3 * Base class for the backend of file upload.
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
25 * @defgroup Upload Upload related
31 * UploadBase and subclasses are the backend of MediaWiki's file uploads.
32 * The frontends are formed by ApiUpload and SpecialUpload.
34 * @author Brion Vibber
35 * @author Bryan Tong Minh
36 * @author Michael Dale
38 abstract class UploadBase
{
39 /** @var string Local file system path to the file to upload (or a local copy) */
41 /** @var TempFSFile|null Wrapper to handle deleting the temp file */
42 protected $tempFileObj;
44 protected $mDesiredDestName, $mDestName, $mRemoveTempFile, $mSourceType;
45 protected $mTitle = false, $mTitleError = 0;
46 protected $mFilteredName, $mFinalExtension;
47 protected $mLocalFile, $mFileSize, $mFileProps;
48 protected $mBlackListedExtensions;
49 protected $mJavaDetected, $mSVGNSError;
51 protected static $safeXmlEncodings = [
62 const MIN_LENGTH_PARTNAME
= 4;
63 const ILLEGAL_FILENAME
= 5;
64 const OVERWRITE_EXISTING_FILE
= 7; # Not used anymore; handled by verifyTitlePermissions()
65 const FILETYPE_MISSING
= 8;
66 const FILETYPE_BADTYPE
= 9;
67 const VERIFICATION_ERROR
= 10;
68 const HOOK_ABORTED
= 11;
69 const FILE_TOO_LARGE
= 12;
70 const WINDOWS_NONASCII_FILENAME
= 13;
71 const FILENAME_TOO_LONG
= 14;
77 public function getVerificationErrorCode( $error ) {
79 self
::EMPTY_FILE
=> 'empty-file',
80 self
::FILE_TOO_LARGE
=> 'file-too-large',
81 self
::FILETYPE_MISSING
=> 'filetype-missing',
82 self
::FILETYPE_BADTYPE
=> 'filetype-banned',
83 self
::MIN_LENGTH_PARTNAME
=> 'filename-tooshort',
84 self
::ILLEGAL_FILENAME
=> 'illegal-filename',
85 self
::OVERWRITE_EXISTING_FILE
=> 'overwrite',
86 self
::VERIFICATION_ERROR
=> 'verification-error',
87 self
::HOOK_ABORTED
=> 'hookaborted',
88 self
::WINDOWS_NONASCII_FILENAME
=> 'windows-nonascii-filename',
89 self
::FILENAME_TOO_LONG
=> 'filename-toolong',
91 if ( isset( $code_to_status[$error] ) ) {
92 return $code_to_status[$error];
95 return 'unknown-error';
99 * Returns true if uploads are enabled.
100 * Can be override by subclasses.
103 public static function isEnabled() {
104 global $wgEnableUploads;
106 if ( !$wgEnableUploads ) {
110 # Check php's file_uploads setting
111 return wfIsHHVM() ||
wfIniGetBool( 'file_uploads' );
115 * Returns true if the user can use this upload module or else a string
116 * identifying the missing permission.
117 * Can be overridden by subclasses.
120 * @return bool|string
122 public static function isAllowed( $user ) {
123 foreach ( [ 'upload', 'edit' ] as $permission ) {
124 if ( !$user->isAllowed( $permission ) ) {
133 * Returns true if the user has surpassed the upload rate limit, false otherwise.
138 public static function isThrottled( $user ) {
139 return $user->pingLimiter( 'upload' );
142 // Upload handlers. Should probably just be a global.
143 private static $uploadHandlers = [ 'Stash', 'File', 'Url' ];
146 * Create a form of UploadBase depending on wpSourceType and initializes it
148 * @param WebRequest $request
149 * @param string|null $type
150 * @return null|UploadBase
152 public static function createFromRequest( &$request, $type = null ) {
153 $type = $type ?
$type : $request->getVal( 'wpSourceType', 'File' );
159 // Get the upload class
160 $type = ucfirst( $type );
162 // Give hooks the chance to handle this request
164 Hooks
::run( 'UploadCreateFromRequest', [ $type, &$className ] );
165 if ( is_null( $className ) ) {
166 $className = 'UploadFrom' . $type;
167 wfDebug( __METHOD__
. ": class name: $className\n" );
168 if ( !in_array( $type, self
::$uploadHandlers ) ) {
173 // Check whether this upload class is enabled
174 if ( !call_user_func( [ $className, 'isEnabled' ] ) ) {
178 // Check whether the request is valid
179 if ( !call_user_func( [ $className, 'isValidRequest' ], $request ) ) {
183 /** @var UploadBase $handler */
184 $handler = new $className;
186 $handler->initializeFromRequest( $request );
192 * Check whether a request if valid for this handler
193 * @param WebRequest $request
196 public static function isValidRequest( $request ) {
200 public function __construct() {
204 * Returns the upload type. Should be overridden by child classes
209 public function getSourceType() {
214 * Initialize the path information
215 * @param string $name The desired destination name
216 * @param string $tempPath The temporary path
217 * @param int $fileSize The file size
218 * @param bool $removeTempFile (false) remove the temporary file?
219 * @throws MWException
221 public function initializePathInfo( $name, $tempPath, $fileSize, $removeTempFile = false ) {
222 $this->mDesiredDestName
= $name;
223 if ( FileBackend
::isStoragePath( $tempPath ) ) {
224 throw new MWException( __METHOD__
. " given storage path `$tempPath`." );
227 $this->setTempFile( $tempPath, $fileSize );
228 $this->mRemoveTempFile
= $removeTempFile;
232 * Initialize from a WebRequest. Override this in a subclass.
234 * @param WebRequest $request
236 abstract public function initializeFromRequest( &$request );
239 * @param string $tempPath File system path to temporary file containing the upload
240 * @param integer $fileSize
242 protected function setTempFile( $tempPath, $fileSize = null ) {
243 $this->mTempPath
= $tempPath;
244 if ( strlen( $this->mTempPath
) && file_exists( $this->mTempPath
) ) {
245 $this->tempFileObj
= new TempFSFile( $this->mTempPath
);
246 $this->mFileSize
= $fileSize ?
: filesize( $this->mTempPath
);
248 $this->tempFileObj
= null;
249 $this->mFileSize
= null;
254 * Fetch the file. Usually a no-op
257 public function fetchFile() {
258 return Status
::newGood();
262 * Return true if the file is empty
265 public function isEmptyFile() {
266 return empty( $this->mFileSize
);
270 * Return the file size
273 public function getFileSize() {
274 return $this->mFileSize
;
278 * Get the base 36 SHA1 of the file
281 public function getTempFileSha1Base36() {
282 return FSFile
::getSha1Base36FromPath( $this->mTempPath
);
286 * @param string $srcPath The source path
287 * @return string|bool The real path if it was a virtual URL Returns false on failure
289 function getRealPath( $srcPath ) {
290 $repo = RepoGroup
::singleton()->getLocalRepo();
291 if ( $repo->isVirtualUrl( $srcPath ) ) {
292 /** @todo Just make uploads work with storage paths UploadFromStash
293 * loads files via virtual URLs.
295 $tmpFile = $repo->getLocalCopy( $srcPath );
297 $tmpFile->bind( $this ); // keep alive with $this
299 $path = $tmpFile ?
$tmpFile->getPath() : false;
308 * Verify whether the upload is sane.
309 * @return mixed Const self::OK or else an array with error information
311 public function verifyUpload() {
314 * If there was no filename or a zero size given, give up quick.
316 if ( $this->isEmptyFile() ) {
317 return [ 'status' => self
::EMPTY_FILE
];
321 * Honor $wgMaxUploadSize
323 $maxSize = self
::getMaxUploadSize( $this->getSourceType() );
324 if ( $this->mFileSize
> $maxSize ) {
326 'status' => self
::FILE_TOO_LARGE
,
332 * Look at the contents of the file; if we can recognize the
333 * type but it's corrupt or data of the wrong type, we should
334 * probably not accept it.
336 $verification = $this->verifyFile();
337 if ( $verification !== true ) {
339 'status' => self
::VERIFICATION_ERROR
,
340 'details' => $verification
345 * Make sure this file can be created
347 $result = $this->validateName();
348 if ( $result !== true ) {
353 if ( !Hooks
::run( 'UploadVerification',
354 [ $this->mDestName
, $this->mTempPath
, &$error ] )
356 return [ 'status' => self
::HOOK_ABORTED
, 'error' => $error ];
359 return [ 'status' => self
::OK
];
363 * Verify that the name is valid and, if necessary, that we can overwrite
365 * @return mixed True if valid, otherwise and array with 'status'
368 public function validateName() {
369 $nt = $this->getTitle();
370 if ( is_null( $nt ) ) {
371 $result = [ 'status' => $this->mTitleError
];
372 if ( $this->mTitleError
== self
::ILLEGAL_FILENAME
) {
373 $result['filtered'] = $this->mFilteredName
;
375 if ( $this->mTitleError
== self
::FILETYPE_BADTYPE
) {
376 $result['finalExt'] = $this->mFinalExtension
;
377 if ( count( $this->mBlackListedExtensions
) ) {
378 $result['blacklistedExt'] = $this->mBlackListedExtensions
;
384 $this->mDestName
= $this->getLocalFile()->getName();
390 * Verify the MIME type.
392 * @note Only checks that it is not an evil MIME. The "does it have
393 * correct extension given its MIME type?" check is in verifyFile.
394 * in `verifyFile()` that MIME type and file extension correlate.
395 * @param string $mime Representing the MIME
396 * @return mixed True if the file is verified, an array otherwise
398 protected function verifyMimeType( $mime ) {
399 global $wgVerifyMimeType;
400 if ( $wgVerifyMimeType ) {
401 wfDebug( "mime: <$mime> extension: <{$this->mFinalExtension}>\n" );
402 global $wgMimeTypeBlacklist;
403 if ( $this->checkFileExtension( $mime, $wgMimeTypeBlacklist ) ) {
404 return [ 'filetype-badmime', $mime ];
407 # Check what Internet Explorer would detect
408 $fp = fopen( $this->mTempPath
, 'rb' );
409 $chunk = fread( $fp, 256 );
412 $magic = MimeMagic
::singleton();
413 $extMime = $magic->guessTypesForExtension( $this->mFinalExtension
);
414 $ieTypes = $magic->getIEMimeTypes( $this->mTempPath
, $chunk, $extMime );
415 foreach ( $ieTypes as $ieType ) {
416 if ( $this->checkFileExtension( $ieType, $wgMimeTypeBlacklist ) ) {
417 return [ 'filetype-bad-ie-mime', $ieType ];
426 * Verifies that it's ok to include the uploaded file
428 * @return mixed True of the file is verified, array otherwise.
430 protected function verifyFile() {
431 global $wgVerifyMimeType, $wgDisableUploadScriptChecks;
433 $status = $this->verifyPartialFile();
434 if ( $status !== true ) {
438 $this->mFileProps
= FSFile
::getPropsFromPath( $this->mTempPath
, $this->mFinalExtension
);
439 $mime = $this->mFileProps
['mime'];
441 if ( $wgVerifyMimeType ) {
442 # XXX: Missing extension will be caught by validateName() via getTitle()
443 if ( $this->mFinalExtension
!= '' && !$this->verifyExtension( $mime, $this->mFinalExtension
) ) {
444 return [ 'filetype-mime-mismatch', $this->mFinalExtension
, $mime ];
448 # check for htmlish code and javascript
449 if ( !$wgDisableUploadScriptChecks ) {
450 if ( $this->mFinalExtension
== 'svg' ||
$mime == 'image/svg+xml' ) {
451 $svgStatus = $this->detectScriptInSvg( $this->mTempPath
, false );
452 if ( $svgStatus !== false ) {
458 $handler = MediaHandler
::getHandler( $mime );
460 $handlerStatus = $handler->verifyUpload( $this->mTempPath
);
461 if ( !$handlerStatus->isOK() ) {
462 $errors = $handlerStatus->getErrorsArray();
464 return reset( $errors );
468 Hooks
::run( 'UploadVerifyFile', [ $this, $mime, &$status ] );
469 if ( $status !== true ) {
473 wfDebug( __METHOD__
. ": all clear; passing.\n" );
479 * A verification routine suitable for partial files
481 * Runs the blacklist checks, but not any checks that may
482 * assume the entire file is present.
484 * @return mixed True for valid or array with error message key.
486 protected function verifyPartialFile() {
487 global $wgAllowJavaUploads, $wgDisableUploadScriptChecks;
489 # getTitle() sets some internal parameters like $this->mFinalExtension
492 $this->mFileProps
= FSFile
::getPropsFromPath( $this->mTempPath
, $this->mFinalExtension
);
494 # check MIME type, if desired
495 $mime = $this->mFileProps
['file-mime'];
496 $status = $this->verifyMimeType( $mime );
497 if ( $status !== true ) {
501 # check for htmlish code and javascript
502 if ( !$wgDisableUploadScriptChecks ) {
503 if ( self
::detectScript( $this->mTempPath
, $mime, $this->mFinalExtension
) ) {
504 return [ 'uploadscripted' ];
506 if ( $this->mFinalExtension
== 'svg' ||
$mime == 'image/svg+xml' ) {
507 $svgStatus = $this->detectScriptInSvg( $this->mTempPath
, true );
508 if ( $svgStatus !== false ) {
514 # Check for Java applets, which if uploaded can bypass cross-site
516 if ( !$wgAllowJavaUploads ) {
517 $this->mJavaDetected
= false;
518 $zipStatus = ZipDirectoryReader
::read( $this->mTempPath
,
519 [ $this, 'zipEntryCallback' ] );
520 if ( !$zipStatus->isOK() ) {
521 $errors = $zipStatus->getErrorsArray();
522 $error = reset( $errors );
523 if ( $error[0] !== 'zip-wrong-format' ) {
527 if ( $this->mJavaDetected
) {
528 return [ 'uploadjava' ];
532 # Scan the uploaded file for viruses
533 $virus = $this->detectVirus( $this->mTempPath
);
535 return [ 'uploadvirus', $virus ];
542 * Callback for ZipDirectoryReader to detect Java class files.
544 * @param array $entry
546 function zipEntryCallback( $entry ) {
547 $names = [ $entry['name'] ];
549 // If there is a null character, cut off the name at it, because JDK's
550 // ZIP_GetEntry() uses strcmp() if the name hashes match. If a file name
551 // were constructed which had ".class\0" followed by a string chosen to
552 // make the hash collide with the truncated name, that file could be
553 // returned in response to a request for the .class file.
554 $nullPos = strpos( $entry['name'], "\000" );
555 if ( $nullPos !== false ) {
556 $names[] = substr( $entry['name'], 0, $nullPos );
559 // If there is a trailing slash in the file name, we have to strip it,
560 // because that's what ZIP_GetEntry() does.
561 if ( preg_grep( '!\.class/?$!', $names ) ) {
562 $this->mJavaDetected
= true;
567 * Alias for verifyTitlePermissions. The function was originally
568 * 'verifyPermissions', but that suggests it's checking the user, when it's
569 * really checking the title + user combination.
571 * @param User $user User object to verify the permissions against
572 * @return mixed An array as returned by getUserPermissionsErrors or true
573 * in case the user has proper permissions.
575 public function verifyPermissions( $user ) {
576 return $this->verifyTitlePermissions( $user );
580 * Check whether the user can edit, upload and create the image. This
581 * checks only against the current title; if it returns errors, it may
582 * very well be that another title will not give errors. Therefore
583 * isAllowed() should be called as well for generic is-user-blocked or
584 * can-user-upload checking.
586 * @param User $user User object to verify the permissions against
587 * @return mixed An array as returned by getUserPermissionsErrors or true
588 * in case the user has proper permissions.
590 public function verifyTitlePermissions( $user ) {
592 * If the image is protected, non-sysop users won't be able
593 * to modify it by uploading a new revision.
595 $nt = $this->getTitle();
596 if ( is_null( $nt ) ) {
599 $permErrors = $nt->getUserPermissionsErrors( 'edit', $user );
600 $permErrorsUpload = $nt->getUserPermissionsErrors( 'upload', $user );
601 if ( !$nt->exists() ) {
602 $permErrorsCreate = $nt->getUserPermissionsErrors( 'create', $user );
604 $permErrorsCreate = [];
606 if ( $permErrors ||
$permErrorsUpload ||
$permErrorsCreate ) {
607 $permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsUpload, $permErrors ) );
608 $permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsCreate, $permErrors ) );
613 $overwriteError = $this->checkOverwrite( $user );
614 if ( $overwriteError !== true ) {
615 return [ $overwriteError ];
622 * Check for non fatal problems with the file.
624 * This should not assume that mTempPath is set.
626 * @return array Array of warnings
628 public function checkWarnings() {
633 $localFile = $this->getLocalFile();
634 $localFile->load( File
::READ_LATEST
);
635 $filename = $localFile->getName();
638 * Check whether the resulting filename is different from the desired one,
639 * but ignore things like ucfirst() and spaces/underscore things
641 $comparableName = str_replace( ' ', '_', $this->mDesiredDestName
);
642 $comparableName = Title
::capitalize( $comparableName, NS_FILE
);
644 if ( $this->mDesiredDestName
!= $filename && $comparableName != $filename ) {
645 $warnings['badfilename'] = $filename;
646 // Debugging for bug 62241
647 wfDebugLog( 'upload', "Filename: '$filename', mDesiredDestName: "
648 . "'$this->mDesiredDestName', comparableName: '$comparableName'" );
651 // Check whether the file extension is on the unwanted list
652 global $wgCheckFileExtensions, $wgFileExtensions;
653 if ( $wgCheckFileExtensions ) {
654 $extensions = array_unique( $wgFileExtensions );
655 if ( !$this->checkFileExtension( $this->mFinalExtension
, $extensions ) ) {
656 $warnings['filetype-unwanted-type'] = [ $this->mFinalExtension
,
657 $wgLang->commaList( $extensions ), count( $extensions ) ];
661 global $wgUploadSizeWarning;
662 if ( $wgUploadSizeWarning && ( $this->mFileSize
> $wgUploadSizeWarning ) ) {
663 $warnings['large-file'] = [ $wgUploadSizeWarning, $this->mFileSize
];
666 if ( $this->mFileSize
== 0 ) {
667 $warnings['emptyfile'] = true;
670 $exists = self
::getExistsWarning( $localFile );
671 if ( $exists !== false ) {
672 $warnings['exists'] = $exists;
675 if ( $localFile->wasDeleted() && !$localFile->exists() ) {
676 $warnings['was-deleted'] = $filename;
679 // Check dupes against existing files
680 $hash = $this->getTempFileSha1Base36();
681 $dupes = RepoGroup
::singleton()->findBySha1( $hash );
682 $title = $this->getTitle();
683 // Remove all matches against self
684 foreach ( $dupes as $key => $dupe ) {
685 if ( $title->equals( $dupe->getTitle() ) ) {
686 unset( $dupes[$key] );
690 $warnings['duplicate'] = $dupes;
693 // Check dupes against archives
694 $archivedFile = new ArchivedFile( null, 0, '', $hash );
695 if ( $archivedFile->getID() > 0 ) {
696 if ( $archivedFile->userCan( File
::DELETED_FILE
) ) {
697 $warnings['duplicate-archive'] = $archivedFile->getName();
699 $warnings['duplicate-archive'] = '';
707 * Really perform the upload. Stores the file in the local repo, watches
708 * if necessary and runs the UploadComplete hook.
710 * @param string $comment
711 * @param string $pageText
712 * @param bool $watch Whether the file page should be added to user's watchlist.
713 * (This doesn't check $user's permissions.)
715 * @param string[] $tags Change tags to add to the log entry and page revision.
716 * (This doesn't check $user's permissions.)
717 * @return Status Indicating the whether the upload succeeded.
719 public function performUpload( $comment, $pageText, $watch, $user, $tags = [] ) {
720 $this->getLocalFile()->load( File
::READ_LATEST
);
722 $status = $this->getLocalFile()->upload(
733 if ( $status->isGood() ) {
735 WatchAction
::doWatch(
736 $this->getLocalFile()->getTitle(),
738 User
::IGNORE_USER_RIGHTS
741 Hooks
::run( 'UploadComplete', [ &$this ] );
743 $this->postProcessUpload();
750 * Perform extra steps after a successful upload.
754 public function postProcessUpload() {
755 global $wgUploadThumbnailRenderMap;
759 $sizes = $wgUploadThumbnailRenderMap;
762 $file = $this->getLocalFile();
764 foreach ( $sizes as $size ) {
765 if ( $file->isVectorized() ||
$file->getWidth() > $size ) {
766 $jobs[] = new ThumbnailRenderJob(
768 [ 'transformParams' => [ 'width' => $size ] ]
774 JobQueueGroup
::singleton()->push( $jobs );
779 * Returns the title of the file to be uploaded. Sets mTitleError in case
780 * the name was illegal.
782 * @return Title The title of the file or null in case the name was illegal
784 public function getTitle() {
785 if ( $this->mTitle
!== false ) {
786 return $this->mTitle
;
788 if ( !is_string( $this->mDesiredDestName
) ) {
789 $this->mTitleError
= self
::ILLEGAL_FILENAME
;
790 $this->mTitle
= null;
792 return $this->mTitle
;
794 /* Assume that if a user specified File:Something.jpg, this is an error
795 * and that the namespace prefix needs to be stripped of.
797 $title = Title
::newFromText( $this->mDesiredDestName
);
798 if ( $title && $title->getNamespace() == NS_FILE
) {
799 $this->mFilteredName
= $title->getDBkey();
801 $this->mFilteredName
= $this->mDesiredDestName
;
804 # oi_archive_name is max 255 bytes, which include a timestamp and an
805 # exclamation mark, so restrict file name to 240 bytes.
806 if ( strlen( $this->mFilteredName
) > 240 ) {
807 $this->mTitleError
= self
::FILENAME_TOO_LONG
;
808 $this->mTitle
= null;
810 return $this->mTitle
;
814 * Chop off any directories in the given filename. Then
815 * filter out illegal characters, and try to make a legible name
816 * out of it. We'll strip some silently that Title would die on.
818 $this->mFilteredName
= wfStripIllegalFilenameChars( $this->mFilteredName
);
819 /* Normalize to title form before we do any further processing */
820 $nt = Title
::makeTitleSafe( NS_FILE
, $this->mFilteredName
);
821 if ( is_null( $nt ) ) {
822 $this->mTitleError
= self
::ILLEGAL_FILENAME
;
823 $this->mTitle
= null;
825 return $this->mTitle
;
827 $this->mFilteredName
= $nt->getDBkey();
830 * We'll want to blacklist against *any* 'extension', and use
831 * only the final one for the whitelist.
833 list( $partname, $ext ) = $this->splitExtensions( $this->mFilteredName
);
835 if ( count( $ext ) ) {
836 $this->mFinalExtension
= trim( $ext[count( $ext ) - 1] );
838 $this->mFinalExtension
= '';
840 # No extension, try guessing one
841 $magic = MimeMagic
::singleton();
842 $mime = $magic->guessMimeType( $this->mTempPath
);
843 if ( $mime !== 'unknown/unknown' ) {
844 # Get a space separated list of extensions
845 $extList = $magic->getExtensionsForType( $mime );
847 # Set the extension to the canonical extension
848 $this->mFinalExtension
= strtok( $extList, ' ' );
850 # Fix up the other variables
851 $this->mFilteredName
.= ".{$this->mFinalExtension}";
852 $nt = Title
::makeTitleSafe( NS_FILE
, $this->mFilteredName
);
853 $ext = [ $this->mFinalExtension
];
858 /* Don't allow users to override the blacklist (check file extension) */
859 global $wgCheckFileExtensions, $wgStrictFileExtensions;
860 global $wgFileExtensions, $wgFileBlacklist;
862 $blackListedExtensions = $this->checkFileExtensionList( $ext, $wgFileBlacklist );
864 if ( $this->mFinalExtension
== '' ) {
865 $this->mTitleError
= self
::FILETYPE_MISSING
;
866 $this->mTitle
= null;
868 return $this->mTitle
;
869 } elseif ( $blackListedExtensions ||
870 ( $wgCheckFileExtensions && $wgStrictFileExtensions &&
871 !$this->checkFileExtension( $this->mFinalExtension
, $wgFileExtensions ) )
873 $this->mBlackListedExtensions
= $blackListedExtensions;
874 $this->mTitleError
= self
::FILETYPE_BADTYPE
;
875 $this->mTitle
= null;
877 return $this->mTitle
;
880 // Windows may be broken with special characters, see bug 1780
881 if ( !preg_match( '/^[\x0-\x7f]*$/', $nt->getText() )
882 && !RepoGroup
::singleton()->getLocalRepo()->backendSupportsUnicodePaths()
884 $this->mTitleError
= self
::WINDOWS_NONASCII_FILENAME
;
885 $this->mTitle
= null;
887 return $this->mTitle
;
890 # If there was more than one "extension", reassemble the base
891 # filename to prevent bogus complaints about length
892 if ( count( $ext ) > 1 ) {
893 $iterations = count( $ext ) - 1;
894 for ( $i = 0; $i < $iterations; $i++
) {
895 $partname .= '.' . $ext[$i];
899 if ( strlen( $partname ) < 1 ) {
900 $this->mTitleError
= self
::MIN_LENGTH_PARTNAME
;
901 $this->mTitle
= null;
903 return $this->mTitle
;
908 return $this->mTitle
;
912 * Return the local file and initializes if necessary.
914 * @return LocalFile|UploadStashFile|null
916 public function getLocalFile() {
917 if ( is_null( $this->mLocalFile
) ) {
918 $nt = $this->getTitle();
919 $this->mLocalFile
= is_null( $nt ) ?
null : wfLocalFile( $nt );
922 return $this->mLocalFile
;
926 * If the user does not supply all necessary information in the first upload
927 * form submission (either by accident or by design) then we may want to
928 * stash the file temporarily, get more information, and publish the file
931 * This method will stash a file in a temporary directory for later
932 * processing, and save the necessary descriptive info into the database.
933 * This method returns the file object, which also has a 'fileKey' property
934 * which can be passed through a form or API request to find this stashed
938 * @return UploadStashFile Stashed file
940 public function stashFile( User
$user = null ) {
941 // was stashSessionFile
943 $stash = RepoGroup
::singleton()->getLocalRepo()->getUploadStash( $user );
944 $file = $stash->stashFile( $this->mTempPath
, $this->getSourceType() );
945 $this->mLocalFile
= $file;
951 * Stash a file in a temporary directory, returning a key which can be used
952 * to find the file again. See stashFile().
954 * @return string File key
956 public function stashFileGetKey() {
957 return $this->stashFile()->getFileKey();
961 * alias for stashFileGetKey, for backwards compatibility
963 * @return string File key
965 public function stashSession() {
966 return $this->stashFileGetKey();
970 * If we've modified the upload file we need to manually remove it
971 * on exit to clean up.
973 public function cleanupTempFile() {
974 if ( $this->mRemoveTempFile
&& $this->tempFileObj
) {
975 // Delete when all relevant TempFSFile handles go out of scope
976 wfDebug( __METHOD__
. ": Marked temporary file '{$this->mTempPath}' for removal\n" );
977 $this->tempFileObj
->autocollect();
981 public function getTempPath() {
982 return $this->mTempPath
;
986 * Split a file into a base name and all dot-delimited 'extensions'
987 * on the end. Some web server configurations will fall back to
988 * earlier pseudo-'extensions' to determine type and execute
989 * scripts, so the blacklist needs to check them all.
991 * @param string $filename
994 public static function splitExtensions( $filename ) {
995 $bits = explode( '.', $filename );
996 $basename = array_shift( $bits );
998 return [ $basename, $bits ];
1002 * Perform case-insensitive match against a list of file extensions.
1003 * Returns true if the extension is in the list.
1005 * @param string $ext
1006 * @param array $list
1009 public static function checkFileExtension( $ext, $list ) {
1010 return in_array( strtolower( $ext ), $list );
1014 * Perform case-insensitive match against a list of file extensions.
1015 * Returns an array of matching extensions.
1018 * @param array $list
1021 public static function checkFileExtensionList( $ext, $list ) {
1022 return array_intersect( array_map( 'strtolower', $ext ), $list );
1026 * Checks if the MIME type of the uploaded file matches the file extension.
1028 * @param string $mime The MIME type of the uploaded file
1029 * @param string $extension The filename extension that the file is to be served with
1032 public static function verifyExtension( $mime, $extension ) {
1033 $magic = MimeMagic
::singleton();
1035 if ( !$mime ||
$mime == 'unknown' ||
$mime == 'unknown/unknown' ) {
1036 if ( !$magic->isRecognizableExtension( $extension ) ) {
1037 wfDebug( __METHOD__
. ": passing file with unknown detected mime type; " .
1038 "unrecognized extension '$extension', can't verify\n" );
1042 wfDebug( __METHOD__
. ": rejecting file with unknown detected mime type; " .
1043 "recognized extension '$extension', so probably invalid file\n" );
1049 $match = $magic->isMatchingExtension( $extension, $mime );
1051 if ( $match === null ) {
1052 if ( $magic->getTypesForExtension( $extension ) !== null ) {
1053 wfDebug( __METHOD__
. ": No extension known for $mime, but we know a mime for $extension\n" );
1057 wfDebug( __METHOD__
. ": no file extension known for mime type $mime, passing file\n" );
1061 } elseif ( $match === true ) {
1062 wfDebug( __METHOD__
. ": mime type $mime matches extension $extension, passing file\n" );
1064 /** @todo If it's a bitmap, make sure PHP or ImageMagick resp. can handle it! */
1068 . ": mime type $mime mismatches file extension $extension, rejecting file\n" );
1075 * Heuristic for detecting files that *could* contain JavaScript instructions or
1076 * things that may look like HTML to a browser and are thus
1077 * potentially harmful. The present implementation will produce false
1078 * positives in some situations.
1080 * @param string $file Pathname to the temporary upload file
1081 * @param string $mime The MIME type of the file
1082 * @param string $extension The extension of the file
1083 * @return bool True if the file contains something looking like embedded scripts
1085 public static function detectScript( $file, $mime, $extension ) {
1086 global $wgAllowTitlesInSVG;
1088 # ugly hack: for text files, always look at the entire file.
1089 # For binary field, just check the first K.
1091 if ( strpos( $mime, 'text/' ) === 0 ) {
1092 $chunk = file_get_contents( $file );
1094 $fp = fopen( $file, 'rb' );
1095 $chunk = fread( $fp, 1024 );
1099 $chunk = strtolower( $chunk );
1105 # decode from UTF-16 if needed (could be used for obfuscation).
1106 if ( substr( $chunk, 0, 2 ) == "\xfe\xff" ) {
1108 } elseif ( substr( $chunk, 0, 2 ) == "\xff\xfe" ) {
1115 $chunk = iconv( $enc, "ASCII//IGNORE", $chunk );
1118 $chunk = trim( $chunk );
1120 /** @todo FIXME: Convert from UTF-16 if necessary! */
1121 wfDebug( __METHOD__
. ": checking for embedded scripts and HTML stuff\n" );
1123 # check for HTML doctype
1124 if ( preg_match( "/<!DOCTYPE *X?HTML/i", $chunk ) ) {
1128 // Some browsers will interpret obscure xml encodings as UTF-8, while
1129 // PHP/expat will interpret the given encoding in the xml declaration (bug 47304)
1130 if ( $extension == 'svg' ||
strpos( $mime, 'image/svg' ) === 0 ) {
1131 if ( self
::checkXMLEncodingMissmatch( $file ) ) {
1137 * Internet Explorer for Windows performs some really stupid file type
1138 * autodetection which can cause it to interpret valid image files as HTML
1139 * and potentially execute JavaScript, creating a cross-site scripting
1142 * Apple's Safari browser also performs some unsafe file type autodetection
1143 * which can cause legitimate files to be interpreted as HTML if the
1144 * web server is not correctly configured to send the right content-type
1145 * (or if you're really uploading plain text and octet streams!)
1147 * Returns true if IE is likely to mistake the given file for HTML.
1148 * Also returns true if Safari would mistake the given file for HTML
1149 * when served with a generic content-type.
1155 '<html', # also in safari
1158 '<script', # also in safari
1162 if ( !$wgAllowTitlesInSVG && $extension !== 'svg' && $mime !== 'image/svg' ) {
1166 foreach ( $tags as $tag ) {
1167 if ( false !== strpos( $chunk, $tag ) ) {
1168 wfDebug( __METHOD__
. ": found something that may make it be mistaken for html: $tag\n" );
1175 * look for JavaScript
1178 # resolve entity-refs to look at attributes. may be harsh on big files... cache result?
1179 $chunk = Sanitizer
::decodeCharReferences( $chunk );
1181 # look for script-types
1182 if ( preg_match( '!type\s*=\s*[\'"]?\s*(?:\w*/)?(?:ecma|java)!sim', $chunk ) ) {
1183 wfDebug( __METHOD__
. ": found script types\n" );
1188 # look for html-style script-urls
1189 if ( preg_match( '!(?:href|src|data)\s*=\s*[\'"]?\s*(?:ecma|java)script:!sim', $chunk ) ) {
1190 wfDebug( __METHOD__
. ": found html-style script urls\n" );
1195 # look for css-style script-urls
1196 if ( preg_match( '!url\s*\(\s*[\'"]?\s*(?:ecma|java)script:!sim', $chunk ) ) {
1197 wfDebug( __METHOD__
. ": found css-style script urls\n" );
1202 wfDebug( __METHOD__
. ": no scripts found\n" );
1208 * Check a whitelist of xml encodings that are known not to be interpreted differently
1209 * by the server's xml parser (expat) and some common browsers.
1211 * @param string $file Pathname to the temporary upload file
1212 * @return bool True if the file contains an encoding that could be misinterpreted
1214 public static function checkXMLEncodingMissmatch( $file ) {
1215 global $wgSVGMetadataCutoff;
1216 $contents = file_get_contents( $file, false, null, -1, $wgSVGMetadataCutoff );
1217 $encodingRegex = '!encoding[ \t\n\r]*=[ \t\n\r]*[\'"](.*?)[\'"]!si';
1219 if ( preg_match( "!<\?xml\b(.*?)\?>!si", $contents, $matches ) ) {
1220 if ( preg_match( $encodingRegex, $matches[1], $encMatch )
1221 && !in_array( strtoupper( $encMatch[1] ), self
::$safeXmlEncodings )
1223 wfDebug( __METHOD__
. ": Found unsafe XML encoding '{$encMatch[1]}'\n" );
1227 } elseif ( preg_match( "!<\?xml\b!si", $contents ) ) {
1228 // Start of XML declaration without an end in the first $wgSVGMetadataCutoff
1229 // bytes. There shouldn't be a legitimate reason for this to happen.
1230 wfDebug( __METHOD__
. ": Unmatched XML declaration start\n" );
1233 } elseif ( substr( $contents, 0, 4 ) == "\x4C\x6F\xA7\x94" ) {
1234 // EBCDIC encoded XML
1235 wfDebug( __METHOD__
. ": EBCDIC Encoded XML\n" );
1240 // It's possible the file is encoded with multi-byte encoding, so re-encode attempt to
1241 // detect the encoding in case is specifies an encoding not whitelisted in self::$safeXmlEncodings
1242 $attemptEncodings = [ 'UTF-16', 'UTF-16BE', 'UTF-32', 'UTF-32BE' ];
1243 foreach ( $attemptEncodings as $encoding ) {
1244 MediaWiki\
suppressWarnings();
1245 $str = iconv( $encoding, 'UTF-8', $contents );
1246 MediaWiki\restoreWarnings
();
1247 if ( $str != '' && preg_match( "!<\?xml\b(.*?)\?>!si", $str, $matches ) ) {
1248 if ( preg_match( $encodingRegex, $matches[1], $encMatch )
1249 && !in_array( strtoupper( $encMatch[1] ), self
::$safeXmlEncodings )
1251 wfDebug( __METHOD__
. ": Found unsafe XML encoding '{$encMatch[1]}'\n" );
1255 } elseif ( $str != '' && preg_match( "!<\?xml\b!si", $str ) ) {
1256 // Start of XML declaration without an end in the first $wgSVGMetadataCutoff
1257 // bytes. There shouldn't be a legitimate reason for this to happen.
1258 wfDebug( __METHOD__
. ": Unmatched XML declaration start\n" );
1268 * @param string $filename
1269 * @param bool $partial
1270 * @return mixed False of the file is verified (does not contain scripts), array otherwise.
1272 protected function detectScriptInSvg( $filename, $partial ) {
1273 $this->mSVGNSError
= false;
1274 $check = new XmlTypeCheck(
1276 [ $this, 'checkSvgScriptCallback' ],
1278 [ 'processing_instruction_handler' => 'UploadBase::checkSvgPICallback' ]
1280 if ( $check->wellFormed
!== true ) {
1281 // Invalid xml (bug 58553)
1282 // But only when non-partial (bug 65724)
1283 return $partial ?
false : [ 'uploadinvalidxml' ];
1284 } elseif ( $check->filterMatch
) {
1285 if ( $this->mSVGNSError
) {
1286 return [ 'uploadscriptednamespace', $this->mSVGNSError
];
1289 return $check->filterMatchType
;
1296 * Callback to filter SVG Processing Instructions.
1297 * @param string $target Processing instruction name
1298 * @param string $data Processing instruction attribute and value
1299 * @return bool (true if the filter identified something bad)
1301 public static function checkSvgPICallback( $target, $data ) {
1302 // Don't allow external stylesheets (bug 57550)
1303 if ( preg_match( '/xml-stylesheet/i', $target ) ) {
1304 return [ 'upload-scripted-pi-callback' ];
1311 * @todo Replace this with a whitelist filter!
1312 * @param string $element
1313 * @param array $attribs
1316 public function checkSvgScriptCallback( $element, $attribs, $data = null ) {
1318 list( $namespace, $strippedElement ) = $this->splitXmlNamespace( $element );
1320 // We specifically don't include:
1321 // http://www.w3.org/1999/xhtml (bug 60771)
1322 static $validNamespaces = [
1325 'http://creativecommons.org/ns#',
1326 'http://inkscape.sourceforge.net/dtd/sodipodi-0.dtd',
1327 'http://ns.adobe.com/adobeillustrator/10.0/',
1328 'http://ns.adobe.com/adobesvgviewerextensions/3.0/',
1329 'http://ns.adobe.com/extensibility/1.0/',
1330 'http://ns.adobe.com/flows/1.0/',
1331 'http://ns.adobe.com/illustrator/1.0/',
1332 'http://ns.adobe.com/imagereplacement/1.0/',
1333 'http://ns.adobe.com/pdf/1.3/',
1334 'http://ns.adobe.com/photoshop/1.0/',
1335 'http://ns.adobe.com/saveforweb/1.0/',
1336 'http://ns.adobe.com/variables/1.0/',
1337 'http://ns.adobe.com/xap/1.0/',
1338 'http://ns.adobe.com/xap/1.0/g/',
1339 'http://ns.adobe.com/xap/1.0/g/img/',
1340 'http://ns.adobe.com/xap/1.0/mm/',
1341 'http://ns.adobe.com/xap/1.0/rights/',
1342 'http://ns.adobe.com/xap/1.0/stype/dimensions#',
1343 'http://ns.adobe.com/xap/1.0/stype/font#',
1344 'http://ns.adobe.com/xap/1.0/stype/manifestitem#',
1345 'http://ns.adobe.com/xap/1.0/stype/resourceevent#',
1346 'http://ns.adobe.com/xap/1.0/stype/resourceref#',
1347 'http://ns.adobe.com/xap/1.0/t/pg/',
1348 'http://purl.org/dc/elements/1.1/',
1349 'http://purl.org/dc/elements/1.1',
1350 'http://schemas.microsoft.com/visio/2003/svgextensions/',
1351 'http://sodipodi.sourceforge.net/dtd/sodipodi-0.dtd',
1352 'http://taptrix.com/inkpad/svg_extensions',
1353 'http://web.resource.org/cc/',
1354 'http://www.freesoftware.fsf.org/bkchem/cdml',
1355 'http://www.inkscape.org/namespaces/inkscape',
1356 'http://www.opengis.net/gml',
1357 'http://www.w3.org/1999/02/22-rdf-syntax-ns#',
1358 'http://www.w3.org/2000/svg',
1359 'http://www.w3.org/tr/rec-rdf-syntax/',
1362 if ( !in_array( $namespace, $validNamespaces ) ) {
1363 wfDebug( __METHOD__
. ": Non-svg namespace '$namespace' in uploaded file.\n" );
1364 /** @todo Return a status object to a closure in XmlTypeCheck, for MW1.21+ */
1365 $this->mSVGNSError
= $namespace;
1371 * check for elements that can contain javascript
1373 if ( $strippedElement == 'script' ) {
1374 wfDebug( __METHOD__
. ": Found script element '$element' in uploaded file.\n" );
1376 return [ 'uploaded-script-svg', $strippedElement ];
1379 # e.g., <svg xmlns="http://www.w3.org/2000/svg">
1380 # <handler xmlns:ev="http://www.w3.org/2001/xml-events" ev:event="load">alert(1)</handler> </svg>
1381 if ( $strippedElement == 'handler' ) {
1382 wfDebug( __METHOD__
. ": Found scriptable element '$element' in uploaded file.\n" );
1384 return [ 'uploaded-script-svg', $strippedElement ];
1387 # SVG reported in Feb '12 that used xml:stylesheet to generate javascript block
1388 if ( $strippedElement == 'stylesheet' ) {
1389 wfDebug( __METHOD__
. ": Found scriptable element '$element' in uploaded file.\n" );
1391 return [ 'uploaded-script-svg', $strippedElement ];
1394 # Block iframes, in case they pass the namespace check
1395 if ( $strippedElement == 'iframe' ) {
1396 wfDebug( __METHOD__
. ": iframe in uploaded file.\n" );
1398 return [ 'uploaded-script-svg', $strippedElement ];
1402 if ( $strippedElement == 'style'
1403 && self
::checkCssFragment( Sanitizer
::normalizeCss( $data ) )
1405 wfDebug( __METHOD__
. ": hostile css in style element.\n" );
1406 return [ 'uploaded-hostile-svg' ];
1409 foreach ( $attribs as $attrib => $value ) {
1410 $stripped = $this->stripXmlNamespace( $attrib );
1411 $value = strtolower( $value );
1413 if ( substr( $stripped, 0, 2 ) == 'on' ) {
1415 . ": Found event-handler attribute '$attrib'='$value' in uploaded file.\n" );
1417 return [ 'uploaded-event-handler-on-svg', $attrib, $value ];
1420 # href with non-local target (don't allow http://, javascript:, etc)
1421 if ( $stripped == 'href'
1422 && strpos( $value, 'data:' ) !== 0
1423 && strpos( $value, '#' ) !== 0
1425 if ( !( $strippedElement === 'a'
1426 && preg_match( '!^https?://!im', $value ) )
1428 wfDebug( __METHOD__
. ": Found href attribute <$strippedElement "
1429 . "'$attrib'='$value' in uploaded file.\n" );
1431 return [ 'uploaded-href-attribute-svg', $strippedElement, $attrib, $value ];
1435 # only allow data: targets that should be safe. This prevents vectors like,
1436 # image/svg, text/xml, application/xml, and text/html, which can contain scripts
1437 if ( $stripped == 'href' && strncasecmp( 'data:', $value, 5 ) === 0 ) {
1438 // rfc2397 parameters. This is only slightly slower than (;[\w;]+)*.
1439 // @codingStandardsIgnoreStart Generic.Files.LineLength
1440 $parameters = '(?>;[a-zA-Z0-9\!#$&\'*+.^_`{|}~-]+=(?>[a-zA-Z0-9\!#$&\'*+.^_`{|}~-]+|"(?>[\0-\x0c\x0e-\x21\x23-\x5b\x5d-\x7f]+|\\\\[\0-\x7f])*"))*(?:;base64)?';
1441 // @codingStandardsIgnoreEnd
1443 if ( !preg_match( "!^data:\s*image/(gif|jpeg|jpg|png)$parameters,!i", $value ) ) {
1444 wfDebug( __METHOD__
. ": Found href to unwhitelisted data: uri "
1445 . "\"<$strippedElement '$attrib'='$value'...\" in uploaded file.\n" );
1446 return [ 'uploaded-href-unsafe-target-svg', $strippedElement, $attrib, $value ];
1450 # Change href with animate from (http://html5sec.org/#137).
1451 if ( $stripped === 'attributename'
1452 && $strippedElement === 'animate'
1453 && $this->stripXmlNamespace( $value ) == 'href'
1455 wfDebug( __METHOD__
. ": Found animate that might be changing href using from "
1456 . "\"<$strippedElement '$attrib'='$value'...\" in uploaded file.\n" );
1458 return [ 'uploaded-animate-svg', $strippedElement, $attrib, $value ];
1461 # use set/animate to add event-handler attribute to parent
1462 if ( ( $strippedElement == 'set' ||
$strippedElement == 'animate' )
1463 && $stripped == 'attributename'
1464 && substr( $value, 0, 2 ) == 'on'
1466 wfDebug( __METHOD__
. ": Found svg setting event-handler attribute with "
1467 . "\"<$strippedElement $stripped='$value'...\" in uploaded file.\n" );
1469 return [ 'uploaded-setting-event-handler-svg', $strippedElement, $stripped, $value ];
1472 # use set to add href attribute to parent element
1473 if ( $strippedElement == 'set'
1474 && $stripped == 'attributename'
1475 && strpos( $value, 'href' ) !== false
1477 wfDebug( __METHOD__
. ": Found svg setting href attribute '$value' in uploaded file.\n" );
1479 return [ 'uploaded-setting-href-svg' ];
1482 # use set to add a remote / data / script target to an element
1483 if ( $strippedElement == 'set'
1484 && $stripped == 'to'
1485 && preg_match( '!(http|https|data|script):!sim', $value )
1487 wfDebug( __METHOD__
. ": Found svg setting attribute to '$value' in uploaded file.\n" );
1489 return [ 'uploaded-wrong-setting-svg', $value ];
1492 # use handler attribute with remote / data / script
1493 if ( $stripped == 'handler' && preg_match( '!(http|https|data|script):!sim', $value ) ) {
1494 wfDebug( __METHOD__
. ": Found svg setting handler with remote/data/script "
1495 . "'$attrib'='$value' in uploaded file.\n" );
1497 return [ 'uploaded-setting-handler-svg', $attrib, $value ];
1500 # use CSS styles to bring in remote code
1501 if ( $stripped == 'style'
1502 && self
::checkCssFragment( Sanitizer
::normalizeCss( $value ) )
1504 wfDebug( __METHOD__
. ": Found svg setting a style with "
1505 . "remote url '$attrib'='$value' in uploaded file.\n" );
1506 return [ 'uploaded-remote-url-svg', $attrib, $value ];
1509 # Several attributes can include css, css character escaping isn't allowed
1510 $cssAttrs = [ 'font', 'clip-path', 'fill', 'filter', 'marker',
1511 'marker-end', 'marker-mid', 'marker-start', 'mask', 'stroke' ];
1512 if ( in_array( $stripped, $cssAttrs )
1513 && self
::checkCssFragment( $value )
1515 wfDebug( __METHOD__
. ": Found svg setting a style with "
1516 . "remote url '$attrib'='$value' in uploaded file.\n" );
1517 return [ 'uploaded-remote-url-svg', $attrib, $value ];
1520 # image filters can pull in url, which could be svg that executes scripts
1521 if ( $strippedElement == 'image'
1522 && $stripped == 'filter'
1523 && preg_match( '!url\s*\(!sim', $value )
1525 wfDebug( __METHOD__
. ": Found image filter with url: "
1526 . "\"<$strippedElement $stripped='$value'...\" in uploaded file.\n" );
1528 return [ 'uploaded-image-filter-svg', $strippedElement, $stripped, $value ];
1532 return false; // No scripts detected
1536 * Check a block of CSS or CSS fragment for anything that looks like
1537 * it is bringing in remote code.
1538 * @param string $value a string of CSS
1539 * @param bool $propOnly only check css properties (start regex with :)
1540 * @return bool true if the CSS contains an illegal string, false if otherwise
1542 private static function checkCssFragment( $value ) {
1544 # Forbid external stylesheets, for both reliability and to protect viewer's privacy
1545 if ( stripos( $value, '@import' ) !== false ) {
1549 # We allow @font-face to embed fonts with data: urls, so we snip the string
1550 # 'url' out so this case won't match when we check for urls below
1551 $pattern = '!(@font-face\s*{[^}]*src:)url(\("data:;base64,)!im';
1552 $value = preg_replace( $pattern, '$1$2', $value );
1554 # Check for remote and executable CSS. Unlike in Sanitizer::checkCss, the CSS
1555 # properties filter and accelerator don't seem to be useful for xss in SVG files.
1556 # Expression and -o-link don't seem to work either, but filtering them here in case.
1557 # Additionally, we catch remote urls like url("http:..., url('http:..., url(http:...,
1558 # but not local ones such as url("#..., url('#..., url(#....
1559 if ( preg_match( '!expression
1561 | -o-link-source\s*:
1562 | -o-replace\s*:!imx', $value ) ) {
1566 if ( preg_match_all(
1567 "!(\s*(url|image|image-set)\s*\(\s*[\"']?\s*[^#]+.*?\))!sim",
1572 # TODO: redo this in one regex. Until then, url("#whatever") matches the first
1573 foreach ( $matches[1] as $match ) {
1574 if ( !preg_match( "!\s*(url|image|image-set)\s*\(\s*(#|'#|\"#)!im", $match ) ) {
1580 if ( preg_match( '/[\000-\010\013\016-\037\177]/', $value ) ) {
1588 * Divide the element name passed by the xml parser to the callback into URI and prifix.
1589 * @param string $element
1590 * @return array Containing the namespace URI and prefix
1592 private static function splitXmlNamespace( $element ) {
1593 // 'http://www.w3.org/2000/svg:script' -> array( 'http://www.w3.org/2000/svg', 'script' )
1594 $parts = explode( ':', strtolower( $element ) );
1595 $name = array_pop( $parts );
1596 $ns = implode( ':', $parts );
1598 return [ $ns, $name ];
1602 * @param string $name
1605 private function stripXmlNamespace( $name ) {
1606 // 'http://www.w3.org/2000/svg:script' -> 'script'
1607 $parts = explode( ':', strtolower( $name ) );
1609 return array_pop( $parts );
1613 * Generic wrapper function for a virus scanner program.
1614 * This relies on the $wgAntivirus and $wgAntivirusSetup variables.
1615 * $wgAntivirusRequired may be used to deny upload if the scan fails.
1617 * @param string $file Pathname to the temporary upload file
1618 * @return mixed False if not virus is found, null if the scan fails or is disabled,
1619 * or a string containing feedback from the virus scanner if a virus was found.
1620 * If textual feedback is missing but a virus was found, this function returns true.
1622 public static function detectVirus( $file ) {
1623 global $wgAntivirus, $wgAntivirusSetup, $wgAntivirusRequired, $wgOut;
1625 if ( !$wgAntivirus ) {
1626 wfDebug( __METHOD__
. ": virus scanner disabled\n" );
1631 if ( !$wgAntivirusSetup[$wgAntivirus] ) {
1632 wfDebug( __METHOD__
. ": unknown virus scanner: $wgAntivirus\n" );
1633 $wgOut->wrapWikiMsg( "<div class=\"error\">\n$1\n</div>",
1634 [ 'virus-badscanner', $wgAntivirus ] );
1636 return wfMessage( 'virus-unknownscanner' )->text() . " $wgAntivirus";
1639 # look up scanner configuration
1640 $command = $wgAntivirusSetup[$wgAntivirus]['command'];
1641 $exitCodeMap = $wgAntivirusSetup[$wgAntivirus]['codemap'];
1642 $msgPattern = isset( $wgAntivirusSetup[$wgAntivirus]['messagepattern'] ) ?
1643 $wgAntivirusSetup[$wgAntivirus]['messagepattern'] : null;
1645 if ( strpos( $command, "%f" ) === false ) {
1646 # simple pattern: append file to scan
1647 $command .= " " . wfEscapeShellArg( $file );
1649 # complex pattern: replace "%f" with file to scan
1650 $command = str_replace( "%f", wfEscapeShellArg( $file ), $command );
1653 wfDebug( __METHOD__
. ": running virus scan: $command \n" );
1655 # execute virus scanner
1658 # NOTE: there's a 50 line workaround to make stderr redirection work on windows, too.
1659 # that does not seem to be worth the pain.
1660 # Ask me (Duesentrieb) about it if it's ever needed.
1661 $output = wfShellExecWithStderr( $command, $exitCode );
1663 # map exit code to AV_xxx constants.
1664 $mappedCode = $exitCode;
1665 if ( $exitCodeMap ) {
1666 if ( isset( $exitCodeMap[$exitCode] ) ) {
1667 $mappedCode = $exitCodeMap[$exitCode];
1668 } elseif ( isset( $exitCodeMap["*"] ) ) {
1669 $mappedCode = $exitCodeMap["*"];
1673 /* NB: AV_NO_VIRUS is 0 but AV_SCAN_FAILED is false,
1674 * so we need the strict equalities === and thus can't use a switch here
1676 if ( $mappedCode === AV_SCAN_FAILED
) {
1677 # scan failed (code was mapped to false by $exitCodeMap)
1678 wfDebug( __METHOD__
. ": failed to scan $file (code $exitCode).\n" );
1680 $output = $wgAntivirusRequired
1681 ?
wfMessage( 'virus-scanfailed', [ $exitCode ] )->text()
1683 } elseif ( $mappedCode === AV_SCAN_ABORTED
) {
1684 # scan failed because filetype is unknown (probably imune)
1685 wfDebug( __METHOD__
. ": unsupported file type $file (code $exitCode).\n" );
1687 } elseif ( $mappedCode === AV_NO_VIRUS
) {
1689 wfDebug( __METHOD__
. ": file passed virus scan.\n" );
1692 $output = trim( $output );
1695 $output = true; # if there's no output, return true
1696 } elseif ( $msgPattern ) {
1698 if ( preg_match( $msgPattern, $output, $groups ) ) {
1700 $output = $groups[1];
1705 wfDebug( __METHOD__
. ": FOUND VIRUS! scanner feedback: $output \n" );
1712 * Check if there's an overwrite conflict and, if so, if restrictions
1713 * forbid this user from performing the upload.
1717 * @return mixed True on success, array on failure
1719 private function checkOverwrite( $user ) {
1720 // First check whether the local file can be overwritten
1721 $file = $this->getLocalFile();
1722 $file->load( File
::READ_LATEST
);
1723 if ( $file->exists() ) {
1724 if ( !self
::userCanReUpload( $user, $file ) ) {
1725 return [ 'fileexists-forbidden', $file->getName() ];
1731 /* Check shared conflicts: if the local file does not exist, but
1732 * wfFindFile finds a file, it exists in a shared repository.
1734 $file = wfFindFile( $this->getTitle(), [ 'latest' => true ] );
1735 if ( $file && !$user->isAllowed( 'reupload-shared' ) ) {
1736 return [ 'fileexists-shared-forbidden', $file->getName() ];
1743 * Check if a user is the last uploader
1749 public static function userCanReUpload( User
$user, File
$img ) {
1750 if ( $user->isAllowed( 'reupload' ) ) {
1751 return true; // non-conditional
1752 } elseif ( !$user->isAllowed( 'reupload-own' ) ) {
1756 if ( !( $img instanceof LocalFile
) ) {
1762 return $user->getId() == $img->getUser( 'id' );
1766 * Helper function that does various existence checks for a file.
1767 * The following checks are performed:
1769 * - Article with the same name as the file exists
1770 * - File exists with normalized extension
1771 * - The file looks like a thumbnail and the original exists
1773 * @param File $file The File object to check
1774 * @return mixed False if the file does not exists, else an array
1776 public static function getExistsWarning( $file ) {
1777 if ( $file->exists() ) {
1778 return [ 'warning' => 'exists', 'file' => $file ];
1781 if ( $file->getTitle()->getArticleID() ) {
1782 return [ 'warning' => 'page-exists', 'file' => $file ];
1785 if ( strpos( $file->getName(), '.' ) == false ) {
1786 $partname = $file->getName();
1789 $n = strrpos( $file->getName(), '.' );
1790 $extension = substr( $file->getName(), $n +
1 );
1791 $partname = substr( $file->getName(), 0, $n );
1793 $normalizedExtension = File
::normalizeExtension( $extension );
1795 if ( $normalizedExtension != $extension ) {
1796 // We're not using the normalized form of the extension.
1797 // Normal form is lowercase, using most common of alternate
1798 // extensions (eg 'jpg' rather than 'JPEG').
1800 // Check for another file using the normalized form...
1801 $nt_lc = Title
::makeTitle( NS_FILE
, "{$partname}.{$normalizedExtension}" );
1802 $file_lc = wfLocalFile( $nt_lc );
1804 if ( $file_lc->exists() ) {
1806 'warning' => 'exists-normalized',
1808 'normalizedFile' => $file_lc
1813 // Check for files with the same name but a different extension
1814 $similarFiles = RepoGroup
::singleton()->getLocalRepo()->findFilesByPrefix(
1815 "{$partname}.", 1 );
1816 if ( count( $similarFiles ) ) {
1818 'warning' => 'exists-normalized',
1820 'normalizedFile' => $similarFiles[0],
1824 if ( self
::isThumbName( $file->getName() ) ) {
1825 # Check for filenames like 50px- or 180px-, these are mostly thumbnails
1826 $nt_thb = Title
::newFromText(
1827 substr( $partname, strpos( $partname, '-' ) +
1 ) . '.' . $extension,
1830 $file_thb = wfLocalFile( $nt_thb );
1831 if ( $file_thb->exists() ) {
1833 'warning' => 'thumb',
1835 'thumbFile' => $file_thb
1838 // File does not exist, but we just don't like the name
1840 'warning' => 'thumb-name',
1842 'thumbFile' => $file_thb
1847 foreach ( self
::getFilenamePrefixBlacklist() as $prefix ) {
1848 if ( substr( $partname, 0, strlen( $prefix ) ) == $prefix ) {
1850 'warning' => 'bad-prefix',
1861 * Helper function that checks whether the filename looks like a thumbnail
1862 * @param string $filename
1865 public static function isThumbName( $filename ) {
1866 $n = strrpos( $filename, '.' );
1867 $partname = $n ?
substr( $filename, 0, $n ) : $filename;
1870 substr( $partname, 3, 3 ) == 'px-' ||
1871 substr( $partname, 2, 3 ) == 'px-'
1873 preg_match( "/[0-9]{2}/", substr( $partname, 0, 2 ) );
1877 * Get a list of blacklisted filename prefixes from [[MediaWiki:Filename-prefix-blacklist]]
1879 * @return array List of prefixes
1881 public static function getFilenamePrefixBlacklist() {
1883 $message = wfMessage( 'filename-prefix-blacklist' )->inContentLanguage();
1884 if ( !$message->isDisabled() ) {
1885 $lines = explode( "\n", $message->plain() );
1886 foreach ( $lines as $line ) {
1887 // Remove comment lines
1888 $comment = substr( trim( $line ), 0, 1 );
1889 if ( $comment == '#' ||
$comment == '' ) {
1892 // Remove additional comments after a prefix
1893 $comment = strpos( $line, '#' );
1894 if ( $comment > 0 ) {
1895 $line = substr( $line, 0, $comment - 1 );
1897 $blacklist[] = trim( $line );
1905 * Gets image info about the file just uploaded.
1907 * Also has the effect of setting metadata to be an 'indexed tag name' in
1908 * returned API result if 'metadata' was requested. Oddly, we have to pass
1909 * the "result" object down just so it can do that with the appropriate
1910 * format, presumably.
1912 * @param ApiResult $result
1913 * @return array Image info
1915 public function getImageInfo( $result ) {
1916 $file = $this->getLocalFile();
1917 /** @todo This cries out for refactoring.
1918 * We really want to say $file->getAllInfo(); here.
1919 * Perhaps "info" methods should be moved into files, and the API should
1920 * just wrap them in queries.
1922 if ( $file instanceof UploadStashFile
) {
1923 $imParam = ApiQueryStashImageInfo
::getPropertyNames();
1924 $info = ApiQueryStashImageInfo
::getInfo( $file, array_flip( $imParam ), $result );
1926 $imParam = ApiQueryImageInfo
::getPropertyNames();
1927 $info = ApiQueryImageInfo
::getInfo( $file, array_flip( $imParam ), $result );
1934 * @param array $error
1937 public function convertVerifyErrorToStatus( $error ) {
1938 $code = $error['status'];
1939 unset( $code['status'] );
1941 return Status
::newFatal( $this->getVerificationErrorCode( $code ), $error );
1945 * Get the MediaWiki maximum uploaded file size for given type of upload, based on
1948 * @param null|string $forType
1951 public static function getMaxUploadSize( $forType = null ) {
1952 global $wgMaxUploadSize;
1954 if ( is_array( $wgMaxUploadSize ) ) {
1955 if ( !is_null( $forType ) && isset( $wgMaxUploadSize[$forType] ) ) {
1956 return $wgMaxUploadSize[$forType];
1958 return $wgMaxUploadSize['*'];
1961 return intval( $wgMaxUploadSize );
1966 * Get the PHP maximum uploaded file size, based on ini settings. If there is no limit or the
1967 * limit can't be guessed, returns a very large number (PHP_INT_MAX).
1972 public static function getMaxPhpUploadSize() {
1973 $phpMaxFileSize = wfShorthandToInteger(
1974 ini_get( 'upload_max_filesize' ) ?
: ini_get( 'hhvm.server.upload.upload_max_file_size' ),
1977 $phpMaxPostSize = wfShorthandToInteger(
1978 ini_get( 'post_max_size' ) ?
: ini_get( 'hhvm.server.max_post_size' ),
1981 return min( $phpMaxFileSize, $phpMaxPostSize );
1985 * Get the current status of a chunked upload (used for polling)
1987 * The value will be read from cache.
1990 * @param string $statusKey
1991 * @return Status[]|bool
1993 public static function getSessionStatus( User
$user, $statusKey ) {
1994 $key = wfMemcKey( 'uploadstatus', $user->getId() ?
: md5( $user->getName() ), $statusKey );
1996 return ObjectCache
::getMainStashInstance()->get( $key );
2000 * Set the current status of a chunked upload (used for polling)
2002 * The value will be set in cache for 1 day
2005 * @param string $statusKey
2006 * @param array|bool $value
2009 public static function setSessionStatus( User
$user, $statusKey, $value ) {
2010 $key = wfMemcKey( 'uploadstatus', $user->getId() ?
: md5( $user->getName() ), $statusKey );
2012 $cache = ObjectCache
::getMainStashInstance();
2013 if ( $value === false ) {
2014 $cache->delete( $key );
2016 $cache->set( $key, $value, $cache::TTL_DAY
);