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         /** @var integer|null */
  37         private $ticket;
  38
  39         const ENQUEUE_FUDGE_SEC = 60;
  40
  41         public function __construct( Title $title, array $params ) {
  42                 parent::__construct( 'categoryMembershipChange', $title, $params );
  43                 // Only need one job per page. Note that ENQUEUE_FUDGE_SEC handles races where an
  44                 // older revision job gets inserted while the newer revision job is de-duplicated.
  45                 $this->removeDuplicates = true;
  46         }
  47
  48         public function run() {
  49                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  50                 $lb = $lbFactory->getMainLB();
  51                 $dbw = $lb->getConnection( DB_MASTER );
  52
  53                 $this->ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
  54
  55                 $page = WikiPage::newFromID( $this->params['pageId'], WikiPage::READ_LATEST );
  56                 if ( !$page ) {
  57                         $this->setLastError( "Could not find page #{$this->params['pageId']}" );
  58                         return false; // deleted?
  59                 }
  60
  61                 // Use a named lock so that jobs for this page see each others' changes
  62                 $lockKey = "CategoryMembershipUpdates:{$page->getId()}";
  63                 $scopedLock = $dbw->getScopedLockAndFlush( $lockKey, __METHOD__, 3 );
  64                 if ( !$scopedLock ) {
  65                         $this->setLastError( "Could not acquire lock '$lockKey'" );
  66                         return false;
  67                 }
  68
  69                 $dbr = $lb->getConnection( DB_REPLICA, [ 'recentchanges' ] );
  70                 // Wait till the replica DB is caught up so that jobs for this page see each others' changes
  71                 if ( !$lb->safeWaitForMasterPos( $dbr ) ) {
  72                         $this->setLastError( "Timed out while waiting for replica DB to catch up" );
  73                         return false;
  74                 }
  75                 // Clear any stale REPEATABLE-READ snapshot
  76                 $dbr->flushSnapshot( __METHOD__ );
  77
  78                 $cutoffUnix = wfTimestamp( TS_UNIX, $this->params['revTimestamp'] );
  79                 // Using ENQUEUE_FUDGE_SEC handles jobs inserted out of revision order due to the delay
  80                 // between COMMIT and actual enqueueing of the CategoryMembershipChangeJob job.
  81                 $cutoffUnix -= self::ENQUEUE_FUDGE_SEC;
  82
  83                 // Get the newest revision that has a SRC_CATEGORIZE row...
  84                 $row = $dbr->selectRow(
  85                         [ 'revision', 'recentchanges' ],
  86                         [ 'rev_timestamp', 'rev_id' ],
  87                         [
  88                                 'rev_page' => $page->getId(),
  89                                 'rev_timestamp >= ' . $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) )
  90                         ],
  91                         __METHOD__,
  92                         [ 'ORDER BY' => 'rev_timestamp DESC, rev_id DESC' ],
  93                         [
  94                                 'recentchanges' => [
  95                                         'INNER JOIN',
  96                                         [
  97                                                 'rc_this_oldid = rev_id',
  98                                                 'rc_source' => RecentChange::SRC_CATEGORIZE,
  99                                                 // Allow rc_cur_id or rc_timestamp index usage
 100                                                 'rc_cur_id = rev_page',
 101                                                 'rc_timestamp >= rev_timestamp'
 102                                         ]
 103                                 ]
 104                         ]
 105                 );
 106                 // Only consider revisions newer than any such revision
 107                 if ( $row ) {
 108                         $cutoffUnix = wfTimestamp( TS_UNIX, $row->rev_timestamp );
 109                         $lastRevId = (int)$row->rev_id;
 110                 } else {
 111                         $lastRevId = 0;
 112                 }
 113
 114                 // Find revisions to this page made around and after this revision which lack category
 115                 // notifications in recent changes. This lets jobs pick up were the last one left off.
 116                 $encCutoff = $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) );
 117                 $res = $dbr->select(
 118                         'revision',
 119                         Revision::selectFields(),
 120                         [
 121                                 'rev_page' => $page->getId(),
 122                                 "rev_timestamp > $encCutoff" .
 123                                         " OR (rev_timestamp = $encCutoff AND rev_id > $lastRevId)"
 124                         ],
 125                         __METHOD__,
 126                         [ 'ORDER BY' => 'rev_timestamp ASC, rev_id ASC' ]
 127                 );
 128
 129                 // Apply all category updates in revision timestamp order
 130                 foreach ( $res as $row ) {
 131                         $this->notifyUpdatesForRevision( $lbFactory, $page, Revision::newFromRow( $row ) );
 132                 }
 133
 134                 return true;
 135         }
 136
 137         /**
 138          * @param LBFactory $lbFactory
 139          * @param WikiPage $page
 140          * @param Revision $newRev
 141          * @throws MWException
 142          */
 143         protected function notifyUpdatesForRevision(
 144                 LBFactory $lbFactory, WikiPage $page, Revision $newRev
 145         ) {
 146                 $config = RequestContext::getMain()->getConfig();
 147                 $title = $page->getTitle();
 148
 149                 // Get the new revision
 150                 if ( !$newRev->getContent() ) {
 151                         return; // deleted?
 152                 }
 153
 154                 // Get the prior revision (the same for null edits)
 155                 if ( $newRev->getParentId() ) {
 156                         $oldRev = Revision::newFromId( $newRev->getParentId(), Revision::READ_LATEST );
 157                         if ( !$oldRev->getContent() ) {
 158                                 return; // deleted?
 159                         }
 160                 } else {
 161                         $oldRev = null;
 162                 }
 163
 164                 // Parse the new revision and get the categories
 165                 $categoryChanges = $this->getExplicitCategoriesChanges( $title, $newRev, $oldRev );
 166                 list( $categoryInserts, $categoryDeletes ) = $categoryChanges;
 167                 if ( !$categoryInserts && !$categoryDeletes ) {
 168                         return; // nothing to do
 169                 }
 170
 171                 $catMembChange = new CategoryMembershipChange( $title, $newRev );
 172                 $catMembChange->checkTemplateLinks();
 173
 174                 $batchSize = $config->get( 'UpdateRowsPerQuery' );
 175                 $insertCount = 0;
 176
 177                 foreach ( $categoryInserts as $categoryName ) {
 178                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 179                         $catMembChange->triggerCategoryAddedNotification( $categoryTitle );
 180                         if ( $insertCount++ && ( $insertCount % $batchSize ) == 0 ) {
 181                                 $lbFactory->commitAndWaitForReplication( __METHOD__, $this->ticket );
 182                         }
 183                 }
 184
 185                 foreach ( $categoryDeletes as $categoryName ) {
 186                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 187                         $catMembChange->triggerCategoryRemovedNotification( $categoryTitle );
 188                         if ( $insertCount++ && ( $insertCount++ % $batchSize ) == 0 ) {
 189                                 $lbFactory->commitAndWaitForReplication( __METHOD__, $this->ticket );
 190                         }
 191                 }
 192         }
 193
 194         private function getExplicitCategoriesChanges(
 195                 Title $title, Revision $newRev, Revision $oldRev = null
 196         ) {
 197                 // Inject the same timestamp for both revision parses to avoid seeing category changes
 198                 // due to time-based parser functions. Inject the same page title for the parses too.
 199                 // Note that REPEATABLE-READ makes template/file pages appear unchanged between parses.
 200                 $parseTimestamp = $newRev->getTimestamp();
 201                 // Parse the old rev and get the categories. Do not use link tables as that
 202                 // assumes these updates are perfectly FIFO and that link tables are always
 203                 // up to date, neither of which are true.
 204                 $oldCategories = $oldRev
 205                         ? $this->getCategoriesAtRev( $title, $oldRev, $parseTimestamp )
 206                         : [];
 207                 // Parse the new revision and get the categories
 208                 $newCategories = $this->getCategoriesAtRev( $title, $newRev, $parseTimestamp );
 209
 210                 $categoryInserts = array_values( array_diff( $newCategories, $oldCategories ) );
 211                 $categoryDeletes = array_values( array_diff( $oldCategories, $newCategories ) );
 212
 213                 return [ $categoryInserts, $categoryDeletes ];
 214         }
 215
 216         /**
 217          * @param Title $title
 218          * @param Revision $rev
 219          * @param string $parseTimestamp TS_MW
 220          *
 221          * @return string[] category names
 222          */
 223         private function getCategoriesAtRev( Title $title, Revision $rev, $parseTimestamp ) {
 224                 $content = $rev->getContent();
 225                 $options = $content->getContentHandler()->makeParserOptions( 'canonical' );
 226                 $options->setTimestamp( $parseTimestamp );
 227                 // This could possibly use the parser cache if it checked the revision ID,
 228                 // but that's more complicated than it's worth.
 229                 $output = $content->getParserOutput( $title, $rev->getId(), $options );
 230
 231                 // array keys will cast numeric category names to ints
 232                 // so we need to cast them back to strings to avoid breaking things!
 233                 return array_map( 'strval', array_keys( $output->getCategories() ) );
 234         }
 235
 236         public function getDeduplicationInfo() {
 237                 $info = parent::getDeduplicationInfo();
 238                 unset( $info['params']['revTimestamp'] ); // first job wins
 239
 240                 return $info;
 241         }
 242 }