'empty-file',
self::FILE_TOO_LARGE => 'file-too-large',
self::FILETYPE_MISSING => 'filetype-missing',
self::FILETYPE_BADTYPE => 'filetype-banned',
self::MIN_LENGTH_PARTNAME => 'filename-tooshort',
self::ILLEGAL_FILENAME => 'illegal-filename',
self::OVERWRITE_EXISTING_FILE => 'overwrite',
self::VERIFICATION_ERROR => 'verification-error',
self::HOOK_ABORTED => 'hookaborted',
self::WINDOWS_NONASCII_FILENAME => 'windows-nonascii-filename',
self::FILENAME_TOO_LONG => 'filename-toolong',
];
return $code_to_status[$error] ?? 'unknown-error';
}
/**
* Returns true if uploads are enabled.
* Can be override by subclasses.
* @stable to override
* @return bool
*/
public static function isEnabled() {
global $wgEnableUploads;
return $wgEnableUploads && wfIniGetBool( 'file_uploads' );
}
/**
* Returns true if the user can use this upload module or else a string
* identifying the missing permission.
* Can be overridden by subclasses.
*
* @param UserIdentity $user
* @return bool|string
*/
public static function isAllowed( UserIdentity $user ) {
$permissionManager = MediaWikiServices::getInstance()->getPermissionManager();
foreach ( [ 'upload', 'edit' ] as $permission ) {
if ( !$permissionManager->userHasRight( $user, $permission ) ) {
return $permission;
}
}
return true;
}
/**
* Returns true if the user has surpassed the upload rate limit, false otherwise.
*
* @param User $user
* @return bool
*/
public static function isThrottled( $user ) {
return $user->pingLimiter( 'upload' );
}
// Upload handlers. Should probably just be a global.
private static $uploadHandlers = [ 'Stash', 'File', 'Url' ];
/**
* Create a form of UploadBase depending on wpSourceType and initializes it
*
* @param WebRequest &$request
* @param string|null $type
* @return null|self
*/
public static function createFromRequest( &$request, $type = null ) {
$type = $type ?: $request->getVal( 'wpSourceType', 'File' );
if ( !$type ) {
return null;
}
// Get the upload class
$type = ucfirst( $type );
// Give hooks the chance to handle this request
/** @var self|null $className */
$className = null;
Hooks::runner()->onUploadCreateFromRequest( $type, $className );
if ( $className === null ) {
$className = 'UploadFrom' . $type;
wfDebug( __METHOD__ . ": class name: $className" );
if ( !in_array( $type, self::$uploadHandlers ) ) {
return null;
}
}
// Check whether this upload class is enabled
if ( !$className::isEnabled() ) {
return null;
}
// Check whether the request is valid
if ( !$className::isValidRequest( $request ) ) {
return null;
}
/** @var self $handler */
$handler = new $className;
$handler->initializeFromRequest( $request );
return $handler;
}
/**
* Check whether a request if valid for this handler
* @param WebRequest $request
* @return bool
*/
public static function isValidRequest( $request ) {
return false;
}
/**
* @stable to call
*/
public function __construct() {
}
/**
* Returns the upload type. Should be overridden by child classes
*
* @since 1.18
* @stable to override
* @return string
*/
public function getSourceType() {
return null;
}
/**
* @param string $name The desired destination name
* @param string $tempPath
* @param int|null $fileSize
* @param bool $removeTempFile (false) remove the temporary file?
* @throws MWException
*/
public function initializePathInfo( $name, $tempPath, $fileSize, $removeTempFile = false ) {
$this->mDesiredDestName = $name;
if ( FileBackend::isStoragePath( $tempPath ) ) {
throw new MWException( __METHOD__ . " given storage path `$tempPath`." );
}
$this->setTempFile( $tempPath, $fileSize );
$this->mRemoveTempFile = $removeTempFile;
}
/**
* Initialize from a WebRequest. Override this in a subclass.
*
* @param WebRequest &$request
*/
abstract public function initializeFromRequest( &$request );
/**
* @param string $tempPath File system path to temporary file containing the upload
* @param int|null $fileSize
*/
protected function setTempFile( $tempPath, $fileSize = null ) {
$this->mTempPath = $tempPath;
$this->mFileSize = $fileSize ?: null;
if ( strlen( $this->mTempPath ) && file_exists( $this->mTempPath ) ) {
$this->tempFileObj = new TempFSFile( $this->mTempPath );
if ( !$fileSize ) {
$this->mFileSize = filesize( $this->mTempPath );
}
} else {
$this->tempFileObj = null;
}
}
/**
* Fetch the file. Usually a no-op
* @stable to override
* @return Status
*/
public function fetchFile() {
return Status::newGood();
}
/**
* Return true if the file is empty
* @return bool
*/
public function isEmptyFile() {
return empty( $this->mFileSize );
}
/**
* Return the file size
* @return int
*/
public function getFileSize() {
return $this->mFileSize;
}
/**
* Get the base 36 SHA1 of the file
* @stable to override
* @return string
*/
public function getTempFileSha1Base36() {
return FSFile::getSha1Base36FromPath( $this->mTempPath );
}
/**
* @param string $srcPath The source path
* @return string|bool The real path if it was a virtual URL Returns false on failure
*/
public function getRealPath( $srcPath ) {
$repo = MediaWikiServices::getInstance()->getRepoGroup()->getLocalRepo();
if ( FileRepo::isVirtualUrl( $srcPath ) ) {
/** @todo Just make uploads work with storage paths UploadFromStash
* loads files via virtual URLs.
*/
$tmpFile = $repo->getLocalCopy( $srcPath );
if ( $tmpFile ) {
$tmpFile->bind( $this ); // keep alive with $this
}
$path = $tmpFile ? $tmpFile->getPath() : false;
} else {
$path = $srcPath;
}
return $path;
}
/**
* Verify whether the upload is sane.
*
* Return a status array representing the outcome of the verification.
* Possible keys are:
* - 'status': set to self::OK in case of success, or to one of the error constants defined in
* this class in case of failure
* - 'max': set to the maximum allowed file size ($wgMaxUploadSize) if the upload is too large
* - 'details': set to error details if the file type is valid but contents are corrupt
* - 'filtered': set to the sanitized file name if the requested file name is invalid
* - 'finalExt': set to the file's file extension if it is not an allowed file extension
* - 'blacklistedExt': set to the list of blacklisted file extensions if the current file extension
* is not allowed for uploads and the blacklist is not empty
*
* @stable to override
* @return mixed[] array representing the result of the verification
*/
public function verifyUpload() {
/**
* If there was no filename or a zero size given, give up quick.
*/
if ( $this->isEmptyFile() ) {
return [ 'status' => self::EMPTY_FILE ];
}
/**
* Honor $wgMaxUploadSize
*/
$maxSize = self::getMaxUploadSize( $this->getSourceType() );
if ( $this->mFileSize > $maxSize ) {
return [
'status' => self::FILE_TOO_LARGE,
'max' => $maxSize,
];
}
/**
* Look at the contents of the file; if we can recognize the
* type but it's corrupt or data of the wrong type, we should
* probably not accept it.
*/
$verification = $this->verifyFile();
if ( $verification !== true ) {
return [
'status' => self::VERIFICATION_ERROR,
'details' => $verification
];
}
/**
* Make sure this file can be created
*/
$result = $this->validateName();
if ( $result !== true ) {
return $result;
}
return [ 'status' => self::OK ];
}
/**
* Verify that the name is valid and, if necessary, that we can overwrite
*
* @return array|bool True if valid, otherwise an array with 'status'
* and other keys
*/
public function validateName() {
$nt = $this->getTitle();
if ( $nt === null ) {
$result = [ 'status' => $this->mTitleError ];
if ( $this->mTitleError == self::ILLEGAL_FILENAME ) {
$result['filtered'] = $this->mFilteredName;
}
if ( $this->mTitleError == self::FILETYPE_BADTYPE ) {
$result['finalExt'] = $this->mFinalExtension;
if ( count( $this->mBlackListedExtensions ) ) {
$result['blacklistedExt'] = $this->mBlackListedExtensions;
}
}
return $result;
}
$this->mDestName = $this->getLocalFile()->getName();
return true;
}
/**
* Verify the MIME type.
*
* @note Only checks that it is not an evil MIME. The "does it have
* correct extension given its MIME type?" check is in verifyFile.
* in `verifyFile()` that MIME type and file extension correlate.
* @param string $mime Representing the MIME
* @return array|bool True if the file is verified, an array otherwise
*/
protected function verifyMimeType( $mime ) {
global $wgVerifyMimeType, $wgVerifyMimeTypeIE;
if ( $wgVerifyMimeType ) {
wfDebug( "mime: <$mime> extension: <{$this->mFinalExtension}>" );
global $wgMimeTypeBlacklist;
if ( $this->checkFileExtension( $mime, $wgMimeTypeBlacklist ) ) {
return [ 'filetype-badmime', $mime ];
}
if ( $wgVerifyMimeTypeIE ) {
# Check what Internet Explorer would detect
$fp = fopen( $this->mTempPath, 'rb' );
$chunk = fread( $fp, 256 );
fclose( $fp );
$magic = MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer();
$extMime = $magic->getMimeTypeFromExtensionOrNull( $this->mFinalExtension );
$ieTypes = $magic->getIEMimeTypes( $this->mTempPath, $chunk, $extMime );
foreach ( $ieTypes as $ieType ) {
if ( $this->checkFileExtension( $ieType, $wgMimeTypeBlacklist ) ) {
return [ 'filetype-bad-ie-mime', $ieType ];
}
}
}
}
return true;
}
/**
* Verifies that it's ok to include the uploaded file
*
* @return array|bool True of the file is verified, array otherwise.
*/
protected function verifyFile() {
global $wgVerifyMimeType, $wgDisableUploadScriptChecks;
$status = $this->verifyPartialFile();
if ( $status !== true ) {
return $status;
}
$mwProps = new MWFileProps( MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer() );
$this->mFileProps = $mwProps->getPropsFromPath( $this->mTempPath, $this->mFinalExtension );
$mime = $this->mFileProps['mime'];
if ( $wgVerifyMimeType ) {
# XXX: Missing extension will be caught by validateName() via getTitle()
if ( $this->mFinalExtension != '' && !$this->verifyExtension( $mime, $this->mFinalExtension ) ) {
return [ 'filetype-mime-mismatch', $this->mFinalExtension, $mime ];
}
}
# check for htmlish code and javascript
if ( !$wgDisableUploadScriptChecks ) {
if ( $this->mFinalExtension == 'svg' || $mime == 'image/svg+xml' ) {
$svgStatus = $this->detectScriptInSvg( $this->mTempPath, false );
if ( $svgStatus !== false ) {
return $svgStatus;
}
}
}
$handler = MediaHandler::getHandler( $mime );
if ( $handler ) {
$handlerStatus = $handler->verifyUpload( $this->mTempPath );
if ( !$handlerStatus->isOK() ) {
$errors = $handlerStatus->getErrorsArray();
return reset( $errors );
}
}
$error = true;
$this->getHookRunner()->onUploadVerifyFile( $this, $mime, $error );
if ( $error !== true ) {
if ( !is_array( $error ) ) {
$error = [ $error ];
}
return $error;
}
wfDebug( __METHOD__ . ": all clear; passing." );
return true;
}
/**
* A verification routine suitable for partial files
*
* Runs the blacklist checks, but not any checks that may
* assume the entire file is present.
*
* @return array|bool True if the file is valid, else an array with error message key.
*/
protected function verifyPartialFile() {
global $wgAllowJavaUploads, $wgDisableUploadScriptChecks;
# getTitle() sets some internal parameters like $this->mFinalExtension
$this->getTitle();
$mwProps = new MWFileProps( MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer() );
$this->mFileProps = $mwProps->getPropsFromPath( $this->mTempPath, $this->mFinalExtension );
# check MIME type, if desired
$mime = $this->mFileProps['file-mime'];
$status = $this->verifyMimeType( $mime );
if ( $status !== true ) {
return $status;
}
# check for htmlish code and javascript
if ( !$wgDisableUploadScriptChecks ) {
if ( self::detectScript( $this->mTempPath, $mime, $this->mFinalExtension ) ) {
return [ 'uploadscripted' ];
}
if ( $this->mFinalExtension == 'svg' || $mime == 'image/svg+xml' ) {
$svgStatus = $this->detectScriptInSvg( $this->mTempPath, true );
if ( $svgStatus !== false ) {
return $svgStatus;
}
}
}
# Check for Java applets, which if uploaded can bypass cross-site
# restrictions.
if ( !$wgAllowJavaUploads ) {
$this->mJavaDetected = false;
$zipStatus = ZipDirectoryReader::read( $this->mTempPath,
[ $this, 'zipEntryCallback' ] );
if ( !$zipStatus->isOK() ) {
$errors = $zipStatus->getErrorsArray();
$error = reset( $errors );
if ( $error[0] !== 'zip-wrong-format' ) {
return $error;
}
}
if ( $this->mJavaDetected ) {
return [ 'uploadjava' ];
}
}
# Scan the uploaded file for viruses
$virus = $this->detectVirus( $this->mTempPath );
if ( $virus ) {
return [ 'uploadvirus', $virus ];
}
return true;
}
/**
* Callback for ZipDirectoryReader to detect Java class files.
*
* @param array $entry
*/
public function zipEntryCallback( $entry ) {
$names = [ $entry['name'] ];
// If there is a null character, cut off the name at it, because JDK's
// ZIP_GetEntry() uses strcmp() if the name hashes match. If a file name
// were constructed which had ".class\0" followed by a string chosen to
// make the hash collide with the truncated name, that file could be
// returned in response to a request for the .class file.
$nullPos = strpos( $entry['name'], "\000" );
if ( $nullPos !== false ) {
$names[] = substr( $entry['name'], 0, $nullPos );
}
// If there is a trailing slash in the file name, we have to strip it,
// because that's what ZIP_GetEntry() does.
if ( preg_grep( '!\.class/?$!', $names ) ) {
$this->mJavaDetected = true;
}
}
/**
* Alias for verifyTitlePermissions. The function was originally
* 'verifyPermissions', but that suggests it's checking the user, when it's
* really checking the title + user combination.
*
* @param User $user User object to verify the permissions against
* @return array|bool An array as returned by getPermissionErrors or true
* in case the user has proper permissions.
*/
public function verifyPermissions( $user ) {
return $this->verifyTitlePermissions( $user );
}
/**
* Check whether the user can edit, upload and create the image. This
* checks only against the current title; if it returns errors, it may
* very well be that another title will not give errors. Therefore
* isAllowed() should be called as well for generic is-user-blocked or
* can-user-upload checking.
*
* @param User $user User object to verify the permissions against
* @return array|bool An array as returned by getPermissionErrors or true
* in case the user has proper permissions.
*/
public function verifyTitlePermissions( $user ) {
/**
* If the image is protected, non-sysop users won't be able
* to modify it by uploading a new revision.
*/
$nt = $this->getTitle();
if ( $nt === null ) {
return true;
}
$permManager = MediaWikiServices::getInstance()->getPermissionManager();
$permErrors = $permManager->getPermissionErrors( 'edit', $user, $nt );
$permErrorsUpload = $permManager->getPermissionErrors( 'upload', $user, $nt );
if ( !$nt->exists() ) {
$permErrorsCreate = $permManager->getPermissionErrors( 'create', $user, $nt );
} else {
$permErrorsCreate = [];
}
if ( $permErrors || $permErrorsUpload || $permErrorsCreate ) {
$permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsUpload, $permErrors ) );
$permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsCreate, $permErrors ) );
return $permErrors;
}
$overwriteError = $this->checkOverwrite( $user );
if ( $overwriteError !== true ) {
return [ $overwriteError ];
}
return true;
}
/**
* Check for non fatal problems with the file.
*
* This should not assume that mTempPath is set.
*
* @param User|null $user Accepted since 1.35
*
* @return mixed[] Array of warnings
*/
public function checkWarnings( $user = null ) {
if ( $user === null ) {
// TODO check uses and hard deprecate
$user = RequestContext::getMain()->getUser();
}
$warnings = [];
$localFile = $this->getLocalFile();
$localFile->load( File::READ_LATEST );
$filename = $localFile->getName();
$hash = $this->getTempFileSha1Base36();
$badFileName = $this->checkBadFileName( $filename, $this->mDesiredDestName );
if ( $badFileName !== null ) {
$warnings['badfilename'] = $badFileName;
}
$unwantedFileExtensionDetails = $this->checkUnwantedFileExtensions( $this->mFinalExtension );
if ( $unwantedFileExtensionDetails !== null ) {
$warnings['filetype-unwanted-type'] = $unwantedFileExtensionDetails;
}
$fileSizeWarnings = $this->checkFileSize( $this->mFileSize );
if ( $fileSizeWarnings ) {
$warnings = array_merge( $warnings, $fileSizeWarnings );
}
$localFileExistsWarnings = $this->checkLocalFileExists( $localFile, $hash );
if ( $localFileExistsWarnings ) {
$warnings = array_merge( $warnings, $localFileExistsWarnings );
}
if ( $this->checkLocalFileWasDeleted( $localFile ) ) {
$warnings['was-deleted'] = $filename;
}
// If a file with the same name exists locally then the local file has already been tested
// for duplication of content
$ignoreLocalDupes = isset( $warnings['exists'] );
$dupes = $this->checkAgainstExistingDupes( $hash, $ignoreLocalDupes );
if ( $dupes ) {
$warnings['duplicate'] = $dupes;
}
$archivedDupes = $this->checkAgainstArchiveDupes( $hash, $user );
if ( $archivedDupes !== null ) {
$warnings['duplicate-archive'] = $archivedDupes;
}
return $warnings;
}
/**
* Convert the warnings array returned by checkWarnings() to something that
* can be serialized. File objects will be converted to an associative array
* with the following keys:
*
* - fileName: The name of the file
* - timestamp: The upload timestamp
*
* @param mixed[] $warnings
* @return mixed[]
*/
public static function makeWarningsSerializable( $warnings ) {
array_walk_recursive( $warnings, function ( &$param, $key ) {
if ( $param instanceof File ) {
$param = [
'fileName' => $param->getName(),
'timestamp' => $param->getTimestamp()
];
} elseif ( is_object( $param ) ) {
throw new InvalidArgumentException(
'UploadBase::makeWarningsSerializable: ' .
'Unexpected object of class ' . get_class( $param ) );
}
} );
return $warnings;
}
/**
* Check whether the resulting filename is different from the desired one,
* but ignore things like ucfirst() and spaces/underscore things
*
* @param string $filename
* @param string $desiredFileName
*
* @return string|null String that was determined to be bad or null if the filename is okay
*/
private function checkBadFileName( $filename, $desiredFileName ) {
$comparableName = str_replace( ' ', '_', $desiredFileName );
$comparableName = Title::capitalize( $comparableName, NS_FILE );
if ( $desiredFileName != $filename && $comparableName != $filename ) {
return $filename;
}
return null;
}
/**
* @param string $fileExtension The file extension to check
*
* @return array|null array with the following keys:
* 0 => string The final extension being used
* 1 => string[] The extensions that are allowed
* 2 => int The number of extensions that are allowed.
*/
private function checkUnwantedFileExtensions( $fileExtension ) {
global $wgCheckFileExtensions, $wgFileExtensions, $wgLang;
if ( $wgCheckFileExtensions ) {
$extensions = array_unique( $wgFileExtensions );
if ( !$this->checkFileExtension( $fileExtension, $extensions ) ) {
return [
$fileExtension,
$wgLang->commaList( $extensions ),
count( $extensions )
];
}
}
return null;
}
/**
* @param int $fileSize
*
* @return array warnings
*/
private function checkFileSize( $fileSize ) {
global $wgUploadSizeWarning;
$warnings = [];
if ( $wgUploadSizeWarning && ( $fileSize > $wgUploadSizeWarning ) ) {
$warnings['large-file'] = [
Message::sizeParam( $wgUploadSizeWarning ),
Message::sizeParam( $fileSize ),
];
}
if ( $fileSize == 0 ) {
$warnings['empty-file'] = true;
}
return $warnings;
}
/**
* @param LocalFile $localFile
* @param string $hash sha1 hash of the file to check
*
* @return array warnings
*/
private function checkLocalFileExists( LocalFile $localFile, $hash ) {
$warnings = [];
$exists = self::getExistsWarning( $localFile );
if ( $exists !== false ) {
$warnings['exists'] = $exists;
// check if file is an exact duplicate of current file version
if ( $hash === $localFile->getSha1() ) {
$warnings['no-change'] = $localFile;
}
// check if file is an exact duplicate of older versions of this file
$history = $localFile->getHistory();
foreach ( $history as $oldFile ) {
if ( $hash === $oldFile->getSha1() ) {
$warnings['duplicate-version'][] = $oldFile;
}
}
}
return $warnings;
}
private function checkLocalFileWasDeleted( LocalFile $localFile ) {
return $localFile->wasDeleted() && !$localFile->exists();
}
/**
* @param string $hash sha1 hash of the file to check
* @param bool $ignoreLocalDupes True to ignore local duplicates
*
* @return File[] Duplicate files, if found.
*/
private function checkAgainstExistingDupes( $hash, $ignoreLocalDupes ) {
$dupes = MediaWikiServices::getInstance()->getRepoGroup()->findBySha1( $hash );
$title = $this->getTitle();
foreach ( $dupes as $key => $dupe ) {
if (
( $dupe instanceof LocalFile ) &&
$ignoreLocalDupes &&
$title->equals( $dupe->getTitle() )
) {
unset( $dupes[$key] );
}
}
return $dupes;
}
/**
* @param string $hash sha1 hash of the file to check
* @param User $user
*
* @return string|null Name of the dupe or empty string if discovered (depending on visibility)
* null if the check discovered no dupes.
*/
private function checkAgainstArchiveDupes( $hash, User $user ) {
$archivedFile = new ArchivedFile( null, 0, '', $hash );
if ( $archivedFile->getID() > 0 ) {
if ( $archivedFile->userCan( File::DELETED_FILE, $user ) ) {
return $archivedFile->getName();
} else {
return '';
}
}
return null;
}
/**
* Really perform the upload. Stores the file in the local repo, watches
* if necessary and runs the UploadComplete hook.
*
* @param string $comment
* @param string $pageText
* @param bool $watch Whether the file page should be added to user's watchlist.
* (This doesn't check $user's permissions.)
* @param User $user
* @param string[] $tags Change tags to add to the log entry and page revision.
* (This doesn't check $user's permissions.)
* @param string|null $watchlistExpiry Optional watchlist expiry timestamp in any format
* acceptable to wfTimestamp().
* @return Status Indicating the whether the upload succeeded.
*
* @since 1.35 Accepts $watchlistExpiry parameter.
*/
public function performUpload(
$comment, $pageText, $watch, $user, $tags = [], ?string $watchlistExpiry = null
) {
$this->getLocalFile()->load( File::READ_LATEST );
$props = $this->mFileProps;
$error = null;
$this->getHookRunner()->onUploadVerifyUpload( $this, $user, $props, $comment, $pageText, $error );
if ( $error ) {
if ( !is_array( $error ) ) {
$error = [ $error ];
}
return Status::newFatal( ...$error );
}
$status = $this->getLocalFile()->upload(
$this->mTempPath,
$comment,
$pageText,
File::DELETE_SOURCE,
$props,
false,
$user,
$tags
);
if ( $status->isGood() ) {
if ( $watch ) {
WatchAction::doWatch(
$this->getLocalFile()->getTitle(),
$user,
User::IGNORE_USER_RIGHTS,
$watchlistExpiry
);
}
$this->getHookRunner()->onUploadComplete( $this );
$this->postProcessUpload();
}
return $status;
}
/**
* Perform extra steps after a successful upload.
*
* @stable to override
* @since 1.25
*/
public function postProcessUpload() {
}
/**
* Returns the title of the file to be uploaded. Sets mTitleError in case
* the name was illegal.
*
* @return Title|null The title of the file or null in case the name was illegal
*/
public function getTitle() {
if ( $this->mTitle !== false ) {
return $this->mTitle;
}
if ( !is_string( $this->mDesiredDestName ) ) {
$this->mTitleError = self::ILLEGAL_FILENAME;
$this->mTitle = null;
return $this->mTitle;
}
/* Assume that if a user specified File:Something.jpg, this is an error
* and that the namespace prefix needs to be stripped of.
*/
$title = Title::newFromText( $this->mDesiredDestName );
if ( $title && $title->getNamespace() == NS_FILE ) {
$this->mFilteredName = $title->getDBkey();
} else {
$this->mFilteredName = $this->mDesiredDestName;
}
# oi_archive_name is max 255 bytes, which include a timestamp and an
# exclamation mark, so restrict file name to 240 bytes.
if ( strlen( $this->mFilteredName ) > 240 ) {
$this->mTitleError = self::FILENAME_TOO_LONG;
$this->mTitle = null;
return $this->mTitle;
}
/**
* Chop off any directories in the given filename. Then
* filter out illegal characters, and try to make a legible name
* out of it. We'll strip some silently that Title would die on.
*/
$this->mFilteredName = wfStripIllegalFilenameChars( $this->mFilteredName );
/* Normalize to title form before we do any further processing */
$nt = Title::makeTitleSafe( NS_FILE, $this->mFilteredName );
if ( $nt === null ) {
$this->mTitleError = self::ILLEGAL_FILENAME;
$this->mTitle = null;
return $this->mTitle;
}
$this->mFilteredName = $nt->getDBkey();
/**
* We'll want to blacklist against *any* 'extension', and use
* only the final one for the whitelist.
*/
list( $partname, $ext ) = $this->splitExtensions( $this->mFilteredName );
if ( $ext !== [] ) {
$this->mFinalExtension = trim( end( $ext ) );
} else {
$this->mFinalExtension = '';
# No extension, try guessing one
$magic = MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer();
$mime = $magic->guessMimeType( $this->mTempPath );
if ( $mime !== 'unknown/unknown' ) {
# Get a space separated list of extensions
$mimeExt = $magic->getExtensionFromMimeTypeOrNull( $mime );
if ( $mimeExt !== null ) {
# Set the extension to the canonical extension
$this->mFinalExtension = $mimeExt;
# Fix up the other variables
$this->mFilteredName .= ".{$this->mFinalExtension}";
$nt = Title::makeTitleSafe( NS_FILE, $this->mFilteredName );
$ext = [ $this->mFinalExtension ];
}
}
}
/* Don't allow users to override the blacklist (check file extension) */
global $wgCheckFileExtensions, $wgStrictFileExtensions;
global $wgFileExtensions, $wgFileBlacklist;
$blackListedExtensions = $this->checkFileExtensionList( $ext, $wgFileBlacklist );
if ( $this->mFinalExtension == '' ) {
$this->mTitleError = self::FILETYPE_MISSING;
$this->mTitle = null;
return $this->mTitle;
} elseif ( $blackListedExtensions ||
( $wgCheckFileExtensions && $wgStrictFileExtensions &&
!$this->checkFileExtension( $this->mFinalExtension, $wgFileExtensions ) )
) {
$this->mBlackListedExtensions = $blackListedExtensions;
$this->mTitleError = self::FILETYPE_BADTYPE;
$this->mTitle = null;
return $this->mTitle;
}
// Windows may be broken with special characters, see T3780
if ( !preg_match( '/^[\x0-\x7f]*$/', $nt->getText() )
&& !MediaWikiServices::getInstance()->getRepoGroup()
->getLocalRepo()->backendSupportsUnicodePaths()
) {
$this->mTitleError = self::WINDOWS_NONASCII_FILENAME;
$this->mTitle = null;
return $this->mTitle;
}
# If there was more than one "extension", reassemble the base
# filename to prevent bogus complaints about length
if ( count( $ext ) > 1 ) {
$iterations = count( $ext ) - 1;
for ( $i = 0; $i < $iterations; $i++ ) {
$partname .= '.' . $ext[$i];
}
}
if ( strlen( $partname ) < 1 ) {
$this->mTitleError = self::MIN_LENGTH_PARTNAME;
$this->mTitle = null;
return $this->mTitle;
}
$this->mTitle = $nt;
return $this->mTitle;
}
/**
* Return the local file and initializes if necessary.
*
* @stable to override
* @return LocalFile|null
*/
public function getLocalFile() {
if ( $this->mLocalFile === null ) {
$nt = $this->getTitle();
$this->mLocalFile = $nt === null
? null
: MediaWikiServices::getInstance()->getRepoGroup()->getLocalRepo()->newFile( $nt );
}
return $this->mLocalFile;
}
/**
* @return UploadStashFile|null
*/
public function getStashFile() {
return $this->mStashFile;
}
/**
* Like stashFile(), but respects extensions' wishes to prevent the stashing. verifyUpload() must
* be called before calling this method (unless $isPartial is true).
*
* Upload stash exceptions are also caught and converted to an error status.
*
* @since 1.28
* @stable to override
* @param User $user
* @param bool $isPartial Pass `true` if this is a part of a chunked upload (not a complete file).
* @return Status If successful, value is an UploadStashFile instance
*/
public function tryStashFile( User $user, $isPartial = false ) {
if ( !$isPartial ) {
$error = $this->runUploadStashFileHook( $user );
if ( $error ) {
return Status::newFatal( ...$error );
}
}
try {
$file = $this->doStashFile( $user );
return Status::newGood( $file );
} catch ( UploadStashException $e ) {
return Status::newFatal( 'uploadstash-exception', get_class( $e ), $e->getMessage() );
}
}
/**
* @param User $user
* @return array|null Error message and parameters, null if there's no error
*/
protected function runUploadStashFileHook( User $user ) {
$props = $this->mFileProps;
$error = null;
$this->getHookRunner()->onUploadStashFile( $this, $user, $props, $error );
if ( $error && !is_array( $error ) ) {
$error = [ $error ];
}
return $error;
}
/**
* If the user does not supply all necessary information in the first upload
* form submission (either by accident or by design) then we may want to
* stash the file temporarily, get more information, and publish the file
* later.
*
* This method will stash a file in a temporary directory for later
* processing, and save the necessary descriptive info into the database.
* This method returns the file object, which also has a 'fileKey' property
* which can be passed through a form or API request to find this stashed
* file again.
*
* @deprecated since 1.28 Use tryStashFile() instead
* @param User|null $user
* @return UploadStashFile Stashed file
* @throws UploadStashBadPathException
* @throws UploadStashFileException
* @throws UploadStashNotLoggedInException
*/
public function stashFile( User $user = null ) {
wfDeprecated( __METHOD__, '1.28' );
return $this->doStashFile( $user );
}
/**
* Implementation for stashFile() and tryStashFile().
*
* @stable to override
* @param User|null $user
* @return UploadStashFile Stashed file
*/
protected function doStashFile( User $user = null ) {
$stash = MediaWikiServices::getInstance()->getRepoGroup()
->getLocalRepo()->getUploadStash( $user );
$file = $stash->stashFile( $this->mTempPath, $this->getSourceType() );
$this->mStashFile = $file;
return $file;
}
/**
* If we've modified the upload file we need to manually remove it
* on exit to clean up.
*/
public function cleanupTempFile() {
if ( $this->mRemoveTempFile && $this->tempFileObj ) {
// Delete when all relevant TempFSFile handles go out of scope
wfDebug( __METHOD__ . ": Marked temporary file '{$this->mTempPath}' for removal" );
$this->tempFileObj->autocollect();
}
}
public function getTempPath() {
return $this->mTempPath;
}
/**
* Split a file into a base name and all dot-delimited 'extensions'
* on the end. Some web server configurations will fall back to
* earlier pseudo-'extensions' to determine type and execute
* scripts, so the blacklist needs to check them all.
*
* @param string $filename
* @return array [ string, string[] ]
*/
public static function splitExtensions( $filename ) {
$bits = explode( '.', $filename );
$basename = array_shift( $bits );
return [ $basename, $bits ];
}
/**
* Perform case-insensitive match against a list of file extensions.
* Returns true if the extension is in the list.
*
* @param string $ext
* @param array $list
* @return bool
*/
public static function checkFileExtension( $ext, $list ) {
return in_array( strtolower( $ext ), $list );
}
/**
* Perform case-insensitive match against a list of file extensions.
* Returns an array of matching extensions.
*
* @param string[] $ext
* @param string[] $list
* @return string[]
*/
public static function checkFileExtensionList( $ext, $list ) {
return array_intersect( array_map( 'strtolower', $ext ), $list );
}
/**
* Checks if the MIME type of the uploaded file matches the file extension.
*
* @param string $mime The MIME type of the uploaded file
* @param string $extension The filename extension that the file is to be served with
* @return bool
*/
public static function verifyExtension( $mime, $extension ) {
$magic = MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer();
if ( !$mime || $mime == 'unknown' || $mime == 'unknown/unknown' ) {
if ( !$magic->isRecognizableExtension( $extension ) ) {
wfDebug( __METHOD__ . ": passing file with unknown detected mime type; " .
"unrecognized extension '$extension', can't verify" );
return true;
} else {
wfDebug( __METHOD__ . ": rejecting file with unknown detected mime type; " .
"recognized extension '$extension', so probably invalid file" );
return false;
}
}
$match = $magic->isMatchingExtension( $extension, $mime );
if ( $match === null ) {
if ( $magic->getTypesForExtension( $extension ) !== null ) {
wfDebug( __METHOD__ . ": No extension known for $mime, but we know a mime for $extension" );
return false;
} else {
wfDebug( __METHOD__ . ": no file extension known for mime type $mime, passing file" );
return true;
}
} elseif ( $match === true ) {
wfDebug( __METHOD__ . ": mime type $mime matches extension $extension, passing file" );
/** @todo If it's a bitmap, make sure PHP or ImageMagick resp. can handle it! */
return true;
} else {
wfDebug( __METHOD__
. ": mime type $mime mismatches file extension $extension, rejecting file" );
return false;
}
}
/**
* Heuristic for detecting files that *could* contain JavaScript instructions or
* things that may look like HTML to a browser and are thus
* potentially harmful. The present implementation will produce false
* positives in some situations.
*
* @param string $file Pathname to the temporary upload file
* @param string $mime The MIME type of the file
* @param string $extension The extension of the file
* @return bool True if the file contains something looking like embedded scripts
*/
public static function detectScript( $file, $mime, $extension ) {
# ugly hack: for text files, always look at the entire file.
# For binary field, just check the first K.
$isText = strpos( $mime, 'text/' ) === 0;
if ( $isText ) {
$chunk = file_get_contents( $file );
} else {
$fp = fopen( $file, 'rb' );
$chunk = fread( $fp, 1024 );
fclose( $fp );
}
$chunk = strtolower( $chunk );
if ( !$chunk ) {
return false;
}
# decode from UTF-16 if needed (could be used for obfuscation).
if ( substr( $chunk, 0, 2 ) == "\xfe\xff" ) {
$enc = 'UTF-16BE';
} elseif ( substr( $chunk, 0, 2 ) == "\xff\xfe" ) {
$enc = 'UTF-16LE';
} else {
$enc = null;
}
if ( $enc !== null ) {
$chunk = iconv( $enc, "ASCII//IGNORE", $chunk );
}
$chunk = trim( $chunk );
/** @todo FIXME: Convert from UTF-16 if necessary! */
wfDebug( __METHOD__ . ": checking for embedded scripts and HTML stuff" );
# check for HTML doctype
if ( preg_match( "/!si", $contents, $matches ) ) {
if ( preg_match( $encodingRegex, $matches[1], $encMatch )
&& !in_array( strtoupper( $encMatch[1] ), self::$safeXmlEncodings )
) {
wfDebug( __METHOD__ . ": Found unsafe XML encoding '{$encMatch[1]}'" );
return true;
}
} elseif ( preg_match( "!<\?xml\b!si", $contents ) ) {
// Start of XML declaration without an end in the first $wgSVGMetadataCutoff
// bytes. There shouldn't be a legitimate reason for this to happen.
wfDebug( __METHOD__ . ": Unmatched XML declaration start" );
return true;
} elseif ( substr( $contents, 0, 4 ) == "\x4C\x6F\xA7\x94" ) {
// EBCDIC encoded XML
wfDebug( __METHOD__ . ": EBCDIC Encoded XML" );
return true;
}
// It's possible the file is encoded with multi-byte encoding, so re-encode attempt to
// detect the encoding in case is specifies an encoding not whitelisted in self::$safeXmlEncodings
$attemptEncodings = [ 'UTF-16', 'UTF-16BE', 'UTF-32', 'UTF-32BE' ];
foreach ( $attemptEncodings as $encoding ) {
Wikimedia\suppressWarnings();
$str = iconv( $encoding, 'UTF-8', $contents );
Wikimedia\restoreWarnings();
if ( $str != '' && preg_match( "!<\?xml\b(.*?)\?>!si", $str, $matches ) ) {
if ( preg_match( $encodingRegex, $matches[1], $encMatch )
&& !in_array( strtoupper( $encMatch[1] ), self::$safeXmlEncodings )
) {
wfDebug( __METHOD__ . ": Found unsafe XML encoding '{$encMatch[1]}'" );
return true;
}
} elseif ( $str != '' && preg_match( "!<\?xml\b!si", $str ) ) {
// Start of XML declaration without an end in the first $wgSVGMetadataCutoff
// bytes. There shouldn't be a legitimate reason for this to happen.
wfDebug( __METHOD__ . ": Unmatched XML declaration start" );
return true;
}
}
return false;
}
/**
* @param string $filename
* @param bool $partial
* @return bool|array
*/
protected function detectScriptInSvg( $filename, $partial ) {
$this->mSVGNSError = false;
$check = new XmlTypeCheck(
$filename,
[ $this, 'checkSvgScriptCallback' ],
true,
[
'processing_instruction_handler' => [ __CLASS__, 'checkSvgPICallback' ],
'external_dtd_handler' => [ __CLASS__, 'checkSvgExternalDTD' ],
]
);
if ( $check->wellFormed !== true ) {
// Invalid xml (T60553)
// But only when non-partial (T67724)
return $partial ? false : [ 'uploadinvalidxml' ];
} elseif ( $check->filterMatch ) {
if ( $this->mSVGNSError ) {
return [ 'uploadscriptednamespace', $this->mSVGNSError ];
}
return $check->filterMatchType;
}
return false;
}
/**
* Callback to filter SVG Processing Instructions.
* @param string $target Processing instruction name
* @param string $data Processing instruction attribute and value
* @return bool|array
*/
public static function checkSvgPICallback( $target, $data ) {
// Don't allow external stylesheets (T59550)
if ( preg_match( '/xml-stylesheet/i', $target ) ) {
return [ 'upload-scripted-pi-callback' ];
}
return false;
}
/**
* Verify that DTD urls referenced are only the standard dtds
*
* Browsers seem to ignore external dtds. However just to be on the
* safe side, only allow dtds from the svg standard.
*
* @param string $type PUBLIC or SYSTEM
* @param string $publicId The well-known public identifier for the dtd
* @param string $systemId The url for the external dtd
* @return bool|array
*/
public static function checkSvgExternalDTD( $type, $publicId, $systemId ) {
// This doesn't include the XHTML+MathML+SVG doctype since we don't
// allow XHTML anyways.
$allowedDTDs = [
'http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd',
'http://www.w3.org/TR/2001/REC-SVG-20010904/DTD/svg10.dtd',
'http://www.w3.org/Graphics/SVG/1.1/DTD/svg11-basic.dtd',
'http://www.w3.org/Graphics/SVG/1.1/DTD/svg11-tiny.dtd',
// https://phabricator.wikimedia.org/T168856
'http://www.w3.org/TR/2001/PR-SVG-20010719/DTD/svg10.dtd',
];
if ( $type !== 'PUBLIC'
|| !in_array( $systemId, $allowedDTDs )
|| strpos( $publicId, "-//W3C//" ) !== 0
) {
return [ 'upload-scripted-dtd' ];
}
return false;
}
/**
* @todo Replace this with a whitelist filter!
* @param string $element
* @param array $attribs
* @param string|null $data
* @return bool|array
*/
public function checkSvgScriptCallback( $element, $attribs, $data = null ) {
list( $namespace, $strippedElement ) = $this->splitXmlNamespace( $element );
// We specifically don't include:
// http://www.w3.org/1999/xhtml (T62771)
static $validNamespaces = [
'',
'adobe:ns:meta/',
'http://creativecommons.org/ns#',
'http://inkscape.sourceforge.net/dtd/sodipodi-0.dtd',
'http://ns.adobe.com/adobeillustrator/10.0/',
'http://ns.adobe.com/adobesvgviewerextensions/3.0/',
'http://ns.adobe.com/extensibility/1.0/',
'http://ns.adobe.com/flows/1.0/',
'http://ns.adobe.com/illustrator/1.0/',
'http://ns.adobe.com/imagereplacement/1.0/',
'http://ns.adobe.com/pdf/1.3/',
'http://ns.adobe.com/photoshop/1.0/',
'http://ns.adobe.com/saveforweb/1.0/',
'http://ns.adobe.com/variables/1.0/',
'http://ns.adobe.com/xap/1.0/',
'http://ns.adobe.com/xap/1.0/g/',
'http://ns.adobe.com/xap/1.0/g/img/',
'http://ns.adobe.com/xap/1.0/mm/',
'http://ns.adobe.com/xap/1.0/rights/',
'http://ns.adobe.com/xap/1.0/stype/dimensions#',
'http://ns.adobe.com/xap/1.0/stype/font#',
'http://ns.adobe.com/xap/1.0/stype/manifestitem#',
'http://ns.adobe.com/xap/1.0/stype/resourceevent#',
'http://ns.adobe.com/xap/1.0/stype/resourceref#',
'http://ns.adobe.com/xap/1.0/t/pg/',
'http://purl.org/dc/elements/1.1/',
'http://purl.org/dc/elements/1.1',
'http://schemas.microsoft.com/visio/2003/svgextensions/',
'http://sodipodi.sourceforge.net/dtd/sodipodi-0.dtd',
'http://taptrix.com/inkpad/svg_extensions',
'http://web.resource.org/cc/',
'http://www.freesoftware.fsf.org/bkchem/cdml',
'http://www.inkscape.org/namespaces/inkscape',
'http://www.opengis.net/gml',
'http://www.w3.org/1999/02/22-rdf-syntax-ns#',
'http://www.w3.org/2000/svg',
'http://www.w3.org/tr/rec-rdf-syntax/',
'http://www.w3.org/2000/01/rdf-schema#',
];
// Inkscape mangles namespace definitions created by Adobe Illustrator.
// This is nasty but harmless. (T144827)
$isBuggyInkscape = preg_match( '/^&(#38;)*ns_[a-z_]+;$/', $namespace );
if ( !( $isBuggyInkscape || in_array( $namespace, $validNamespaces ) ) ) {
wfDebug( __METHOD__ . ": Non-svg namespace '$namespace' in uploaded file." );
/** @todo Return a status object to a closure in XmlTypeCheck, for MW1.21+ */
$this->mSVGNSError = $namespace;
return true;
}
/*
* check for elements that can contain javascript
*/
if ( $strippedElement == 'script' ) {
wfDebug( __METHOD__ . ": Found script element '$element' in uploaded file." );
return [ 'uploaded-script-svg', $strippedElement ];
}
# e.g.,
if ( $strippedElement == 'handler' ) {
wfDebug( __METHOD__ . ": Found scriptable element '$element' in uploaded file." );
return [ 'uploaded-script-svg', $strippedElement ];
}
# SVG reported in Feb '12 that used xml:stylesheet to generate javascript block
if ( $strippedElement == 'stylesheet' ) {
wfDebug( __METHOD__ . ": Found scriptable element '$element' in uploaded file." );
return [ 'uploaded-script-svg', $strippedElement ];
}
# Block iframes, in case they pass the namespace check
if ( $strippedElement == 'iframe' ) {
wfDebug( __METHOD__ . ": iframe in uploaded file." );
return [ 'uploaded-script-svg', $strippedElement ];
}
# Check