includes/jobqueue/jobs/RefreshLinksJob.php

   1 <?php
   2 /**
   3  * Job to update link tables for pages
   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  * @ingroup JobQueue
  22  */
  23 use MediaWiki\MediaWikiServices;
  24
  25 /**
  26  * Job to update link tables for pages
  27  *
  28  * This job comes in a few variants:
  29  *   - a) Recursive jobs to update links for backlink pages for a given title.
  30  *        These jobs have (recursive:true,table:<table>) set.
  31  *   - b) Jobs to update links for a set of pages (the job title is ignored).
  32  *            These jobs have (pages:(<page ID>:(<namespace>,<title>),...) set.
  33  *   - c) Jobs to update links for a single page (the job title)
  34  *        These jobs need no extra fields set.
  35  *
  36  * @ingroup JobQueue
  37  */
  38 class RefreshLinksJob extends Job {
  39         /** @var float Cache parser output when it takes this long to render */
  40         const PARSE_THRESHOLD_SEC = 1.0;
  41         /** @var integer Lag safety margin when comparing root job times to last-refresh times */
  42         const CLOCK_FUDGE = 10;
  43         /** @var integer How many seconds to wait for slaves to catch up */
  44         const LAG_WAIT_TIMEOUT = 15;
  45
  46         function __construct( Title $title, array $params ) {
  47                 parent::__construct( 'refreshLinks', $title, $params );
  48                 // Avoid the overhead of de-duplication when it would be pointless
  49                 $this->removeDuplicates = (
  50                         // Ranges rarely will line up
  51                         !isset( $params['range'] ) &&
  52                         // Multiple pages per job make matches unlikely
  53                         !( isset( $params['pages'] ) && count( $params['pages'] ) != 1 )
  54                 );
  55         }
  56
  57         /**
  58          * @param Title $title
  59          * @param array $params
  60          * @return RefreshLinksJob
  61          */
  62         public static function newPrioritized( Title $title, array $params ) {
  63                 $job = new self( $title, $params );
  64                 $job->command = 'refreshLinksPrioritized';
  65
  66                 return $job;
  67         }
  68
  69         /**
  70          * @param Title $title
  71          * @param array $params
  72          * @return RefreshLinksJob
  73          */
  74         public static function newDynamic( Title $title, array $params ) {
  75                 $job = new self( $title, $params );
  76                 $job->command = 'refreshLinksDynamic';
  77
  78                 return $job;
  79         }
  80
  81         function run() {
  82                 global $wgUpdateRowsPerJob;
  83
  84                 // Job to update all (or a range of) backlink pages for a page
  85                 if ( !empty( $this->params['recursive'] ) ) {
  86                         // When the base job branches, wait for the slaves to catch up to the master.
  87                         // From then on, we know that any template changes at the time the base job was
  88                         // enqueued will be reflected in backlink page parses when the leaf jobs run.
  89                         if ( !isset( $params['range'] ) ) {
  90                                 try {
  91                                         wfGetLBFactory()->waitForReplication( [
  92                                                 'wiki'    => wfWikiID(),
  93                                                 'timeout' => self::LAG_WAIT_TIMEOUT
  94                                         ] );
  95                                 } catch ( DBReplicationWaitError $e ) { // only try so hard
  96                                         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
  97                                         $stats->increment( 'refreshlinks.lag_wait_failed' );
  98                                 }
  99                         }
 100                         // Carry over information for de-duplication
 101                         $extraParams = $this->getRootJobParams();
 102                         $extraParams['triggeredRecursive'] = true;
 103                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 104                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 105                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 106                                 $this,
 107                                 $wgUpdateRowsPerJob,
 108                                 1, // job-per-title
 109                                 [ 'params' => $extraParams ]
 110                         );
 111                         JobQueueGroup::singleton()->push( $jobs );
 112                 // Job to update link tables for a set of titles
 113                 } elseif ( isset( $this->params['pages'] ) ) {
 114                         foreach ( $this->params['pages'] as $pageId => $nsAndKey ) {
 115                                 list( $ns, $dbKey ) = $nsAndKey;
 116                                 $this->runForTitle( Title::makeTitleSafe( $ns, $dbKey ) );
 117                         }
 118                 // Job to update link tables for a given title
 119                 } else {
 120                         $this->runForTitle( $this->title );
 121                 }
 122
 123                 return true;
 124         }
 125
 126         /**
 127          * @param Title $title
 128          * @return bool
 129          */
 130         protected function runForTitle( Title $title ) {
 131                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 132
 133                 $page = WikiPage::factory( $title );
 134                 $page->loadPageData( WikiPage::READ_LATEST );
 135
 136                 // Serialize links updates by page ID so they see each others' changes
 137                 $scopedLock = LinksUpdate::acquirePageLock( wfGetDB( DB_MASTER ), $page->getId(), 'job' );
 138                 // Get the latest ID *after* acquirePageLock() flushed the transaction.
 139                 // This is used to detect edits/moves after loadPageData() but before the scope lock.
 140                 // The works around the chicken/egg problem of determining the scope lock key.
 141                 $latest = $title->getLatestRevID( Title::GAID_FOR_UPDATE );
 142
 143                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 144                         // Fetch the specified revision; lockAndGetLatest() below detects if the page
 145                         // was edited since and aborts in order to avoid corrupting the link tables
 146                         $revision = Revision::newFromId(
 147                                 $this->params['triggeringRevisionId'],
 148                                 Revision::READ_LATEST
 149                         );
 150                 } else {
 151                         // Fetch current revision; READ_LATEST reduces lockAndGetLatest() check failures
 152                         $revision = Revision::newFromTitle( $title, false, Revision::READ_LATEST );
 153                 }
 154
 155                 if ( !$revision ) {
 156                         $stats->increment( 'refreshlinks.rev_not_found' );
 157                         $this->setLastError( "Revision not found for {$title->getPrefixedDBkey()}" );
 158                         return false; // just deleted?
 159                 } elseif ( $revision->getId() != $latest || $revision->getPage() !== $page->getId() ) {
 160                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 161                         // serialized, it would be OK to update links based on older revisions since it
 162                         // would eventually get to the latest. Since that is not the case (by design),
 163                         // only update the link tables to a state matching the current revision's output.
 164                         $stats->increment( 'refreshlinks.rev_not_current' );
 165                         $this->setLastError( "Revision {$revision->getId()} is not current" );
 166                         return false;
 167                 }
 168
 169                 $content = $revision->getContent( Revision::RAW );
 170                 if ( !$content ) {
 171                         // If there is no content, pretend the content is empty
 172                         $content = $revision->getContentHandler()->makeEmptyContent();
 173                 }
 174
 175                 $parserOutput = false;
 176                 $parserOptions = $page->makeParserOptions( 'canonical' );
 177                 // If page_touched changed after this root job, then it is likely that
 178                 // any views of the pages already resulted in re-parses which are now in
 179                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 180                 if ( isset( $this->params['rootJobTimestamp'] ) ) {
 181                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 182
 183                         $skewedTimestamp = $this->params['rootJobTimestamp'];
 184                         if ( $opportunistic ) {
 185                                 // Neither clock skew nor DB snapshot/slave lag matter much for such
 186                                 // updates; focus on reusing the (often recently updated) cache
 187                         } else {
 188                                 // For transclusion updates, the template changes must be reflected
 189                                 $skewedTimestamp = wfTimestamp( TS_MW,
 190                                         wfTimestamp( TS_UNIX, $skewedTimestamp ) + self::CLOCK_FUDGE
 191                                 );
 192                         }
 193
 194                         if ( $page->getLinksTimestamp() > $skewedTimestamp ) {
 195                                 // Something already updated the backlinks since this job was made
 196                                 $stats->increment( 'refreshlinks.update_skipped' );
 197                                 return true;
 198                         }
 199
 200                         if ( $page->getTouched() >= $this->params['rootJobTimestamp'] || $opportunistic ) {
 201                                 // Cache is suspected to be up-to-date. As long as the cache rev ID matches
 202                                 // and it reflects the job's triggering change, then it is usable.
 203                                 $parserOutput = ParserCache::singleton()->getDirty( $page, $parserOptions );
 204                                 if ( !$parserOutput
 205                                         || $parserOutput->getCacheRevisionId() != $revision->getId()
 206                                         || $parserOutput->getCacheTime() < $skewedTimestamp
 207                                 ) {
 208                                         $parserOutput = false; // too stale
 209                                 }
 210                         }
 211                 }
 212
 213                 // Fetch the current revision and parse it if necessary...
 214                 if ( $parserOutput ) {
 215                         $stats->increment( 'refreshlinks.parser_cached' );
 216                 } else {
 217                         $start = microtime( true );
 218                         // Revision ID must be passed to the parser output to get revision variables correct
 219                         $parserOutput = $content->getParserOutput(
 220                                 $title, $revision->getId(), $parserOptions, false );
 221                         $elapsed = microtime( true ) - $start;
 222                         // If it took a long time to render, then save this back to the cache to avoid
 223                         // wasted CPU by other apaches or job runners. We don't want to always save to
 224                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 225                         if ( $elapsed >= self::PARSE_THRESHOLD_SEC
 226                                 && $page->shouldCheckParserCache( $parserOptions, $revision->getId() )
 227                                 && $parserOutput->isCacheable()
 228                         ) {
 229                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 230                                 ParserCache::singleton()->save(
 231                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 232                                 );
 233                         }
 234                         $stats->increment( 'refreshlinks.parser_uncached' );
 235                 }
 236
 237                 $updates = $content->getSecondaryDataUpdates(
 238                         $title,
 239                         null,
 240                         !empty( $this->params['useRecursiveLinksUpdate'] ),
 241                         $parserOutput
 242                 );
 243
 244                 $factory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 245                 $ticket = $factory->getEmptyTransactionTicket( __METHOD__ );
 246                 foreach ( $updates as $key => $update ) {
 247                         $update->setTransactionTicket( $ticket );
 248                         // FIXME: This code probably shouldn't be here?
 249                         // Needed by things like Echo notifications which need
 250                         // to know which user caused the links update
 251                         if ( $update instanceof LinksUpdate ) {
 252                                 $update->setRevision( $revision );
 253                                 if ( !empty( $this->params['triggeringUser'] ) ) {
 254                                         $userInfo = $this->params['triggeringUser'];
 255                                         if ( $userInfo['userId'] ) {
 256                                                 $user = User::newFromId( $userInfo['userId'] );
 257                                         } else {
 258                                                 // Anonymous, use the username
 259                                                 $user = User::newFromName( $userInfo['userName'], false );
 260                                         }
 261                                         $update->setTriggeringUser( $user );
 262                                 }
 263                         }
 264                 }
 265
 266                 DataUpdate::runUpdates( $updates );
 267
 268                 InfoAction::invalidateCache( $title );
 269
 270                 return true;
 271         }
 272
 273         public function getDeduplicationInfo() {
 274                 $info = parent::getDeduplicationInfo();
 275                 if ( is_array( $info['params'] ) ) {
 276                         // For per-pages jobs, the job title is that of the template that changed
 277                         // (or similar), so remove that since it ruins duplicate detection
 278                         if ( isset( $info['pages'] ) ) {
 279                                 unset( $info['namespace'] );
 280                                 unset( $info['title'] );
 281                         }
 282                 }
 283
 284                 return $info;
 285         }
 286
 287         public function workItemCount() {
 288                 return isset( $this->params['pages'] ) ? count( $this->params['pages'] ) : 1;
 289         }
 290 }