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