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