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