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
  23 /**
  24  * Job to add recent change entries mentioning category membership changes
  25  *
  26  * Parameters include:
  27  *   - pageId : page ID
  28  *   - revTimestamp : timestamp of the triggering revision
  29  *
  30  * Category changes will be mentioned for revisions at/after the timestamp for this page
  31  *
  32  * @since 1.27
  33  */
  34 class CategoryMembershipChangeJob extends Job {
  35         const ENQUEUE_FUDGE_SEC = 60;
  36
  37         public function __construct( Title $title, array $params ) {
  38                 parent::__construct( 'categoryMembershipChange', $title, $params );
  39                 // Only need one job per page. Note that ENQUEUE_FUDGE_SEC handles races where an
  40                 // older revision job gets inserted while the newer revision job is de-duplicated.
  41                 $this->removeDuplicates = true;
  42         }
  43
  44         public function run() {
  45                 $page = WikiPage::newFromID( $this->params['pageId'], WikiPage::READ_LATEST );
  46                 if ( !$page ) {
  47                         $this->setLastError( "Could not find page #{$this->params['pageId']}" );
  48                         return false; // deleted?
  49                 }
  50
  51                 $dbw = wfGetDB( DB_MASTER );
  52                 // Use a named lock so that jobs for this page see each others' changes
  53                 $lockKey = "CategoryMembershipUpdates:{$page->getId()}";
  54                 $scopedLock = $dbw->getScopedLockAndFlush( $lockKey, __METHOD__, 10 );
  55                 if ( !$scopedLock ) {
  56                         $this->setLastError( "Could not acquire lock '$lockKey'" );
  57                         return false;
  58                 }
  59
  60                 $dbr = wfGetDB( DB_SLAVE, [ 'recentchanges' ] );
  61                 // Wait till the slave is caught up so that jobs for this page see each others' changes
  62                 if ( !wfGetLB()->safeWaitForMasterPos( $dbr ) ) {
  63                         $this->setLastError( "Timed out while waiting for slave to catch up" );
  64                         return false;
  65                 }
  66                 // Clear any stale REPEATABLE-READ snapshot
  67                 $dbr->commit( __METHOD__, 'flush' );
  68
  69                 $cutoffUnix = wfTimestamp( TS_UNIX, $this->params['revTimestamp'] );
  70                 // Using ENQUEUE_FUDGE_SEC handles jobs inserted out of revision order due to the delay
  71                 // between COMMIT and actual enqueueing of the CategoryMembershipChangeJob job.
  72                 $cutoffUnix -= self::ENQUEUE_FUDGE_SEC;
  73
  74                 // Get the newest revision that has a SRC_CATEGORIZE row...
  75                 $row = $dbr->selectRow(
  76                         [ 'revision', 'recentchanges' ],
  77                         [ 'rev_timestamp', 'rev_id' ],
  78                         [
  79                                 'rev_page' => $page->getId(),
  80                                 'rev_timestamp >= ' . $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) )
  81                         ],
  82                         __METHOD__,
  83                         [ 'ORDER BY' => 'rev_timestamp DESC, rev_id DESC' ],
  84                         [
  85                                 'recentchanges' => [
  86                                         'INNER JOIN',
  87                                         [
  88                                                 'rc_this_oldid = rev_id',
  89                                                 'rc_source' => RecentChange::SRC_CATEGORIZE,
  90                                                 // Allow rc_cur_id or rc_timestamp index usage
  91                                                 'rc_cur_id = rev_page',
  92                                                 'rc_timestamp >= rev_timestamp'
  93                                         ]
  94                                 ]
  95                         ]
  96                 );
  97                 // Only consider revisions newer than any such revision
  98                 if ( $row ) {
  99                         $cutoffUnix = wfTimestamp( TS_UNIX, $row->rev_timestamp );
 100                         $lastRevId = (int)$row->rev_id;
 101                 } else {
 102                         $lastRevId = 0;
 103                 }
 104
 105                 // Find revisions to this page made around and after this revision which lack category
 106                 // notifications in recent changes. This lets jobs pick up were the last one left off.
 107                 $encCutoff = $dbr->addQuotes( $dbr->timestamp( $cutoffUnix ) );
 108                 $res = $dbr->select(
 109                         'revision',
 110                         Revision::selectFields(),
 111                         [
 112                                 'rev_page' => $page->getId(),
 113                                 "rev_timestamp > $encCutoff" .
 114                                         " OR (rev_timestamp = $encCutoff AND rev_id > $lastRevId)"
 115                         ],
 116                         __METHOD__,
 117                         [ 'ORDER BY' => 'rev_timestamp ASC, rev_id ASC' ]
 118                 );
 119
 120                 // Apply all category updates in revision timestamp order
 121                 foreach ( $res as $row ) {
 122                         $this->notifyUpdatesForRevision( $page, Revision::newFromRow( $row ) );
 123                 }
 124
 125                 return true;
 126         }
 127
 128         /**
 129          * @param WikiPage $page
 130          * @param Revision $newRev
 131          * @throws MWException
 132          */
 133         protected function notifyUpdatesForRevision( WikiPage $page, Revision $newRev ) {
 134                 $config = RequestContext::getMain()->getConfig();
 135                 $title = $page->getTitle();
 136
 137                 // Get the new revision
 138                 if ( !$newRev->getContent() ) {
 139                         return; // deleted?
 140                 }
 141
 142                 // Get the prior revision (the same for null edits)
 143                 if ( $newRev->getParentId() ) {
 144                         $oldRev = Revision::newFromId( $newRev->getParentId(), Revision::READ_LATEST );
 145                         if ( !$oldRev->getContent() ) {
 146                                 return; // deleted?
 147                         }
 148                 } else {
 149                         $oldRev = null;
 150                 }
 151
 152                 // Parse the new revision and get the categories
 153                 $categoryChanges = $this->getExplicitCategoriesChanges( $title, $newRev, $oldRev );
 154                 list( $categoryInserts, $categoryDeletes ) = $categoryChanges;
 155                 if ( !$categoryInserts && !$categoryDeletes ) {
 156                         return; // nothing to do
 157                 }
 158
 159                 $dbw = wfGetDB( DB_MASTER );
 160                 $catMembChange = new CategoryMembershipChange( $title, $newRev );
 161                 $catMembChange->checkTemplateLinks();
 162
 163                 $batchSize = $config->get( 'UpdateRowsPerQuery' );
 164                 $insertCount = 0;
 165
 166                 foreach ( $categoryInserts as $categoryName ) {
 167                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 168                         $catMembChange->triggerCategoryAddedNotification( $categoryTitle );
 169                         if ( $insertCount++ && ( $insertCount % $batchSize ) == 0 ) {
 170                                 $dbw->commit( __METHOD__, 'flush' );
 171                                 wfGetLBFactory()->waitForReplication();
 172                         }
 173                 }
 174
 175                 foreach ( $categoryDeletes as $categoryName ) {
 176                         $categoryTitle = Title::makeTitle( NS_CATEGORY, $categoryName );
 177                         $catMembChange->triggerCategoryRemovedNotification( $categoryTitle );
 178                         if ( $insertCount++ && ( $insertCount++ % $batchSize ) == 0 ) {
 179                                 $dbw->commit( __METHOD__, 'flush' );
 180                                 wfGetLBFactory()->waitForReplication();
 181                         }
 182                 }
 183         }
 184
 185         private function getExplicitCategoriesChanges(
 186                 Title $title, Revision $newRev, Revision $oldRev = null
 187         ) {
 188                 // Inject the same timestamp for both revision parses to avoid seeing category changes
 189                 // due to time-based parser functions. Inject the same page title for the parses too.
 190                 // Note that REPEATABLE-READ makes template/file pages appear unchanged between parses.
 191                 $parseTimestamp = $newRev->getTimestamp();
 192                 // Parse the old rev and get the categories. Do not use link tables as that
 193                 // assumes these updates are perfectly FIFO and that link tables are always
 194                 // up to date, neither of which are true.
 195                 $oldCategories = $oldRev
 196                         ? $this->getCategoriesAtRev( $title, $oldRev, $parseTimestamp )
 197                         : [];
 198                 // Parse the new revision and get the categories
 199                 $newCategories = $this->getCategoriesAtRev( $title, $newRev, $parseTimestamp );
 200
 201                 $categoryInserts = array_values( array_diff( $newCategories, $oldCategories ) );
 202                 $categoryDeletes = array_values( array_diff( $oldCategories, $newCategories ) );
 203
 204                 return [ $categoryInserts, $categoryDeletes ];
 205         }
 206
 207         /**
 208          * @param Title $title
 209          * @param Revision $rev
 210          * @param string $parseTimestamp TS_MW
 211          *
 212          * @return string[] category names
 213          */
 214         private function getCategoriesAtRev( Title $title, Revision $rev, $parseTimestamp ) {
 215                 $content = $rev->getContent();
 216                 $options = $content->getContentHandler()->makeParserOptions( 'canonical' );
 217                 $options->setTimestamp( $parseTimestamp );
 218                 // This could possibly use the parser cache if it checked the revision ID,
 219                 // but that's more complicated than it's worth.
 220                 $output = $content->getParserOutput( $title, $rev->getId(), $options );
 221
 222                 // array keys will cast numeric category names to ints
 223                 // so we need to cast them back to strings to avoid breaking things!
 224                 return array_map( 'strval', array_keys( $output->getCategories() ) );
 225         }
 226
 227         public function getDeduplicationInfo() {
 228                 $info = parent::getDeduplicationInfo();
 229                 unset( $info['params']['revTimestamp'] ); // first job wins
 230
 231                 return $info;
 232         }
 233 }