includes/upload/UploadFromChunks.php

   1 <?php
   2 /**
   3  * Backend for uploading files from chunks.
   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 chunks
  26  *
  27  * @ingroup Upload
  28  * @author Michael Dale
  29  */
  30 class UploadFromChunks extends UploadFromFile {
  31         protected $mOffset;
  32         protected $mChunkIndex;
  33         protected $mFileKey;
  34         protected $mVirtualTempPath;
  35
  36         /**
  37          * Setup local pointers to stash, repo and user (similar to UploadFromStash)
  38          *
  39          * @param $user User|null Default: null
  40          * @param $stash UploadStash|bool Default: false
  41          * @param $repo FileRepo|bool Default: false
  42          */
  43         public function __construct( $user = null, $stash = false, $repo = false ) {
  44                 // user object. sometimes this won't exist, as when running from cron.
  45                 $this->user = $user;
  46
  47                 if ( $repo ) {
  48                         $this->repo = $repo;
  49                 } else {
  50                         $this->repo = RepoGroup::singleton()->getLocalRepo();
  51                 }
  52
  53                 if ( $stash ) {
  54                         $this->stash = $stash;
  55                 } else {
  56                         if ( $user ) {
  57                                 wfDebug( __METHOD__ . " creating new UploadFromChunks instance for " . $user->getId() . "\n" );
  58                         } else {
  59                                 wfDebug( __METHOD__ . " creating new UploadFromChunks instance with no user\n" );
  60                         }
  61                         $this->stash = new UploadStash( $this->repo, $this->user );
  62                 }
  63
  64                 return true;
  65         }
  66
  67         /**
  68          * Calls the parent stashFile and updates the uploadsession table to handle "chunks"
  69          *
  70          * @return UploadStashFile stashed file
  71          */
  72         public function stashFile( User $user = null ) {
  73                 // Stash file is the called on creating a new chunk session:
  74                 $this->mChunkIndex = 0;
  75                 $this->mOffset = 0;
  76
  77                 $this->verifyChunk();
  78                 // Create a local stash target
  79                 $this->mLocalFile = parent::stashFile();
  80                 // Update the initial file offset (based on file size)
  81                 $this->mOffset = $this->mLocalFile->getSize();
  82                 $this->mFileKey = $this->mLocalFile->getFileKey();
  83
  84                 // Output a copy of this first to chunk 0 location:
  85                 $this->outputChunk( $this->mLocalFile->getPath() );
  86
  87                 // Update db table to reflect initial "chunk" state
  88                 $this->updateChunkStatus();
  89                 return $this->mLocalFile;
  90         }
  91
  92         /**
  93          * Continue chunk uploading
  94          */
  95         public function continueChunks( $name, $key, $webRequestUpload ) {
  96                 $this->mFileKey = $key;
  97                 $this->mUpload = $webRequestUpload;
  98                 // Get the chunk status form the db:
  99                 $this->getChunkStatus();
 100
 101                 $metadata = $this->stash->getMetadata( $key );
 102                 $this->initializePathInfo( $name,
 103                         $this->getRealPath( $metadata['us_path'] ),
 104                         $metadata['us_size'],
 105                         false
 106                 );
 107         }
 108
 109         /**
 110          * Append the final chunk and ready file for parent::performUpload()
 111          * @return FileRepoStatus
 112          */
 113         public function concatenateChunks() {
 114                 $chunkIndex = $this->getChunkIndex();
 115                 wfDebug( __METHOD__ . " concatenate {$this->mChunkIndex} chunks:" .
 116                         $this->getOffset() . ' inx:' . $chunkIndex . "\n" );
 117
 118                 // Concatenate all the chunks to mVirtualTempPath
 119                 $fileList = array();
 120                 // The first chunk is stored at the mVirtualTempPath path so we start on "chunk 1"
 121                 for ( $i = 0; $i <= $chunkIndex; $i++ ) {
 122                         $fileList[] = $this->getVirtualChunkLocation( $i );
 123                 }
 124
 125                 // Get the file extension from the last chunk
 126                 $ext = FileBackend::extensionFromPath( $this->mVirtualTempPath );
 127                 // Get a 0-byte temp file to perform the concatenation at
 128                 $tmpFile = TempFSFile::factory( 'chunkedupload_', $ext );
 129                 $tmpPath = false; // fail in concatenate()
 130                 if ( $tmpFile ) {
 131                         // keep alive with $this
 132                         $tmpPath = $tmpFile->bind( $this )->getPath();
 133                 }
 134
 135                 // Concatenate the chunks at the temp file
 136                 $tStart = microtime( true );
 137                 $status = $this->repo->concatenate( $fileList, $tmpPath, FileRepo::DELETE_SOURCE );
 138                 $tAmount = microtime( true ) - $tStart;
 139                 if ( !$status->isOk() ) {
 140                         return $status;
 141                 }
 142                 wfDebugLog( 'fileconcatenate', "Combined $i chunks in $tAmount seconds.\n" );
 143
 144                 // File system path
 145                 $this->mTempPath = $tmpPath;
 146                 // Since this was set for the last chunk previously
 147                 $this->mFileSize = filesize( $this->mTempPath );
 148                 $ret = $this->verifyUpload();
 149                 if ( $ret['status'] !== UploadBase::OK ) {
 150                         wfDebugLog( 'fileconcatenate', "Verification failed for chunked upload" );
 151                         $status->fatal( $this->getVerificationErrorCode( $ret['status'] ) );
 152                         return $status;
 153                 }
 154
 155                 // Update the mTempPath and mLocalFile
 156                 // (for FileUpload or normal Stash to take over)
 157                 $tStart = microtime( true );
 158                 $this->mLocalFile = parent::stashFile( $this->user );
 159                 $tAmount = microtime( true ) - $tStart;
 160                 $this->mLocalFile->setLocalReference( $tmpFile ); // reuse (e.g. for getImageInfo())
 161                 wfDebugLog( 'fileconcatenate', "Stashed combined file ($i chunks) in $tAmount seconds.\n" );
 162
 163                 return $status;
 164         }
 165
 166         /**
 167          * Perform the upload, then remove the temp copy afterward
 168          * @param $comment string
 169          * @param $pageText string
 170          * @param $watch bool
 171          * @param $user User
 172          * @return Status
 173          */
 174         public function performUpload( $comment, $pageText, $watch, $user ) {
 175                 $rv = parent::performUpload( $comment, $pageText, $watch, $user );
 176                 return $rv;
 177         }
 178
 179         /**
 180          * Returns the virtual chunk location:
 181          * @param $index
 182          * @return string
 183          */
 184         function getVirtualChunkLocation( $index ) {
 185                 return $this->repo->getVirtualUrl( 'temp' ) .
 186                                 '/' .
 187                                 $this->repo->getHashPath(
 188                                         $this->getChunkFileKey( $index )
 189                                 ) .
 190                                 $this->getChunkFileKey( $index );
 191         }
 192
 193         /**
 194          * Add a chunk to the temporary directory
 195          *
 196          * @param string $chunkPath path to temporary chunk file
 197          * @param int $chunkSize size of the current chunk
 198          * @param int $offset offset of current chunk ( mutch match database chunk offset )
 199          * @return Status
 200          */
 201         public function addChunk( $chunkPath, $chunkSize, $offset ) {
 202                 // Get the offset before we add the chunk to the file system
 203                 $preAppendOffset = $this->getOffset();
 204
 205                 if ( $preAppendOffset + $chunkSize > $this->getMaxUploadSize() ) {
 206                         $status = Status::newFatal( 'file-too-large' );
 207                 } else {
 208                         // Make sure the client is uploading the correct chunk with a matching offset.
 209                         if ( $preAppendOffset == $offset ) {
 210                                 // Update local chunk index for the current chunk
 211                                 $this->mChunkIndex++;
 212                                 try {
 213                                         # For some reason mTempPath is set to first part
 214                                         $oldTemp = $this->mTempPath;
 215                                         $this->mTempPath = $chunkPath;
 216                                         $this->verifyChunk();
 217                                         $this->mTempPath = $oldTemp;
 218                                 } catch ( UploadChunkVerificationException $e ) {
 219                                         return Status::newFatal( $e->getMessage() );
 220                                 }
 221                                 $status = $this->outputChunk( $chunkPath );
 222                                 if ( $status->isGood() ) {
 223                                         // Update local offset:
 224                                         $this->mOffset = $preAppendOffset + $chunkSize;
 225                                         // Update chunk table status db
 226                                         $this->updateChunkStatus();
 227                                 }
 228                         } else {
 229                                 $status = Status::newFatal( 'invalid-chunk-offset' );
 230                         }
 231                 }
 232                 return $status;
 233         }
 234
 235         /**
 236          * Update the chunk db table with the current status:
 237          */
 238         private function updateChunkStatus() {
 239                 wfDebug( __METHOD__ . " update chunk status for {$this->mFileKey} offset:" .
 240                                         $this->getOffset() . ' inx:' . $this->getChunkIndex() . "\n" );
 241
 242                 $dbw = $this->repo->getMasterDb();
 243                 // Use a quick transaction since we will upload the full temp file into shared
 244                 // storage, which takes time for large files. We don't want to hold locks then.
 245                 $dbw->begin( __METHOD__ );
 246                 $dbw->update(
 247                         'uploadstash',
 248                         array(
 249                                 'us_status' => 'chunks',
 250                                 'us_chunk_inx' => $this->getChunkIndex(),
 251                                 'us_size' => $this->getOffset()
 252                         ),
 253                         array( 'us_key' => $this->mFileKey ),
 254                         __METHOD__
 255                 );
 256                 $dbw->commit( __METHOD__ );
 257         }
 258
 259         /**
 260          * Get the chunk db state and populate update relevant local values
 261          */
 262         private function getChunkStatus() {
 263                 // get Master db to avoid race conditions.
 264                 // Otherwise, if chunk upload time < replag there will be spurious errors
 265                 $dbw = $this->repo->getMasterDb();
 266                 $row = $dbw->selectRow(
 267                         'uploadstash',
 268                         array(
 269                                 'us_chunk_inx',
 270                                 'us_size',
 271                                 'us_path',
 272                         ),
 273                         array( 'us_key' => $this->mFileKey ),
 274                         __METHOD__
 275                 );
 276                 // Handle result:
 277                 if ( $row ) {
 278                         $this->mChunkIndex = $row->us_chunk_inx;
 279                         $this->mOffset = $row->us_size;
 280                         $this->mVirtualTempPath = $row->us_path;
 281                 }
 282         }
 283
 284         /**
 285          * Get the current Chunk index
 286          * @return Integer index of the current chunk
 287          */
 288         private function getChunkIndex() {
 289                 if ( $this->mChunkIndex !== null ) {
 290                         return $this->mChunkIndex;
 291                 }
 292                 return 0;
 293         }
 294
 295         /**
 296          * Gets the current offset in fromt the stashedupload table
 297          * @return Integer current byte offset of the chunk file set
 298          */
 299         private function getOffset() {
 300                 if ( $this->mOffset !== null ) {
 301                         return $this->mOffset;
 302                 }
 303                 return 0;
 304         }
 305
 306         /**
 307          * Output the chunk to disk
 308          *
 309          * @param $chunkPath string
 310          * @throws UploadChunkFileException
 311          * @return FileRepoStatus
 312          */
 313         private function outputChunk( $chunkPath ) {
 314                 // Key is fileKey + chunk index
 315                 $fileKey = $this->getChunkFileKey();
 316
 317                 // Store the chunk per its indexed fileKey:
 318                 $hashPath = $this->repo->getHashPath( $fileKey );
 319                 $storeStatus = $this->repo->quickImport( $chunkPath,
 320                         $this->repo->getZonePath( 'temp' ) . "/{$hashPath}{$fileKey}" );
 321
 322                 // Check for error in stashing the chunk:
 323                 if ( ! $storeStatus->isOK() ) {
 324                         $error = $storeStatus->getErrorsArray();
 325                         $error = reset( $error );
 326                         if ( ! count( $error ) ) {
 327                                 $error = $storeStatus->getWarningsArray();
 328                                 $error = reset( $error );
 329                                 if ( ! count( $error ) ) {
 330                                         $error = array( 'unknown', 'no error recorded' );
 331                                 }
 332                         }
 333                         throw new UploadChunkFileException( "Error storing file in '$chunkPath': " .
 334                                 implode( '; ', $error ) );
 335                 }
 336                 return $storeStatus;
 337         }
 338
 339         private function getChunkFileKey( $index = null ) {
 340                 if ( $index === null ) {
 341                         $index = $this->getChunkIndex();
 342                 }
 343                 return $this->mFileKey . '.' . $index;
 344         }
 345
 346         /**
 347          * Verify that the chunk isn't really an evil html file
 348          *
 349          * @throws UploadChunkVerificationException
 350          */
 351         private function verifyChunk() {
 352                 // Rest mDesiredDestName here so we verify the name as if it were mFileKey
 353                 $oldDesiredDestName = $this->mDesiredDestName;
 354                 $this->mDesiredDestName = $this->mFileKey;
 355                 $this->mTitle = false;
 356                 $res = $this->verifyPartialFile();
 357                 $this->mDesiredDestName = $oldDesiredDestName;
 358                 $this->mTitle = false;
 359                 if ( is_array( $res ) ) {
 360                         throw new UploadChunkVerificationException( $res[0] );
 361                 }
 362         }
 363 }
 364
 365 class UploadChunkZeroLengthFileException extends MWException {
 366 }
 367
 368 class UploadChunkFileException extends MWException {
 369 }
 370
 371 class UploadChunkVerificationException extends MWException {
 372 }