropriately if NUM_METHODS changes */
  66
  67 #define PK_OK             0    /* no error */
  68 #define PK_COOL           0    /* no error */
  69 #define PK_GNARLY         0    /* no error */
  70 #define PK_WARN           1    /* warning error */
  71 #define PK_ERR            2    /* error in zipfile */
  72 #define PK_BADERR         3    /* severe error in zipfile */
  73 #define PK_MEM            4    /* insufficient memory */
  74 #define PK_MEM2           5    /* insufficient memory */
  75 #define PK_MEM3           6    /* insufficient memory */
  76 #define PK_MEM4           7    /* insufficient memory */
  77 #define PK_MEM5           8    /* insufficient memory */
  78 #define PK_NOZIP          9    /* zipfile not found */
  79 #define PK_PARAM          10   /* bad or illegal parameters specified */
  80 #define PK_FIND           11   /* no files found */
  81 #define PK_DISK           50   /* disk full */
  82 #define PK_EOF            51   /* unexpected EOF */
  83
  84 /*---------------------------------------------------------------------------
  85     True sizes of the various headers, as defined by PKWARE--so it is not
  86     likely that these will ever change.  But if they do, make sure both these
  87     defines AND the typedefs below get updated accordingly.
  88   ---------------------------------------------------------------------------*/
  89 #define LREC_SIZE     26    /* lengths of local file headers, central */
  90 #define CREC_SIZE     42    /*  directory headers, and the end-of-    */
  91 #define ECREC_SIZE    18    /*  central-dir record, respectively      */
  92
  93
  94 #ifndef SEEK_SET
  95 #  define SEEK_SET  0
  96 #  define SEEK_CUR  1
  97 #  define SEEK_END  2
  98 #endif
  99
 100 /**************/
 101 /*  Typedefs  */
 102 /**************/
 103
 104 typedef char              boolean;
 105 typedef unsigned char     uch;  /* code assumes unsigned bytes; these type-  */
 106 typedef unsigned short    ush;  /*  defs replace byte/UWORD/ULONG (which are */
 107 typedef unsigned long     ulg;  /*  predefined on some systems) & match zip  */
 108
 109 /*---------------------------------------------------------------------------
 110     Zipfile layout declarations.  If these headers ever change, make sure the
 111     xxREC_SIZE defines (above) change with them!
 112   ---------------------------------------------------------------------------*/
 113
 114    typedef uch   local_byte_hdr[ LREC_SIZE ];
 115 #      define L_VERSION_NEEDED_TO_EXTRACT_0     0
 116 #      define L_VERSION_NEEDED_TO_EXTRACT_1     1
 117 #      define L_GENERAL_PURPOSE_BIT_FLAG        2
 118 #      define L_COMPRESSION_METHOD              4
 119 #      define L_LAST_MOD_FILE_TIME              6
 120 #      define L_LAST_MOD_FILE_DATE              8
 121 #      define L_CRC32                           10
 122 #      define L_COMPRESSED_SIZE                 14
 123 #      define L_UNCOMPRESSED_SIZE               18
 124 #      define L_FILENAME_LENGTH                 22
 125 #      define L_EXTRA_FIELD_LENGTH              24
 126
 127   typedef uch   cdir_byte_hdr[ CREC_SIZE ];
 128 #      define C_VERSION_MADE_BY_0               0
 129 #      define C_VERSION_MADE_BY_1               1
 130 #      define C_VERSION_NEEDED_TO_EXTRACT_0     2
 131 #      define C_VERSION_NEEDED_TO_EXTRACT_1     3
 132 #      define C_GENERAL_PURPOSE_BIT_FLAG        4
 133 #      define C_COMPRESSION_METHOD              6
 134 #      define C_LAST_MOD_FILE_TIME              8
 135 #      define C_LAST_MOD_FILE_DATE              10
 136 #      define C_CRC32                           12
 137 #      define C_COMPRESSED_SIZE                 16
 138 #      define C_UNCOMPRESSED_SIZE               20
 139 #      define C_FILENAME_LENGTH                 24
 140 #      define C_EXTRA_FIELD_LENGTH              26
 141 #      define C_FILE_COMMENT_LENGTH             28
 142 #      define C_DISK_NUMBER_START               30
 143 #      define C_INTERNAL_FILE_ATTRIBUTES        32
 144 #      define C_EXTERNAL_FILE_ATTRIBUTES        34
 145 #      define C_RELATIVE_OFFSET_LOCAL_HEADER    38
 146
 147    typedef uch   ec_byte_rec[ ECREC_SIZE+4 ];
 148 /*     define SIGNATURE                         0   space-holder only */
 149 #      define NUMBER_THIS_DISK                  4
 150 #      define NUM_DISK_WITH_START_CENTRAL_DIR   6
 151 #      define NUM_ENTRIES_CENTRL_DIR_THS_DISK   8
 152 #      define TOTAL_ENTRIES_CENTRAL_DIR         10
 153 #      define SIZE_CENTRAL_DIRECTORY            12
 154 #      define OFFSET_START_CENTRAL_DIRECTORY    16
 155 #      define ZIPFILE_COMMENT_LENGTH            20
 156
 157
 158    typedef struct local_file_header {                 /* LOCAL */
 159        uch version_needed_to_extract[2];
 160        ush general_purpose_bit_flag;
 161        ush compression_method;
 162        ush last_mod_file_time;
 163        ush last_mod_file_date;
 164        ulg crc32;
 165        ulg csize;
 166        ulg ucsize;
 167        ush filename_length;
 168        ush extra_field_length;
 169    } local_file_hdr;
 170
 171    typedef struct central_directory_file_header {     /* CENTRAL */
 172        uch version_made_by[2];
 173        uch version_needed_to_extract[2];
 174        ush general_purpose_bit_flag;
 175        ush compression_method;
 176        ush last_mod_file_time;
 177        ush last_mod_file_date;
 178        ulg crc32;
 179        ulg csize;
 180        ulg ucsize;
 181        ush filename_length;
 182        ush extra_field_length;
 183        ush file_comment_length;
 184        ush disk_number_start;
 185        ush internal_file_attributes;
 186        ulg external_file_attributes;
 187        ulg relative_offset_local_header;
 188    } cdir_file_hdr;
 189
 190    typedef struct end_central_dir_record {            /* END CENTRAL */
 191        ush number_this_disk;
 192        ush num_disk_with_start_central_dir;
 193        ush num_entries_centrl_dir_ths_disk;
 194        ush total_entries_central_dir;
 195        ulg size_central_directory;
 196        ulg offset_start_central_directory;
 197        ush zipfile_comment_length;
 198    } ecdir_rec;
 199
 200
 201 /************/
 202 /*  Macros  */
 203 /************/
 204
 205 #ifndef MAX
 206 #  define MAX(a,b)   ((a) > (b) ? (a) : (b))
 207 #endif
 208 #ifndef MIN
 209 #  define MIN(a,b)   ((a) < (b) ? (a) : (b))
 210 #endif
 211
 212
 213 /***********************/
 214 /* Prototypes          */
 215 /***********************/
 216
 217 static ush makeword (const uch *);
 218 static ulg makelong (const uch *);
 219 static long find_zip_file_start (int fd, long offset);
 220
 221 /***********************/
 222 /* Function makeword() */
 223 /***********************/
 224
 225 static ush makeword(const uch *b)
 226 {
 227     /*
 228      * Convert Intel style 'short' integer to non-Intel non-16-bit
 229      * host format.  This routine also takes care of byte-ordering.
 230      */
 231     return (ush)((b[1] << 8) | b[0]);
 232 }
 233
 234
 235 /***********************/
 236 /* Function makelong() */
 237 /***********************/
 238
 239 static ulg
 240 makelong (const uch *sig)
 241 {
 242     /*
 243      * Convert intel style 'long' variable to non-Intel non-16-bit
 244      * host format.  This routine also takes care of byte-ordering.
 245      */
 246     return (((ulg)sig[3]) << 24)
 247         + (((ulg)sig[2]) << 16)
 248         + (((ulg)sig[1]) << 8)
 249         + ((ulg)sig[0]);
 250 }
 251
 252 /* Examine file's header in zip file and return the offset of the
 253    start of the actual data.  Return -1 on error.  OFFSET is the
 254    offset from the beginning of the zip file of the file's header.  */
 255 static long
 256 find_zip_file_start (int fd, long offset)
 257 {
 258   int filename_length, extra_field_length;
 259   unsigned char buffer[LREC_SIZE + 4];
 260
 261   if (lseek (fd, offset, SEEK_SET) < 0)
 262     return -1;
 263
 264   if (read (fd, buffer, LREC_SIZE + 4) != LREC_SIZE + 4)
 265     return -1;
 266
 267   if (buffer[0] != 'P' || strncmp ((const char *) &buffer[1], LOCAL_HDR_SIG, 3))
 268     return -1;
 269
 270   filename_length = makeword (&buffer[4 + L_FILENAME_LENGTH]);
 271   extra_field_length = makeword (&buffer[4 + L_EXTRA_FIELD_LENGTH]);
 272
 273   return offset + (4 + LREC_SIZE) + filename_length + extra_field_length;
 274 }
 275
 276 int
 277 read_zip_archive (ZipFile *zipf)
 278 {
 279   int i;
 280   int dir_last_pad;
 281   char *dir_ptr;
 282   char buffer[100];
 283
 284   zipf->size = lseek (zipf->fd, 0L, SEEK_END);
 285
 286   if (zipf->size < (ECREC_SIZE+4) || lseek (zipf->fd, (long)(-(ECREC_SIZE+4)), SEEK_CUR) <= 0)
 287     return -1;
 288   if (read (zipf->fd, buffer, ECREC_SIZE+4) != ECREC_SIZE+4)
 289     return -2;
 290   if (buffer[0] != 'P'
 291       || strncmp ((const char *) &buffer[1], END_CENTRAL_SIG, 3))
 292     {
 293       /* We could not find the end-central-header signature, probably
 294          because a zipfile comment is present. Scan backwards until we
 295          find the signature. */
 296       if (lseek (zipf->fd, (long)(-ECREC_SIZE), SEEK_END) <= 0)
 297         return -2;
 298       while (buffer[0] != 'P'
 299              || strncmp ((const char *) &buffer[1], END_CENTRAL_SIG, 3))
 300         {
 301           if (lseek (zipf->fd, -5, SEEK_CUR) < 0)
 302             return -2;
 303           if (read (zipf->fd, buffer, 4) != 4)
 304             return -2;
 305         }
 306       if (read (zipf->fd, buffer + 4, ECREC_SIZE) != ECREC_SIZE)
 307         return -2;
 308     }
 309   zipf->count = makeword((const uch *) &buffer[TOTAL_ENTRIES_CENTRAL_DIR]);
 310   zipf->dir_size = makelong((const uch *) &buffer[SIZE_CENTRAL_DIRECTORY]);
 311   /* Allocate 1 more to allow appending '\0' to last filename. */
 312   zipf->central_directory = XNEWVEC (char, zipf->dir_size + 1);
 313   if (lseek (zipf->fd, -(zipf->dir_size+ECREC_SIZE+4), SEEK_CUR) < 0)
 314     return -2;
 315   if (read (zipf->fd, zipf->central_directory, zipf->dir_size) < 0)
 316     return -2;
 317
 318 #ifdef TEST
 319   printf ("number_this_disk = %d\n", makeword(&buffer[NUMBER_THIS_DISK]));
 320   printf ("num_disk_with_start_central_dir = %d\n", makeword(&buffer[NUM_DISK_WITH_START_CENTRAL_DIR]));
 321
 322   printf ("num_entries_centrl_dir_ths_disk = %d\n",
 323         makeword(&buffer[NUM_ENTRIES_CENTRL_DIR_THS_DISK]));
 324   printf ("total_entries_central_dir = %d\n",
 325         makeword(&buffer[TOTAL_ENTRIES_CENTRAL_DIR]));
 326   printf ("size_central_directory = %d\n",
 327         makelong((const uch *) &buffer[SIZE_CENTRAL_DIRECTORY]));
 328   printf ("offset_start_central_directory = %d\n",
 329         makelong((const uch *) &buffer[OFFSET_START_CENTRAL_DIRECTORY]));
 330   printf ("zipfile_comment_length = %d\n",
 331         makeword(&buffer[ZIPFILE_COMMENT_LENGTH]));
 332 #endif
 333
 334   dir_last_pad = 0;
 335   dir_ptr = zipf->central_directory;
 336   for (i = 0; i < zipf->count; i++)
 337     {
 338       ZipDirectory *zipd = (ZipDirectory*)(dir_ptr + dir_last_pad);
 339       int compression_method = (int) dir_ptr[4+C_COMPRESSION_METHOD];
 340       long size = makelong ((const uch *) &dir_ptr[4+C_COMPRESSED_SIZE]);
 341       long uncompressed_size = makelong ((const uch *) &dir_ptr[4+C_UNCOMPRESSED_SIZE]);
 342       long filename_length = makeword ((const uch *) &dir_ptr[4+C_FILENAME_LENGTH]);
 343       long extra_field_length = makeword ((const uch *) &dir_ptr[4+C_EXTRA_FIELD_LENGTH]);
 344       long file_offset = makelong ((const uch *) &dir_ptr[4+C_RELATIVE_OFFSET_LOCAL_HEADER]);
 345       int unpadded_direntry_length;
 346       if ((dir_ptr-zipf->central_directory)+filename_length+CREC_SIZE+4>zipf->dir_size)
 347         return -1;
 348
 349       zipd->filename_length = filename_length;
 350       zipd->compression_method = compression_method;
 351       zipd->size = size;
 352       zipd->uncompressed_size = uncompressed_size;
 353       zipd->zipf = zipf;
 354 #ifdef __GNUC__
 355 #define DIR_ALIGN __alignof__(ZipDirectory)
 356 #else
 357 #define DIR_ALIGN sizeof(long)
 358 #endif
 359       zipd->filestart = find_zip_file_start (zipf->fd, file_offset);
 360       zipd->filename_offset = CREC_SIZE+4 - dir_last_pad;
 361       unpadded_direntry_length
 362           = zipd->filename_offset + zipd->filename_length + extra_field_length;
 363       zipd->direntry_size =
 364         ((unpadded_direntry_length + DIR_ALIGN) / DIR_ALIGN) * DIR_ALIGN;
 365       dir_last_pad = zipd->direntry_size - unpadded_direntry_length;
 366       dir_ptr = (char*)zipd + unpadded_direntry_length;
 367       *dir_ptr = '\0';
 368     }
 369   return 0;
 370 }
 371
 372 #ifdef TEST
 373 main (void)
 374 {
 375   ZipFile zipf[1];
 376   ZipDirectory *zipd;
 377   int i;
 378
 379   zipf->fd = 0;
 380
 381   i = read_zip_archive (zipf);
 382   if (i)
 383     {
 384       fprintf (stderr, "Bad zip file.\n");
 385       exit (i);
 386     }
 387
 388   zipd = (ZipDirectory*) zipf->central_directory;
 389   for (i = 0; i < zipf->count; i++, zipd = ZIPDIR_NEXT (zipd))
 390     {
 391       printf ("%d: size:%d, name(#%d)%s, offset:%d\n",
 392               i, zipd->size, zipd->filename_length,
 393               ZIPDIR_FILENAME (zipd),
 394               zipd->filestart);
 395     }
 396 }
 397 #endif