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
  24 /**
  25  * Job to add recent change entries mentioning category membership changes
  26  *
  27  * Parameters include:
  28  *   - pageId : page ID
  29  *   - revTimestamp : timestamp of the triggering revision
  30  *
  31  * Category changes will be mentioned for revisions at/after the timestamp for this page
  32  *
  33  * @since 1.27
  34  */
  35 class CategoryMembershipChangeJob extends Job {
  36         const ENQUEUE_FUDGE_SEC = 60;
  37
  38         public function __construct( Title $title, array $params ) {
  39                 parent::__construct( 'categoryMembershipChange', $title, $params );
  40                 // Only need one job per page. Note that ENQUEUE_FUDGE_SEC handles races where an
  41                 // older revision job gets inserted while the newer revision job is de-duplicated.
  42                 $this->removeDuplicates = true;
  43         }
  44
  45         public function run() {
  46                 $page = WikiPage::newFromID( $this->params['pageId'], WikiPage::READ_LATEST );
  47                 if ( !$page ) {
  48                         $this->setLastError( "Could not find page #{$this->params['pageId']}" );
  49                         return false; // deleted?
  50                 }
  51
  52                 $lb = MediaWikiServices::getInstance()->getDBLoadBalancer();
  53                 $dbw = $lb->getConnection( DB_MASTER );
  54                 // Use a named lock so that jobs for this page see each others' changes
  55                 $lockKey = "CategoryMembershipUpdates:{$page->getId()}";
  56                 $scopedLock = $dbw->getScopedLockAndFlush( $lockKey, __METHOD__, 10 );
  57                 if ( !$scopedLock ) {
  58                         $this->setLastError( "Could not acquire lock '$lockKey'" );
  59                         return false;
  60                 }
  61
  62                 $dbr = wfGetDB( DB_REPLICA, [ 'recentchanges' ] );
  63                 // Wait till the replica DB is caught up so that jobs for this page see each others' changes
  64                 if ( !$lb->safeWaitForMasterPos( $dbr ) ) {
  65                         $this->setLastError( "Timed out while waiting for replica DB to catch up" );
  66                         return false;
  67                 }
  68                 // Clear any stale REPEATABLE-READ snapshot
  69                 $dbr->flushSnapshot( __METHOD__ );
  70
  71                 $cutoffUnix = wfTimestamp( TS_UNIX, $this->params['revTimestamp'] );
  72                 // Using ENQUEUE_FUDGE_SEC handles jobs inserted out of revision order due to the delay
  73                 // between COMMIT and actual enqueueing of the CategoryMembershipChangeJob job.
  74                 $cutoffUnix -= self::ENQUEUE_FUDGE_SEC;
  75
  76                 // Get the newest revision that has a SRC_CATEGORIZE row...
  77                 $row = $dbr->selectRow(
  78                         [ 'revision', 'recentchanges' ],
  79                         [ 'rev_timestamp', 'rev_id' ],
  80                         [
  81                                 'rev_page' => $page->getId(),
  82                                 'rev_timestamp >= ' . $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) )
  83                         ],
  84                         __METHOD__,
  85                         [ 'ORDER BY' => 'rev_timestamp DESC, rev_id DESC' ],
  86                         [
  87                                 'recentchanges' => [
  88                                         'INNER JOIN',
  89                                         [
  90                                                 'rc_this_oldid = rev_id',
  91                                                 'rc_source' => RecentChange::SRC_CATEGORIZE,
  92                                                 // Allow rc_cur_id or rc_timestamp index usage
  93                                                 'rc_cur_id = rev_page',
  94                                                 'rc_timestamp >= rev_timestamp'
  95                                         ]
  96                                 ]
  97                         ]
  98                 );
  99                 // Only consider revisions newer than any such revision
 100                 if ( $row ) {
 101                         $cutoffUnix = wfTimestamp( TS_UNIX, $row->rev_timestamp );
 102                         $lastRevId = (int)$row->rev_id;
 103                 } else {
 104                         $lastRevId = 0;
 105                 }
 106
 107                 // Find revisions to this page made around and after this revision which lack category
 108                 // notifications in recent changes. This lets jobs pick up were the last one left off.
 109                 $encCutoff = $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) );
 110                 $res = $dbr->select(
 111                         'revision',
 112                         Revision::selectFields(),
 113                         [
 114                                 'rev_page' => $page->getId(),
 115                                 "rev_timestamp > $encCutoff" .
 116                                         " OR (rev_timestamp = $encCutoff AND rev_id > $lastRevId)"
 117                         ],
 118                         __METHOD__,
 119                         [ 'ORDER BY' => 'rev_timestamp ASC, rev_id ASC' ]
 120                 );
 121
 122                 // Apply all category updates in revision timestamp order
 123                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 124                 foreach ( $res as $row ) {
 125                         $this->notifyUpdatesForRevision( $lbFactory, $page, Revision::newFromRow( $row ) );
 126                 }
 127
 128                 return true;
 129         }
 130
 131         /**
 132          * @param LBFactory $lbFactory
 133          * @param WikiPage $page
 134          * @param Revision $newRev
 135          * @throws MWException
 136          */
 137         protected function notifyUpdatesForRevision(
 138                 LBFactory $lbFactory, WikiPage $page, Revision $newRev
 139         ) {
 140                 $config = RequestContext::getMain()->getConfig();
 141                 $title = $page->getTitle();
 142
 143                 // Get the new revision
 144                 if ( !$newRev->getContent() ) {
 145                         return; // deleted?
 146                 }
 147
 148                 // Get the prior revision (the same for null edits)
 149                 if ( $newRev->getParentId() ) {
 150                         $oldRev = Revision::newFromId( $newRev->getParentId(), Revision::READ_LATEST );
 151                         if ( !$oldRev->getContent() ) {
 152                                 return; // deleted?
 153                         }
 154                 } else {
 155                         $oldRev = null;
 156                 }
 157
 158                 // Parse the new revision and get the categories
 159                 $categoryChanges = $this->getExplicitCategoriesChanges( $title, $newRev, $oldRev );
 160                 list( $categoryInserts, $categoryDeletes ) = $categoryChanges;
 161                 if ( !$categoryInserts && !$categoryDeletes ) {
 162                         return; // nothing to do
 163                 }
 164
 165                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 166
 167                 $catMembChange = new CategoryMembershipChange( $title, $newRev );
 168                 $catMembChange->checkTemplateLinks();
 169
 170                 $batchSize = $config->get( 'UpdateRowsPerQuery' );
 171                 $insertCount = 0;
 172
 173                 foreach ( $categoryInserts as $categoryName ) {
 174                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 175                         $catMembChange->triggerCategoryAddedNotification( $categoryTitle );
 176                         if ( $insertCount++ && ( $insertCount % $batchSize ) == 0 ) {
 177                                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 178                         }
 179                 }
 180
 181                 foreach ( $categoryDeletes as $categoryName ) {
 182                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 183                         $catMembChange->triggerCategoryRemovedNotification( $categoryTitle );
 184                         if ( $insertCount++ && ( $insertCount++ % $batchSize ) == 0 ) {
 185                                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 186                         }
 187                 }
 188         }
 189
 190         private function getExplicitCategoriesChanges(
 191                 Title $title, Revision $newRev, Revision $oldRev = null
 192         ) {
 193                 // Inject the same timestamp for both revision parses to avoid seeing category changes
 194                 // due to time-based parser functions. Inject the same page title for the parses too.
 195                 // Note that REPEATABLE-READ makes template/file pages appear unchanged between parses.
 196                 $parseTimestamp = $newRev->getTimestamp();
 197                 // Parse the old rev and get the categories. Do not use link tables as that
 198                 // assumes these updates are perfectly FIFO and that link tables are always
 199                 // up to date, neither of which are true.
 200                 $oldCategories = $oldRev
 201                         ? $this->getCategoriesAtRev( $title, $oldRev, $parseTimestamp )
 202                         : [];
 203                 // Parse the new revision and get the categories
 204                 $newCategories = $this->getCategoriesAtRev( $title, $newRev, $parseTimestamp );
 205
 206                 $categoryInserts = array_values( array_diff( $newCategories, $oldCategories ) );
 207                 $categoryDeletes = array_values( array_diff( $oldCategories, $newCategories ) );
 208
 209                 return [ $categoryInserts, $categoryDeletes ];
 210         }
 211
 212         /**
 213          * @param Title $title
 214          * @param Revision $rev
 215          * @param string $parseTimestamp TS_MW
 216          *
 217          * @return string[] category names
 218          */
 219         private function getCategoriesAtRev( Title $title, Revision $rev, $parseTimestamp ) {
 220                 $content = $rev->getContent();
 221                 $options = $content->getContentHandler()->makeParserOptions( 'canonical' );
 222                 $options->setTimestamp( $parseTimestamp );
 223                 // This could possibly use the parser cache if it checked the revision ID,
 224                 // but that's more complicated than it's worth.
 225                 $output = $content->getParserOutput( $title, $rev->getId(), $options );
 226
 227                 // array keys will cast numeric category names to ints
 228                 // so we need to cast them back to strings to avoid breaking things!
 229                 return array_map( 'strval', array_keys( $output->getCategories() ) );
 230         }
 231
 232         public function getDeduplicationInfo() {
 233                 $info = parent::getDeduplicationInfo();
 234                 unset( $info['params']['revTimestamp'] ); // first job wins
 235
 236                 return $info;
 237         }
 238 }