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
   5    Copyright (C) 1996, 1997, 1998, 1999, 2000  Free Software Foundation, Inc.
   6
   7 This program is free software; you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License as published by
   9 the Free Software Foundation; either version 2, or (at your option)
  10 any later version.
  11
  12 This program is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with GNU CC; see the file COPYING.  If not, write to
  19 the Free Software Foundation, 59 Temple Place - Suite 330,
  20 Boston, MA 02111-1307, USA.
  21
  22 Java and all Java-based marks are trademarks or registered trademarks
  23 of Sun Microsystems, Inc. in the United States and other countries.
  24 The Free Software Foundation is independent of Sun Microsystems, Inc.  */
  25
  26 /* Written by Per Bothner <bothner@cygnus.com>, February 1996. */
  27
  28 #include "config.h"
  29 #include "system.h"
  30 #include "zipfile.h"
  31
  32 /* This stuff is partly based on the 28 August 1994 public release of the
  33 Info-ZIP group's portable UnZip zipfile-extraction program (and related
  34 utilities). */
  35
  36 /*************/
  37 /*  Defines  */
  38 /*************/
  39
  40 #define UNZIP
  41 #define UNZIP_VERSION     20   /* compatible with PKUNZIP 2.0 */
  42 #define VMS_UNZIP_VERSION 42   /* if OS-needed-to-extract is VMS:  can do */
  43
  44
  45 #define ZSUFX             ".zip"
  46 #define CENTRAL_HDR_SIG   "\113\001\002"   /* the infamous "PK" signature */
  47 #define LOCAL_HDR_SIG     "\113\003\004"   /*  bytes, sans "P" (so unzip */
  48 #define END_CENTRAL_SIG   "\113\005\006"   /*  executable not mistaken for */
  49 #define EXTD_LOCAL_SIG    "\113\007\010"   /*  zipfile itself) */
  50
  51 #define STORED            0    /* compression methods */
  52 #define SHRUNK            1
  53 #define REDUCED1          2
  54 #define REDUCED2          3
  55 #define REDUCED3          4
  56 #define REDUCED4          5
  57 #define IMPLODED          6
  58 #define TOKENIZED         7
  59 #define DEFLATED          8
  60 #define NUM_METHODS       9    /* index of last method + 1 */
  61 /* don't forget to update list_files() appropriately if NUM_METHODS changes */
  62
  63 #define PK_OK             0    /* no error */
  64 #define PK_COOL           0    /* no error */
  65 #define PK_GNARLY         0    /* no error */
  66 #define PK_WARN           1    /* warning error */
  67 #define PK_ERR            2    /* error in zipfile */
  68 #define PK_BADERR         3    /* severe error in zipfile */
  69 #define PK_MEM            4    /* insufficient memory */
  70 #define PK_MEM2           5    /* insufficient memory */
  71 #define PK_MEM3           6    /* insufficient memory */
  72 #define PK_MEM4           7    /* insufficient memory */
  73 #define PK_MEM5           8    /* insufficient memory */
  74 #define PK_NOZIP          9    /* zipfile not found */
  75 #define PK_PARAM          10   /* bad or illegal parameters specified */
  76 #define PK_FIND           11   /* no files found */
  77 #define PK_DISK           50   /* disk full */
  78 #define PK_EOF            51   /* unexpected EOF */
  79
  80 /*---------------------------------------------------------------------------
  81     True sizes of the various headers, as defined by PKWARE--so it is not
  82     likely that these will ever change.  But if they do, make sure both these
  83     defines AND the typedefs below get updated accordingly.
  84   ---------------------------------------------------------------------------*/
  85 #define LREC_SIZE     26    /* lengths of local file headers, central */
  86 #define CREC_SIZE     42    /*  directory headers, and the end-of-    */
  87 #define ECREC_SIZE    18    /*  central-dir record, respectively      */
  88
  89
  90 #ifndef SEEK_SET
  91 #  define SEEK_SET  0
  92 #  define SEEK_CUR  1
  93 #  define SEEK_END  2
  94 #endif
  95
  96 /**************/
  97 /*  Typedefs  */
  98 /**************/
  99
 100 typedef char              boolean;
 101 typedef unsigned char     uch;  /* code assumes unsigned bytes; these type-  */
 102 typedef unsigned short    ush;  /*  defs replace byte/UWORD/ULONG (which are */
 103 typedef unsigned long     ulg;  /*  predefined on some systems) & match zip  */
 104
 105 /*---------------------------------------------------------------------------
 106     Zipfile layout declarations.  If these headers ever change, make sure the
 107     xxREC_SIZE defines (above) change with them!
 108   ---------------------------------------------------------------------------*/
 109
 110    typedef uch   local_byte_hdr[ LREC_SIZE ];
 111 #      define L_VERSION_NEEDED_TO_EXTRACT_0     0
 112 #      define L_VERSION_NEEDED_TO_EXTRACT_1     1
 113 #      define L_GENERAL_PURPOSE_BIT_FLAG        2
 114 #      define L_COMPRESSION_METHOD              4
 115 #      define L_LAST_MOD_FILE_TIME              6
 116 #      define L_LAST_MOD_FILE_DATE              8
 117 #      define L_CRC32                           10
 118 #      define L_COMPRESSED_SIZE                 14
 119 #      define L_UNCOMPRESSED_SIZE               18
 120 #      define L_FILENAME_LENGTH                 22
 121 #      define L_EXTRA_FIELD_LENGTH              24
 122
 123   typedef uch   cdir_byte_hdr[ CREC_SIZE ];
 124 #      define C_VERSION_MADE_BY_0               0
 125 #      define C_VERSION_MADE_BY_1               1
 126 #      define C_VERSION_NEEDED_TO_EXTRACT_0     2
 127 #      define C_VERSION_NEEDED_TO_EXTRACT_1     3
 128 #      define C_GENERAL_PURPOSE_BIT_FLAG        4
 129 #      define C_COMPRESSION_METHOD              6
 130 #      define C_LAST_MOD_FILE_TIME              8
 131 #      define C_LAST_MOD_FILE_DATE              10
 132 #      define C_CRC32                           12
 133 #      define C_COMPRESSED_SIZE                 16
 134 #      define C_UNCOMPRESSED_SIZE               20
 135 #      define C_FILENAME_LENGTH                 24
 136 #      define C_EXTRA_FIELD_LENGTH              26
 137 #      define C_FILE_COMMENT_LENGTH             28
 138 #      define C_DISK_NUMBER_START               30
 139 #      define C_INTERNAL_FILE_ATTRIBUTES        32
 140 #      define C_EXTERNAL_FILE_ATTRIBUTES        34
 141 #      define C_RELATIVE_OFFSET_LOCAL_HEADER    38
 142
 143    typedef uch   ec_byte_rec[ ECREC_SIZE+4 ];
 144 /*     define SIGNATURE                         0   space-holder only */
 145 #      define NUMBER_THIS_DISK                  4
 146 #      define NUM_DISK_WITH_START_CENTRAL_DIR   6
 147 #      define NUM_ENTRIES_CENTRL_DIR_THS_DISK   8
 148 #      define TOTAL_ENTRIES_CENTRAL_DIR         10
 149 #      define SIZE_CENTRAL_DIRECTORY            12
 150 #      define OFFSET_START_CENTRAL_DIRECTORY    16
 151 #      define ZIPFILE_COMMENT_LENGTH            20
 152
 153
 154    typedef struct local_file_header {                 /* LOCAL */
 155        uch version_needed_to_extract[2];
 156        ush general_purpose_bit_flag;
 157        ush compression_method;
 158        ush last_mod_file_time;
 159        ush last_mod_file_date;
 160        ulg crc32;
 161        ulg csize;
 162        ulg ucsize;
 163        ush filename_length;
 164        ush extra_field_length;
 165    } local_file_hdr;
 166
 167    typedef struct central_directory_file_header {     /* CENTRAL */
 168        uch version_made_by[2];
 169        uch version_needed_to_extract[2];
 170        ush general_purpose_bit_flag;
 171        ush compression_method;
 172        ush last_mod_file_time;
 173        ush last_mod_file_date;
 174        ulg crc32;
 175        ulg csize;
 176        ulg ucsize;
 177        ush filename_length;
 178        ush extra_field_length;
 179        ush file_comment_length;
 180        ush disk_number_start;
 181        ush internal_file_attributes;
 182        ulg external_file_attributes;
 183        ulg relative_offset_local_header;
 184    } cdir_file_hdr;
 185
 186    typedef struct end_central_dir_record {            /* END CENTRAL */
 187        ush number_this_disk;
 188        ush num_disk_with_start_central_dir;
 189        ush num_entries_centrl_dir_ths_disk;
 190        ush total_entries_central_dir;
 191        ulg size_central_directory;
 192        ulg offset_start_central_directory;
 193        ush zipfile_comment_length;
 194    } ecdir_rec;
 195
 196
 197 /************/
 198 /*  Macros  */
 199 /************/
 200
 201 #ifndef MAX
 202 #  define MAX(a,b)   ((a) > (b) ? (a) : (b))
 203 #endif
 204 #ifndef MIN
 205 #  define MIN(a,b)   ((a) < (b) ? (a) : (b))
 206 #endif
 207
 208
 209 /***********************/
 210 /* Prototypes          */
 211 /***********************/
 212
 213 static ush makeword PARAMS ((const uch *));
 214 static ulg makelong PARAMS ((const uch *));
 215 static long find_zip_file_start PARAMS ((int fd, long offset));
 216
 217 /***********************/
 218 /* Function makeword() */
 219 /***********************/
 220
 221 static ush makeword(b)
 222     const uch *b;
 223 {
 224     /*
 225      * Convert Intel style 'short' integer to non-Intel non-16-bit
 226      * host format.  This routine also takes care of byte-ordering.
 227      */
 228     return (ush)((b[1] << 8) | b[0]);
 229 }
 230
 231
 232 /***********************/
 233 /* Function makelong() */
 234 /***********************/
 235
 236 static ulg makelong(sig)
 237     const uch *sig;
 238 {
 239     /*
 240      * Convert intel style 'long' variable to non-Intel non-16-bit
 241      * host format.  This routine also takes care of byte-ordering.
 242      */
 243     return (((ulg)sig[3]) << 24)
 244         + (((ulg)sig[2]) << 16)
 245         + (((ulg)sig[1]) << 8)
 246         + ((ulg)sig[0]);
 247 }
 248
 249 /* Examine file's header in zip file and return the offset of the
 250    start of the actual data.  Return -1 on error.  OFFSET is the
 251    offset from the beginning of the zip file of the file's header.  */
 252 static long
 253 find_zip_file_start (fd, offset)
 254      int fd;
 255      long offset;
 256 {
 257   int filename_length, extra_field_length;
 258   unsigned char buffer[LREC_SIZE + 4];
 259
 260   if (lseek (fd, offset, SEEK_SET) < 0)
 261     return -1;
 262
 263   if (read (fd, buffer, LREC_SIZE + 4) != LREC_SIZE + 4)
 264     return -1;
 265
 266   if (buffer[0] != 'P' || strncmp (&buffer[1], LOCAL_HDR_SIG, 3))
 267     return -1;
 268
 269   filename_length = makeword (&buffer[4 + L_FILENAME_LENGTH]);
 270   extra_field_length = makeword (&buffer[4 + L_EXTRA_FIELD_LENGTH]);
 271
 272   return offset + (4 + LREC_SIZE) + filename_length + extra_field_length;
 273 }
 274
 275 int
 276 read_zip_archive (zipf)
 277      register 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   zipf->count = makeword(&buffer[TOTAL_ENTRIES_CENTRAL_DIR]);
 291   zipf->dir_size = makelong(&buffer[SIZE_CENTRAL_DIRECTORY]);
 292 #define ALLOC xmalloc
 293   /* Allocate 1 more to allow appending '\0' to last filename. */
 294   zipf->central_directory = ALLOC (zipf->dir_size+1);
 295   if (lseek (zipf->fd, -(zipf->dir_size+ECREC_SIZE+4), SEEK_CUR) < 0)
 296     return -2;
 297   if (read (zipf->fd, zipf->central_directory, zipf->dir_size) < 0)
 298     return -2;
 299
 300 #ifdef TEST
 301   printf ("number_this_disk = %d\n", makeword(&buffer[NUMBER_THIS_DISK]));
 302   printf ("num_disk_with_start_central_dir = %d\n", makeword(&buffer[NUM_DISK_WITH_START_CENTRAL_DIR]));
 303
 304   printf ("num_entries_centrl_dir_ths_disk = %d\n",
 305         makeword(&buffer[NUM_ENTRIES_CENTRL_DIR_THS_DISK]));
 306   printf ("total_entries_central_dir = %d\n",
 307         makeword(&buffer[TOTAL_ENTRIES_CENTRAL_DIR]));
 308   printf ("size_central_directory = %d\n",
 309         makelong(&buffer[SIZE_CENTRAL_DIRECTORY]));
 310   printf ("offset_start_central_directory = %d\n",
 311         makelong(&buffer[OFFSET_START_CENTRAL_DIRECTORY]));
 312   printf ("zipfile_comment_length = %d\n",
 313         makeword(&buffer[ZIPFILE_COMMENT_LENGTH]));
 314 #endif
 315
 316   dir_last_pad = 0;
 317   dir_ptr = zipf->central_directory;
 318   for (i = 0; i < zipf->count; i++)
 319     {
 320       ZipDirectory *zipd = (ZipDirectory*)(dir_ptr + dir_last_pad);
 321       int compression_method = (int) dir_ptr[4+C_COMPRESSION_METHOD];
 322       long size = makelong (&dir_ptr[4+C_COMPRESSED_SIZE]);
 323       long uncompressed_size = makelong (&dir_ptr[4+C_UNCOMPRESSED_SIZE]);
 324       long filename_length = makeword (&dir_ptr[4+C_FILENAME_LENGTH]);
 325       long extra_field_length = makeword (&dir_ptr[4+C_EXTRA_FIELD_LENGTH]);
 326       long file_offset = makelong (&dir_ptr[4+C_RELATIVE_OFFSET_LOCAL_HEADER]);
 327       int unpadded_direntry_length;
 328       if ((dir_ptr-zipf->central_directory)+filename_length+CREC_SIZE+4>zipf->dir_size)
 329         return -1;
 330
 331       zipd->filename_length = filename_length;
 332       zipd->compression_method = compression_method;
 333       zipd->size = size;
 334       zipd->uncompressed_size = uncompressed_size;
 335       zipd->zipf = zipf;
 336 #ifdef __GNUC__
 337 #define DIR_ALIGN __alignof__(ZipDirectory)
 338 #else
 339 #define DIR_ALIGN sizeof(long)
 340 #endif
 341       zipd->filestart = find_zip_file_start (zipf->fd, file_offset);
 342       zipd->filename_offset = CREC_SIZE+4 - dir_last_pad;
 343       unpadded_direntry_length
 344           = zipd->filename_offset + zipd->filename_length + extra_field_length;
 345       zipd->direntry_size =
 346         ((unpadded_direntry_length + DIR_ALIGN) / DIR_ALIGN) * DIR_ALIGN;
 347       dir_last_pad = zipd->direntry_size - unpadded_direntry_length;
 348       dir_ptr = (char*)zipd + unpadded_direntry_length;
 349       *dir_ptr = '\0';
 350     }
 351   return 0;
 352 }
 353
 354 #ifdef TEST
 355 main ()
 356 {
 357   ZipFile zipf[1];
 358   ZipDirectory *zipd;
 359   int i;
 360
 361   zipf->fd = 0;
 362
 363   i = read_zip_archive (zipf);
 364   if (i)
 365     {
 366       fprintf (stderr, "Bad zip file.\n");
 367       exit (i);
 368     }
 369
 370   zipd = (ZipDirectory*) zipf->central_directory;
 371   for (i = 0; i < zipf->count; i++, zipd = ZIPDIR_NEXT (zipd))
 372     {
 373       printf ("%d: size:%d, name(#%d)%s, offset:%d\n",
 374               i, zipd->size, zipd->filename_length,
 375               ZIPDIR_FILENAME (zipd),
 376               zipd->filestart);
 377     }
 378 }
 379 #endif