includes/upload/UploadFromUrl.php

   1 <?php
   2 /**
   3  * Backend for uploading files from a HTTP resource.
   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  * Implements uploading from a HTTP resource.
  26  *
  27  * @ingroup Upload
  28  * @author Bryan Tong Minh
  29  * @author Michael Dale
  30  */
  31 class UploadFromUrl extends UploadBase {
  32         protected $mAsync, $mUrl;
  33         protected $mIgnoreWarnings = true;
  34
  35         protected $mTempPath, $mTmpHandle;
  36
  37         protected static $allowedUrls = array();
  38
  39         /**
  40          * Checks if the user is allowed to use the upload-by-URL feature. If the
  41          * user is not allowed, return the name of the user right as a string. If
  42          * the user is allowed, have the parent do further permissions checking.
  43          *
  44          * @param User $user
  45          *
  46          * @return bool|string
  47          */
  48         public static function isAllowed( $user ) {
  49                 if ( !$user->isAllowed( 'upload_by_url' ) ) {
  50                         return 'upload_by_url';
  51                 }
  52
  53                 return parent::isAllowed( $user );
  54         }
  55
  56         /**
  57          * Checks if the upload from URL feature is enabled
  58          * @return bool
  59          */
  60         public static function isEnabled() {
  61                 global $wgAllowCopyUploads;
  62
  63                 return $wgAllowCopyUploads && parent::isEnabled();
  64         }
  65
  66         /**
  67          * Checks whether the URL is for an allowed host
  68          * The domains in the whitelist can include wildcard characters (*) in place
  69          * of any of the domain levels, e.g. '*.flickr.com' or 'upload.*.gov.uk'.
  70          *
  71          * @param string $url
  72          * @return bool
  73          */
  74         public static function isAllowedHost( $url ) {
  75                 global $wgCopyUploadsDomains;
  76                 if ( !count( $wgCopyUploadsDomains ) ) {
  77                         return true;
  78                 }
  79                 $parsedUrl = wfParseUrl( $url );
  80                 if ( !$parsedUrl ) {
  81                         return false;
  82                 }
  83                 $valid = false;
  84                 foreach ( $wgCopyUploadsDomains as $domain ) {
  85                         // See if the domain for the upload matches this whitelisted domain
  86                         $whitelistedDomainPieces = explode( '.', $domain );
  87                         $uploadDomainPieces = explode( '.', $parsedUrl['host'] );
  88                         if ( count( $whitelistedDomainPieces ) === count( $uploadDomainPieces ) ) {
  89                                 $valid = true;
  90                                 // See if all the pieces match or not (excluding wildcards)
  91                                 foreach ( $whitelistedDomainPieces as $index => $piece ) {
  92                                         if ( $piece !== '*' && $piece !== $uploadDomainPieces[$index] ) {
  93                                                 $valid = false;
  94                                         }
  95                                 }
  96                                 if ( $valid ) {
  97                                         // We found a match, so quit comparing against the list
  98                                         break;
  99                                 }
 100                         }
 101                         /* Non-wildcard test
 102                         if ( $parsedUrl['host'] === $domain ) {
 103                                 $valid = true;
 104                                 break;
 105                         }
 106                         */
 107                 }
 108
 109                 return $valid;
 110         }
 111
 112         /**
 113          * Checks whether the URL is not allowed.
 114          *
 115          * @param string $url
 116          * @return bool
 117          */
 118         public static function isAllowedUrl( $url ) {
 119                 if ( !isset( self::$allowedUrls[$url] ) ) {
 120                         $allowed = true;
 121                         Hooks::run( 'IsUploadAllowedFromUrl', array( $url, &$allowed ) );
 122                         self::$allowedUrls[$url] = $allowed;
 123                 }
 124
 125                 return self::$allowedUrls[$url];
 126         }
 127
 128         /**
 129          * Entry point for API upload
 130          *
 131          * @param string $name
 132          * @param string $url
 133          * @param bool|string $async Whether the download should be performed
 134          * asynchronous. False for synchronous, async or async-leavemessage for
 135          * asynchronous download.
 136          * @throws MWException
 137          */
 138         public function initialize( $name, $url, $async = false ) {
 139                 global $wgAllowAsyncCopyUploads;
 140
 141                 $this->mUrl = $url;
 142                 $this->mAsync = $wgAllowAsyncCopyUploads ? $async : false;
 143                 if ( $async ) {
 144                         throw new MWException( 'Asynchronous copy uploads are no longer possible as of r81612.' );
 145                 }
 146
 147                 $tempPath = $this->mAsync ? null : $this->makeTemporaryFile();
 148                 # File size and removeTempFile will be filled in later
 149                 $this->initializePathInfo( $name, $tempPath, 0, false );
 150         }
 151
 152         /**
 153          * Entry point for SpecialUpload
 154          * @param WebRequest $request
 155          */
 156         public function initializeFromRequest( &$request ) {
 157                 $desiredDestName = $request->getText( 'wpDestFile' );
 158                 if ( !$desiredDestName ) {
 159                         $desiredDestName = $request->getText( 'wpUploadFileURL' );
 160                 }
 161                 $this->initialize(
 162                         $desiredDestName,
 163                         trim( $request->getVal( 'wpUploadFileURL' ) ),
 164                         false
 165                 );
 166         }
 167
 168         /**
 169          * @param WebRequest $request
 170          * @return bool
 171          */
 172         public static function isValidRequest( $request ) {
 173                 global $wgUser;
 174
 175                 $url = $request->getVal( 'wpUploadFileURL' );
 176
 177                 return !empty( $url )
 178                         && Http::isValidURI( $url )
 179                         && $wgUser->isAllowed( 'upload_by_url' );
 180         }
 181
 182         /**
 183          * @return string
 184          */
 185         public function getSourceType() {
 186                 return 'url';
 187         }
 188
 189         /**
 190          * Download the file (if not async)
 191          *
 192          * @param array $httpOptions Array of options for MWHttpRequest. Ignored if async.
 193          *   This could be used to override the timeout on the http request.
 194          * @return Status
 195          */
 196         public function fetchFile( $httpOptions = array() ) {
 197                 if ( !Http::isValidURI( $this->mUrl ) ) {
 198                         return Status::newFatal( 'http-invalid-url', $this->mUrl );
 199                 }
 200
 201                 if ( !self::isAllowedHost( $this->mUrl ) ) {
 202                         return Status::newFatal( 'upload-copy-upload-invalid-domain' );
 203                 }
 204                 if ( !self::isAllowedUrl( $this->mUrl ) ) {
 205                         return Status::newFatal( 'upload-copy-upload-invalid-url' );
 206                 }
 207                 if ( !$this->mAsync ) {
 208                         return $this->reallyFetchFile( $httpOptions );
 209                 }
 210
 211                 return Status::newGood();
 212         }
 213
 214         /**
 215          * Create a new temporary file in the URL subdirectory of wfTempDir().
 216          *
 217          * @return string Path to the file
 218          */
 219         protected function makeTemporaryFile() {
 220                 $tmpFile = TempFSFile::factory( 'URL' );
 221                 $tmpFile->bind( $this );
 222
 223                 return $tmpFile->getPath();
 224         }
 225
 226         /**
 227          * Callback: save a chunk of the result of a HTTP request to the temporary file
 228          *
 229          * @param mixed $req
 230          * @param string $buffer
 231          * @return int Number of bytes handled
 232          */
 233         public function saveTempFileChunk( $req, $buffer ) {
 234                 wfDebugLog( 'fileupload', 'Received chunk of ' . strlen( $buffer ) . ' bytes' );
 235                 $nbytes = fwrite( $this->mTmpHandle, $buffer );
 236
 237                 if ( $nbytes == strlen( $buffer ) ) {
 238                         $this->mFileSize += $nbytes;
 239                 } else {
 240                         // Well... that's not good!
 241                         wfDebugLog(
 242                                 'fileupload',
 243                                 'Short write ' . $this->nbytes . '/' . strlen( $buffer ) .
 244                                         ' bytes, aborting with ' . $this->mFileSize . ' uploaded so far'
 245                         );
 246                         fclose( $this->mTmpHandle );
 247                         $this->mTmpHandle = false;
 248                 }
 249
 250                 return $nbytes;
 251         }
 252
 253         /**
 254          * Download the file, save it to the temporary file and update the file
 255          * size and set $mRemoveTempFile to true.
 256          *
 257          * @param array $httpOptions Array of options for MWHttpRequest
 258          * @return Status
 259          */
 260         protected function reallyFetchFile( $httpOptions = array() ) {
 261                 global $wgCopyUploadProxy, $wgCopyUploadTimeout;
 262                 if ( $this->mTempPath === false ) {
 263                         return Status::newFatal( 'tmp-create-error' );
 264                 }
 265
 266                 // Note the temporary file should already be created by makeTemporaryFile()
 267                 $this->mTmpHandle = fopen( $this->mTempPath, 'wb' );
 268                 if ( !$this->mTmpHandle ) {
 269                         return Status::newFatal( 'tmp-create-error' );
 270                 }
 271                 wfDebugLog( 'fileupload', 'Temporary file created "' . $this->mTempPath . '"' );
 272
 273                 $this->mRemoveTempFile = true;
 274                 $this->mFileSize = 0;
 275
 276                 $options = $httpOptions + array( 'followRedirects' => true );
 277
 278                 if ( $wgCopyUploadProxy !== false ) {
 279                         $options['proxy'] = $wgCopyUploadProxy;
 280                 }
 281
 282                 if ( $wgCopyUploadTimeout && !isset( $options['timeout'] ) ) {
 283                         $options['timeout'] = $wgCopyUploadTimeout;
 284                 }
 285                 wfDebugLog(
 286                         'fileupload',
 287                         'Starting download from "' . $this->mUrl . '" ' .
 288                                 '<' . implode( ',', array_keys( array_filter( $options ) ) ) . '>'
 289                 );
 290                 $req = MWHttpRequest::factory( $this->mUrl, $options, __METHOD__ );
 291                 $req->setCallback( array( $this, 'saveTempFileChunk' ) );
 292                 $status = $req->execute();
 293
 294                 if ( $this->mTmpHandle ) {
 295                         // File got written ok...
 296                         fclose( $this->mTmpHandle );
 297                         $this->mTmpHandle = null;
 298                 } else {
 299                         // We encountered a write error during the download...
 300                         return Status::newFatal( 'tmp-write-error' );
 301                 }
 302
 303                 wfDebugLog( 'fileupload', $status );
 304                 if ( $status->isOk() ) {
 305                         wfDebugLog( 'fileupload', 'Download by URL completed successfuly.' );
 306                 } else {
 307                         wfDebugLog(
 308                                 'fileupload',
 309                                 'Download by URL completed with HTTP status ' . $req->getStatus()
 310                         );
 311                 }
 312
 313                 return $status;
 314         }
 315
 316         /**
 317          * Wrapper around the parent function in order to defer verifying the
 318          * upload until the file really has been fetched.
 319          * @return array|mixed
 320          */
 321         public function verifyUpload() {
 322                 if ( $this->mAsync ) {
 323                         return array( 'status' => UploadBase::OK );
 324                 }
 325
 326                 return parent::verifyUpload();
 327         }
 328
 329         /**
 330          * Wrapper around the parent function in order to defer checking warnings
 331          * until the file really has been fetched.
 332          * @return array
 333          */
 334         public function checkWarnings() {
 335                 if ( $this->mAsync ) {
 336                         $this->mIgnoreWarnings = false;
 337
 338                         return array();
 339                 }
 340
 341                 return parent::checkWarnings();
 342         }
 343
 344         /**
 345          * Wrapper around the parent function in order to defer checking protection
 346          * until we are sure that the file can actually be uploaded
 347          * @param User $user
 348          * @return bool|mixed
 349          */
 350         public function verifyTitlePermissions( $user ) {
 351                 if ( $this->mAsync ) {
 352                         return true;
 353                 }
 354
 355                 return parent::verifyTitlePermissions( $user );
 356         }
 357
 358         /**
 359          * Wrapper around the parent function in order to defer uploading to the
 360          * job queue for asynchronous uploads
 361          * @param string $comment
 362          * @param string $pageText
 363          * @param bool $watch
 364          * @param User $user
 365          * @return Status
 366          */
 367         public function performUpload( $comment, $pageText, $watch, $user ) {
 368                 if ( $this->mAsync ) {
 369                         $sessionKey = $this->insertJob( $comment, $pageText, $watch, $user );
 370
 371                         return Status::newFatal( 'async', $sessionKey );
 372                 }
 373
 374                 return parent::performUpload( $comment, $pageText, $watch, $user );
 375         }
 376
 377         /**
 378          * @param string $comment
 379          * @param string $pageText
 380          * @param bool $watch
 381          * @param User $user
 382          * @return string
 383          */
 384         protected function insertJob( $comment, $pageText, $watch, $user ) {
 385                 $sessionKey = $this->stashSession();
 386                 $job = new UploadFromUrlJob( $this->getTitle(), array(
 387                         'url' => $this->mUrl,
 388                         'comment' => $comment,
 389                         'pageText' => $pageText,
 390                         'watch' => $watch,
 391                         'userName' => $user->getName(),
 392                         'leaveMessage' => $this->mAsync == 'async-leavemessage',
 393                         'ignoreWarnings' => $this->mIgnoreWarnings,
 394                         'sessionId' => session_id(),
 395                         'sessionKey' => $sessionKey,
 396                 ) );
 397                 $job->initializeSessionData();
 398                 JobQueueGroup::singleton()->push( $job );
 399
 400                 return $sessionKey;
 401         }
 402 }