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