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