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
  24 /**
  25  * Job to update link tables for pages
  26  *
  27  * This job comes in a few variants:
  28  *   - a) Recursive jobs to update links for backlink pages for a given title.
  29  *        These jobs have (recursive:true,table:<table>) set.
  30  *   - b) Jobs to update links for a set of pages (the job title is ignored).
  31  *            These jobs have (pages:(<page ID>:(<namespace>,<title>),...) set.
  32  *   - c) Jobs to update links for a single page (the job title)
  33  *        These jobs need no extra fields set.
  34  *
  35  * @ingroup JobQueue
  36  */
  37 class RefreshLinksJob extends Job {
  38         const PARSE_THRESHOLD_SEC = 1.0;
  39
  40         const CLOCK_FUDGE = 10;
  41
  42         function __construct( Title $title, array $params ) {
  43                 parent::__construct( 'refreshLinks', $title, $params );
  44                 // A separate type is used just for cascade-protected backlinks
  45                 if ( !empty( $this->params['prioritize'] ) ) {
  46                         $this->command .= 'Prioritized';
  47                 }
  48                 // Base backlink update jobs and per-title update jobs can be de-duplicated.
  49                 // If template A changes twice before any jobs run, a clean queue will have:
  50                 //              (A base, A base)
  51                 // The second job is ignored by the queue on insertion.
  52                 // Suppose, many pages use template A, and that template itself uses template B.
  53                 // An edit to both will first create two base jobs. A clean FIFO queue will have:
  54                 //              (A base, B base)
  55                 // When these jobs run, the queue will have per-title and remnant partition jobs:
  56                 //              (titleX,titleY,titleZ,...,A remnant,titleM,titleN,titleO,...,B remnant)
  57                 // Some these jobs will be the same, and will automatically be ignored by
  58                 // the queue upon insertion. Some title jobs will run before the duplicate is
  59                 // inserted, so the work will still be done twice in those cases. More titles
  60                 // can be de-duplicated as the remnant jobs continue to be broken down. This
  61                 // works best when $wgUpdateRowsPerJob, and either the pages have few backlinks
  62                 // and/or the backlink sets for pages A and B are almost identical.
  63                 $this->removeDuplicates = !isset( $params['range'] )
  64                         && ( !isset( $params['pages'] ) || count( $params['pages'] ) == 1 );
  65         }
  66
  67         function run() {
  68                 global $wgUpdateRowsPerJob;
  69
  70                 // Job to update all (or a range of) backlink pages for a page
  71                 if ( !empty( $this->params['recursive'] ) ) {
  72                         // Carry over information for de-duplication
  73                         $extraParams = $this->getRootJobParams();
  74                         // Avoid slave lag when fetching templates.
  75                         // When the outermost job is run, we know that the caller that enqueued it must have
  76                         // committed the relevant changes to the DB by now. At that point, record the master
  77                         // position and pass it along as the job recursively breaks into smaller range jobs.
  78                         // Hopefully, when leaf jobs are popped, the slaves will have reached that position.
  79                         if ( isset( $this->params['masterPos'] ) ) {
  80                                 $extraParams['masterPos'] = $this->params['masterPos'];
  81                         } elseif ( wfGetLB()->getServerCount() > 1 ) {
  82                                 $extraParams['masterPos'] = wfGetLB()->getMasterPos();
  83                         } else {
  84                                 $extraParams['masterPos'] = false;
  85                         }
  86                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
  87                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
  88                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
  89                                 $this,
  90                                 $wgUpdateRowsPerJob,
  91                                 1, // job-per-title
  92                                 array( 'params' => $extraParams )
  93                         );
  94                         JobQueueGroup::singleton()->push( $jobs );
  95                 // Job to update link tables for a set of titles
  96                 } elseif ( isset( $this->params['pages'] ) ) {
  97                         foreach ( $this->params['pages'] as $pageId => $nsAndKey ) {
  98                                 list( $ns, $dbKey ) = $nsAndKey;
  99                                 $this->runForTitle( Title::makeTitleSafe( $ns, $dbKey ) );
 100                         }
 101                 // Job to update link tables for a given title
 102                 } else {
 103                         $this->runForTitle( $this->title );
 104                 }
 105
 106                 return true;
 107         }
 108
 109         /**
 110          * @param Title $title
 111          * @return bool
 112          */
 113         protected function runForTitle( Title $title = null ) {
 114                 if ( is_null( $title ) ) {
 115                         $this->setLastError( "refreshLinks: Invalid title" );
 116                         return false;
 117                 }
 118
 119                 // Wait for the DB of the current/next slave DB handle to catch up to the master.
 120                 // This way, we get the correct page_latest for templates or files that just changed
 121                 // milliseconds ago, having triggered this job to begin with.
 122                 if ( isset( $this->params['masterPos'] ) && $this->params['masterPos'] !== false ) {
 123                         wfGetLB()->waitFor( $this->params['masterPos'] );
 124                 }
 125
 126                 // Clear out title cache data from prior job transaction snapshots
 127                 $linkCache = LinkCache::singleton();
 128                 $linkCache->clear();
 129
 130                 // Fetch the current page and revision...
 131                 $page = WikiPage::factory( $title );
 132                 $revision = Revision::newFromTitle( $title, false, Revision::READ_NORMAL );
 133                 if ( !$revision ) {
 134                         $this->setLastError( "refreshLinks: Article not found {$title->getPrefixedDBkey()}" );
 135                         return false; // XXX: what if it was just deleted?
 136                 }
 137
 138                 $content = $revision->getContent( Revision::RAW );
 139                 if ( !$content ) {
 140                         // If there is no content, pretend the content is empty
 141                         $content = $revision->getContentHandler()->makeEmptyContent();
 142                 }
 143
 144                 $parserOutput = false;
 145                 $parserOptions = $page->makeParserOptions( 'canonical' );
 146                 // If page_touched changed after this root job, then it is likely that
 147                 // any views of the pages already resulted in re-parses which are now in
 148                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 149                 if ( isset( $this->params['rootJobTimestamp'] ) ) {
 150                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 151
 152                         $skewedTimestamp = $this->params['rootJobTimestamp'];
 153                         if ( $opportunistic ) {
 154                                 // Neither clock skew nor DB snapshot/slave lag matter much for such
 155                                 // updates; focus on reusing the (often recently updated) cache
 156                         } else {
 157                                 // For transclusion updates, the template changes must be reflected
 158                                 $skewedTimestamp = wfTimestamp( TS_MW,
 159                                         wfTimestamp( TS_UNIX, $skewedTimestamp ) + self::CLOCK_FUDGE
 160                                 );
 161                         }
 162
 163                         if ( $page->getLinksTimestamp() > $skewedTimestamp ) {
 164                                 // Something already updated the backlinks since this job was made
 165                                 return true;
 166                         }
 167
 168                         if ( $page->getTouched() >= $skewedTimestamp || $opportunistic ) {
 169                                 // Something bumped page_touched since this job was made
 170                                 // or the cache is otherwise suspected to be up-to-date
 171                                 $parserOutput = ParserCache::singleton()->getDirty( $page, $parserOptions );
 172                                 if ( $parserOutput && $parserOutput->getCacheTime() < $skewedTimestamp ) {
 173                                         $parserOutput = false; // too stale
 174                                 }
 175                         }
 176                 }
 177
 178                 // Fetch the current revision and parse it if necessary...
 179                 if ( $parserOutput == false ) {
 180                         $start = microtime( true );
 181                         // Revision ID must be passed to the parser output to get revision variables correct
 182                         $parserOutput = $content->getParserOutput(
 183                                 $title, $revision->getId(), $parserOptions, false );
 184                         $elapsed = microtime( true ) - $start;
 185                         // If it took a long time to render, then save this back to the cache to avoid
 186                         // wasted CPU by other apaches or job runners. We don't want to always save to
 187                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 188                         if ( $elapsed >= self::PARSE_THRESHOLD_SEC
 189                                 && $page->shouldCheckParserCache( $parserOptions, $revision->getId() )
 190                                 && $parserOutput->isCacheable()
 191                         ) {
 192                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 193                                 ParserCache::singleton()->save(
 194                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 195                                 );
 196                         }
 197                 }
 198
 199                 $updates = $content->getSecondaryDataUpdates( $title, null, false, $parserOutput );
 200                 DataUpdate::runUpdates( $updates );
 201
 202                 InfoAction::invalidateCache( $title );
 203
 204                 return true;
 205         }
 206
 207         public function getDeduplicationInfo() {
 208                 $info = parent::getDeduplicationInfo();
 209                 if ( is_array( $info['params'] ) ) {
 210                         // Don't let highly unique "masterPos" values ruin duplicate detection
 211                         unset( $info['params']['masterPos'] );
 212                         // For per-pages jobs, the job title is that of the template that changed
 213                         // (or similar), so remove that since it ruins duplicate detection
 214                         if ( isset( $info['pages'] ) ) {
 215                                 unset( $info['namespace'] );
 216                                 unset( $info['title'] );
 217                         }
 218                 }
 219
 220                 return $info;
 221         }
 222
 223         public function workItemCount() {
 224                 return isset( $this->params['pages'] ) ? count( $this->params['pages'] ) : 1;
 225         }
 226 }