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
  13 application/pdf pdf
  14 application/x-javascript js
  15 application/x-shockwave-flash swf
  16 audio/midi mid midi kar
  17 audio/mpeg mpga mpa mp2 mp3
  18 audio/x-aiff aif aiff aifc
  19 audio/x-wav wav
  20 audio/ogg ogg
  21 image/x-bmp bmp
  22 image/gif gif
  23 image/jpeg jpeg jpg jpe
  24 image/png png
  25 image/svg+xml image/svg svg
  26 image/tiff tiff tif
  27 image/vnd.djvu djvu
  28 image/x-portable-pixmap ppm
  29 text/plain txt
  30 text/html html htm
  31 video/ogg ogm ogg
  32 video/mpeg mpg mpeg
  33 END_STRING
  34 );
  35
  36  /** Defines a set of well known mime info entries
  37  * This is used as a fallback to mime.info files.
  38  * An extensive list of well known mime types is provided by
  39  * the file mime.info in the includes directory.
  40  */
  41 define('MM_WELL_KNOWN_MIME_INFO', <<<END_STRING
  42 application/pdf [OFFICE]
  43 text/javascript application/x-javascript [EXECUTABLE]
  44 application/x-shockwave-flash [MULTIMEDIA]
  45 audio/midi [AUDIO]
  46 audio/x-aiff [AUDIO]
  47 audio/x-wav [AUDIO]
  48 audio/mp3 audio/mpeg [AUDIO]
  49 application/ogg audio/ogg video/ogg [MULTIMEDIA]
  50 image/x-bmp image/bmp [BITMAP]
  51 image/gif [BITMAP]
  52 image/jpeg [BITMAP]
  53 image/png [BITMAP]
  54 image/svg+xml [DRAWING]
  55 image/tiff [BITMAP]
  56 image/vnd.djvu [BITMAP]
  57 image/x-portable-pixmap [BITMAP]
  58 text/plain [TEXT]
  59 text/html [TEXT]
  60 video/ogg [VIDEO]
  61 video/mpeg [VIDEO]
  62 unknown/unknown application/octet-stream application/x-empty [UNKNOWN]
  63 END_STRING
  64 );
  65
  66 #note: because this file is possibly included by a function,
  67 #we need to access the global scope explicitely!
  68 global $wgLoadFileinfoExtension;
  69
  70 if ($wgLoadFileinfoExtension) {
  71         if(!extension_loaded('fileinfo')) dl('fileinfo.' . PHP_SHLIB_SUFFIX);
  72 }
  73
  74 /**
  75  * Implements functions related to mime types such as detection and mapping to
  76  * file extension.
  77  *
  78  * Instances of this class are stateles, there only needs to be one global instance
  79  * of MimeMagic. Please use MimeMagic::singleton() to get that instance.
  80  */
  81 class MimeMagic {
  82
  83         /**
  84         * Mapping of media types to arrays of mime types.
  85         * This is used by findMediaType and getMediaType, respectively
  86         */
  87         var $mMediaTypes= NULL;
  88
  89         /** Map of mime type aliases
  90         */
  91         var $mMimeTypeAliases= NULL;
  92
  93         /** map of mime types to file extensions (as a space seprarated list)
  94         */
  95         var $mMimeToExt= NULL;
  96
  97         /** map of file extensions types to mime types (as a space seprarated list)
  98         */
  99         var $mExtToMime= NULL;
 100
 101         /** The singleton instance
 102          */
 103         private static $instance;
 104
 105         /** Initializes the MimeMagic object. This is called by MimeMagic::singleton().
 106         *
 107         * This constructor parses the mime.types and mime.info files and build internal mappings.
 108         */
 109         function __construct() {
 110                 /*
 111                 *   --- load mime.types ---
 112                 */
 113
 114                 global $wgMimeTypeFile, $IP;
 115
 116                 $types = MM_WELL_KNOWN_MIME_TYPES;
 117
 118                 if ( $wgMimeTypeFile == 'includes/mime.types' ) {
 119                         $wgMimeTypeFile = "$IP/$wgMimeTypeFile";
 120                 }
 121
 122                 if ( $wgMimeTypeFile ) {
 123                         if ( is_file( $wgMimeTypeFile ) and is_readable( $wgMimeTypeFile ) ) {
 124                                 wfDebug( __METHOD__.": loading mime types from $wgMimeTypeFile\n" );
 125                                 $types .= "\n";
 126                                 $types .= file_get_contents( $wgMimeTypeFile );
 127                         } else {
 128                                 wfDebug( __METHOD__.": can't load mime types from $wgMimeTypeFile\n" );
 129                         }
 130                 } else {
 131                         wfDebug( __METHOD__.": no mime types file defined, using build-ins only.\n" );
 132                 }
 133
 134                 $types = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $types );
 135                 $types = str_replace( "\t", " ", $types );
 136
 137                 $this->mMimeToExt = array();
 138                 $this->mToMime = array();
 139
 140                 $lines = explode( "\n",$types );
 141                 foreach ( $lines as $s ) {
 142                         $s = trim( $s );
 143                         if ( empty( $s ) ) continue;
 144                         if ( strpos( $s, '#' ) === 0 ) continue;
 145
 146                         $s = strtolower( $s );
 147                         $i = strpos( $s, ' ' );
 148
 149                         if ( $i === false ) continue;
 150
 151                         #print "processing MIME line $s<br>";
 152
 153                         $mime = substr( $s, 0, $i );
 154                         $ext = trim( substr($s, $i+1 ) );
 155
 156                         if ( empty( $ext ) ) continue;
 157
 158                         if ( !empty( $this->mMimeToExt[$mime] ) ) {
 159                                 $this->mMimeToExt[$mime] .= ' ' . $ext;
 160                         } else {
 161                                 $this->mMimeToExt[$mime] = $ext;
 162                         }
 163
 164                         $extensions = explode( ' ', $ext );
 165
 166                         foreach ( $extensions as $e ) {
 167                                 $e = trim( $e );
 168                                 if ( empty( $e ) ) continue;
 169
 170                                 if ( !empty( $this->mExtToMime[$e] ) ) {
 171                                         $this->mExtToMime[$e] .= ' ' . $mime;
 172                                 } else {
 173                                         $this->mExtToMime[$e] = $mime;
 174                                 }
 175                         }
 176                 }
 177
 178                 /*
 179                 *   --- load mime.info ---
 180                 */
 181
 182                 global $wgMimeInfoFile;
 183                 if ( $wgMimeInfoFile == 'includes/mime.info' ) {
 184                         $wgMimeInfoFile = "$IP/$wgMimeInfoFile";
 185                 }
 186
 187                 $info = MM_WELL_KNOWN_MIME_INFO;
 188
 189                 if ( $wgMimeInfoFile ) {
 190                         if ( is_file( $wgMimeInfoFile ) and is_readable( $wgMimeInfoFile ) ) {
 191                                 wfDebug( __METHOD__.": loading mime info from $wgMimeInfoFile\n" );
 192                                 $info .= "\n";
 193                                 $info .= file_get_contents( $wgMimeInfoFile );
 194                         } else {
 195                                 wfDebug(__METHOD__.": can't load mime info from $wgMimeInfoFile\n");
 196                         }
 197                 } else {
 198                         wfDebug(__METHOD__.": no mime info file defined, using build-ins only.\n");
 199                 }
 200
 201                 $info = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $info);
 202                 $info = str_replace( "\t", " ", $info );
 203
 204                 $this->mMimeTypeAliases = array();
 205                 $this->mMediaTypes = array();
 206
 207                 $lines = explode( "\n", $info );
 208                 foreach ( $lines as $s ) {
 209                         $s = trim( $s );
 210                         if ( empty( $s ) ) continue;
 211                         if ( strpos( $s, '#' ) === 0 ) continue;
 212
 213                         $s = strtolower( $s );
 214                         $i = strpos( $s, ' ' );
 215
 216                         if ( $i === false ) continue;
 217
 218                         #print "processing MIME INFO line $s<br>";
 219
 220                         $match = array();
 221                         if ( preg_match( '!\[\s*(\w+)\s*\]!', $s, $match ) ) {
 222                                 $s = preg_replace( '!\[\s*(\w+)\s*\]!', '', $s );
 223                                 $mtype = trim( strtoupper( $match[1] ) );
 224                         } else {
 225                                 $mtype = MEDIATYPE_UNKNOWN;
 226                         }
 227
 228                         $m = explode( ' ', $s );
 229
 230                         if ( !isset( $this->mMediaTypes[$mtype] ) ) {
 231                                 $this->mMediaTypes[$mtype] = array();
 232                         }
 233
 234                         foreach ( $m as $mime ) {
 235                                 $mime = trim( $mime );
 236                                 if ( empty( $mime ) ) continue;
 237
 238                                 $this->mMediaTypes[$mtype][] = $mime;
 239                         }
 240
 241                         if ( sizeof( $m ) > 1 ) {
 242                                 $main = $m[0];
 243                                 for ( $i=1; $i<sizeof($m); $i += 1 ) {
 244                                         $mime = $m[$i];
 245                                         $this->mMimeTypeAliases[$mime] = $main;
 246                                 }
 247                         }
 248                 }
 249
 250         }
 251
 252         /**
 253          * Get an instance of this class
 254          */
 255         static function &singleton() {
 256                 if ( !isset( self::$instance ) ) {
 257                         self::$instance = new MimeMagic;
 258                 }
 259                 return self::$instance;
 260         }
 261
 262         /** returns a list of file extensions for a given mime type
 263         * as a space separated string.
 264         */
 265         function getExtensionsForType( $mime ) {
 266                 $mime = strtolower( $mime );
 267
 268                 $r = @$this->mMimeToExt[$mime];
 269
 270                 if ( @!$r and isset( $this->mMimeTypeAliases[$mime] ) ) {
 271                         $mime = $this->mMimeTypeAliases[$mime];
 272                         $r = @$this->mMimeToExt[$mime];
 273                 }
 274
 275                 return $r;
 276         }
 277
 278         /** returns a list of mime types for a given file extension
 279         * as a space separated string.
 280         */
 281         function getTypesForExtension( $ext ) {
 282                 $ext = strtolower( $ext );
 283
 284                 $r = isset( $this->mExtToMime[$ext] ) ? $this->mExtToMime[$ext] : null;
 285                 return $r;
 286         }
 287
 288         /** returns a single mime type for a given file extension.
 289         * This is always the first type from the list returned by getTypesForExtension($ext).
 290         */
 291         function guessTypesForExtension( $ext ) {
 292                 $m = $this->getTypesForExtension( $ext );
 293                 if ( is_null( $m ) ) return NULL;
 294
 295                 $m = trim( $m );
 296                 $m = preg_replace( '/\s.*$/', '', $m );
 297
 298                 return $m;
 299         }
 300
 301
 302         /** tests if the extension matches the given mime type.
 303         * returns true if a match was found, NULL if the mime type is unknown,
 304         * and false if the mime type is known but no matches where found.
 305         */
 306         function isMatchingExtension( $extension, $mime ) {
 307                 $ext = $this->getExtensionsForType( $mime );
 308
 309                 if ( !$ext ) {
 310                         return NULL;  //unknown
 311                 }
 312
 313                 $ext = explode( ' ', $ext );
 314
 315                 $extension = strtolower( $extension );
 316                 if ( in_array( $extension, $ext ) ) {
 317                         return true;
 318                 }
 319
 320                 return false;
 321         }
 322
 323         /** returns true if the mime type is known to represent
 324         * an image format supported by the PHP GD library.
 325         */
 326         function isPHPImageType( $mime ) {
 327                 #as defined by imagegetsize and image_type_to_mime
 328                 static $types = array(
 329                         'image/gif', 'image/jpeg', 'image/png',
 330                         'image/x-bmp', 'image/xbm', 'image/tiff',
 331                         'image/jp2', 'image/jpeg2000', 'image/iff',
 332                         'image/xbm', 'image/x-xbitmap',
 333                         'image/vnd.wap.wbmp', 'image/vnd.xiff',
 334                         'image/x-photoshop',
 335                         'application/x-shockwave-flash',
 336                 );
 337
 338                 return in_array( $mime, $types );
 339         }
 340
 341         /**
 342          * Returns true if the extension represents a type which can
 343          * be reliably detected from its content. Use this to determine
 344          * whether strict content checks should be applied to reject
 345          * invalid uploads; if we can't identify the type we won't
 346          * be able to say if it's invalid.
 347          *
 348          * @todo Be more accurate when using fancy mime detector plugins;
 349          *       right now this is the bare minimum getimagesize() list.
 350          * @return bool
 351          */
 352         function isRecognizableExtension( $extension ) {
 353                 static $types = array(
 354                         'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
 355                         'bmp', 'tiff', 'tif', 'jpc', 'jp2',
 356                         'jpx', 'jb2', 'swc', 'iff', 'wbmp',
 357                         'xbm', 'djvu'
 358                 );
 359                 return in_array( strtolower( $extension ), $types );
 360         }
 361
 362
 363         /** mime type detection. This uses detectMimeType to detect the mime type of the file,
 364         * but applies additional checks to determine some well known file formats that may be missed
 365         * or misinterpreter by the default mime detection (namely xml based formats like XHTML or SVG).
 366         *
 367         * @param string $file The file to check
 368         * @param mixed $ext The file extension, or true to extract it from the filename.
 369         *                   Set it to false to ignore the extension.
 370         *
 371         * @return string the mime type of $file
 372         */
 373         function guessMimeType( $file, $ext = true ) {
 374                 $mime = $this->detectMimeType( $file, $ext );
 375
 376                 // Read a chunk of the file
 377                 $f = fopen( $file, "rt" );
 378                 if( !$f ) return "unknown/unknown";
 379                 $head = fread( $f, 1024 );
 380                 fclose( $f );
 381
 382                 $sub4 =  substr( $head, 0, 4 );
 383                 if ( $sub4 == "\x01\x00\x09\x00" || $sub4 == "\xd7\xcd\xc6\x9a" ) {
 384                         // WMF kill kill kill
 385                         // Note that WMF may have a bare header, no magic number.
 386                         // The former of the above two checks is theoretically prone to false positives
 387                         $mime = "application/x-msmetafile";
 388                 }
 389
 390                 if ( strpos( $mime, "text/" ) === 0 || $mime === "application/xml" ) {
 391
 392                         $xml_type = NULL;
 393                         $script_type = NULL;
 394
 395                         /*
 396                         * look for XML formats (XHTML and SVG)
 397                         */
 398                         if ($mime === "text/sgml" ||
 399                             $mime === "text/plain" ||
 400                             $mime === "text/html" ||
 401                             $mime === "text/xml" ||
 402                             $mime === "application/xml") {
 403
 404                                 if ( substr( $head, 0, 5 ) == "<?xml" ) {
 405                                         $xml_type = "ASCII";
 406                                 } elseif ( substr( $head, 0, 8 ) == "\xef\xbb\xbf<?xml") {
 407                                         $xml_type = "UTF-8";
 408                                 } elseif ( substr( $head, 0, 10 ) == "\xfe\xff\x00<\x00?\x00x\x00m\x00l" ) {
 409                                         $xml_type = "UTF-16BE";
 410                                 } elseif ( substr( $head, 0, 10 ) == "\xff\xfe<\x00?\x00x\x00m\x00l\x00") {
 411                                         $xml_type = "UTF-16LE";
 412                                 }
 413
 414                                 if ( $xml_type ) {
 415                                         if ( $xml_type !== "UTF-8" && $xml_type !== "ASCII" ) {
 416                                                 $head = iconv( $xml_type, "ASCII//IGNORE", $head );
 417                                         }
 418
 419                                         $match = array();
 420                                         $doctype = "";
 421                                         $tag = "";
 422
 423                                         if ( preg_match( '%<!DOCTYPE\s+[\w-]+\s+PUBLIC\s+["'."'".'"](.*?)["'."'".'"].*>%sim',
 424                                                 $head, $match ) ) {
 425                                                         $doctype = $match[1];
 426                                                 }
 427                                         if ( preg_match( '%<(\w+).*>%sim', $head, $match ) ) {
 428                                                 $tag = $match[1];
 429                                         }
 430
 431                                         #print "<br>ANALYSING $file ($mime): doctype= $doctype; tag= $tag<br>";
 432
 433                                         if ( strpos( $doctype, "-//W3C//DTD SVG" ) === 0 ) {
 434                                                 $mime = "image/svg+xml";
 435                                         } elseif ( $tag === "svg" ) {
 436                                                 $mime = "image/svg+xml";
 437                                         } elseif ( strpos( $doctype, "-//W3C//DTD XHTML" ) === 0 ) {
 438                                                 $mime = "text/html";
 439                                         } elseif ( $tag === "html" ) {
 440                                                 $mime = "text/html";
 441                                         }
 442                                 }
 443                         }
 444
 445                         /*
 446                         * look for shell scripts
 447                         */
 448                         if ( !$xml_type ) {
 449                                 $script_type = NULL;
 450
 451                                 # detect by shebang
 452                                 if ( substr( $head, 0, 2) == "#!" ) {
 453                                         $script_type = "ASCII";
 454                                 } elseif ( substr( $head, 0, 5) == "\xef\xbb\xbf#!" ) {
 455                                         $script_type = "UTF-8";
 456                                 } elseif ( substr( $head, 0, 7) == "\xfe\xff\x00#\x00!" ) {
 457                                         $script_type = "UTF-16BE";
 458                                 } elseif ( substr( $head, 0, 7 ) == "\xff\xfe#\x00!" ) {
 459                                         $script_type= "UTF-16LE";
 460                                 }
 461
 462                                 if ( $script_type ) {
 463                                         if ( $script_type !== "UTF-8" && $script_type !== "ASCII") {
 464                                                 $head = iconv( $script_type, "ASCII//IGNORE", $head);
 465                                         }
 466
 467                                         $match = array();
 468
 469                                         if ( preg_match( '%/?([^\s]+/)(\w+)%', $head, $match ) ) {
 470                                                 $mime = "application/x-{$match[2]}";
 471                                         }
 472                                 }
 473                         }
 474
 475                         /*
 476                         * look for PHP
 477                         */
 478                         if( !$xml_type && !$script_type ) {
 479
 480                                 if( ( strpos( $head, '<?php' ) !== false ) ||
 481                                     ( strpos( $head, '<? ' ) !== false ) ||
 482                                     ( strpos( $head, "<?\n" ) !== false ) ||
 483                                     ( strpos( $head, "<?\t" ) !== false ) ||
 484                                     ( strpos( $head, "<?=" ) !== false ) ||
 485
 486                                     ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
 487                                     ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
 488                                     ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
 489                                     ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
 490                                     ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {
 491
 492                                         $mime = "application/x-php";
 493                                 }
 494                         }
 495
 496                 }
 497
 498                 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
 499                         $mime = $this->mMimeTypeAliases[$mime];
 500                 }
 501
 502                 wfDebug(__METHOD__.": final mime type of $file: $mime\n");
 503                 return $mime;
 504         }
 505
 506         /** Internal mime type detection, please use guessMimeType() for application code instead.
 507         * Detection is done using an external program, if $wgMimeDetectorCommand is set.
 508         * Otherwise, the fileinfo extension and mime_content_type are tried (in this order), if they are available.
 509         * If the dections fails and $ext is not false, the mime type is guessed from the file extension, using
 510         * guessTypesForExtension.
 511         * If the mime type is still unknown, getimagesize is used to detect the mime type if the file is an image.
 512         * If no mime type can be determined, this function returns "unknown/unknown".
 513         *
 514         * @param string $file The file to check
 515         * @param mixed $ext The file extension, or true to extract it from the filename.
 516         *                   Set it to false to ignore the extension.
 517         *
 518         * @return string the mime type of $file
 519         * @access private
 520         */
 521         function detectMimeType( $file, $ext = true ) {
 522                 global $wgMimeDetectorCommand;
 523
 524                 $m = NULL;
 525                 if ( $wgMimeDetectorCommand ) {
 526                         $fn = wfEscapeShellArg( $file );
 527                         $m = `$wgMimeDetectorCommand $fn`;
 528                 } elseif ( function_exists( "finfo_open" ) && function_exists( "finfo_file" ) ) {
 529
 530                         # This required the fileinfo extension by PECL,
 531                         # see http://pecl.php.net/package/fileinfo
 532                         # This must be compiled into PHP
 533                         #
 534                         # finfo is the official replacement for the deprecated
 535                         # mime_content_type function, see below.
 536                         #
 537                         # If you may need to load the fileinfo extension at runtime, set
 538                         # $wgLoadFileinfoExtension in LocalSettings.php
 539
 540                         $mime_magic_resource = finfo_open(FILEINFO_MIME); /* return mime type ala mimetype extension */
 541
 542                         if ($mime_magic_resource) {
 543                                 $m = finfo_file( $mime_magic_resource, $file );
 544                                 finfo_close( $mime_magic_resource );
 545                         } else {
 546                                 wfDebug( __METHOD__.": finfo_open failed on ".FILEINFO_MIME."!\n" );
 547                         }
 548                 } elseif ( function_exists( "mime_content_type" ) ) {
 549
 550                         # NOTE: this function is available since PHP 4.3.0, but only if
 551                         # PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
 552                         #
 553                         # On Windows, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
 554                         # sometimes, this may even be needed under linus/unix.
 555                         #
 556                         # Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
 557                         # see http://www.php.net/manual/en/ref.mime-magic.php for details.
 558
 559                         $m = mime_content_type($file);
 560
 561                         if ( $m == 'text/plain' ) {
 562                                 // mime_content_type sometimes considers DJVU files to be text/plain.
 563                                 $deja = new DjVuImage( $file );
 564                                 if( $deja->isValid() ) {
 565                                         wfDebug( __METHOD__.": (re)detected $file as image/vnd.djvu\n" );
 566                                         $m = 'image/vnd.djvu';
 567                                 }
 568                         }
 569                 } else {
 570                         wfDebug( __METHOD__.": no magic mime detector found!\n" );
 571                 }
 572
 573                 if ( $m ) {
 574                         # normalize
 575                         $m = preg_replace( '![;, ].*$!', '', $m ); #strip charset, etc
 576                         $m = trim( $m );
 577                         $m = strtolower( $m );
 578
 579                         if ( strpos( $m, 'unknown' ) !== false ) {
 580                                 $m = NULL;
 581                         } else {
 582                                 wfDebug( __METHOD__.": magic mime type of $file: $m\n" );
 583                                 return $m;
 584                         }
 585                 }
 586
 587                 # if still not known, use getimagesize to find out the type of image
 588                 # TODO: skip things that do not have a well-known image extension? Would that be safe?
 589                 wfSuppressWarnings();
 590                 $gis = getimagesize( $file );
 591                 wfRestoreWarnings();
 592
 593                 $notAnImage = false;
 594
 595                 if ( $gis && is_array($gis) && $gis[2] ) {
 596
 597                         switch ( $gis[2] ) {
 598                                 case IMAGETYPE_GIF: $m = "image/gif"; break;
 599                                 case IMAGETYPE_JPEG: $m = "image/jpeg"; break;
 600                                 case IMAGETYPE_PNG: $m = "image/png"; break;
 601                                 case IMAGETYPE_SWF: $m = "application/x-shockwave-flash"; break;
 602                                 case IMAGETYPE_PSD: $m = "application/photoshop"; break;
 603                                 case IMAGETYPE_BMP: $m = "image/bmp"; break;
 604                                 case IMAGETYPE_TIFF_II: $m = "image/tiff"; break;
 605                                 case IMAGETYPE_TIFF_MM: $m = "image/tiff"; break;
 606                                 case IMAGETYPE_JPC: $m = "image"; break;
 607                                 case IMAGETYPE_JP2: $m = "image/jpeg2000"; break;
 608                                 case IMAGETYPE_JPX: $m = "image/jpeg2000"; break;
 609                                 case IMAGETYPE_JB2: $m = "image"; break;
 610                                 case IMAGETYPE_SWC: $m = "application/x-shockwave-flash"; break;
 611                                 case IMAGETYPE_IFF: $m = "image/vnd.xiff"; break;
 612                                 case IMAGETYPE_WBMP: $m = "image/vnd.wap.wbmp"; break;
 613                                 case IMAGETYPE_XBM: $m = "image/x-xbitmap"; break;
 614                         }
 615
 616                         if ( $m ) {
 617                                 wfDebug( __METHOD__.": image mime type of $file: $m\n" );
 618                                 return $m;
 619                         }
 620                         else {
 621                                 $notAnImage = true;
 622                         }
 623                 } else {
 624                         // Also test DjVu
 625                         $deja = new DjVuImage( $file );
 626                         if( $deja->isValid() ) {
 627                                 wfDebug( __METHOD__.": detected $file as image/vnd.djvu\n" );
 628                                 return 'image/vnd.djvu';
 629                         }
 630                 }
 631
 632                 # if desired, look at extension as a fallback.
 633                 if ( $ext === true ) {
 634                         $i = strrpos( $file, '.' );
 635                         $ext = strtolower( $i ? substr( $file, $i + 1 ) : '' );
 636                 }
 637                 if ( $ext ) {
 638                         $m = $this->guessTypesForExtension( $ext );
 639
 640                         # TODO: if $notAnImage is set, do not trust the file extension if
 641                         # the results is one of the image types that should have been recognized
 642                         # by getimagesize
 643
 644                         if ( $m ) {
 645                                 wfDebug( __METHOD__.": extension mime type of $file: $m\n" );
 646                                 return $m;
 647                         }
 648                 }
 649
 650                 #unknown type
 651                 wfDebug( __METHOD__.": failed to guess mime type for $file!\n" );
 652                 return "unknown/unknown";
 653         }
 654
 655         /**
 656         * Determine the media type code for a file, using its mime type, name and possibly
 657         * its contents.
 658         *
 659         * This function relies on the findMediaType(), mapping extensions and mime
 660         * types to media types.
 661         *
 662         * @todo analyse file if need be
 663         * @todo look at multiple extension, separately and together.
 664         *
 665         * @param string $path full path to the image file, in case we have to look at the contents
 666         *        (if null, only the mime type is used to determine the media type code).
 667         * @param string $mime mime type. If null it will be guessed using guessMimeType.
 668         *
 669         * @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
 670         */
 671         function getMediaType( $path = NULL, $mime = NULL ) {
 672                 if( !$mime && !$path ) return MEDIATYPE_UNKNOWN;
 673
 674                 # If mime type is unknown, guess it
 675                 if( !$mime ) $mime = $this->guessMimeType( $path, false );
 676
 677                 # Special code for ogg - detect if it's video (theora),
 678                 # else label it as sound.
 679                 if( $mime == "application/ogg" && file_exists( $path ) ) {
 680
 681                         // Read a chunk of the file
 682                         $f = fopen( $path, "rt" );
 683                         if ( !$f ) return MEDIATYPE_UNKNOWN;
 684                         $head = fread( $f, 256 );
 685                         fclose( $f );
 686
 687                         $head = strtolower( $head );
 688
 689                         # This is an UGLY HACK, file should be parsed correctly
 690                         if ( strpos( $head, 'theora' ) !== false ) return MEDIATYPE_VIDEO;
 691                         elseif ( strpos( $head, 'vorbis' ) !== false ) return MEDIATYPE_AUDIO;
 692                         elseif ( strpos( $head, 'flac' ) !== false ) return MEDIATYPE_AUDIO;
 693                         elseif ( strpos( $head, 'speex' ) !== false ) return MEDIATYPE_AUDIO;
 694                         else return MEDIATYPE_MULTIMEDIA;
 695                 }
 696
 697                 # check for entry for full mime type
 698                 if( $mime ) {
 699                         $type = $this->findMediaType( $mime );
 700                         if( $type !== MEDIATYPE_UNKNOWN ) return $type;
 701                 }
 702
 703                 # Check for entry for file extension
 704                 $e = NULL;
 705                 if ( $path ) {
 706                         $i = strrpos( $path, '.' );
 707                         $e = strtolower( $i ? substr( $path, $i + 1 ) : '' );
 708
 709                         # TODO: look at multi-extension if this fails, parse from full path
 710
 711                         $type = $this->findMediaType( '.' . $e );
 712                         if ( $type !== MEDIATYPE_UNKNOWN ) return $type;
 713                 }
 714
 715                 # Check major mime type
 716                 if( $mime ) {
 717                         $i = strpos( $mime, '/' );
 718                         if( $i !== false ) {
 719                                 $major = substr( $mime, 0, $i );
 720                                 $type = $this->findMediaType( $major );
 721                                 if( $type !== MEDIATYPE_UNKNOWN ) return $type;
 722                         }
 723                 }
 724
 725                 if( !$type ) $type = MEDIATYPE_UNKNOWN;
 726
 727                 return $type;
 728         }
 729
 730         /** returns a media code matching the given mime type or file extension.
 731         * File extensions are represented by a string starting with a dot (.) to
 732         * distinguish them from mime types.
 733         *
 734         * This funktion relies on the mapping defined by $this->mMediaTypes
 735         * @access private
 736         */
 737         function findMediaType( $extMime ) {
 738                 if ( strpos( $extMime, '.' ) === 0 ) { #if it's an extension, look up the mime types
 739                         $m = $this->getTypesForExtension( substr( $extMime, 1 ) );
 740                         if ( !$m ) return MEDIATYPE_UNKNOWN;
 741
 742                         $m = explode( ' ', $m );
 743                 } else {
 744                         # Normalize mime type
 745                         if ( isset( $this->mMimeTypeAliases[$extMime] ) ) {
 746                                 $extMime = $this->mMimeTypeAliases[$extMime];
 747                         }
 748
 749                         $m = array($extMime);
 750                 }
 751
 752                 foreach ( $m as $mime ) {
 753                         foreach ( $this->mMediaTypes as $type => $codes ) {
 754                                 if ( in_array($mime, $codes, true ) ) {
 755                                         return $type;
 756                                 }
 757                         }
 758                 }
 759
 760                 return MEDIATYPE_UNKNOWN;
 761         }
 762 }
 763
 764