3 * Module defining helper functions for detecting and dealing with MIME types.
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
22 use Psr\Log\LoggerAwareInterface
;
23 use Psr\Log\LoggerInterface
;
26 * Implements functions related to MIME types such as detection and mapping to file extension
30 class MimeAnalyzer
implements LoggerAwareInterface
{
38 protected $initCallback;
40 protected $detectCallback;
42 protected $guessCallback;
44 protected $extCallback;
45 /** @var array Mapping of media types to arrays of MIME types */
46 protected $mediaTypes = null;
47 /** @var array Map of MIME type aliases */
48 protected $mimeTypeAliases = null;
49 /** @var array Map of MIME types to file extensions (as a space separated list) */
50 protected $mimetoExt = null;
52 /** @var array Map of file extensions types to MIME types (as a space separated list) */
53 public $mExtToMime = null; // legacy name; field accessed by hooks
55 /** @var IEContentAnalyzer */
56 protected $IEAnalyzer;
58 /** @var string Extra MIME types, set for example by media handling extensions */
59 private $extraTypes = '';
60 /** @var string Extra MIME info, set for example by media handling extensions */
61 private $extraInfo = '';
63 /** @var LoggerInterface */
67 * Defines a set of well known MIME types
68 * This is used as a fallback to mime.types files.
69 * An extensive list of well known MIME types is provided by
70 * the file mime.types in the includes directory.
72 * This list concatenated with mime.types is used to create a MIME <-> ext
73 * map. Each line contains a MIME type followed by a space separated list of
74 * extensions. If multiple extensions for a single MIME type exist or if
75 * multiple MIME types exist for a single extension then in most cases
76 * MediaWiki assumes that the first extension following the MIME type is the
77 * canonical extension, and the first time a MIME type appears for a certain
78 * extension is considered the canonical MIME type.
80 * (Note that appending the type file list to the end of self::$wellKnownTypes
81 * sucks because you can't redefine canonical types. This could be fixed by
82 * appending self::$wellKnownTypes behind type file list, but who knows
83 * what will break? In practice this probably isn't a problem anyway -- Bryan)
85 protected static $wellKnownTypes = <<<EOT
86 application/ogg ogx ogg ogm ogv oga spx opus
88 application/vnd.oasis.opendocument.chart odc
89 application/vnd.oasis.opendocument.chart-template otc
90 application/vnd.oasis.opendocument.database odb
91 application/vnd.oasis.opendocument.formula odf
92 application/vnd.oasis.opendocument.formula-template otf
93 application/vnd.oasis.opendocument.graphics odg
94 application/vnd.oasis.opendocument.graphics-template otg
95 application/vnd.oasis.opendocument.image odi
96 application/vnd.oasis.opendocument.image-template oti
97 application/vnd.oasis.opendocument.presentation odp
98 application/vnd.oasis.opendocument.presentation-template otp
99 application/vnd.oasis.opendocument.spreadsheet ods
100 application/vnd.oasis.opendocument.spreadsheet-template ots
101 application/vnd.oasis.opendocument.text odt
102 application/vnd.oasis.opendocument.text-master otm
103 application/vnd.oasis.opendocument.text-template ott
104 application/vnd.oasis.opendocument.text-web oth
105 application/javascript js
106 application/x-shockwave-flash swf
107 audio/midi mid midi kar
108 audio/mpeg mpga mpa mp2 mp3
109 audio/x-aiff aif aiff aifc
111 audio/ogg oga spx ogg opus
112 audio/opus opus ogg oga ogg spx
115 image/jpeg jpeg jpg jpe
123 image/x-portable-pixmap ppm
127 video/ogg ogv ogm ogg
132 * Defines a set of well known MIME info entries
133 * This is used as a fallback to mime.info files.
134 * An extensive list of well known MIME types is provided by
135 * the file mime.info in the includes directory.
137 protected static $wellKnownInfo = <<<EOT
138 application/pdf [OFFICE]
139 application/vnd.oasis.opendocument.chart [OFFICE]
140 application/vnd.oasis.opendocument.chart-template [OFFICE]
141 application/vnd.oasis.opendocument.database [OFFICE]
142 application/vnd.oasis.opendocument.formula [OFFICE]
143 application/vnd.oasis.opendocument.formula-template [OFFICE]
144 application/vnd.oasis.opendocument.graphics [OFFICE]
145 application/vnd.oasis.opendocument.graphics-template [OFFICE]
146 application/vnd.oasis.opendocument.image [OFFICE]
147 application/vnd.oasis.opendocument.image-template [OFFICE]
148 application/vnd.oasis.opendocument.presentation [OFFICE]
149 application/vnd.oasis.opendocument.presentation-template [OFFICE]
150 application/vnd.oasis.opendocument.spreadsheet [OFFICE]
151 application/vnd.oasis.opendocument.spreadsheet-template [OFFICE]
152 application/vnd.oasis.opendocument.text [OFFICE]
153 application/vnd.oasis.opendocument.text-template [OFFICE]
154 application/vnd.oasis.opendocument.text-master [OFFICE]
155 application/vnd.oasis.opendocument.text-web [OFFICE]
156 application/javascript text/javascript application/x-javascript [EXECUTABLE]
157 application/x-shockwave-flash [MULTIMEDIA]
161 audio/mp3 audio/mpeg [AUDIO]
162 application/ogg audio/ogg video/ogg [MULTIMEDIA]
163 image/x-bmp image/x-ms-bmp image/bmp [BITMAP]
167 image/svg+xml [DRAWING]
169 image/vnd.djvu [BITMAP]
171 image/x-portable-pixmap [BITMAP]
176 unknown/unknown application/octet-stream application/x-empty [UNKNOWN]
180 * @param array $params Configuration map, includes:
181 * - typeFile: path to file with the list of known MIME types
182 * - infoFile: path to file with the MIME type info
183 * - xmlTypes: map of root element names to XML MIME types
184 * - initCallback: initialization callback that is passed this object [optional]
185 * - detectCallback: alternative to finfo that returns the mime type for a file.
186 * For example, the callback can return the output of "file -bi". [optional]
187 * - guessCallback: callback to improve the guessed MIME type using the file data.
188 * This is intended for fixing mistakes in fileinfo or "detectCallback". [optional]
189 * - extCallback: callback to improve the guessed MIME type using the extension. [optional]
190 * - logger: PSR-3 logger [optional]
191 * @note Constructing these instances is expensive due to file reads.
192 * A service or singleton pattern should be used to avoid creating instances again and again.
194 public function __construct( array $params ) {
195 $this->typeFile
= $params['typeFile'];
196 $this->infoFile
= $params['infoFile'];
197 $this->xmlTypes
= $params['xmlTypes'];
198 $this->initCallback
= $params['initCallback'] ??
null;
199 $this->detectCallback
= $params['detectCallback'] ??
null;
200 $this->guessCallback
= $params['guessCallback'] ??
null;
201 $this->extCallback
= $params['extCallback'] ??
null;
202 $this->logger
= $params['logger'] ??
new \Psr\Log\
NullLogger();
207 protected function loadFiles() {
209 * --- load mime.types ---
212 # Allow media handling extensions adding MIME-types and MIME-info
213 if ( $this->initCallback
) {
214 call_user_func( $this->initCallback
, $this );
217 $types = self
::$wellKnownTypes;
219 $mimeTypeFile = $this->typeFile
;
220 if ( $mimeTypeFile ) {
221 if ( is_file( $mimeTypeFile ) && is_readable( $mimeTypeFile ) ) {
222 $this->logger
->info( __METHOD__
. ": loading mime types from $mimeTypeFile\n" );
224 $types .= file_get_contents( $mimeTypeFile );
226 $this->logger
->info( __METHOD__
. ": can't load mime types from $mimeTypeFile\n" );
229 $this->logger
->info( __METHOD__
.
230 ": no mime types file defined, using built-ins only.\n" );
233 $types .= "\n" . $this->extraTypes
;
235 $types = str_replace( [ "\r\n", "\n\r", "\n\n", "\r\r", "\r" ], "\n", $types );
236 $types = str_replace( "\t", " ", $types );
238 $this->mimetoExt
= [];
239 $this->mExtToMime
= [];
241 $lines = explode( "\n", $types );
242 foreach ( $lines as $s ) {
247 if ( strpos( $s, '#' ) === 0 ) {
251 $s = strtolower( $s );
252 $i = strpos( $s, ' ' );
254 if ( $i === false ) {
258 $mime = substr( $s, 0, $i );
259 $ext = trim( substr( $s, $i +
1 ) );
261 if ( empty( $ext ) ) {
265 if ( !empty( $this->mimetoExt
[$mime] ) ) {
266 $this->mimetoExt
[$mime] .= ' ' . $ext;
268 $this->mimetoExt
[$mime] = $ext;
271 $extensions = explode( ' ', $ext );
273 foreach ( $extensions as $e ) {
279 if ( !empty( $this->mExtToMime
[$e] ) ) {
280 $this->mExtToMime
[$e] .= ' ' . $mime;
282 $this->mExtToMime
[$e] = $mime;
288 * --- load mime.info ---
291 $mimeInfoFile = $this->infoFile
;
293 $info = self
::$wellKnownInfo;
295 if ( $mimeInfoFile ) {
296 if ( is_file( $mimeInfoFile ) && is_readable( $mimeInfoFile ) ) {
297 $this->logger
->info( __METHOD__
. ": loading mime info from $mimeInfoFile\n" );
299 $info .= file_get_contents( $mimeInfoFile );
301 $this->logger
->info( __METHOD__
. ": can't load mime info from $mimeInfoFile\n" );
304 $this->logger
->info( __METHOD__
.
305 ": no mime info file defined, using built-ins only.\n" );
308 $info .= "\n" . $this->extraInfo
;
310 $info = str_replace( [ "\r\n", "\n\r", "\n\n", "\r\r", "\r" ], "\n", $info );
311 $info = str_replace( "\t", " ", $info );
313 $this->mimeTypeAliases
= [];
314 $this->mediaTypes
= [];
316 $lines = explode( "\n", $info );
317 foreach ( $lines as $s ) {
322 if ( strpos( $s, '#' ) === 0 ) {
326 $s = strtolower( $s );
327 $i = strpos( $s, ' ' );
329 if ( $i === false ) {
333 # print "processing MIME INFO line $s<br>";
336 if ( preg_match( '!\[\s*(\w+)\s*\]!', $s, $match ) ) {
337 $s = preg_replace( '!\[\s*(\w+)\s*\]!', '', $s );
338 $mtype = trim( strtoupper( $match[1] ) );
340 $mtype = MEDIATYPE_UNKNOWN
;
343 $m = explode( ' ', $s );
345 if ( !isset( $this->mediaTypes
[$mtype] ) ) {
346 $this->mediaTypes
[$mtype] = [];
349 foreach ( $m as $mime ) {
350 $mime = trim( $mime );
351 if ( empty( $mime ) ) {
355 $this->mediaTypes
[$mtype][] = $mime;
358 if ( count( $m ) > 1 ) {
360 $mCount = count( $m );
361 for ( $i = 1; $i < $mCount; $i +
= 1 ) {
363 $this->mimeTypeAliases
[$mime] = $main;
369 public function setLogger( LoggerInterface
$logger ) {
370 $this->logger
= $logger;
374 * Adds to the list mapping MIME to file extensions.
375 * As an extension author, you are encouraged to submit patches to
376 * MediaWiki's core to add new MIME types to mime.types.
377 * @param string $types
379 public function addExtraTypes( $types ) {
380 $this->extraTypes
.= "\n" . $types;
384 * Adds to the list mapping MIME to media type.
385 * As an extension author, you are encouraged to submit patches to
386 * MediaWiki's core to add new MIME info to mime.info.
387 * @param string $info
389 public function addExtraInfo( $info ) {
390 $this->extraInfo
.= "\n" . $info;
394 * Returns a list of file extensions for a given MIME type as a space
395 * separated string or null if the MIME type was unrecognized. Resolves
398 * @param string $mime
399 * @return string|null
401 public function getExtensionsForType( $mime ) {
402 $mime = strtolower( $mime );
404 // Check the mime-to-ext map
405 if ( isset( $this->mimetoExt
[$mime] ) ) {
406 return $this->mimetoExt
[$mime];
409 // Resolve the MIME type to the canonical type
410 if ( isset( $this->mimeTypeAliases
[$mime] ) ) {
411 $mime = $this->mimeTypeAliases
[$mime];
412 if ( isset( $this->mimetoExt
[$mime] ) ) {
413 return $this->mimetoExt
[$mime];
421 * Returns a list of MIME types for a given file extension as a space
422 * separated string or null if the extension was unrecognized.
425 * @return string|null
427 public function getTypesForExtension( $ext ) {
428 $ext = strtolower( $ext );
430 $r = $this->mExtToMime
[$ext] ??
null;
435 * Returns a single MIME type for a given file extension or null if unknown.
436 * This is always the first type from the list returned by getTypesForExtension($ext).
439 * @return string|null
441 public function guessTypesForExtension( $ext ) {
442 $m = $this->getTypesForExtension( $ext );
443 if ( is_null( $m ) ) {
447 // TODO: Check if this is needed; strtok( $m, ' ' ) should be sufficient
449 $m = preg_replace( '/\s.*$/', '', $m );
455 * Tests if the extension matches the given MIME type. Returns true if a
456 * match was found, null if the MIME type is unknown, and false if the
457 * MIME type is known but no matches where found.
459 * @param string $extension
460 * @param string $mime
463 public function isMatchingExtension( $extension, $mime ) {
464 $ext = $this->getExtensionsForType( $mime );
467 return null; // Unknown MIME type
470 $ext = explode( ' ', $ext );
472 $extension = strtolower( $extension );
473 return in_array( $extension, $ext );
477 * Returns true if the MIME type is known to represent an image format
478 * supported by the PHP GD library.
480 * @param string $mime
484 public function isPHPImageType( $mime ) {
485 // As defined by imagegetsize and image_type_to_mime
487 'image/gif', 'image/jpeg', 'image/png',
488 'image/x-bmp', 'image/xbm', 'image/tiff',
489 'image/jp2', 'image/jpeg2000', 'image/iff',
490 'image/xbm', 'image/x-xbitmap',
491 'image/vnd.wap.wbmp', 'image/vnd.xiff',
493 'application/x-shockwave-flash',
496 return in_array( $mime, $types );
500 * Returns true if the extension represents a type which can
501 * be reliably detected from its content. Use this to determine
502 * whether strict content checks should be applied to reject
503 * invalid uploads; if we can't identify the type we won't
504 * be able to say if it's invalid.
506 * @todo Be more accurate when using fancy MIME detector plugins;
507 * right now this is the bare minimum getimagesize() list.
508 * @param string $extension
511 function isRecognizableExtension( $extension ) {
513 // Types recognized by getimagesize()
514 'gif', 'jpeg', 'jpg', 'png', 'swf', 'psd',
515 'bmp', 'tiff', 'tif', 'jpc', 'jp2',
516 'jpx', 'jb2', 'swc', 'iff', 'wbmp',
519 // Formats we recognize magic numbers for
520 'djvu', 'ogx', 'ogg', 'ogv', 'oga', 'spx', 'opus',
521 'mid', 'pdf', 'wmf', 'xcf', 'webm', 'mkv', 'mka',
524 // XML formats we sure hope we recognize reliably
530 return in_array( strtolower( $extension ), $types );
534 * Improves a MIME type using the file extension. Some file formats are very generic,
535 * so their MIME type is not very meaningful. A more useful MIME type can be derived
536 * by looking at the file extension. Typically, this method would be called on the
537 * result of guessMimeType().
539 * @param string $mime The MIME type, typically guessed from a file's content.
540 * @param string $ext The file extension, as taken from the file name
542 * @return string The MIME type
544 public function improveTypeFromExtension( $mime, $ext ) {
545 if ( $mime === 'unknown/unknown' ) {
546 if ( $this->isRecognizableExtension( $ext ) ) {
547 $this->logger
->info( __METHOD__
. ': refusing to guess mime type for .' .
548 "$ext file, we should have recognized it\n" );
550 // Not something we can detect, so simply
551 // trust the file extension
552 $mime = $this->guessTypesForExtension( $ext );
554 } elseif ( $mime === 'application/x-opc+zip' ) {
555 if ( $this->isMatchingExtension( $ext, $mime ) ) {
556 // A known file extension for an OPC file,
557 // find the proper MIME type for that file extension
558 $mime = $this->guessTypesForExtension( $ext );
560 $this->logger
->info( __METHOD__
.
561 ": refusing to guess better type for $mime file, " .
562 ".$ext is not a known OPC extension.\n" );
563 $mime = 'application/zip';
565 } elseif ( $mime === 'text/plain' && $this->findMediaType( ".$ext" ) === MEDIATYPE_TEXT
) {
566 // Textual types are sometimes not recognized properly.
567 // If detected as text/plain, and has an extension which is textual
568 // improve to the extension's type. For example, csv and json are often
569 // misdetected as text/plain.
570 $mime = $this->guessTypesForExtension( $ext );
573 # Media handling extensions can improve the MIME detected
574 $callback = $this->extCallback
;
576 $callback( $this, $ext, $mime /* by reference */ );
579 if ( isset( $this->mimeTypeAliases
[$mime] ) ) {
580 $mime = $this->mimeTypeAliases
[$mime];
583 $this->logger
->info( __METHOD__
. ": improved mime type for .$ext: $mime\n" );
588 * MIME type detection. This uses detectMimeType to detect the MIME type
589 * of the file, but applies additional checks to determine some well known
590 * file formats that may be missed or misinterpreted by the default MIME
591 * detection (namely XML based formats like XHTML or SVG, as well as ZIP
592 * based formats like OPC/ODF files).
594 * @param string $file The file to check
595 * @param string|bool $ext The file extension, or true (default) to extract
596 * it from the filename. Set it to false to ignore the extension. DEPRECATED!
597 * Set to false, use improveTypeFromExtension($mime, $ext) later to improve MIME type.
599 * @return string The MIME type of $file
601 public function guessMimeType( $file, $ext = true ) {
602 if ( $ext ) { // TODO: make $ext default to false. Or better, remove it.
603 $this->logger
->info( __METHOD__
.
604 ": WARNING: use of the \$ext parameter is deprecated. " .
605 "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
608 $mime = $this->doGuessMimeType( $file, $ext );
611 $this->logger
->info( __METHOD__
.
612 ": internal type detection failed for $file (.$ext)...\n" );
613 $mime = $this->detectMimeType( $file, $ext );
616 if ( isset( $this->mimeTypeAliases
[$mime] ) ) {
617 $mime = $this->mimeTypeAliases
[$mime];
620 $this->logger
->info( __METHOD__
. ": guessed mime type of $file: $mime\n" );
625 * Guess the MIME type from the file contents.
627 * @todo Remove $ext param
629 * @param string $file
631 * @return bool|string
632 * @throws UnexpectedValueException
634 private function doGuessMimeType( $file, $ext ) {
635 // Read a chunk of the file
636 Wikimedia\
suppressWarnings();
637 $f = fopen( $file, 'rb' );
638 Wikimedia\restoreWarnings
();
641 return 'unknown/unknown';
644 $fsize = filesize( $file );
645 if ( $fsize === false ) {
646 return 'unknown/unknown';
649 $head = fread( $f, 1024 );
650 $tailLength = min( 65558, $fsize ); // 65558 = maximum size of a zip EOCDR
651 if ( fseek( $f, -1 * $tailLength, SEEK_END
) === -1 ) {
652 throw new UnexpectedValueException(
653 "Seeking $tailLength bytes from EOF failed in " . __METHOD__
);
655 $tail = $tailLength ?
fread( $f, $tailLength ) : '';
658 $this->logger
->info( __METHOD__
.
659 ": analyzing head and tail of $file for magic numbers.\n" );
661 // Hardcode a few magic number checks...
664 'MThd' => 'audio/midi',
665 'OggS' => 'application/ogg',
666 'ID3' => 'audio/mpeg',
667 "\xff\xfb" => 'audio/mpeg', // MPEG-1 layer 3
668 "\xff\xf3" => 'audio/mpeg', // MPEG-2 layer 3 (lower sample rates)
669 "\xff\xe3" => 'audio/mpeg', // MPEG-2.5 layer 3 (very low sample rates)
672 // Note that WMF may have a bare header, no magic number.
673 "\x01\x00\x09\x00" => 'application/x-msmetafile', // Possibly prone to false positives?
674 "\xd7\xcd\xc6\x9a" => 'application/x-msmetafile',
675 '%PDF' => 'application/pdf',
676 'gimp xcf' => 'image/x-xcf',
678 // Some forbidden fruit...
679 'MZ' => 'application/octet-stream', // DOS/Windows executable
680 "\xca\xfe\xba\xbe" => 'application/octet-stream', // Mach-O binary
681 "\x7fELF" => 'application/octet-stream', // ELF binary
684 foreach ( $headers as $magic => $candidate ) {
685 if ( strncmp( $head, $magic, strlen( $magic ) ) == 0 ) {
686 $this->logger
->info( __METHOD__
.
687 ": magic header in $file recognized as $candidate\n" );
692 /* Look for WebM and Matroska files */
693 if ( strncmp( $head, pack( "C4", 0x1a, 0x45, 0xdf, 0xa3 ), 4 ) == 0 ) {
694 $doctype = strpos( $head, "\x42\x82" );
696 // Next byte is datasize, then data (sizes larger than 1 byte are stupid muxers)
697 $data = substr( $head, $doctype +
3, 8 );
698 if ( strncmp( $data, "matroska", 8 ) == 0 ) {
699 $this->logger
->info( __METHOD__
. ": recognized file as video/x-matroska\n" );
700 return "video/x-matroska";
701 } elseif ( strncmp( $data, "webm", 4 ) == 0 ) {
702 // XXX HACK look for a video track, if we don't find it, this is an audio file
703 $videotrack = strpos( $head, "\x86\x85V_VP" );
706 // There is a video track, so this is a video file.
707 $this->logger
->info( __METHOD__
. ": recognized file as video/webm\n" );
711 $this->logger
->info( __METHOD__
. ": recognized file as audio/webm\n" );
715 $this->logger
->info( __METHOD__
. ": unknown EBML file\n" );
716 return "unknown/unknown";
720 if ( strncmp( $head, "RIFF", 4 ) == 0 &&
721 strncmp( substr( $head, 8, 7 ), "WEBPVP8", 7 ) == 0
723 $this->logger
->info( __METHOD__
. ": recognized file as image/webp\n" );
728 * Look for PHP. Check for this before HTML/XML... Warning: this is a
729 * heuristic, and won't match a file with a lot of non-PHP before. It
730 * will also match text files which could be PHP. :)
732 * @todo FIXME: For this reason, the check is probably useless -- an attacker
733 * could almost certainly just pad the file with a lot of nonsense to
734 * circumvent the check in any case where it would be a security
735 * problem. On the other hand, it causes harmful false positives (bug
736 * 16583). The heuristic has been cut down to exclude three-character
737 * strings like "<? ", but should it be axed completely?
739 if ( ( strpos( $head, '<?php' ) !== false ) ||
740 ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
741 ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
742 ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
743 ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
744 ( strpos( $head, "<\x00?\x00=" ) !== false )
746 $this->logger
->info( __METHOD__
. ": recognized $file as application/x-php\n" );
747 return 'application/x-php';
751 * look for XML formats (XHTML and SVG)
753 $xml = new XmlTypeCheck( $file );
754 if ( $xml->wellFormed
) {
755 $xmlTypes = $this->xmlTypes
;
756 if ( isset( $xmlTypes[$xml->getRootElement()] ) ) {
757 return $xmlTypes[$xml->getRootElement()];
759 return 'application/xml';
764 * look for shell scripts
769 if ( substr( $head, 0, 2 ) == "#!" ) {
770 $script_type = "ASCII";
771 } elseif ( substr( $head, 0, 5 ) == "\xef\xbb\xbf#!" ) {
772 $script_type = "UTF-8";
773 } elseif ( substr( $head, 0, 7 ) == "\xfe\xff\x00#\x00!" ) {
774 $script_type = "UTF-16BE";
775 } elseif ( substr( $head, 0, 7 ) == "\xff\xfe#\x00!" ) {
776 $script_type = "UTF-16LE";
779 if ( $script_type ) {
780 if ( $script_type !== "UTF-8" && $script_type !== "ASCII" ) {
781 // Quick and dirty fold down to ASCII!
782 $pack = [ 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' ];
783 $chars = unpack( $pack[$script_type], substr( $head, 2 ) );
785 foreach ( $chars as $codepoint ) {
786 if ( $codepoint < 128 ) {
787 $head .= chr( $codepoint );
796 if ( preg_match( '%/?([^\s]+/)(\w+)%', $head, $match ) ) {
797 $mime = "application/x-{$match[2]}";
798 $this->logger
->info( __METHOD__
. ": shell script recognized as $mime\n" );
803 // Check for ZIP variants (before getimagesize)
804 if ( strpos( $tail, "PK\x05\x06" ) !== false ) {
805 $this->logger
->info( __METHOD__
. ": ZIP header present in $file\n" );
806 return $this->detectZipType( $head, $tail, $ext );
809 // Check for STL (3D) files
810 // @see https://en.wikipedia.org/wiki/STL_(file_format)
812 stripos( $head, 'SOLID ' ) === 0 &&
813 preg_match( '/\RENDSOLID .*$/i', $tail ) ) {
815 return 'application/sla';
816 } elseif ( $fsize > 84 ) {
818 $triangles = substr( $head, 80, 4 );
819 $triangles = unpack( 'V', $triangles );
820 $triangles = reset( $triangles );
821 if ( $triangles !== false && $fsize === 84 +
( $triangles * 50 ) ) {
822 return 'application/sla';
826 Wikimedia\
suppressWarnings();
827 $gis = getimagesize( $file );
828 Wikimedia\restoreWarnings
();
830 if ( $gis && isset( $gis['mime'] ) ) {
831 $mime = $gis['mime'];
832 $this->logger
->info( __METHOD__
. ": getimagesize detected $file as $mime\n" );
836 # Media handling extensions can guess the MIME by content
837 # It's intentionally here so that if core is wrong about a type (false positive),
838 # people will hopefully nag and submit patches :)
840 # Some strings by reference for performance - assuming well-behaved hooks
841 $callback = $this->guessCallback
;
843 $callback( $this, $head, $tail, $file, $mime /* by reference */ );
850 * Detect application-specific file type of a given ZIP file from its
851 * header data. Currently works for OpenDocument and OpenXML types...
852 * If can't tell, returns 'application/zip'.
854 * @param string $header Some reasonably-sized chunk of file header
855 * @param string|null $tail The tail of the file
856 * @param string|bool $ext The file extension, or true to extract it from the filename.
857 * Set it to false (default) to ignore the extension. DEPRECATED! Set to false,
858 * use improveTypeFromExtension($mime, $ext) later to improve MIME type.
862 function detectZipType( $header, $tail = null, $ext = false ) {
863 if ( $ext ) { # TODO: remove $ext param
864 $this->logger
->info( __METHOD__
.
865 ": WARNING: use of the \$ext parameter is deprecated. " .
866 "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
869 $mime = 'application/zip';
879 'presentation-template',
881 'spreadsheet-template',
888 // https://lists.oasis-open.org/archives/office/200505/msg00006.html
889 $types = '(?:' . implode( '|', $opendocTypes ) . ')';
890 $opendocRegex = "/^mimetype(application\/vnd\.oasis\.opendocument\.$types)/";
892 $openxmlRegex = "/^\[Content_Types\].xml/";
894 if ( preg_match( $opendocRegex, substr( $header, 30 ), $matches ) ) {
896 $this->logger
->info( __METHOD__
. ": detected $mime from ZIP archive\n" );
897 } elseif ( preg_match( $openxmlRegex, substr( $header, 30 ) ) ) {
898 $mime = "application/x-opc+zip";
899 # TODO: remove the block below, as soon as improveTypeFromExtension is used everywhere
900 if ( $ext !== true && $ext !== false ) {
901 /** This is the mode used by getPropsFromPath
902 * These MIME's are stored in the database, where we don't really want
903 * x-opc+zip, because we use it only for internal purposes
905 if ( $this->isMatchingExtension( $ext, $mime ) ) {
906 /* A known file extension for an OPC file,
907 * find the proper mime type for that file extension
909 $mime = $this->guessTypesForExtension( $ext );
911 $mime = "application/zip";
914 $this->logger
->info( __METHOD__
.
915 ": detected an Open Packaging Conventions archive: $mime\n" );
916 } elseif ( substr( $header, 0, 8 ) == "\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1" &&
917 ( $headerpos = strpos( $tail, "PK\x03\x04" ) ) !== false &&
918 preg_match( $openxmlRegex, substr( $tail, $headerpos +
30 ) ) ) {
919 if ( substr( $header, 512, 4 ) == "\xEC\xA5\xC1\x00" ) {
920 $mime = "application/msword";
922 switch ( substr( $header, 512, 6 ) ) {
923 case "\xEC\xA5\xC1\x00\x0E\x00":
924 case "\xEC\xA5\xC1\x00\x1C\x00":
925 case "\xEC\xA5\xC1\x00\x43\x00":
926 $mime = "application/vnd.ms-powerpoint";
928 case "\xFD\xFF\xFF\xFF\x10\x00":
929 case "\xFD\xFF\xFF\xFF\x1F\x00":
930 case "\xFD\xFF\xFF\xFF\x22\x00":
931 case "\xFD\xFF\xFF\xFF\x23\x00":
932 case "\xFD\xFF\xFF\xFF\x28\x00":
933 case "\xFD\xFF\xFF\xFF\x29\x00":
934 case "\xFD\xFF\xFF\xFF\x10\x02":
935 case "\xFD\xFF\xFF\xFF\x1F\x02":
936 case "\xFD\xFF\xFF\xFF\x22\x02":
937 case "\xFD\xFF\xFF\xFF\x23\x02":
938 case "\xFD\xFF\xFF\xFF\x28\x02":
939 case "\xFD\xFF\xFF\xFF\x29\x02":
940 $mime = "application/vnd.msexcel";
944 $this->logger
->info( __METHOD__
.
945 ": detected a MS Office document with OPC trailer\n" );
947 $this->logger
->info( __METHOD__
. ": unable to identify type of ZIP archive\n" );
953 * Internal MIME type detection. Detection is done using the fileinfo
954 * extension if it is available. It can be overriden by callback, which could
955 * use an external program, for example. If detection fails and $ext is not false,
956 * the MIME type is guessed from the file extension, using guessTypesForExtension.
958 * If the MIME type is still unknown, getimagesize is used to detect the
959 * MIME type if the file is an image. If no MIME type can be determined,
960 * this function returns 'unknown/unknown'.
962 * @param string $file The file to check
963 * @param string|bool $ext The file extension, or true (default) to extract it from the filename.
964 * Set it to false to ignore the extension. DEPRECATED! Set to false, use
965 * improveTypeFromExtension($mime, $ext) later to improve MIME type.
967 * @return string The MIME type of $file
969 private function detectMimeType( $file, $ext = true ) {
970 /** @todo Make $ext default to false. Or better, remove it. */
972 $this->logger
->info( __METHOD__
.
973 ": WARNING: use of the \$ext parameter is deprecated. "
974 . "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
977 $callback = $this->detectCallback
;
980 $m = $callback( $file );
982 $m = mime_content_type( $file );
987 $m = preg_replace( '![;, ].*$!', '', $m ); # strip charset, etc
989 $m = strtolower( $m );
991 if ( strpos( $m, 'unknown' ) !== false ) {
994 $this->logger
->info( __METHOD__
. ": magic mime type of $file: $m\n" );
999 // If desired, look at extension as a fallback.
1000 if ( $ext === true ) {
1001 $i = strrpos( $file, '.' );
1002 $ext = strtolower( $i ?
substr( $file, $i +
1 ) : '' );
1005 if ( $this->isRecognizableExtension( $ext ) ) {
1006 $this->logger
->info( __METHOD__
. ": refusing to guess mime type for .$ext file, "
1007 . "we should have recognized it\n" );
1009 $m = $this->guessTypesForExtension( $ext );
1011 $this->logger
->info( __METHOD__
. ": extension mime type of $file: $m\n" );
1018 $this->logger
->info( __METHOD__
. ": failed to guess mime type for $file!\n" );
1019 return 'unknown/unknown';
1023 * Determine the media type code for a file, using its MIME type, name and
1024 * possibly its contents.
1026 * This function relies on the findMediaType(), mapping extensions and MIME
1027 * types to media types.
1029 * @todo analyse file if need be
1030 * @todo look at multiple extension, separately and together.
1032 * @param string|null $path Full path to the image file, in case we have to look at the contents
1033 * (if null, only the MIME type is used to determine the media type code).
1034 * @param string|null $mime MIME type. If null it will be guessed using guessMimeType.
1036 * @return string A value to be used with the MEDIATYPE_xxx constants.
1038 function getMediaType( $path = null, $mime = null ) {
1039 if ( !$mime && !$path ) {
1040 return MEDIATYPE_UNKNOWN
;
1043 // If MIME type is unknown, guess it
1045 $mime = $this->guessMimeType( $path, false );
1048 // Special code for ogg - detect if it's video (theora),
1049 // else label it as sound.
1050 if ( $mime == 'application/ogg' && file_exists( $path ) ) {
1051 // Read a chunk of the file
1052 $f = fopen( $path, "rt" );
1054 return MEDIATYPE_UNKNOWN
;
1056 $head = fread( $f, 256 );
1059 $head = str_replace( 'ffmpeg2theora', '', strtolower( $head ) );
1061 // This is an UGLY HACK, file should be parsed correctly
1062 if ( strpos( $head, 'theora' ) !== false ) {
1063 return MEDIATYPE_VIDEO
;
1064 } elseif ( strpos( $head, 'vorbis' ) !== false ) {
1065 return MEDIATYPE_AUDIO
;
1066 } elseif ( strpos( $head, 'flac' ) !== false ) {
1067 return MEDIATYPE_AUDIO
;
1068 } elseif ( strpos( $head, 'speex' ) !== false ) {
1069 return MEDIATYPE_AUDIO
;
1070 } elseif ( strpos( $head, 'opus' ) !== false ) {
1071 return MEDIATYPE_AUDIO
;
1073 return MEDIATYPE_MULTIMEDIA
;
1078 // Check for entry for full MIME type
1080 $type = $this->findMediaType( $mime );
1081 if ( $type !== MEDIATYPE_UNKNOWN
) {
1086 // Check for entry for file extension
1088 $i = strrpos( $path, '.' );
1089 $e = strtolower( $i ?
substr( $path, $i +
1 ) : '' );
1091 // TODO: look at multi-extension if this fails, parse from full path
1092 $type = $this->findMediaType( '.' . $e );
1093 if ( $type !== MEDIATYPE_UNKNOWN
) {
1098 // Check major MIME type
1100 $i = strpos( $mime, '/' );
1101 if ( $i !== false ) {
1102 $major = substr( $mime, 0, $i );
1103 $type = $this->findMediaType( $major );
1104 if ( $type !== MEDIATYPE_UNKNOWN
) {
1111 $type = MEDIATYPE_UNKNOWN
;
1118 * Returns a media code matching the given MIME type or file extension.
1119 * File extensions are represented by a string starting with a dot (.) to
1120 * distinguish them from MIME types.
1122 * This function relies on the mapping defined by $this->mMediaTypes
1124 * @param string $extMime
1125 * @return int|string
1127 function findMediaType( $extMime ) {
1128 if ( strpos( $extMime, '.' ) === 0 ) {
1129 // If it's an extension, look up the MIME types
1130 $m = $this->getTypesForExtension( substr( $extMime, 1 ) );
1132 return MEDIATYPE_UNKNOWN
;
1135 $m = explode( ' ', $m );
1137 // Normalize MIME type
1138 if ( isset( $this->mimeTypeAliases
[$extMime] ) ) {
1139 $extMime = $this->mimeTypeAliases
[$extMime];
1145 foreach ( $m as $mime ) {
1146 foreach ( $this->mediaTypes
as $type => $codes ) {
1147 if ( in_array( $mime, $codes, true ) ) {
1153 return MEDIATYPE_UNKNOWN
;
1157 * Returns an array of media types (MEDIATYPE_xxx constants)
1161 public function getMediaTypes() {
1162 return array_keys( $this->mediaTypes
);
1166 * Get the MIME types that various versions of Internet Explorer would
1167 * detect from a chunk of the content.
1169 * @param string $fileName The file name (unused at present)
1170 * @param string $chunk The first 256 bytes of the file
1171 * @param string $proposed The MIME type proposed by the server
1174 public function getIEMimeTypes( $fileName, $chunk, $proposed ) {
1175 $ca = $this->getIEContentAnalyzer();
1176 return $ca->getRealMimesFromData( $fileName, $chunk, $proposed );
1180 * Get a cached instance of IEContentAnalyzer
1182 * @return IEContentAnalyzer
1184 protected function getIEContentAnalyzer() {
1185 if ( is_null( $this->IEAnalyzer
) ) {
1186 $this->IEAnalyzer
= new IEContentAnalyzer
;
1188 return $this->IEAnalyzer
;