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         /** @var LocalRepo */
  36         private $repo;
  37
  38         /**
  39          * Setup local pointers to stash, repo and user (similar to UploadFromStash)
  40          *
  41          * @param User|null $user Default: null
  42          * @param UploadStash|bool $stash Default: false
  43          * @param FileRepo|bool $repo Default: false
  44          */
  45         public function __construct( $user = null, $stash = false, $repo = false ) {
  46                 // user object. sometimes this won't exist, as when running from cron.
  47                 $this->user = $user;
  48
  49                 if ( $repo ) {
  50                         $this->repo = $repo;
  51                 } else {
  52                         $this->repo = RepoGroup::singleton()->getLocalRepo();
  53                 }
  54
  55                 if ( $stash ) {
  56                         $this->stash = $stash;
  57                 } else {
  58                         if ( $user ) {
  59                                 wfDebug( __METHOD__ . " creating new UploadFromChunks instance for " . $user->getId() . "\n" );
  60                         } else {
  61                                 wfDebug( __METHOD__ . " creating new UploadFromChunks instance with no user\n" );
  62                         }
  63                         $this->stash = new UploadStash( $this->repo, $this->user );
  64                 }
  65
  66                 return true;
  67         }
  68
  69         /**
  70          * Calls the parent stashFile and updates the uploadsession table to handle "chunks"
  71          *
  72          * @param User|null $user
  73          * @return UploadStashFile Stashed file
  74          */
  75         public function stashFile( User $user = null ) {
  76                 // Stash file is the called on creating a new chunk session:
  77                 $this->mChunkIndex = 0;
  78                 $this->mOffset = 0;
  79
  80                 $this->verifyChunk();
  81                 // Create a local stash target
  82                 $this->mLocalFile = parent::stashFile();
  83                 // Update the initial file offset (based on file size)
  84                 $this->mOffset = $this->mLocalFile->getSize();
  85                 $this->mFileKey = $this->mLocalFile->getFileKey();
  86
  87                 // Output a copy of this first to chunk 0 location:
  88                 $this->outputChunk( $this->mLocalFile->getPath() );
  89
  90                 // Update db table to reflect initial "chunk" state
  91                 $this->updateChunkStatus();
  92
  93                 return $this->mLocalFile;
  94         }
  95
  96         /**
  97          * Continue chunk uploading
  98          *
  99          * @param string $name
 100          * @param string $key
 101          * @param WebRequestUpload $webRequestUpload
 102          */
 103         public function continueChunks( $name, $key, $webRequestUpload ) {
 104                 $this->mFileKey = $key;
 105                 $this->mUpload = $webRequestUpload;
 106                 // Get the chunk status form the db:
 107                 $this->getChunkStatus();
 108
 109                 $metadata = $this->stash->getMetadata( $key );
 110                 $this->initializePathInfo( $name,
 111                         $this->getRealPath( $metadata['us_path'] ),
 112                         $metadata['us_size'],
 113                         false
 114                 );
 115         }
 116
 117         /**
 118          * Append the final chunk and ready file for parent::performUpload()
 119          * @return FileRepoStatus
 120          */
 121         public function concatenateChunks() {
 122                 $chunkIndex = $this->getChunkIndex();
 123                 wfDebug( __METHOD__ . " concatenate {$this->mChunkIndex} chunks:" .
 124                         $this->getOffset() . ' inx:' . $chunkIndex . "\n" );
 125
 126                 // Concatenate all the chunks to mVirtualTempPath
 127                 $fileList = array();
 128                 // The first chunk is stored at the mVirtualTempPath path so we start on "chunk 1"
 129                 for ( $i = 0; $i <= $chunkIndex; $i++ ) {
 130                         $fileList[] = $this->getVirtualChunkLocation( $i );
 131                 }
 132
 133                 // Get the file extension from the last chunk
 134                 $ext = FileBackend::extensionFromPath( $this->mVirtualTempPath );
 135                 // Get a 0-byte temp file to perform the concatenation at
 136                 $tmpFile = TempFSFile::factory( 'chunkedupload_', $ext );
 137                 $tmpPath = false; // fail in concatenate()
 138                 if ( $tmpFile ) {
 139                         // keep alive with $this
 140                         $tmpPath = $tmpFile->bind( $this )->getPath();
 141                 }
 142
 143                 // Concatenate the chunks at the temp file
 144                 $tStart = microtime( true );
 145                 $status = $this->repo->concatenate( $fileList, $tmpPath, FileRepo::DELETE_SOURCE );
 146                 $tAmount = microtime( true ) - $tStart;
 147                 if ( !$status->isOk() ) {
 148                         return $status;
 149                 }
 150                 wfDebugLog( 'fileconcatenate', "Combined $i chunks in $tAmount seconds." );
 151
 152                 // File system path
 153                 $this->mTempPath = $tmpPath;
 154                 // Since this was set for the last chunk previously
 155                 $this->mFileSize = filesize( $this->mTempPath );
 156                 $ret = $this->verifyUpload();
 157                 if ( $ret['status'] !== UploadBase::OK ) {
 158                         wfDebugLog( 'fileconcatenate', "Verification failed for chunked upload" );
 159                         $status->fatal( $this->getVerificationErrorCode( $ret['status'] ) );
 160
 161                         return $status;
 162                 }
 163
 164                 // Update the mTempPath and mLocalFile
 165                 // (for FileUpload or normal Stash to take over)
 166                 $tStart = microtime( true );
 167                 $this->mLocalFile = parent::stashFile( $this->user );
 168                 $tAmount = microtime( true ) - $tStart;
 169                 $this->mLocalFile->setLocalReference( $tmpFile ); // reuse (e.g. for getImageInfo())
 170                 wfDebugLog( 'fileconcatenate', "Stashed combined file ($i chunks) in $tAmount seconds." );
 171
 172                 return $status;
 173         }
 174
 175         /**
 176          * Perform the upload, then remove the temp copy afterward
 177          * @param string $comment
 178          * @param string $pageText
 179          * @param bool $watch
 180          * @param User $user
 181          * @return Status
 182          */
 183         public function performUpload( $comment, $pageText, $watch, $user ) {
 184                 $rv = parent::performUpload( $comment, $pageText, $watch, $user );
 185
 186                 return $rv;
 187         }
 188
 189         /**
 190          * Returns the virtual chunk location:
 191          * @param int $index
 192          * @return string
 193          */
 194         function getVirtualChunkLocation( $index ) {
 195                 return $this->repo->getVirtualUrl( 'temp' ) .
 196                         '/' .
 197                         $this->repo->getHashPath(
 198                                 $this->getChunkFileKey( $index )
 199                         ) .
 200                         $this->getChunkFileKey( $index );
 201         }
 202
 203         /**
 204          * Add a chunk to the temporary directory
 205          *
 206          * @param string $chunkPath Path to temporary chunk file
 207          * @param int $chunkSize Size of the current chunk
 208          * @param int $offset Offset of current chunk ( mutch match database chunk offset )
 209          * @return Status
 210          */
 211         public function addChunk( $chunkPath, $chunkSize, $offset ) {
 212                 // Get the offset before we add the chunk to the file system
 213                 $preAppendOffset = $this->getOffset();
 214
 215                 if ( $preAppendOffset + $chunkSize > $this->getMaxUploadSize() ) {
 216                         $status = Status::newFatal( 'file-too-large' );
 217                 } else {
 218                         // Make sure the client is uploading the correct chunk with a matching offset.
 219                         if ( $preAppendOffset == $offset ) {
 220                                 // Update local chunk index for the current chunk
 221                                 $this->mChunkIndex++;
 222                                 try {
 223                                         # For some reason mTempPath is set to first part
 224                                         $oldTemp = $this->mTempPath;
 225                                         $this->mTempPath = $chunkPath;
 226                                         $this->verifyChunk();
 227                                         $this->mTempPath = $oldTemp;
 228                                 } catch ( UploadChunkVerificationException $e ) {
 229                                         return Status::newFatal( $e->getMessage() );
 230                                 }
 231                                 $status = $this->outputChunk( $chunkPath );
 232                                 if ( $status->isGood() ) {
 233                                         // Update local offset:
 234                                         $this->mOffset = $preAppendOffset + $chunkSize;
 235                                         // Update chunk table status db
 236                                         $this->updateChunkStatus();
 237                                 }
 238                         } else {
 239                                 $status = Status::newFatal( 'invalid-chunk-offset' );
 240                         }
 241                 }
 242
 243                 return $status;
 244         }
 245
 246         /**
 247          * Update the chunk db table with the current status:
 248          */
 249         private function updateChunkStatus() {
 250                 wfDebug( __METHOD__ . " update chunk status for {$this->mFileKey} offset:" .
 251                         $this->getOffset() . ' inx:' . $this->getChunkIndex() . "\n" );
 252
 253                 $dbw = $this->repo->getMasterDb();
 254                 // Use a quick transaction since we will upload the full temp file into shared
 255                 // storage, which takes time for large files. We don't want to hold locks then.
 256                 $dbw->begin( __METHOD__ );
 257                 $dbw->update(
 258                         'uploadstash',
 259                         array(
 260                                 'us_status' => 'chunks',
 261                                 'us_chunk_inx' => $this->getChunkIndex(),
 262                                 'us_size' => $this->getOffset()
 263                         ),
 264                         array( 'us_key' => $this->mFileKey ),
 265                         __METHOD__
 266                 );
 267                 $dbw->commit( __METHOD__ );
 268         }
 269
 270         /**
 271          * Get the chunk db state and populate update relevant local values
 272          */
 273         private function getChunkStatus() {
 274                 // get Master db to avoid race conditions.
 275                 // Otherwise, if chunk upload time < replag there will be spurious errors
 276                 $dbw = $this->repo->getMasterDb();
 277                 $row = $dbw->selectRow(
 278                         'uploadstash',
 279                         array(
 280                                 'us_chunk_inx',
 281                                 'us_size',
 282                                 'us_path',
 283                         ),
 284                         array( 'us_key' => $this->mFileKey ),
 285                         __METHOD__
 286                 );
 287                 // Handle result:
 288                 if ( $row ) {
 289                         $this->mChunkIndex = $row->us_chunk_inx;
 290                         $this->mOffset = $row->us_size;
 291                         $this->mVirtualTempPath = $row->us_path;
 292                 }
 293         }
 294
 295         /**
 296          * Get the current Chunk index
 297          * @return int Index of the current chunk
 298          */
 299         private function getChunkIndex() {
 300                 if ( $this->mChunkIndex !== null ) {
 301                         return $this->mChunkIndex;
 302                 }
 303
 304                 return 0;
 305         }
 306
 307         /**
 308          * Gets the current offset in fromt the stashedupload table
 309          * @return int Current byte offset of the chunk file set
 310          */
 311         private function getOffset() {
 312                 if ( $this->mOffset !== null ) {
 313                         return $this->mOffset;
 314                 }
 315
 316                 return 0;
 317         }
 318
 319         /**
 320          * Output the chunk to disk
 321          *
 322          * @param string $chunkPath
 323          * @throws UploadChunkFileException
 324          * @return FileRepoStatus
 325          */
 326         private function outputChunk( $chunkPath ) {
 327                 // Key is fileKey + chunk index
 328                 $fileKey = $this->getChunkFileKey();
 329
 330                 // Store the chunk per its indexed fileKey:
 331                 $hashPath = $this->repo->getHashPath( $fileKey );
 332                 $storeStatus = $this->repo->quickImport( $chunkPath,
 333                         $this->repo->getZonePath( 'temp' ) . "/{$hashPath}{$fileKey}" );
 334
 335                 // Check for error in stashing the chunk:
 336                 if ( !$storeStatus->isOK() ) {
 337                         $error = $storeStatus->getErrorsArray();
 338                         $error = reset( $error );
 339                         if ( !count( $error ) ) {
 340                                 $error = $storeStatus->getWarningsArray();
 341                                 $error = reset( $error );
 342                                 if ( !count( $error ) ) {
 343                                         $error = array( 'unknown', 'no error recorded' );
 344                                 }
 345                         }
 346                         throw new UploadChunkFileException( "Error storing file in '$chunkPath': " .
 347                                 implode( '; ', $error ) );
 348                 }
 349
 350                 return $storeStatus;
 351         }
 352
 353         private function getChunkFileKey( $index = null ) {
 354                 if ( $index === null ) {
 355                         $index = $this->getChunkIndex();
 356                 }
 357
 358                 return $this->mFileKey . '.' . $index;
 359         }
 360
 361         /**
 362          * Verify that the chunk isn't really an evil html file
 363          *
 364          * @throws UploadChunkVerificationException
 365          */
 366         private function verifyChunk() {
 367                 // Rest mDesiredDestName here so we verify the name as if it were mFileKey
 368                 $oldDesiredDestName = $this->mDesiredDestName;
 369                 $this->mDesiredDestName = $this->mFileKey;
 370                 $this->mTitle = false;
 371                 $res = $this->verifyPartialFile();
 372                 $this->mDesiredDestName = $oldDesiredDestName;
 373                 $this->mTitle = false;
 374                 if ( is_array( $res ) ) {
 375                         throw new UploadChunkVerificationException( $res[0] );
 376                 }
 377         }
 378 }
 379
 380 class UploadChunkZeroLengthFileException extends MWException {
 381 }
 382
 383 class UploadChunkFileException extends MWException {
 384 }
 385
 386 class UploadChunkVerificationException extends MWException {
 387 }