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                 // Base backlink update jobs and per-title update jobs can be de-duplicated.
  45                 // If template A changes twice before any jobs run, a clean queue will have:
  46                 //              (A base, A base)
  47                 // The second job is ignored by the queue on insertion.
  48                 // Suppose, many pages use template A, and that template itself uses template B.
  49                 // An edit to both will first create two base jobs. A clean FIFO queue will have:
  50                 //              (A base, B base)
  51                 // When these jobs run, the queue will have per-title and remnant partition jobs:
  52                 //              (titleX,titleY,titleZ,...,A remnant,titleM,titleN,titleO,...,B remnant)
  53                 // Some these jobs will be the same, and will automatically be ignored by
  54                 // the queue upon insertion. Some title jobs will run before the duplicate is
  55                 // inserted, so the work will still be done twice in those cases. More titles
  56                 // can be de-duplicated as the remnant jobs continue to be broken down. This
  57                 // works best when $wgUpdateRowsPerJob, and either the pages have few backlinks
  58                 // and/or the backlink sets for pages A and B are almost identical.
  59                 $this->removeDuplicates = !isset( $params['range'] )
  60                         && ( !isset( $params['pages'] ) || count( $params['pages'] ) == 1 );
  61         }
  62
  63         /**
  64          * @param Title $title
  65          * @param array $params
  66          * @return RefreshLinksJob
  67          */
  68         public static function newPrioritized( Title $title, array $params ) {
  69                 $job = new self( $title, $params );
  70                 $job->command = 'refreshLinksPrioritized';
  71
  72                 return $job;
  73         }
  74
  75         /**
  76          * @param Title $title
  77          * @param array $params
  78          * @return RefreshLinksJob
  79          */
  80         public static function newDynamic( Title $title, array $params ) {
  81                 $job = new self( $title, $params );
  82                 $job->command = 'refreshLinksDynamic';
  83
  84                 return $job;
  85         }
  86
  87         function run() {
  88                 global $wgUpdateRowsPerJob;
  89
  90                 // Job to update all (or a range of) backlink pages for a page
  91                 if ( !empty( $this->params['recursive'] ) ) {
  92                         // Carry over information for de-duplication
  93                         $extraParams = $this->getRootJobParams();
  94                         // Avoid slave lag when fetching templates.
  95                         // When the outermost job is run, we know that the caller that enqueued it must have
  96                         // committed the relevant changes to the DB by now. At that point, record the master
  97                         // position and pass it along as the job recursively breaks into smaller range jobs.
  98                         // Hopefully, when leaf jobs are popped, the slaves will have reached that position.
  99                         if ( isset( $this->params['masterPos'] ) ) {
 100                                 $extraParams['masterPos'] = $this->params['masterPos'];
 101                         } elseif ( wfGetLB()->getServerCount() > 1 ) {
 102                                 $extraParams['masterPos'] = wfGetLB()->getMasterPos();
 103                         } else {
 104                                 $extraParams['masterPos'] = false;
 105                         }
 106                         $extraParams['triggeredRecursive'] = true;
 107                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 108                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 109                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 110                                 $this,
 111                                 $wgUpdateRowsPerJob,
 112                                 1, // job-per-title
 113                                 array( 'params' => $extraParams )
 114                         );
 115                         JobQueueGroup::singleton()->push( $jobs );
 116                 // Job to update link tables for a set of titles
 117                 } elseif ( isset( $this->params['pages'] ) ) {
 118                         $this->waitForMasterPosition();
 119                         foreach ( $this->params['pages'] as $pageId => $nsAndKey ) {
 120                                 list( $ns, $dbKey ) = $nsAndKey;
 121                                 $this->runForTitle( Title::makeTitleSafe( $ns, $dbKey ) );
 122                         }
 123                 // Job to update link tables for a given title
 124                 } else {
 125                         $this->waitForMasterPosition();
 126                         $this->runForTitle( $this->title );
 127                 }
 128
 129                 return true;
 130         }
 131
 132         protected function waitForMasterPosition() {
 133                 if ( !empty( $this->params['masterPos'] ) && wfGetLB()->getServerCount() > 1 ) {
 134                         // Wait for the current/next slave DB handle to catch up to the master.
 135                         // This way, we get the correct page_latest for templates or files that just
 136                         // changed milliseconds ago, having triggered this job to begin with.
 137                         wfGetLB()->waitFor( $this->params['masterPos'] );
 138                 }
 139         }
 140
 141         /**
 142          * @param Title $title
 143          * @return bool
 144          */
 145         protected function runForTitle( Title $title ) {
 146                 $page = WikiPage::factory( $title );
 147                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 148                         // Fetch the specified revision; lockAndGetLatest() below detects if the page
 149                         // was edited since and aborts in order to avoid corrupting the link tables
 150                         $revision = Revision::newFromId(
 151                                 $this->params['triggeringRevisionId'],
 152                                 Revision::READ_LATEST
 153                         );
 154                 } else {
 155                         // Fetch current revision; READ_LATEST reduces lockAndGetLatest() check failures
 156                         $revision = Revision::newFromTitle( $title, false, Revision::READ_LATEST );
 157                 }
 158
 159                 if ( !$revision ) {
 160                         $this->setLastError( "Revision not found for {$title->getPrefixedDBkey()}" );
 161                         return false; // just deleted?
 162                 }
 163
 164                 $content = $revision->getContent( Revision::RAW );
 165                 if ( !$content ) {
 166                         // If there is no content, pretend the content is empty
 167                         $content = $revision->getContentHandler()->makeEmptyContent();
 168                 }
 169
 170                 $parserOutput = false;
 171                 $parserOptions = $page->makeParserOptions( 'canonical' );
 172                 // If page_touched changed after this root job, then it is likely that
 173                 // any views of the pages already resulted in re-parses which are now in
 174                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 175                 if ( isset( $this->params['rootJobTimestamp'] ) ) {
 176                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 177
 178                         $skewedTimestamp = $this->params['rootJobTimestamp'];
 179                         if ( $opportunistic ) {
 180                                 // Neither clock skew nor DB snapshot/slave lag matter much for such
 181                                 // updates; focus on reusing the (often recently updated) cache
 182                         } else {
 183                                 // For transclusion updates, the template changes must be reflected
 184                                 $skewedTimestamp = wfTimestamp( TS_MW,
 185                                         wfTimestamp( TS_UNIX, $skewedTimestamp ) + self::CLOCK_FUDGE
 186                                 );
 187                         }
 188
 189                         if ( $page->getLinksTimestamp() > $skewedTimestamp ) {
 190                                 // Something already updated the backlinks since this job was made
 191                                 return true;
 192                         }
 193
 194                         if ( $page->getTouched() >= $skewedTimestamp || $opportunistic ) {
 195                                 // Something bumped page_touched since this job was made
 196                                 // or the cache is otherwise suspected to be up-to-date
 197                                 $parserOutput = ParserCache::singleton()->getDirty( $page, $parserOptions );
 198                                 if ( $parserOutput && $parserOutput->getCacheTime() < $skewedTimestamp ) {
 199                                         $parserOutput = false; // too stale
 200                                 }
 201                         }
 202                 }
 203
 204                 // Fetch the current revision and parse it if necessary...
 205                 if ( $parserOutput == false ) {
 206                         $start = microtime( true );
 207                         // Revision ID must be passed to the parser output to get revision variables correct
 208                         $parserOutput = $content->getParserOutput(
 209                                 $title, $revision->getId(), $parserOptions, false );
 210                         $elapsed = microtime( true ) - $start;
 211                         // If it took a long time to render, then save this back to the cache to avoid
 212                         // wasted CPU by other apaches or job runners. We don't want to always save to
 213                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 214                         if ( $elapsed >= self::PARSE_THRESHOLD_SEC
 215                                 && $page->shouldCheckParserCache( $parserOptions, $revision->getId() )
 216                                 && $parserOutput->isCacheable()
 217                         ) {
 218                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 219                                 ParserCache::singleton()->save(
 220                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 221                                 );
 222                         }
 223                 }
 224
 225                 $updates = $content->getSecondaryDataUpdates(
 226                         $title,
 227                         null,
 228                         !empty( $this->params['useRecursiveLinksUpdate'] ),
 229                         $parserOutput
 230                 );
 231
 232                 foreach ( $updates as $key => $update ) {
 233                         // FIXME: move category change RC stuff to a separate update.
 234                         // RC entry addition aborts if edits where since made, which is not necessary.
 235                         // It's also an SoC violation for links update code to care about RC.
 236                         if ( $update instanceof LinksUpdate ) {
 237                                 if ( !empty( $this->params['triggeredRecursive'] ) ) {
 238                                         $update->setTriggeredRecursive();
 239                                 }
 240                                 if ( !empty( $this->params['triggeringUser'] ) ) {
 241                                         $userInfo = $this->params['triggeringUser'];
 242                                         if ( $userInfo['userId'] ) {
 243                                                 $user = User::newFromId( $userInfo['userId'] );
 244                                         } else {
 245                                                 // Anonymous, use the username
 246                                                 $user = User::newFromName( $userInfo['userName'], false );
 247                                         }
 248                                         $update->setTriggeringUser( $user );
 249                                 }
 250                                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 251                                         $update->setRevision( $revision );
 252                                 }
 253                         }
 254                 }
 255
 256                 $latestNow = $page->lockAndGetLatest();
 257                 if ( !$latestNow || $revision->getId() != $latestNow ) {
 258                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 259                         // serialized, it would be OK to update links based on older revisions since it
 260                         // would eventually get to the latest. Since that is not the case (by design),
 261                         // only update the link tables to a state matching the current revision's output.
 262                         $this->setLastError( "page_latest changed from {$revision->getId()} to $latestNow" );
 263                         return false;
 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                         // Don't let highly unique "masterPos" values ruin duplicate detection
 277                         unset( $info['params']['masterPos'] );
 278                         // For per-pages jobs, the job title is that of the template that changed
 279                         // (or similar), so remove that since it ruins duplicate detection
 280                         if ( isset( $info['pages'] ) ) {
 281                                 unset( $info['namespace'] );
 282                                 unset( $info['title'] );
 283                         }
 284                 }
 285
 286                 return $info;
 287         }
 288
 289         public function workItemCount() {
 290                 return isset( $this->params['pages'] ) ? count( $this->params['pages'] ) : 1;
 291         }
 292 }