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