f86b6051cce1eb0c2768617e10be1ea7eb16cacd
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 stateles, 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 #print "processing MIME line $s<br>";
220 $mime = substr( $s, 0, $i );
221 $ext = trim( substr($s, $i+1 ) );
223 if ( empty( $ext ) ) {
227 if ( !empty( $this->mMimeToExt[$mime] ) ) {
228 $this->mMimeToExt[$mime] .= ' ' . $ext;
230 $this->mMimeToExt[$mime] = $ext;
233 $extensions = explode( ' ', $ext );
235 foreach ( $extensions as $e ) {
241 if ( !empty( $this->mExtToMime[$e] ) ) {
242 $this->mExtToMime[$e] .= ' ' . $mime;
244 $this->mExtToMime[$e] = $mime;
250 * --- load mime.info ---
253 global $wgMimeInfoFile;
254 if ( $wgMimeInfoFile == 'includes/mime.info' ) {
255 $wgMimeInfoFile = "$IP/$wgMimeInfoFile";
258 $info = MM_WELL_KNOWN_MIME_INFO;
260 if ( $wgMimeInfoFile ) {
261 if ( is_file( $wgMimeInfoFile ) and is_readable( $wgMimeInfoFile ) ) {
262 wfDebug( __METHOD__.": loading mime info from $wgMimeInfoFile\n" );
264 $info .= file_get_contents( $wgMimeInfoFile );
266 wfDebug(__METHOD__.": can't load mime info from $wgMimeInfoFile\n");
269 wfDebug(__METHOD__.": no mime info file defined, using build-ins only.\n");
272 $info = str_replace( array( "\r\n", "\n\r", "\n\n", "\r\r", "\r" ), "\n", $info);
273 $info = str_replace( "\t", " ", $info );
275 $this->mMimeTypeAliases = array();
276 $this->mMediaTypes = array();
278 $lines = explode( "\n", $info );
279 foreach ( $lines as $s ) {
284 if ( strpos( $s, '#' ) === 0 ) {
288 $s = strtolower( $s );
289 $i = strpos( $s, ' ' );
291 if ( $i === false ) {
295 #print "processing MIME INFO line $s<br>";
298 if ( preg_match( '!\[\s*(\w+)\s*\]!', $s, $match ) ) {
299 $s = preg_replace( '!\[\s*(\w+)\s*\]!', '', $s );
300 $mtype = trim( strtoupper( $match[1] ) );
302 $mtype = MEDIATYPE_UNKNOWN;
305 $m = explode( ' ', $s );
307 if ( !isset( $this->mMediaTypes[$mtype] ) ) {
308 $this->mMediaTypes[$mtype] = array();
311 foreach ( $m as $mime ) {
312 $mime = trim( $mime );
313 if ( empty( $mime ) ) {
317 $this->mMediaTypes[$mtype][] = $mime;
320 if ( sizeof( $m ) > 1 ) {
322 for ( $i=1; $i<sizeof($m); $i += 1 ) {
324 $this->mMimeTypeAliases[$mime] = $main;
332 * Get an instance of this class
335 public static function &singleton() {
336 if ( !isset( self::$instance ) ) {
337 self::$instance = new MimeMagic;
339 return self::$instance;
343 * Returns a list of file extensions for a given mime type as a space
344 * separated string or null if the mime type was unrecognized. Resolves
347 * @param $mime string
348 * @return string|null
350 public function getExtensionsForType( $mime ) {
351 $mime = strtolower( $mime );
353 // Check the mime-to-ext map
354 if ( isset( $this->mMimeToExt[$mime] ) ) {
355 return $this->mMimeToExt[$mime];
358 // Resolve the mime type to the canonical type
359 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
360 $mime = $this->mMimeTypeAliases[$mime];
361 if ( isset( $this->mMimeToExt[$mime] ) ) {
362 return $this->mMimeToExt[$mime];
370 * Returns a list of mime types for a given file extension as a space
371 * separated string or null if the extension was unrecognized.
374 * @return string|null
376 public function getTypesForExtension( $ext ) {
377 $ext = strtolower( $ext );
379 $r = isset( $this->mExtToMime[$ext] ) ? $this->mExtToMime[$ext] : null;
384 * Returns a single mime type for a given file extension or null if unknown.
385 * This is always the first type from the list returned by getTypesForExtension($ext).
388 * @return string|null
390 public function guessTypesForExtension( $ext ) {
391 $m = $this->getTypesForExtension( $ext );
392 if ( is_null( $m ) ) {
396 // TODO: Check if this is needed; strtok( $m, ' ' ) should be sufficient
398 $m = preg_replace( '/\s.*$/', '', $m );
405 * Tests if the extension matches the given mime type. Returns true if a
406 * match was found, null if the mime type is unknown, and false if the
407 * mime type is known but no matches where found.
409 * @param $extension string
410 * @param $mime string
413 public function isMatchingExtension( $extension, $mime ) {
414 $ext = $this->getExtensionsForType( $mime );
417 return null; // Unknown mime type
420 $ext = explode( ' ', $ext );
422 $extension = strtolower( $extension );
423 return in_array( $extension, $ext );
427 * Returns true if the mime type is known to represent an image format
428 * supported by the PHP GD library.
430 * @param $mime string
434 public function isPHPImageType( $mime ) {
435 // As defined by imagegetsize and image_type_to_mime
436 static $types = array(
437 'image/gif', 'image/jpeg', 'image/png',
438 'image/x-bmp', 'image/xbm', 'image/tiff',
439 'image/jp2', 'image/jpeg2000', 'image/iff',
440 'image/xbm', 'image/x-xbitmap',
441 'image/vnd.wap.wbmp', 'image/vnd.xiff',
443 'application/x-shockwave-flash',
446 return in_array( $mime, $types );
450 * Returns true if the extension represents a type which can
451 * be reliably detected from its content. Use this to determine
452 * whether strict content checks should be applied to reject
453 * invalid uploads; if we can't identify the type we won't
454 * be able to say if it's invalid.
456 * @todo Be more accurate when using fancy mime detector plugins;
457 * right now this is the bare minimum getimagesize() list.
460 function isRecognizableExtension( $extension ) {
461 static $types = array(
462 // Types recognized by getimagesize()
463 'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
464 'bmp', 'tiff', 'tif', 'jpc', 'jp2',
465 'jpx', 'jb2', 'swc', 'iff', 'wbmp',
468 // Formats we recognize magic numbers for
469 'djvu', 'ogx', 'ogg', 'ogv', 'oga', 'spx',
470 'mid', 'pdf', 'wmf', 'xcf', 'webm', 'mkv', 'mka',
473 // XML formats we sure hope we recognize reliably
476 return in_array( strtolower( $extension ), $types );
480 * Improves a mime type using the file extension. Some file formats are very generic,
481 * so their mime type is not very meaningful. A more useful mime type can be derived
482 * by looking at the file extension. Typically, this method would be called on the
483 * result of guessMimeType().
485 * Currently, this method does the following:
487 * If $mime is "unknown/unknown" and isRecognizableExtension( $ext ) returns false,
488 * return the result of guessTypesForExtension($ext).
490 * If $mime is "application/x-opc+zip" and isMatchingExtension( $ext, $mime )
491 * gives true, return the result of guessTypesForExtension($ext).
493 * @param $mime String: the mime type, typically guessed from a file's content.
494 * @param $ext String: the file extension, as taken from the file name
496 * @return string the mime type
498 public function improveTypeFromExtension( $mime, $ext ) {
499 if ( $mime === 'unknown/unknown' ) {
500 if ( $this->isRecognizableExtension( $ext ) ) {
501 wfDebug( __METHOD__. ': refusing to guess mime type for .' .
502 "$ext file, we should have recognized it\n" );
504 // Not something we can detect, so simply
505 // trust the file extension
506 $mime = $this->guessTypesForExtension( $ext );
509 elseif ( $mime === 'application/x-opc+zip' ) {
510 if ( $this->isMatchingExtension( $ext, $mime ) ) {
511 // A known file extension for an OPC file,
512 // find the proper mime type for that file extension
513 $mime = $this->guessTypesForExtension( $ext );
515 wfDebug( __METHOD__. ": refusing to guess better type for $mime file, " .
516 ".$ext is not a known OPC extension.\n" );
517 $mime = 'application/zip';
521 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
522 $mime = $this->mMimeTypeAliases[$mime];
525 wfDebug(__METHOD__.": improved mime type for .$ext: $mime\n");
530 * Mime type detection. This uses detectMimeType to detect the mime type
531 * of the file, but applies additional checks to determine some well known
532 * file formats that may be missed or misinterpreter by the default mime
533 * detection (namely XML based formats like XHTML or SVG, as well as ZIP
534 * based formats like OPC/ODF files).
536 * @param $file String: the file to check
537 * @param $ext Mixed: the file extension, or true (default) to extract it from the filename.
538 * Set it to false to ignore the extension. DEPRECATED! Set to false, use
539 * improveTypeFromExtension($mime, $ext) later to improve mime type.
541 * @return string the mime type of $file
543 public function guessMimeType( $file, $ext = true ) {
544 if ( $ext ) { // TODO: make $ext default to false. Or better, remove it.
545 wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. " .
546 "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
549 $mime = $this->doGuessMimeType( $file, $ext );
552 wfDebug( __METHOD__.": internal type detection failed for $file (.$ext)...\n" );
553 $mime = $this->detectMimeType( $file, $ext );
556 if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
557 $mime = $this->mMimeTypeAliases[$mime];
560 wfDebug(__METHOD__.": guessed mime type of $file: $mime\n");
565 * Guess the mime type from the file contents.
567 * @param string $file
570 private function doGuessMimeType( $file, $ext ) { // TODO: remove $ext param
571 // Read a chunk of the file
572 wfSuppressWarnings();
573 // @todo FIXME: Shouldn't this be rb?
574 $f = fopen( $file, 'rt' );
578 return 'unknown/unknown';
580 $head = fread( $f, 1024 );
581 fseek( $f, -65558, SEEK_END );
582 $tail = fread( $f, 65558 ); // 65558 = maximum size of a zip EOCDR
585 wfDebug( __METHOD__ . ": analyzing head and tail of $file for magic numbers.\n" );
587 // Hardcode a few magic number checks...
590 'MThd' => 'audio/midi',
591 'OggS' => 'application/ogg',
594 // Note that WMF may have a bare header, no magic number.
595 "\x01\x00\x09\x00" => 'application/x-msmetafile', // Possibly prone to false positives?
596 "\xd7\xcd\xc6\x9a" => 'application/x-msmetafile',
597 '%PDF' => 'application/pdf',
598 'gimp xcf' => 'image/x-xcf',
600 // Some forbidden fruit...
601 'MZ' => 'application/octet-stream', // DOS/Windows executable
602 "\xca\xfe\xba\xbe" => 'application/octet-stream', // Mach-O binary
603 "\x7fELF" => 'application/octet-stream', // ELF binary
606 foreach ( $headers as $magic => $candidate ) {
607 if ( strncmp( $head, $magic, strlen( $magic ) ) == 0 ) {
608 wfDebug( __METHOD__ . ": magic header in $file recognized as $candidate\n" );
613 /* Look for WebM and Matroska files */
614 if ( strncmp( $head, pack( "C4", 0x1a, 0x45, 0xdf, 0xa3 ), 4 ) == 0 ) {
615 $doctype = strpos( $head, "\x42\x82" );
617 // Next byte is datasize, then data (sizes larger than 1 byte are very stupid muxers)
618 $data = substr($head, $doctype+3, 8);
619 if ( strncmp( $data, "matroska", 8 ) == 0 ) {
620 wfDebug( __METHOD__ . ": recognized file as video/x-matroska\n" );
621 return "video/x-matroska";
622 } elseif ( strncmp( $data, "webm", 4 ) == 0 ) {
623 wfDebug( __METHOD__ . ": recognized file as video/webm\n" );
627 wfDebug( __METHOD__ . ": unknown EBML file\n" );
628 return "unknown/unknown";
632 if ( strncmp( $head, "RIFF", 4 ) == 0 && strncmp( substr( $head, 8, 8), "WEBPVP8 ", 8 ) == 0 ) {
633 wfDebug( __METHOD__ . ": recognized file as image/webp\n" );
638 * Look for PHP. Check for this before HTML/XML... Warning: this is a
639 * heuristic, and won't match a file with a lot of non-PHP before. It
640 * will also match text files which could be PHP. :)
642 * @todo FIXME: For this reason, the check is probably useless -- an attacker
643 * could almost certainly just pad the file with a lot of nonsense to
644 * circumvent the check in any case where it would be a security
645 * problem. On the other hand, it causes harmful false positives (bug
646 * 16583). The heuristic has been cut down to exclude three-character
647 * strings like "<? ", but should it be axed completely?
649 if ( ( strpos( $head, '<?php' ) !== false ) ||
651 ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
652 ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
653 ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
654 ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
655 ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {
657 wfDebug( __METHOD__ . ": recognized $file as application/x-php\n" );
658 return 'application/x-php';
662 * look for XML formats (XHTML and SVG)
664 $xml = new XmlTypeCheck( $file );
665 if ( $xml->wellFormed ) {
666 global $wgXMLMimeTypes;
667 if ( isset( $wgXMLMimeTypes[$xml->getRootElement()] ) ) {
668 return $wgXMLMimeTypes[$xml->getRootElement()];
670 return 'application/xml';
675 * look for shell scripts
680 if ( substr( $head, 0, 2) == "#!" ) {
681 $script_type = "ASCII";
682 } elseif ( substr( $head, 0, 5) == "\xef\xbb\xbf#!" ) {
683 $script_type = "UTF-8";
684 } elseif ( substr( $head, 0, 7) == "\xfe\xff\x00#\x00!" ) {
685 $script_type = "UTF-16BE";
686 } elseif ( substr( $head, 0, 7 ) == "\xff\xfe#\x00!" ) {
687 $script_type= "UTF-16LE";
690 if ( $script_type ) {
691 if ( $script_type !== "UTF-8" && $script_type !== "ASCII") {
692 // Quick and dirty fold down to ASCII!
693 $pack = array( 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' );
694 $chars = unpack( $pack[$script_type], substr( $head, 2 ) );
696 foreach( $chars as $codepoint ) {
697 if( $codepoint < 128 ) {
698 $head .= chr( $codepoint );
707 if ( preg_match( '%/?([^\s]+/)(\w+)%', $head, $match ) ) {
708 $mime = "application/x-{$match[2]}";
709 wfDebug( __METHOD__.": shell script recognized as $mime\n" );
714 // Check for ZIP variants (before getimagesize)
715 if ( strpos( $tail, "PK\x05\x06" ) !== false ) {
716 wfDebug( __METHOD__.": ZIP header present in $file\n" );
717 return $this->detectZipType( $head, $tail, $ext );
720 wfSuppressWarnings();
721 $gis = getimagesize( $file );
724 if( $gis && isset( $gis['mime'] ) ) {
725 $mime = $gis['mime'];
726 wfDebug( __METHOD__.": getimagesize detected $file as $mime\n" );
731 $deja = new DjVuImage( $file );
732 if( $deja->isValid() ) {
733 wfDebug( __METHOD__.": detected $file as image/vnd.djvu\n" );
734 return 'image/vnd.djvu';
741 * Detect application-specific file type of a given ZIP file from its
742 * header data. Currently works for OpenDocument and OpenXML types...
743 * If can't tell, returns 'application/zip'.
745 * @param $header String: some reasonably-sized chunk of file header
746 * @param $tail String: the tail of the file
747 * @param $ext Mixed: the file extension, or true to extract it from the filename.
748 * Set it to false (default) to ignore the extension. DEPRECATED! Set to false,
749 * use improveTypeFromExtension($mime, $ext) later to improve mime type.
753 function detectZipType( $header, $tail = null, $ext = false ) {
754 if( $ext ) { # TODO: remove $ext param
755 wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. " .
756 "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
759 $mime = 'application/zip';
760 $opendocTypes = array(
769 'presentation-template',
771 'spreadsheet-template',
778 // http://lists.oasis-open.org/archives/office/200505/msg00006.html
779 $types = '(?:' . implode( '|', $opendocTypes ) . ')';
780 $opendocRegex = "/^mimetype(application\/vnd\.oasis\.opendocument\.$types)/";
782 $openxmlRegex = "/^\[Content_Types\].xml/";
784 if ( preg_match( $opendocRegex, substr( $header, 30 ), $matches ) ) {
786 wfDebug( __METHOD__.": detected $mime from ZIP archive\n" );
787 } elseif ( preg_match( $openxmlRegex, substr( $header, 30 ) ) ) {
788 $mime = "application/x-opc+zip";
789 # TODO: remove the block below, as soon as improveTypeFromExtension is used everywhere
790 if ( $ext !== true && $ext !== false ) {
791 /** This is the mode used by getPropsFromPath
792 * These mime's are stored in the database, where we don't really want
793 * x-opc+zip, because we use it only for internal purposes
795 if ( $this->isMatchingExtension( $ext, $mime) ) {
796 /* A known file extension for an OPC file,
797 * find the proper mime type for that file extension */
798 $mime = $this->guessTypesForExtension( $ext );
800 $mime = "application/zip";
803 wfDebug( __METHOD__.": detected an Open Packaging Conventions archive: $mime\n" );
804 } elseif ( substr( $header, 0, 8 ) == "\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1" &&
805 ($headerpos = strpos( $tail, "PK\x03\x04" ) ) !== false &&
806 preg_match( $openxmlRegex, substr( $tail, $headerpos + 30 ) ) ) {
807 if ( substr( $header, 512, 4) == "\xEC\xA5\xC1\x00" ) {
808 $mime = "application/msword";
810 switch( substr( $header, 512, 6) ) {
811 case "\xEC\xA5\xC1\x00\x0E\x00":
812 case "\xEC\xA5\xC1\x00\x1C\x00":
813 case "\xEC\xA5\xC1\x00\x43\x00":
814 $mime = "application/vnd.ms-powerpoint";
816 case "\xFD\xFF\xFF\xFF\x10\x00":
817 case "\xFD\xFF\xFF\xFF\x1F\x00":
818 case "\xFD\xFF\xFF\xFF\x22\x00":
819 case "\xFD\xFF\xFF\xFF\x23\x00":
820 case "\xFD\xFF\xFF\xFF\x28\x00":
821 case "\xFD\xFF\xFF\xFF\x29\x00":
822 case "\xFD\xFF\xFF\xFF\x10\x02":
823 case "\xFD\xFF\xFF\xFF\x1F\x02":
824 case "\xFD\xFF\xFF\xFF\x22\x02":
825 case "\xFD\xFF\xFF\xFF\x23\x02":
826 case "\xFD\xFF\xFF\xFF\x28\x02":
827 case "\xFD\xFF\xFF\xFF\x29\x02":
828 $mime = "application/vnd.msexcel";
832 wfDebug( __METHOD__.": detected a MS Office document with OPC trailer\n");
834 wfDebug( __METHOD__.": unable to identify type of ZIP archive\n" );
840 * Internal mime type detection. Detection is done using an external
841 * program, if $wgMimeDetectorCommand is set. Otherwise, the fileinfo
842 * extension and mime_content_type are tried (in this order), if they
843 * are available. If the dections fails and $ext is not false, the mime
844 * type is guessed from the file extension, using guessTypesForExtension.
846 * If the mime type is still unknown, getimagesize is used to detect the
847 * mime type if the file is an image. If no mime type can be determined,
848 * this function returns 'unknown/unknown'.
850 * @param $file String: the file to check
851 * @param $ext Mixed: the file extension, or true (default) to extract it from the filename.
852 * Set it to false to ignore the extension. DEPRECATED! Set to false, use
853 * improveTypeFromExtension($mime, $ext) later to improve mime type.
855 * @return string the mime type of $file
857 private function detectMimeType( $file, $ext = true ) {
858 global $wgMimeDetectorCommand;
860 if ( $ext ) { # TODO: make $ext default to false. Or better, remove it.
861 wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
865 if ( $wgMimeDetectorCommand ) {
866 // @todo FIXME: Use wfShellExec
867 $fn = wfEscapeShellArg( $file );
868 $m = `$wgMimeDetectorCommand $fn`;
869 } elseif ( function_exists( "finfo_open" ) && function_exists( "finfo_file" ) ) {
871 # This required the fileinfo extension by PECL,
872 # see http://pecl.php.net/package/fileinfo
873 # This must be compiled into PHP
875 # finfo is the official replacement for the deprecated
876 # mime_content_type function, see below.
878 # If you may need to load the fileinfo extension at runtime, set
879 # $wgLoadFileinfoExtension in LocalSettings.php
881 $mime_magic_resource = finfo_open( FILEINFO_MIME ); /* return mime type ala mimetype extension */
883 if ( $mime_magic_resource ) {
884 $m = finfo_file( $mime_magic_resource, $file );
885 finfo_close( $mime_magic_resource );
887 wfDebug( __METHOD__.": finfo_open failed on ".FILEINFO_MIME."!\n" );
889 } elseif ( function_exists( "mime_content_type" ) ) {
891 # NOTE: this function is available since PHP 4.3.0, but only if
892 # PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
894 # On Windows, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
895 # sometimes, this may even be needed under linus/unix.
897 # Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
898 # see http://www.php.net/manual/en/ref.mime-magic.php for details.
900 $m = mime_content_type($file);
902 wfDebug( __METHOD__.": no magic mime detector found!\n" );
907 $m = preg_replace( '![;, ].*$!', '', $m ); #strip charset, etc
909 $m = strtolower( $m );
911 if ( strpos( $m, 'unknown' ) !== false ) {
914 wfDebug( __METHOD__.": magic mime type of $file: $m\n" );
919 // If desired, look at extension as a fallback.
920 if ( $ext === true ) {
921 $i = strrpos( $file, '.' );
922 $ext = strtolower( $i ? substr( $file, $i + 1 ) : '' );
925 if( $this->isRecognizableExtension( $ext ) ) {
926 wfDebug( __METHOD__. ": refusing to guess mime type for .$ext file, we should have recognized it\n" );
928 $m = $this->guessTypesForExtension( $ext );
930 wfDebug( __METHOD__.": extension mime type of $file: $m\n" );
937 wfDebug( __METHOD__ . ": failed to guess mime type for $file!\n" );
938 return 'unknown/unknown';
942 * Determine the media type code for a file, using its mime type, name and
943 * possibly its contents.
945 * This function relies on the findMediaType(), mapping extensions and mime
946 * types to media types.
948 * @todo analyse file if need be
949 * @todo look at multiple extension, separately and together.
951 * @param $path String: full path to the image file, in case we have to look at the contents
952 * (if null, only the mime type is used to determine the media type code).
953 * @param $mime String: mime type. If null it will be guessed using guessMimeType.
955 * @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
957 function getMediaType( $path = null, $mime = null ) {
958 if( !$mime && !$path ) {
959 return MEDIATYPE_UNKNOWN;
962 // If mime type is unknown, guess it
964 $mime = $this->guessMimeType( $path, false );
967 // Special code for ogg - detect if it's video (theora),
968 // else label it as sound.
969 if ( $mime == 'application/ogg' && file_exists( $path ) ) {
971 // Read a chunk of the file
972 $f = fopen( $path, "rt" );
973 if ( !$f ) return MEDIATYPE_UNKNOWN;
974 $head = fread( $f, 256 );
977 $head = strtolower( $head );
979 // This is an UGLY HACK, file should be parsed correctly
980 if ( strpos( $head, 'theora' ) !== false ) return MEDIATYPE_VIDEO;
981 elseif ( strpos( $head, 'vorbis' ) !== false ) return MEDIATYPE_AUDIO;
982 elseif ( strpos( $head, 'flac' ) !== false ) return MEDIATYPE_AUDIO;
983 elseif ( strpos( $head, 'speex' ) !== false ) return MEDIATYPE_AUDIO;
984 else return MEDIATYPE_MULTIMEDIA;
987 // Check for entry for full mime type
989 $type = $this->findMediaType( $mime );
990 if ( $type !== MEDIATYPE_UNKNOWN ) {
995 // Check for entry for file extension
997 $i = strrpos( $path, '.' );
998 $e = strtolower( $i ? substr( $path, $i + 1 ) : '' );
1000 // TODO: look at multi-extension if this fails, parse from full path
1001 $type = $this->findMediaType( '.' . $e );
1002 if ( $type !== MEDIATYPE_UNKNOWN ) {
1007 // Check major mime type
1009 $i = strpos( $mime, '/' );
1010 if ( $i !== false ) {
1011 $major = substr( $mime, 0, $i );
1012 $type = $this->findMediaType( $major );
1013 if ( $type !== MEDIATYPE_UNKNOWN ) {
1020 $type = MEDIATYPE_UNKNOWN;
1027 * Returns a media code matching the given mime type or file extension.
1028 * File extensions are represented by a string starting with a dot (.) to
1029 * distinguish them from mime types.
1031 * This funktion relies on the mapping defined by $this->mMediaTypes
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
1071 public function getIEMimeTypes( $fileName, $chunk, $proposed ) {
1072 $ca = $this->getIEContentAnalyzer();
1073 return $ca->getRealMimesFromData( $fileName, $chunk, $proposed );
1077 * Get a cached instance of IEContentAnalyzer
1079 * @return IEContentAnalyzer
1081 protected function getIEContentAnalyzer() {
1082 if ( is_null( $this->mIEAnalyzer ) ) {
1083 $this->mIEAnalyzer = new IEContentAnalyzer;
1085 return $this->mIEAnalyzer;