3 * Module defining helper functions for detecting and dealing with mime types.
9 * Defines a set of well known mime types
10 * This is used as a fallback to mime.types files.
11 * An extensive list of well known mime types is provided by
12 * the file mime.types in the includes directory.
14 * This list concatenated with mime.types is used to create a mime <-> ext
15 * map. Each line contains a mime type followed by a space separated list of
16 * extensions. If multiple extensions for a single mime type exist or if
17 * multiple mime types exist for a single extension then in most cases
18 * MediaWiki assumes that the first extension following the mime type is the
19 * canonical extension, and the first time a mime type appears for a certain
20 * extension is considered the canonical mime type.
22 * (Note that appending $wgMimeTypeFile to the end of MM_WELL_KNOWN_MIME_TYPES
23 * sucks because you can't redefine canonical types. This could be fixed by
24 * appending MM_WELL_KNOWN_MIME_TYPES behind $wgMimeTypeFile, but who knows
25 * what will break? In practice this probably isn't a problem anyway -- Bryan)
27 define('MM_WELL_KNOWN_MIME_TYPES',<<<END_STRING
28 application/ogg ogx ogg ogm ogv oga spx
30 application/vnd.oasis.opendocument.chart odc
31 application/vnd.oasis.opendocument.chart-template otc
32 application/vnd.oasis.opendocument.database odb
33 application/vnd.oasis.opendocument.formula odf
34 application/vnd.oasis.opendocument.formula-template otf
35 application/vnd.oasis.opendocument.graphics odg
36 application/vnd.oasis.opendocument.graphics-template otg
37 application/vnd.oasis.opendocument.image odi
38 application/vnd.oasis.opendocument.image-template oti
39 application/vnd.oasis.opendocument.presentation odp
40 application/vnd.oasis.opendocument.presentation-template otp
41 application/vnd.oasis.opendocument.spreadsheet ods
42 application/vnd.oasis.opendocument.spreadsheet-template ots
43 application/vnd.oasis.opendocument.text odt
44 application/vnd.oasis.opendocument.text-master otm
45 application/vnd.oasis.opendocument.text-template ott
46 application/vnd.oasis.opendocument.text-web oth
47 application/x-javascript js
48 application/x-shockwave-flash swf
49 audio/midi mid midi kar
50 audio/mpeg mpga mpa mp2 mp3
51 audio/x-aiff aif aiff aifc
56 image/jpeg jpeg jpg jpe
64 image/x-portable-pixmap ppm
74 * Defines a set of well known mime info entries
75 * This is used as a fallback to mime.info files.
76 * An extensive list of well known mime types is provided by
77 * the file mime.info in the includes directory.
79 define('MM_WELL_KNOWN_MIME_INFO', <<<END_STRING
80 application/pdf [OFFICE]
81 application/vnd.oasis.opendocument.chart [OFFICE]
82 application/vnd.oasis.opendocument.chart-template [OFFICE]
83 application/vnd.oasis.opendocument.database [OFFICE]
84 application/vnd.oasis.opendocument.formula [OFFICE]
85 application/vnd.oasis.opendocument.formula-template [OFFICE]
86 application/vnd.oasis.opendocument.graphics [OFFICE]
87 application/vnd.oasis.opendocument.graphics-template [OFFICE]
88 application/vnd.oasis.opendocument.image [OFFICE]
89 application/vnd.oasis.opendocument.image-template [OFFICE]
90 application/vnd.oasis.opendocument.presentation [OFFICE]
91 application/vnd.oasis.opendocument.presentation-template [OFFICE]
92 application/vnd.oasis.opendocument.spreadsheet [OFFICE]
93 application/vnd.oasis.opendocument.spreadsheet-template [OFFICE]
94 application/vnd.oasis.opendocument.text [OFFICE]
95 application/vnd.oasis.opendocument.text-template [OFFICE]
96 application/vnd.oasis.opendocument.text-master [OFFICE]
97 application/vnd.oasis.opendocument.text-web [OFFICE]
98 text/javascript application/x-javascript [EXECUTABLE]
99 application/x-shockwave-flash [MULTIMEDIA]
103 audio/mp3 audio/mpeg [AUDIO]
104 application/ogg audio/ogg video/ogg [MULTIMEDIA]
105 image/x-bmp image/x-ms-bmp image/bmp [BITMAP]
109 image/svg+xml [DRAWING]
111 image/vnd.djvu [BITMAP]
113 image/x-portable-pixmap [BITMAP]
118 unknown/unknown application/octet-stream application/x-empty [UNKNOWN]
123 * Implements functions related to mime types such as detection and mapping to
126 * Instances of this class are stateless, there only needs to be one global instance
127 * of MimeMagic. Please use MimeMagic::singleton() to get that instance.
132 * Mapping of media types to arrays of mime types.
133 * This is used by findMediaType and getMediaType, respectively
135 var $mMediaTypes = null;
137 /** Map of mime type aliases
139 var $mMimeTypeAliases = null;
141 /** map of mime types to file extensions (as a space seprarated list)
143 var $mMimeToExt = null;
145 /** map of file extensions types to mime types (as a space seprarated list)
147 var $mExtToMime = null;
149 /** IEContentAnalyzer instance
153 /** The singleton instance
155 private static $instance;
157 /** True if the fileinfo extension has been loaded
159 private static $extensionLoaded = false;
161 /** Initializes the MimeMagic object. This is called by MimeMagic::singleton().
163 * This constructor parses the mime.types and mime.info files and build internal mappings.
165 function __construct() {
167 * --- load mime.types ---
170 global $wgMimeTypeFile, $IP, $wgLoadFileinfoExtension;
172 $types = MM_WELL_KNOWN_MIME_TYPES;
174 if ( $wgMimeTypeFile == 'includes/mime.types' ) {
175 $wgMimeTypeFile = "$IP/$wgMimeTypeFile";
178 if ( $wgLoadFileinfoExtension && !self::$extensionLoaded ) {
179 self::$extensionLoaded = true;
183 if ( $wgMimeTypeFile ) {
184 if ( is_file( $wgMimeTypeFile ) and is_readable( $wgMimeTypeFile ) ) {
185 wfDebug( __METHOD__.": loading mime types from $wgMimeTypeFile\n" );
187 $types .= file_get_contents( $wgMimeTypeFile );
189 wfDebug( __METHOD__.": can't load mime types from $wgMimeTypeFile\n" );
192 wfDebug( __METHOD__.": no mime types file defined, using build-ins only.\n" );
195 $types = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $types );
196 $types = str_replace( "\t", " ", $types );
198 $this->mMimeToExt = array();
199 $this->mToMime = array();
201 $lines = explode( "\n",$types );
202 foreach ( $lines as $s ) {
207 if ( strpos( $s, '#' ) === 0 ) {
211 $s = strtolower( $s );
212 $i = strpos( $s, ' ' );
214 if ( $i === false ) {
218 $mime = substr( $s, 0, $i );
219 $ext = trim( substr($s, $i+1 ) );
221 if ( empty( $ext ) ) {
225 if ( !empty( $this->mMimeToExt[$mime] ) ) {
226 $this->mMimeToExt[$mime] .= ' ' . $ext;
228 $this->mMimeToExt[$mime] = $ext;
231 $extensions = explode( ' ', $ext );
233 foreach ( $extensions as $e ) {
239 if ( !empty( $this->mExtToMime[$e] ) ) {
240 $this->mExtToMime[$e] .= ' ' . $mime;
242 $this->mExtToMime[$e] = $mime;
248 * --- load mime.info ---
251 global $wgMimeInfoFile;
252 if ( $wgMimeInfoFile == 'includes/mime.info' ) {
253 $wgMimeInfoFile = "$IP/$wgMimeInfoFile";
256 $info = MM_WELL_KNOWN_MIME_INFO;
258 if ( $wgMimeInfoFile ) {
259 if ( is_file( $wgMimeInfoFile ) and is_readable( $wgMimeInfoFile ) ) {
260 wfDebug( __METHOD__.": loading mime info from $wgMimeInfoFile\n" );
262 $info .= file_get_contents( $wgMimeInfoFile );
264 wfDebug(__METHOD__.": can't load mime info from $wgMimeInfoFile\n");
267 wfDebug(__METHOD__.": no mime info file defined, using build-ins only.\n");
270 $info = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $info);
271 $info = str_replace( "\t", " ", $info );
273 $this->mMimeTypeAliases = array();
274 $this->mMediaTypes = array();
276 $lines = explode( "\n", $info );
277 foreach ( $lines as $s ) {
282 if ( strpos( $s, '#' ) === 0 ) {
286 $s = strtolower( $s );
287 $i = strpos( $s, ' ' );
289 if ( $i === false ) {
293 #print "processing MIME INFO line $s<br>";
296 if ( preg_match( '!\[\s*(\w+)\s*\]!', $s, $match ) ) {
297 $s = preg_replace( '!\[\s*(\w+)\s*\]!', '', $s );
298 $mtype = trim( strtoupper( $match[1] ) );
300 $mtype = MEDIATYPE_UNKNOWN;
303 $m = explode( ' ', $s );
305 if ( !isset( $this->mMediaTypes[$mtype] ) ) {
306 $this->mMediaTypes[$mtype] = array();
309 foreach ( $m as $mime ) {
310 $mime = trim( $mime );
311 if ( empty( $mime ) ) {
315 $this->mMediaTypes[$mtype][] = $mime;
318 if ( sizeof( $m ) > 1 ) {
320 for ( $i=1; $i<sizeof($m); $i += 1 ) {
322 $this->mMimeTypeAliases[$mime] = $main;
330 * Get an instance of this class
333 public static function &singleton() {
334 if ( !isset( self::$instance ) ) {
335 self::$instance = new MimeMagic;
337 return self::$instance;
341 * Returns a list of file extensions for a given mime type as a space
342 * separated string or null if the mime type was unrecognized. Resolves
345 * @param $mime string
346 * @return string|null
348 public function getExtensionsForType( $mime ) {
349 $mime = strtolower( $mime );
351 // Check the mime-to-ext map
352 if ( isset( $this->mMimeToExt[$mime] ) ) {
353 return $this->mMimeToExt[$mime];
356 // Resolve the mime type to the canonical type
357 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
358 $mime = $this->mMimeTypeAliases[$mime];
359 if ( isset( $this->mMimeToExt[$mime] ) ) {
360 return $this->mMimeToExt[$mime];
368 * Returns a list of mime types for a given file extension as a space
369 * separated string or null if the extension was unrecognized.
372 * @return string|null
374 public function getTypesForExtension( $ext ) {
375 $ext = strtolower( $ext );
377 $r = isset( $this->mExtToMime[$ext] ) ? $this->mExtToMime[$ext] : null;
382 * Returns a single mime type for a given file extension or null if unknown.
383 * This is always the first type from the list returned by getTypesForExtension($ext).
386 * @return string|null
388 public function guessTypesForExtension( $ext ) {
389 $m = $this->getTypesForExtension( $ext );
390 if ( is_null( $m ) ) {
394 // TODO: Check if this is needed; strtok( $m, ' ' ) should be sufficient
396 $m = preg_replace( '/\s.*$/', '', $m );
403 * Tests if the extension matches the given mime type. Returns true if a
404 * match was found, null if the mime type is unknown, and false if the
405 * mime type is known but no matches where found.
407 * @param $extension string
408 * @param $mime string
411 public function isMatchingExtension( $extension, $mime ) {
412 $ext = $this->getExtensionsForType( $mime );
415 return null; // Unknown mime type
418 $ext = explode( ' ', $ext );
420 $extension = strtolower( $extension );
421 return in_array( $extension, $ext );
425 * Returns true if the mime type is known to represent an image format
426 * supported by the PHP GD library.
428 * @param $mime string
432 public function isPHPImageType( $mime ) {
433 // As defined by imagegetsize and image_type_to_mime
434 static $types = array(
435 'image/gif', 'image/jpeg', 'image/png',
436 'image/x-bmp', 'image/xbm', 'image/tiff',
437 'image/jp2', 'image/jpeg2000', 'image/iff',
438 'image/xbm', 'image/x-xbitmap',
439 'image/vnd.wap.wbmp', 'image/vnd.xiff',
441 'application/x-shockwave-flash',
444 return in_array( $mime, $types );
448 * Returns true if the extension represents a type which can
449 * be reliably detected from its content. Use this to determine
450 * whether strict content checks should be applied to reject
451 * invalid uploads; if we can't identify the type we won't
452 * be able to say if it's invalid.
454 * @todo Be more accurate when using fancy mime detector plugins;
455 * right now this is the bare minimum getimagesize() list.
458 function isRecognizableExtension( $extension ) {
459 static $types = array(
460 // Types recognized by getimagesize()
461 'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
462 'bmp', 'tiff', 'tif', 'jpc', 'jp2',
463 'jpx', 'jb2', 'swc', 'iff', 'wbmp',
466 // Formats we recognize magic numbers for
467 'djvu', 'ogx', 'ogg', 'ogv', 'oga', 'spx',
468 'mid', 'pdf', 'wmf', 'xcf', 'webm', 'mkv', 'mka',
471 // XML formats we sure hope we recognize reliably
474 return in_array( strtolower( $extension ), $types );
478 * Improves a mime type using the file extension. Some file formats are very generic,
479 * so their mime type is not very meaningful. A more useful mime type can be derived
480 * by looking at the file extension. Typically, this method would be called on the
481 * result of guessMimeType().
483 * Currently, this method does the following:
485 * If $mime is "unknown/unknown" and isRecognizableExtension( $ext ) returns false,
486 * return the result of guessTypesForExtension($ext).
488 * If $mime is "application/x-opc+zip" and isMatchingExtension( $ext, $mime )
489 * gives true, return the result of guessTypesForExtension($ext).
491 * @param $mime String: the mime type, typically guessed from a file's content.
492 * @param $ext String: the file extension, as taken from the file name
494 * @return string the mime type
496 public function improveTypeFromExtension( $mime, $ext ) {
497 if ( $mime === 'unknown/unknown' ) {
498 if ( $this->isRecognizableExtension( $ext ) ) {
499 wfDebug( __METHOD__. ': refusing to guess mime type for .' .
500 "$ext file, we should have recognized it\n" );
502 // Not something we can detect, so simply
503 // trust the file extension
504 $mime = $this->guessTypesForExtension( $ext );
507 elseif ( $mime === 'application/x-opc+zip' ) {
508 if ( $this->isMatchingExtension( $ext, $mime ) ) {
509 // A known file extension for an OPC file,
510 // find the proper mime type for that file extension
511 $mime = $this->guessTypesForExtension( $ext );
513 wfDebug( __METHOD__. ": refusing to guess better type for $mime file, " .
514 ".$ext is not a known OPC extension.\n" );
515 $mime = 'application/zip';
519 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
520 $mime = $this->mMimeTypeAliases[$mime];
523 wfDebug(__METHOD__.": improved mime type for .$ext: $mime\n");
528 * Mime type detection. This uses detectMimeType to detect the mime type
529 * of the file, but applies additional checks to determine some well known
530 * file formats that may be missed or misinterpreter by the default mime
531 * detection (namely XML based formats like XHTML or SVG, as well as ZIP
532 * based formats like OPC/ODF files).
534 * @param $file String: the file to check
535 * @param $ext Mixed: the file extension, or true (default) to extract it from the filename.
536 * Set it to false to ignore the extension. DEPRECATED! Set to false, use
537 * improveTypeFromExtension($mime, $ext) later to improve mime type.
539 * @return string the mime type of $file
541 public function guessMimeType( $file, $ext = true ) {
542 if ( $ext ) { // TODO: make $ext default to false. Or better, remove it.
543 wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. " .
544 "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
547 $mime = $this->doGuessMimeType( $file, $ext );
550 wfDebug( __METHOD__.": internal type detection failed for $file (.$ext)...\n" );
551 $mime = $this->detectMimeType( $file, $ext );
554 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
555 $mime = $this->mMimeTypeAliases[$mime];
558 wfDebug(__METHOD__.": guessed mime type of $file: $mime\n");
563 * Guess the mime type from the file contents.
565 * @param string $file
567 * @return bool|string
569 private function doGuessMimeType( $file, $ext ) { // TODO: remove $ext param
570 // Read a chunk of the file
571 wfSuppressWarnings();
572 // @todo FIXME: Shouldn't this be rb?
573 $f = fopen( $file, 'rt' );
577 return 'unknown/unknown';
579 $head = fread( $f, 1024 );
580 fseek( $f, -65558, SEEK_END );
581 $tail = fread( $f, 65558 ); // 65558 = maximum size of a zip EOCDR
584 wfDebug( __METHOD__ . ": analyzing head and tail of $file for magic numbers.\n" );
586 // Hardcode a few magic number checks...
589 'MThd' => 'audio/midi',
590 'OggS' => 'application/ogg',
593 // Note that WMF may have a bare header, no magic number.
594 "\x01\x00\x09\x00" => 'application/x-msmetafile', // Possibly prone to false positives?
595 "\xd7\xcd\xc6\x9a" => 'application/x-msmetafile',
596 '%PDF' => 'application/pdf',
597 'gimp xcf' => 'image/x-xcf',
599 // Some forbidden fruit...
600 'MZ' => 'application/octet-stream', // DOS/Windows executable
601 "\xca\xfe\xba\xbe" => 'application/octet-stream', // Mach-O binary
602 "\x7fELF" => 'application/octet-stream', // ELF binary
605 foreach ( $headers as $magic => $candidate ) {
606 if ( strncmp( $head, $magic, strlen( $magic ) ) == 0 ) {
607 wfDebug( __METHOD__ . ": magic header in $file recognized as $candidate\n" );
612 /* Look for WebM and Matroska files */
613 if ( strncmp( $head, pack( "C4", 0x1a, 0x45, 0xdf, 0xa3 ), 4 ) == 0 ) {
614 $doctype = strpos( $head, "\x42\x82" );
616 // Next byte is datasize, then data (sizes larger than 1 byte are very stupid muxers)
617 $data = substr($head, $doctype+3, 8);
618 if ( strncmp( $data, "matroska", 8 ) == 0 ) {
619 wfDebug( __METHOD__ . ": recognized file as video/x-matroska\n" );
620 return "video/x-matroska";
621 } elseif ( strncmp( $data, "webm", 4 ) == 0 ) {
622 wfDebug( __METHOD__ . ": recognized file as video/webm\n" );
626 wfDebug( __METHOD__ . ": unknown EBML file\n" );
627 return "unknown/unknown";
631 if ( strncmp( $head, "RIFF", 4 ) == 0 && strncmp( substr( $head, 8, 8), "WEBPVP8 ", 8 ) == 0 ) {
632 wfDebug( __METHOD__ . ": recognized file as image/webp\n" );
637 * Look for PHP. Check for this before HTML/XML... Warning: this is a
638 * heuristic, and won't match a file with a lot of non-PHP before. It
639 * will also match text files which could be PHP. :)
641 * @todo FIXME: For this reason, the check is probably useless -- an attacker
642 * could almost certainly just pad the file with a lot of nonsense to
643 * circumvent the check in any case where it would be a security
644 * problem. On the other hand, it causes harmful false positives (bug
645 * 16583). The heuristic has been cut down to exclude three-character
646 * strings like "<? ", but should it be axed completely?
648 if ( ( strpos( $head, '<?php' ) !== false ) ||
650 ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
651 ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
652 ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
653 ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
654 ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {
656 wfDebug( __METHOD__ . ": recognized $file as application/x-php\n" );
657 return 'application/x-php';
661 * look for XML formats (XHTML and SVG)
663 $xml = new XmlTypeCheck( $file );
664 if ( $xml->wellFormed ) {
665 global $wgXMLMimeTypes;
666 if ( isset( $wgXMLMimeTypes[$xml->getRootElement()] ) ) {
667 return $wgXMLMimeTypes[$xml->getRootElement()];
669 return 'application/xml';
674 * look for shell scripts
679 if ( substr( $head, 0, 2) == "#!" ) {
680 $script_type = "ASCII";
681 } elseif ( substr( $head, 0, 5) == "\xef\xbb\xbf#!" ) {
682 $script_type = "UTF-8";
683 } elseif ( substr( $head, 0, 7) == "\xfe\xff\x00#\x00!" ) {
684 $script_type = "UTF-16BE";
685 } elseif ( substr( $head, 0, 7 ) == "\xff\xfe#\x00!" ) {
686 $script_type= "UTF-16LE";
689 if ( $script_type ) {
690 if ( $script_type !== "UTF-8" && $script_type !== "ASCII") {
691 // Quick and dirty fold down to ASCII!
692 $pack = array( 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' );
693 $chars = unpack( $pack[$script_type], substr( $head, 2 ) );
695 foreach( $chars as $codepoint ) {
696 if( $codepoint < 128 ) {
697 $head .= chr( $codepoint );
706 if ( preg_match( '%/?([^\s]+/)(\w+)%', $head, $match ) ) {
707 $mime = "application/x-{$match[2]}";
708 wfDebug( __METHOD__.": shell script recognized as $mime\n" );
713 // Check for ZIP variants (before getimagesize)
714 if ( strpos( $tail, "PK\x05\x06" ) !== false ) {
715 wfDebug( __METHOD__.": ZIP header present in $file\n" );
716 return $this->detectZipType( $head, $tail, $ext );
719 wfSuppressWarnings();
720 $gis = getimagesize( $file );
723 if( $gis && isset( $gis['mime'] ) ) {
724 $mime = $gis['mime'];
725 wfDebug( __METHOD__.": getimagesize detected $file as $mime\n" );
730 $deja = new DjVuImage( $file );
731 if( $deja->isValid() ) {
732 wfDebug( __METHOD__.": detected $file as image/vnd.djvu\n" );
733 return 'image/vnd.djvu';
740 * Detect application-specific file type of a given ZIP file from its
741 * header data. Currently works for OpenDocument and OpenXML types...
742 * If can't tell, returns 'application/zip'.
744 * @param $header String: some reasonably-sized chunk of file header
745 * @param $tail String: the tail of the file
746 * @param $ext Mixed: the file extension, or true to extract it from the filename.
747 * Set it to false (default) to ignore the extension. DEPRECATED! Set to false,
748 * use improveTypeFromExtension($mime, $ext) later to improve mime type.
752 function detectZipType( $header, $tail = null, $ext = false ) {
753 if( $ext ) { # TODO: remove $ext param
754 wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. " .
755 "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
758 $mime = 'application/zip';
759 $opendocTypes = array(
768 'presentation-template',
770 'spreadsheet-template',
777 // http://lists.oasis-open.org/archives/office/200505/msg00006.html
778 $types = '(?:' . implode( '|', $opendocTypes ) . ')';
779 $opendocRegex = "/^mimetype(application\/vnd\.oasis\.opendocument\.$types)/";
781 $openxmlRegex = "/^\[Content_Types\].xml/";
783 if ( preg_match( $opendocRegex, substr( $header, 30 ), $matches ) ) {
785 wfDebug( __METHOD__.": detected $mime from ZIP archive\n" );
786 } elseif ( preg_match( $openxmlRegex, substr( $header, 30 ) ) ) {
787 $mime = "application/x-opc+zip";
788 # TODO: remove the block below, as soon as improveTypeFromExtension is used everywhere
789 if ( $ext !== true && $ext !== false ) {
790 /** This is the mode used by getPropsFromPath
791 * These mime's are stored in the database, where we don't really want
792 * x-opc+zip, because we use it only for internal purposes
794 if ( $this->isMatchingExtension( $ext, $mime) ) {
795 /* A known file extension for an OPC file,
796 * find the proper mime type for that file extension */
797 $mime = $this->guessTypesForExtension( $ext );
799 $mime = "application/zip";
802 wfDebug( __METHOD__.": detected an Open Packaging Conventions archive: $mime\n" );
803 } elseif ( substr( $header, 0, 8 ) == "\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1" &&
804 ($headerpos = strpos( $tail, "PK\x03\x04" ) ) !== false &&
805 preg_match( $openxmlRegex, substr( $tail, $headerpos + 30 ) ) ) {
806 if ( substr( $header, 512, 4) == "\xEC\xA5\xC1\x00" ) {
807 $mime = "application/msword";
809 switch( substr( $header, 512, 6) ) {
810 case "\xEC\xA5\xC1\x00\x0E\x00":
811 case "\xEC\xA5\xC1\x00\x1C\x00":
812 case "\xEC\xA5\xC1\x00\x43\x00":
813 $mime = "application/vnd.ms-powerpoint";
815 case "\xFD\xFF\xFF\xFF\x10\x00":
816 case "\xFD\xFF\xFF\xFF\x1F\x00":
817 case "\xFD\xFF\xFF\xFF\x22\x00":
818 case "\xFD\xFF\xFF\xFF\x23\x00":
819 case "\xFD\xFF\xFF\xFF\x28\x00":
820 case "\xFD\xFF\xFF\xFF\x29\x00":
821 case "\xFD\xFF\xFF\xFF\x10\x02":
822 case "\xFD\xFF\xFF\xFF\x1F\x02":
823 case "\xFD\xFF\xFF\xFF\x22\x02":
824 case "\xFD\xFF\xFF\xFF\x23\x02":
825 case "\xFD\xFF\xFF\xFF\x28\x02":
826 case "\xFD\xFF\xFF\xFF\x29\x02":
827 $mime = "application/vnd.msexcel";
831 wfDebug( __METHOD__.": detected a MS Office document with OPC trailer\n");
833 wfDebug( __METHOD__.": unable to identify type of ZIP archive\n" );
839 * Internal mime type detection. Detection is done using an external
840 * program, if $wgMimeDetectorCommand is set. Otherwise, the fileinfo
841 * extension and mime_content_type are tried (in this order), if they
842 * are available. If the dections fails and $ext is not false, the mime
843 * type is guessed from the file extension, using guessTypesForExtension.
845 * If the mime type is still unknown, getimagesize is used to detect the
846 * mime type if the file is an image. If no mime type can be determined,
847 * this function returns 'unknown/unknown'.
849 * @param $file String: the file to check
850 * @param $ext Mixed: the file extension, or true (default) to extract it from the filename.
851 * Set it to false to ignore the extension. DEPRECATED! Set to false, use
852 * improveTypeFromExtension($mime, $ext) later to improve mime type.
854 * @return string the mime type of $file
856 private function detectMimeType( $file, $ext = true ) {
857 global $wgMimeDetectorCommand;
859 if ( $ext ) { # TODO: make $ext default to false. Or better, remove it.
860 wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
864 if ( $wgMimeDetectorCommand ) {
865 // @todo FIXME: Use wfShellExec
866 $fn = wfEscapeShellArg( $file );
867 $m = `$wgMimeDetectorCommand $fn`;
868 } elseif ( function_exists( "finfo_open" ) && function_exists( "finfo_file" ) ) {
870 # This required the fileinfo extension by PECL,
871 # see http://pecl.php.net/package/fileinfo
872 # This must be compiled into PHP
874 # finfo is the official replacement for the deprecated
875 # mime_content_type function, see below.
877 # If you may need to load the fileinfo extension at runtime, set
878 # $wgLoadFileinfoExtension in LocalSettings.php
880 $mime_magic_resource = finfo_open( FILEINFO_MIME ); /* return mime type ala mimetype extension */
882 if ( $mime_magic_resource ) {
883 $m = finfo_file( $mime_magic_resource, $file );
884 finfo_close( $mime_magic_resource );
886 wfDebug( __METHOD__.": finfo_open failed on ".FILEINFO_MIME."!\n" );
888 } elseif ( function_exists( "mime_content_type" ) ) {
890 # NOTE: this function is available since PHP 4.3.0, but only if
891 # PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
893 # On Windows, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
894 # sometimes, this may even be needed under linus/unix.
896 # Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
897 # see http://www.php.net/manual/en/ref.mime-magic.php for details.
899 $m = mime_content_type($file);
901 wfDebug( __METHOD__.": no magic mime detector found!\n" );
906 $m = preg_replace( '![;, ].*$!', '', $m ); #strip charset, etc
908 $m = strtolower( $m );
910 if ( strpos( $m, 'unknown' ) !== false ) {
913 wfDebug( __METHOD__.": magic mime type of $file: $m\n" );
918 // If desired, look at extension as a fallback.
919 if ( $ext === true ) {
920 $i = strrpos( $file, '.' );
921 $ext = strtolower( $i ? substr( $file, $i + 1 ) : '' );
924 if( $this->isRecognizableExtension( $ext ) ) {
925 wfDebug( __METHOD__. ": refusing to guess mime type for .$ext file, we should have recognized it\n" );
927 $m = $this->guessTypesForExtension( $ext );
929 wfDebug( __METHOD__.": extension mime type of $file: $m\n" );
936 wfDebug( __METHOD__ . ": failed to guess mime type for $file!\n" );
937 return 'unknown/unknown';
941 * Determine the media type code for a file, using its mime type, name and
942 * possibly its contents.
944 * This function relies on the findMediaType(), mapping extensions and mime
945 * types to media types.
947 * @todo analyse file if need be
948 * @todo look at multiple extension, separately and together.
950 * @param $path String: full path to the image file, in case we have to look at the contents
951 * (if null, only the mime type is used to determine the media type code).
952 * @param $mime String: mime type. If null it will be guessed using guessMimeType.
954 * @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
956 function getMediaType( $path = null, $mime = null ) {
957 if( !$mime && !$path ) {
958 return MEDIATYPE_UNKNOWN;
961 // If mime type is unknown, guess it
963 $mime = $this->guessMimeType( $path, false );
966 // Special code for ogg - detect if it's video (theora),
967 // else label it as sound.
968 if ( $mime == 'application/ogg' && file_exists( $path ) ) {
970 // Read a chunk of the file
971 $f = fopen( $path, "rt" );
972 if ( !$f ) return MEDIATYPE_UNKNOWN;
973 $head = fread( $f, 256 );
976 $head = strtolower( $head );
978 // This is an UGLY HACK, file should be parsed correctly
979 if ( strpos( $head, 'theora' ) !== false ) return MEDIATYPE_VIDEO;
980 elseif ( strpos( $head, 'vorbis' ) !== false ) return MEDIATYPE_AUDIO;
981 elseif ( strpos( $head, 'flac' ) !== false ) return MEDIATYPE_AUDIO;
982 elseif ( strpos( $head, 'speex' ) !== false ) return MEDIATYPE_AUDIO;
983 else return MEDIATYPE_MULTIMEDIA;
986 // Check for entry for full mime type
988 $type = $this->findMediaType( $mime );
989 if ( $type !== MEDIATYPE_UNKNOWN ) {
994 // Check for entry for file extension
996 $i = strrpos( $path, '.' );
997 $e = strtolower( $i ? substr( $path, $i + 1 ) : '' );
999 // TODO: look at multi-extension if this fails, parse from full path
1000 $type = $this->findMediaType( '.' . $e );
1001 if ( $type !== MEDIATYPE_UNKNOWN ) {
1006 // Check major mime type
1008 $i = strpos( $mime, '/' );
1009 if ( $i !== false ) {
1010 $major = substr( $mime, 0, $i );
1011 $type = $this->findMediaType( $major );
1012 if ( $type !== MEDIATYPE_UNKNOWN ) {
1019 $type = MEDIATYPE_UNKNOWN;
1026 * Returns a media code matching the given mime type or file extension.
1027 * File extensions are represented by a string starting with a dot (.) to
1028 * distinguish them from mime types.
1030 * This funktion relies on the mapping defined by $this->mMediaTypes
1032 * @return int|string
1034 function findMediaType( $extMime ) {
1035 if ( strpos( $extMime, '.' ) === 0 ) {
1036 // If it's an extension, look up the mime types
1037 $m = $this->getTypesForExtension( substr( $extMime, 1 ) );
1039 return MEDIATYPE_UNKNOWN;
1042 $m = explode( ' ', $m );
1044 // Normalize mime type
1045 if ( isset( $this->mMimeTypeAliases[$extMime] ) ) {
1046 $extMime = $this->mMimeTypeAliases[$extMime];
1049 $m = array( $extMime );
1052 foreach ( $m as $mime ) {
1053 foreach ( $this->mMediaTypes as $type => $codes ) {
1054 if ( in_array($mime, $codes, true ) ) {
1060 return MEDIATYPE_UNKNOWN;
1064 * Get the MIME types that various versions of Internet Explorer would
1065 * detect from a chunk of the content.
1067 * @param $fileName String: the file name (unused at present)
1068 * @param $chunk String: the first 256 bytes of the file
1069 * @param $proposed String: the MIME type proposed by the server
1072 public function getIEMimeTypes( $fileName, $chunk, $proposed ) {
1073 $ca = $this->getIEContentAnalyzer();
1074 return $ca->getRealMimesFromData( $fileName, $chunk, $proposed );
1078 * Get a cached instance of IEContentAnalyzer
1080 * @return IEContentAnalyzer
1082 protected function getIEContentAnalyzer() {
1083 if ( is_null( $this->mIEAnalyzer ) ) {
1084 $this->mIEAnalyzer = new IEContentAnalyzer;
1086 return $this->mIEAnalyzer;