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