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( $user );
  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 = [];
 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
 149                 wfDebugLog( 'fileconcatenate', "Combined $i chunks in $tAmount seconds." );
 150
 151                 // File system path of the actual full temp file
 152                 $this->setTempFile( $tmpPath );
 153
 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          * Returns the virtual chunk location:
 175          * @param int $index
 176          * @return string
 177          */
 178         function getVirtualChunkLocation( $index ) {
 179                 return $this->repo->getVirtualUrl( 'temp' ) .
 180                         '/' .
 181                         $this->repo->getHashPath(
 182                                 $this->getChunkFileKey( $index )
 183                         ) .
 184                         $this->getChunkFileKey( $index );
 185         }
 186
 187         /**
 188          * Add a chunk to the temporary directory
 189          *
 190          * @param string $chunkPath Path to temporary chunk file
 191          * @param int $chunkSize Size of the current chunk
 192          * @param int $offset Offset of current chunk ( mutch match database chunk offset )
 193          * @return Status
 194          */
 195         public function addChunk( $chunkPath, $chunkSize, $offset ) {
 196                 // Get the offset before we add the chunk to the file system
 197                 $preAppendOffset = $this->getOffset();
 198
 199                 if ( $preAppendOffset + $chunkSize > $this->getMaxUploadSize() ) {
 200                         $status = Status::newFatal( 'file-too-large' );
 201                 } else {
 202                         // Make sure the client is uploading the correct chunk with a matching offset.
 203                         if ( $preAppendOffset == $offset ) {
 204                                 // Update local chunk index for the current chunk
 205                                 $this->mChunkIndex++;
 206                                 try {
 207                                         # For some reason mTempPath is set to first part
 208                                         $oldTemp = $this->mTempPath;
 209                                         $this->mTempPath = $chunkPath;
 210                                         $this->verifyChunk();
 211                                         $this->mTempPath = $oldTemp;
 212                                 } catch ( UploadChunkVerificationException $e ) {
 213                                         return Status::newFatal( $e->getMessage() );
 214                                 }
 215                                 $status = $this->outputChunk( $chunkPath );
 216                                 if ( $status->isGood() ) {
 217                                         // Update local offset:
 218                                         $this->mOffset = $preAppendOffset + $chunkSize;
 219                                         // Update chunk table status db
 220                                         $this->updateChunkStatus();
 221                                 }
 222                         } else {
 223                                 $status = Status::newFatal( 'invalid-chunk-offset' );
 224                         }
 225                 }
 226
 227                 return $status;
 228         }
 229
 230         /**
 231          * Update the chunk db table with the current status:
 232          */
 233         private function updateChunkStatus() {
 234                 wfDebug( __METHOD__ . " update chunk status for {$this->mFileKey} offset:" .
 235                         $this->getOffset() . ' inx:' . $this->getChunkIndex() . "\n" );
 236
 237                 $dbw = $this->repo->getMasterDB();
 238                 // Use a quick transaction since we will upload the full temp file into shared
 239                 // storage, which takes time for large files. We don't want to hold locks then.
 240                 $dbw->update(
 241                         'uploadstash',
 242                         [
 243                                 'us_status' => 'chunks',
 244                                 'us_chunk_inx' => $this->getChunkIndex(),
 245                                 'us_size' => $this->getOffset()
 246                         ],
 247                         [ 'us_key' => $this->mFileKey ],
 248                         __METHOD__
 249                 );
 250                 $dbw->commit( __METHOD__, 'flush' );
 251         }
 252
 253         /**
 254          * Get the chunk db state and populate update relevant local values
 255          */
 256         private function getChunkStatus() {
 257                 // get Master db to avoid race conditions.
 258                 // Otherwise, if chunk upload time < replag there will be spurious errors
 259                 $dbw = $this->repo->getMasterDB();
 260                 $row = $dbw->selectRow(
 261                         'uploadstash',
 262                         [
 263                                 'us_chunk_inx',
 264                                 'us_size',
 265                                 'us_path',
 266                         ],
 267                         [ 'us_key' => $this->mFileKey ],
 268                         __METHOD__
 269                 );
 270                 // Handle result:
 271                 if ( $row ) {
 272                         $this->mChunkIndex = $row->us_chunk_inx;
 273                         $this->mOffset = $row->us_size;
 274                         $this->mVirtualTempPath = $row->us_path;
 275                 }
 276         }
 277
 278         /**
 279          * Get the current Chunk index
 280          * @return int Index of the current chunk
 281          */
 282         private function getChunkIndex() {
 283                 if ( $this->mChunkIndex !== null ) {
 284                         return $this->mChunkIndex;
 285                 }
 286
 287                 return 0;
 288         }
 289
 290         /**
 291          * Get the offset at which the next uploaded chunk will be appended to
 292          * @return int Current byte offset of the chunk file set
 293          */
 294         public function getOffset() {
 295                 if ( $this->mOffset !== null ) {
 296                         return $this->mOffset;
 297                 }
 298
 299                 return 0;
 300         }
 301
 302         /**
 303          * Output the chunk to disk
 304          *
 305          * @param string $chunkPath
 306          * @throws UploadChunkFileException
 307          * @return FileRepoStatus
 308          */
 309         private function outputChunk( $chunkPath ) {
 310                 // Key is fileKey + chunk index
 311                 $fileKey = $this->getChunkFileKey();
 312
 313                 // Store the chunk per its indexed fileKey:
 314                 $hashPath = $this->repo->getHashPath( $fileKey );
 315                 $storeStatus = $this->repo->quickImport( $chunkPath,
 316                         $this->repo->getZonePath( 'temp' ) . "/{$hashPath}{$fileKey}" );
 317
 318                 // Check for error in stashing the chunk:
 319                 if ( !$storeStatus->isOK() ) {
 320                         $error = $storeStatus->getErrorsArray();
 321                         $error = reset( $error );
 322                         if ( !count( $error ) ) {
 323                                 $error = $storeStatus->getWarningsArray();
 324                                 $error = reset( $error );
 325                                 if ( !count( $error ) ) {
 326                                         $error = [ 'unknown', 'no error recorded' ];
 327                                 }
 328                         }
 329                         throw new UploadChunkFileException( "Error storing file in '$chunkPath': " .
 330                                 implode( '; ', $error ) );
 331                 }
 332
 333                 return $storeStatus;
 334         }
 335
 336         private function getChunkFileKey( $index = null ) {
 337                 if ( $index === null ) {
 338                         $index = $this->getChunkIndex();
 339                 }
 340
 341                 return $this->mFileKey . '.' . $index;
 342         }
 343
 344         /**
 345          * Verify that the chunk isn't really an evil html file
 346          *
 347          * @throws UploadChunkVerificationException
 348          */
 349         private function verifyChunk() {
 350                 // Rest mDesiredDestName here so we verify the name as if it were mFileKey
 351                 $oldDesiredDestName = $this->mDesiredDestName;
 352                 $this->mDesiredDestName = $this->mFileKey;
 353                 $this->mTitle = false;
 354                 $res = $this->verifyPartialFile();
 355                 $this->mDesiredDestName = $oldDesiredDestName;
 356                 $this->mTitle = false;
 357                 if ( is_array( $res ) ) {
 358                         throw new UploadChunkVerificationException( $res[0] );
 359                 }
 360         }
 361 }
 362
 363 class UploadChunkZeroLengthFileException extends MWException {
 364 }
 365
 366 class UploadChunkFileException extends MWException {
 367 }
 368
 369 class UploadChunkVerificationException extends MWException {
 370 }