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 use MediaWiki\MediaWikiServices;
  25 use MediaWiki\User\UserIdentity;
  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 UserIdentity $user
  47          *
  48          * @return bool|string
  49          */
  50         public static function isAllowed( UserIdentity $user ) {
  51                 if ( !MediaWikiServices::getInstance()
  52                                 ->getPermissionManager()
  53                                 ->userHasRight( $user, 'upload_by_url' )
  54                 ) {
  55                         return 'upload_by_url';
  56                 }
  57
  58                 return parent::isAllowed( $user );
  59         }
  60
  61         /**
  62          * Checks if the upload from URL feature is enabled
  63          * @return bool
  64          */
  65         public static function isEnabled() {
  66                 global $wgAllowCopyUploads;
  67
  68                 return $wgAllowCopyUploads && parent::isEnabled();
  69         }
  70
  71         /**
  72          * Checks whether the URL is for an allowed host
  73          * The domains in the whitelist can include wildcard characters (*) in place
  74          * of any of the domain levels, e.g. '*.flickr.com' or 'upload.*.gov.uk'.
  75          *
  76          * @param string $url
  77          * @return bool
  78          */
  79         public static function isAllowedHost( $url ) {
  80                 global $wgCopyUploadsDomains;
  81                 if ( !count( $wgCopyUploadsDomains ) ) {
  82                         return true;
  83                 }
  84                 $parsedUrl = wfParseUrl( $url );
  85                 if ( !$parsedUrl ) {
  86                         return false;
  87                 }
  88                 $valid = false;
  89                 foreach ( $wgCopyUploadsDomains as $domain ) {
  90                         // See if the domain for the upload matches this whitelisted domain
  91                         $whitelistedDomainPieces = explode( '.', $domain );
  92                         $uploadDomainPieces = explode( '.', $parsedUrl['host'] );
  93                         if ( count( $whitelistedDomainPieces ) === count( $uploadDomainPieces ) ) {
  94                                 $valid = true;
  95                                 // See if all the pieces match or not (excluding wildcards)
  96                                 foreach ( $whitelistedDomainPieces as $index => $piece ) {
  97                                         if ( $piece !== '*' && $piece !== $uploadDomainPieces[$index] ) {
  98                                                 $valid = false;
  99                                         }
 100                                 }
 101                                 if ( $valid ) {
 102                                         // We found a match, so quit comparing against the list
 103                                         break;
 104                                 }
 105                         }
 106                         /* Non-wildcard test
 107                         if ( $parsedUrl['host'] === $domain ) {
 108                                 $valid = true;
 109                                 break;
 110                         }
 111                         */
 112                 }
 113
 114                 return $valid;
 115         }
 116
 117         /**
 118          * Checks whether the URL is not allowed.
 119          *
 120          * @param string $url
 121          * @return bool
 122          */
 123         public static function isAllowedUrl( $url ) {
 124                 if ( !isset( self::$allowedUrls[$url] ) ) {
 125                         $allowed = true;
 126                         Hooks::run( 'IsUploadAllowedFromUrl', [ $url, &$allowed ] );
 127                         self::$allowedUrls[$url] = $allowed;
 128                 }
 129
 130                 return self::$allowedUrls[$url];
 131         }
 132
 133         /**
 134          * Entry point for API upload
 135          *
 136          * @param string $name
 137          * @param string $url
 138          * @throws MWException
 139          */
 140         public function initialize( $name, $url ) {
 141                 $this->mUrl = $url;
 142
 143                 $tempPath = $this->makeTemporaryFile();
 144                 # File size and removeTempFile will be filled in later
 145                 $this->initializePathInfo( $name, $tempPath, 0, false );
 146         }
 147
 148         /**
 149          * Entry point for SpecialUpload
 150          * @param WebRequest &$request
 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                 );
 161         }
 162
 163         /**
 164          * @param WebRequest $request
 165          * @return bool
 166          */
 167         public static function isValidRequest( $request ) {
 168                 global $wgUser;
 169
 170                 $url = $request->getVal( 'wpUploadFileURL' );
 171
 172                 return !empty( $url )
 173                         && MediaWikiServices::getInstance()
 174                                    ->getPermissionManager()
 175                                    ->userHasRight( $wgUser, 'upload_by_url' );
 176         }
 177
 178         /**
 179          * @return string
 180          */
 181         public function getSourceType() {
 182                 return 'url';
 183         }
 184
 185         /**
 186          * Download the file
 187          *
 188          * @param array $httpOptions Array of options for MWHttpRequest.
 189          *   This could be used to override the timeout on the http request.
 190          * @return Status
 191          */
 192         public function fetchFile( $httpOptions = [] ) {
 193                 if ( !Http::isValidURI( $this->mUrl ) ) {
 194                         return Status::newFatal( 'http-invalid-url', $this->mUrl );
 195                 }
 196
 197                 if ( !self::isAllowedHost( $this->mUrl ) ) {
 198                         return Status::newFatal( 'upload-copy-upload-invalid-domain' );
 199                 }
 200                 if ( !self::isAllowedUrl( $this->mUrl ) ) {
 201                         return Status::newFatal( 'upload-copy-upload-invalid-url' );
 202                 }
 203                 return $this->reallyFetchFile( $httpOptions );
 204         }
 205
 206         /**
 207          * Create a new temporary file in the URL subdirectory of wfTempDir().
 208          *
 209          * @return string Path to the file
 210          */
 211         protected function makeTemporaryFile() {
 212                 $tmpFile = MediaWikiServices::getInstance()->getTempFSFileFactory()
 213                         ->newTempFSFile( 'URL', 'urlupload_' );
 214                 $tmpFile->bind( $this );
 215
 216                 return $tmpFile->getPath();
 217         }
 218
 219         /**
 220          * Callback: save a chunk of the result of a HTTP request to the temporary file
 221          *
 222          * @param mixed $req
 223          * @param string $buffer
 224          * @return int Number of bytes handled
 225          */
 226         public function saveTempFileChunk( $req, $buffer ) {
 227                 wfDebugLog( 'fileupload', 'Received chunk of ' . strlen( $buffer ) . ' bytes' );
 228                 $nbytes = fwrite( $this->mTmpHandle, $buffer );
 229
 230                 if ( $nbytes == strlen( $buffer ) ) {
 231                         $this->mFileSize += $nbytes;
 232                 } else {
 233                         // Well... that's not good!
 234                         wfDebugLog(
 235                                 'fileupload',
 236                                 'Short write ' . $nbytes . '/' . strlen( $buffer ) .
 237                                         ' bytes, aborting with ' . $this->mFileSize . ' uploaded so far'
 238                         );
 239                         fclose( $this->mTmpHandle );
 240                         $this->mTmpHandle = false;
 241                 }
 242
 243                 return $nbytes;
 244         }
 245
 246         /**
 247          * Download the file, save it to the temporary file and update the file
 248          * size and set $mRemoveTempFile to true.
 249          *
 250          * @param array $httpOptions Array of options for MWHttpRequest
 251          * @return Status
 252          */
 253         protected function reallyFetchFile( $httpOptions = [] ) {
 254                 global $wgCopyUploadProxy, $wgCopyUploadTimeout;
 255                 if ( $this->mTempPath === false ) {
 256                         return Status::newFatal( 'tmp-create-error' );
 257                 }
 258
 259                 // Note the temporary file should already be created by makeTemporaryFile()
 260                 $this->mTmpHandle = fopen( $this->mTempPath, 'wb' );
 261                 if ( !$this->mTmpHandle ) {
 262                         return Status::newFatal( 'tmp-create-error' );
 263                 }
 264                 wfDebugLog( 'fileupload', 'Temporary file created "' . $this->mTempPath . '"' );
 265
 266                 $this->mRemoveTempFile = true;
 267                 $this->mFileSize = 0;
 268
 269                 $options = $httpOptions + [ 'followRedirects' => true ];
 270
 271                 if ( $wgCopyUploadProxy !== false ) {
 272                         $options['proxy'] = $wgCopyUploadProxy;
 273                 }
 274
 275                 if ( $wgCopyUploadTimeout && !isset( $options['timeout'] ) ) {
 276                         $options['timeout'] = $wgCopyUploadTimeout;
 277                 }
 278                 wfDebugLog(
 279                         'fileupload',
 280                         'Starting download from "' . $this->mUrl . '" ' .
 281                                 '<' . implode( ',', array_keys( array_filter( $options ) ) ) . '>'
 282                 );
 283                 $req = MWHttpRequest::factory( $this->mUrl, $options, __METHOD__ );
 284                 $req->setCallback( [ $this, 'saveTempFileChunk' ] );
 285                 $status = $req->execute();
 286
 287                 if ( $this->mTmpHandle ) {
 288                         // File got written ok...
 289                         fclose( $this->mTmpHandle );
 290                         $this->mTmpHandle = null;
 291                 } else {
 292                         // We encountered a write error during the download...
 293                         return Status::newFatal( 'tmp-write-error' );
 294                 }
 295
 296                 wfDebugLog( 'fileupload', $status );
 297                 if ( $status->isOK() ) {
 298                         wfDebugLog( 'fileupload', 'Download by URL completed successfully.' );
 299                 } else {
 300                         wfDebugLog(
 301                                 'fileupload',
 302                                 'Download by URL completed with HTTP status ' . $req->getStatus()
 303                         );
 304                 }
 305
 306                 return $status;
 307         }
 308 }