Add ability to pre-render thumbnails at upload time
[lhc/web/wiklou.git] / includes / upload / UploadBase.php
1 <?php
2 /**
3 * Base class for the backend of file upload.
4 *
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.
9 *
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.
14 *
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
19 *
20 * @file
21 * @ingroup Upload
22 */
23
24 /**
25 * @defgroup Upload Upload related
26 */
27
28 /**
29 * @ingroup Upload
30 *
31 * UploadBase and subclasses are the backend of MediaWiki's file uploads.
32 * The frontends are formed by ApiUpload and SpecialUpload.
33 *
34 * @author Brion Vibber
35 * @author Bryan Tong Minh
36 * @author Michael Dale
37 */
38 abstract class UploadBase {
39 protected $mTempPath;
40 protected $mDesiredDestName, $mDestName, $mRemoveTempFile, $mSourceType;
41 protected $mTitle = false, $mTitleError = 0;
42 protected $mFilteredName, $mFinalExtension;
43 protected $mLocalFile, $mFileSize, $mFileProps;
44 protected $mBlackListedExtensions;
45 protected $mJavaDetected, $mSVGNSError;
46
47 protected static $safeXmlEncodings = array(
48 'UTF-8',
49 'ISO-8859-1',
50 'ISO-8859-2',
51 'UTF-16',
52 'UTF-32'
53 );
54
55 const SUCCESS = 0;
56 const OK = 0;
57 const EMPTY_FILE = 3;
58 const MIN_LENGTH_PARTNAME = 4;
59 const ILLEGAL_FILENAME = 5;
60 const OVERWRITE_EXISTING_FILE = 7; # Not used anymore; handled by verifyTitlePermissions()
61 const FILETYPE_MISSING = 8;
62 const FILETYPE_BADTYPE = 9;
63 const VERIFICATION_ERROR = 10;
64
65 # HOOK_ABORTED is the new name of UPLOAD_VERIFICATION_ERROR
66 const UPLOAD_VERIFICATION_ERROR = 11;
67 const HOOK_ABORTED = 11;
68 const FILE_TOO_LARGE = 12;
69 const WINDOWS_NONASCII_FILENAME = 13;
70 const FILENAME_TOO_LONG = 14;
71
72 const SESSION_STATUS_KEY = 'wsUploadStatusData';
73
74 /**
75 * @param int $error
76 * @return string
77 */
78 public function getVerificationErrorCode( $error ) {
79 $code_to_status = array(
80 self::EMPTY_FILE => 'empty-file',
81 self::FILE_TOO_LARGE => 'file-too-large',
82 self::FILETYPE_MISSING => 'filetype-missing',
83 self::FILETYPE_BADTYPE => 'filetype-banned',
84 self::MIN_LENGTH_PARTNAME => 'filename-tooshort',
85 self::ILLEGAL_FILENAME => 'illegal-filename',
86 self::OVERWRITE_EXISTING_FILE => 'overwrite',
87 self::VERIFICATION_ERROR => 'verification-error',
88 self::HOOK_ABORTED => 'hookaborted',
89 self::WINDOWS_NONASCII_FILENAME => 'windows-nonascii-filename',
90 self::FILENAME_TOO_LONG => 'filename-toolong',
91 );
92 if ( isset( $code_to_status[$error] ) ) {
93 return $code_to_status[$error];
94 }
95
96 return 'unknown-error';
97 }
98
99 /**
100 * Returns true if uploads are enabled.
101 * Can be override by subclasses.
102 * @return bool
103 */
104 public static function isEnabled() {
105 global $wgEnableUploads;
106
107 if ( !$wgEnableUploads ) {
108 return false;
109 }
110
111 # Check php's file_uploads setting
112 return wfIsHHVM() || wfIniGetBool( 'file_uploads' );
113 }
114
115 /**
116 * Returns true if the user can use this upload module or else a string
117 * identifying the missing permission.
118 * Can be overridden by subclasses.
119 *
120 * @param User $user
121 * @return bool|string
122 */
123 public static function isAllowed( $user ) {
124 foreach ( array( 'upload', 'edit' ) as $permission ) {
125 if ( !$user->isAllowed( $permission ) ) {
126 return $permission;
127 }
128 }
129
130 return true;
131 }
132
133 // Upload handlers. Should probably just be a global.
134 private static $uploadHandlers = array( 'Stash', 'File', 'Url' );
135
136 /**
137 * Create a form of UploadBase depending on wpSourceType and initializes it
138 *
139 * @param WebRequest $request
140 * @param string|null $type
141 * @return null|UploadBase
142 */
143 public static function createFromRequest( &$request, $type = null ) {
144 $type = $type ? $type : $request->getVal( 'wpSourceType', 'File' );
145
146 if ( !$type ) {
147 return null;
148 }
149
150 // Get the upload class
151 $type = ucfirst( $type );
152
153 // Give hooks the chance to handle this request
154 $className = null;
155 wfRunHooks( 'UploadCreateFromRequest', array( $type, &$className ) );
156 if ( is_null( $className ) ) {
157 $className = 'UploadFrom' . $type;
158 wfDebug( __METHOD__ . ": class name: $className\n" );
159 if ( !in_array( $type, self::$uploadHandlers ) ) {
160 return null;
161 }
162 }
163
164 // Check whether this upload class is enabled
165 if ( !call_user_func( array( $className, 'isEnabled' ) ) ) {
166 return null;
167 }
168
169 // Check whether the request is valid
170 if ( !call_user_func( array( $className, 'isValidRequest' ), $request ) ) {
171 return null;
172 }
173
174 /** @var UploadBase $handler */
175 $handler = new $className;
176
177 $handler->initializeFromRequest( $request );
178
179 return $handler;
180 }
181
182 /**
183 * Check whether a request if valid for this handler
184 * @param WebRequest $request
185 * @return bool
186 */
187 public static function isValidRequest( $request ) {
188 return false;
189 }
190
191 public function __construct() {
192 }
193
194 /**
195 * Returns the upload type. Should be overridden by child classes
196 *
197 * @since 1.18
198 * @return string
199 */
200 public function getSourceType() {
201 return null;
202 }
203
204 /**
205 * Initialize the path information
206 * @param string $name The desired destination name
207 * @param string $tempPath The temporary path
208 * @param int $fileSize The file size
209 * @param bool $removeTempFile (false) remove the temporary file?
210 * @throws MWException
211 */
212 public function initializePathInfo( $name, $tempPath, $fileSize, $removeTempFile = false ) {
213 $this->mDesiredDestName = $name;
214 if ( FileBackend::isStoragePath( $tempPath ) ) {
215 throw new MWException( __METHOD__ . " given storage path `$tempPath`." );
216 }
217 $this->mTempPath = $tempPath;
218 $this->mFileSize = $fileSize;
219 $this->mRemoveTempFile = $removeTempFile;
220 }
221
222 /**
223 * Initialize from a WebRequest. Override this in a subclass.
224 *
225 * @param WebRequest $request
226 */
227 abstract public function initializeFromRequest( &$request );
228
229 /**
230 * Fetch the file. Usually a no-op
231 * @return Status
232 */
233 public function fetchFile() {
234 return Status::newGood();
235 }
236
237 /**
238 * Return true if the file is empty
239 * @return bool
240 */
241 public function isEmptyFile() {
242 return empty( $this->mFileSize );
243 }
244
245 /**
246 * Return the file size
247 * @return int
248 */
249 public function getFileSize() {
250 return $this->mFileSize;
251 }
252
253 /**
254 * Get the base 36 SHA1 of the file
255 * @return string
256 */
257 public function getTempFileSha1Base36() {
258 return FSFile::getSha1Base36FromPath( $this->mTempPath );
259 }
260
261 /**
262 * @param string $srcPath The source path
263 * @return string|bool The real path if it was a virtual URL Returns false on failure
264 */
265 function getRealPath( $srcPath ) {
266 wfProfileIn( __METHOD__ );
267 $repo = RepoGroup::singleton()->getLocalRepo();
268 if ( $repo->isVirtualUrl( $srcPath ) ) {
269 /** @todo Just make uploads work with storage paths UploadFromStash
270 * loads files via virtual URLs.
271 */
272 $tmpFile = $repo->getLocalCopy( $srcPath );
273 if ( $tmpFile ) {
274 $tmpFile->bind( $this ); // keep alive with $this
275 }
276 $path = $tmpFile ? $tmpFile->getPath() : false;
277 } else {
278 $path = $srcPath;
279 }
280 wfProfileOut( __METHOD__ );
281
282 return $path;
283 }
284
285 /**
286 * Verify whether the upload is sane.
287 * @return mixed Const self::OK or else an array with error information
288 */
289 public function verifyUpload() {
290 wfProfileIn( __METHOD__ );
291
292 /**
293 * If there was no filename or a zero size given, give up quick.
294 */
295 if ( $this->isEmptyFile() ) {
296 wfProfileOut( __METHOD__ );
297
298 return array( 'status' => self::EMPTY_FILE );
299 }
300
301 /**
302 * Honor $wgMaxUploadSize
303 */
304 $maxSize = self::getMaxUploadSize( $this->getSourceType() );
305 if ( $this->mFileSize > $maxSize ) {
306 wfProfileOut( __METHOD__ );
307
308 return array(
309 'status' => self::FILE_TOO_LARGE,
310 'max' => $maxSize,
311 );
312 }
313
314 /**
315 * Look at the contents of the file; if we can recognize the
316 * type but it's corrupt or data of the wrong type, we should
317 * probably not accept it.
318 */
319 $verification = $this->verifyFile();
320 if ( $verification !== true ) {
321 wfProfileOut( __METHOD__ );
322
323 return array(
324 'status' => self::VERIFICATION_ERROR,
325 'details' => $verification
326 );
327 }
328
329 /**
330 * Make sure this file can be created
331 */
332 $result = $this->validateName();
333 if ( $result !== true ) {
334 wfProfileOut( __METHOD__ );
335
336 return $result;
337 }
338
339 $error = '';
340 if ( !wfRunHooks( 'UploadVerification',
341 array( $this->mDestName, $this->mTempPath, &$error ) )
342 ) {
343 wfProfileOut( __METHOD__ );
344
345 return array( 'status' => self::HOOK_ABORTED, 'error' => $error );
346 }
347
348 wfProfileOut( __METHOD__ );
349
350 return array( 'status' => self::OK );
351 }
352
353 /**
354 * Verify that the name is valid and, if necessary, that we can overwrite
355 *
356 * @return mixed True if valid, otherwise and array with 'status'
357 * and other keys
358 */
359 public function validateName() {
360 $nt = $this->getTitle();
361 if ( is_null( $nt ) ) {
362 $result = array( 'status' => $this->mTitleError );
363 if ( $this->mTitleError == self::ILLEGAL_FILENAME ) {
364 $result['filtered'] = $this->mFilteredName;
365 }
366 if ( $this->mTitleError == self::FILETYPE_BADTYPE ) {
367 $result['finalExt'] = $this->mFinalExtension;
368 if ( count( $this->mBlackListedExtensions ) ) {
369 $result['blacklistedExt'] = $this->mBlackListedExtensions;
370 }
371 }
372
373 return $result;
374 }
375 $this->mDestName = $this->getLocalFile()->getName();
376
377 return true;
378 }
379
380 /**
381 * Verify the MIME type.
382 *
383 * @note Only checks that it is not an evil MIME. The "does it have
384 * correct extension given its MIME type?" check is in verifyFile.
385 * in `verifyFile()` that MIME type and file extension correlate.
386 * @param string $mime Representing the MIME
387 * @return mixed True if the file is verified, an array otherwise
388 */
389 protected function verifyMimeType( $mime ) {
390 global $wgVerifyMimeType;
391 wfProfileIn( __METHOD__ );
392 if ( $wgVerifyMimeType ) {
393 wfDebug( "mime: <$mime> extension: <{$this->mFinalExtension}>\n" );
394 global $wgMimeTypeBlacklist;
395 if ( $this->checkFileExtension( $mime, $wgMimeTypeBlacklist ) ) {
396 wfProfileOut( __METHOD__ );
397
398 return array( 'filetype-badmime', $mime );
399 }
400
401 # Check what Internet Explorer would detect
402 $fp = fopen( $this->mTempPath, 'rb' );
403 $chunk = fread( $fp, 256 );
404 fclose( $fp );
405
406 $magic = MimeMagic::singleton();
407 $extMime = $magic->guessTypesForExtension( $this->mFinalExtension );
408 $ieTypes = $magic->getIEMimeTypes( $this->mTempPath, $chunk, $extMime );
409 foreach ( $ieTypes as $ieType ) {
410 if ( $this->checkFileExtension( $ieType, $wgMimeTypeBlacklist ) ) {
411 wfProfileOut( __METHOD__ );
412
413 return array( 'filetype-bad-ie-mime', $ieType );
414 }
415 }
416 }
417
418 wfProfileOut( __METHOD__ );
419
420 return true;
421 }
422
423 /**
424 * Verifies that it's ok to include the uploaded file
425 *
426 * @return mixed True of the file is verified, array otherwise.
427 */
428 protected function verifyFile() {
429 global $wgVerifyMimeType;
430 wfProfileIn( __METHOD__ );
431
432 $status = $this->verifyPartialFile();
433 if ( $status !== true ) {
434 wfProfileOut( __METHOD__ );
435
436 return $status;
437 }
438
439 $this->mFileProps = FSFile::getPropsFromPath( $this->mTempPath, $this->mFinalExtension );
440 $mime = $this->mFileProps['mime'];
441
442 if ( $wgVerifyMimeType ) {
443 # XXX: Missing extension will be caught by validateName() via getTitle()
444 if ( $this->mFinalExtension != '' && !$this->verifyExtension( $mime, $this->mFinalExtension ) ) {
445 wfProfileOut( __METHOD__ );
446
447 return array( 'filetype-mime-mismatch', $this->mFinalExtension, $mime );
448 }
449 }
450
451 $handler = MediaHandler::getHandler( $mime );
452 if ( $handler ) {
453 $handlerStatus = $handler->verifyUpload( $this->mTempPath );
454 if ( !$handlerStatus->isOK() ) {
455 $errors = $handlerStatus->getErrorsArray();
456 wfProfileOut( __METHOD__ );
457
458 return reset( $errors );
459 }
460 }
461
462 wfRunHooks( 'UploadVerifyFile', array( $this, $mime, &$status ) );
463 if ( $status !== true ) {
464 wfProfileOut( __METHOD__ );
465
466 return $status;
467 }
468
469 wfDebug( __METHOD__ . ": all clear; passing.\n" );
470 wfProfileOut( __METHOD__ );
471
472 return true;
473 }
474
475 /**
476 * A verification routine suitable for partial files
477 *
478 * Runs the blacklist checks, but not any checks that may
479 * assume the entire file is present.
480 *
481 * @return mixed True for valid or array with error message key.
482 */
483 protected function verifyPartialFile() {
484 global $wgAllowJavaUploads, $wgDisableUploadScriptChecks;
485 wfProfileIn( __METHOD__ );
486
487 # getTitle() sets some internal parameters like $this->mFinalExtension
488 $this->getTitle();
489
490 $this->mFileProps = FSFile::getPropsFromPath( $this->mTempPath, $this->mFinalExtension );
491
492 # check MIME type, if desired
493 $mime = $this->mFileProps['file-mime'];
494 $status = $this->verifyMimeType( $mime );
495 if ( $status !== true ) {
496 wfProfileOut( __METHOD__ );
497
498 return $status;
499 }
500
501 # check for htmlish code and javascript
502 if ( !$wgDisableUploadScriptChecks ) {
503 if ( self::detectScript( $this->mTempPath, $mime, $this->mFinalExtension ) ) {
504 wfProfileOut( __METHOD__ );
505
506 return array( 'uploadscripted' );
507 }
508 if ( $this->mFinalExtension == 'svg' || $mime == 'image/svg+xml' ) {
509 $svgStatus = $this->detectScriptInSvg( $this->mTempPath );
510 if ( $svgStatus !== false ) {
511 wfProfileOut( __METHOD__ );
512
513 return $svgStatus;
514 }
515 }
516 }
517
518 # Check for Java applets, which if uploaded can bypass cross-site
519 # restrictions.
520 if ( !$wgAllowJavaUploads ) {
521 $this->mJavaDetected = false;
522 $zipStatus = ZipDirectoryReader::read( $this->mTempPath,
523 array( $this, 'zipEntryCallback' ) );
524 if ( !$zipStatus->isOK() ) {
525 $errors = $zipStatus->getErrorsArray();
526 $error = reset( $errors );
527 if ( $error[0] !== 'zip-wrong-format' ) {
528 wfProfileOut( __METHOD__ );
529
530 return $error;
531 }
532 }
533 if ( $this->mJavaDetected ) {
534 wfProfileOut( __METHOD__ );
535
536 return array( 'uploadjava' );
537 }
538 }
539
540 # Scan the uploaded file for viruses
541 $virus = $this->detectVirus( $this->mTempPath );
542 if ( $virus ) {
543 wfProfileOut( __METHOD__ );
544
545 return array( 'uploadvirus', $virus );
546 }
547
548 wfProfileOut( __METHOD__ );
549
550 return true;
551 }
552
553 /**
554 * Callback for ZipDirectoryReader to detect Java class files.
555 *
556 * @param array $entry
557 */
558 function zipEntryCallback( $entry ) {
559 $names = array( $entry['name'] );
560
561 // If there is a null character, cut off the name at it, because JDK's
562 // ZIP_GetEntry() uses strcmp() if the name hashes match. If a file name
563 // were constructed which had ".class\0" followed by a string chosen to
564 // make the hash collide with the truncated name, that file could be
565 // returned in response to a request for the .class file.
566 $nullPos = strpos( $entry['name'], "\000" );
567 if ( $nullPos !== false ) {
568 $names[] = substr( $entry['name'], 0, $nullPos );
569 }
570
571 // If there is a trailing slash in the file name, we have to strip it,
572 // because that's what ZIP_GetEntry() does.
573 if ( preg_grep( '!\.class/?$!', $names ) ) {
574 $this->mJavaDetected = true;
575 }
576 }
577
578 /**
579 * Alias for verifyTitlePermissions. The function was originally
580 * 'verifyPermissions', but that suggests it's checking the user, when it's
581 * really checking the title + user combination.
582 *
583 * @param User $user User object to verify the permissions against
584 * @return mixed An array as returned by getUserPermissionsErrors or true
585 * in case the user has proper permissions.
586 */
587 public function verifyPermissions( $user ) {
588 return $this->verifyTitlePermissions( $user );
589 }
590
591 /**
592 * Check whether the user can edit, upload and create the image. This
593 * checks only against the current title; if it returns errors, it may
594 * very well be that another title will not give errors. Therefore
595 * isAllowed() should be called as well for generic is-user-blocked or
596 * can-user-upload checking.
597 *
598 * @param User $user User object to verify the permissions against
599 * @return mixed An array as returned by getUserPermissionsErrors or true
600 * in case the user has proper permissions.
601 */
602 public function verifyTitlePermissions( $user ) {
603 /**
604 * If the image is protected, non-sysop users won't be able
605 * to modify it by uploading a new revision.
606 */
607 $nt = $this->getTitle();
608 if ( is_null( $nt ) ) {
609 return true;
610 }
611 $permErrors = $nt->getUserPermissionsErrors( 'edit', $user );
612 $permErrorsUpload = $nt->getUserPermissionsErrors( 'upload', $user );
613 if ( !$nt->exists() ) {
614 $permErrorsCreate = $nt->getUserPermissionsErrors( 'create', $user );
615 } else {
616 $permErrorsCreate = array();
617 }
618 if ( $permErrors || $permErrorsUpload || $permErrorsCreate ) {
619 $permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsUpload, $permErrors ) );
620 $permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsCreate, $permErrors ) );
621
622 return $permErrors;
623 }
624
625 $overwriteError = $this->checkOverwrite( $user );
626 if ( $overwriteError !== true ) {
627 return array( $overwriteError );
628 }
629
630 return true;
631 }
632
633 /**
634 * Check for non fatal problems with the file.
635 *
636 * This should not assume that mTempPath is set.
637 *
638 * @return array Array of warnings
639 */
640 public function checkWarnings() {
641 global $wgLang;
642 wfProfileIn( __METHOD__ );
643
644 $warnings = array();
645
646 $localFile = $this->getLocalFile();
647 $filename = $localFile->getName();
648
649 /**
650 * Check whether the resulting filename is different from the desired one,
651 * but ignore things like ucfirst() and spaces/underscore things
652 */
653 $comparableName = str_replace( ' ', '_', $this->mDesiredDestName );
654 $comparableName = Title::capitalize( $comparableName, NS_FILE );
655
656 if ( $this->mDesiredDestName != $filename && $comparableName != $filename ) {
657 $warnings['badfilename'] = $filename;
658 // Debugging for bug 62241
659 wfDebugLog( 'upload', "Filename: '$filename', mDesiredDestName: "
660 . "'$this->mDesiredDestName', comparableName: '$comparableName'" );
661 }
662
663 // Check whether the file extension is on the unwanted list
664 global $wgCheckFileExtensions, $wgFileExtensions;
665 if ( $wgCheckFileExtensions ) {
666 $extensions = array_unique( $wgFileExtensions );
667 if ( !$this->checkFileExtension( $this->mFinalExtension, $extensions ) ) {
668 $warnings['filetype-unwanted-type'] = array( $this->mFinalExtension,
669 $wgLang->commaList( $extensions ), count( $extensions ) );
670 }
671 }
672
673 global $wgUploadSizeWarning;
674 if ( $wgUploadSizeWarning && ( $this->mFileSize > $wgUploadSizeWarning ) ) {
675 $warnings['large-file'] = array( $wgUploadSizeWarning, $this->mFileSize );
676 }
677
678 if ( $this->mFileSize == 0 ) {
679 $warnings['emptyfile'] = true;
680 }
681
682 $exists = self::getExistsWarning( $localFile );
683 if ( $exists !== false ) {
684 $warnings['exists'] = $exists;
685 }
686
687 // Check dupes against existing files
688 $hash = $this->getTempFileSha1Base36();
689 $dupes = RepoGroup::singleton()->findBySha1( $hash );
690 $title = $this->getTitle();
691 // Remove all matches against self
692 foreach ( $dupes as $key => $dupe ) {
693 if ( $title->equals( $dupe->getTitle() ) ) {
694 unset( $dupes[$key] );
695 }
696 }
697 if ( $dupes ) {
698 $warnings['duplicate'] = $dupes;
699 }
700
701 // Check dupes against archives
702 $archivedImage = new ArchivedFile( null, 0, "{$hash}.{$this->mFinalExtension}" );
703 if ( $archivedImage->getID() > 0 ) {
704 if ( $archivedImage->userCan( File::DELETED_FILE ) ) {
705 $warnings['duplicate-archive'] = $archivedImage->getName();
706 } else {
707 $warnings['duplicate-archive'] = '';
708 }
709 }
710
711 wfProfileOut( __METHOD__ );
712
713 return $warnings;
714 }
715
716 /**
717 * Really perform the upload. Stores the file in the local repo, watches
718 * if necessary and runs the UploadComplete hook.
719 *
720 * @param string $comment
721 * @param string $pageText
722 * @param bool $watch
723 * @param User $user
724 *
725 * @return Status Indicating the whether the upload succeeded.
726 */
727 public function performUpload( $comment, $pageText, $watch, $user ) {
728 wfProfileIn( __METHOD__ );
729
730 $status = $this->getLocalFile()->upload(
731 $this->mTempPath,
732 $comment,
733 $pageText,
734 File::DELETE_SOURCE,
735 $this->mFileProps,
736 false,
737 $user
738 );
739
740 if ( $status->isGood() ) {
741 if ( $watch ) {
742 WatchAction::doWatch(
743 $this->getLocalFile()->getTitle(),
744 $user,
745 WatchedItem::IGNORE_USER_RIGHTS
746 );
747 }
748 wfRunHooks( 'UploadComplete', array( &$this ) );
749
750 $this->postProcessUpload();
751 }
752
753 wfProfileOut( __METHOD__ );
754
755 return $status;
756 }
757
758 /**
759 * Perform extra steps after a successful upload.
760 */
761 public function postProcessUpload() {
762 global $wgUploadThumbnailRenderMap;
763
764 $jobs = array();
765
766 $sizes = $wgUploadThumbnailRenderMap;
767 rsort( $sizes );
768
769 foreach ( $sizes as $size ) {
770 $jobs []= new ThumbnailRenderJob( $this->getLocalFile()->getTitle(), array(
771 'transformParams' => array( 'width' => $size ),
772 ) );
773 }
774
775 JobQueueGroup::singleton()->push( $jobs );
776 }
777
778 /**
779 * Returns the title of the file to be uploaded. Sets mTitleError in case
780 * the name was illegal.
781 *
782 * @return Title The title of the file or null in case the name was illegal
783 */
784 public function getTitle() {
785 if ( $this->mTitle !== false ) {
786 return $this->mTitle;
787 }
788 /* Assume that if a user specified File:Something.jpg, this is an error
789 * and that the namespace prefix needs to be stripped of.
790 */
791 $title = Title::newFromText( $this->mDesiredDestName );
792 if ( $title && $title->getNamespace() == NS_FILE ) {
793 $this->mFilteredName = $title->getDBkey();
794 } else {
795 $this->mFilteredName = $this->mDesiredDestName;
796 }
797
798 # oi_archive_name is max 255 bytes, which include a timestamp and an
799 # exclamation mark, so restrict file name to 240 bytes.
800 if ( strlen( $this->mFilteredName ) > 240 ) {
801 $this->mTitleError = self::FILENAME_TOO_LONG;
802 $this->mTitle = null;
803
804 return $this->mTitle;
805 }
806
807 /**
808 * Chop off any directories in the given filename. Then
809 * filter out illegal characters, and try to make a legible name
810 * out of it. We'll strip some silently that Title would die on.
811 */
812 $this->mFilteredName = wfStripIllegalFilenameChars( $this->mFilteredName );
813 /* Normalize to title form before we do any further processing */
814 $nt = Title::makeTitleSafe( NS_FILE, $this->mFilteredName );
815 if ( is_null( $nt ) ) {
816 $this->mTitleError = self::ILLEGAL_FILENAME;
817 $this->mTitle = null;
818
819 return $this->mTitle;
820 }
821 $this->mFilteredName = $nt->getDBkey();
822
823 /**
824 * We'll want to blacklist against *any* 'extension', and use
825 * only the final one for the whitelist.
826 */
827 list( $partname, $ext ) = $this->splitExtensions( $this->mFilteredName );
828
829 if ( count( $ext ) ) {
830 $this->mFinalExtension = trim( $ext[count( $ext ) - 1] );
831 } else {
832 $this->mFinalExtension = '';
833
834 # No extension, try guessing one
835 $magic = MimeMagic::singleton();
836 $mime = $magic->guessMimeType( $this->mTempPath );
837 if ( $mime !== 'unknown/unknown' ) {
838 # Get a space separated list of extensions
839 $extList = $magic->getExtensionsForType( $mime );
840 if ( $extList ) {
841 # Set the extension to the canonical extension
842 $this->mFinalExtension = strtok( $extList, ' ' );
843
844 # Fix up the other variables
845 $this->mFilteredName .= ".{$this->mFinalExtension}";
846 $nt = Title::makeTitleSafe( NS_FILE, $this->mFilteredName );
847 $ext = array( $this->mFinalExtension );
848 }
849 }
850 }
851
852 /* Don't allow users to override the blacklist (check file extension) */
853 global $wgCheckFileExtensions, $wgStrictFileExtensions;
854 global $wgFileExtensions, $wgFileBlacklist;
855
856 $blackListedExtensions = $this->checkFileExtensionList( $ext, $wgFileBlacklist );
857
858 if ( $this->mFinalExtension == '' ) {
859 $this->mTitleError = self::FILETYPE_MISSING;
860 $this->mTitle = null;
861
862 return $this->mTitle;
863 } elseif ( $blackListedExtensions ||
864 ( $wgCheckFileExtensions && $wgStrictFileExtensions &&
865 !$this->checkFileExtension( $this->mFinalExtension, $wgFileExtensions ) )
866 ) {
867 $this->mBlackListedExtensions = $blackListedExtensions;
868 $this->mTitleError = self::FILETYPE_BADTYPE;
869 $this->mTitle = null;
870
871 return $this->mTitle;
872 }
873
874 // Windows may be broken with special characters, see bug 1780
875 if ( !preg_match( '/^[\x0-\x7f]*$/', $nt->getText() )
876 && !RepoGroup::singleton()->getLocalRepo()->backendSupportsUnicodePaths()
877 ) {
878 $this->mTitleError = self::WINDOWS_NONASCII_FILENAME;
879 $this->mTitle = null;
880
881 return $this->mTitle;
882 }
883
884 # If there was more than one "extension", reassemble the base
885 # filename to prevent bogus complaints about length
886 if ( count( $ext ) > 1 ) {
887 $iterations = count( $ext ) - 1;
888 for ( $i = 0; $i < $iterations; $i++ ) {
889 $partname .= '.' . $ext[$i];
890 }
891 }
892
893 if ( strlen( $partname ) < 1 ) {
894 $this->mTitleError = self::MIN_LENGTH_PARTNAME;
895 $this->mTitle = null;
896
897 return $this->mTitle;
898 }
899
900 $this->mTitle = $nt;
901
902 return $this->mTitle;
903 }
904
905 /**
906 * Return the local file and initializes if necessary.
907 *
908 * @return LocalFile|UploadStashFile|null
909 */
910 public function getLocalFile() {
911 if ( is_null( $this->mLocalFile ) ) {
912 $nt = $this->getTitle();
913 $this->mLocalFile = is_null( $nt ) ? null : wfLocalFile( $nt );
914 }
915
916 return $this->mLocalFile;
917 }
918
919 /**
920 * If the user does not supply all necessary information in the first upload
921 * form submission (either by accident or by design) then we may want to
922 * stash the file temporarily, get more information, and publish the file
923 * later.
924 *
925 * This method will stash a file in a temporary directory for later
926 * processing, and save the necessary descriptive info into the database.
927 * This method returns the file object, which also has a 'fileKey' property
928 * which can be passed through a form or API request to find this stashed
929 * file again.
930 *
931 * @param User $user
932 * @return UploadStashFile Stashed file
933 */
934 public function stashFile( User $user = null ) {
935 // was stashSessionFile
936 wfProfileIn( __METHOD__ );
937
938 $stash = RepoGroup::singleton()->getLocalRepo()->getUploadStash( $user );
939 $file = $stash->stashFile( $this->mTempPath, $this->getSourceType() );
940 $this->mLocalFile = $file;
941
942 wfProfileOut( __METHOD__ );
943
944 return $file;
945 }
946
947 /**
948 * Stash a file in a temporary directory, returning a key which can be used
949 * to find the file again. See stashFile().
950 *
951 * @return string File key
952 */
953 public function stashFileGetKey() {
954 return $this->stashFile()->getFileKey();
955 }
956
957 /**
958 * alias for stashFileGetKey, for backwards compatibility
959 *
960 * @return string File key
961 */
962 public function stashSession() {
963 return $this->stashFileGetKey();
964 }
965
966 /**
967 * If we've modified the upload file we need to manually remove it
968 * on exit to clean up.
969 */
970 public function cleanupTempFile() {
971 if ( $this->mRemoveTempFile && $this->mTempPath && file_exists( $this->mTempPath ) ) {
972 wfDebug( __METHOD__ . ": Removing temporary file {$this->mTempPath}\n" );
973 unlink( $this->mTempPath );
974 }
975 }
976
977 public function getTempPath() {
978 return $this->mTempPath;
979 }
980
981 /**
982 * Split a file into a base name and all dot-delimited 'extensions'
983 * on the end. Some web server configurations will fall back to
984 * earlier pseudo-'extensions' to determine type and execute
985 * scripts, so the blacklist needs to check them all.
986 *
987 * @param string $filename
988 * @return array
989 */
990 public static function splitExtensions( $filename ) {
991 $bits = explode( '.', $filename );
992 $basename = array_shift( $bits );
993
994 return array( $basename, $bits );
995 }
996
997 /**
998 * Perform case-insensitive match against a list of file extensions.
999 * Returns true if the extension is in the list.
1000 *
1001 * @param string $ext
1002 * @param array $list
1003 * @return bool
1004 */
1005 public static function checkFileExtension( $ext, $list ) {
1006 return in_array( strtolower( $ext ), $list );
1007 }
1008
1009 /**
1010 * Perform case-insensitive match against a list of file extensions.
1011 * Returns an array of matching extensions.
1012 *
1013 * @param array $ext
1014 * @param array $list
1015 * @return bool
1016 */
1017 public static function checkFileExtensionList( $ext, $list ) {
1018 return array_intersect( array_map( 'strtolower', $ext ), $list );
1019 }
1020
1021 /**
1022 * Checks if the MIME type of the uploaded file matches the file extension.
1023 *
1024 * @param string $mime The MIME type of the uploaded file
1025 * @param string $extension The filename extension that the file is to be served with
1026 * @return bool
1027 */
1028 public static function verifyExtension( $mime, $extension ) {
1029 $magic = MimeMagic::singleton();
1030
1031 if ( !$mime || $mime == 'unknown' || $mime == 'unknown/unknown' ) {
1032 if ( !$magic->isRecognizableExtension( $extension ) ) {
1033 wfDebug( __METHOD__ . ": passing file with unknown detected mime type; " .
1034 "unrecognized extension '$extension', can't verify\n" );
1035
1036 return true;
1037 } else {
1038 wfDebug( __METHOD__ . ": rejecting file with unknown detected mime type; " .
1039 "recognized extension '$extension', so probably invalid file\n" );
1040
1041 return false;
1042 }
1043 }
1044
1045 $match = $magic->isMatchingExtension( $extension, $mime );
1046
1047 if ( $match === null ) {
1048 if ( $magic->getTypesForExtension( $extension ) !== null ) {
1049 wfDebug( __METHOD__ . ": No extension known for $mime, but we know a mime for $extension\n" );
1050
1051 return false;
1052 } else {
1053 wfDebug( __METHOD__ . ": no file extension known for mime type $mime, passing file\n" );
1054
1055 return true;
1056 }
1057 } elseif ( $match === true ) {
1058 wfDebug( __METHOD__ . ": mime type $mime matches extension $extension, passing file\n" );
1059
1060 /** @todo If it's a bitmap, make sure PHP or ImageMagick resp. can handle it! */
1061 return true;
1062 } else {
1063 wfDebug( __METHOD__
1064 . ": mime type $mime mismatches file extension $extension, rejecting file\n" );
1065
1066 return false;
1067 }
1068 }
1069
1070 /**
1071 * Heuristic for detecting files that *could* contain JavaScript instructions or
1072 * things that may look like HTML to a browser and are thus
1073 * potentially harmful. The present implementation will produce false
1074 * positives in some situations.
1075 *
1076 * @param string $file Pathname to the temporary upload file
1077 * @param string $mime The MIME type of the file
1078 * @param string $extension The extension of the file
1079 * @return bool True if the file contains something looking like embedded scripts
1080 */
1081 public static function detectScript( $file, $mime, $extension ) {
1082 global $wgAllowTitlesInSVG;
1083 wfProfileIn( __METHOD__ );
1084
1085 # ugly hack: for text files, always look at the entire file.
1086 # For binary field, just check the first K.
1087
1088 if ( strpos( $mime, 'text/' ) === 0 ) {
1089 $chunk = file_get_contents( $file );
1090 } else {
1091 $fp = fopen( $file, 'rb' );
1092 $chunk = fread( $fp, 1024 );
1093 fclose( $fp );
1094 }
1095
1096 $chunk = strtolower( $chunk );
1097
1098 if ( !$chunk ) {
1099 wfProfileOut( __METHOD__ );
1100
1101 return false;
1102 }
1103
1104 # decode from UTF-16 if needed (could be used for obfuscation).
1105 if ( substr( $chunk, 0, 2 ) == "\xfe\xff" ) {
1106 $enc = 'UTF-16BE';
1107 } elseif ( substr( $chunk, 0, 2 ) == "\xff\xfe" ) {
1108 $enc = 'UTF-16LE';
1109 } else {
1110 $enc = null;
1111 }
1112
1113 if ( $enc ) {
1114 $chunk = iconv( $enc, "ASCII//IGNORE", $chunk );
1115 }
1116
1117 $chunk = trim( $chunk );
1118
1119 /** @todo FIXME: Convert from UTF-16 if necessary! */
1120 wfDebug( __METHOD__ . ": checking for embedded scripts and HTML stuff\n" );
1121
1122 # check for HTML doctype
1123 if ( preg_match( "/<!DOCTYPE *X?HTML/i", $chunk ) ) {
1124 wfProfileOut( __METHOD__ );
1125
1126 return true;
1127 }
1128
1129 // Some browsers will interpret obscure xml encodings as UTF-8, while
1130 // PHP/expat will interpret the given encoding in the xml declaration (bug 47304)
1131 if ( $extension == 'svg' || strpos( $mime, 'image/svg' ) === 0 ) {
1132 if ( self::checkXMLEncodingMissmatch( $file ) ) {
1133 wfProfileOut( __METHOD__ );
1134
1135 return true;
1136 }
1137 }
1138
1139 /**
1140 * Internet Explorer for Windows performs some really stupid file type
1141 * autodetection which can cause it to interpret valid image files as HTML
1142 * and potentially execute JavaScript, creating a cross-site scripting
1143 * attack vectors.
1144 *
1145 * Apple's Safari browser also performs some unsafe file type autodetection
1146 * which can cause legitimate files to be interpreted as HTML if the
1147 * web server is not correctly configured to send the right content-type
1148 * (or if you're really uploading plain text and octet streams!)
1149 *
1150 * Returns true if IE is likely to mistake the given file for HTML.
1151 * Also returns true if Safari would mistake the given file for HTML
1152 * when served with a generic content-type.
1153 */
1154 $tags = array(
1155 '<a href',
1156 '<body',
1157 '<head',
1158 '<html', #also in safari
1159 '<img',
1160 '<pre',
1161 '<script', #also in safari
1162 '<table'
1163 );
1164
1165 if ( !$wgAllowTitlesInSVG && $extension !== 'svg' && $mime !== 'image/svg' ) {
1166 $tags[] = '<title';
1167 }
1168
1169 foreach ( $tags as $tag ) {
1170 if ( false !== strpos( $chunk, $tag ) ) {
1171 wfDebug( __METHOD__ . ": found something that may make it be mistaken for html: $tag\n" );
1172 wfProfileOut( __METHOD__ );
1173
1174 return true;
1175 }
1176 }
1177
1178 /*
1179 * look for JavaScript
1180 */
1181
1182 # resolve entity-refs to look at attributes. may be harsh on big files... cache result?
1183 $chunk = Sanitizer::decodeCharReferences( $chunk );
1184
1185 # look for script-types
1186 if ( preg_match( '!type\s*=\s*[\'"]?\s*(?:\w*/)?(?:ecma|java)!sim', $chunk ) ) {
1187 wfDebug( __METHOD__ . ": found script types\n" );
1188 wfProfileOut( __METHOD__ );
1189
1190 return true;
1191 }
1192
1193 # look for html-style script-urls
1194 if ( preg_match( '!(?:href|src|data)\s*=\s*[\'"]?\s*(?:ecma|java)script:!sim', $chunk ) ) {
1195 wfDebug( __METHOD__ . ": found html-style script urls\n" );
1196 wfProfileOut( __METHOD__ );
1197
1198 return true;
1199 }
1200
1201 # look for css-style script-urls
1202 if ( preg_match( '!url\s*\(\s*[\'"]?\s*(?:ecma|java)script:!sim', $chunk ) ) {
1203 wfDebug( __METHOD__ . ": found css-style script urls\n" );
1204 wfProfileOut( __METHOD__ );
1205
1206 return true;
1207 }
1208
1209 wfDebug( __METHOD__ . ": no scripts found\n" );
1210 wfProfileOut( __METHOD__ );
1211
1212 return false;
1213 }
1214
1215 /**
1216 * Check a whitelist of xml encodings that are known not to be interpreted differently
1217 * by the server's xml parser (expat) and some common browsers.
1218 *
1219 * @param string $file Pathname to the temporary upload file
1220 * @return bool True if the file contains an encoding that could be misinterpreted
1221 */
1222 public static function checkXMLEncodingMissmatch( $file ) {
1223 global $wgSVGMetadataCutoff;
1224 $contents = file_get_contents( $file, false, null, -1, $wgSVGMetadataCutoff );
1225 $encodingRegex = '!encoding[ \t\n\r]*=[ \t\n\r]*[\'"](.*?)[\'"]!si';
1226
1227 if ( preg_match( "!<\?xml\b(.*?)\?>!si", $contents, $matches ) ) {
1228 if ( preg_match( $encodingRegex, $matches[1], $encMatch )
1229 && !in_array( strtoupper( $encMatch[1] ), self::$safeXmlEncodings )
1230 ) {
1231 wfDebug( __METHOD__ . ": Found unsafe XML encoding '{$encMatch[1]}'\n" );
1232
1233 return true;
1234 }
1235 } elseif ( preg_match( "!<\?xml\b!si", $contents ) ) {
1236 // Start of XML declaration without an end in the first $wgSVGMetadataCutoff
1237 // bytes. There shouldn't be a legitimate reason for this to happen.
1238 wfDebug( __METHOD__ . ": Unmatched XML declaration start\n" );
1239
1240 return true;
1241 } elseif ( substr( $contents, 0, 4 ) == "\x4C\x6F\xA7\x94" ) {
1242 // EBCDIC encoded XML
1243 wfDebug( __METHOD__ . ": EBCDIC Encoded XML\n" );
1244
1245 return true;
1246 }
1247
1248 // It's possible the file is encoded with multi-byte encoding, so re-encode attempt to
1249 // detect the encoding in case is specifies an encoding not whitelisted in self::$safeXmlEncodings
1250 $attemptEncodings = array( 'UTF-16', 'UTF-16BE', 'UTF-32', 'UTF-32BE' );
1251 foreach ( $attemptEncodings as $encoding ) {
1252 wfSuppressWarnings();
1253 $str = iconv( $encoding, 'UTF-8', $contents );
1254 wfRestoreWarnings();
1255 if ( $str != '' && preg_match( "!<\?xml\b(.*?)\?>!si", $str, $matches ) ) {
1256 if ( preg_match( $encodingRegex, $matches[1], $encMatch )
1257 && !in_array( strtoupper( $encMatch[1] ), self::$safeXmlEncodings )
1258 ) {
1259 wfDebug( __METHOD__ . ": Found unsafe XML encoding '{$encMatch[1]}'\n" );
1260
1261 return true;
1262 }
1263 } elseif ( $str != '' && preg_match( "!<\?xml\b!si", $str ) ) {
1264 // Start of XML declaration without an end in the first $wgSVGMetadataCutoff
1265 // bytes. There shouldn't be a legitimate reason for this to happen.
1266 wfDebug( __METHOD__ . ": Unmatched XML declaration start\n" );
1267
1268 return true;
1269 }
1270 }
1271
1272 return false;
1273 }
1274
1275 /**
1276 * @param string $filename
1277 * @return mixed False of the file is verified (does not contain scripts), array otherwise.
1278 */
1279 protected function detectScriptInSvg( $filename ) {
1280 $this->mSVGNSError = false;
1281 $check = new XmlTypeCheck(
1282 $filename,
1283 array( $this, 'checkSvgScriptCallback' ),
1284 true,
1285 array( 'processing_instruction_handler' => 'UploadBase::checkSvgPICallback' )
1286 );
1287 if ( $check->wellFormed !== true ) {
1288 // Invalid xml (bug 58553)
1289 return array( 'uploadinvalidxml' );
1290 } elseif ( $check->filterMatch ) {
1291 if ( $this->mSVGNSError ) {
1292 return array( 'uploadscriptednamespace', $this->mSVGNSError );
1293 }
1294
1295 return array( 'uploadscripted' );
1296 }
1297
1298 return false;
1299 }
1300
1301 /**
1302 * Callback to filter SVG Processing Instructions.
1303 * @param string $target Processing instruction name
1304 * @param string $data Processing instruction attribute and value
1305 * @return bool (true if the filter identified something bad)
1306 */
1307 public static function checkSvgPICallback( $target, $data ) {
1308 // Don't allow external stylesheets (bug 57550)
1309 if ( preg_match( '/xml-stylesheet/i', $target ) ) {
1310 return true;
1311 }
1312
1313 return false;
1314 }
1315
1316 /**
1317 * @todo Replace this with a whitelist filter!
1318 * @param string $element
1319 * @param array $attribs
1320 * @return bool
1321 */
1322 public function checkSvgScriptCallback( $element, $attribs ) {
1323 list( $namespace, $strippedElement ) = $this->splitXmlNamespace( $element );
1324
1325 // We specifically don't include:
1326 // http://www.w3.org/1999/xhtml (bug 60771)
1327 static $validNamespaces = array(
1328 '',
1329 'adobe:ns:meta/',
1330 'http://creativecommons.org/ns#',
1331 'http://inkscape.sourceforge.net/dtd/sodipodi-0.dtd',
1332 'http://ns.adobe.com/adobeillustrator/10.0/',
1333 'http://ns.adobe.com/adobesvgviewerextensions/3.0/',
1334 'http://ns.adobe.com/extensibility/1.0/',
1335 'http://ns.adobe.com/flows/1.0/',
1336 'http://ns.adobe.com/illustrator/1.0/',
1337 'http://ns.adobe.com/imagereplacement/1.0/',
1338 'http://ns.adobe.com/pdf/1.3/',
1339 'http://ns.adobe.com/photoshop/1.0/',
1340 'http://ns.adobe.com/saveforweb/1.0/',
1341 'http://ns.adobe.com/variables/1.0/',
1342 'http://ns.adobe.com/xap/1.0/',
1343 'http://ns.adobe.com/xap/1.0/g/',
1344 'http://ns.adobe.com/xap/1.0/g/img/',
1345 'http://ns.adobe.com/xap/1.0/mm/',
1346 'http://ns.adobe.com/xap/1.0/rights/',
1347 'http://ns.adobe.com/xap/1.0/stype/dimensions#',
1348 'http://ns.adobe.com/xap/1.0/stype/font#',
1349 'http://ns.adobe.com/xap/1.0/stype/manifestitem#',
1350 'http://ns.adobe.com/xap/1.0/stype/resourceevent#',
1351 'http://ns.adobe.com/xap/1.0/stype/resourceref#',
1352 'http://ns.adobe.com/xap/1.0/t/pg/',
1353 'http://purl.org/dc/elements/1.1/',
1354 'http://purl.org/dc/elements/1.1',
1355 'http://schemas.microsoft.com/visio/2003/svgextensions/',
1356 'http://sodipodi.sourceforge.net/dtd/sodipodi-0.dtd',
1357 'http://taptrix.com/inkpad/svg_extensions',
1358 'http://web.resource.org/cc/',
1359 'http://www.freesoftware.fsf.org/bkchem/cdml',
1360 'http://www.inkscape.org/namespaces/inkscape',
1361 'http://www.opengis.net/gml',
1362 'http://www.w3.org/1999/02/22-rdf-syntax-ns#',
1363 'http://www.w3.org/2000/svg',
1364 'http://www.w3.org/tr/rec-rdf-syntax/',
1365 );
1366
1367 if ( !in_array( $namespace, $validNamespaces ) ) {
1368 wfDebug( __METHOD__ . ": Non-svg namespace '$namespace' in uploaded file.\n" );
1369 /** @todo Return a status object to a closure in XmlTypeCheck, for MW1.21+ */
1370 $this->mSVGNSError = $namespace;
1371
1372 return true;
1373 }
1374
1375 /*
1376 * check for elements that can contain javascript
1377 */
1378 if ( $strippedElement == 'script' ) {
1379 wfDebug( __METHOD__ . ": Found script element '$element' in uploaded file.\n" );
1380
1381 return true;
1382 }
1383
1384 # e.g., <svg xmlns="http://www.w3.org/2000/svg">
1385 # <handler xmlns:ev="http://www.w3.org/2001/xml-events" ev:event="load">alert(1)</handler> </svg>
1386 if ( $strippedElement == 'handler' ) {
1387 wfDebug( __METHOD__ . ": Found scriptable element '$element' in uploaded file.\n" );
1388
1389 return true;
1390 }
1391
1392 # SVG reported in Feb '12 that used xml:stylesheet to generate javascript block
1393 if ( $strippedElement == 'stylesheet' ) {
1394 wfDebug( __METHOD__ . ": Found scriptable element '$element' in uploaded file.\n" );
1395
1396 return true;
1397 }
1398
1399 # Block iframes, in case they pass the namespace check
1400 if ( $strippedElement == 'iframe' ) {
1401 wfDebug( __METHOD__ . ": iframe in uploaded file.\n" );
1402
1403 return true;
1404 }
1405
1406 foreach ( $attribs as $attrib => $value ) {
1407 $stripped = $this->stripXmlNamespace( $attrib );
1408 $value = strtolower( $value );
1409
1410 if ( substr( $stripped, 0, 2 ) == 'on' ) {
1411 wfDebug( __METHOD__
1412 . ": Found event-handler attribute '$attrib'='$value' in uploaded file.\n" );
1413
1414 return true;
1415 }
1416
1417 # href with non-local target (don't allow http://, javascript:, etc)
1418 if ( $stripped == 'href'
1419 && strpos( $value, 'data:' ) !== 0
1420 && strpos( $value, '#' ) !== 0
1421 ) {
1422 if ( !( $strippedElement === 'a'
1423 && preg_match( '!^https?://!im', $value ) )
1424 ) {
1425 wfDebug( __METHOD__ . ": Found href attribute <$strippedElement "
1426 . "'$attrib'='$value' in uploaded file.\n" );
1427
1428 return true;
1429 }
1430 }
1431
1432 # href with embedded svg as target
1433 if ( $stripped == 'href' && preg_match( '!data:[^,]*image/svg[^,]*,!sim', $value ) ) {
1434 wfDebug( __METHOD__ . ": Found href to embedded svg "
1435 . "\"<$strippedElement '$attrib'='$value'...\" in uploaded file.\n" );
1436
1437 return true;
1438 }
1439
1440 # href with embedded (text/xml) svg as target
1441 if ( $stripped == 'href' && preg_match( '!data:[^,]*text/xml[^,]*,!sim', $value ) ) {
1442 wfDebug( __METHOD__ . ": Found href to embedded svg "
1443 . "\"<$strippedElement '$attrib'='$value'...\" in uploaded file.\n" );
1444
1445 return true;
1446 }
1447
1448 # use set/animate to add event-handler attribute to parent
1449 if ( ( $strippedElement == 'set' || $strippedElement == 'animate' )
1450 && $stripped == 'attributename'
1451 && substr( $value, 0, 2 ) == 'on'
1452 ) {
1453 wfDebug( __METHOD__ . ": Found svg setting event-handler attribute with "
1454 . "\"<$strippedElement $stripped='$value'...\" in uploaded file.\n" );
1455
1456 return true;
1457 }
1458
1459 # use set to add href attribute to parent element
1460 if ( $strippedElement == 'set'
1461 && $stripped == 'attributename'
1462 && strpos( $value, 'href' ) !== false
1463 ) {
1464 wfDebug( __METHOD__ . ": Found svg setting href attribute '$value' in uploaded file.\n" );
1465
1466 return true;
1467 }
1468
1469 # use set to add a remote / data / script target to an element
1470 if ( $strippedElement == 'set'
1471 && $stripped == 'to'
1472 && preg_match( '!(http|https|data|script):!sim', $value )
1473 ) {
1474 wfDebug( __METHOD__ . ": Found svg setting attribute to '$value' in uploaded file.\n" );
1475
1476 return true;
1477 }
1478
1479 # use handler attribute with remote / data / script
1480 if ( $stripped == 'handler' && preg_match( '!(http|https|data|script):!sim', $value ) ) {
1481 wfDebug( __METHOD__ . ": Found svg setting handler with remote/data/script "
1482 . "'$attrib'='$value' in uploaded file.\n" );
1483
1484 return true;
1485 }
1486
1487 # use CSS styles to bring in remote code
1488 # catch url("http:..., url('http:..., url(http:..., but not url("#..., url('#..., url(#....
1489 $tagsList = "font|clip-path|fill|filter|marker|marker-end|marker-mid|marker-start|mask|stroke";
1490 if ( $stripped == 'style'
1491 && preg_match_all(
1492 '!((?:' . $tagsList . ')\s*:\s*url\s*\(\s*["\']?\s*[^#]+.*?\))!sim',
1493 $value,
1494 $matches
1495 )
1496 ) {
1497 foreach ( $matches[1] as $match ) {
1498 if ( !preg_match( '!(?:' . $tagsList . ')\s*:\s*url\s*\(\s*(#|\'#|"#)!sim', $match ) ) {
1499 wfDebug( __METHOD__ . ": Found svg setting a style with "
1500 . "remote url '$attrib'='$value' in uploaded file.\n" );
1501
1502 return true;
1503 }
1504 }
1505 }
1506
1507 # image filters can pull in url, which could be svg that executes scripts
1508 if ( $strippedElement == 'image'
1509 && $stripped == 'filter'
1510 && preg_match( '!url\s*\(!sim', $value )
1511 ) {
1512 wfDebug( __METHOD__ . ": Found image filter with url: "
1513 . "\"<$strippedElement $stripped='$value'...\" in uploaded file.\n" );
1514
1515 return true;
1516 }
1517 }
1518
1519 return false; //No scripts detected
1520 }
1521
1522 /**
1523 * Divide the element name passed by the xml parser to the callback into URI and prifix.
1524 * @param string $element
1525 * @return array Containing the namespace URI and prefix
1526 */
1527 private static function splitXmlNamespace( $element ) {
1528 // 'http://www.w3.org/2000/svg:script' -> array( 'http://www.w3.org/2000/svg', 'script' )
1529 $parts = explode( ':', strtolower( $element ) );
1530 $name = array_pop( $parts );
1531 $ns = implode( ':', $parts );
1532
1533 return array( $ns, $name );
1534 }
1535
1536 /**
1537 * @param string $name
1538 * @return string
1539 */
1540 private function stripXmlNamespace( $name ) {
1541 // 'http://www.w3.org/2000/svg:script' -> 'script'
1542 $parts = explode( ':', strtolower( $name ) );
1543
1544 return array_pop( $parts );
1545 }
1546
1547 /**
1548 * Generic wrapper function for a virus scanner program.
1549 * This relies on the $wgAntivirus and $wgAntivirusSetup variables.
1550 * $wgAntivirusRequired may be used to deny upload if the scan fails.
1551 *
1552 * @param string $file Pathname to the temporary upload file
1553 * @return mixed False if not virus is found, null if the scan fails or is disabled,
1554 * or a string containing feedback from the virus scanner if a virus was found.
1555 * If textual feedback is missing but a virus was found, this function returns true.
1556 */
1557 public static function detectVirus( $file ) {
1558 global $wgAntivirus, $wgAntivirusSetup, $wgAntivirusRequired, $wgOut;
1559 wfProfileIn( __METHOD__ );
1560
1561 if ( !$wgAntivirus ) {
1562 wfDebug( __METHOD__ . ": virus scanner disabled\n" );
1563 wfProfileOut( __METHOD__ );
1564
1565 return null;
1566 }
1567
1568 if ( !$wgAntivirusSetup[$wgAntivirus] ) {
1569 wfDebug( __METHOD__ . ": unknown virus scanner: $wgAntivirus\n" );
1570 $wgOut->wrapWikiMsg( "<div class=\"error\">\n$1\n</div>",
1571 array( 'virus-badscanner', $wgAntivirus ) );
1572 wfProfileOut( __METHOD__ );
1573
1574 return wfMessage( 'virus-unknownscanner' )->text() . " $wgAntivirus";
1575 }
1576
1577 # look up scanner configuration
1578 $command = $wgAntivirusSetup[$wgAntivirus]['command'];
1579 $exitCodeMap = $wgAntivirusSetup[$wgAntivirus]['codemap'];
1580 $msgPattern = isset( $wgAntivirusSetup[$wgAntivirus]['messagepattern'] ) ?
1581 $wgAntivirusSetup[$wgAntivirus]['messagepattern'] : null;
1582
1583 if ( strpos( $command, "%f" ) === false ) {
1584 # simple pattern: append file to scan
1585 $command .= " " . wfEscapeShellArg( $file );
1586 } else {
1587 # complex pattern: replace "%f" with file to scan
1588 $command = str_replace( "%f", wfEscapeShellArg( $file ), $command );
1589 }
1590
1591 wfDebug( __METHOD__ . ": running virus scan: $command \n" );
1592
1593 # execute virus scanner
1594 $exitCode = false;
1595
1596 # NOTE: there's a 50 line workaround to make stderr redirection work on windows, too.
1597 # that does not seem to be worth the pain.
1598 # Ask me (Duesentrieb) about it if it's ever needed.
1599 $output = wfShellExecWithStderr( $command, $exitCode );
1600
1601 # map exit code to AV_xxx constants.
1602 $mappedCode = $exitCode;
1603 if ( $exitCodeMap ) {
1604 if ( isset( $exitCodeMap[$exitCode] ) ) {
1605 $mappedCode = $exitCodeMap[$exitCode];
1606 } elseif ( isset( $exitCodeMap["*"] ) ) {
1607 $mappedCode = $exitCodeMap["*"];
1608 }
1609 }
1610
1611 /* NB: AV_NO_VIRUS is 0 but AV_SCAN_FAILED is false,
1612 * so we need the strict equalities === and thus can't use a switch here
1613 */
1614 if ( $mappedCode === AV_SCAN_FAILED ) {
1615 # scan failed (code was mapped to false by $exitCodeMap)
1616 wfDebug( __METHOD__ . ": failed to scan $file (code $exitCode).\n" );
1617
1618 $output = $wgAntivirusRequired
1619 ? wfMessage( 'virus-scanfailed', array( $exitCode ) )->text()
1620 : null;
1621 } elseif ( $mappedCode === AV_SCAN_ABORTED ) {
1622 # scan failed because filetype is unknown (probably imune)
1623 wfDebug( __METHOD__ . ": unsupported file type $file (code $exitCode).\n" );
1624 $output = null;
1625 } elseif ( $mappedCode === AV_NO_VIRUS ) {
1626 # no virus found
1627 wfDebug( __METHOD__ . ": file passed virus scan.\n" );
1628 $output = false;
1629 } else {
1630 $output = trim( $output );
1631
1632 if ( !$output ) {
1633 $output = true; #if there's no output, return true
1634 } elseif ( $msgPattern ) {
1635 $groups = array();
1636 if ( preg_match( $msgPattern, $output, $groups ) ) {
1637 if ( $groups[1] ) {
1638 $output = $groups[1];
1639 }
1640 }
1641 }
1642
1643 wfDebug( __METHOD__ . ": FOUND VIRUS! scanner feedback: $output \n" );
1644 }
1645
1646 wfProfileOut( __METHOD__ );
1647
1648 return $output;
1649 }
1650
1651 /**
1652 * Check if there's an overwrite conflict and, if so, if restrictions
1653 * forbid this user from performing the upload.
1654 *
1655 * @param User $user
1656 *
1657 * @return mixed True on success, array on failure
1658 */
1659 private function checkOverwrite( $user ) {
1660 // First check whether the local file can be overwritten
1661 $file = $this->getLocalFile();
1662 if ( $file->exists() ) {
1663 if ( !self::userCanReUpload( $user, $file ) ) {
1664 return array( 'fileexists-forbidden', $file->getName() );
1665 } else {
1666 return true;
1667 }
1668 }
1669
1670 /* Check shared conflicts: if the local file does not exist, but
1671 * wfFindFile finds a file, it exists in a shared repository.
1672 */
1673 $file = wfFindFile( $this->getTitle() );
1674 if ( $file && !$user->isAllowed( 'reupload-shared' ) ) {
1675 return array( 'fileexists-shared-forbidden', $file->getName() );
1676 }
1677
1678 return true;
1679 }
1680
1681 /**
1682 * Check if a user is the last uploader
1683 *
1684 * @param User $user
1685 * @param string $img Image name
1686 * @return bool
1687 */
1688 public static function userCanReUpload( User $user, $img ) {
1689 if ( $user->isAllowed( 'reupload' ) ) {
1690 return true; // non-conditional
1691 }
1692 if ( !$user->isAllowed( 'reupload-own' ) ) {
1693 return false;
1694 }
1695 if ( is_string( $img ) ) {
1696 $img = wfLocalFile( $img );
1697 }
1698 if ( !( $img instanceof LocalFile ) ) {
1699 return false;
1700 }
1701
1702 return $user->getId() == $img->getUser( 'id' );
1703 }
1704
1705 /**
1706 * Helper function that does various existence checks for a file.
1707 * The following checks are performed:
1708 * - The file exists
1709 * - Article with the same name as the file exists
1710 * - File exists with normalized extension
1711 * - The file looks like a thumbnail and the original exists
1712 *
1713 * @param File $file The File object to check
1714 * @return mixed False if the file does not exists, else an array
1715 */
1716 public static function getExistsWarning( $file ) {
1717 if ( $file->exists() ) {
1718 return array( 'warning' => 'exists', 'file' => $file );
1719 }
1720
1721 if ( $file->getTitle()->getArticleID() ) {
1722 return array( 'warning' => 'page-exists', 'file' => $file );
1723 }
1724
1725 if ( $file->wasDeleted() && !$file->exists() ) {
1726 return array( 'warning' => 'was-deleted', 'file' => $file );
1727 }
1728
1729 if ( strpos( $file->getName(), '.' ) == false ) {
1730 $partname = $file->getName();
1731 $extension = '';
1732 } else {
1733 $n = strrpos( $file->getName(), '.' );
1734 $extension = substr( $file->getName(), $n + 1 );
1735 $partname = substr( $file->getName(), 0, $n );
1736 }
1737 $normalizedExtension = File::normalizeExtension( $extension );
1738
1739 if ( $normalizedExtension != $extension ) {
1740 // We're not using the normalized form of the extension.
1741 // Normal form is lowercase, using most common of alternate
1742 // extensions (eg 'jpg' rather than 'JPEG').
1743 //
1744 // Check for another file using the normalized form...
1745 $nt_lc = Title::makeTitle( NS_FILE, "{$partname}.{$normalizedExtension}" );
1746 $file_lc = wfLocalFile( $nt_lc );
1747
1748 if ( $file_lc->exists() ) {
1749 return array(
1750 'warning' => 'exists-normalized',
1751 'file' => $file,
1752 'normalizedFile' => $file_lc
1753 );
1754 }
1755 }
1756
1757 // Check for files with the same name but a different extension
1758 $similarFiles = RepoGroup::singleton()->getLocalRepo()->findFilesByPrefix(
1759 "{$partname}.", 1 );
1760 if ( count( $similarFiles ) ) {
1761 return array(
1762 'warning' => 'exists-normalized',
1763 'file' => $file,
1764 'normalizedFile' => $similarFiles[0],
1765 );
1766 }
1767
1768 if ( self::isThumbName( $file->getName() ) ) {
1769 # Check for filenames like 50px- or 180px-, these are mostly thumbnails
1770 $nt_thb = Title::newFromText(
1771 substr( $partname, strpos( $partname, '-' ) + 1 ) . '.' . $extension,
1772 NS_FILE
1773 );
1774 $file_thb = wfLocalFile( $nt_thb );
1775 if ( $file_thb->exists() ) {
1776 return array(
1777 'warning' => 'thumb',
1778 'file' => $file,
1779 'thumbFile' => $file_thb
1780 );
1781 } else {
1782 // File does not exist, but we just don't like the name
1783 return array(
1784 'warning' => 'thumb-name',
1785 'file' => $file,
1786 'thumbFile' => $file_thb
1787 );
1788 }
1789 }
1790
1791 foreach ( self::getFilenamePrefixBlacklist() as $prefix ) {
1792 if ( substr( $partname, 0, strlen( $prefix ) ) == $prefix ) {
1793 return array(
1794 'warning' => 'bad-prefix',
1795 'file' => $file,
1796 'prefix' => $prefix
1797 );
1798 }
1799 }
1800
1801 return false;
1802 }
1803
1804 /**
1805 * Helper function that checks whether the filename looks like a thumbnail
1806 * @param string $filename
1807 * @return bool
1808 */
1809 public static function isThumbName( $filename ) {
1810 $n = strrpos( $filename, '.' );
1811 $partname = $n ? substr( $filename, 0, $n ) : $filename;
1812
1813 return (
1814 substr( $partname, 3, 3 ) == 'px-' ||
1815 substr( $partname, 2, 3 ) == 'px-'
1816 ) &&
1817 preg_match( "/[0-9]{2}/", substr( $partname, 0, 2 ) );
1818 }
1819
1820 /**
1821 * Get a list of blacklisted filename prefixes from [[MediaWiki:Filename-prefix-blacklist]]
1822 *
1823 * @return array List of prefixes
1824 */
1825 public static function getFilenamePrefixBlacklist() {
1826 $blacklist = array();
1827 $message = wfMessage( 'filename-prefix-blacklist' )->inContentLanguage();
1828 if ( !$message->isDisabled() ) {
1829 $lines = explode( "\n", $message->plain() );
1830 foreach ( $lines as $line ) {
1831 // Remove comment lines
1832 $comment = substr( trim( $line ), 0, 1 );
1833 if ( $comment == '#' || $comment == '' ) {
1834 continue;
1835 }
1836 // Remove additional comments after a prefix
1837 $comment = strpos( $line, '#' );
1838 if ( $comment > 0 ) {
1839 $line = substr( $line, 0, $comment - 1 );
1840 }
1841 $blacklist[] = trim( $line );
1842 }
1843 }
1844
1845 return $blacklist;
1846 }
1847
1848 /**
1849 * Gets image info about the file just uploaded.
1850 *
1851 * Also has the effect of setting metadata to be an 'indexed tag name' in
1852 * returned API result if 'metadata' was requested. Oddly, we have to pass
1853 * the "result" object down just so it can do that with the appropriate
1854 * format, presumably.
1855 *
1856 * @param ApiResult $result
1857 * @return array Image info
1858 */
1859 public function getImageInfo( $result ) {
1860 $file = $this->getLocalFile();
1861 /** @todo This cries out for refactoring.
1862 * We really want to say $file->getAllInfo(); here.
1863 * Perhaps "info" methods should be moved into files, and the API should
1864 * just wrap them in queries.
1865 */
1866 if ( $file instanceof UploadStashFile ) {
1867 $imParam = ApiQueryStashImageInfo::getPropertyNames();
1868 $info = ApiQueryStashImageInfo::getInfo( $file, array_flip( $imParam ), $result );
1869 } else {
1870 $imParam = ApiQueryImageInfo::getPropertyNames();
1871 $info = ApiQueryImageInfo::getInfo( $file, array_flip( $imParam ), $result );
1872 }
1873
1874 return $info;
1875 }
1876
1877 /**
1878 * @param array $error
1879 * @return Status
1880 */
1881 public function convertVerifyErrorToStatus( $error ) {
1882 $code = $error['status'];
1883 unset( $code['status'] );
1884
1885 return Status::newFatal( $this->getVerificationErrorCode( $code ), $error );
1886 }
1887
1888 /**
1889 * @param null|string $forType
1890 * @return int
1891 */
1892 public static function getMaxUploadSize( $forType = null ) {
1893 global $wgMaxUploadSize;
1894
1895 if ( is_array( $wgMaxUploadSize ) ) {
1896 if ( !is_null( $forType ) && isset( $wgMaxUploadSize[$forType] ) ) {
1897 return $wgMaxUploadSize[$forType];
1898 } else {
1899 return $wgMaxUploadSize['*'];
1900 }
1901 } else {
1902 return intval( $wgMaxUploadSize );
1903 }
1904 }
1905
1906 /**
1907 * Get the current status of a chunked upload (used for polling).
1908 * The status will be read from the *current* user session.
1909 * @param string $statusKey
1910 * @return Status[]|bool
1911 */
1912 public static function getSessionStatus( $statusKey ) {
1913 return isset( $_SESSION[self::SESSION_STATUS_KEY][$statusKey] )
1914 ? $_SESSION[self::SESSION_STATUS_KEY][$statusKey]
1915 : false;
1916 }
1917
1918 /**
1919 * Set the current status of a chunked upload (used for polling).
1920 * The status will be stored in the *current* user session.
1921 * @param string $statusKey
1922 * @param array|bool $value
1923 * @return void
1924 */
1925 public static function setSessionStatus( $statusKey, $value ) {
1926 if ( $value === false ) {
1927 unset( $_SESSION[self::SESSION_STATUS_KEY][$statusKey] );
1928 } else {
1929 $_SESSION[self::SESSION_STATUS_KEY][$statusKey] = $value;
1930 }
1931 }
1932 }