From 48a77e1d830e93e28375e1be2252c572af44eae5 Mon Sep 17 00:00:00 2001
From: Aaron Schulz <aschulz@wikimedia.org>
Date: Wed, 27 Nov 2013 22:43:00 -0800
Subject: [PATCH] Avoid parsing more in refreshLinksJobs

* This reuses the parser cache in some cases when possible
* Clarified the return value of CacheTime::getCacheTime()
* A few documentation tweaks

Change-Id: I80b7c6404b3f8c48b53c3bba96115dbf94d80873
---
 includes/job/Job.php                  |  4 ++-
 includes/job/jobs/RefreshLinksJob.php | 40 +++++++++++++++++++--------
 includes/parser/CacheTime.php         |  7 +++--
 3 files changed, 37 insertions(+), 14 deletions(-)

diff --git a/includes/job/Job.php b/includes/job/Job.php
index 3f44a918ff..e33baf535a 100644
--- a/includes/job/Job.php
+++ b/includes/job/Job.php
@@ -241,7 +241,9 @@ abstract class Job {
 	/**
 	 * @see JobQueue::deduplicateRootJob()
 	 * @param string $key A key that identifies the task
-	 * @return array
+	 * @return array Map of:
+	 *   - rootJobSignature : hash (e.g. SHA1) that identifies the task
+	 *   - rootJobTimestamp : TS_MW timestamp of this instance of the task
 	 * @since 1.21
 	 */
 	public static function newRootJobParams( $key ) {
diff --git a/includes/job/jobs/RefreshLinksJob.php b/includes/job/jobs/RefreshLinksJob.php
index 0372d85927..ea1d5962ec 100644
--- a/includes/job/jobs/RefreshLinksJob.php
+++ b/includes/job/jobs/RefreshLinksJob.php
@@ -119,20 +119,38 @@ class RefreshLinksJob extends Job {
 			wfGetLB()->waitFor( $this->params['masterPos'] );
 		}
 
-		$revision = Revision::newFromTitle( $title, false, Revision::READ_NORMAL );
-		if ( !$revision ) {
-			$this->setLastError( "refreshLinks: Article not found {$title->getPrefixedDBkey()}" );
-			return false; // XXX: what if it was just deleted?
+		$parserOutput = false;
+		// If page_touched changed after this root job (with a good slave lag skew factor),
+		// then it is likely that any views of the pages already resulted in re-parses which
+		// are now in cache. This can be reused to avoid expensive parsing in some cases.
+		if ( isset( $this->params['rootJobTimestamp'] ) ) {
+			$page = WikiPage::factory( $title );
+			$skewedTimestamp = wfTimestamp( TS_UNIX, $this->params['rootJobTimestamp'] ) + 5;
+			if ( $page->getTouched() > wfTimestamp( TS_MW, $skewedTimestamp ) ) {
+				$parserOptions = $page->makeParserOptions( 'canonical' );
+				$parserOutput = ParserCache::singleton()->getDirty( $page, $parserOptions );
+				if ( $parserOutput->getCacheTime() <= $skewedTimestamp ) {
+					$parserOutput = false; // too stale
+				}
+			}
 		}
+		// Fetch the current revision and parse it if necessary...
+		if ( $parserOutput == false ) {
+			$revision = Revision::newFromTitle( $title, false, Revision::READ_NORMAL );
+			if ( !$revision ) {
+				$this->setLastError( "refreshLinks: Article not found {$title->getPrefixedDBkey()}" );
+				return false; // XXX: what if it was just deleted?
+			}
 
-		$content = $revision->getContent( Revision::RAW );
-		if ( !$content ) {
-			// If there is no content, pretend the content is empty
-			$content = $revision->getContentHandler()->makeEmptyContent();
-		}
+			$content = $revision->getContent( Revision::RAW );
+			if ( !$content ) {
+				// If there is no content, pretend the content is empty
+				$content = $revision->getContentHandler()->makeEmptyContent();
+			}
 
-		// Revision ID must be passed to the parser output to get revision variables correct
-		$parserOutput = $content->getParserOutput( $title, $revision->getId(), null, false );
+			// Revision ID must be passed to the parser output to get revision variables correct
+			$parserOutput = $content->getParserOutput( $title, $revision->getId(), null, false );
+		}
 
 		$updates = $content->getSecondaryDataUpdates( $title, null, false, $parserOutput );
 		DataUpdate::runUpdates( $updates );
diff --git a/includes/parser/CacheTime.php b/includes/parser/CacheTime.php
index a4203b0894..7b8935a646 100644
--- a/includes/parser/CacheTime.php
+++ b/includes/parser/CacheTime.php
@@ -37,14 +37,17 @@ class CacheTime {
 		$mCacheExpiry = null,         # Seconds after which the object should expire, use 0 for uncachable. Used in ParserCache.
 		$mContainsOldMagic;           # Boolean variable indicating if the input contained variables like {{CURRENTDAY}}
 
-	function getCacheTime()              { return $this->mCacheTime; }
+	/**
+	 * @return string TS_MW timestamp
+	 */
+	function getCacheTime()              { return wfTimestamp( TS_MW, $this->mCacheTime ); }
 
 	function containsOldMagic()          { return $this->mContainsOldMagic; }
 	function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
 
 	/**
 	 * setCacheTime() sets the timestamp expressing when the page has been rendered.
-	 * This doesn not control expiry, see updateCacheExpiry() for that!
+	 * This does not control expiry, see updateCacheExpiry() for that!
 	 * @param $t string
 	 * @return string
 	 */
-- 
2.20.1