includes/MimeMagic.php

   1 <?php
   2 /** Module defining helper functions for detecting and dealing with mime types.
   3  *
   4  * @package MediaWiki
   5  */
   6
   7  /** Defines a set of well known mime types
   8  * This is used as a fallback to mime.types files.
   9  * An extensive list of well known mime types is provided by
  10  * the file mime.types in the includes directory.
  11  */
  12 define('MM_WELL_KNOWN_MIME_TYPES',<<<END_STRING
  13 application/ogg ogg ogm
  14 application/pdf pdf
  15 application/x-javascript js
  16 application/x-shockwave-flash swf
  17 audio/midi mid midi kar
  18 audio/mpeg mpga mpa mp2 mp3
  19 audio/x-aiff aif aiff aifc
  20 audio/x-wav wav
  21 audio/ogg ogg
  22 image/x-bmp bmp
  23 image/gif gif
  24 image/jpeg jpeg jpg jpe
  25 image/png png
  26 image/svg+xml svg
  27 image/tiff tiff tif
  28 image/vnd.djvu djvu
  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 image/svg+xml [DRAWING]
  55 image/tiff [BITMAP]
  56 image/vnd.djvu [BITMAP]
  57 text/plain [TEXT]
  58 text/html [TEXT]
  59 video/ogg [VIDEO]
  60 video/mpeg [VIDEO]
  61 unknown/unknown application/octet-stream application/x-empty [UNKNOWN]
  62 END_STRING
  63 );
  64
  65 #note: because this file is possibly included by a function,
  66 #we need to access the global scope explicitely!
  67 global $wgLoadFileinfoExtension;
  68
  69 if ($wgLoadFileinfoExtension) {
  70         if(!extension_loaded('fileinfo')) dl('fileinfo.' . PHP_SHLIB_SUFFIX);
  71 }
  72
  73 /** Implements functions related to mime types such as detection and mapping to
  74 * file extension,
  75 *
  76 * Instances of this class are stateles, there only needs to be one global instance
  77 * of MimeMagic. Please use wfGetMimeMagic to get that instance.
  78 * @package MediaWiki
  79 */
  80 class MimeMagic {
  81
  82         /**
  83         * Mapping of media types to arrays of mime types.
  84         * This is used by findMediaType and getMediaType, respectively
  85         */
  86         var $mMediaTypes= NULL;
  87
  88         /** Map of mime type aliases
  89         */
  90         var $mMimeTypeAliases= NULL;
  91
  92         /** map of mime types to file extensions (as a space seprarated list)
  93         */
  94         var $mMimeToExt= NULL;
  95
  96         /** map of file extensions types to mime types (as a space seprarated list)
  97         */
  98         var $mExtToMime= NULL;
  99
 100         /** Initializes the MimeMagic object. This is called by wfGetMimeMagic when instantiation
 101         * the global MimeMagic singleton object.
 102         *
 103         * This constructor parses the mime.types and mime.info files and build internal mappings.
 104         */
 105         function MimeMagic() {
 106                 /*
 107                 *   --- load mime.types ---
 108                 */
 109
 110                 global $wgMimeTypeFile;
 111
 112                 $types= MM_WELL_KNOWN_MIME_TYPES;
 113
 114                 if ($wgMimeTypeFile) {
 115                         if (is_file($wgMimeTypeFile) and is_readable($wgMimeTypeFile)) {
 116                                 wfDebug("MimeMagic::MimeMagic: loading mime types from $wgMimeTypeFile\n");
 117
 118                                 $types.= "\n";
 119                                 $types.= file_get_contents($wgMimeTypeFile);
 120                         }
 121                         else wfDebug("MimeMagic::MimeMagic: can't load mime types from $wgMimeTypeFile\n");
 122                 }
 123                 else wfDebug("MimeMagic::MimeMagic: no mime types file defined, using build-ins only.\n");
 124
 125                 $types= str_replace(array("\r\n","\n\r","\n\n","\r\r","\r"),"\n",$types);
 126                 $types= str_replace("\t"," ",$types);
 127
 128                 $this->mMimeToExt= array();
 129                 $this->mToMime= array();
 130
 131                 $lines= explode("\n",$types);
 132                 foreach ($lines as $s) {
 133                         $s= trim($s);
 134                         if (empty($s)) continue;
 135                         if (strpos($s,'#')===0) continue;
 136
 137                         $s= strtolower($s);
 138                         $i= strpos($s,' ');
 139
 140                         if ($i===false) continue;
 141
 142                         #print "processing MIME line $s<br>";
 143
 144                         $mime= substr($s,0,$i);
 145                         $ext= trim(substr($s,$i+1));
 146
 147                         if (empty($ext)) continue;
 148
 149                         if (@$this->mMimeToExt[$mime]) $this->mMimeToExt[$mime] .= ' '.$ext;
 150                         else $this->mMimeToExt[$mime]= $ext;
 151
 152                         $extensions= explode(' ',$ext);
 153
 154                         foreach ($extensions as $e) {
 155                                 $e= trim($e);
 156                                 if (empty($e)) continue;
 157
 158                                 if (@$this->mExtToMime[$e]) $this->mExtToMime[$e] .= ' '.$mime;
 159                                 else $this->mExtToMime[$e]= $mime;
 160                         }
 161                 }
 162
 163                 /*
 164                 *   --- load mime.info ---
 165                 */
 166
 167                 global $wgMimeInfoFile;
 168
 169                 $info= MM_WELL_KNOWN_MIME_INFO;
 170
 171                 if ($wgMimeInfoFile) {
 172                         if (is_file($wgMimeInfoFile) and is_readable($wgMimeInfoFile)) {
 173                                 wfDebug("MimeMagic::MimeMagic: loading mime info from $wgMimeInfoFile\n");
 174
 175                                 $info.= "\n";
 176                                 $info.= file_get_contents($wgMimeInfoFile);
 177                         }
 178                         else wfDebug("MimeMagic::MimeMagic: can't load mime info from $wgMimeInfoFile\n");
 179                 }
 180                 else wfDebug("MimeMagic::MimeMagic: no mime info file defined, using build-ins only.\n");
 181
 182                 $info= str_replace(array("\r\n","\n\r","\n\n","\r\r","\r"),"\n",$info);
 183                 $info= str_replace("\t"," ",$info);
 184
 185                 $this->mMimeTypeAliases= array();
 186                 $this->mMediaTypes= array();
 187
 188                 $lines= explode("\n",$info);
 189                 foreach ($lines as $s) {
 190                         $s= trim($s);
 191                         if (empty($s)) continue;
 192                         if (strpos($s,'#')===0) continue;
 193
 194                         $s= strtolower($s);
 195                         $i= strpos($s,' ');
 196
 197                         if ($i===false) continue;
 198
 199                         #print "processing MIME INFO line $s<br>";
 200
 201                         $match= array();
 202                         if (preg_match('!\[\s*(\w+)\s*\]!',$s,$match)) {
 203                                 $s= preg_replace('!\[\s*(\w+)\s*\]!','',$s);
 204                                 $mtype= trim(strtoupper($match[1]));
 205                         }
 206                         else $mtype= MEDIATYPE_UNKNOWN;
 207
 208                         $m= explode(' ',$s);
 209
 210                         if (!isset($this->mMediaTypes[$mtype])) $this->mMediaTypes[$mtype]= array();
 211
 212                         foreach ($m as $mime) {
 213                                 $mime= trim($mime);
 214                                 if (empty($mime)) continue;
 215
 216                                 $this->mMediaTypes[$mtype][]= $mime;
 217                         }
 218
 219                         if (sizeof($m)>1) {
 220                                 $main= $m[0];
 221                                 for ($i=1; $i<sizeof($m); $i+= 1) {
 222                                         $mime= $m[$i];
 223                                         $this->mMimeTypeAliases[$mime]= $main;
 224                                 }
 225                         }
 226                 }
 227
 228         }
 229
 230         /** returns a list of file extensions for a given mime type
 231         * as a space separated string.
 232         */
 233         function getExtensionsForType($mime) {
 234                 $mime= strtolower($mime);
 235
 236                 $r= @$this->mMimeToExt[$mime];
 237
 238                 if (@!$r and isset($this->mMimeTypeAliases[$mime])) {
 239                         $mime= $this->mMimeTypeAliases[$mime];
 240                         $r= @$this->mMimeToExt[$mime];
 241                 }
 242
 243                 return $r;
 244         }
 245
 246         /** returns a list of mime types for a given file extension
 247         * as a space separated string.
 248         */
 249         function getTypesForExtension($ext) {
 250                 $ext= strtolower($ext);
 251
 252                 $r= @$this->mExtToMime[$ext];
 253                 return $r;
 254         }
 255
 256         /** returns a single mime type for a given file extension.
 257         * This is always the first type from the list returned by getTypesForExtension($ext).
 258         */
 259         function guessTypesForExtension($ext) {
 260                 $m= $this->getTypesForExtension( $ext );
 261                 if( is_null($m) ) return NULL;
 262
 263                 $m= trim( $m );
 264                 $m= preg_replace('/\s.*$/','',$m);
 265
 266                 return $m;
 267         }
 268
 269
 270         /** tests if the extension matches the given mime type.
 271         * returns true if a match was found, NULL if the mime type is unknown,
 272         * and false if the mime type is known but no matches where found.
 273         */
 274         function isMatchingExtension($extension,$mime) {
 275                 $ext= $this->getExtensionsForType($mime);
 276
 277                 if (!$ext) {
 278                         return NULL;  //unknown
 279                 }
 280
 281                 $ext= explode(' ',$ext);
 282
 283                 $extension= strtolower($extension);
 284                 if (in_array($extension,$ext)) {
 285                         return true;
 286                 }
 287
 288                 return false;
 289         }
 290
 291         /** returns true if the mime type is known to represent
 292         * an image format supported by the PHP GD library.
 293         */
 294         function isPHPImageType( $mime ) {
 295                 #as defined by imagegetsize and image_type_to_mime
 296                 static $types = array(
 297                         'image/gif', 'image/jpeg', 'image/png',
 298                         'image/x-bmp', 'image/xbm', 'image/tiff',
 299                         'image/jp2', 'image/jpeg2000', 'image/iff',
 300                         'image/xbm', 'image/x-xbitmap',
 301                         'image/vnd.wap.wbmp', 'image/vnd.xiff',
 302                         'image/x-photoshop',
 303                         'application/x-shockwave-flash',
 304                 );
 305
 306                 return in_array( $mime, $types );
 307         }
 308
 309         /**
 310          * Returns true if the extension represents a type which can
 311          * be reliably detected from its content. Use this to determine
 312          * whether strict content checks should be applied to reject
 313          * invalid uploads; if we can't identify the type we won't
 314          * be able to say if it's invalid.
 315          *
 316          * @todo Be more accurate when using fancy mime detector plugins;
 317          *       right now this is the bare minimum getimagesize() list.
 318          * @return bool
 319          */
 320         function isRecognizableExtension( $extension ) {
 321                 static $types = array(
 322                         'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
 323                         'bmp', 'tiff', 'tif', 'jpc', 'jp2',
 324                         'jpx', 'jb2', 'swc', 'iff', 'wbmp',
 325                         'xbm', 'djvu'
 326                 );
 327                 return in_array( strtolower( $extension ), $types );
 328         }
 329
 330
 331         /** mime type detection. This uses detectMimeType to detect the mim type of the file,
 332         * but applies additional checks to determine some well known file formats that may be missed
 333         * or misinterpreter by the default mime detection (namely xml based formats like XHTML or SVG).
 334         *
 335         * @param string $file The file to check
 336         * @param bool $useExt switch for allowing to use the file extension to guess the mime type. true by default.
 337         *
 338         * @return string the mime type of $file
 339         */
 340         function guessMimeType( $file, $useExt=true ) {
 341                 $fname = 'MimeMagic::guessMimeType';
 342                 $mime= $this->detectMimeType($file,$useExt);
 343
 344                 // Read a chunk of the file
 345                 $f = fopen( $file, "rt" );
 346                 if( !$f ) return "unknown/unknown";
 347                 $head = fread( $f, 1024 );
 348                 fclose( $f );
 349
 350                 $sub4 =  substr( $head, 0, 4 );
 351                 if ( $sub4 == "\x01\x00\x09\x00" || $sub4 == "\xd7\xcd\xc6\x9a" ) {
 352                         // WMF kill kill kill
 353                         // Note that WMF may have a bare header, no magic number.
 354                         // The former of the above two checks is theoretically prone to false positives
 355                         $mime = "application/x-msmetafile";
 356                 }
 357
 358                 if (strpos($mime,"text/")===0 || $mime==="application/xml") {
 359
 360                         $xml_type= NULL;
 361                         $script_type= NULL;
 362
 363                         /*
 364                         * look for XML formats (XHTML and SVG)
 365                         */
 366                         if ($mime==="text/sgml" ||
 367                             $mime==="text/plain" ||
 368                             $mime==="text/html" ||
 369                             $mime==="text/xml" ||
 370                             $mime==="application/xml") {
 371
 372                                 if (substr($head,0,5)=="<?xml") $xml_type= "ASCII";
 373                                 elseif (substr($head,0,8)=="\xef\xbb\xbf<?xml") $xml_type= "UTF-8";
 374                                 elseif (substr($head,0,10)=="\xfe\xff\x00<\x00?\x00x\x00m\x00l") $xml_type= "UTF-16BE";
 375                                 elseif (substr($head,0,10)=="\xff\xfe<\x00?\x00x\x00m\x00l\x00") $xml_type= "UTF-16LE";
 376
 377                                 if ($xml_type) {
 378                                         if ($xml_type!=="UTF-8" && $xml_type!=="ASCII") $head= iconv($xml_type,"ASCII//IGNORE",$head);
 379
 380                                         $match= array();
 381                                         $doctype= "";
 382                                         $tag= "";
 383
 384                                         if (preg_match('%<!DOCTYPE\s+[\w-]+\s+PUBLIC\s+["'."'".'"](.*?)["'."'".'"].*>%sim',$head,$match)) $doctype= $match[1];
 385                                         if (preg_match('%<(\w+).*>%sim',$head,$match)) $tag= $match[1];
 386
 387                                         #print "<br>ANALYSING $file ($mime): doctype= $doctype; tag= $tag<br>";
 388
 389                                         if (strpos($doctype,"-//W3C//DTD SVG")===0) $mime= "image/svg";
 390                                         elseif ($tag==="svg") $mime= "image/svg";
 391                                         elseif (strpos($doctype,"-//W3C//DTD XHTML")===0) $mime= "text/html";
 392                                         elseif ($tag==="html") $mime= "text/html";
 393
 394                                         $test_more= false;
 395                                 }
 396                         }
 397
 398                         /*
 399                         * look for shell scripts
 400                         */
 401                         if (!$xml_type) {
 402                                 $script_type= NULL;
 403
 404                                 #detect by shebang
 405                                 if (substr($head,0,2)=="#!") $script_type= "ASCII";
 406                                 elseif (substr($head,0,5)=="\xef\xbb\xbf#!") $script_type= "UTF-8";
 407                                 elseif (substr($head,0,7)=="\xfe\xff\x00#\x00!") $script_type= "UTF-16BE";
 408                                 elseif (substr($head,0,7)=="\xff\xfe#\x00!") $script_type= "UTF-16LE";
 409
 410                                 if ($script_type) {
 411                                         if ($script_type!=="UTF-8" && $script_type!=="ASCII") $head= iconv($script_type,"ASCII//IGNORE",$head);
 412
 413                                         $match= array();
 414                                         $prog= "";
 415
 416                                         if (preg_match('%/?([^\s]+/)(w+)%sim',$head,$match)) $script= $match[2];
 417
 418                                         $mime= "application/x-$prog";
 419                                 }
 420                         }
 421
 422                         /*
 423                         * look for PHP
 424                         */
 425                         if( !$xml_type && !$script_type ) {
 426
 427                                 if( ( strpos( $head, '<?php' ) !== false ) ||
 428                                     ( strpos( $head, '<? ' ) !== false ) ||
 429                                     ( strpos( $head, "<?\n" ) !== false ) ||
 430                                     ( strpos( $head, "<?\t" ) !== false ) ||
 431                                     ( strpos( $head, "<?=" ) !== false ) ||
 432
 433                                     ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
 434                                     ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
 435                                     ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
 436                                     ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
 437                                     ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {
 438
 439                                         $mime= "application/x-php";
 440                                 }
 441                         }
 442
 443                 }
 444
 445                 if (isset($this->mMimeTypeAliases[$mime])) $mime= $this->mMimeTypeAliases[$mime];
 446
 447                 wfDebug("$fname: final mime type of $file: $mime\n");
 448                 return $mime;
 449         }
 450
 451         /** Internal mime type detection, please use guessMimeType() for application code instead.
 452         * Detection is done using an external program, if $wgMimeDetectorCommand is set.
 453         * Otherwise, the fileinfo extension and mime_content_type are tried (in this order), if they are available.
 454         * If the dections fails and $useExt is true, the mime type is guessed from the file extension, using guessTypesForExtension.
 455         * If the mime type is still unknown, getimagesize is used to detect the mime type if the file is an image.
 456         * If no mime type can be determined, this function returns "unknown/unknown".
 457         *
 458         * @param string $file The file to check
 459         * @param bool $useExt switch for allowing to use the file extension to guess the mime type. true by default.
 460         *
 461         * @return string the mime type of $file
 462         * @access private
 463         */
 464         function detectMimeType( $file, $useExt=true ) {
 465                 $fname = 'MimeMagic::detectMimeType';
 466
 467                 global $wgMimeDetectorCommand;
 468
 469                 $m= NULL;
 470                 if ($wgMimeDetectorCommand) {
 471                         $fn= wfEscapeShellArg($file);
 472                         $m= `$wgMimeDetectorCommand $fn`;
 473                 }
 474                 else if (function_exists("finfo_open") && function_exists("finfo_file")) {
 475
 476                         # This required the fileinfo extension by PECL,
 477                         # see http://pecl.php.net/package/fileinfo
 478                         # This must be compiled into PHP
 479                         #
 480                         # finfo is the official replacement for the deprecated
 481                         # mime_content_type function, see below.
 482                         #
 483                         # If you may need to load the fileinfo extension at runtime, set
 484                         # $wgLoadFileinfoExtension in LocalSettings.php
 485
 486                         $mime_magic_resource = finfo_open(FILEINFO_MIME); /* return mime type ala mimetype extension */
 487
 488                         if ($mime_magic_resource) {
 489                                 $m= finfo_file($mime_magic_resource, $file);
 490
 491                                 finfo_close($mime_magic_resource);
 492                         }
 493                         else wfDebug("$fname: finfo_open failed on ".FILEINFO_MIME."!\n");
 494                 }
 495                 else if (function_exists("mime_content_type")) {
 496
 497                         # NOTE: this function is available since PHP 4.3.0, but only if
 498                         # PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
 499                         #
 500                         # On Windows, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
 501                         # sometimes, this may even be needed under linus/unix.
 502                         #
 503                         # Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
 504                         # see http://www.php.net/manual/en/ref.mime-magic.php for details.
 505
 506                         $m= mime_content_type($file);
 507
 508                         if ( $m == 'text/plain' ) {
 509                                 // mime_content_type sometimes considers DJVU files to be text/plain.
 510                                 $deja = new DjVuImage( $file );
 511                                 if( $deja->isValid() ) {
 512                                         wfDebug("$fname: (re)detected $file as image/vnd.djvu\n");
 513                                         $m = 'image/vnd.djvu';
 514                                 }
 515                         }
 516                 }
 517                 else wfDebug("$fname: no magic mime detector found!\n");
 518
 519                 if ($m) {
 520                         #normalize
 521                         $m= preg_replace('![;, ].*$!','',$m); #strip charset, etc
 522                         $m= trim($m);
 523                         $m= strtolower($m);
 524
 525                         if (strpos($m,'unknown')!==false) $m= NULL;
 526                         else {
 527                                 wfDebug("$fname: magic mime type of $file: $m\n");
 528                                 return $m;
 529                         }
 530                 }
 531
 532                 #if still not known, use getimagesize to find out the type of image
 533                 #TODO: skip things that do not have a well-known image extension? Would that be safe?
 534                 wfSuppressWarnings();
 535                 $gis = getimagesize( $file );
 536                 wfRestoreWarnings();
 537
 538                 $notAnImage= false;
 539
 540                 if ($gis && is_array($gis) && $gis[2]) {
 541                         switch ($gis[2]) {
 542                                 case IMAGETYPE_GIF: $m= "image/gif"; break;
 543                                 case IMAGETYPE_JPEG: $m= "image/jpeg"; break;
 544                                 case IMAGETYPE_PNG: $m= "image/png"; break;
 545                                 case IMAGETYPE_SWF: $m= "application/x-shockwave-flash"; break;
 546                                 case IMAGETYPE_PSD: $m= "application/photoshop"; break;
 547                                 case IMAGETYPE_BMP: $m= "image/bmp"; break;
 548                                 case IMAGETYPE_TIFF_II: $m= "image/tiff"; break;
 549                                 case IMAGETYPE_TIFF_MM: $m= "image/tiff"; break;
 550                                 case IMAGETYPE_JPC: $m= "image"; break;
 551                                 case IMAGETYPE_JP2: $m= "image/jpeg2000"; break;
 552                                 case IMAGETYPE_JPX: $m= "image/jpeg2000"; break;
 553                                 case IMAGETYPE_JB2: $m= "image"; break;
 554                                 case IMAGETYPE_SWC: $m= "application/x-shockwave-flash"; break;
 555                                 case IMAGETYPE_IFF: $m= "image/vnd.xiff"; break;
 556                                 case IMAGETYPE_WBMP: $m= "image/vnd.wap.wbmp"; break;
 557                                 case IMAGETYPE_XBM: $m= "image/x-xbitmap"; break;
 558                         }
 559
 560                         if ($m) {
 561                                 wfDebug("$fname: image mime type of $file: $m\n");
 562                                 return $m;
 563                         }
 564                         else $notAnImage= true;
 565                 } else {
 566                         // Also test DjVu
 567                         $deja = new DjVuImage( $file );
 568                         if( $deja->isValid() ) {
 569                                 wfDebug("$fname: detected $file as image/vnd.djvu\n");
 570                                 return 'image/vnd.djvu';
 571                         }
 572                 }
 573
 574                 #if desired, look at extension as a fallback.
 575                 if ($useExt) {
 576                         $i = strrpos( $file, '.' );
 577                         $e= strtolower( $i ? substr( $file, $i + 1 ) : '' );
 578
 579                         $m= $this->guessTypesForExtension($e);
 580
 581                         #TODO: if $notAnImage is set, do not trust the file extension if
 582                         # the results is one of the image types that should have been recognized
 583                         # by getimagesize
 584
 585                         if ($m) {
 586                                 wfDebug("$fname: extension mime type of $file: $m\n");
 587                                 return $m;
 588                         }
 589                 }
 590
 591                 #unknown type
 592                 wfDebug("$fname: failed to guess mime type for $file!\n");
 593                 return "unknown/unknown";
 594         }
 595
 596         /**
 597         * Determine the media type code for a file, using its mime type, name and possibly
 598         * its contents.
 599         *
 600         * This function relies on the findMediaType(), mapping extensions and mime
 601         * types to media types.
 602         *
 603         * @todo analyse file if need be
 604         * @todo look at multiple extension, separately and together.
 605         *
 606         * @param string $path full path to the image file, in case we have to look at the contents
 607         *        (if null, only the mime type is used to determine the media type code).
 608         * @param string $mime mime type. If null it will be guessed using guessMimeType.
 609         *
 610         * @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
 611         */
 612         function getMediaType($path=NULL,$mime=NULL) {
 613                 if( !$mime && !$path ) return MEDIATYPE_UNKNOWN;
 614
 615                 #if mime type is unknown, guess it
 616                 if( !$mime ) $mime= $this->guessMimeType($path,false);
 617
 618                 #special code for ogg - detect if it's video (theora),
 619                 #else label it as sound.
 620                 if( $mime=="application/ogg" && file_exists($path) ) {
 621
 622                         // Read a chunk of the file
 623                         $f = fopen( $path, "rt" );
 624                         if( !$f ) return MEDIATYPE_UNKNOWN;
 625                         $head = fread( $f, 256 );
 626                         fclose( $f );
 627
 628                         $head= strtolower( $head );
 629
 630                         #This is an UGLY HACK, file should be parsed correctly
 631                         if( strpos($head,'theora')!==false ) return MEDIATYPE_VIDEO;
 632                         elseif( strpos($head,'vorbis')!==false ) return MEDIATYPE_AUDIO;
 633                         elseif( strpos($head,'flac')!==false ) return MEDIATYPE_AUDIO;
 634                         elseif( strpos($head,'speex')!==false ) return MEDIATYPE_AUDIO;
 635                         else return MEDIATYPE_MULTIMEDIA;
 636                 }
 637
 638                 #check for entry for full mime type
 639                 if( $mime ) {
 640                         $type= $this->findMediaType($mime);
 641                         if( $type!==MEDIATYPE_UNKNOWN ) return $type;
 642                 }
 643
 644                 #check for entry for file extension
 645                 $e= NULL;
 646                 if( $path ) {
 647                         $i = strrpos( $path, '.' );
 648                         $e= strtolower( $i ? substr( $path, $i + 1 ) : '' );
 649
 650                         #TODO: look at multi-extension if this fails, parse from full path
 651
 652                         $type= $this->findMediaType('.'.$e);
 653                         if( $type!==MEDIATYPE_UNKNOWN ) return $type;
 654                 }
 655
 656                 #check major mime type
 657                 if( $mime ) {
 658                         $i= strpos($mime,'/');
 659                         if( $i !== false ) {
 660                                 $major= substr($mime,0,$i);
 661                                 $type= $this->findMediaType($major);
 662                                 if( $type!==MEDIATYPE_UNKNOWN ) return $type;
 663                         }
 664                 }
 665
 666                 if( !$type ) $type= MEDIATYPE_UNKNOWN;
 667
 668                 return $type;
 669         }
 670
 671         /** returns a media code matching the given mime type or file extension.
 672         * File extensions are represented by a string starting with a dot (.) to
 673         * distinguish them from mime types.
 674         *
 675         * This funktion relies on the mapping defined by $this->mMediaTypes
 676         * @access private
 677         */
 678         function findMediaType($extMime) {
 679
 680                 if (strpos($extMime,'.')===0) { #if it's an extension, look up the mime types
 681                         $m= $this->getTypesForExtension(substr($extMime,1));
 682                         if (!$m) return MEDIATYPE_UNKNOWN;
 683
 684                         $m= explode(' ',$m);
 685                 }
 686                 else { #normalize mime type
 687                         if (isset($this->mMimeTypeAliases[$extMime])) {
 688                                 $extMime= $this->mMimeTypeAliases[$extMime];
 689                         }
 690
 691                         $m= array($extMime);
 692                 }
 693
 694                 foreach ($m as $mime) {
 695                         foreach ($this->mMediaTypes as $type => $codes) {
 696                                 if (in_array($mime,$codes,true)) return $type;
 697                         }
 698                 }
 699
 700                 return MEDIATYPE_UNKNOWN;
 701         }
 702 }
 703
 704 ?>