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
  44         function __construct( Title $title, array $params ) {
  45                 parent::__construct( 'refreshLinks', $title, $params );
  46                 // Avoid the overhead of de-duplication when it would be pointless
  47                 $this->removeDuplicates = (
  48                         // Master positions won't match
  49                         !isset( $params['masterPos'] ) &&
  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                         // Carry over information for de-duplication
  87                         $extraParams = $this->getRootJobParams();
  88                         // Avoid slave lag when fetching templates.
  89                         // When the outermost job is run, we know that the caller that enqueued it must have
  90                         // committed the relevant changes to the DB by now. At that point, record the master
  91                         // position and pass it along as the job recursively breaks into smaller range jobs.
  92                         // Hopefully, when leaf jobs are popped, the slaves will have reached that position.
  93                         if ( isset( $this->params['masterPos'] ) ) {
  94                                 $extraParams['masterPos'] = $this->params['masterPos'];
  95                         } elseif ( wfGetLB()->getServerCount() > 1 ) {
  96                                 $extraParams['masterPos'] = wfGetLB()->getMasterPos();
  97                         } else {
  98                                 $extraParams['masterPos'] = false;
  99                         }
 100                         $extraParams['triggeredRecursive'] = true;
 101                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 102                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 103                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 104                                 $this,
 105                                 $wgUpdateRowsPerJob,
 106                                 1, // job-per-title
 107                                 [ 'params' => $extraParams ]
 108                         );
 109                         JobQueueGroup::singleton()->push( $jobs );
 110                 // Job to update link tables for a set of titles
 111                 } elseif ( isset( $this->params['pages'] ) ) {
 112                         $this->waitForMasterPosition();
 113                         foreach ( $this->params['pages'] as $pageId => $nsAndKey ) {
 114                                 list( $ns, $dbKey ) = $nsAndKey;
 115                                 $this->runForTitle( Title::makeTitleSafe( $ns, $dbKey ) );
 116                         }
 117                 // Job to update link tables for a given title
 118                 } else {
 119                         $this->waitForMasterPosition();
 120                         $this->runForTitle( $this->title );
 121                 }
 122
 123                 return true;
 124         }
 125
 126         protected function waitForMasterPosition() {
 127                 if ( !empty( $this->params['masterPos'] ) && wfGetLB()->getServerCount() > 1 ) {
 128                         // Wait for the current/next slave DB handle to catch up to the master.
 129                         // This way, we get the correct page_latest for templates or files that just
 130                         // changed milliseconds ago, having triggered this job to begin with.
 131                         wfGetLB()->waitFor( $this->params['masterPos'] );
 132                 }
 133         }
 134
 135         /**
 136          * @param Title $title
 137          * @return bool
 138          */
 139         protected function runForTitle( Title $title ) {
 140                 $page = WikiPage::factory( $title );
 141                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 142                         // Fetch the specified revision; lockAndGetLatest() below detects if the page
 143                         // was edited since and aborts in order to avoid corrupting the link tables
 144                         $revision = Revision::newFromId(
 145                                 $this->params['triggeringRevisionId'],
 146                                 Revision::READ_LATEST
 147                         );
 148                 } else {
 149                         // Fetch current revision; READ_LATEST reduces lockAndGetLatest() check failures
 150                         $revision = Revision::newFromTitle( $title, false, Revision::READ_LATEST );
 151                 }
 152
 153                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 154
 155                 if ( !$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->isCurrent() ) {
 160                         // If the revision isn't current, there's no point in doing a bunch
 161                         // of work just to fail at the lockAndGetLatest() check later.
 162                         $stats->increment( 'refreshlinks.rev_not_current' );
 163                         $this->setLastError( "Revision {$revision->getId()} is not current" );
 164                         return false;
 165                 }
 166
 167                 $content = $revision->getContent( Revision::RAW );
 168                 if ( !$content ) {
 169                         // If there is no content, pretend the content is empty
 170                         $content = $revision->getContentHandler()->makeEmptyContent();
 171                 }
 172
 173                 $parserOutput = false;
 174                 $parserOptions = $page->makeParserOptions( 'canonical' );
 175                 // If page_touched changed after this root job, then it is likely that
 176                 // any views of the pages already resulted in re-parses which are now in
 177                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 178                 if ( isset( $this->params['rootJobTimestamp'] ) ) {
 179                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 180
 181                         $skewedTimestamp = $this->params['rootJobTimestamp'];
 182                         if ( $opportunistic ) {
 183                                 // Neither clock skew nor DB snapshot/slave lag matter much for such
 184                                 // updates; focus on reusing the (often recently updated) cache
 185                         } else {
 186                                 // For transclusion updates, the template changes must be reflected
 187                                 $skewedTimestamp = wfTimestamp( TS_MW,
 188                                         wfTimestamp( TS_UNIX, $skewedTimestamp ) + self::CLOCK_FUDGE
 189                                 );
 190                         }
 191
 192                         if ( $page->getLinksTimestamp() > $skewedTimestamp ) {
 193                                 // Something already updated the backlinks since this job was made
 194                                 $stats->increment( 'refreshlinks.update_skipped' );
 195                                 return true;
 196                         }
 197
 198                         if ( $page->getTouched() >= $this->params['rootJobTimestamp'] || $opportunistic ) {
 199                                 // Cache is suspected to be up-to-date. As long as the cache rev ID matches
 200                                 // and it reflects the job's triggering change, then it is usable.
 201                                 $parserOutput = ParserCache::singleton()->getDirty( $page, $parserOptions );
 202                                 if ( !$parserOutput
 203                                         || $parserOutput->getCacheRevisionId() != $revision->getId()
 204                                         || $parserOutput->getCacheTime() < $skewedTimestamp
 205                                 ) {
 206                                         $parserOutput = false; // too stale
 207                                 }
 208                         }
 209                 }
 210
 211                 // Fetch the current revision and parse it if necessary...
 212                 if ( $parserOutput ) {
 213                         $stats->increment( 'refreshlinks.parser_cached' );
 214                 } else {
 215                         $start = microtime( true );
 216                         // Revision ID must be passed to the parser output to get revision variables correct
 217                         $parserOutput = $content->getParserOutput(
 218                                 $title, $revision->getId(), $parserOptions, false );
 219                         $elapsed = microtime( true ) - $start;
 220                         // If it took a long time to render, then save this back to the cache to avoid
 221                         // wasted CPU by other apaches or job runners. We don't want to always save to
 222                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 223                         if ( $elapsed >= self::PARSE_THRESHOLD_SEC
 224                                 && $page->shouldCheckParserCache( $parserOptions, $revision->getId() )
 225                                 && $parserOutput->isCacheable()
 226                         ) {
 227                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 228                                 ParserCache::singleton()->save(
 229                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 230                                 );
 231                         }
 232                         $stats->increment( 'refreshlinks.parser_uncached' );
 233                 }
 234
 235                 $updates = $content->getSecondaryDataUpdates(
 236                         $title,
 237                         null,
 238                         !empty( $this->params['useRecursiveLinksUpdate'] ),
 239                         $parserOutput
 240                 );
 241
 242                 foreach ( $updates as $key => $update ) {
 243                         // FIXME: This code probably shouldn't be here?
 244                         // Needed by things like Echo notifications which need
 245                         // to know which user caused the links update
 246                         if ( $update instanceof LinksUpdate ) {
 247                                 $update->setRevision( $revision );
 248                                 if ( !empty( $this->params['triggeringUser'] ) ) {
 249                                         $userInfo = $this->params['triggeringUser'];
 250                                         if ( $userInfo['userId'] ) {
 251                                                 $user = User::newFromId( $userInfo['userId'] );
 252                                         } else {
 253                                                 // Anonymous, use the username
 254                                                 $user = User::newFromName( $userInfo['userName'], false );
 255                                         }
 256                                         $update->setTriggeringUser( $user );
 257                                 }
 258                         }
 259                 }
 260
 261                 $latestNow = $page->lockAndGetLatest();
 262                 if ( !$latestNow || $revision->getId() != $latestNow ) {
 263                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 264                         // serialized, it would be OK to update links based on older revisions since it
 265                         // would eventually get to the latest. Since that is not the case (by design),
 266                         // only update the link tables to a state matching the current revision's output.
 267                         $stats->increment( 'refreshlinks.rev_cas_failure' );
 268                         $this->setLastError( "page_latest changed from {$revision->getId()} to $latestNow" );
 269                         return false;
 270                 }
 271
 272                 DataUpdate::runUpdates( $updates );
 273
 274                 InfoAction::invalidateCache( $title );
 275
 276                 return true;
 277         }
 278
 279         public function getDeduplicationInfo() {
 280                 $info = parent::getDeduplicationInfo();
 281                 if ( is_array( $info['params'] ) ) {
 282                         // For per-pages jobs, the job title is that of the template that changed
 283                         // (or similar), so remove that since it ruins duplicate detection
 284                         if ( isset( $info['pages'] ) ) {
 285                                 unset( $info['namespace'] );
 286                                 unset( $info['title'] );
 287                         }
 288                 }
 289
 290                 return $info;
 291         }
 292
 293         public function workItemCount() {
 294                 return isset( $this->params['pages'] ) ? count( $this->params['pages'] ) : 1;
 295         }
 296 }