includes/filerepo/LocalRepo.php

   1 <?php
   2 /**
   3  * Local repository that stores files in the local filesystem and registers them
   4  * in the wiki's own database.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup FileRepo
  23  */
  24
  25 /**
  26  * A repository that stores files in the local filesystem and registers them
  27  * in the wiki's own database. This is the most commonly used repository class.
  28  *
  29  * @ingroup FileRepo
  30  */
  31 class LocalRepo extends FileRepo {
  32         /** @var bool */
  33         protected $hasSha1Storage = false;
  34
  35         /** @var array */
  36         protected $fileFactory = array( 'LocalFile', 'newFromTitle' );
  37
  38         /** @var array */
  39         protected $fileFactoryKey = array( 'LocalFile', 'newFromKey' );
  40
  41         /** @var array */
  42         protected $fileFromRowFactory = array( 'LocalFile', 'newFromRow' );
  43
  44         /** @var array */
  45         protected $oldFileFromRowFactory = array( 'OldLocalFile', 'newFromRow' );
  46
  47         /** @var array */
  48         protected $oldFileFactory = array( 'OldLocalFile', 'newFromTitle' );
  49
  50         /** @var array */
  51         protected $oldFileFactoryKey = array( 'OldLocalFile', 'newFromKey' );
  52
  53         function __construct( array $info = null ) {
  54                 parent::__construct( $info );
  55
  56                 $this->hasSha1Storage = isset( $info['storageLayout'] ) && $info['storageLayout'] === 'sha1';
  57
  58                 if ( $this->hasSha1Storage() ) {
  59                         $this->backend = new FileBackendDBRepoWrapper( array(
  60                                 'backend'         => $this->backend,
  61                                 'repoName'        => $this->name,
  62                                 'dbHandleFactory' => $this->getDBFactory()
  63                         ) );
  64                 }
  65         }
  66
  67         /**
  68          * @throws MWException
  69          * @param stdClass $row
  70          * @return LocalFile
  71          */
  72         function newFileFromRow( $row ) {
  73                 if ( isset( $row->img_name ) ) {
  74                         return call_user_func( $this->fileFromRowFactory, $row, $this );
  75                 } elseif ( isset( $row->oi_name ) ) {
  76                         return call_user_func( $this->oldFileFromRowFactory, $row, $this );
  77                 } else {
  78                         throw new MWException( __METHOD__ . ': invalid row' );
  79                 }
  80         }
  81
  82         /**
  83          * @param Title $title
  84          * @param string $archiveName
  85          * @return OldLocalFile
  86          */
  87         function newFromArchiveName( $title, $archiveName ) {
  88                 return OldLocalFile::newFromArchiveName( $title, $this, $archiveName );
  89         }
  90
  91         /**
  92          * Delete files in the deleted directory if they are not referenced in the
  93          * filearchive table. This needs to be done in the repo because it needs to
  94          * interleave database locks with file operations, which is potentially a
  95          * remote operation.
  96          *
  97          * @param array $storageKeys
  98          *
  99          * @return FileRepoStatus
 100          */
 101         function cleanupDeletedBatch( array $storageKeys ) {
 102                 if ( $this->hasSha1Storage() ) {
 103                         wfDebug( __METHOD__ . ": skipped because storage uses sha1 paths\n" );
 104                         return Status::newGood();
 105                 }
 106
 107                 $backend = $this->backend; // convenience
 108                 $root = $this->getZonePath( 'deleted' );
 109                 $dbw = $this->getMasterDB();
 110                 $status = $this->newGood();
 111                 $storageKeys = array_unique( $storageKeys );
 112                 foreach ( $storageKeys as $key ) {
 113                         $hashPath = $this->getDeletedHashPath( $key );
 114                         $path = "$root/$hashPath$key";
 115                         $dbw->startAtomic( __METHOD__ );
 116                         // Check for usage in deleted/hidden files and preemptively
 117                         // lock the key to avoid any future use until we are finished.
 118                         $deleted = $this->deletedFileHasKey( $key, 'lock' );
 119                         $hidden = $this->hiddenFileHasKey( $key, 'lock' );
 120                         if ( !$deleted && !$hidden ) { // not in use now
 121                                 wfDebug( __METHOD__ . ": deleting $key\n" );
 122                                 $op = array( 'op' => 'delete', 'src' => $path );
 123                                 if ( !$backend->doOperation( $op )->isOK() ) {
 124                                         $status->error( 'undelete-cleanup-error', $path );
 125                                         $status->failCount++;
 126                                 }
 127                         } else {
 128                                 wfDebug( __METHOD__ . ": $key still in use\n" );
 129                                 $status->successCount++;
 130                         }
 131                         $dbw->endAtomic( __METHOD__ );
 132                 }
 133
 134                 return $status;
 135         }
 136
 137         /**
 138          * Check if a deleted (filearchive) file has this sha1 key
 139          *
 140          * @param string $key File storage key (base-36 sha1 key with file extension)
 141          * @param string|null $lock Use "lock" to lock the row via FOR UPDATE
 142          * @return bool File with this key is in use
 143          */
 144         protected function deletedFileHasKey( $key, $lock = null ) {
 145                 $options = ( $lock === 'lock' ) ? array( 'FOR UPDATE' ) : array();
 146
 147                 $dbw = $this->getMasterDB();
 148
 149                 return (bool)$dbw->selectField( 'filearchive', '1',
 150                         array( 'fa_storage_group' => 'deleted', 'fa_storage_key' => $key ),
 151                         __METHOD__, $options
 152                 );
 153         }
 154
 155         /**
 156          * Check if a hidden (revision delete) file has this sha1 key
 157          *
 158          * @param string $key File storage key (base-36 sha1 key with file extension)
 159          * @param string|null $lock Use "lock" to lock the row via FOR UPDATE
 160          * @return bool File with this key is in use
 161          */
 162         protected function hiddenFileHasKey( $key, $lock = null ) {
 163                 $options = ( $lock === 'lock' ) ? array( 'FOR UPDATE' ) : array();
 164
 165                 $sha1 = self::getHashFromKey( $key );
 166                 $ext = File::normalizeExtension( substr( $key, strcspn( $key, '.' ) + 1 ) );
 167
 168                 $dbw = $this->getMasterDB();
 169
 170                 return (bool)$dbw->selectField( 'oldimage', '1',
 171                         array( 'oi_sha1' => $sha1,
 172                                 'oi_archive_name ' . $dbw->buildLike( $dbw->anyString(), ".$ext" ),
 173                                 $dbw->bitAnd( 'oi_deleted', File::DELETED_FILE ) => File::DELETED_FILE ),
 174                         __METHOD__, $options
 175                 );
 176         }
 177
 178         /**
 179          * Gets the SHA1 hash from a storage key
 180          *
 181          * @param string $key
 182          * @return string
 183          */
 184         public static function getHashFromKey( $key ) {
 185                 return strtok( $key, '.' );
 186         }
 187
 188         /**
 189          * Checks if there is a redirect named as $title
 190          *
 191          * @param Title $title Title of file
 192          * @return bool|Title
 193          */
 194         function checkRedirect( Title $title ) {
 195                 $cache = ObjectCache::getMainWANInstance();
 196
 197                 $title = File::normalizeTitle( $title, 'exception' );
 198
 199                 $memcKey = $this->getSharedCacheKey( 'image_redirect', md5( $title->getDBkey() ) );
 200                 if ( $memcKey === false ) {
 201                         $memcKey = $this->getLocalCacheKey( 'image_redirect', md5( $title->getDBkey() ) );
 202                         $expiry = 300; // no invalidation, 5 minutes
 203                 } else {
 204                         $expiry = 86400; // has invalidation, 1 day
 205                 }
 206                 $cachedValue = $cache->get( $memcKey );
 207                 if ( $cachedValue === ' ' || $cachedValue === '' ) {
 208                         // Does not exist
 209                         return false;
 210                 } elseif ( strval( $cachedValue ) !== '' && $cachedValue !== ' PURGED' ) {
 211                         return Title::newFromText( $cachedValue, NS_FILE );
 212                 } // else $cachedValue is false or null: cache miss
 213
 214                 $id = $this->getArticleID( $title );
 215                 if ( !$id ) {
 216                         $cache->set( $memcKey, " ", $expiry );
 217
 218                         return false;
 219                 }
 220                 $dbr = $this->getSlaveDB();
 221                 $row = $dbr->selectRow(
 222                         'redirect',
 223                         array( 'rd_title', 'rd_namespace' ),
 224                         array( 'rd_from' => $id ),
 225                         __METHOD__
 226                 );
 227
 228                 if ( $row && $row->rd_namespace == NS_FILE ) {
 229                         $targetTitle = Title::makeTitle( $row->rd_namespace, $row->rd_title );
 230                         $cache->set( $memcKey, $targetTitle->getDBkey(), $expiry );
 231
 232                         return $targetTitle;
 233                 } else {
 234                         $cache->set( $memcKey, '', $expiry );
 235
 236                         return false;
 237                 }
 238         }
 239
 240         /**
 241          * Function link Title::getArticleID().
 242          * We can't say Title object, what database it should use, so we duplicate that function here.
 243          *
 244          * @param Title $title
 245          * @return bool|int|mixed
 246          */
 247         protected function getArticleID( $title ) {
 248                 if ( !$title instanceof Title ) {
 249                         return 0;
 250                 }
 251                 $dbr = $this->getSlaveDB();
 252                 $id = $dbr->selectField(
 253                         'page', // Table
 254                         'page_id', //Field
 255                         array( //Conditions
 256                                 'page_namespace' => $title->getNamespace(),
 257                                 'page_title' => $title->getDBkey(),
 258                         ),
 259                         __METHOD__ //Function name
 260                 );
 261
 262                 return $id;
 263         }
 264
 265         public function findFiles( array $items, $flags = 0 ) {
 266                 $finalFiles = array(); // map of (DB key => corresponding File) for matches
 267
 268                 $searchSet = array(); // map of (normalized DB key => search params)
 269                 foreach ( $items as $item ) {
 270                         if ( is_array( $item ) ) {
 271                                 $title = File::normalizeTitle( $item['title'] );
 272                                 if ( $title ) {
 273                                         $searchSet[$title->getDBkey()] = $item;
 274                                 }
 275                         } else {
 276                                 $title = File::normalizeTitle( $item );
 277                                 if ( $title ) {
 278                                         $searchSet[$title->getDBkey()] = array();
 279                                 }
 280                         }
 281                 }
 282
 283                 $fileMatchesSearch = function ( File $file, array $search ) {
 284                         // Note: file name comparison done elsewhere (to handle redirects)
 285                         $user = ( !empty( $search['private'] ) && $search['private'] instanceof User )
 286                                 ? $search['private']
 287                                 : null;
 288
 289                         return (
 290                                 $file->exists() &&
 291                                 (
 292                                         ( empty( $search['time'] ) && !$file->isOld() ) ||
 293                                         ( !empty( $search['time'] ) && $search['time'] === $file->getTimestamp() )
 294                                 ) &&
 295                                 ( !empty( $search['private'] ) || !$file->isDeleted( File::DELETED_FILE ) ) &&
 296                                 $file->userCan( File::DELETED_FILE, $user )
 297                         );
 298                 };
 299
 300                 $that = $this;
 301                 $applyMatchingFiles = function ( ResultWrapper $res, &$searchSet, &$finalFiles )
 302                         use ( $that, $fileMatchesSearch, $flags )
 303                 {
 304                         global $wgContLang;
 305                         $info = $that->getInfo();
 306                         foreach ( $res as $row ) {
 307                                 $file = $that->newFileFromRow( $row );
 308                                 // There must have been a search for this DB key, but this has to handle the
 309                                 // cases were title capitalization is different on the client and repo wikis.
 310                                 $dbKeysLook = array( strtr( $file->getName(), ' ', '_' ) );
 311                                 if ( !empty( $info['initialCapital'] ) ) {
 312                                         // Search keys for "hi.png" and "Hi.png" should use the "Hi.png file"
 313                                         $dbKeysLook[] = $wgContLang->lcfirst( $file->getName() );
 314                                 }
 315                                 foreach ( $dbKeysLook as $dbKey ) {
 316                                         if ( isset( $searchSet[$dbKey] )
 317                                                 && $fileMatchesSearch( $file, $searchSet[$dbKey] )
 318                                         ) {
 319                                                 $finalFiles[$dbKey] = ( $flags & FileRepo::NAME_AND_TIME_ONLY )
 320                                                         ? array( 'title' => $dbKey, 'timestamp' => $file->getTimestamp() )
 321                                                         : $file;
 322                                                 unset( $searchSet[$dbKey] );
 323                                         }
 324                                 }
 325                         }
 326                 };
 327
 328                 $dbr = $this->getSlaveDB();
 329
 330                 // Query image table
 331                 $imgNames = array();
 332                 foreach ( array_keys( $searchSet ) as $dbKey ) {
 333                         $imgNames[] = $this->getNameFromTitle( File::normalizeTitle( $dbKey ) );
 334                 }
 335
 336                 if ( count( $imgNames ) ) {
 337                         $res = $dbr->select( 'image',
 338                                 LocalFile::selectFields(), array( 'img_name' => $imgNames ), __METHOD__ );
 339                         $applyMatchingFiles( $res, $searchSet, $finalFiles );
 340                 }
 341
 342                 // Query old image table
 343                 $oiConds = array(); // WHERE clause array for each file
 344                 foreach ( $searchSet as $dbKey => $search ) {
 345                         if ( isset( $search['time'] ) ) {
 346                                 $oiConds[] = $dbr->makeList(
 347                                         array(
 348                                                 'oi_name' => $this->getNameFromTitle( File::normalizeTitle( $dbKey ) ),
 349                                                 'oi_timestamp' => $dbr->timestamp( $search['time'] )
 350                                         ),
 351                                         LIST_AND
 352                                 );
 353                         }
 354                 }
 355
 356                 if ( count( $oiConds ) ) {
 357                         $res = $dbr->select( 'oldimage',
 358                                 OldLocalFile::selectFields(), $dbr->makeList( $oiConds, LIST_OR ), __METHOD__ );
 359                         $applyMatchingFiles( $res, $searchSet, $finalFiles );
 360                 }
 361
 362                 // Check for redirects...
 363                 foreach ( $searchSet as $dbKey => $search ) {
 364                         if ( !empty( $search['ignoreRedirect'] ) ) {
 365                                 continue;
 366                         }
 367
 368                         $title = File::normalizeTitle( $dbKey );
 369                         $redir = $this->checkRedirect( $title ); // hopefully hits memcached
 370
 371                         if ( $redir && $redir->getNamespace() == NS_FILE ) {
 372                                 $file = $this->newFile( $redir );
 373                                 if ( $file && $fileMatchesSearch( $file, $search ) ) {
 374                                         $file->redirectedFrom( $title->getDBkey() );
 375                                         if ( $flags & FileRepo::NAME_AND_TIME_ONLY ) {
 376                                                 $finalFiles[$dbKey] = array(
 377                                                         'title' => $file->getTitle()->getDBkey(),
 378                                                         'timestamp' => $file->getTimestamp()
 379                                                 );
 380                                         } else {
 381                                                 $finalFiles[$dbKey] = $file;
 382                                         }
 383                                 }
 384                         }
 385                 }
 386
 387                 return $finalFiles;
 388         }
 389
 390         /**
 391          * Get an array or iterator of file objects for files that have a given
 392          * SHA-1 content hash.
 393          *
 394          * @param string $hash A sha1 hash to look for
 395          * @return File[]
 396          */
 397         function findBySha1( $hash ) {
 398                 $dbr = $this->getSlaveDB();
 399                 $res = $dbr->select(
 400                         'image',
 401                         LocalFile::selectFields(),
 402                         array( 'img_sha1' => $hash ),
 403                         __METHOD__,
 404                         array( 'ORDER BY' => 'img_name' )
 405                 );
 406
 407                 $result = array();
 408                 foreach ( $res as $row ) {
 409                         $result[] = $this->newFileFromRow( $row );
 410                 }
 411                 $res->free();
 412
 413                 return $result;
 414         }
 415
 416         /**
 417          * Get an array of arrays or iterators of file objects for files that
 418          * have the given SHA-1 content hashes.
 419          *
 420          * Overrides generic implementation in FileRepo for performance reason
 421          *
 422          * @param array $hashes An array of hashes
 423          * @return array An Array of arrays or iterators of file objects and the hash as key
 424          */
 425         function findBySha1s( array $hashes ) {
 426                 if ( !count( $hashes ) ) {
 427                         return array(); //empty parameter
 428                 }
 429
 430                 $dbr = $this->getSlaveDB();
 431                 $res = $dbr->select(
 432                         'image',
 433                         LocalFile::selectFields(),
 434                         array( 'img_sha1' => $hashes ),
 435                         __METHOD__,
 436                         array( 'ORDER BY' => 'img_name' )
 437                 );
 438
 439                 $result = array();
 440                 foreach ( $res as $row ) {
 441                         $file = $this->newFileFromRow( $row );
 442                         $result[$file->getSha1()][] = $file;
 443                 }
 444                 $res->free();
 445
 446                 return $result;
 447         }
 448
 449         /**
 450          * Return an array of files where the name starts with $prefix.
 451          *
 452          * @param string $prefix The prefix to search for
 453          * @param int $limit The maximum amount of files to return
 454          * @return array
 455          */
 456         public function findFilesByPrefix( $prefix, $limit ) {
 457                 $selectOptions = array( 'ORDER BY' => 'img_name', 'LIMIT' => intval( $limit ) );
 458
 459                 // Query database
 460                 $dbr = $this->getSlaveDB();
 461                 $res = $dbr->select(
 462                         'image',
 463                         LocalFile::selectFields(),
 464                         'img_name ' . $dbr->buildLike( $prefix, $dbr->anyString() ),
 465                         __METHOD__,
 466                         $selectOptions
 467                 );
 468
 469                 // Build file objects
 470                 $files = array();
 471                 foreach ( $res as $row ) {
 472                         $files[] = $this->newFileFromRow( $row );
 473                 }
 474
 475                 return $files;
 476         }
 477
 478         /**
 479          * Get a connection to the slave DB
 480          * @return DatabaseBase
 481          */
 482         function getSlaveDB() {
 483                 return wfGetDB( DB_SLAVE );
 484         }
 485
 486         /**
 487          * Get a connection to the master DB
 488          * @return DatabaseBase
 489          */
 490         function getMasterDB() {
 491                 return wfGetDB( DB_MASTER );
 492         }
 493
 494         /**
 495          * Get a callback to get a DB handle given an index (DB_SLAVE/DB_MASTER)
 496          * @return Closure
 497          */
 498         protected function getDBFactory() {
 499                 return function( $index ) {
 500                         return wfGetDB( $index );
 501                 };
 502         }
 503
 504         /**
 505          * Get a key on the primary cache for this repository.
 506          * Returns false if the repository's cache is not accessible at this site.
 507          * The parameters are the parts of the key, as for wfMemcKey().
 508          *
 509          * @return string
 510          */
 511         function getSharedCacheKey( /*...*/ ) {
 512                 $args = func_get_args();
 513
 514                 return call_user_func_array( 'wfMemcKey', $args );
 515         }
 516
 517         /**
 518          * Invalidates image redirect cache related to that image
 519          *
 520          * @param Title $title Title of page
 521          * @return void
 522          */
 523         function invalidateImageRedirect( Title $title ) {
 524                 $key = $this->getSharedCacheKey( 'image_redirect', md5( $title->getDBkey() ) );
 525                 if ( $key ) {
 526                         $this->getMasterDB()->onTransactionPreCommitOrIdle( function() use ( $key ) {
 527                                 ObjectCache::getMainWANInstance()->delete( $key );
 528                         } );
 529                 }
 530         }
 531
 532         /**
 533          * Return information about the repository.
 534          *
 535          * @return array
 536          * @since 1.22
 537          */
 538         function getInfo() {
 539                 global $wgFavicon;
 540
 541                 return array_merge( parent::getInfo(), array(
 542                         'favicon' => wfExpandUrl( $wgFavicon ),
 543                 ) );
 544         }
 545
 546         public function store( $srcPath, $dstZone, $dstRel, $flags = 0 ) {
 547                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 548         }
 549
 550         public function storeBatch( array $triplets, $flags = 0 ) {
 551                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 552         }
 553
 554         public function cleanupBatch( array $files, $flags = 0 ) {
 555                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 556         }
 557
 558         public function publish(
 559                 $srcPath,
 560                 $dstRel,
 561                 $archiveRel,
 562                 $flags = 0,
 563                 array $options = array()
 564         ) {
 565                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 566         }
 567
 568         public function publishBatch( array $ntuples, $flags = 0 ) {
 569                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 570         }
 571
 572         public function delete( $srcRel, $archiveRel ) {
 573                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 574         }
 575
 576         public function deleteBatch( array $sourceDestPairs ) {
 577                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 578         }
 579
 580         /**
 581          * Skips the write operation if storage is sha1-based, executes it normally otherwise
 582          *
 583          * @param string $function
 584          * @param array $args
 585          * @return FileRepoStatus
 586          */
 587         protected function skipWriteOperationIfSha1( $function, array $args ) {
 588                 $this->assertWritableRepo(); // fail out if read-only
 589
 590                 if ( $this->hasSha1Storage() ) {
 591                         wfDebug( __METHOD__ . ": skipped because storage uses sha1 paths\n" );
 592                         return Status::newGood();
 593                 } else {
 594                         return call_user_func_array( 'parent::' . $function, $args );
 595                 }
 596         }
 597 }