includes/filerepo/LocalRepo.php

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