includes/jobqueue/jobs/CategoryMembershipChangeJob.php

   1 <?php
   2 /**
   3  * Updater for link tracking tables after a page edit.
   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  */
  22 use MediaWiki\MediaWikiServices;
  23 use Wikimedia\Rdbms\LBFactory;
  24
  25 /**
  26  * Job to add recent change entries mentioning category membership changes
  27  *
  28  * This allows users to easily scan categories for recent page membership changes
  29  *
  30  * Parameters include:
  31  *   - pageId : page ID
  32  *   - revTimestamp : timestamp of the triggering revision
  33  *
  34  * Category changes will be mentioned for revisions at/after the timestamp for this page
  35  *
  36  * @since 1.27
  37  */
  38 class CategoryMembershipChangeJob extends Job {
  39         /** @var int|null */
  40         private $ticket;
  41
  42         const ENQUEUE_FUDGE_SEC = 60;
  43
  44         /**
  45          * @var ParserCache
  46          */
  47         private $parserCache;
  48
  49         /**
  50          * @param Title $title The title of the page for which to update category emmbership.
  51          * @param string $revisionTimestamp The timestamp of the new revision that triggered the job.
  52          * @return JobSpecification
  53          */
  54         public static function newSpec( Title $title, $revisionTimestamp ) {
  55                 return new JobSpecification(
  56                         'categoryMembershipChange',
  57                         [
  58                                 'pageId' => $title->getArticleID(),
  59                                 'revTimestamp' => $revisionTimestamp,
  60                         ],
  61                         [],
  62                         $title
  63                 );
  64         }
  65
  66         /**
  67          * Constructor for use by the Job Queue infrastructure.
  68          * @note Don't call this when queueing a new instance, use newSpec() instead.
  69          */
  70         public function __construct( ParserCache $parserCache, Title $title, array $params ) {
  71                 parent::__construct( 'categoryMembershipChange', $title, $params );
  72                 // Only need one job per page. Note that ENQUEUE_FUDGE_SEC handles races where an
  73                 // older revision job gets inserted while the newer revision job is de-duplicated.
  74                 $this->removeDuplicates = true;
  75                 $this->parserCache = $parserCache;
  76         }
  77
  78         public function run() {
  79                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  80                 $lb = $lbFactory->getMainLB();
  81                 $dbw = $lb->getConnection( DB_MASTER );
  82
  83                 $this->ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
  84
  85                 $page = WikiPage::newFromID( $this->params['pageId'], WikiPage::READ_LATEST );
  86                 if ( !$page ) {
  87                         $this->setLastError( "Could not find page #{$this->params['pageId']}" );
  88                         return false; // deleted?
  89                 }
  90
  91                 // Cut down on the time spent in safeWaitForMasterPos() in the critical section
  92                 $dbr = $lb->getConnection( DB_REPLICA, [ 'recentchanges' ] );
  93                 if ( !$lb->safeWaitForMasterPos( $dbr ) ) {
  94                         $this->setLastError( "Timed out while pre-waiting for replica DB to catch up" );
  95                         return false;
  96                 }
  97
  98                 // Use a named lock so that jobs for this page see each others' changes
  99                 $lockKey = "CategoryMembershipUpdates:{$page->getId()}";
 100                 $scopedLock = $dbw->getScopedLockAndFlush( $lockKey, __METHOD__, 3 );
 101                 if ( !$scopedLock ) {
 102                         $this->setLastError( "Could not acquire lock '$lockKey'" );
 103                         return false;
 104                 }
 105
 106                 // Wait till replica DB is caught up so that jobs for this page see each others' changes
 107                 if ( !$lb->safeWaitForMasterPos( $dbr ) ) {
 108                         $this->setLastError( "Timed out while waiting for replica DB to catch up" );
 109                         return false;
 110                 }
 111                 // Clear any stale REPEATABLE-READ snapshot
 112                 $dbr->flushSnapshot( __METHOD__ );
 113
 114                 $cutoffUnix = wfTimestamp( TS_UNIX, $this->params['revTimestamp'] );
 115                 // Using ENQUEUE_FUDGE_SEC handles jobs inserted out of revision order due to the delay
 116                 // between COMMIT and actual enqueueing of the CategoryMembershipChangeJob job.
 117                 $cutoffUnix -= self::ENQUEUE_FUDGE_SEC;
 118
 119                 // Get the newest page revision that has a SRC_CATEGORIZE row.
 120                 // Assume that category changes before it were already handled.
 121                 $row = $dbr->selectRow(
 122                         'revision',
 123                         [ 'rev_timestamp', 'rev_id' ],
 124                         [
 125                                 'rev_page' => $page->getId(),
 126                                 'rev_timestamp >= ' . $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) ),
 127                                 'EXISTS (' . $dbr->selectSQLText(
 128                                         'recentchanges',
 129                                         '1',
 130                                         [
 131                                                 'rc_this_oldid = rev_id',
 132                                                 'rc_source' => RecentChange::SRC_CATEGORIZE,
 133                                                 // Allow rc_cur_id or rc_timestamp index usage
 134                                                 'rc_cur_id = rev_page',
 135                                                 'rc_timestamp = rev_timestamp'
 136                                         ]
 137                                 ) . ')'
 138                         ],
 139                         __METHOD__,
 140                         [ 'ORDER BY' => 'rev_timestamp DESC, rev_id DESC' ]
 141                 );
 142                 // Only consider revisions newer than any such revision
 143                 if ( $row ) {
 144                         $cutoffUnix = wfTimestamp( TS_UNIX, $row->rev_timestamp );
 145                         $lastRevId = (int)$row->rev_id;
 146                 } else {
 147                         $lastRevId = 0;
 148                 }
 149
 150                 // Find revisions to this page made around and after this revision which lack category
 151                 // notifications in recent changes. This lets jobs pick up were the last one left off.
 152                 $encCutoff = $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) );
 153                 $revQuery = Revision::getQueryInfo();
 154                 $res = $dbr->select(
 155                         $revQuery['tables'],
 156                         $revQuery['fields'],
 157                         [
 158                                 'rev_page' => $page->getId(),
 159                                 "rev_timestamp > $encCutoff" .
 160                                         " OR (rev_timestamp = $encCutoff AND rev_id > $lastRevId)"
 161                         ],
 162                         __METHOD__,
 163                         [ 'ORDER BY' => 'rev_timestamp ASC, rev_id ASC' ],
 164                         $revQuery['joins']
 165                 );
 166
 167                 // Apply all category updates in revision timestamp order
 168                 foreach ( $res as $row ) {
 169                         $this->notifyUpdatesForRevision( $lbFactory, $page, Revision::newFromRow( $row ) );
 170                 }
 171
 172                 return true;
 173         }
 174
 175         /**
 176          * @param LBFactory $lbFactory
 177          * @param WikiPage $page
 178          * @param Revision $newRev
 179          * @throws MWException
 180          */
 181         protected function notifyUpdatesForRevision(
 182                 LBFactory $lbFactory, WikiPage $page, Revision $newRev
 183         ) {
 184                 $config = RequestContext::getMain()->getConfig();
 185                 $title = $page->getTitle();
 186
 187                 // Get the new revision
 188                 if ( !$newRev->getContent() ) {
 189                         return; // deleted?
 190                 }
 191
 192                 // Get the prior revision (the same for null edits)
 193                 if ( $newRev->getParentId() ) {
 194                         $oldRev = Revision::newFromId( $newRev->getParentId(), Revision::READ_LATEST );
 195                         if ( !$oldRev->getContent() ) {
 196                                 return; // deleted?
 197                         }
 198                 } else {
 199                         $oldRev = null;
 200                 }
 201
 202                 // Parse the new revision and get the categories
 203                 $categoryChanges = $this->getExplicitCategoriesChanges( $page, $newRev, $oldRev );
 204                 list( $categoryInserts, $categoryDeletes ) = $categoryChanges;
 205                 if ( !$categoryInserts && !$categoryDeletes ) {
 206                         return; // nothing to do
 207                 }
 208
 209                 $catMembChange = new CategoryMembershipChange( $title, $newRev );
 210                 $catMembChange->checkTemplateLinks();
 211
 212                 $batchSize = $config->get( 'UpdateRowsPerQuery' );
 213                 $insertCount = 0;
 214
 215                 foreach ( $categoryInserts as $categoryName ) {
 216                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 217                         $catMembChange->triggerCategoryAddedNotification( $categoryTitle );
 218                         if ( $insertCount++ && ( $insertCount % $batchSize ) == 0 ) {
 219                                 $lbFactory->commitAndWaitForReplication( __METHOD__, $this->ticket );
 220                         }
 221                 }
 222
 223                 foreach ( $categoryDeletes as $categoryName ) {
 224                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 225                         $catMembChange->triggerCategoryRemovedNotification( $categoryTitle );
 226                         if ( $insertCount++ && ( $insertCount++ % $batchSize ) == 0 ) {
 227                                 $lbFactory->commitAndWaitForReplication( __METHOD__, $this->ticket );
 228                         }
 229                 }
 230         }
 231
 232         private function getExplicitCategoriesChanges(
 233                 WikiPage $page, Revision $newRev, Revision $oldRev = null
 234         ) {
 235                 // Inject the same timestamp for both revision parses to avoid seeing category changes
 236                 // due to time-based parser functions. Inject the same page title for the parses too.
 237                 // Note that REPEATABLE-READ makes template/file pages appear unchanged between parses.
 238                 $parseTimestamp = $newRev->getTimestamp();
 239                 // Parse the old rev and get the categories. Do not use link tables as that
 240                 // assumes these updates are perfectly FIFO and that link tables are always
 241                 // up to date, neither of which are true.
 242                 $oldCategories = $oldRev
 243                         ? $this->getCategoriesAtRev( $page, $oldRev, $parseTimestamp )
 244                         : [];
 245                 // Parse the new revision and get the categories
 246                 $newCategories = $this->getCategoriesAtRev( $page, $newRev, $parseTimestamp );
 247
 248                 $categoryInserts = array_values( array_diff( $newCategories, $oldCategories ) );
 249                 $categoryDeletes = array_values( array_diff( $oldCategories, $newCategories ) );
 250
 251                 return [ $categoryInserts, $categoryDeletes ];
 252         }
 253
 254         /**
 255          * @param WikiPage $page
 256          * @param Revision $rev
 257          * @param string $parseTimestamp TS_MW
 258          *
 259          * @return string[] category names
 260          */
 261         private function getCategoriesAtRev( WikiPage $page, Revision $rev, $parseTimestamp ) {
 262                 $renderer = MediaWikiServices::getInstance()->getRevisionRenderer();
 263                 $options = $page->makeParserOptions( 'canonical' );
 264                 $options->setTimestamp( $parseTimestamp );
 265
 266                 $output = $rev->isCurrent() ? $this->parserCache->get( $page, $options ) : null;
 267
 268                 if ( !$output || $output->getCacheRevisionId() !== $rev->getId() ) {
 269                         $output = $renderer->getRenderedRevision( $rev->getRevisionRecord(), $options )
 270                                 ->getRevisionParserOutput();
 271                 }
 272
 273                 // array keys will cast numeric category names to ints
 274                 // so we need to cast them back to strings to avoid breaking things!
 275                 return array_map( 'strval', array_keys( $output->getCategories() ) );
 276         }
 277
 278         public function getDeduplicationInfo() {
 279                 $info = parent::getDeduplicationInfo();
 280                 unset( $info['params']['revTimestamp'] ); // first job wins
 281
 282                 return $info;
 283         }
 284 }