(bug 28556) Remove MacBinary
[lhc/web/wiklou.git] / includes / upload / UploadBase.php
1 <?php
2 /**
3 * @file
4 * @ingroup upload
5 *
6 * UploadBase and subclasses are the backend of MediaWiki's file uploads.
7 * The frontends are formed by ApiUpload and SpecialUpload.
8 *
9 * See also includes/docs/upload.txt
10 *
11 * @author Brion Vibber
12 * @author Bryan Tong Minh
13 * @author Michael Dale
14 */
15
16 abstract class UploadBase {
17 protected $mTempPath;
18 protected $mDesiredDestName, $mDestName, $mRemoveTempFile, $mSourceType;
19 protected $mTitle = false, $mTitleError = 0;
20 protected $mFilteredName, $mFinalExtension;
21 protected $mLocalFile, $mFileSize, $mFileProps;
22 protected $mBlackListedExtensions;
23 protected $mJavaDetected;
24
25 const SUCCESS = 0;
26 const OK = 0;
27 const EMPTY_FILE = 3;
28 const MIN_LENGTH_PARTNAME = 4;
29 const ILLEGAL_FILENAME = 5;
30 const OVERWRITE_EXISTING_FILE = 7; # Not used anymore; handled by verifyTitlePermissions()
31 const FILETYPE_MISSING = 8;
32 const FILETYPE_BADTYPE = 9;
33 const VERIFICATION_ERROR = 10;
34
35 # HOOK_ABORTED is the new name of UPLOAD_VERIFICATION_ERROR
36 const UPLOAD_VERIFICATION_ERROR = 11;
37 const HOOK_ABORTED = 11;
38 const FILE_TOO_LARGE = 12;
39 const WINDOWS_NONASCII_FILENAME = 13;
40
41 const SESSION_VERSION = 2;
42 const SESSION_KEYNAME = 'wsUploadData';
43
44 static public function getSessionKeyname() {
45 return self::SESSION_KEYNAME;
46 }
47
48 public function getVerificationErrorCode( $error ) {
49 $code_to_status = array(self::EMPTY_FILE => 'empty-file',
50 self::FILE_TOO_LARGE => 'file-too-large',
51 self::FILETYPE_MISSING => 'filetype-missing',
52 self::FILETYPE_BADTYPE => 'filetype-banned',
53 self::MIN_LENGTH_PARTNAME => 'filename-tooshort',
54 self::ILLEGAL_FILENAME => 'illegal-filename',
55 self::OVERWRITE_EXISTING_FILE => 'overwrite',
56 self::VERIFICATION_ERROR => 'verification-error',
57 self::HOOK_ABORTED => 'hookaborted',
58 self::WINDOWS_NONASCII_FILENAME => 'windows-nonascii-filename',
59 );
60 if( isset( $code_to_status[$error] ) ) {
61 return $code_to_status[$error];
62 }
63
64 return 'unknown-error';
65 }
66
67 /**
68 * Returns true if uploads are enabled.
69 * Can be override by subclasses.
70 */
71 public static function isEnabled() {
72 global $wgEnableUploads;
73
74 if ( !$wgEnableUploads ) {
75 return false;
76 }
77
78 # Check php's file_uploads setting
79 return wfIsHipHop() || wfIniGetBool( 'file_uploads' );
80 }
81
82 /**
83 * Returns true if the user can use this upload module or else a string
84 * identifying the missing permission.
85 * Can be overriden by subclasses.
86 *
87 * @param $user User
88 */
89 public static function isAllowed( $user ) {
90 foreach ( array( 'upload', 'edit' ) as $permission ) {
91 if ( !$user->isAllowed( $permission ) ) {
92 return $permission;
93 }
94 }
95 return true;
96 }
97
98 // Upload handlers. Should probably just be a global.
99 static $uploadHandlers = array( 'Stash', 'File', 'Url' );
100
101 /**
102 * Create a form of UploadBase depending on wpSourceType and initializes it
103 *
104 * @param $request WebRequest
105 * @param $type
106 */
107 public static function createFromRequest( &$request, $type = null ) {
108 $type = $type ? $type : $request->getVal( 'wpSourceType', 'File' );
109
110 if( !$type ) {
111 return null;
112 }
113
114 // Get the upload class
115 $type = ucfirst( $type );
116
117 // Give hooks the chance to handle this request
118 $className = null;
119 wfRunHooks( 'UploadCreateFromRequest', array( $type, &$className ) );
120 if ( is_null( $className ) ) {
121 $className = 'UploadFrom' . $type;
122 wfDebug( __METHOD__ . ": class name: $className\n" );
123 if( !in_array( $type, self::$uploadHandlers ) ) {
124 return null;
125 }
126 }
127
128 // Check whether this upload class is enabled
129 if( !call_user_func( array( $className, 'isEnabled' ) ) ) {
130 return null;
131 }
132
133 // Check whether the request is valid
134 if( !call_user_func( array( $className, 'isValidRequest' ), $request ) ) {
135 return null;
136 }
137
138 $handler = new $className;
139
140 $handler->initializeFromRequest( $request );
141 return $handler;
142 }
143
144 /**
145 * Check whether a request if valid for this handler
146 */
147 public static function isValidRequest( $request ) {
148 return false;
149 }
150
151 public function __construct() {}
152
153 /**
154 * Returns the upload type. Should be overridden by child classes
155 *
156 * @since 1.18
157 * @return string
158 */
159 public function getSourceType() { return null; }
160
161 /**
162 * Initialize the path information
163 * @param $name string the desired destination name
164 * @param $tempPath string the temporary path
165 * @param $fileSize int the file size
166 * @param $removeTempFile bool (false) remove the temporary file?
167 * @return null
168 */
169 public function initializePathInfo( $name, $tempPath, $fileSize, $removeTempFile = false ) {
170 $this->mDesiredDestName = $name;
171 $this->mTempPath = $tempPath;
172 $this->mFileSize = $fileSize;
173 $this->mRemoveTempFile = $removeTempFile;
174 }
175
176 /**
177 * Initialize from a WebRequest. Override this in a subclass.
178 */
179 public abstract function initializeFromRequest( &$request );
180
181 /**
182 * Fetch the file. Usually a no-op
183 */
184 public function fetchFile() {
185 return Status::newGood();
186 }
187
188 /**
189 * Return true if the file is empty
190 * @return bool
191 */
192 public function isEmptyFile() {
193 return empty( $this->mFileSize );
194 }
195
196 /**
197 * Return the file size
198 * @return integer
199 */
200 public function getFileSize() {
201 return $this->mFileSize;
202 }
203
204 /**
205 * Append a file to the Repo file
206 *
207 * @param $srcPath String: path to source file
208 * @param $toAppendPath String: path to the Repo file that will be appended to.
209 * @return Status Status
210 */
211 protected function appendToUploadFile( $srcPath, $toAppendPath ) {
212 $repo = RepoGroup::singleton()->getLocalRepo();
213 $status = $repo->append( $srcPath, $toAppendPath );
214 return $status;
215 }
216
217 /**
218 * Finish appending to the Repo file
219 *
220 * @param $toAppendPath String: path to the Repo file that will be appended to.
221 * @return Status Status
222 */
223 protected function appendFinish( $toAppendPath ) {
224 $repo = RepoGroup::singleton()->getLocalRepo();
225 $status = $repo->appendFinish( $toAppendPath );
226 return $status;
227 }
228
229
230 /**
231 * @param $srcPath String: the source path
232 * @return the real path if it was a virtual URL
233 */
234 function getRealPath( $srcPath ) {
235 $repo = RepoGroup::singleton()->getLocalRepo();
236 if ( $repo->isVirtualUrl( $srcPath ) ) {
237 return $repo->resolveVirtualUrl( $srcPath );
238 }
239 return $srcPath;
240 }
241
242 /**
243 * Verify whether the upload is sane.
244 * @return mixed self::OK or else an array with error information
245 */
246 public function verifyUpload() {
247 /**
248 * If there was no filename or a zero size given, give up quick.
249 */
250 if( $this->isEmptyFile() ) {
251 return array( 'status' => self::EMPTY_FILE );
252 }
253
254 /**
255 * Honor $wgMaxUploadSize
256 */
257 $maxSize = self::getMaxUploadSize( $this->getSourceType() );
258 if( $this->mFileSize > $maxSize ) {
259 return array(
260 'status' => self::FILE_TOO_LARGE,
261 'max' => $maxSize,
262 );
263 }
264
265 /**
266 * Look at the contents of the file; if we can recognize the
267 * type but it's corrupt or data of the wrong type, we should
268 * probably not accept it.
269 */
270 $verification = $this->verifyFile();
271 if( $verification !== true ) {
272 return array(
273 'status' => self::VERIFICATION_ERROR,
274 'details' => $verification
275 );
276 }
277
278 /**
279 * Make sure this file can be created
280 */
281 $result = $this->validateName();
282 if( $result !== true ) {
283 return $result;
284 }
285
286 $error = '';
287 if( !wfRunHooks( 'UploadVerification',
288 array( $this->mDestName, $this->mTempPath, &$error ) ) ) {
289 return array( 'status' => self::HOOK_ABORTED, 'error' => $error );
290 }
291
292 return array( 'status' => self::OK );
293 }
294
295 /**
296 * Verify that the name is valid and, if necessary, that we can overwrite
297 *
298 * @return mixed true if valid, otherwise and array with 'status'
299 * and other keys
300 **/
301 protected function validateName() {
302 $nt = $this->getTitle();
303 if( is_null( $nt ) ) {
304 $result = array( 'status' => $this->mTitleError );
305 if( $this->mTitleError == self::ILLEGAL_FILENAME ) {
306 $result['filtered'] = $this->mFilteredName;
307 }
308 if ( $this->mTitleError == self::FILETYPE_BADTYPE ) {
309 $result['finalExt'] = $this->mFinalExtension;
310 if ( count( $this->mBlackListedExtensions ) ) {
311 $result['blacklistedExt'] = $this->mBlackListedExtensions;
312 }
313 }
314 return $result;
315 }
316 $this->mDestName = $this->getLocalFile()->getName();
317
318 return true;
319 }
320
321 /**
322 * Verify the mime type
323 *
324 * @param $mime string representing the mime
325 * @return mixed true if the file is verified, an array otherwise
326 */
327 protected function verifyMimeType( $mime ) {
328 global $wgVerifyMimeType;
329 if ( $wgVerifyMimeType ) {
330 wfDebug ( "\n\nmime: <$mime> extension: <{$this->mFinalExtension}>\n\n");
331 global $wgMimeTypeBlacklist;
332 if ( $this->checkFileExtension( $mime, $wgMimeTypeBlacklist ) ) {
333 return array( 'filetype-badmime', $mime );
334 }
335
336 # XXX: Missing extension will be caught by validateName() via getTitle()
337 if ( $this->mFinalExtension != '' && !$this->verifyExtension( $mime, $this->mFinalExtension ) ) {
338 return array( 'filetype-mime-mismatch', $this->mFinalExtension, $mime );
339 }
340
341 # Check IE type
342 $fp = fopen( $this->mTempPath, 'rb' );
343 $chunk = fread( $fp, 256 );
344 fclose( $fp );
345
346 $magic = MimeMagic::singleton();
347 $extMime = $magic->guessTypesForExtension( $this->mFinalExtension );
348 $ieTypes = $magic->getIEMimeTypes( $this->mTempPath, $chunk, $extMime );
349 foreach ( $ieTypes as $ieType ) {
350 if ( $this->checkFileExtension( $ieType, $wgMimeTypeBlacklist ) ) {
351 return array( 'filetype-bad-ie-mime', $ieType );
352 }
353 }
354 }
355
356 return true;
357 }
358
359 /**
360 * Verifies that it's ok to include the uploaded file
361 *
362 * @return mixed true of the file is verified, array otherwise.
363 */
364 protected function verifyFile() {
365 global $wgAllowJavaUploads;
366 # get the title, even though we are doing nothing with it, because
367 # we need to populate mFinalExtension
368 $this->getTitle();
369
370 $this->mFileProps = File::getPropsFromPath( $this->mTempPath, $this->mFinalExtension );
371
372 # check mime type, if desired
373 $mime = $this->mFileProps[ 'file-mime' ];
374 $status = $this->verifyMimeType( $mime );
375 if ( $status !== true ) {
376 return $status;
377 }
378
379 # check for htmlish code and javascript
380 if( self::detectScript( $this->mTempPath, $mime, $this->mFinalExtension ) ) {
381 return array( 'uploadscripted' );
382 }
383 if( $this->mFinalExtension == 'svg' || $mime == 'image/svg+xml' ) {
384 if( $this->detectScriptInSvg( $this->mTempPath ) ) {
385 return array( 'uploadscripted' );
386 }
387 }
388
389 # Check for Java applets, which if uploaded can bypass cross-site
390 # restrictions.
391 if ( !$wgAllowJavaUploads ) {
392 $this->mJavaDetected = false;
393 $zipStatus = ZipDirectoryReader::read( $this->mTempPath,
394 array( $this, 'zipEntryCallback' ) );
395 if ( !$zipStatus->isOK() ) {
396 $errors = $zipStatus->getErrorsArray();
397 $error = reset( $errors );
398 if ( $error[0] !== 'zip-wrong-format' ) {
399 return $error;
400 }
401 }
402 if ( $this->mJavaDetected ) {
403 return array( 'uploadjava' );
404 }
405 }
406
407 # Scan the uploaded file for viruses
408 $virus = $this->detectVirus( $this->mTempPath );
409 if ( $virus ) {
410 return array( 'uploadvirus', $virus );
411 }
412
413 $handler = MediaHandler::getHandler( $mime );
414 if ( $handler ) {
415 $handlerStatus = $handler->verifyUpload( $this->mTempPath );
416 if ( !$handlerStatus->isOK() ) {
417 $errors = $handlerStatus->getErrorsArray();
418 return reset( $errors );
419 }
420 }
421
422 wfRunHooks( 'UploadVerifyFile', array( $this, $mime, &$status ) );
423 if ( $status !== true ) {
424 return $status;
425 }
426
427 wfDebug( __METHOD__ . ": all clear; passing.\n" );
428 return true;
429 }
430
431 /**
432 * Callback for ZipDirectoryReader to detect Java class files.
433 */
434 function zipEntryCallback( $entry ) {
435 $names = array( $entry['name'] );
436
437 // If there is a null character, cut off the name at it, because JDK's
438 // ZIP_GetEntry() uses strcmp() if the name hashes match. If a file name
439 // were constructed which had ".class\0" followed by a string chosen to
440 // make the hash collide with the truncated name, that file could be
441 // returned in response to a request for the .class file.
442 $nullPos = strpos( $entry['name'], "\000" );
443 if ( $nullPos !== false ) {
444 $names[] = substr( $entry['name'], 0, $nullPos );
445 }
446
447 // If there is a trailing slash in the file name, we have to strip it,
448 // because that's what ZIP_GetEntry() does.
449 if ( preg_grep( '!\.class/?$!', $names ) ) {
450 $this->mJavaDetected = true;
451 }
452 }
453
454 /**
455 * Alias for verifyTitlePermissions. The function was originally 'verifyPermissions'
456 * but that suggests it's checking the user, when it's really checking the title + user combination.
457 * @param $user User object to verify the permissions against
458 * @return mixed An array as returned by getUserPermissionsErrors or true
459 * in case the user has proper permissions.
460 */
461 public function verifyPermissions( $user ) {
462 return $this->verifyTitlePermissions( $user );
463 }
464
465 /**
466 * Check whether the user can edit, upload and create the image. This
467 * checks only against the current title; if it returns errors, it may
468 * very well be that another title will not give errors. Therefore
469 * isAllowed() should be called as well for generic is-user-blocked or
470 * can-user-upload checking.
471 *
472 * @param $user User object to verify the permissions against
473 * @return mixed An array as returned by getUserPermissionsErrors or true
474 * in case the user has proper permissions.
475 */
476 public function verifyTitlePermissions( $user ) {
477 /**
478 * If the image is protected, non-sysop users won't be able
479 * to modify it by uploading a new revision.
480 */
481 $nt = $this->getTitle();
482 if( is_null( $nt ) ) {
483 return true;
484 }
485 $permErrors = $nt->getUserPermissionsErrors( 'edit', $user );
486 $permErrorsUpload = $nt->getUserPermissionsErrors( 'upload', $user );
487 if ( !$nt->exists() ) {
488 $permErrorsCreate = $nt->getUserPermissionsErrors( 'createpage', $user );
489 } else {
490 $permErrorsCreate = array();
491 }
492 if( $permErrors || $permErrorsUpload || $permErrorsCreate ) {
493 $permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsUpload, $permErrors ) );
494 $permErrors = array_merge( $permErrors, wfArrayDiff2( $permErrorsCreate, $permErrors ) );
495 return $permErrors;
496 }
497
498 $overwriteError = $this->checkOverwrite( $user );
499 if ( $overwriteError !== true ) {
500 return array( $overwriteError );
501 }
502
503 return true;
504 }
505
506 /**
507 * Check for non fatal problems with the file
508 *
509 * @return Array of warnings
510 */
511 public function checkWarnings() {
512 $warnings = array();
513
514 $localFile = $this->getLocalFile();
515 $filename = $localFile->getName();
516
517 /**
518 * Check whether the resulting filename is different from the desired one,
519 * but ignore things like ucfirst() and spaces/underscore things
520 */
521 $comparableName = str_replace( ' ', '_', $this->mDesiredDestName );
522 $comparableName = Title::capitalize( $comparableName, NS_FILE );
523
524 if( $this->mDesiredDestName != $filename && $comparableName != $filename ) {
525 $warnings['badfilename'] = $filename;
526 }
527
528 // Check whether the file extension is on the unwanted list
529 global $wgCheckFileExtensions, $wgFileExtensions;
530 if ( $wgCheckFileExtensions ) {
531 if ( !$this->checkFileExtension( $this->mFinalExtension, $wgFileExtensions ) ) {
532 $warnings['filetype-unwanted-type'] = $this->mFinalExtension;
533 }
534 }
535
536 global $wgUploadSizeWarning;
537 if ( $wgUploadSizeWarning && ( $this->mFileSize > $wgUploadSizeWarning ) ) {
538 $warnings['large-file'] = $wgUploadSizeWarning;
539 }
540
541 if ( $this->mFileSize == 0 ) {
542 $warnings['emptyfile'] = true;
543 }
544
545 $exists = self::getExistsWarning( $localFile );
546 if( $exists !== false ) {
547 $warnings['exists'] = $exists;
548 }
549
550 // Check dupes against existing files
551 $hash = File::sha1Base36( $this->mTempPath );
552 $dupes = RepoGroup::singleton()->findBySha1( $hash );
553 $title = $this->getTitle();
554 // Remove all matches against self
555 foreach ( $dupes as $key => $dupe ) {
556 if( $title->equals( $dupe->getTitle() ) ) {
557 unset( $dupes[$key] );
558 }
559 }
560 if( $dupes ) {
561 $warnings['duplicate'] = $dupes;
562 }
563
564 // Check dupes against archives
565 $archivedImage = new ArchivedFile( null, 0, "{$hash}.{$this->mFinalExtension}" );
566 if ( $archivedImage->getID() > 0 ) {
567 $warnings['duplicate-archive'] = $archivedImage->getName();
568 }
569
570 return $warnings;
571 }
572
573 /**
574 * Really perform the upload. Stores the file in the local repo, watches
575 * if necessary and runs the UploadComplete hook.
576 *
577 * @param $user User
578 *
579 * @return Status indicating the whether the upload succeeded.
580 */
581 public function performUpload( $comment, $pageText, $watch, $user ) {
582 $status = $this->getLocalFile()->upload(
583 $this->mTempPath,
584 $comment,
585 $pageText,
586 File::DELETE_SOURCE,
587 $this->mFileProps,
588 false,
589 $user
590 );
591
592 if( $status->isGood() ) {
593 if ( $watch ) {
594 $user->addWatch( $this->getLocalFile()->getTitle() );
595 }
596
597 wfRunHooks( 'UploadComplete', array( &$this ) );
598 }
599
600 return $status;
601 }
602
603 /**
604 * Returns the title of the file to be uploaded. Sets mTitleError in case
605 * the name was illegal.
606 *
607 * @return Title The title of the file or null in case the name was illegal
608 */
609 public function getTitle() {
610 if ( $this->mTitle !== false ) {
611 return $this->mTitle;
612 }
613
614 /* Assume that if a user specified File:Something.jpg, this is an error
615 * and that the namespace prefix needs to be stripped of.
616 */
617 $title = Title::newFromText( $this->mDesiredDestName );
618 if ( $title->getNamespace() == NS_FILE ) {
619 $this->mFilteredName = $title->getDBkey();
620 } else {
621 $this->mFilteredName = $this->mDesiredDestName;
622 }
623
624 /**
625 * Chop off any directories in the given filename. Then
626 * filter out illegal characters, and try to make a legible name
627 * out of it. We'll strip some silently that Title would die on.
628 */
629 $this->mFilteredName = wfStripIllegalFilenameChars( $this->mFilteredName );
630 /* Normalize to title form before we do any further processing */
631 $nt = Title::makeTitleSafe( NS_FILE, $this->mFilteredName );
632 if( is_null( $nt ) ) {
633 $this->mTitleError = self::ILLEGAL_FILENAME;
634 return $this->mTitle = null;
635 }
636 $this->mFilteredName = $nt->getDBkey();
637
638 /**
639 * We'll want to blacklist against *any* 'extension', and use
640 * only the final one for the whitelist.
641 */
642 list( $partname, $ext ) = $this->splitExtensions( $this->mFilteredName );
643
644 if( count( $ext ) ) {
645 $this->mFinalExtension = trim( $ext[count( $ext ) - 1] );
646 } else {
647 $this->mFinalExtension = '';
648
649 # No extension, try guessing one
650 $magic = MimeMagic::singleton();
651 $mime = $magic->guessMimeType( $this->mTempPath );
652 if ( $mime !== 'unknown/unknown' ) {
653 # Get a space separated list of extensions
654 $extList = $magic->getExtensionsForType( $mime );
655 if ( $extList ) {
656 # Set the extension to the canonical extension
657 $this->mFinalExtension = strtok( $extList, ' ' );
658
659 # Fix up the other variables
660 $this->mFilteredName .= ".{$this->mFinalExtension}";
661 $nt = Title::makeTitleSafe( NS_FILE, $this->mFilteredName );
662 $ext = array( $this->mFinalExtension );
663 }
664 }
665
666 }
667
668 /* Don't allow users to override the blacklist (check file extension) */
669 global $wgCheckFileExtensions, $wgStrictFileExtensions;
670 global $wgFileExtensions, $wgFileBlacklist;
671
672 $blackListedExtensions = $this->checkFileExtensionList( $ext, $wgFileBlacklist );
673
674 if ( $this->mFinalExtension == '' ) {
675 $this->mTitleError = self::FILETYPE_MISSING;
676 return $this->mTitle = null;
677 } elseif ( $blackListedExtensions ||
678 ( $wgCheckFileExtensions && $wgStrictFileExtensions &&
679 !$this->checkFileExtension( $this->mFinalExtension, $wgFileExtensions ) ) ) {
680 $this->mBlackListedExtensions = $blackListedExtensions;
681 $this->mTitleError = self::FILETYPE_BADTYPE;
682 return $this->mTitle = null;
683 }
684
685 // Windows may be broken with special characters, see bug XXX
686 if ( wfIsWindows() && !preg_match( '/^[\x0-\x7f]*$/', $nt->getText() ) ) {
687 $this->mTitleError = self::WINDOWS_NONASCII_FILENAME;
688 return $this->mTitle = null;
689 }
690
691 # If there was more than one "extension", reassemble the base
692 # filename to prevent bogus complaints about length
693 if( count( $ext ) > 1 ) {
694 for( $i = 0; $i < count( $ext ) - 1; $i++ ) {
695 $partname .= '.' . $ext[$i];
696 }
697 }
698
699 if( strlen( $partname ) < 1 ) {
700 $this->mTitleError = self::MIN_LENGTH_PARTNAME;
701 return $this->mTitle = null;
702 }
703
704 return $this->mTitle = $nt;
705 }
706
707 /**
708 * Return the local file and initializes if necessary.
709 *
710 * @return LocalFile
711 */
712 public function getLocalFile() {
713 if( is_null( $this->mLocalFile ) ) {
714 $nt = $this->getTitle();
715 $this->mLocalFile = is_null( $nt ) ? null : wfLocalFile( $nt );
716 }
717 return $this->mLocalFile;
718 }
719
720 /**
721 * NOTE: Probably should be deprecated in favor of UploadStash, but this is sometimes
722 * called outside that context.
723 *
724 * Stash a file in a temporary directory for later processing
725 * after the user has confirmed it.
726 *
727 * If the user doesn't explicitly cancel or accept, these files
728 * can accumulate in the temp directory.
729 *
730 * @param $saveName String: the destination filename
731 * @param $tempSrc String: the source temporary file to save
732 * @return String: full path the stashed file, or false on failure
733 */
734 protected function saveTempUploadedFile( $saveName, $tempSrc ) {
735 $repo = RepoGroup::singleton()->getLocalRepo();
736 $status = $repo->storeTemp( $saveName, $tempSrc );
737 return $status;
738 }
739
740 /**
741 * If the user does not supply all necessary information in the first upload form submission (either by accident or
742 * by design) then we may want to stash the file temporarily, get more information, and publish the file later.
743 *
744 * This method will stash a file in a temporary directory for later processing, and save the necessary descriptive info
745 * into the user's session.
746 * This method returns the file object, which also has a 'sessionKey' property which can be passed through a form or
747 * API request to find this stashed file again.
748 *
749 * @param $key String: (optional) the session key used to find the file info again. If not supplied, a key will be autogenerated.
750 * @return UploadStashFile stashed file
751 */
752 public function stashSessionFile( $key = null ) {
753 $stash = RepoGroup::singleton()->getLocalRepo()->getUploadStash();
754 $data = array(
755 'mFileProps' => $this->mFileProps,
756 'mSourceType' => $this->getSourceType(),
757 );
758 $file = $stash->stashFile( $this->mTempPath, $data, $key );
759 $this->mLocalFile = $file;
760 return $file;
761 }
762
763 /**
764 * Stash a file in a temporary directory, returning a key which can be used to find the file again. See stashSessionFile().
765 *
766 * @param $key String: (optional) the session key used to find the file info again. If not supplied, a key will be autogenerated.
767 * @return String: session key
768 */
769 public function stashSession( $key = null ) {
770 return $this->stashSessionFile( $key )->getSessionKey();
771 }
772
773 /**
774 * If we've modified the upload file we need to manually remove it
775 * on exit to clean up.
776 */
777 public function cleanupTempFile() {
778 if ( $this->mRemoveTempFile && $this->mTempPath && file_exists( $this->mTempPath ) ) {
779 wfDebug( __METHOD__ . ": Removing temporary file {$this->mTempPath}\n" );
780 unlink( $this->mTempPath );
781 }
782 }
783
784 public function getTempPath() {
785 return $this->mTempPath;
786 }
787
788 /**
789 * Split a file into a base name and all dot-delimited 'extensions'
790 * on the end. Some web server configurations will fall back to
791 * earlier pseudo-'extensions' to determine type and execute
792 * scripts, so the blacklist needs to check them all.
793 *
794 * @return array
795 */
796 public static function splitExtensions( $filename ) {
797 $bits = explode( '.', $filename );
798 $basename = array_shift( $bits );
799 return array( $basename, $bits );
800 }
801
802 /**
803 * Perform case-insensitive match against a list of file extensions.
804 * Returns true if the extension is in the list.
805 *
806 * @param $ext String
807 * @param $list Array
808 * @return Boolean
809 */
810 public static function checkFileExtension( $ext, $list ) {
811 return in_array( strtolower( $ext ), $list );
812 }
813
814 /**
815 * Perform case-insensitive match against a list of file extensions.
816 * Returns an array of matching extensions.
817 *
818 * @param $ext Array
819 * @param $list Array
820 * @return Boolean
821 */
822 public static function checkFileExtensionList( $ext, $list ) {
823 return array_intersect( array_map( 'strtolower', $ext ), $list );
824 }
825
826 /**
827 * Checks if the mime type of the uploaded file matches the file extension.
828 *
829 * @param $mime String: the mime type of the uploaded file
830 * @param $extension String: the filename extension that the file is to be served with
831 * @return Boolean
832 */
833 public static function verifyExtension( $mime, $extension ) {
834 $magic = MimeMagic::singleton();
835
836 if ( !$mime || $mime == 'unknown' || $mime == 'unknown/unknown' )
837 if ( !$magic->isRecognizableExtension( $extension ) ) {
838 wfDebug( __METHOD__ . ": passing file with unknown detected mime type; " .
839 "unrecognized extension '$extension', can't verify\n" );
840 return true;
841 } else {
842 wfDebug( __METHOD__ . ": rejecting file with unknown detected mime type; ".
843 "recognized extension '$extension', so probably invalid file\n" );
844 return false;
845 }
846
847 $match = $magic->isMatchingExtension( $extension, $mime );
848
849 if ( $match === null ) {
850 wfDebug( __METHOD__ . ": no file extension known for mime type $mime, passing file\n" );
851 return true;
852 } elseif( $match === true ) {
853 wfDebug( __METHOD__ . ": mime type $mime matches extension $extension, passing file\n" );
854
855 #TODO: if it's a bitmap, make sure PHP or ImageMagic resp. can handle it!
856 return true;
857
858 } else {
859 wfDebug( __METHOD__ . ": mime type $mime mismatches file extension $extension, rejecting file\n" );
860 return false;
861 }
862 }
863
864 /**
865 * Heuristic for detecting files that *could* contain JavaScript instructions or
866 * things that may look like HTML to a browser and are thus
867 * potentially harmful. The present implementation will produce false
868 * positives in some situations.
869 *
870 * @param $file String: pathname to the temporary upload file
871 * @param $mime String: the mime type of the file
872 * @param $extension String: the extension of the file
873 * @return Boolean: true if the file contains something looking like embedded scripts
874 */
875 public static function detectScript( $file, $mime, $extension ) {
876 global $wgAllowTitlesInSVG;
877
878 # ugly hack: for text files, always look at the entire file.
879 # For binary field, just check the first K.
880
881 if( strpos( $mime,'text/' ) === 0 ) {
882 $chunk = file_get_contents( $file );
883 } else {
884 $fp = fopen( $file, 'rb' );
885 $chunk = fread( $fp, 1024 );
886 fclose( $fp );
887 }
888
889 $chunk = strtolower( $chunk );
890
891 if( !$chunk ) {
892 return false;
893 }
894
895 # decode from UTF-16 if needed (could be used for obfuscation).
896 if( substr( $chunk, 0, 2 ) == "\xfe\xff" ) {
897 $enc = 'UTF-16BE';
898 } elseif( substr( $chunk, 0, 2 ) == "\xff\xfe" ) {
899 $enc = 'UTF-16LE';
900 } else {
901 $enc = null;
902 }
903
904 if( $enc ) {
905 $chunk = iconv( $enc, "ASCII//IGNORE", $chunk );
906 }
907
908 $chunk = trim( $chunk );
909
910 # @todo FIXME: Convert from UTF-16 if necessarry!
911 wfDebug( __METHOD__ . ": checking for embedded scripts and HTML stuff\n" );
912
913 # check for HTML doctype
914 if ( preg_match( "/<!DOCTYPE *X?HTML/i", $chunk ) ) {
915 return true;
916 }
917
918 /**
919 * Internet Explorer for Windows performs some really stupid file type
920 * autodetection which can cause it to interpret valid image files as HTML
921 * and potentially execute JavaScript, creating a cross-site scripting
922 * attack vectors.
923 *
924 * Apple's Safari browser also performs some unsafe file type autodetection
925 * which can cause legitimate files to be interpreted as HTML if the
926 * web server is not correctly configured to send the right content-type
927 * (or if you're really uploading plain text and octet streams!)
928 *
929 * Returns true if IE is likely to mistake the given file for HTML.
930 * Also returns true if Safari would mistake the given file for HTML
931 * when served with a generic content-type.
932 */
933 $tags = array(
934 '<a href',
935 '<body',
936 '<head',
937 '<html', #also in safari
938 '<img',
939 '<pre',
940 '<script', #also in safari
941 '<table'
942 );
943
944 if( !$wgAllowTitlesInSVG && $extension !== 'svg' && $mime !== 'image/svg' ) {
945 $tags[] = '<title';
946 }
947
948 foreach( $tags as $tag ) {
949 if( false !== strpos( $chunk, $tag ) ) {
950 wfDebug( __METHOD__ . ": found something that may make it be mistaken for html: $tag\n" );
951 return true;
952 }
953 }
954
955 /*
956 * look for JavaScript
957 */
958
959 # resolve entity-refs to look at attributes. may be harsh on big files... cache result?
960 $chunk = Sanitizer::decodeCharReferences( $chunk );
961
962 # look for script-types
963 if( preg_match( '!type\s*=\s*[\'"]?\s*(?:\w*/)?(?:ecma|java)!sim', $chunk ) ) {
964 wfDebug( __METHOD__ . ": found script types\n" );
965 return true;
966 }
967
968 # look for html-style script-urls
969 if( preg_match( '!(?:href|src|data)\s*=\s*[\'"]?\s*(?:ecma|java)script:!sim', $chunk ) ) {
970 wfDebug( __METHOD__ . ": found html-style script urls\n" );
971 return true;
972 }
973
974 # look for css-style script-urls
975 if( preg_match( '!url\s*\(\s*[\'"]?\s*(?:ecma|java)script:!sim', $chunk ) ) {
976 wfDebug( __METHOD__ . ": found css-style script urls\n" );
977 return true;
978 }
979
980 wfDebug( __METHOD__ . ": no scripts found\n" );
981 return false;
982 }
983
984 protected function detectScriptInSvg( $filename ) {
985 $check = new XmlTypeCheck( $filename, array( $this, 'checkSvgScriptCallback' ) );
986 return $check->filterMatch;
987 }
988
989 /**
990 * @todo Replace this with a whitelist filter!
991 */
992 public function checkSvgScriptCallback( $element, $attribs ) {
993 $stripped = $this->stripXmlNamespace( $element );
994
995 if( $stripped == 'script' ) {
996 wfDebug( __METHOD__ . ": Found script element '$element' in uploaded file.\n" );
997 return true;
998 }
999
1000 foreach( $attribs as $attrib => $value ) {
1001 $stripped = $this->stripXmlNamespace( $attrib );
1002 if( substr( $stripped, 0, 2 ) == 'on' ) {
1003 wfDebug( __METHOD__ . ": Found script attribute '$attrib'='value' in uploaded file.\n" );
1004 return true;
1005 }
1006 if( $stripped == 'href' && strpos( strtolower( $value ), 'javascript:' ) !== false ) {
1007 wfDebug( __METHOD__ . ": Found script href attribute '$attrib'='$value' in uploaded file.\n" );
1008 return true;
1009 }
1010 }
1011 }
1012
1013 private function stripXmlNamespace( $name ) {
1014 // 'http://www.w3.org/2000/svg:script' -> 'script'
1015 $parts = explode( ':', strtolower( $name ) );
1016 return array_pop( $parts );
1017 }
1018
1019 /**
1020 * Generic wrapper function for a virus scanner program.
1021 * This relies on the $wgAntivirus and $wgAntivirusSetup variables.
1022 * $wgAntivirusRequired may be used to deny upload if the scan fails.
1023 *
1024 * @param $file String: pathname to the temporary upload file
1025 * @return mixed false if not virus is found, NULL if the scan fails or is disabled,
1026 * or a string containing feedback from the virus scanner if a virus was found.
1027 * If textual feedback is missing but a virus was found, this function returns true.
1028 */
1029 public static function detectVirus( $file ) {
1030 global $wgAntivirus, $wgAntivirusSetup, $wgAntivirusRequired, $wgOut;
1031
1032 if ( !$wgAntivirus ) {
1033 wfDebug( __METHOD__ . ": virus scanner disabled\n" );
1034 return null;
1035 }
1036
1037 if ( !$wgAntivirusSetup[$wgAntivirus] ) {
1038 wfDebug( __METHOD__ . ": unknown virus scanner: $wgAntivirus\n" );
1039 $wgOut->wrapWikiMsg( "<div class=\"error\">\n$1\n</div>",
1040 array( 'virus-badscanner', $wgAntivirus ) );
1041 return wfMsg( 'virus-unknownscanner' ) . " $wgAntivirus";
1042 }
1043
1044 # look up scanner configuration
1045 $command = $wgAntivirusSetup[$wgAntivirus]['command'];
1046 $exitCodeMap = $wgAntivirusSetup[$wgAntivirus]['codemap'];
1047 $msgPattern = isset( $wgAntivirusSetup[$wgAntivirus]['messagepattern'] ) ?
1048 $wgAntivirusSetup[$wgAntivirus]['messagepattern'] : null;
1049
1050 if ( strpos( $command, "%f" ) === false ) {
1051 # simple pattern: append file to scan
1052 $command .= " " . wfEscapeShellArg( $file );
1053 } else {
1054 # complex pattern: replace "%f" with file to scan
1055 $command = str_replace( "%f", wfEscapeShellArg( $file ), $command );
1056 }
1057
1058 wfDebug( __METHOD__ . ": running virus scan: $command \n" );
1059
1060 # execute virus scanner
1061 $exitCode = false;
1062
1063 # NOTE: there's a 50 line workaround to make stderr redirection work on windows, too.
1064 # that does not seem to be worth the pain.
1065 # Ask me (Duesentrieb) about it if it's ever needed.
1066 $output = wfShellExec( "$command 2>&1", $exitCode );
1067
1068 # map exit code to AV_xxx constants.
1069 $mappedCode = $exitCode;
1070 if ( $exitCodeMap ) {
1071 if ( isset( $exitCodeMap[$exitCode] ) ) {
1072 $mappedCode = $exitCodeMap[$exitCode];
1073 } elseif ( isset( $exitCodeMap["*"] ) ) {
1074 $mappedCode = $exitCodeMap["*"];
1075 }
1076 }
1077
1078 if ( $mappedCode === AV_SCAN_FAILED ) {
1079 # scan failed (code was mapped to false by $exitCodeMap)
1080 wfDebug( __METHOD__ . ": failed to scan $file (code $exitCode).\n" );
1081
1082 if ( $wgAntivirusRequired ) {
1083 return wfMsg( 'virus-scanfailed', array( $exitCode ) );
1084 } else {
1085 return null;
1086 }
1087 } elseif ( $mappedCode === AV_SCAN_ABORTED ) {
1088 # scan failed because filetype is unknown (probably imune)
1089 wfDebug( __METHOD__ . ": unsupported file type $file (code $exitCode).\n" );
1090 return null;
1091 } elseif ( $mappedCode === AV_NO_VIRUS ) {
1092 # no virus found
1093 wfDebug( __METHOD__ . ": file passed virus scan.\n" );
1094 return false;
1095 } else {
1096 $output = trim( $output );
1097
1098 if ( !$output ) {
1099 $output = true; #if there's no output, return true
1100 } elseif ( $msgPattern ) {
1101 $groups = array();
1102 if ( preg_match( $msgPattern, $output, $groups ) ) {
1103 if ( $groups[1] ) {
1104 $output = $groups[1];
1105 }
1106 }
1107 }
1108
1109 wfDebug( __METHOD__ . ": FOUND VIRUS! scanner feedback: $output \n" );
1110 return $output;
1111 }
1112 }
1113
1114 /**
1115 * Check if there's an overwrite conflict and, if so, if restrictions
1116 * forbid this user from performing the upload.
1117 *
1118 * @param $user User
1119 *
1120 * @return mixed true on success, array on failure
1121 */
1122 private function checkOverwrite( $user ) {
1123 // First check whether the local file can be overwritten
1124 $file = $this->getLocalFile();
1125 if( $file->exists() ) {
1126 if( !self::userCanReUpload( $user, $file ) ) {
1127 return array( 'fileexists-forbidden', $file->getName() );
1128 } else {
1129 return true;
1130 }
1131 }
1132
1133 /* Check shared conflicts: if the local file does not exist, but
1134 * wfFindFile finds a file, it exists in a shared repository.
1135 */
1136 $file = wfFindFile( $this->getTitle() );
1137 if ( $file && !$user->isAllowed( 'reupload-shared' ) ) {
1138 return array( 'fileexists-shared-forbidden', $file->getName() );
1139 }
1140
1141 return true;
1142 }
1143
1144 /**
1145 * Check if a user is the last uploader
1146 *
1147 * @param $user User object
1148 * @param $img String: image name
1149 * @return Boolean
1150 */
1151 public static function userCanReUpload( User $user, $img ) {
1152 if( $user->isAllowed( 'reupload' ) ) {
1153 return true; // non-conditional
1154 }
1155 if( !$user->isAllowed( 'reupload-own' ) ) {
1156 return false;
1157 }
1158 if( is_string( $img ) ) {
1159 $img = wfLocalFile( $img );
1160 }
1161 if ( !( $img instanceof LocalFile ) ) {
1162 return false;
1163 }
1164
1165 return $user->getId() == $img->getUser( 'id' );
1166 }
1167
1168 /**
1169 * Helper function that does various existence checks for a file.
1170 * The following checks are performed:
1171 * - The file exists
1172 * - Article with the same name as the file exists
1173 * - File exists with normalized extension
1174 * - The file looks like a thumbnail and the original exists
1175 *
1176 * @param $file File The File object to check
1177 * @return mixed False if the file does not exists, else an array
1178 */
1179 public static function getExistsWarning( $file ) {
1180 if( $file->exists() ) {
1181 return array( 'warning' => 'exists', 'file' => $file );
1182 }
1183
1184 if( $file->getTitle()->getArticleID() ) {
1185 return array( 'warning' => 'page-exists', 'file' => $file );
1186 }
1187
1188 if ( $file->wasDeleted() && !$file->exists() ) {
1189 return array( 'warning' => 'was-deleted', 'file' => $file );
1190 }
1191
1192 if( strpos( $file->getName(), '.' ) == false ) {
1193 $partname = $file->getName();
1194 $extension = '';
1195 } else {
1196 $n = strrpos( $file->getName(), '.' );
1197 $extension = substr( $file->getName(), $n + 1 );
1198 $partname = substr( $file->getName(), 0, $n );
1199 }
1200 $normalizedExtension = File::normalizeExtension( $extension );
1201
1202 if ( $normalizedExtension != $extension ) {
1203 // We're not using the normalized form of the extension.
1204 // Normal form is lowercase, using most common of alternate
1205 // extensions (eg 'jpg' rather than 'JPEG').
1206 //
1207 // Check for another file using the normalized form...
1208 $nt_lc = Title::makeTitle( NS_FILE, "{$partname}.{$normalizedExtension}" );
1209 $file_lc = wfLocalFile( $nt_lc );
1210
1211 if( $file_lc->exists() ) {
1212 return array(
1213 'warning' => 'exists-normalized',
1214 'file' => $file,
1215 'normalizedFile' => $file_lc
1216 );
1217 }
1218 }
1219
1220 if ( self::isThumbName( $file->getName() ) ) {
1221 # Check for filenames like 50px- or 180px-, these are mostly thumbnails
1222 $nt_thb = Title::newFromText( substr( $partname , strpos( $partname , '-' ) +1 ) . '.' . $extension, NS_FILE );
1223 $file_thb = wfLocalFile( $nt_thb );
1224 if( $file_thb->exists() ) {
1225 return array(
1226 'warning' => 'thumb',
1227 'file' => $file,
1228 'thumbFile' => $file_thb
1229 );
1230 } else {
1231 // File does not exist, but we just don't like the name
1232 return array(
1233 'warning' => 'thumb-name',
1234 'file' => $file,
1235 'thumbFile' => $file_thb
1236 );
1237 }
1238 }
1239
1240
1241 foreach( self::getFilenamePrefixBlacklist() as $prefix ) {
1242 if ( substr( $partname, 0, strlen( $prefix ) ) == $prefix ) {
1243 return array(
1244 'warning' => 'bad-prefix',
1245 'file' => $file,
1246 'prefix' => $prefix
1247 );
1248 }
1249 }
1250
1251 return false;
1252 }
1253
1254 /**
1255 * Helper function that checks whether the filename looks like a thumbnail
1256 */
1257 public static function isThumbName( $filename ) {
1258 $n = strrpos( $filename, '.' );
1259 $partname = $n ? substr( $filename, 0, $n ) : $filename;
1260 return (
1261 substr( $partname , 3, 3 ) == 'px-' ||
1262 substr( $partname , 2, 3 ) == 'px-'
1263 ) &&
1264 preg_match( "/[0-9]{2}/" , substr( $partname , 0, 2 ) );
1265 }
1266
1267 /**
1268 * Get a list of blacklisted filename prefixes from [[MediaWiki:Filename-prefix-blacklist]]
1269 *
1270 * @return array list of prefixes
1271 */
1272 public static function getFilenamePrefixBlacklist() {
1273 $blacklist = array();
1274 $message = wfMessage( 'filename-prefix-blacklist' )->inContentLanguage();
1275 if( !$message->isDisabled() ) {
1276 $lines = explode( "\n", $message->plain() );
1277 foreach( $lines as $line ) {
1278 // Remove comment lines
1279 $comment = substr( trim( $line ), 0, 1 );
1280 if ( $comment == '#' || $comment == '' ) {
1281 continue;
1282 }
1283 // Remove additional comments after a prefix
1284 $comment = strpos( $line, '#' );
1285 if ( $comment > 0 ) {
1286 $line = substr( $line, 0, $comment-1 );
1287 }
1288 $blacklist[] = trim( $line );
1289 }
1290 }
1291 return $blacklist;
1292 }
1293
1294 /**
1295 * Gets image info about the file just uploaded.
1296 *
1297 * Also has the effect of setting metadata to be an 'indexed tag name' in returned API result if
1298 * 'metadata' was requested. Oddly, we have to pass the "result" object down just so it can do that
1299 * with the appropriate format, presumably.
1300 *
1301 * @param $result ApiResult:
1302 * @return Array: image info
1303 */
1304 public function getImageInfo( $result ) {
1305 $file = $this->getLocalFile();
1306 // TODO This cries out for refactoring. We really want to say $file->getAllInfo(); here.
1307 // Perhaps "info" methods should be moved into files, and the API should just wrap them in queries.
1308 if ( $file instanceof UploadStashFile ) {
1309 $imParam = ApiQueryStashImageInfo::getPropertyNames();
1310 $info = ApiQueryStashImageInfo::getInfo( $file, array_flip( $imParam ), $result );
1311 } else {
1312 $imParam = ApiQueryImageInfo::getPropertyNames();
1313 $info = ApiQueryImageInfo::getInfo( $file, array_flip( $imParam ), $result );
1314 }
1315 return $info;
1316 }
1317
1318
1319 public function convertVerifyErrorToStatus( $error ) {
1320 $code = $error['status'];
1321 unset( $code['status'] );
1322 return Status::newFatal( $this->getVerificationErrorCode( $code ), $error );
1323 }
1324
1325 public static function getMaxUploadSize( $forType = null ) {
1326 global $wgMaxUploadSize;
1327
1328 if ( is_array( $wgMaxUploadSize ) ) {
1329 if ( !is_null( $forType ) && isset( $wgMaxUploadSize[$forType] ) ) {
1330 return $wgMaxUploadSize[$forType];
1331 } else {
1332 return $wgMaxUploadSize['*'];
1333 }
1334 } else {
1335 return intval( $wgMaxUploadSize );
1336 }
1337
1338 }
1339 }