includes/cache/LinkCache.php

   1 <?php
   2 /**
   3  * Page existence cache.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Cache
  22  */
  23
  24 use Wikimedia\Rdbms\Database;
  25 use Wikimedia\Rdbms\IDatabase;
  26 use MediaWiki\Linker\LinkTarget;
  27 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * Cache for article titles (prefixed DB keys) and ids linked from one source
  31  *
  32  * @ingroup Cache
  33  */
  34 class LinkCache {
  35         /** @var MapCacheLRU */
  36         private $goodLinks;
  37         /** @var MapCacheLRU */
  38         private $badLinks;
  39         /** @var WANObjectCache */
  40         private $wanCache;
  41
  42         /** @var bool */
  43         private $mForUpdate = false;
  44
  45         /** @var TitleFormatter */
  46         private $titleFormatter;
  47
  48         /**
  49          * How many Titles to store. There are two caches, so the amount actually
  50          * stored in memory can be up to twice this.
  51          */
  52         const MAX_SIZE = 10000;
  53
  54         public function __construct( TitleFormatter $titleFormatter, WANObjectCache $cache ) {
  55                 $this->goodLinks = new MapCacheLRU( self::MAX_SIZE );
  56                 $this->badLinks = new MapCacheLRU( self::MAX_SIZE );
  57                 $this->wanCache = $cache;
  58                 $this->titleFormatter = $titleFormatter;
  59         }
  60
  61         /**
  62          * Get an instance of this class.
  63          *
  64          * @return LinkCache
  65          * @deprecated since 1.28, use MediaWikiServices instead
  66          */
  67         public static function singleton() {
  68                 return MediaWikiServices::getInstance()->getLinkCache();
  69         }
  70
  71         /**
  72          * General accessor to get/set whether the master DB should be used
  73          *
  74          * This used to also set the FOR UPDATE option (locking the rows read
  75          * in order to avoid link table inconsistency), which was later removed
  76          * for performance on wikis with a high edit rate.
  77          *
  78          * @param bool|null $update
  79          * @return bool
  80          */
  81         public function forUpdate( $update = null ) {
  82                 return wfSetVar( $this->mForUpdate, $update );
  83         }
  84
  85         /**
  86          * @param string $title Prefixed DB key
  87          * @return int Page ID or zero
  88          */
  89         public function getGoodLinkID( $title ) {
  90                 $info = $this->goodLinks->get( $title );
  91                 if ( !$info ) {
  92                         return 0;
  93                 }
  94                 return $info['id'];
  95         }
  96
  97         /**
  98          * Get a field of a title object from cache.
  99          * If this link is not a cached good title, it will return NULL.
 100          * @param LinkTarget $target
 101          * @param string $field ('length','redirect','revision','model')
 102          * @return string|int|null
 103          */
 104         public function getGoodLinkFieldObj( LinkTarget $target, $field ) {
 105                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 106                 $info = $this->goodLinks->get( $dbkey );
 107                 if ( !$info ) {
 108                         return null;
 109                 }
 110                 return $info[$field];
 111         }
 112
 113         /**
 114          * @param string $title Prefixed DB key
 115          * @return bool
 116          */
 117         public function isBadLink( $title ) {
 118                 // Use get() to ensure it records as used for LRU.
 119                 return $this->badLinks->has( $title );
 120         }
 121
 122         /**
 123          * Add a link for the title to the link cache
 124          *
 125          * @param int $id Page's ID
 126          * @param LinkTarget $target
 127          * @param int $len Text's length
 128          * @param int|null $redir Whether the page is a redirect
 129          * @param int $revision Latest revision's ID
 130          * @param string|null $model Latest revision's content model ID
 131          * @param string|null $lang Language code of the page, if not the content language
 132          */
 133         public function addGoodLinkObj( $id, LinkTarget $target, $len = -1, $redir = null,
 134                 $revision = 0, $model = null, $lang = null
 135         ) {
 136                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 137                 $this->goodLinks->set( $dbkey, [
 138                         'id' => (int)$id,
 139                         'length' => (int)$len,
 140                         'redirect' => (int)$redir,
 141                         'revision' => (int)$revision,
 142                         'model' => $model ? (string)$model : null,
 143                         'lang' => $lang ? (string)$lang : null,
 144                         'restrictions' => null
 145                 ] );
 146         }
 147
 148         /**
 149          * Same as above with better interface.
 150          * @since 1.19
 151          * @param LinkTarget $target
 152          * @param stdClass $row Object which has the fields page_id, page_is_redirect,
 153          *  page_latest and page_content_model
 154          */
 155         public function addGoodLinkObjFromRow( LinkTarget $target, $row ) {
 156                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 157                 $this->goodLinks->set( $dbkey, [
 158                         'id' => intval( $row->page_id ),
 159                         'length' => intval( $row->page_len ),
 160                         'redirect' => intval( $row->page_is_redirect ),
 161                         'revision' => intval( $row->page_latest ),
 162                         'model' => !empty( $row->page_content_model )
 163                                 ? strval( $row->page_content_model )
 164                                 : null,
 165                         'lang' => !empty( $row->page_lang )
 166                                 ? strval( $row->page_lang )
 167                                 : null,
 168                         'restrictions' => !empty( $row->page_restrictions )
 169                                 ? strval( $row->page_restrictions )
 170                                 : null
 171                 ] );
 172         }
 173
 174         /**
 175          * @param LinkTarget $target
 176          */
 177         public function addBadLinkObj( LinkTarget $target ) {
 178                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 179                 if ( !$this->isBadLink( $dbkey ) ) {
 180                         $this->badLinks->set( $dbkey, 1 );
 181                 }
 182         }
 183
 184         /**
 185          * @param string $title Prefixed DB key
 186          */
 187         public function clearBadLink( $title ) {
 188                 $this->badLinks->clear( $title );
 189         }
 190
 191         /**
 192          * @param LinkTarget $target
 193          */
 194         public function clearLink( LinkTarget $target ) {
 195                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 196                 $this->badLinks->clear( $dbkey );
 197                 $this->goodLinks->clear( $dbkey );
 198         }
 199
 200         /**
 201          * Fields that LinkCache needs to select
 202          *
 203          * @since 1.28
 204          * @return array
 205          */
 206         public static function getSelectFields() {
 207                 global $wgContentHandlerUseDB, $wgPageLanguageUseDB;
 208
 209                 $fields = [
 210                         'page_id',
 211                         'page_len',
 212                         'page_is_redirect',
 213                         'page_latest',
 214                         'page_restrictions'
 215                 ];
 216                 if ( $wgContentHandlerUseDB ) {
 217                         $fields[] = 'page_content_model';
 218                 }
 219                 if ( $wgPageLanguageUseDB ) {
 220                         $fields[] = 'page_lang';
 221                 }
 222
 223                 return $fields;
 224         }
 225
 226         /**
 227          * Add a title to the link cache, return the page_id or zero if non-existent
 228          *
 229          * @param LinkTarget $nt LinkTarget object to add
 230          * @return int Page ID or zero
 231          */
 232         public function addLinkObj( LinkTarget $nt ) {
 233                 $key = $this->titleFormatter->getPrefixedDBkey( $nt );
 234                 if ( $this->isBadLink( $key ) || $nt->isExternal()
 235                         || $nt->inNamespace( NS_SPECIAL )
 236                 ) {
 237                         return 0;
 238                 }
 239                 $id = $this->getGoodLinkID( $key );
 240                 if ( $id != 0 ) {
 241                         return $id;
 242                 }
 243
 244                 if ( $key === '' ) {
 245                         return 0;
 246                 }
 247
 248                 // Cache template/file pages as they are less often viewed but heavily used
 249                 if ( $this->mForUpdate ) {
 250                         $row = $this->fetchPageRow( wfGetDB( DB_MASTER ), $nt );
 251                 } elseif ( $this->isCacheable( $nt ) ) {
 252                         // These pages are often transcluded heavily, so cache them
 253                         $cache = $this->wanCache;
 254                         $row = $cache->getWithSetCallback(
 255                                 $cache->makeKey( 'page', $nt->getNamespace(), sha1( $nt->getDBkey() ) ),
 256                                 $cache::TTL_DAY,
 257                                 function ( $curValue, &$ttl, array &$setOpts ) use ( $cache, $nt ) {
 258                                         $dbr = wfGetDB( DB_REPLICA );
 259                                         $setOpts += Database::getCacheSetOptions( $dbr );
 260
 261                                         $row = $this->fetchPageRow( $dbr, $nt );
 262                                         $mtime = $row ? wfTimestamp( TS_UNIX, $row->page_touched ) : false;
 263                                         $ttl = $cache->adaptiveTTL( $mtime, $ttl );
 264
 265                                         return $row;
 266                                 }
 267                         );
 268                 } else {
 269                         $row = $this->fetchPageRow( wfGetDB( DB_REPLICA ), $nt );
 270                 }
 271
 272                 if ( $row ) {
 273                         $this->addGoodLinkObjFromRow( $nt, $row );
 274                         $id = intval( $row->page_id );
 275                 } else {
 276                         $this->addBadLinkObj( $nt );
 277                         $id = 0;
 278                 }
 279
 280                 return $id;
 281         }
 282
 283         /**
 284          * @param WANObjectCache $cache
 285          * @param LinkTarget $t
 286          * @return string[]
 287          * @since 1.28
 288          */
 289         public function getMutableCacheKeys( WANObjectCache $cache, LinkTarget $t ) {
 290                 if ( $this->isCacheable( $t ) ) {
 291                         return [ $cache->makeKey( 'page', $t->getNamespace(), sha1( $t->getDBkey() ) ) ];
 292                 }
 293
 294                 return [];
 295         }
 296
 297         private function isCacheable( LinkTarget $title ) {
 298                 $ns = $title->getNamespace();
 299                 if ( in_array( $ns, [ NS_TEMPLATE, NS_FILE, NS_CATEGORY ] ) ) {
 300                         return true;
 301                 }
 302                 // Focus on transcluded pages more than the main content
 303                 if ( MWNamespace::isContent( $ns ) ) {
 304                         return false;
 305                 }
 306                 // Non-talk extension namespaces (e.g. NS_MODULE)
 307                 return ( $ns >= 100 && MWNamespace::isSubject( $ns ) );
 308         }
 309
 310         private function fetchPageRow( IDatabase $db, LinkTarget $nt ) {
 311                 $fields = self::getSelectFields();
 312                 if ( $this->isCacheable( $nt ) ) {
 313                         $fields[] = 'page_touched';
 314                 }
 315
 316                 return $db->selectRow(
 317                         'page',
 318                         $fields,
 319                         [ 'page_namespace' => $nt->getNamespace(), 'page_title' => $nt->getDBkey() ],
 320                         __METHOD__
 321                 );
 322         }
 323
 324         /**
 325          * Purge the link cache for a title
 326          *
 327          * @param LinkTarget $title
 328          * @since 1.28
 329          */
 330         public function invalidateTitle( LinkTarget $title ) {
 331                 if ( $this->isCacheable( $title ) ) {
 332                         $cache = $this->wanCache;
 333                         $cache->delete(
 334                                 $cache->makeKey( 'page', $title->getNamespace(), sha1( $title->getDBkey() ) )
 335                         );
 336                 }
 337         }
 338
 339         /**
 340          * Clears cache
 341          */
 342         public function clear() {
 343                 $this->goodLinks->clear();
 344                 $this->badLinks->clear();
 345         }
 346 }