From: Aaron Schulz <aschulz@wikimedia.org>
Date: Thu, 14 Sep 2017 09:25:23 +0000 (+0200)
Subject: De-duplicate HTMLCacheUpdate jobs with a page array of size 1
X-Git-Tag: 1.31.0-rc.0~2102^2
X-Git-Url: http://git.cyclocoop.org/%22%2C%20generer_url_ecrire%28?a=commitdiff_plain;h=d0c373fdf2f28e9b541daa68e5b0e97666099dcf;p=lhc%2Fweb%2Fwiklou.git

De-duplicate HTMLCacheUpdate jobs with a page array of size 1

BacklinkJobUtils consistently uses the "pages" field in leaf
jobs, even when there is only one page per leaf job.

RefreshLinksJob already has this logic for de-duplication.

Change-Id: Ia189bbc9df44f2161cfed4192c23b2ac3cfa65ce
---

diff --git a/includes/jobqueue/jobs/HTMLCacheUpdateJob.php b/includes/jobqueue/jobs/HTMLCacheUpdateJob.php
index 4c16d7f25b..0aa33cac18 100644
--- a/includes/jobqueue/jobs/HTMLCacheUpdateJob.php
+++ b/includes/jobqueue/jobs/HTMLCacheUpdateJob.php
@@ -38,8 +38,15 @@ use MediaWiki\MediaWikiServices;
 class HTMLCacheUpdateJob extends Job {
 	function __construct( Title $title, array $params ) {
 		parent::__construct( 'htmlCacheUpdate', $title, $params );
-		// Base backlink purge jobs can be de-duplicated
-		$this->removeDuplicates = ( !isset( $params['range'] ) && !isset( $params['pages'] ) );
+		// Avoid the overhead of de-duplication when it would be pointless.
+		// Note that these jobs always set page_touched to the current time,
+		// so letting the older existing job "win" is still correct.
+		$this->removeDuplicates = (
+			// Ranges rarely will line up
+			!isset( $params['range'] ) &&
+			// Multiple pages per job make matches unlikely
+			!( isset( $params['pages'] ) && count( $params['pages'] ) != 1 )
+		);
 	}
 
 	/**