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 image/x.djvu image/x-djvu djvu
  28 image/x-portable-pixmap ppm
  29 image/x-xcf xcf
  30 text/plain txt
  31 text/html html htm
  32 video/ogg ogm ogg
  33 video/mpeg mpg mpeg
  34 END_STRING
  35 );
  36
  37  /** Defines a set of well known mime info entries
  38  * This is used as a fallback to mime.info files.
  39  * An extensive list of well known mime types is provided by
  40  * the file mime.info in the includes directory.
  41  */
  42 define('MM_WELL_KNOWN_MIME_INFO', <<<END_STRING
  43 application/pdf [OFFICE]
  44 text/javascript application/x-javascript [EXECUTABLE]
  45 application/x-shockwave-flash [MULTIMEDIA]
  46 audio/midi [AUDIO]
  47 audio/x-aiff [AUDIO]
  48 audio/x-wav [AUDIO]
  49 audio/mp3 audio/mpeg [AUDIO]
  50 application/ogg audio/ogg video/ogg [MULTIMEDIA]
  51 image/x-bmp image/bmp [BITMAP]
  52 image/gif [BITMAP]
  53 image/jpeg [BITMAP]
  54 image/png [BITMAP]
  55 image/svg+xml [DRAWING]
  56 image/tiff [BITMAP]
  57 image/vnd.djvu [BITMAP]
  58 image/x-xcf [BITMAP]
  59 image/x-portable-pixmap [BITMAP]
  60 text/plain [TEXT]
  61 text/html [TEXT]
  62 video/ogg [VIDEO]
  63 video/mpeg [VIDEO]
  64 unknown/unknown application/octet-stream application/x-empty [UNKNOWN]
  65 END_STRING
  66 );
  67
  68 #note: because this file is possibly included by a function,
  69 #we need to access the global scope explicitely!
  70 global $wgLoadFileinfoExtension;
  71
  72 if ($wgLoadFileinfoExtension) {
  73         if(!extension_loaded('fileinfo')) dl('fileinfo.' . PHP_SHLIB_SUFFIX);
  74 }
  75
  76 /**
  77  * Implements functions related to mime types such as detection and mapping to
  78  * file extension.
  79  *
  80  * Instances of this class are stateles, there only needs to be one global instance
  81  * of MimeMagic. Please use MimeMagic::singleton() to get that instance.
  82  */
  83 class MimeMagic {
  84
  85         /**
  86         * Mapping of media types to arrays of mime types.
  87         * This is used by findMediaType and getMediaType, respectively
  88         */
  89         var $mMediaTypes= NULL;
  90
  91         /** Map of mime type aliases
  92         */
  93         var $mMimeTypeAliases= NULL;
  94
  95         /** map of mime types to file extensions (as a space seprarated list)
  96         */
  97         var $mMimeToExt= NULL;
  98
  99         /** map of file extensions types to mime types (as a space seprarated list)
 100         */
 101         var $mExtToMime= NULL;
 102
 103         /** The singleton instance
 104          */
 105         private static $instance;
 106
 107         /** Initializes the MimeMagic object. This is called by MimeMagic::singleton().
 108         *
 109         * This constructor parses the mime.types and mime.info files and build internal mappings.
 110         */
 111         function __construct() {
 112                 /*
 113                 *   --- load mime.types ---
 114                 */
 115
 116                 global $wgMimeTypeFile, $IP;
 117
 118                 $types = MM_WELL_KNOWN_MIME_TYPES;
 119
 120                 if ( $wgMimeTypeFile == 'includes/mime.types' ) {
 121                         $wgMimeTypeFile = "$IP/$wgMimeTypeFile";
 122                 }
 123
 124                 if ( $wgMimeTypeFile ) {
 125                         if ( is_file( $wgMimeTypeFile ) and is_readable( $wgMimeTypeFile ) ) {
 126                                 wfDebug( __METHOD__.": loading mime types from $wgMimeTypeFile\n" );
 127                                 $types .= "\n";
 128                                 $types .= file_get_contents( $wgMimeTypeFile );
 129                         } else {
 130                                 wfDebug( __METHOD__.": can't load mime types from $wgMimeTypeFile\n" );
 131                         }
 132                 } else {
 133                         wfDebug( __METHOD__.": no mime types file defined, using build-ins only.\n" );
 134                 }
 135
 136                 $types = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $types );
 137                 $types = str_replace( "\t", " ", $types );
 138
 139                 $this->mMimeToExt = array();
 140                 $this->mToMime = array();
 141
 142                 $lines = explode( "\n",$types );
 143                 foreach ( $lines as $s ) {
 144                         $s = trim( $s );
 145                         if ( empty( $s ) ) continue;
 146                         if ( strpos( $s, '#' ) === 0 ) continue;
 147
 148                         $s = strtolower( $s );
 149                         $i = strpos( $s, ' ' );
 150
 151                         if ( $i === false ) continue;
 152
 153                         #print "processing MIME line $s<br>";
 154
 155                         $mime = substr( $s, 0, $i );
 156                         $ext = trim( substr($s, $i+1 ) );
 157
 158                         if ( empty( $ext ) ) continue;
 159
 160                         if ( !empty( $this->mMimeToExt[$mime] ) ) {
 161                                 $this->mMimeToExt[$mime] .= ' ' . $ext;
 162                         } else {
 163                                 $this->mMimeToExt[$mime] = $ext;
 164                         }
 165
 166                         $extensions = explode( ' ', $ext );
 167
 168                         foreach ( $extensions as $e ) {
 169                                 $e = trim( $e );
 170                                 if ( empty( $e ) ) continue;
 171
 172                                 if ( !empty( $this->mExtToMime[$e] ) ) {
 173                                         $this->mExtToMime[$e] .= ' ' . $mime;
 174                                 } else {
 175                                         $this->mExtToMime[$e] = $mime;
 176                                 }
 177                         }
 178                 }
 179
 180                 /*
 181                 *   --- load mime.info ---
 182                 */
 183
 184                 global $wgMimeInfoFile;
 185                 if ( $wgMimeInfoFile == 'includes/mime.info' ) {
 186                         $wgMimeInfoFile = "$IP/$wgMimeInfoFile";
 187                 }
 188
 189                 $info = MM_WELL_KNOWN_MIME_INFO;
 190
 191                 if ( $wgMimeInfoFile ) {
 192                         if ( is_file( $wgMimeInfoFile ) and is_readable( $wgMimeInfoFile ) ) {
 193                                 wfDebug( __METHOD__.": loading mime info from $wgMimeInfoFile\n" );
 194                                 $info .= "\n";
 195                                 $info .= file_get_contents( $wgMimeInfoFile );
 196                         } else {
 197                                 wfDebug(__METHOD__.": can't load mime info from $wgMimeInfoFile\n");
 198                         }
 199                 } else {
 200                         wfDebug(__METHOD__.": no mime info file defined, using build-ins only.\n");
 201                 }
 202
 203                 $info = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $info);
 204                 $info = str_replace( "\t", " ", $info );
 205
 206                 $this->mMimeTypeAliases = array();
 207                 $this->mMediaTypes = array();
 208
 209                 $lines = explode( "\n", $info );
 210                 foreach ( $lines as $s ) {
 211                         $s = trim( $s );
 212                         if ( empty( $s ) ) continue;
 213                         if ( strpos( $s, '#' ) === 0 ) continue;
 214
 215                         $s = strtolower( $s );
 216                         $i = strpos( $s, ' ' );
 217
 218                         if ( $i === false ) continue;
 219
 220                         #print "processing MIME INFO line $s<br>";
 221
 222                         $match = array();
 223                         if ( preg_match( '!\[\s*(\w+)\s*\]!', $s, $match ) ) {
 224                                 $s = preg_replace( '!\[\s*(\w+)\s*\]!', '', $s );
 225                                 $mtype = trim( strtoupper( $match[1] ) );
 226                         } else {
 227                                 $mtype = MEDIATYPE_UNKNOWN;
 228                         }
 229
 230                         $m = explode( ' ', $s );
 231
 232                         if ( !isset( $this->mMediaTypes[$mtype] ) ) {
 233                                 $this->mMediaTypes[$mtype] = array();
 234                         }
 235
 236                         foreach ( $m as $mime ) {
 237                                 $mime = trim( $mime );
 238                                 if ( empty( $mime ) ) continue;
 239
 240                                 $this->mMediaTypes[$mtype][] = $mime;
 241                         }
 242
 243                         if ( sizeof( $m ) > 1 ) {
 244                                 $main = $m[0];
 245                                 for ( $i=1; $i<sizeof($m); $i += 1 ) {
 246                                         $mime = $m[$i];
 247                                         $this->mMimeTypeAliases[$mime] = $main;
 248                                 }
 249                         }
 250                 }
 251
 252         }
 253
 254         /**
 255          * Get an instance of this class
 256          */
 257         static function &singleton() {
 258                 if ( !isset( self::$instance ) ) {
 259                         self::$instance = new MimeMagic;
 260                 }
 261                 return self::$instance;
 262         }
 263
 264         /** returns a list of file extensions for a given mime type
 265         * as a space separated string.
 266         */
 267         function getExtensionsForType( $mime ) {
 268                 $mime = strtolower( $mime );
 269
 270                 $r = @$this->mMimeToExt[$mime];
 271
 272                 if ( @!$r and isset( $this->mMimeTypeAliases[$mime] ) ) {
 273                         $mime = $this->mMimeTypeAliases[$mime];
 274                         $r = @$this->mMimeToExt[$mime];
 275                 }
 276
 277                 return $r;
 278         }
 279
 280         /** returns a list of mime types for a given file extension
 281         * as a space separated string.
 282         */
 283         function getTypesForExtension( $ext ) {
 284                 $ext = strtolower( $ext );
 285
 286                 $r = isset( $this->mExtToMime[$ext] ) ? $this->mExtToMime[$ext] : null;
 287                 return $r;
 288         }
 289
 290         /** returns a single mime type for a given file extension.
 291         * This is always the first type from the list returned by getTypesForExtension($ext).
 292         */
 293         function guessTypesForExtension( $ext ) {
 294                 $m = $this->getTypesForExtension( $ext );
 295                 if ( is_null( $m ) ) return NULL;
 296
 297                 $m = trim( $m );
 298                 $m = preg_replace( '/\s.*$/', '', $m );
 299
 300                 return $m;
 301         }
 302
 303
 304         /** tests if the extension matches the given mime type.
 305         * returns true if a match was found, NULL if the mime type is unknown,
 306         * and false if the mime type is known but no matches where found.
 307         */
 308         function isMatchingExtension( $extension, $mime ) {
 309                 $ext = $this->getExtensionsForType( $mime );
 310
 311                 if ( !$ext ) {
 312                         return NULL;  //unknown
 313                 }
 314
 315                 $ext = explode( ' ', $ext );
 316
 317                 $extension = strtolower( $extension );
 318                 if ( in_array( $extension, $ext ) ) {
 319                         return true;
 320                 }
 321
 322                 return false;
 323         }
 324
 325         /** returns true if the mime type is known to represent
 326         * an image format supported by the PHP GD library.
 327         */
 328         function isPHPImageType( $mime ) {
 329                 #as defined by imagegetsize and image_type_to_mime
 330                 static $types = array(
 331                         'image/gif', 'image/jpeg', 'image/png',
 332                         'image/x-bmp', 'image/xbm', 'image/tiff',
 333                         'image/jp2', 'image/jpeg2000', 'image/iff',
 334                         'image/xbm', 'image/x-xbitmap',
 335                         'image/vnd.wap.wbmp', 'image/vnd.xiff',
 336                         'image/x-photoshop',
 337                         'application/x-shockwave-flash',
 338                 );
 339
 340                 return in_array( $mime, $types );
 341         }
 342
 343         /**
 344          * Returns true if the extension represents a type which can
 345          * be reliably detected from its content. Use this to determine
 346          * whether strict content checks should be applied to reject
 347          * invalid uploads; if we can't identify the type we won't
 348          * be able to say if it's invalid.
 349          *
 350          * @todo Be more accurate when using fancy mime detector plugins;
 351          *       right now this is the bare minimum getimagesize() list.
 352          * @return bool
 353          */
 354         function isRecognizableExtension( $extension ) {
 355                 static $types = array(
 356                         // Types recognized by getimagesize()
 357                         'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
 358                         'bmp', 'tiff', 'tif', 'jpc', 'jp2',
 359                         'jpx', 'jb2', 'swc', 'iff', 'wbmp',
 360                         'xbm',
 361
 362                         // Formats we recognize magic numbers for
 363                         'djvu', 'ogg', 'mid', 'pdf', 'wmf', 'xcf',
 364
 365                         // XML formats we sure hope we recognize reliably
 366                         'svg',
 367                 );
 368                 return in_array( strtolower( $extension ), $types );
 369         }
 370
 371
 372         /** mime type detection. This uses detectMimeType to detect the mime type of the file,
 373         * but applies additional checks to determine some well known file formats that may be missed
 374         * or misinterpreter by the default mime detection (namely xml based formats like XHTML or SVG).
 375         *
 376         * @param string $file The file to check
 377         * @param mixed $ext The file extension, or true to extract it from the filename.
 378         *                   Set it to false to ignore the extension.
 379         *
 380         * @return string the mime type of $file
 381         */
 382         function guessMimeType( $file, $ext = true ) {
 383                 $mime = $this->doGuessMimeType( $file, $ext );
 384
 385                 if( !$mime ) {
 386                         wfDebug( __METHOD__.": internal type detection failed for $file (.$ext)...\n" );
 387                         $mime = $this->detectMimeType( $file, $ext );
 388                 }
 389
 390                 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
 391                         $mime = $this->mMimeTypeAliases[$mime];
 392                 }
 393
 394                 wfDebug(__METHOD__.": final mime type of $file: $mime\n");
 395                 return $mime;
 396         }
 397
 398         function doGuessMimeType( $file, $ext = true ) {
 399                 // Read a chunk of the file
 400                 wfSuppressWarnings();
 401                 $f = fopen( $file, "rt" );
 402                 wfRestoreWarnings();
 403                 if( !$f ) return "unknown/unknown";
 404                 $head = fread( $f, 1024 );
 405                 fclose( $f );
 406
 407                 // Hardcode a few magic number checks...
 408                 $headers = array(
 409                         // Multimedia...
 410                         'MThd'             => 'audio/midi',
 411                         'OggS'             => 'application/ogg',
 412
 413                         // Image formats...
 414                         // Note that WMF may have a bare header, no magic number.
 415                         "\x01\x00\x09\x00" => 'application/x-msmetafile', // Possibly prone to false positives?
 416                         "\xd7\xcd\xc6\x9a" => 'application/x-msmetafile',
 417                         'PDF%'             => 'application/pdf',
 418                         'gimp xcf'         => 'image/x-xcf',
 419
 420                         // Some forbidden fruit...
 421                         'MZ'               => 'application/octet-stream', // DOS/Windows executable
 422                         "\xca\xfe\xba\xbe" => 'application/octet-stream', // Mach-O binary
 423                         "\x7fELF"          => 'application/octet-stream', // ELF binary
 424                 );
 425
 426                 foreach( $headers as $magic => $candidate ) {
 427                         if( strncmp( $head, $magic, strlen( $magic ) ) == 0 ) {
 428                                 wfDebug( __METHOD__ . ": magic header in $file recognized as $candidate\n" );
 429                                 return $candidate;
 430                         }
 431                 }
 432
 433                 /*
 434                  * look for PHP
 435                  * Check for this before HTML/XML...
 436                  * Warning: this is a heuristic, and won't match a file with a lot of non-PHP before.
 437                  * It will also match text files which could be PHP. :)
 438                  */
 439                 if( ( strpos( $head, '<?php' ) !== false ) ||
 440                     ( strpos( $head, '<? ' ) !== false ) ||
 441                     ( strpos( $head, "<?\n" ) !== false ) ||
 442                     ( strpos( $head, "<?\t" ) !== false ) ||
 443                     ( strpos( $head, "<?=" ) !== false ) ||
 444
 445                     ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
 446                     ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
 447                     ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
 448                     ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
 449                     ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {
 450
 451                         wfDebug( __METHOD__ . ": recognized $file as application/x-php\n" );
 452                         return "application/x-php";
 453                 }
 454
 455                 /*
 456                  * look for XML formats (XHTML and SVG)
 457                  */
 458                 $xml_type = NULL;
 459                 if ( substr( $head, 0, 5 ) == "<?xml" ) {
 460                         $xml_type = "ASCII";
 461                 } elseif ( substr( $head, 0, 8 ) == "\xef\xbb\xbf<?xml") {
 462                         $xml_type = "UTF-8";
 463                 } elseif ( substr( $head, 0, 12 ) == "\xfe\xff\x00<\x00?\x00x\x00m\x00l" ) {
 464                         $xml_type = "UTF-16BE";
 465                 } elseif ( substr( $head, 0, 12 ) == "\xff\xfe<\x00?\x00x\x00m\x00l\x00") {
 466                         $xml_type = "UTF-16LE";
 467                 } else {
 468                         /*
 469                         echo "WARNING: Undetected xml_type ...\n";
 470                         for( $i = 0; $i < 10; $i++ ) {
 471                                 $c = ord( $head{$i} );
 472                                 if( $c < 32 || $c > 126 ) {
 473                                         printf( "\\x%02x", $c );
 474                                 } else {
 475                                         print $head{$i};
 476                                 }
 477                         }
 478                         echo "\n";
 479                         */
 480                 }
 481
 482                 if( $xml_type == 'UTF-16BE' || $xml_type == 'UTF-16LE' ) {
 483                         // Quick and dirty fold down to ASCII!
 484                         $pack = array( 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' );
 485                         $chars = unpack( $pack[$xml_type], substr( $head, 2 ) );
 486                         $head = '';
 487                         foreach( $chars as $codepoint ) {
 488                                 if( $codepoint < 128 ) {
 489                                         $head .= chr( $codepoint );
 490                                 } else {
 491                                         $head .= '?';
 492                                 }
 493                         }
 494                 }
 495
 496                 $match = array();
 497                 $doctype = "";
 498                 $tag = "";
 499
 500                 if ( preg_match( '%<!DOCTYPE\s+[\w-]+\s+PUBLIC\s+["'."'".'"](.*?)["'."'".'"].*>%siD',
 501                         $head, $match ) ) {
 502                                 $doctype = $match[1];
 503                         }
 504
 505                 if( $xml_type || $doctype ) {
 506                         if ( preg_match( '%<(\w+)\b%si', $head, $match ) ) {
 507                                 $tag = $match[1];
 508                         }
 509
 510                         #print "<br>ANALYSING $file: doctype= $doctype; tag= $tag<br>";
 511
 512                         if ( strpos( $doctype, "-//W3C//DTD SVG" ) === 0 ) {
 513                                 return "image/svg+xml";
 514                         } elseif ( $tag === "svg" ) {
 515                                 return "image/svg+xml";
 516                         } elseif ( strpos( $doctype, "-//W3C//DTD XHTML" ) === 0 ) {
 517                                 return "text/html";
 518                         } elseif ( $tag === "html" ) {
 519                                 return "text/html";
 520                         } else {
 521                                 /// Fixme -- this would be the place to allow additional XML type checks
 522                                 return "application/xml";
 523                         }
 524                 }
 525
 526                 /*
 527                  * look for shell scripts
 528                  */
 529                 $script_type = NULL;
 530
 531                 # detect by shebang
 532                 if ( substr( $head, 0, 2) == "#!" ) {
 533                         $script_type = "ASCII";
 534                 } elseif ( substr( $head, 0, 5) == "\xef\xbb\xbf#!" ) {
 535                         $script_type = "UTF-8";
 536                 } elseif ( substr( $head, 0, 7) == "\xfe\xff\x00#\x00!" ) {
 537                         $script_type = "UTF-16BE";
 538                 } elseif ( substr( $head, 0, 7 ) == "\xff\xfe#\x00!" ) {
 539                         $script_type= "UTF-16LE";
 540                 }
 541
 542                 if ( $script_type ) {
 543                         if ( $script_type !== "UTF-8" && $script_type !== "ASCII") {
 544                                 $head = iconv( $script_type, "ASCII//IGNORE", $head);
 545                         }
 546
 547                         $match = array();
 548
 549                         if ( preg_match( '%/?([^\s]+/)(\w+)%', $head, $match ) ) {
 550                                 $mime = "application/x-{$match[2]}";
 551                                 wfDebug( __METHOD__.": shell script recognized as $mime\n" );
 552                                 return $mime;
 553                         }
 554                 }
 555
 556                 wfSuppressWarnings();
 557                 $gis = getimagesize( $file );
 558                 wfRestoreWarnings();
 559
 560                 if( $gis && isset( $gis['mime'] ) ) {
 561                         $mime = $gis['mime'];
 562                         wfDebug( __METHOD__.": getimagesize detected $file as $mime\n" );
 563                         return $mime;
 564                 } else {
 565                         return false;
 566                 }
 567
 568                 // Also test DjVu
 569                 $deja = new DjVuImage( $file );
 570                 if( $deja->isValid() ) {
 571                         wfDebug( __METHOD__.": detected $file as image/vnd.djvu\n" );
 572                         return 'image/vnd.djvu';
 573                 }
 574         }
 575
 576         /** Internal mime type detection, please use guessMimeType() for application code instead.
 577         * Detection is done using an external program, if $wgMimeDetectorCommand is set.
 578         * Otherwise, the fileinfo extension and mime_content_type are tried (in this order), if they are available.
 579         * If the dections fails and $ext is not false, the mime type is guessed from the file extension, using
 580         * guessTypesForExtension.
 581         * If the mime type is still unknown, getimagesize is used to detect the mime type if the file is an image.
 582         * If no mime type can be determined, this function returns "unknown/unknown".
 583         *
 584         * @param string $file The file to check
 585         * @param mixed $ext The file extension, or true to extract it from the filename.
 586         *                   Set it to false to ignore the extension.
 587         *
 588         * @return string the mime type of $file
 589         * @access private
 590         */
 591         function detectMimeType( $file, $ext = true ) {
 592                 global $wgMimeDetectorCommand;
 593
 594                 $m = NULL;
 595                 if ( $wgMimeDetectorCommand ) {
 596                         $fn = wfEscapeShellArg( $file );
 597                         $m = `$wgMimeDetectorCommand $fn`;
 598                 } elseif ( function_exists( "finfo_open" ) && function_exists( "finfo_file" ) ) {
 599
 600                         # This required the fileinfo extension by PECL,
 601                         # see http://pecl.php.net/package/fileinfo
 602                         # This must be compiled into PHP
 603                         #
 604                         # finfo is the official replacement for the deprecated
 605                         # mime_content_type function, see below.
 606                         #
 607                         # If you may need to load the fileinfo extension at runtime, set
 608                         # $wgLoadFileinfoExtension in LocalSettings.php
 609
 610                         $mime_magic_resource = finfo_open(FILEINFO_MIME); /* return mime type ala mimetype extension */
 611
 612                         if ($mime_magic_resource) {
 613                                 $m = finfo_file( $mime_magic_resource, $file );
 614                                 finfo_close( $mime_magic_resource );
 615                         } else {
 616                                 wfDebug( __METHOD__.": finfo_open failed on ".FILEINFO_MIME."!\n" );
 617                         }
 618                 } elseif ( function_exists( "mime_content_type" ) ) {
 619
 620                         # NOTE: this function is available since PHP 4.3.0, but only if
 621                         # PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
 622                         #
 623                         # On Windows, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
 624                         # sometimes, this may even be needed under linus/unix.
 625                         #
 626                         # Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
 627                         # see http://www.php.net/manual/en/ref.mime-magic.php for details.
 628
 629                         $m = mime_content_type($file);
 630                 } else {
 631                         wfDebug( __METHOD__.": no magic mime detector found!\n" );
 632                 }
 633
 634                 if ( $m ) {
 635                         # normalize
 636                         $m = preg_replace( '![;, ].*$!', '', $m ); #strip charset, etc
 637                         $m = trim( $m );
 638                         $m = strtolower( $m );
 639
 640                         if ( strpos( $m, 'unknown' ) !== false ) {
 641                                 $m = NULL;
 642                         } else {
 643                                 wfDebug( __METHOD__.": magic mime type of $file: $m\n" );
 644                                 return $m;
 645                         }
 646                 }
 647
 648                 # if desired, look at extension as a fallback.
 649                 if ( $ext === true ) {
 650                         $i = strrpos( $file, '.' );
 651                         $ext = strtolower( $i ? substr( $file, $i + 1 ) : '' );
 652                 }
 653                 if ( $ext ) {
 654                         if( $this->isRecognizableExtension( $ext ) ) {
 655                                 wfDebug( __METHOD__. ": refusing to guess mime type for .$ext file, we should have recognized it\n" );
 656                         } else {
 657                                 $m = $this->guessTypesForExtension( $ext );
 658                                 if ( $m ) {
 659                                         wfDebug( __METHOD__.": extension mime type of $file: $m\n" );
 660                                         return $m;
 661                                 }
 662                         }
 663                 }
 664
 665                 #unknown type
 666                 wfDebug( __METHOD__.": failed to guess mime type for $file!\n" );
 667                 return "unknown/unknown";
 668         }
 669
 670         /**
 671         * Determine the media type code for a file, using its mime type, name and possibly
 672         * its contents.
 673         *
 674         * This function relies on the findMediaType(), mapping extensions and mime
 675         * types to media types.
 676         *
 677         * @todo analyse file if need be
 678         * @todo look at multiple extension, separately and together.
 679         *
 680         * @param string $path full path to the image file, in case we have to look at the contents
 681         *        (if null, only the mime type is used to determine the media type code).
 682         * @param string $mime mime type. If null it will be guessed using guessMimeType.
 683         *
 684         * @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
 685         */
 686         function getMediaType( $path = NULL, $mime = NULL ) {
 687                 if( !$mime && !$path ) return MEDIATYPE_UNKNOWN;
 688
 689                 # If mime type is unknown, guess it
 690                 if( !$mime ) $mime = $this->guessMimeType( $path, false );
 691
 692                 # Special code for ogg - detect if it's video (theora),
 693                 # else label it as sound.
 694                 if( $mime == "application/ogg" && file_exists( $path ) ) {
 695
 696                         // Read a chunk of the file
 697                         $f = fopen( $path, "rt" );
 698                         if ( !$f ) return MEDIATYPE_UNKNOWN;
 699                         $head = fread( $f, 256 );
 700                         fclose( $f );
 701
 702                         $head = strtolower( $head );
 703
 704                         # This is an UGLY HACK, file should be parsed correctly
 705                         if ( strpos( $head, 'theora' ) !== false ) return MEDIATYPE_VIDEO;
 706                         elseif ( strpos( $head, 'vorbis' ) !== false ) return MEDIATYPE_AUDIO;
 707                         elseif ( strpos( $head, 'flac' ) !== false ) return MEDIATYPE_AUDIO;
 708                         elseif ( strpos( $head, 'speex' ) !== false ) return MEDIATYPE_AUDIO;
 709                         else return MEDIATYPE_MULTIMEDIA;
 710                 }
 711
 712                 # check for entry for full mime type
 713                 if( $mime ) {
 714                         $type = $this->findMediaType( $mime );
 715                         if( $type !== MEDIATYPE_UNKNOWN ) return $type;
 716                 }
 717
 718                 # Check for entry for file extension
 719                 $e = NULL;
 720                 if ( $path ) {
 721                         $i = strrpos( $path, '.' );
 722                         $e = strtolower( $i ? substr( $path, $i + 1 ) : '' );
 723
 724                         # TODO: look at multi-extension if this fails, parse from full path
 725
 726                         $type = $this->findMediaType( '.' . $e );
 727                         if ( $type !== MEDIATYPE_UNKNOWN ) return $type;
 728                 }
 729
 730                 # Check major mime type
 731                 if( $mime ) {
 732                         $i = strpos( $mime, '/' );
 733                         if( $i !== false ) {
 734                                 $major = substr( $mime, 0, $i );
 735                                 $type = $this->findMediaType( $major );
 736                                 if( $type !== MEDIATYPE_UNKNOWN ) return $type;
 737                         }
 738                 }
 739
 740                 if( !$type ) $type = MEDIATYPE_UNKNOWN;
 741
 742                 return $type;
 743         }
 744
 745         /** returns a media code matching the given mime type or file extension.
 746         * File extensions are represented by a string starting with a dot (.) to
 747         * distinguish them from mime types.
 748         *
 749         * This funktion relies on the mapping defined by $this->mMediaTypes
 750         * @access private
 751         */
 752         function findMediaType( $extMime ) {
 753                 if ( strpos( $extMime, '.' ) === 0 ) { #if it's an extension, look up the mime types
 754                         $m = $this->getTypesForExtension( substr( $extMime, 1 ) );
 755                         if ( !$m ) return MEDIATYPE_UNKNOWN;
 756
 757                         $m = explode( ' ', $m );
 758                 } else {
 759                         # Normalize mime type
 760                         if ( isset( $this->mMimeTypeAliases[$extMime] ) ) {
 761                                 $extMime = $this->mMimeTypeAliases[$extMime];
 762                         }
 763
 764                         $m = array($extMime);
 765                 }
 766
 767                 foreach ( $m as $mime ) {
 768                         foreach ( $this->mMediaTypes as $type => $codes ) {
 769                                 if ( in_array($mime, $codes, true ) ) {
 770                                         return $type;
 771                                 }
 772                         }
 773                 }
 774
 775                 return MEDIATYPE_UNKNOWN;
 776         }
 777 }
 778
 779