includes/MacBinary.php

   1 <?php
   2 /**
   3  * MacBinary signature checker and data fork extractor, for files
   4  * uploaded from Internet Explorer for Mac.
   5  *
   6  * Copyright (C) 2005 Brion Vibber <brion@pobox.com>
   7  * Portions based on Convert::BinHex by Eryq et al
   8  * http://www.mediawiki.org/
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License along
  21  * with this program; if not, write to the Free Software Foundation, Inc.,
  22  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  *
  25  * @file
  26  */
  27
  28 class MacBinary {
  29         function __construct( $filename ) {
  30                 $this->open( $filename );
  31                 $this->loadHeader();
  32         }
  33
  34         private $valid, $version, $filename, $dataLength, $resourceLength, $handle;
  35
  36         /**
  37          * The file must be seekable, such as local filesystem.
  38          * Remote URLs probably won't work.
  39          *
  40          * @param $filename String
  41          */
  42         function open( $filename ) {
  43                 $this->valid = false;
  44                 $this->version = 0;
  45                 $this->filename = '';
  46                 $this->dataLength = 0;
  47                 $this->resourceLength = 0;
  48                 $this->handle = fopen( $filename, 'rb' );
  49         }
  50
  51         /**
  52          * Does this appear to be a valid MacBinary archive?
  53          *
  54          * @return Boolean
  55          */
  56         function isValid() {
  57                 return $this->valid;
  58         }
  59
  60         /**
  61          * Get length of data fork
  62          *
  63          * @return Integer
  64          */
  65         function dataForkLength() {
  66                 return $this->dataLength;
  67         }
  68
  69         /**
  70          * Copy the data fork to an external file or resource.
  71          *
  72          * @param $destination Ressource
  73          * @return Boolean
  74          */
  75         function extractData( $destination ) {
  76                 if( !$this->isValid() ) {
  77                         return false;
  78                 }
  79
  80                 // Data fork appears immediately after header
  81                 fseek( $this->handle, 128 );
  82                 return $this->copyBytesTo( $destination, $this->dataLength );
  83         }
  84
  85         /**
  86          *
  87          */
  88         function close() {
  89                 fclose( $this->handle );
  90         }
  91
  92         // --------------------------------------------------------------
  93
  94         /**
  95          * Check if the given file appears to be MacBinary-encoded,
  96          * as Internet Explorer on Mac OS may provide for unknown types.
  97          * http://www.lazerware.com/formats/macbinary/macbinary_iii.html
  98          * If ok, load header data.
  99          *
 100          * @return bool
 101          * @access private
 102          */
 103         function loadHeader() {
 104                 $fname = 'MacBinary::loadHeader';
 105
 106                 fseek( $this->handle, 0 );
 107                 $head = fread( $this->handle, 128 );
 108                 #$this->hexdump( $head );
 109
 110                 if( strlen( $head ) < 128 ) {
 111                         wfDebug( "$fname: couldn't read full MacBinary header\n" );
 112                         return false;
 113                 }
 114
 115                 if( $head[0] != "\x00" || $head[74] != "\x00" ) {
 116                         wfDebug( "$fname: header bytes 0 and 74 not null\n" );
 117                         return false;
 118                 }
 119
 120                 $signature = substr( $head, 102, 4 );
 121                 $a = unpack( "ncrc", substr( $head, 124, 2 ) );
 122                 $storedCRC = $a['crc'];
 123                 $calculatedCRC = $this->calcCRC( substr( $head, 0, 124 ) );
 124                 if( $storedCRC == $calculatedCRC ) {
 125                         if( $signature == 'mBIN' ) {
 126                                 $this->version = 3;
 127                         } else {
 128                                 $this->version = 2;
 129                         }
 130                 } else {
 131                         $crc = sprintf( "%x != %x", $storedCRC, $calculatedCRC );
 132                         if( $storedCRC == 0 && $head[82] == "\x00" &&
 133                                 substr( $head, 101, 24 ) == str_repeat( "\x00", 24 ) ) {
 134                                 wfDebug( "$fname: no CRC, looks like MacBinary I\n" );
 135                                 $this->version = 1;
 136                         } elseif( $signature == 'mBIN' && $storedCRC == 0x185 ) {
 137                                 // Mac IE 5.0 seems to insert this value in the CRC field.
 138                                 // 5.2.3 works correctly; don't know about other versions.
 139                                 wfDebug( "$fname: CRC doesn't match ($crc), looks like Mac IE 5.0\n" );
 140                                 $this->version = 3;
 141                         } else {
 142                                 wfDebug( "$fname: CRC doesn't match ($crc) and not MacBinary I\n" );
 143                                 return false;
 144                         }
 145                 }
 146
 147                 $nameLength = ord( $head[1] );
 148                 if( $nameLength < 1 || $nameLength > 63 ) {
 149                         wfDebug( "$fname: invalid filename size $nameLength\n" );
 150                         return false;
 151                 }
 152                 $this->filename = substr( $head, 2, $nameLength );
 153
 154                 $forks = unpack( "Ndata/Nresource", substr( $head, 83, 8 ) );
 155                 $this->dataLength = $forks['data'];
 156                 $this->resourceLength = $forks['resource'];
 157                 $maxForkLength = 0x7fffff;
 158
 159                 if( $this->dataLength < 0 || $this->dataLength > $maxForkLength ) {
 160                         wfDebug( "$fname: invalid data fork length $this->dataLength\n" );
 161                         return false;
 162                 }
 163
 164                 if( $this->resourceLength < 0 || $this->resourceLength > $maxForkLength ) {
 165                         wfDebug( "$fname: invalid resource fork size $this->resourceLength\n" );
 166                         return false;
 167                 }
 168
 169                 wfDebug( "$fname: appears to be MacBinary $this->version, data length $this->dataLength\n" );
 170                 $this->valid = true;
 171                 return true;
 172         }
 173
 174         /**
 175          * Calculate a 16-bit CRC value as for MacBinary headers.
 176          * Adapted from perl5 Convert::BinHex by Eryq,
 177          * based on the mcvert utility (Doug Moore, April '87),
 178          * with magic array thingy by Jim Van Verth.
 179          * http://search.cpan.org/~eryq/Convert-BinHex-1.119/lib/Convert/BinHex.pm
 180          *
 181          * @param $data String
 182          * @param $seed Integer
 183          * @return Integer
 184          * @access private
 185          */
 186         function calcCRC( $data, $seed = 0 ) {
 187                 # An array useful for CRC calculations that use 0x1021 as the "seed":
 188                 $MAGIC = array(
 189                         0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50a5, 0x60c6, 0x70e7,
 190                         0x8108, 0x9129, 0xa14a, 0xb16b, 0xc18c, 0xd1ad, 0xe1ce, 0xf1ef,
 191                         0x1231, 0x0210, 0x3273, 0x2252, 0x52b5, 0x4294, 0x72f7, 0x62d6,
 192                         0x9339, 0x8318, 0xb37b, 0xa35a, 0xd3bd, 0xc39c, 0xf3ff, 0xe3de,
 193                         0x2462, 0x3443, 0x0420, 0x1401, 0x64e6, 0x74c7, 0x44a4, 0x5485,
 194                         0xa56a, 0xb54b, 0x8528, 0x9509, 0xe5ee, 0xf5cf, 0xc5ac, 0xd58d,
 195                         0x3653, 0x2672, 0x1611, 0x0630, 0x76d7, 0x66f6, 0x5695, 0x46b4,
 196                         0xb75b, 0xa77a, 0x9719, 0x8738, 0xf7df, 0xe7fe, 0xd79d, 0xc7bc,
 197                         0x48c4, 0x58e5, 0x6886, 0x78a7, 0x0840, 0x1861, 0x2802, 0x3823,
 198                         0xc9cc, 0xd9ed, 0xe98e, 0xf9af, 0x8948, 0x9969, 0xa90a, 0xb92b,
 199                         0x5af5, 0x4ad4, 0x7ab7, 0x6a96, 0x1a71, 0x0a50, 0x3a33, 0x2a12,
 200                         0xdbfd, 0xcbdc, 0xfbbf, 0xeb9e, 0x9b79, 0x8b58, 0xbb3b, 0xab1a,
 201                         0x6ca6, 0x7c87, 0x4ce4, 0x5cc5, 0x2c22, 0x3c03, 0x0c60, 0x1c41,
 202                         0xedae, 0xfd8f, 0xcdec, 0xddcd, 0xad2a, 0xbd0b, 0x8d68, 0x9d49,
 203                         0x7e97, 0x6eb6, 0x5ed5, 0x4ef4, 0x3e13, 0x2e32, 0x1e51, 0x0e70,
 204                         0xff9f, 0xefbe, 0xdfdd, 0xcffc, 0xbf1b, 0xaf3a, 0x9f59, 0x8f78,
 205                         0x9188, 0x81a9, 0xb1ca, 0xa1eb, 0xd10c, 0xc12d, 0xf14e, 0xe16f,
 206                         0x1080, 0x00a1, 0x30c2, 0x20e3, 0x5004, 0x4025, 0x7046, 0x6067,
 207                         0x83b9, 0x9398, 0xa3fb, 0xb3da, 0xc33d, 0xd31c, 0xe37f, 0xf35e,
 208                         0x02b1, 0x1290, 0x22f3, 0x32d2, 0x4235, 0x5214, 0x6277, 0x7256,
 209                         0xb5ea, 0xa5cb, 0x95a8, 0x8589, 0xf56e, 0xe54f, 0xd52c, 0xc50d,
 210                         0x34e2, 0x24c3, 0x14a0, 0x0481, 0x7466, 0x6447, 0x5424, 0x4405,
 211                         0xa7db, 0xb7fa, 0x8799, 0x97b8, 0xe75f, 0xf77e, 0xc71d, 0xd73c,
 212                         0x26d3, 0x36f2, 0x0691, 0x16b0, 0x6657, 0x7676, 0x4615, 0x5634,
 213                         0xd94c, 0xc96d, 0xf90e, 0xe92f, 0x99c8, 0x89e9, 0xb98a, 0xa9ab,
 214                         0x5844, 0x4865, 0x7806, 0x6827, 0x18c0, 0x08e1, 0x3882, 0x28a3,
 215                         0xcb7d, 0xdb5c, 0xeb3f, 0xfb1e, 0x8bf9, 0x9bd8, 0xabbb, 0xbb9a,
 216                         0x4a75, 0x5a54, 0x6a37, 0x7a16, 0x0af1, 0x1ad0, 0x2ab3, 0x3a92,
 217                         0xfd2e, 0xed0f, 0xdd6c, 0xcd4d, 0xbdaa, 0xad8b, 0x9de8, 0x8dc9,
 218                         0x7c26, 0x6c07, 0x5c64, 0x4c45, 0x3ca2, 0x2c83, 0x1ce0, 0x0cc1,
 219                         0xef1f, 0xff3e, 0xcf5d, 0xdf7c, 0xaf9b, 0xbfba, 0x8fd9, 0x9ff8,
 220                         0x6e17, 0x7e36, 0x4e55, 0x5e74, 0x2e93, 0x3eb2, 0x0ed1, 0x1ef0
 221                 );
 222                 $len = strlen( $data );
 223                 $crc = $seed;
 224                 for( $i = 0; $i < $len; $i++ ) {
 225                         $crc ^= ord( $data[$i] ) << 8;
 226                         $crc &= 0xFFFF;
 227                         $crc = ($crc << 8) ^ $MAGIC[$crc >> 8];
 228                         $crc &= 0xFFFF;
 229                 }
 230                 return $crc;
 231         }
 232
 233         /**
 234          * @param $destination Resource
 235          * @param $bytesToCopy Integer
 236          * @return Boolean
 237          * @access private
 238          */
 239         function copyBytesTo( $destination, $bytesToCopy ) {
 240                 $bufferSize = 65536;
 241                 for( $remaining = $bytesToCopy; $remaining > 0; $remaining -= $bufferSize ) {
 242                         $thisChunkSize = min( $remaining, $bufferSize );
 243                         $buffer = fread( $this->handle, $thisChunkSize );
 244                         fwrite( $destination, $buffer );
 245                 }
 246         }
 247
 248         /**
 249          * Hex dump of the header for debugging
 250          * @access private
 251          */
 252         function hexdump( $data ) {
 253                 global $wgDebugLogFile;
 254                 if( !$wgDebugLogFile ) return;
 255
 256                 $width = 16;
 257                 $at = 0;
 258                 for( $remaining = strlen( $data ); $remaining > 0; $remaining -= $width ) {
 259                         $line = sprintf( "%04x:", $at );
 260                         $printable = '';
 261                         for( $i = 0; $i < $width && $remaining - $i > 0; $i++ ) {
 262                                 $byte = ord( $data[$at++] );
 263                                 $line .= sprintf( " %02x", $byte );
 264                                 $printable .= ($byte >= 32 && $byte <= 126 )
 265                                         ? chr( $byte )
 266                                         : '.';
 267                         }
 268                         if( $i < $width ) {
 269                                 $line .= str_repeat( '   ', $width - $i );
 270                         }
 271                         wfDebug( "MacBinary: $line $printable\n" );
 272                 }
 273         }
 274 }