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 use MediaWiki\MediaWikiServices;
  26 use Wikimedia\Rdbms\ResultWrapper;
  27 use Wikimedia\Rdbms\Database;
  28 use Wikimedia\Rdbms\IDatabase;
  29
  30 /**
  31  * A repository that stores files in the local filesystem and registers them
  32  * in the wiki's own database. This is the most commonly used repository class.
  33  *
  34  * @ingroup FileRepo
  35  */
  36 class LocalRepo extends FileRepo {
  37         /** @var callable */
  38         protected $fileFactory = [ LocalFile::class, 'newFromTitle' ];
  39         /** @var callable */
  40         protected $fileFactoryKey = [ LocalFile::class, 'newFromKey' ];
  41         /** @var callable */
  42         protected $fileFromRowFactory = [ LocalFile::class, 'newFromRow' ];
  43         /** @var callable */
  44         protected $oldFileFromRowFactory = [ OldLocalFile::class, 'newFromRow' ];
  45         /** @var callable */
  46         protected $oldFileFactory = [ OldLocalFile::class, 'newFromTitle' ];
  47         /** @var callable */
  48         protected $oldFileFactoryKey = [ OldLocalFile::class, 'newFromKey' ];
  49
  50         function __construct( array $info = null ) {
  51                 parent::__construct( $info );
  52
  53                 $this->hasSha1Storage = isset( $info['storageLayout'] )
  54                         && $info['storageLayout'] === 'sha1';
  55
  56                 if ( $this->hasSha1Storage() ) {
  57                         $this->backend = new FileBackendDBRepoWrapper( [
  58                                 'backend'         => $this->backend,
  59                                 'repoName'        => $this->name,
  60                                 'dbHandleFactory' => $this->getDBFactory()
  61                         ] );
  62                 }
  63         }
  64
  65         /**
  66          * @throws MWException
  67          * @param stdClass $row
  68          * @return LocalFile
  69          */
  70         function newFileFromRow( $row ) {
  71                 if ( isset( $row->img_name ) ) {
  72                         return call_user_func( $this->fileFromRowFactory, $row, $this );
  73                 } elseif ( isset( $row->oi_name ) ) {
  74                         return call_user_func( $this->oldFileFromRowFactory, $row, $this );
  75                 } else {
  76                         throw new MWException( __METHOD__ . ': invalid row' );
  77                 }
  78         }
  79
  80         /**
  81          * @param Title $title
  82          * @param string $archiveName
  83          * @return OldLocalFile
  84          */
  85         function newFromArchiveName( $title, $archiveName ) {
  86                 return OldLocalFile::newFromArchiveName( $title, $this, $archiveName );
  87         }
  88
  89         /**
  90          * Delete files in the deleted directory if they are not referenced in the
  91          * filearchive table. This needs to be done in the repo because it needs to
  92          * interleave database locks with file operations, which is potentially a
  93          * remote operation.
  94          *
  95          * @param string[] $storageKeys
  96          *
  97          * @return Status
  98          */
  99         function cleanupDeletedBatch( array $storageKeys ) {
 100                 if ( $this->hasSha1Storage() ) {
 101                         wfDebug( __METHOD__ . ": skipped because storage uses sha1 paths\n" );
 102                         return Status::newGood();
 103                 }
 104
 105                 $backend = $this->backend; // convenience
 106                 $root = $this->getZonePath( 'deleted' );
 107                 $dbw = $this->getMasterDB();
 108                 $status = $this->newGood();
 109                 $storageKeys = array_unique( $storageKeys );
 110                 foreach ( $storageKeys as $key ) {
 111                         $hashPath = $this->getDeletedHashPath( $key );
 112                         $path = "$root/$hashPath$key";
 113                         $dbw->startAtomic( __METHOD__ );
 114                         // Check for usage in deleted/hidden files and preemptively
 115                         // lock the key to avoid any future use until we are finished.
 116                         $deleted = $this->deletedFileHasKey( $key, 'lock' );
 117                         $hidden = $this->hiddenFileHasKey( $key, 'lock' );
 118                         if ( !$deleted && !$hidden ) { // not in use now
 119                                 wfDebug( __METHOD__ . ": deleting $key\n" );
 120                                 $op = [ 'op' => 'delete', 'src' => $path ];
 121                                 if ( !$backend->doOperation( $op )->isOK() ) {
 122                                         $status->error( 'undelete-cleanup-error', $path );
 123                                         $status->failCount++;
 124                                 }
 125                         } else {
 126                                 wfDebug( __METHOD__ . ": $key still in use\n" );
 127                                 $status->successCount++;
 128                         }
 129                         $dbw->endAtomic( __METHOD__ );
 130                 }
 131
 132                 return $status;
 133         }
 134
 135         /**
 136          * Check if a deleted (filearchive) file has this sha1 key
 137          *
 138          * @param string $key File storage key (base-36 sha1 key with file extension)
 139          * @param string|null $lock Use "lock" to lock the row via FOR UPDATE
 140          * @return bool File with this key is in use
 141          */
 142         protected function deletedFileHasKey( $key, $lock = null ) {
 143                 $options = ( $lock === 'lock' ) ? [ 'FOR UPDATE' ] : [];
 144
 145                 $dbw = $this->getMasterDB();
 146
 147                 return (bool)$dbw->selectField( 'filearchive', '1',
 148                         [ 'fa_storage_group' => 'deleted', 'fa_storage_key' => $key ],
 149                         __METHOD__, $options
 150                 );
 151         }
 152
 153         /**
 154          * Check if a hidden (revision delete) file has this sha1 key
 155          *
 156          * @param string $key File storage key (base-36 sha1 key with file extension)
 157          * @param string|null $lock Use "lock" to lock the row via FOR UPDATE
 158          * @return bool File with this key is in use
 159          */
 160         protected function hiddenFileHasKey( $key, $lock = null ) {
 161                 $options = ( $lock === 'lock' ) ? [ 'FOR UPDATE' ] : [];
 162
 163                 $sha1 = self::getHashFromKey( $key );
 164                 $ext = File::normalizeExtension( substr( $key, strcspn( $key, '.' ) + 1 ) );
 165
 166                 $dbw = $this->getMasterDB();
 167
 168                 return (bool)$dbw->selectField( 'oldimage', '1',
 169                         [ 'oi_sha1' => $sha1,
 170                                 'oi_archive_name ' . $dbw->buildLike( $dbw->anyString(), ".$ext" ),
 171                                 $dbw->bitAnd( 'oi_deleted', File::DELETED_FILE ) => File::DELETED_FILE ],
 172                         __METHOD__, $options
 173                 );
 174         }
 175
 176         /**
 177          * Gets the SHA1 hash from a storage key
 178          *
 179          * @param string $key
 180          * @return string
 181          */
 182         public static function getHashFromKey( $key ) {
 183                 return strtok( $key, '.' );
 184         }
 185
 186         /**
 187          * Checks if there is a redirect named as $title
 188          *
 189          * @param Title $title Title of file
 190          * @return bool|Title
 191          */
 192         function checkRedirect( Title $title ) {
 193                 $title = File::normalizeTitle( $title, 'exception' );
 194
 195                 $memcKey = $this->getSharedCacheKey( 'image_redirect', md5( $title->getDBkey() ) );
 196                 if ( $memcKey === false ) {
 197                         $memcKey = $this->getLocalCacheKey( 'image_redirect', md5( $title->getDBkey() ) );
 198                         $expiry = 300; // no invalidation, 5 minutes
 199                 } else {
 200                         $expiry = 86400; // has invalidation, 1 day
 201                 }
 202
 203                 $method = __METHOD__;
 204                 $redirDbKey = MediaWikiServices::getInstance()->getMainWANObjectCache()->getWithSetCallback(
 205                         $memcKey,
 206                         $expiry,
 207                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $method, $title ) {
 208                                 $dbr = $this->getReplicaDB(); // possibly remote DB
 209
 210                                 $setOpts += Database::getCacheSetOptions( $dbr );
 211
 212                                 if ( $title instanceof Title ) {
 213                                         $row = $dbr->selectRow(
 214                                                 [ 'page', 'redirect' ],
 215                                                 [ 'rd_namespace', 'rd_title' ],
 216                                                 [
 217                                                         'page_namespace' => $title->getNamespace(),
 218                                                         'page_title' => $title->getDBkey(),
 219                                                         'rd_from = page_id'
 220                                                 ],
 221                                                 $method
 222                                         );
 223                                 } else {
 224                                         $row = false;
 225                                 }
 226
 227                                 return ( $row && $row->rd_namespace == NS_FILE )
 228                                         ? Title::makeTitle( $row->rd_namespace, $row->rd_title )->getDBkey()
 229                                         : ''; // negative cache
 230                         },
 231                         [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 232                 );
 233
 234                 // @note: also checks " " for b/c
 235                 if ( $redirDbKey !== ' ' && strval( $redirDbKey ) !== '' ) {
 236                         // Page is a redirect to another file
 237                         return Title::newFromText( $redirDbKey, NS_FILE );
 238                 }
 239
 240                 return false; // no redirect
 241         }
 242
 243         public function findFiles( array $items, $flags = 0 ) {
 244                 $finalFiles = []; // map of (DB key => corresponding File) for matches
 245
 246                 $searchSet = []; // map of (normalized DB key => search params)
 247                 foreach ( $items as $item ) {
 248                         if ( is_array( $item ) ) {
 249                                 $title = File::normalizeTitle( $item['title'] );
 250                                 if ( $title ) {
 251                                         $searchSet[$title->getDBkey()] = $item;
 252                                 }
 253                         } else {
 254                                 $title = File::normalizeTitle( $item );
 255                                 if ( $title ) {
 256                                         $searchSet[$title->getDBkey()] = [];
 257                                 }
 258                         }
 259                 }
 260
 261                 $fileMatchesSearch = function ( File $file, array $search ) {
 262                         // Note: file name comparison done elsewhere (to handle redirects)
 263                         $user = ( !empty( $search['private'] ) && $search['private'] instanceof User )
 264                                 ? $search['private']
 265                                 : null;
 266
 267                         return (
 268                                 $file->exists() &&
 269                                 (
 270                                         ( empty( $search['time'] ) && !$file->isOld() ) ||
 271                                         ( !empty( $search['time'] ) && $search['time'] === $file->getTimestamp() )
 272                                 ) &&
 273                                 ( !empty( $search['private'] ) || !$file->isDeleted( File::DELETED_FILE ) ) &&
 274                                 $file->userCan( File::DELETED_FILE, $user )
 275                         );
 276                 };
 277
 278                 $applyMatchingFiles = function ( ResultWrapper $res, &$searchSet, &$finalFiles )
 279                         use ( $fileMatchesSearch, $flags )
 280                 {
 281                         $contLang = MediaWikiServices::getInstance()->getContentLanguage();
 282                         $info = $this->getInfo();
 283                         foreach ( $res as $row ) {
 284                                 $file = $this->newFileFromRow( $row );
 285                                 // There must have been a search for this DB key, but this has to handle the
 286                                 // cases were title capitalization is different on the client and repo wikis.
 287                                 $dbKeysLook = [ strtr( $file->getName(), ' ', '_' ) ];
 288                                 if ( !empty( $info['initialCapital'] ) ) {
 289                                         // Search keys for "hi.png" and "Hi.png" should use the "Hi.png file"
 290                                         $dbKeysLook[] = $contLang->lcfirst( $file->getName() );
 291                                 }
 292                                 foreach ( $dbKeysLook as $dbKey ) {
 293                                         if ( isset( $searchSet[$dbKey] )
 294                                                 && $fileMatchesSearch( $file, $searchSet[$dbKey] )
 295                                         ) {
 296                                                 $finalFiles[$dbKey] = ( $flags & FileRepo::NAME_AND_TIME_ONLY )
 297                                                         ? [ 'title' => $dbKey, 'timestamp' => $file->getTimestamp() ]
 298                                                         : $file;
 299                                                 unset( $searchSet[$dbKey] );
 300                                         }
 301                                 }
 302                         }
 303                 };
 304
 305                 $dbr = $this->getReplicaDB();
 306
 307                 // Query image table
 308                 $imgNames = [];
 309                 foreach ( array_keys( $searchSet ) as $dbKey ) {
 310                         $imgNames[] = $this->getNameFromTitle( File::normalizeTitle( $dbKey ) );
 311                 }
 312
 313                 if ( count( $imgNames ) ) {
 314                         $fileQuery = LocalFile::getQueryInfo();
 315                         $res = $dbr->select( $fileQuery['tables'], $fileQuery['fields'], [ 'img_name' => $imgNames ],
 316                                 __METHOD__, [], $fileQuery['joins'] );
 317                         $applyMatchingFiles( $res, $searchSet, $finalFiles );
 318                 }
 319
 320                 // Query old image table
 321                 $oiConds = []; // WHERE clause array for each file
 322                 foreach ( $searchSet as $dbKey => $search ) {
 323                         if ( isset( $search['time'] ) ) {
 324                                 $oiConds[] = $dbr->makeList(
 325                                         [
 326                                                 'oi_name' => $this->getNameFromTitle( File::normalizeTitle( $dbKey ) ),
 327                                                 'oi_timestamp' => $dbr->timestamp( $search['time'] )
 328                                         ],
 329                                         LIST_AND
 330                                 );
 331                         }
 332                 }
 333
 334                 if ( count( $oiConds ) ) {
 335                         $fileQuery = OldLocalFile::getQueryInfo();
 336                         $res = $dbr->select( $fileQuery['tables'], $fileQuery['fields'],
 337                                 $dbr->makeList( $oiConds, LIST_OR ),
 338                                 __METHOD__, [], $fileQuery['joins'] );
 339                         $applyMatchingFiles( $res, $searchSet, $finalFiles );
 340                 }
 341
 342                 // Check for redirects...
 343                 foreach ( $searchSet as $dbKey => $search ) {
 344                         if ( !empty( $search['ignoreRedirect'] ) ) {
 345                                 continue;
 346                         }
 347
 348                         $title = File::normalizeTitle( $dbKey );
 349                         $redir = $this->checkRedirect( $title ); // hopefully hits memcached
 350
 351                         if ( $redir && $redir->getNamespace() == NS_FILE ) {
 352                                 $file = $this->newFile( $redir );
 353                                 if ( $file && $fileMatchesSearch( $file, $search ) ) {
 354                                         $file->redirectedFrom( $title->getDBkey() );
 355                                         if ( $flags & FileRepo::NAME_AND_TIME_ONLY ) {
 356                                                 $finalFiles[$dbKey] = [
 357                                                         'title' => $file->getTitle()->getDBkey(),
 358                                                         'timestamp' => $file->getTimestamp()
 359                                                 ];
 360                                         } else {
 361                                                 $finalFiles[$dbKey] = $file;
 362                                         }
 363                                 }
 364                         }
 365                 }
 366
 367                 return $finalFiles;
 368         }
 369
 370         /**
 371          * Get an array or iterator of file objects for files that have a given
 372          * SHA-1 content hash.
 373          *
 374          * @param string $hash A sha1 hash to look for
 375          * @return LocalFile[]
 376          */
 377         function findBySha1( $hash ) {
 378                 $dbr = $this->getReplicaDB();
 379                 $fileQuery = LocalFile::getQueryInfo();
 380                 $res = $dbr->select(
 381                         $fileQuery['tables'],
 382                         $fileQuery['fields'],
 383                         [ 'img_sha1' => $hash ],
 384                         __METHOD__,
 385                         [ 'ORDER BY' => 'img_name' ],
 386                         $fileQuery['joins']
 387                 );
 388
 389                 $result = [];
 390                 foreach ( $res as $row ) {
 391                         $result[] = $this->newFileFromRow( $row );
 392                 }
 393                 $res->free();
 394
 395                 return $result;
 396         }
 397
 398         /**
 399          * Get an array of arrays or iterators of file objects for files that
 400          * have the given SHA-1 content hashes.
 401          *
 402          * Overrides generic implementation in FileRepo for performance reason
 403          *
 404          * @param string[] $hashes An array of hashes
 405          * @return array[] An Array of arrays or iterators of file objects and the hash as key
 406          */
 407         function findBySha1s( array $hashes ) {
 408                 if ( !count( $hashes ) ) {
 409                         return []; // empty parameter
 410                 }
 411
 412                 $dbr = $this->getReplicaDB();
 413                 $fileQuery = LocalFile::getQueryInfo();
 414                 $res = $dbr->select(
 415                         $fileQuery['tables'],
 416                         $fileQuery['fields'],
 417                         [ 'img_sha1' => $hashes ],
 418                         __METHOD__,
 419                         [ 'ORDER BY' => 'img_name' ],
 420                         $fileQuery['joins']
 421                 );
 422
 423                 $result = [];
 424                 foreach ( $res as $row ) {
 425                         $file = $this->newFileFromRow( $row );
 426                         $result[$file->getSha1()][] = $file;
 427                 }
 428                 $res->free();
 429
 430                 return $result;
 431         }
 432
 433         /**
 434          * Return an array of files where the name starts with $prefix.
 435          *
 436          * @param string $prefix The prefix to search for
 437          * @param int $limit The maximum amount of files to return
 438          * @return LocalFile[]
 439          */
 440         public function findFilesByPrefix( $prefix, $limit ) {
 441                 $selectOptions = [ 'ORDER BY' => 'img_name', 'LIMIT' => intval( $limit ) ];
 442
 443                 // Query database
 444                 $dbr = $this->getReplicaDB();
 445                 $fileQuery = LocalFile::getQueryInfo();
 446                 $res = $dbr->select(
 447                         $fileQuery['tables'],
 448                         $fileQuery['fields'],
 449                         'img_name ' . $dbr->buildLike( $prefix, $dbr->anyString() ),
 450                         __METHOD__,
 451                         $selectOptions,
 452                         $fileQuery['joins']
 453                 );
 454
 455                 // Build file objects
 456                 $files = [];
 457                 foreach ( $res as $row ) {
 458                         $files[] = $this->newFileFromRow( $row );
 459                 }
 460
 461                 return $files;
 462         }
 463
 464         /**
 465          * Get a connection to the replica DB
 466          * @return IDatabase
 467          */
 468         function getReplicaDB() {
 469                 return wfGetDB( DB_REPLICA );
 470         }
 471
 472         /**
 473          * Alias for getReplicaDB()
 474          *
 475          * @return IDatabase
 476          * @deprecated Since 1.29
 477          */
 478         function getSlaveDB() {
 479                 return $this->getReplicaDB();
 480         }
 481
 482         /**
 483          * Get a connection to the master DB
 484          * @return IDatabase
 485          */
 486         function getMasterDB() {
 487                 return wfGetDB( DB_MASTER );
 488         }
 489
 490         /**
 491          * Get a callback to get a DB handle given an index (DB_REPLICA/DB_MASTER)
 492          * @return Closure
 493          */
 494         protected function getDBFactory() {
 495                 return function ( $index ) {
 496                         return wfGetDB( $index );
 497                 };
 498         }
 499
 500         /**
 501          * Get a key on the primary cache for this repository.
 502          * Returns false if the repository's cache is not accessible at this site.
 503          * The parameters are the parts of the key, as for wfMemcKey().
 504          *
 505          * @return string
 506          */
 507         function getSharedCacheKey( /*...*/ ) {
 508                 $args = func_get_args();
 509
 510                 return wfMemcKey( ...$args );
 511         }
 512
 513         /**
 514          * Invalidates image redirect cache related to that image
 515          *
 516          * @param Title $title Title of page
 517          * @return void
 518          */
 519         function invalidateImageRedirect( Title $title ) {
 520                 $key = $this->getSharedCacheKey( 'image_redirect', md5( $title->getDBkey() ) );
 521                 if ( $key ) {
 522                         $this->getMasterDB()->onTransactionPreCommitOrIdle(
 523                                 function () use ( $key ) {
 524                                         MediaWikiServices::getInstance()->getMainWANObjectCache()->delete( $key );
 525                                 },
 526                                 __METHOD__
 527                         );
 528                 }
 529         }
 530
 531         /**
 532          * Return information about the repository.
 533          *
 534          * @return array
 535          * @since 1.22
 536          */
 537         function getInfo() {
 538                 global $wgFavicon;
 539
 540                 return array_merge( parent::getInfo(), [
 541                         'favicon' => wfExpandUrl( $wgFavicon ),
 542                 ] );
 543         }
 544
 545         public function store( $srcPath, $dstZone, $dstRel, $flags = 0 ) {
 546                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 547         }
 548
 549         public function storeBatch( array $triplets, $flags = 0 ) {
 550                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 551         }
 552
 553         public function cleanupBatch( array $files, $flags = 0 ) {
 554                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 555         }
 556
 557         public function publish(
 558                 $src,
 559                 $dstRel,
 560                 $archiveRel,
 561                 $flags = 0,
 562                 array $options = []
 563         ) {
 564                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 565         }
 566
 567         public function publishBatch( array $ntuples, $flags = 0 ) {
 568                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 569         }
 570
 571         public function delete( $srcRel, $archiveRel ) {
 572                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 573         }
 574
 575         public function deleteBatch( array $sourceDestPairs ) {
 576                 return $this->skipWriteOperationIfSha1( __FUNCTION__, func_get_args() );
 577         }
 578
 579         /**
 580          * Skips the write operation if storage is sha1-based, executes it normally otherwise
 581          *
 582          * @param string $function
 583          * @param array $args
 584          * @return Status
 585          */
 586         protected function skipWriteOperationIfSha1( $function, array $args ) {
 587                 $this->assertWritableRepo(); // fail out if read-only
 588
 589                 if ( $this->hasSha1Storage() ) {
 590                         wfDebug( __METHOD__ . ": skipped because storage uses sha1 paths\n" );
 591                         return Status::newGood();
 592                 } else {
 593                         return parent::$function( ...$args );
 594                 }
 595         }
 596 }