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