includes/MimeMagic.php

   1 <?php
   2 /** Module defining helper functions for detecting and dealing with mime types.
   3  *
   4  */
   5
   6  /** Defines a set of well known mime types
   7  * This is used as a fallback to mime.types files.
   8  * An extensive list of well known mime types is provided by
   9  * the file mime.types in the includes directory.
  10  */
  11 define('MM_WELL_KNOWN_MIME_TYPES',<<<END_STRING
  12 application/ogg ogg ogm ogv
  13 application/pdf pdf
  14 application/vnd.oasis.opendocument.chart odc
  15 application/vnd.oasis.opendocument.chart-template otc
  16 application/vnd.oasis.opendocument.formula odf
  17 application/vnd.oasis.opendocument.formula-template otf
  18 application/vnd.oasis.opendocument.graphics odg
  19 application/vnd.oasis.opendocument.graphics-template otg
  20 application/vnd.oasis.opendocument.image odi
  21 application/vnd.oasis.opendocument.image-template oti
  22 application/vnd.oasis.opendocument.presentation odp
  23 application/vnd.oasis.opendocument.presentation-template otp
  24 application/vnd.oasis.opendocument.spreadsheet ods
  25 application/vnd.oasis.opendocument.spreadsheet-template ots
  26 application/vnd.oasis.opendocument.text odt
  27 application/vnd.oasis.opendocument.text-template ott
  28 application/vnd.oasis.opendocument.text-master otm
  29 application/vnd.oasis.opendocument.text-web oth
  30 application/x-javascript js
  31 application/x-shockwave-flash swf
  32 audio/midi mid midi kar
  33 audio/mpeg mpga mpa mp2 mp3
  34 audio/x-aiff aif aiff aifc
  35 audio/x-wav wav
  36 audio/ogg ogg
  37 image/x-bmp bmp
  38 image/gif gif
  39 image/jpeg jpeg jpg jpe
  40 image/png png
  41 image/svg+xml image/svg svg
  42 image/tiff tiff tif
  43 image/vnd.djvu image/x.djvu image/x-djvu djvu
  44 image/x-portable-pixmap ppm
  45 image/x-xcf xcf
  46 text/plain txt
  47 text/html html htm
  48 video/ogg ogm ogg ogv
  49 video/mpeg mpg mpeg
  50 END_STRING
  51 );
  52
  53  /** Defines a set of well known mime info entries
  54  * This is used as a fallback to mime.info files.
  55  * An extensive list of well known mime types is provided by
  56  * the file mime.info in the includes directory.
  57  */
  58 define('MM_WELL_KNOWN_MIME_INFO', <<<END_STRING
  59 application/pdf [OFFICE]
  60 application/vnd.oasis.opendocument.chart [OFFICE]
  61 application/vnd.oasis.opendocument.chart-template [OFFICE]
  62 application/vnd.oasis.opendocument.formula [OFFICE]
  63 application/vnd.oasis.opendocument.formula-template [OFFICE]
  64 application/vnd.oasis.opendocument.graphics [OFFICE]
  65 application/vnd.oasis.opendocument.graphics-template [OFFICE]
  66 application/vnd.oasis.opendocument.image [OFFICE]
  67 application/vnd.oasis.opendocument.image-template [OFFICE]
  68 application/vnd.oasis.opendocument.presentation [OFFICE]
  69 application/vnd.oasis.opendocument.presentation-template [OFFICE]
  70 application/vnd.oasis.opendocument.spreadsheet [OFFICE]
  71 application/vnd.oasis.opendocument.spreadsheet-template [OFFICE]
  72 application/vnd.oasis.opendocument.text [OFFICE]
  73 application/vnd.oasis.opendocument.text-template [OFFICE]
  74 application/vnd.oasis.opendocument.text-master [OFFICE]
  75 application/vnd.oasis.opendocument.text-web [OFFICE]
  76 text/javascript application/x-javascript [EXECUTABLE]
  77 application/x-shockwave-flash [MULTIMEDIA]
  78 audio/midi [AUDIO]
  79 audio/x-aiff [AUDIO]
  80 audio/x-wav [AUDIO]
  81 audio/mp3 audio/mpeg [AUDIO]
  82 application/ogg audio/ogg video/ogg [MULTIMEDIA]
  83 image/x-bmp image/bmp [BITMAP]
  84 image/gif [BITMAP]
  85 image/jpeg [BITMAP]
  86 image/png [BITMAP]
  87 image/svg+xml [DRAWING]
  88 image/tiff [BITMAP]
  89 image/vnd.djvu [BITMAP]
  90 image/x-xcf [BITMAP]
  91 image/x-portable-pixmap [BITMAP]
  92 text/plain [TEXT]
  93 text/html [TEXT]
  94 video/ogg [VIDEO]
  95 video/mpeg [VIDEO]
  96 unknown/unknown application/octet-stream application/x-empty [UNKNOWN]
  97 END_STRING
  98 );
  99
 100 #note: because this file is possibly included by a function,
 101 #we need to access the global scope explicitely!
 102 global $wgLoadFileinfoExtension;
 103
 104 if ($wgLoadFileinfoExtension) {
 105         if(!extension_loaded('fileinfo')) dl('fileinfo.' . PHP_SHLIB_SUFFIX);
 106 }
 107
 108 /**
 109  * Implements functions related to mime types such as detection and mapping to
 110  * file extension.
 111  *
 112  * Instances of this class are stateles, there only needs to be one global instance
 113  * of MimeMagic. Please use MimeMagic::singleton() to get that instance.
 114  */
 115 class MimeMagic {
 116
 117         /**
 118         * Mapping of media types to arrays of mime types.
 119         * This is used by findMediaType and getMediaType, respectively
 120         */
 121         var $mMediaTypes= NULL;
 122
 123         /** Map of mime type aliases
 124         */
 125         var $mMimeTypeAliases= NULL;
 126
 127         /** map of mime types to file extensions (as a space seprarated list)
 128         */
 129         var $mMimeToExt= NULL;
 130
 131         /** map of file extensions types to mime types (as a space seprarated list)
 132         */
 133         var $mExtToMime= NULL;
 134
 135         /** IEContentAnalyzer instance
 136          */
 137         var $mIEAnalyzer;
 138
 139         /** The singleton instance
 140          */
 141         private static $instance;
 142
 143         /** Initializes the MimeMagic object. This is called by MimeMagic::singleton().
 144         *
 145         * This constructor parses the mime.types and mime.info files and build internal mappings.
 146         */
 147         function __construct() {
 148                 /*
 149                 *   --- load mime.types ---
 150                 */
 151
 152                 global $wgMimeTypeFile, $IP;
 153
 154                 $types = MM_WELL_KNOWN_MIME_TYPES;
 155
 156                 if ( $wgMimeTypeFile == 'includes/mime.types' ) {
 157                         $wgMimeTypeFile = "$IP/$wgMimeTypeFile";
 158                 }
 159
 160                 if ( $wgMimeTypeFile ) {
 161                         if ( is_file( $wgMimeTypeFile ) and is_readable( $wgMimeTypeFile ) ) {
 162                                 wfDebug( __METHOD__.": loading mime types from $wgMimeTypeFile\n" );
 163                                 $types .= "\n";
 164                                 $types .= file_get_contents( $wgMimeTypeFile );
 165                         } else {
 166                                 wfDebug( __METHOD__.": can't load mime types from $wgMimeTypeFile\n" );
 167                         }
 168                 } else {
 169                         wfDebug( __METHOD__.": no mime types file defined, using build-ins only.\n" );
 170                 }
 171
 172                 $types = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $types );
 173                 $types = str_replace( "\t", " ", $types );
 174
 175                 $this->mMimeToExt = array();
 176                 $this->mToMime = array();
 177
 178                 $lines = explode( "\n",$types );
 179                 foreach ( $lines as $s ) {
 180                         $s = trim( $s );
 181                         if ( empty( $s ) ) continue;
 182                         if ( strpos( $s, '#' ) === 0 ) continue;
 183
 184                         $s = strtolower( $s );
 185                         $i = strpos( $s, ' ' );
 186
 187                         if ( $i === false ) continue;
 188
 189                         #print "processing MIME line $s<br>";
 190
 191                         $mime = substr( $s, 0, $i );
 192                         $ext = trim( substr($s, $i+1 ) );
 193
 194                         if ( empty( $ext ) ) continue;
 195
 196                         if ( !empty( $this->mMimeToExt[$mime] ) ) {
 197                                 $this->mMimeToExt[$mime] .= ' ' . $ext;
 198                         } else {
 199                                 $this->mMimeToExt[$mime] = $ext;
 200                         }
 201
 202                         $extensions = explode( ' ', $ext );
 203
 204                         foreach ( $extensions as $e ) {
 205                                 $e = trim( $e );
 206                                 if ( empty( $e ) ) continue;
 207
 208                                 if ( !empty( $this->mExtToMime[$e] ) ) {
 209                                         $this->mExtToMime[$e] .= ' ' . $mime;
 210                                 } else {
 211                                         $this->mExtToMime[$e] = $mime;
 212                                 }
 213                         }
 214                 }
 215
 216                 /*
 217                 *   --- load mime.info ---
 218                 */
 219
 220                 global $wgMimeInfoFile;
 221                 if ( $wgMimeInfoFile == 'includes/mime.info' ) {
 222                         $wgMimeInfoFile = "$IP/$wgMimeInfoFile";
 223                 }
 224
 225                 $info = MM_WELL_KNOWN_MIME_INFO;
 226
 227                 if ( $wgMimeInfoFile ) {
 228                         if ( is_file( $wgMimeInfoFile ) and is_readable( $wgMimeInfoFile ) ) {
 229                                 wfDebug( __METHOD__.": loading mime info from $wgMimeInfoFile\n" );
 230                                 $info .= "\n";
 231                                 $info .= file_get_contents( $wgMimeInfoFile );
 232                         } else {
 233                                 wfDebug(__METHOD__.": can't load mime info from $wgMimeInfoFile\n");
 234                         }
 235                 } else {
 236                         wfDebug(__METHOD__.": no mime info file defined, using build-ins only.\n");
 237                 }
 238
 239                 $info = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $info);
 240                 $info = str_replace( "\t", " ", $info );
 241
 242                 $this->mMimeTypeAliases = array();
 243                 $this->mMediaTypes = array();
 244
 245                 $lines = explode( "\n", $info );
 246                 foreach ( $lines as $s ) {
 247                         $s = trim( $s );
 248                         if ( empty( $s ) ) continue;
 249                         if ( strpos( $s, '#' ) === 0 ) continue;
 250
 251                         $s = strtolower( $s );
 252                         $i = strpos( $s, ' ' );
 253
 254                         if ( $i === false ) continue;
 255
 256                         #print "processing MIME INFO line $s<br>";
 257
 258                         $match = array();
 259                         if ( preg_match( '!\[\s*(\w+)\s*\]!', $s, $match ) ) {
 260                                 $s = preg_replace( '!\[\s*(\w+)\s*\]!', '', $s );
 261                                 $mtype = trim( strtoupper( $match[1] ) );
 262                         } else {
 263                                 $mtype = MEDIATYPE_UNKNOWN;
 264                         }
 265
 266                         $m = explode( ' ', $s );
 267
 268                         if ( !isset( $this->mMediaTypes[$mtype] ) ) {
 269                                 $this->mMediaTypes[$mtype] = array();
 270                         }
 271
 272                         foreach ( $m as $mime ) {
 273                                 $mime = trim( $mime );
 274                                 if ( empty( $mime ) ) continue;
 275
 276                                 $this->mMediaTypes[$mtype][] = $mime;
 277                         }
 278
 279                         if ( sizeof( $m ) > 1 ) {
 280                                 $main = $m[0];
 281                                 for ( $i=1; $i<sizeof($m); $i += 1 ) {
 282                                         $mime = $m[$i];
 283                                         $this->mMimeTypeAliases[$mime] = $main;
 284                                 }
 285                         }
 286                 }
 287
 288         }
 289
 290         /**
 291          * Get an instance of this class
 292          */
 293         static function &singleton() {
 294                 if ( !isset( self::$instance ) ) {
 295                         self::$instance = new MimeMagic;
 296                 }
 297                 return self::$instance;
 298         }
 299
 300         /** returns a list of file extensions for a given mime type
 301         * as a space separated string.
 302         */
 303         function getExtensionsForType( $mime ) {
 304                 $mime = strtolower( $mime );
 305
 306                 $r = @$this->mMimeToExt[$mime];
 307
 308                 if ( @!$r and isset( $this->mMimeTypeAliases[$mime] ) ) {
 309                         $mime = $this->mMimeTypeAliases[$mime];
 310                         $r = @$this->mMimeToExt[$mime];
 311                 }
 312
 313                 return $r;
 314         }
 315
 316         /** returns a list of mime types for a given file extension
 317         * as a space separated string.
 318         */
 319         function getTypesForExtension( $ext ) {
 320                 $ext = strtolower( $ext );
 321
 322                 $r = isset( $this->mExtToMime[$ext] ) ? $this->mExtToMime[$ext] : null;
 323                 return $r;
 324         }
 325
 326         /** returns a single mime type for a given file extension.
 327         * This is always the first type from the list returned by getTypesForExtension($ext).
 328         */
 329         function guessTypesForExtension( $ext ) {
 330                 $m = $this->getTypesForExtension( $ext );
 331                 if ( is_null( $m ) ) return NULL;
 332
 333                 $m = trim( $m );
 334                 $m = preg_replace( '/\s.*$/', '', $m );
 335
 336                 return $m;
 337         }
 338
 339
 340         /** tests if the extension matches the given mime type.
 341         * returns true if a match was found, NULL if the mime type is unknown,
 342         * and false if the mime type is known but no matches where found.
 343         */
 344         function isMatchingExtension( $extension, $mime ) {
 345                 $ext = $this->getExtensionsForType( $mime );
 346
 347                 if ( !$ext ) {
 348                         return NULL;  //unknown
 349                 }
 350
 351                 $ext = explode( ' ', $ext );
 352
 353                 $extension = strtolower( $extension );
 354                 if ( in_array( $extension, $ext ) ) {
 355                         return true;
 356                 }
 357
 358                 return false;
 359         }
 360
 361         /** returns true if the mime type is known to represent
 362         * an image format supported by the PHP GD library.
 363         */
 364         function isPHPImageType( $mime ) {
 365                 #as defined by imagegetsize and image_type_to_mime
 366                 static $types = array(
 367                         'image/gif', 'image/jpeg', 'image/png',
 368                         'image/x-bmp', 'image/xbm', 'image/tiff',
 369                         'image/jp2', 'image/jpeg2000', 'image/iff',
 370                         'image/xbm', 'image/x-xbitmap',
 371                         'image/vnd.wap.wbmp', 'image/vnd.xiff',
 372                         'image/x-photoshop',
 373                         'application/x-shockwave-flash',
 374                 );
 375
 376                 return in_array( $mime, $types );
 377         }
 378
 379         /**
 380          * Returns true if the extension represents a type which can
 381          * be reliably detected from its content. Use this to determine
 382          * whether strict content checks should be applied to reject
 383          * invalid uploads; if we can't identify the type we won't
 384          * be able to say if it's invalid.
 385          *
 386          * @todo Be more accurate when using fancy mime detector plugins;
 387          *       right now this is the bare minimum getimagesize() list.
 388          * @return bool
 389          */
 390         function isRecognizableExtension( $extension ) {
 391                 static $types = array(
 392                         // Types recognized by getimagesize()
 393                         'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
 394                         'bmp', 'tiff', 'tif', 'jpc', 'jp2',
 395                         'jpx', 'jb2', 'swc', 'iff', 'wbmp',
 396                         'xbm',
 397
 398                         // Formats we recognize magic numbers for
 399                         'djvu', 'ogg', 'ogv', 'mid', 'pdf', 'wmf', 'xcf',
 400
 401                         // XML formats we sure hope we recognize reliably
 402                         'svg',
 403                 );
 404                 return in_array( strtolower( $extension ), $types );
 405         }
 406
 407
 408         /** mime type detection. This uses detectMimeType to detect the mime type of the file,
 409         * but applies additional checks to determine some well known file formats that may be missed
 410         * or misinterpreter by the default mime detection (namely xml based formats like XHTML or SVG).
 411         *
 412         * @param string $file The file to check
 413         * @param mixed $ext The file extension, or true to extract it from the filename.
 414         *                   Set it to false to ignore the extension.
 415         *
 416         * @return string the mime type of $file
 417         */
 418         function guessMimeType( $file, $ext = true ) {
 419                 $mime = $this->doGuessMimeType( $file, $ext );
 420
 421                 if( !$mime ) {
 422                         wfDebug( __METHOD__.": internal type detection failed for $file (.$ext)...\n" );
 423                         $mime = $this->detectMimeType( $file, $ext );
 424                 }
 425
 426                 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
 427                         $mime = $this->mMimeTypeAliases[$mime];
 428                 }
 429
 430                 wfDebug(__METHOD__.": final mime type of $file: $mime\n");
 431                 return $mime;
 432         }
 433
 434         function doGuessMimeType( $file, $ext = true ) {
 435                 // Read a chunk of the file
 436                 wfSuppressWarnings();
 437                 $f = fopen( $file, "rt" );
 438                 wfRestoreWarnings();
 439                 if( !$f ) return "unknown/unknown";
 440                 $head = fread( $f, 1024 );
 441                 fseek( $f, -65558, SEEK_END );
 442                 $tail = fread( $f, 65558 ); // 65558 = maximum size of a zip EOCDR
 443                 fclose( $f );
 444
 445                 // Hardcode a few magic number checks...
 446                 $headers = array(
 447                         // Multimedia...
 448                         'MThd'             => 'audio/midi',
 449                         'OggS'             => 'application/ogg',
 450
 451                         // Image formats...
 452                         // Note that WMF may have a bare header, no magic number.
 453                         "\x01\x00\x09\x00" => 'application/x-msmetafile', // Possibly prone to false positives?
 454                         "\xd7\xcd\xc6\x9a" => 'application/x-msmetafile',
 455                         '%PDF'             => 'application/pdf',
 456                         'gimp xcf'         => 'image/x-xcf',
 457
 458                         // Some forbidden fruit...
 459                         'MZ'               => 'application/octet-stream', // DOS/Windows executable
 460                         "\xca\xfe\xba\xbe" => 'application/octet-stream', // Mach-O binary
 461                         "\x7fELF"          => 'application/octet-stream', // ELF binary
 462                 );
 463
 464                 foreach( $headers as $magic => $candidate ) {
 465                         if( strncmp( $head, $magic, strlen( $magic ) ) == 0 ) {
 466                                 wfDebug( __METHOD__ . ": magic header in $file recognized as $candidate\n" );
 467                                 return $candidate;
 468                         }
 469                 }
 470
 471                 /*
 472                  * Look for PHP.  Check for this before HTML/XML...  Warning: this is a
 473                  * heuristic, and won't match a file with a lot of non-PHP before.  It
 474                  * will also match text files which could be PHP. :)
 475                  *
 476                  * FIXME: For this reason, the check is probably useless -- an attacker
 477                  * could almost certainly just pad the file with a lot of nonsense to
 478                  * circumvent the check in any case where it would be a security
 479                  * problem.  On the other hand, it causes harmful false positives (bug
 480                  * 16583).  The heuristic has been cut down to exclude three-character
 481                  * strings like "<? ", but should it be axed completely?
 482                  */
 483                 if( ( strpos( $head, '<?php' ) !== false ) ||
 484
 485                     ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
 486                     ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
 487                     ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
 488                     ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
 489                     ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {
 490
 491                         wfDebug( __METHOD__ . ": recognized $file as application/x-php\n" );
 492                         return "application/x-php";
 493                 }
 494
 495                 /*
 496                  * look for XML formats (XHTML and SVG)
 497                  */
 498                 $xml = new XmlTypeCheck( $file );
 499                 if( $xml->wellFormed ) {
 500                         global $wgXMLMimeTypes;
 501                         if( isset( $wgXMLMimeTypes[$xml->getRootElement()] ) ) {
 502                                 return $wgXMLMimeTypes[$xml->getRootElement()];
 503                         } else {
 504                                 return 'application/xml';
 505                         }
 506                 }
 507
 508                 /*
 509                  * look for shell scripts
 510                  */
 511                 $script_type = NULL;
 512
 513                 # detect by shebang
 514                 if ( substr( $head, 0, 2) == "#!" ) {
 515                         $script_type = "ASCII";
 516                 } elseif ( substr( $head, 0, 5) == "\xef\xbb\xbf#!" ) {
 517                         $script_type = "UTF-8";
 518                 } elseif ( substr( $head, 0, 7) == "\xfe\xff\x00#\x00!" ) {
 519                         $script_type = "UTF-16BE";
 520                 } elseif ( substr( $head, 0, 7 ) == "\xff\xfe#\x00!" ) {
 521                         $script_type= "UTF-16LE";
 522                 }
 523
 524                 if ( $script_type ) {
 525                         if ( $script_type !== "UTF-8" && $script_type !== "ASCII") {
 526                                 // Quick and dirty fold down to ASCII!
 527                                 $pack = array( 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' );
 528                                 $chars = unpack( $pack[$script_type], substr( $head, 2 ) );
 529                                 $head = '';
 530                                 foreach( $chars as $codepoint ) {
 531                                         if( $codepoint < 128 ) {
 532                                                 $head .= chr( $codepoint );
 533                                         } else {
 534                                                 $head .= '?';
 535                                         }
 536                                 }
 537                         }
 538
 539                         $match = array();
 540
 541                         if ( preg_match( '%/?([^\s]+/)(\w+)%', $head, $match ) ) {
 542                                 $mime = "application/x-{$match[2]}";
 543                                 wfDebug( __METHOD__.": shell script recognized as $mime\n" );
 544                                 return $mime;
 545                         }
 546                 }
 547
 548                 // Check for ZIP (before getimagesize)
 549                 if ( strpos( $tail, "PK\x05\x06" ) !== false ) {
 550                         wfDebug( __METHOD__.": ZIP header present at end of $file\n" );
 551                         return $this->detectZipType( $head );
 552                 }
 553
 554                 wfSuppressWarnings();
 555                 $gis = getimagesize( $file );
 556                 wfRestoreWarnings();
 557
 558                 if( $gis && isset( $gis['mime'] ) ) {
 559                         $mime = $gis['mime'];
 560                         wfDebug( __METHOD__.": getimagesize detected $file as $mime\n" );
 561                         return $mime;
 562                 }
 563
 564                 // Also test DjVu
 565                 $deja = new DjVuImage( $file );
 566                 if( $deja->isValid() ) {
 567                         wfDebug( __METHOD__.": detected $file as image/vnd.djvu\n" );
 568                         return 'image/vnd.djvu';
 569                 }
 570
 571                 return false;
 572         }
 573
 574         /**
 575          * Detect application-specific file type of a given ZIP file from its
 576          * header data.  Currently works for OpenDocument types...
 577          * If can't tell, returns 'application/zip'.
 578          *
 579          * @param string $header Some reasonably-sized chunk of file header
 580          * @return string
 581          */
 582         function detectZipType( $header ) {
 583                 $opendocTypes = array(
 584                         'chart-template',
 585                         'chart',
 586                         'formula-template',
 587                         'formula',
 588                         'graphics-template',
 589                         'graphics',
 590                         'image-template',
 591                         'image',
 592                         'presentation-template',
 593                         'presentation',
 594                         'spreadsheet-template',
 595                         'spreadsheet',
 596                         'text-template',
 597                         'text-master',
 598                         'text-web',
 599                         'text' );
 600
 601                 // http://lists.oasis-open.org/archives/office/200505/msg00006.html
 602                 $types = '(?:' . implode( '|', $opendocTypes ) . ')';
 603                 $opendocRegex = "/^mimetype(application\/vnd\.oasis\.opendocument\.$types)/";
 604                 wfDebug( __METHOD__.": $opendocRegex\n" );
 605
 606                 if( preg_match( $opendocRegex, substr( $header, 30 ), $matches ) ) {
 607                         $mime = $matches[1];
 608                         wfDebug( __METHOD__.": detected $mime from ZIP archive\n" );
 609                         return $mime;
 610                 } else {
 611                         wfDebug( __METHOD__.": unable to identify type of ZIP archive\n" );
 612                         return 'application/zip';
 613                 }
 614         }
 615
 616         /** Internal mime type detection, please use guessMimeType() for application code instead.
 617         * Detection is done using an external program, if $wgMimeDetectorCommand is set.
 618         * Otherwise, the fileinfo extension and mime_content_type are tried (in this order), if they are available.
 619         * If the dections fails and $ext is not false, the mime type is guessed from the file extension, using
 620         * guessTypesForExtension.
 621         * If the mime type is still unknown, getimagesize is used to detect the mime type if the file is an image.
 622         * If no mime type can be determined, this function returns "unknown/unknown".
 623         *
 624         * @param string $file The file to check
 625         * @param mixed $ext The file extension, or true to extract it from the filename.
 626         *                   Set it to false to ignore the extension.
 627         *
 628         * @return string the mime type of $file
 629         * @access private
 630         */
 631         function detectMimeType( $file, $ext = true ) {
 632                 global $wgMimeDetectorCommand;
 633
 634                 $m = NULL;
 635                 if ( $wgMimeDetectorCommand ) {
 636                         $fn = wfEscapeShellArg( $file );
 637                         $m = `$wgMimeDetectorCommand $fn`;
 638                 } elseif ( function_exists( "finfo_open" ) && function_exists( "finfo_file" ) ) {
 639
 640                         # This required the fileinfo extension by PECL,
 641                         # see http://pecl.php.net/package/fileinfo
 642                         # This must be compiled into PHP
 643                         #
 644                         # finfo is the official replacement for the deprecated
 645                         # mime_content_type function, see below.
 646                         #
 647                         # If you may need to load the fileinfo extension at runtime, set
 648                         # $wgLoadFileinfoExtension in LocalSettings.php
 649
 650                         $mime_magic_resource = finfo_open(FILEINFO_MIME); /* return mime type ala mimetype extension */
 651
 652                         if ($mime_magic_resource) {
 653                                 $m = finfo_file( $mime_magic_resource, $file );
 654                                 finfo_close( $mime_magic_resource );
 655                         } else {
 656                                 wfDebug( __METHOD__.": finfo_open failed on ".FILEINFO_MIME."!\n" );
 657                         }
 658                 } elseif ( function_exists( "mime_content_type" ) ) {
 659
 660                         # NOTE: this function is available since PHP 4.3.0, but only if
 661                         # PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
 662                         #
 663                         # On Windows, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
 664                         # sometimes, this may even be needed under linus/unix.
 665                         #
 666                         # Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
 667                         # see http://www.php.net/manual/en/ref.mime-magic.php for details.
 668
 669                         $m = mime_content_type($file);
 670                 } else {
 671                         wfDebug( __METHOD__.": no magic mime detector found!\n" );
 672                 }
 673
 674                 if ( $m ) {
 675                         # normalize
 676                         $m = preg_replace( '![;, ].*$!', '', $m ); #strip charset, etc
 677                         $m = trim( $m );
 678                         $m = strtolower( $m );
 679
 680                         if ( strpos( $m, 'unknown' ) !== false ) {
 681                                 $m = NULL;
 682                         } else {
 683                                 wfDebug( __METHOD__.": magic mime type of $file: $m\n" );
 684                                 return $m;
 685                         }
 686                 }
 687
 688                 # if desired, look at extension as a fallback.
 689                 if ( $ext === true ) {
 690                         $i = strrpos( $file, '.' );
 691                         $ext = strtolower( $i ? substr( $file, $i + 1 ) : '' );
 692                 }
 693                 if ( $ext ) {
 694                         if( $this->isRecognizableExtension( $ext ) ) {
 695                                 wfDebug( __METHOD__. ": refusing to guess mime type for .$ext file, we should have recognized it\n" );
 696                         } else {
 697                                 $m = $this->guessTypesForExtension( $ext );
 698                                 if ( $m ) {
 699                                         wfDebug( __METHOD__.": extension mime type of $file: $m\n" );
 700                                         return $m;
 701                                 }
 702                         }
 703                 }
 704
 705                 #unknown type
 706                 wfDebug( __METHOD__.": failed to guess mime type for $file!\n" );
 707                 return "unknown/unknown";
 708         }
 709
 710         /**
 711         * Determine the media type code for a file, using its mime type, name and possibly
 712         * its contents.
 713         *
 714         * This function relies on the findMediaType(), mapping extensions and mime
 715         * types to media types.
 716         *
 717         * @todo analyse file if need be
 718         * @todo look at multiple extension, separately and together.
 719         *
 720         * @param string $path full path to the image file, in case we have to look at the contents
 721         *        (if null, only the mime type is used to determine the media type code).
 722         * @param string $mime mime type. If null it will be guessed using guessMimeType.
 723         *
 724         * @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
 725         */
 726         function getMediaType( $path = NULL, $mime = NULL ) {
 727                 if( !$mime && !$path ) return MEDIATYPE_UNKNOWN;
 728
 729                 # If mime type is unknown, guess it
 730                 if( !$mime ) $mime = $this->guessMimeType( $path, false );
 731
 732                 # Special code for ogg - detect if it's video (theora),
 733                 # else label it as sound.
 734                 if( $mime == "application/ogg" && file_exists( $path ) ) {
 735
 736                         // Read a chunk of the file
 737                         $f = fopen( $path, "rt" );
 738                         if ( !$f ) return MEDIATYPE_UNKNOWN;
 739                         $head = fread( $f, 256 );
 740                         fclose( $f );
 741
 742                         $head = strtolower( $head );
 743
 744                         # This is an UGLY HACK, file should be parsed correctly
 745                         if ( strpos( $head, 'theora' ) !== false ) return MEDIATYPE_VIDEO;
 746                         elseif ( strpos( $head, 'vorbis' ) !== false ) return MEDIATYPE_AUDIO;
 747                         elseif ( strpos( $head, 'flac' ) !== false ) return MEDIATYPE_AUDIO;
 748                         elseif ( strpos( $head, 'speex' ) !== false ) return MEDIATYPE_AUDIO;
 749                         else return MEDIATYPE_MULTIMEDIA;
 750                 }
 751
 752                 # check for entry for full mime type
 753                 if( $mime ) {
 754                         $type = $this->findMediaType( $mime );
 755                         if( $type !== MEDIATYPE_UNKNOWN ) return $type;
 756                 }
 757
 758                 # Check for entry for file extension
 759                 $e = NULL;
 760                 if ( $path ) {
 761                         $i = strrpos( $path, '.' );
 762                         $e = strtolower( $i ? substr( $path, $i + 1 ) : '' );
 763
 764                         # TODO: look at multi-extension if this fails, parse from full path
 765
 766                         $type = $this->findMediaType( '.' . $e );
 767                         if ( $type !== MEDIATYPE_UNKNOWN ) return $type;
 768                 }
 769
 770                 # Check major mime type
 771                 if( $mime ) {
 772                         $i = strpos( $mime, '/' );
 773                         if( $i !== false ) {
 774                                 $major = substr( $mime, 0, $i );
 775                                 $type = $this->findMediaType( $major );
 776                                 if( $type !== MEDIATYPE_UNKNOWN ) return $type;
 777                         }
 778                 }
 779
 780                 if( !$type ) $type = MEDIATYPE_UNKNOWN;
 781
 782                 return $type;
 783         }
 784
 785         /** returns a media code matching the given mime type or file extension.
 786         * File extensions are represented by a string starting with a dot (.) to
 787         * distinguish them from mime types.
 788         *
 789         * This funktion relies on the mapping defined by $this->mMediaTypes
 790         * @access private
 791         */
 792         function findMediaType( $extMime ) {
 793                 if ( strpos( $extMime, '.' ) === 0 ) { #if it's an extension, look up the mime types
 794                         $m = $this->getTypesForExtension( substr( $extMime, 1 ) );
 795                         if ( !$m ) return MEDIATYPE_UNKNOWN;
 796
 797                         $m = explode( ' ', $m );
 798                 } else {
 799                         # Normalize mime type
 800                         if ( isset( $this->mMimeTypeAliases[$extMime] ) ) {
 801                                 $extMime = $this->mMimeTypeAliases[$extMime];
 802                         }
 803
 804                         $m = array($extMime);
 805                 }
 806
 807                 foreach ( $m as $mime ) {
 808                         foreach ( $this->mMediaTypes as $type => $codes ) {
 809                                 if ( in_array($mime, $codes, true ) ) {
 810                                         return $type;
 811                                 }
 812                         }
 813                 }
 814
 815                 return MEDIATYPE_UNKNOWN;
 816         }
 817
 818         /**
 819          * Get the MIME types that various versions of Internet Explorer would
 820          * detect from a chunk of the content.
 821          *
 822          * @param string $fileName The file name (unused at present)
 823          * @param string $chunk The first 256 bytes of the file
 824          * @param string $proposed The MIME type proposed by the server
 825          */
 826         public function getIEMimeTypes( $fileName, $chunk, $proposed ) {
 827                 $ca = $this->getIEContentAnalyzer();
 828                 return $ca->getRealMimesFromData( $fileName, $chunk, $proposed );
 829         }
 830
 831         /**
 832          * Get a cached instance of IEContentAnalyzer
 833          */
 834         protected function getIEContentAnalyzer() {
 835                 if ( is_null( $this->mIEAnalyzer ) ) {
 836                         $this->mIEAnalyzer = new IEContentAnalyzer;
 837                 }
 838                 return $this->mIEAnalyzer;
 839         }
 840 }