includes/upload/UploadFromUrl.php

   1 <?php
   2
   3 use MediaWiki\MediaWikiServices;
   4
   5 /**
   6  * Backend for uploading files from a HTTP resource.
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Upload
  25  */
  26
  27 /**
  28  * Implements uploading from a HTTP resource.
  29  *
  30  * @ingroup Upload
  31  * @author Bryan Tong Minh
  32  * @author Michael Dale
  33  */
  34 class UploadFromUrl extends UploadBase {
  35         protected $mUrl;
  36
  37         protected $mTempPath, $mTmpHandle;
  38
  39         protected static $allowedUrls = [];
  40
  41         /**
  42          * Checks if the user is allowed to use the upload-by-URL feature. If the
  43          * user is not allowed, return the name of the user right as a string. If
  44          * the user is allowed, have the parent do further permissions checking.
  45          *
  46          * @param User $user
  47          *
  48          * @return bool|string
  49          */
  50         public static function isAllowed( $user ) {
  51                 if ( !$user->isAllowed( 'upload_by_url' ) ) {
  52                         return 'upload_by_url';
  53                 }
  54
  55                 return parent::isAllowed( $user );
  56         }
  57
  58         /**
  59          * Checks if the upload from URL feature is enabled
  60          * @return bool
  61          */
  62         public static function isEnabled() {
  63                 global $wgAllowCopyUploads;
  64
  65                 return $wgAllowCopyUploads && parent::isEnabled();
  66         }
  67
  68         /**
  69          * Checks whether the URL is for an allowed host
  70          * The domains in the whitelist can include wildcard characters (*) in place
  71          * of any of the domain levels, e.g. '*.flickr.com' or 'upload.*.gov.uk'.
  72          *
  73          * @param string $url
  74          * @return bool
  75          */
  76         public static function isAllowedHost( $url ) {
  77                 global $wgCopyUploadsDomains;
  78                 if ( !count( $wgCopyUploadsDomains ) ) {
  79                         return true;
  80                 }
  81                 $parsedUrl = wfParseUrl( $url );
  82                 if ( !$parsedUrl ) {
  83                         return false;
  84                 }
  85                 $valid = false;
  86                 foreach ( $wgCopyUploadsDomains as $domain ) {
  87                         // See if the domain for the upload matches this whitelisted domain
  88                         $whitelistedDomainPieces = explode( '.', $domain );
  89                         $uploadDomainPieces = explode( '.', $parsedUrl['host'] );
  90                         if ( count( $whitelistedDomainPieces ) === count( $uploadDomainPieces ) ) {
  91                                 $valid = true;
  92                                 // See if all the pieces match or not (excluding wildcards)
  93                                 foreach ( $whitelistedDomainPieces as $index => $piece ) {
  94                                         if ( $piece !== '*' && $piece !== $uploadDomainPieces[$index] ) {
  95                                                 $valid = false;
  96                                         }
  97                                 }
  98                                 if ( $valid ) {
  99                                         // We found a match, so quit comparing against the list
 100                                         break;
 101                                 }
 102                         }
 103                         /* Non-wildcard test
 104                         if ( $parsedUrl['host'] === $domain ) {
 105                                 $valid = true;
 106                                 break;
 107                         }
 108                         */
 109                 }
 110
 111                 return $valid;
 112         }
 113
 114         /**
 115          * Checks whether the URL is not allowed.
 116          *
 117          * @param string $url
 118          * @return bool
 119          */
 120         public static function isAllowedUrl( $url ) {
 121                 if ( !isset( self::$allowedUrls[$url] ) ) {
 122                         $allowed = true;
 123                         Hooks::run( 'IsUploadAllowedFromUrl', [ $url, &$allowed ] );
 124                         self::$allowedUrls[$url] = $allowed;
 125                 }
 126
 127                 return self::$allowedUrls[$url];
 128         }
 129
 130         /**
 131          * Entry point for API upload
 132          *
 133          * @param string $name
 134          * @param string $url
 135          * @throws MWException
 136          */
 137         public function initialize( $name, $url ) {
 138                 $this->mUrl = $url;
 139
 140                 $tempPath = $this->makeTemporaryFile();
 141                 # File size and removeTempFile will be filled in later
 142                 $this->initializePathInfo( $name, $tempPath, 0, false );
 143         }
 144
 145         /**
 146          * Entry point for SpecialUpload
 147          * @param WebRequest &$request
 148          */
 149         public function initializeFromRequest( &$request ) {
 150                 $desiredDestName = $request->getText( 'wpDestFile' );
 151                 if ( !$desiredDestName ) {
 152                         $desiredDestName = $request->getText( 'wpUploadFileURL' );
 153                 }
 154                 $this->initialize(
 155                         $desiredDestName,
 156                         trim( $request->getVal( 'wpUploadFileURL' ) )
 157                 );
 158         }
 159
 160         /**
 161          * @param WebRequest $request
 162          * @return bool
 163          */
 164         public static function isValidRequest( $request ) {
 165                 global $wgUser;
 166
 167                 $url = $request->getVal( 'wpUploadFileURL' );
 168
 169                 return !empty( $url )
 170                         && $wgUser->isAllowed( 'upload_by_url' );
 171         }
 172
 173         /**
 174          * @return string
 175          */
 176         public function getSourceType() {
 177                 return 'url';
 178         }
 179
 180         /**
 181          * Download the file
 182          *
 183          * @param array $httpOptions Array of options for MWHttpRequest.
 184          *   This could be used to override the timeout on the http request.
 185          * @return Status
 186          */
 187         public function fetchFile( $httpOptions = [] ) {
 188                 if ( !Http::isValidURI( $this->mUrl ) ) {
 189                         return Status::newFatal( 'http-invalid-url', $this->mUrl );
 190                 }
 191
 192                 if ( !self::isAllowedHost( $this->mUrl ) ) {
 193                         return Status::newFatal( 'upload-copy-upload-invalid-domain' );
 194                 }
 195                 if ( !self::isAllowedUrl( $this->mUrl ) ) {
 196                         return Status::newFatal( 'upload-copy-upload-invalid-url' );
 197                 }
 198                 return $this->reallyFetchFile( $httpOptions );
 199         }
 200
 201         /**
 202          * Create a new temporary file in the URL subdirectory of wfTempDir().
 203          *
 204          * @return string Path to the file
 205          */
 206         protected function makeTemporaryFile() {
 207                 $tmpFile = MediaWikiServices::getInstance()->getTempFSFileFactory()
 208                         ->newTempFSFile( 'URL', 'urlupload_' );
 209                 $tmpFile->bind( $this );
 210
 211                 return $tmpFile->getPath();
 212         }
 213
 214         /**
 215          * Callback: save a chunk of the result of a HTTP request to the temporary file
 216          *
 217          * @param mixed $req
 218          * @param string $buffer
 219          * @return int Number of bytes handled
 220          */
 221         public function saveTempFileChunk( $req, $buffer ) {
 222                 wfDebugLog( 'fileupload', 'Received chunk of ' . strlen( $buffer ) . ' bytes' );
 223                 $nbytes = fwrite( $this->mTmpHandle, $buffer );
 224
 225                 if ( $nbytes == strlen( $buffer ) ) {
 226                         $this->mFileSize += $nbytes;
 227                 } else {
 228                         // Well... that's not good!
 229                         wfDebugLog(
 230                                 'fileupload',
 231                                 'Short write ' . $nbytes . '/' . strlen( $buffer ) .
 232                                         ' bytes, aborting with ' . $this->mFileSize . ' uploaded so far'
 233                         );
 234                         fclose( $this->mTmpHandle );
 235                         $this->mTmpHandle = false;
 236                 }
 237
 238                 return $nbytes;
 239         }
 240
 241         /**
 242          * Download the file, save it to the temporary file and update the file
 243          * size and set $mRemoveTempFile to true.
 244          *
 245          * @param array $httpOptions Array of options for MWHttpRequest
 246          * @return Status
 247          */
 248         protected function reallyFetchFile( $httpOptions = [] ) {
 249                 global $wgCopyUploadProxy, $wgCopyUploadTimeout;
 250                 if ( $this->mTempPath === false ) {
 251                         return Status::newFatal( 'tmp-create-error' );
 252                 }
 253
 254                 // Note the temporary file should already be created by makeTemporaryFile()
 255                 $this->mTmpHandle = fopen( $this->mTempPath, 'wb' );
 256                 if ( !$this->mTmpHandle ) {
 257                         return Status::newFatal( 'tmp-create-error' );
 258                 }
 259                 wfDebugLog( 'fileupload', 'Temporary file created "' . $this->mTempPath . '"' );
 260
 261                 $this->mRemoveTempFile = true;
 262                 $this->mFileSize = 0;
 263
 264                 $options = $httpOptions + [ 'followRedirects' => true ];
 265
 266                 if ( $wgCopyUploadProxy !== false ) {
 267                         $options['proxy'] = $wgCopyUploadProxy;
 268                 }
 269
 270                 if ( $wgCopyUploadTimeout && !isset( $options['timeout'] ) ) {
 271                         $options['timeout'] = $wgCopyUploadTimeout;
 272                 }
 273                 wfDebugLog(
 274                         'fileupload',
 275                         'Starting download from "' . $this->mUrl . '" ' .
 276                                 '<' . implode( ',', array_keys( array_filter( $options ) ) ) . '>'
 277                 );
 278                 $req = MWHttpRequest::factory( $this->mUrl, $options, __METHOD__ );
 279                 $req->setCallback( [ $this, 'saveTempFileChunk' ] );
 280                 $status = $req->execute();
 281
 282                 if ( $this->mTmpHandle ) {
 283                         // File got written ok...
 284                         fclose( $this->mTmpHandle );
 285                         $this->mTmpHandle = null;
 286                 } else {
 287                         // We encountered a write error during the download...
 288                         return Status::newFatal( 'tmp-write-error' );
 289                 }
 290
 291                 wfDebugLog( 'fileupload', $status );
 292                 if ( $status->isOK() ) {
 293                         wfDebugLog( 'fileupload', 'Download by URL completed successfully.' );
 294                 } else {
 295                         wfDebugLog(
 296                                 'fileupload',
 297                                 'Download by URL completed with HTTP status ' . $req->getStatus()
 298                         );
 299                 }
 300
 301                 return $status;
 302         }
 303 }