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