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) 1996-2013 Free Software Foundation, Inc.
   5
   6 This file is part of GCC.
   7
   8 GCC is free software; you can redistribute it and/or modify
   9 it under the terms of the GNU General Public License as published by
  10 the Free Software Foundation; either version 3, or (at your option)
  11 any later version.
  12
  13 GCC is distributed in the hope that it will be useful,
  14 but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 GNU General Public License for more details.
  17
  18 You should have received a copy of the GNU General Public License
  19 along with GCC; see the file COPYING3.  If not see
  20 <http://www.gnu.org/licenses/>.
  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 "coretypes.h"
  31 #include "tm.h"
  32 #include "zipfile.h"
  33
  34 /* This stuff is partly based on the 28 August 1994 public release of the
  35 Info-ZIP group's portable UnZip zipfile-extraction program (and related
  36 utilities). */
  37
  38 /*************/
  39 /*  Defines  */
  40 /*************/
  41
  42 #define UNZIP
  43 #define UNZIP_VERSION     20   /* compatible with PKUNZIP 2.0 */
  44 #define VMS_UNZIP_VERSION 42   /* if OS-needed-to-extract is VMS:  can do */
  45
  46
  47 #define ZSUFX             ".zip"
  48 #define CENTRAL_HDR_SIG   "\113\001\002"   /* the infamous "PK" signature */
  49 #define LOCAL_HDR_SIG     "\113\003\004"   /*  bytes, sans "P" (so unzip */
  50 #define END_CENTRAL_SIG   "\113\005\006"   /*  executable not mistaken for */
  51 #define EXTD_LOCAL_SIG    "\113\007\010"   /*  zipfile itself) */
  52
  53 #define STORED            0    /* compression methods */
  54 #define SHRUNK            1
  55 #define REDUCED1          2
  56 #define REDUCED2          3
  57 #define REDUCED3          4
  58 #define REDUCED4          5
  59 #define IMPLODED          6
  60 #define TOKENIZED         7
  61 #define DEFLATED          8
  62 #define NUM_METHODS       9    /* index of last method + 1 */
  63 /* don't forget to update list_files() appropriately if NUM_METHODS changes */
  64
  65 #define PK_OK             0    /* no error */
  66 #define PK_COOL           0    /* no error */
  67 #define PK_GNARLY         0    /* no error */
  68 #define PK_WARN           1    /* warning error */
  69 #define PK_ERR            2    /* error in zipfile */
  70 #define PK_BADERR         3    /* severe error in zipfile */
  71 #define PK_MEM            4    /* insufficient memory */
  72 #define PK_MEM2           5    /* insufficient memory */
  73 #define PK_MEM3           6    /* insufficient memory */
  74 #define PK_MEM4           7    /* insufficient memory */
  75 #define PK_MEM5           8    /* insufficient memory */
  76 #define PK_NOZIP          9    /* zipfile not found */
  77 #define PK_PARAM          10   /* bad or illegal parameters specified */
  78 #define PK_FIND           11   /* no files found */
  79 #define PK_DISK           50   /* disk full */
  80 #define PK_EOF            51   /* unexpected EOF */
  81
  82 /*---------------------------------------------------------------------------
  83     True sizes of the various headers, as defined by PKWARE--so it is not
  84     likely that these will ever change.  But if they do, make sure both these
  85     defines AND the typedefs below get updated accordingly.
  86   ---------------------------------------------------------------------------*/
  87 #define LREC_SIZE     26    /* lengths of local file headers, central */
  88 #define CREC_SIZE     42    /*  directory headers, and the end-of-    */
  89 #define ECREC_SIZE    18    /*  central-dir record, respectively      */
  90
  91
  92 #ifndef SEEK_SET
  93 #  define SEEK_SET  0
  94 #  define SEEK_CUR  1
  95 #  define SEEK_END  2
  96 #endif
  97
  98 /**************/
  99 /*  Typedefs  */
 100 /**************/
 101
 102 typedef char              boolean;
 103 typedef unsigned char     uch;  /* code assumes unsigned bytes; these type-  */
 104 typedef unsigned short    ush;  /*  defs replace byte/UWORD/ULONG (which are */
 105 typedef unsigned long     ulg;  /*  predefined on some systems) & match zip  */
 106
 107 /*---------------------------------------------------------------------------
 108     Zipfile layout declarations.  If these headers ever change, make sure the
 109     xxREC_SIZE defines (above) change with them!
 110   ---------------------------------------------------------------------------*/
 111
 112    typedef uch   local_byte_hdr[ LREC_SIZE ];
 113 #      define L_VERSION_NEEDED_TO_EXTRACT_0     0
 114 #      define L_VERSION_NEEDED_TO_EXTRACT_1     1
 115 #      define L_GENERAL_PURPOSE_BIT_FLAG        2
 116 #      define L_COMPRESSION_METHOD              4
 117 #      define L_LAST_MOD_FILE_TIME              6
 118 #      define L_LAST_MOD_FILE_DATE              8
 119 #      define L_CRC32                           10
 120 #      define L_COMPRESSED_SIZE                 14
 121 #      define L_UNCOMPRESSED_SIZE               18
 122 #      define L_FILENAME_LENGTH                 22
 123 #      define L_EXTRA_FIELD_LENGTH              24
 124
 125   typedef uch   cdir_byte_hdr[ CREC_SIZE ];
 126 #      define C_VERSION_MADE_BY_0               0
 127 #      define C_VERSION_MADE_BY_1               1
 128 #      define C_VERSION_NEEDED_TO_EXTRACT_0     2
 129 #      define C_VERSION_NEEDED_TO_EXTRACT_1     3
 130 #      define C_GENERAL_PURPOSE_BIT_FLAG        4
 131 #      define C_COMPRESSION_METHOD              6
 132 #      define C_LAST_MOD_FILE_TIME              8
 133 #      define C_LAST_MOD_FILE_DATE              10
 134 #      define C_CRC32                           12
 135 #      define C_COMPRESSED_SIZE                 16
 136 #      define C_UNCOMPRESSED_SIZE               20
 137 #      define C_FILENAME_LENGTH                 24
 138 #      define C_EXTRA_FIELD_LENGTH              26
 139 #      define C_FILE_COMMENT_LENGTH             28
 140 #      define C_DISK_NUMBER_START               30
 141 #      define C_INTERNAL_FILE_ATTRIBUTES        32
 142 #      define C_EXTERNAL_FILE_ATTRIBUTES        34
 143 #      define C_RELATIVE_OFFSET_LOCAL_HEADER    38
 144
 145    typedef uch   ec_byte_rec[ ECREC_SIZE+4 ];
 146 /*     define SIGNATURE                         0   space-holder only */
 147 #      define NUMBER_THIS_DISK                  4
 148 #      define NUM_DISK_WITH_START_CENTRAL_DIR   6
 149 #      define NUM_ENTRIES_CENTRL_DIR_THS_DISK   8
 150 #      define TOTAL_ENTRIES_CENTRAL_DIR         10
 151 #      define SIZE_CENTRAL_DIRECTORY            12
 152 #      define OFFSET_START_CENTRAL_DIRECTORY    16
 153 #      define ZIPFILE_COMMENT_LENGTH            20
 154
 155
 156    typedef struct local_file_header {                 /* LOCAL */
 157        uch version_needed_to_extract[2];
 158        ush general_purpose_bit_flag;
 159        ush compression_method;
 160        ush last_mod_file_time;
 161        ush last_mod_file_date;
 162        ulg crc32;
 163        ulg csize;
 164        ulg ucsize;
 165        ush filename_length;
 166        ush extra_field_length;
 167    } local_file_hdr;
 168
 169    typedef struct central_directory_file_header {     /* CENTRAL */
 170        uch version_made_by[2];
 171        uch version_needed_to_extract[2];
 172        ush general_purpose_bit_flag;
 173        ush compression_method;
 174        ush last_mod_file_time;
 175        ush last_mod_file_date;
 176        ulg crc32;
 177        ulg csize;
 178        ulg ucsize;
 179        ush filename_length;
 180        ush extra_field_length;
 181        ush file_comment_length;
 182        ush disk_number_start;
 183        ush internal_file_attributes;
 184        ulg external_file_attributes;
 185        ulg relative_offset_local_header;
 186    } cdir_file_hdr;
 187
 188    typedef struct end_central_dir_record {            /* END CENTRAL */
 189        ush number_this_disk;
 190        ush num_disk_with_start_central_dir;
 191        ush num_entries_centrl_dir_ths_disk;
 192        ush total_entries_central_dir;
 193        ulg size_central_directory;
 194        ulg offset_start_central_directory;
 195        ush zipfile_comment_length;
 196    } ecdir_rec;
 197
 198
 199 /************/
 200 /*  Macros  */
 201 /************/
 202
 203 #ifndef MAX
 204 #  define MAX(a,b)   ((a) > (b) ? (a) : (b))
 205 #endif
 206 #ifndef MIN
 207 #  define MIN(a,b)   ((a) < (b) ? (a) : (b))
 208 #endif
 209
 210
 211 /***********************/
 212 /* Prototypes          */
 213 /***********************/
 214
 215 static ush makeword (const uch *);
 216 static ulg makelong (const uch *);
 217 static long find_zip_file_start (int fd, long offset);
 218
 219 /***********************/
 220 /* Function makeword() */
 221 /***********************/
 222
 223 static ush makeword(const uch *b)
 224 {
 225     /*
 226      * Convert Intel style 'short' integer to non-Intel non-16-bit
 227      * host format.  This routine also takes care of byte-ordering.
 228      */
 229     return (ush)((b[1] << 8) | b[0]);
 230 }
 231
 232
 233 /***********************/
 234 /* Function makelong() */
 235 /***********************/
 236
 237 static ulg
 238 makelong (const uch *sig)
 239 {
 240     /*
 241      * Convert intel style 'long' variable to non-Intel non-16-bit
 242      * host format.  This routine also takes care of byte-ordering.
 243      */
 244     return (((ulg)sig[3]) << 24)
 245         + (((ulg)sig[2]) << 16)
 246         + (((ulg)sig[1]) << 8)
 247         + ((ulg)sig[0]);
 248 }
 249
 250 /* Examine file's header in zip file and return the offset of the
 251    start of the actual data.  Return -1 on error.  OFFSET is the
 252    offset from the beginning of the zip file of the file's header.  */
 253 static long
 254 find_zip_file_start (int fd, long offset)
 255 {
 256   int filename_length, extra_field_length;
 257   unsigned char buffer[LREC_SIZE + 4];
 258
 259   if (lseek (fd, offset, SEEK_SET) < 0)
 260     return -1;
 261
 262   if (read (fd, buffer, LREC_SIZE + 4) != LREC_SIZE + 4)
 263     return -1;
 264
 265   if (buffer[0] != 'P' || strncmp ((const char *) &buffer[1], LOCAL_HDR_SIG, 3))
 266     return -1;
 267
 268   filename_length = makeword (&buffer[4 + L_FILENAME_LENGTH]);
 269   extra_field_length = makeword (&buffer[4 + L_EXTRA_FIELD_LENGTH]);
 270
 271   return offset + (4 + LREC_SIZE) + filename_length + extra_field_length;
 272 }
 273
 274 int
 275 read_zip_archive (ZipFile *zipf)
 276 {
 277   int i;
 278   int dir_last_pad;
 279   char *dir_ptr;
 280   char buffer[100];
 281
 282   zipf->size = lseek (zipf->fd, 0L, SEEK_END);
 283
 284   if (zipf->size < (ECREC_SIZE+4) || lseek (zipf->fd, (long)(-(ECREC_SIZE+4)), SEEK_CUR) <= 0)
 285     return -1;
 286   if (read (zipf->fd, buffer, ECREC_SIZE+4) != ECREC_SIZE+4)
 287     return -2;
 288   if (buffer[0] != 'P'
 289       || strncmp ((const char *) &buffer[1], END_CENTRAL_SIG, 3))
 290     {
 291       /* We could not find the end-central-header signature, probably
 292          because a zipfile comment is present. Scan backwards until we
 293          find the signature. */
 294       if (lseek (zipf->fd, (long)(-ECREC_SIZE), SEEK_END) <= 0)
 295         return -2;
 296       while (buffer[0] != 'P'
 297              || strncmp ((const char *) &buffer[1], END_CENTRAL_SIG, 3))
 298         {
 299           if (lseek (zipf->fd, -5, SEEK_CUR) < 0)
 300             return -2;
 301           if (read (zipf->fd, buffer, 4) != 4)
 302             return -2;
 303         }
 304       if (read (zipf->fd, buffer + 4, ECREC_SIZE) != ECREC_SIZE)
 305         return -2;
 306     }
 307   zipf->count = makeword((const uch *) &buffer[TOTAL_ENTRIES_CENTRAL_DIR]);
 308   zipf->dir_size = makelong((const uch *) &buffer[SIZE_CENTRAL_DIRECTORY]);
 309   /* Allocate 1 more to allow appending '\0' to last filename. */
 310   zipf->central_directory = XNEWVEC (char, zipf->dir_size + 1);
 311   if (lseek (zipf->fd, -(zipf->dir_size+ECREC_SIZE+4), SEEK_CUR) < 0)
 312     return -2;
 313   if (read (zipf->fd, zipf->central_directory, zipf->dir_size) < 0)
 314     return -2;
 315
 316 #ifdef TEST
 317   printf ("number_this_disk = %d\n", makeword(&buffer[NUMBER_THIS_DISK]));
 318   printf ("num_disk_with_start_central_dir = %d\n", makeword(&buffer[NUM_DISK_WITH_START_CENTRAL_DIR]));
 319
 320   printf ("num_entries_centrl_dir_ths_disk = %d\n",
 321         makeword(&buffer[NUM_ENTRIES_CENTRL_DIR_THS_DISK]));
 322   printf ("total_entries_central_dir = %d\n",
 323         makeword(&buffer[TOTAL_ENTRIES_CENTRAL_DIR]));
 324   printf ("size_central_directory = %d\n",
 325         makelong((const uch *) &buffer[SIZE_CENTRAL_DIRECTORY]));
 326   printf ("offset_start_central_directory = %d\n",
 327         makelong((const uch *) &buffer[OFFSET_START_CENTRAL_DIRECTORY]));
 328   printf ("zipfile_comment_length = %d\n",
 329         makeword(&buffer[ZIPFILE_COMMENT_LENGTH]));
 330 #endif
 331
 332   dir_last_pad = 0;
 333   dir_ptr = zipf->central_directory;
 334   for (i = 0; i < zipf->count; i++)
 335     {
 336       ZipDirectory *zipd = (ZipDirectory*)(dir_ptr + dir_last_pad);
 337       int compression_method = (int) dir_ptr[4+C_COMPRESSION_METHOD];
 338       long size = makelong ((const uch *) &dir_ptr[4+C_COMPRESSED_SIZE]);
 339       long uncompressed_size = makelong ((const uch *) &dir_ptr[4+C_UNCOMPRESSED_SIZE]);
 340       long filename_length = makeword ((const uch *) &dir_ptr[4+C_FILENAME_LENGTH]);
 341       long extra_field_length = makeword ((const uch *) &dir_ptr[4+C_EXTRA_FIELD_LENGTH]);
 342       long file_offset = makelong ((const uch *) &dir_ptr[4+C_RELATIVE_OFFSET_LOCAL_HEADER]);
 343       int unpadded_direntry_length;
 344       if ((dir_ptr-zipf->central_directory)+filename_length+CREC_SIZE+4>zipf->dir_size)
 345         return -1;
 346
 347       zipd->filename_length = filename_length;
 348       zipd->compression_method = compression_method;
 349       zipd->size = size;
 350       zipd->uncompressed_size = uncompressed_size;
 351       zipd->zipf = zipf;
 352 #ifdef __GNUC__
 353 #define DIR_ALIGN __alignof__(ZipDirectory)
 354 #else
 355 #define DIR_ALIGN sizeof(long)
 356 #endif
 357       zipd->filestart = find_zip_file_start (zipf->fd, file_offset);
 358       zipd->filename_offset = CREC_SIZE+4 - dir_last_pad;
 359       unpadded_direntry_length
 360           = zipd->filename_offset + zipd->filename_length + extra_field_length;
 361       zipd->direntry_size =
 362         ((unpadded_direntry_length + DIR_ALIGN) / DIR_ALIGN) * DIR_ALIGN;
 363       dir_last_pad = zipd->direntry_size - unpadded_direntry_length;
 364       dir_ptr = (char*)zipd + unpadded_direntry_length;
 365       *dir_ptr = '\0';
 366     }
 367   return 0;
 368 }
 369
 370 #ifdef TEST
 371 main (void)
 372 {
 373   ZipFile zipf[1];
 374   ZipDirectory *zipd;
 375   int i;
 376
 377   zipf->fd = 0;
 378
 379   i = read_zip_archive (zipf);
 380   if (i)
 381     {
 382       fprintf (stderr, "Bad zip file.\n");
 383       exit (i);
 384     }
 385
 386   zipd = (ZipDirectory*) zipf->central_directory;
 387   for (i = 0; i < zipf->count; i++, zipd = ZIPDIR_NEXT (zipd))
 388     {
 389       printf ("%d: size:%d, name(#%d)%s, offset:%d\n",
 390               i, zipd->size, zipd->filename_length,
 391               ZIPDIR_FILENAME (zipd),
 392               zipd->filestart);
 393     }
 394 }
 395 #endif