Merge "Http::getProxy() method to get proxy configuration"
[lhc/web/wiklou.git] / includes / jobqueue / jobs / HTMLCacheUpdateJob.php
1 <?php
2 /**
3 * HTML cache invalidation of all pages linking to a given title.
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 * @ingroup Cache
23 */
24
25 /**
26 * Job to purge the cache for all pages that link to or use another page or file
27 *
28 * This job comes in a few variants:
29 * - a) Recursive jobs to purge caches for backlink pages for a given title.
30 * These jobs have (recursive:true,table:<table>) set.
31 * - b) Jobs to purge caches for a set of titles (the job title is ignored).
32 * These jobs have (pages:(<page ID>:(<namespace>,<title>),...) set.
33 *
34 * @ingroup JobQueue
35 */
36 class HTMLCacheUpdateJob extends Job {
37 function __construct( Title $title, array $params ) {
38 parent::__construct( 'htmlCacheUpdate', $title, $params );
39 // Base backlink purge jobs can be de-duplicated
40 $this->removeDuplicates = ( !isset( $params['range'] ) && !isset( $params['pages'] ) );
41 }
42
43 /**
44 * @param Title $title Title to purge backlink pages from
45 * @param string $table Backlink table name
46 * @return HTMLCacheUpdateJob
47 */
48 public static function newForBacklinks( Title $title, $table ) {
49 return new self(
50 $title,
51 [
52 'table' => $table,
53 'recursive' => true
54 ] + Job::newRootJobParams( // "overall" refresh links job info
55 "htmlCacheUpdate:{$table}:{$title->getPrefixedText()}"
56 )
57 );
58 }
59
60 function run() {
61 global $wgUpdateRowsPerJob, $wgUpdateRowsPerQuery;
62
63 if ( isset( $this->params['table'] ) && !isset( $this->params['pages'] ) ) {
64 $this->params['recursive'] = true; // b/c; base job
65 }
66
67 // Job to purge all (or a range of) backlink pages for a page
68 if ( !empty( $this->params['recursive'] ) ) {
69 // Convert this into no more than $wgUpdateRowsPerJob HTMLCacheUpdateJob per-title
70 // jobs and possibly a recursive HTMLCacheUpdateJob job for the rest of the backlinks
71 $jobs = BacklinkJobUtils::partitionBacklinkJob(
72 $this,
73 $wgUpdateRowsPerJob,
74 $wgUpdateRowsPerQuery, // jobs-per-title
75 // Carry over information for de-duplication
76 [ 'params' => $this->getRootJobParams() ]
77 );
78 JobQueueGroup::singleton()->push( $jobs );
79 // Job to purge pages for a set of titles
80 } elseif ( isset( $this->params['pages'] ) ) {
81 $this->invalidateTitles( $this->params['pages'] );
82 // Job to update a single title
83 } else {
84 $t = $this->title;
85 $this->invalidateTitles( [
86 $t->getArticleID() => [ $t->getNamespace(), $t->getDBkey() ]
87 ] );
88 }
89
90 return true;
91 }
92
93 /**
94 * @param array $pages Map of (page ID => (namespace, DB key)) entries
95 */
96 protected function invalidateTitles( array $pages ) {
97 global $wgUpdateRowsPerQuery, $wgUseFileCache;
98
99 // Get all page IDs in this query into an array
100 $pageIds = array_keys( $pages );
101 if ( !$pageIds ) {
102 return;
103 }
104
105 // The page_touched field will need to be bumped for these pages.
106 // Only bump it to the present time if no "rootJobTimestamp" was known.
107 // If it is known, it can be used instead, which avoids invalidating output
108 // that was in fact generated *after* the relevant dependency change time
109 // (e.g. template edit). This is particularily useful since refreshLinks jobs
110 // save back parser output and usually run along side htmlCacheUpdate jobs;
111 // their saved output would be invalidated by using the current timestamp.
112 if ( isset( $this->params['rootJobTimestamp'] ) ) {
113 $touchTimestamp = $this->params['rootJobTimestamp'];
114 } else {
115 $touchTimestamp = wfTimestampNow();
116 }
117
118 $dbw = wfGetDB( DB_MASTER );
119 // Update page_touched (skipping pages already touched since the root job).
120 // Check $wgUpdateRowsPerQuery for sanity; batch jobs are sized by that already.
121 foreach ( array_chunk( $pageIds, $wgUpdateRowsPerQuery ) as $batch ) {
122 $dbw->commit( __METHOD__, 'flush' );
123 wfGetLBFactory()->waitForReplication();
124
125 $dbw->update( 'page',
126 [ 'page_touched' => $dbw->timestamp( $touchTimestamp ) ],
127 [ 'page_id' => $batch,
128 // don't invalidated pages that were already invalidated
129 "page_touched < " . $dbw->addQuotes( $dbw->timestamp( $touchTimestamp ) )
130 ],
131 __METHOD__
132 );
133 }
134 // Get the list of affected pages (races only mean something else did the purge)
135 $titleArray = TitleArray::newFromResult( $dbw->select(
136 'page',
137 [ 'page_namespace', 'page_title' ],
138 [ 'page_id' => $pageIds, 'page_touched' => $dbw->timestamp( $touchTimestamp ) ],
139 __METHOD__
140 ) );
141
142 // Update CDN
143 $u = CdnCacheUpdate::newFromTitles( $titleArray );
144 $u->doUpdate();
145
146 // Update file cache
147 if ( $wgUseFileCache ) {
148 foreach ( $titleArray as $title ) {
149 HTMLFileCache::clearFileCache( $title );
150 }
151 }
152 }
153
154 public function workItemCount() {
155 return isset( $this->params['pages'] ) ? count( $this->params['pages'] ) : 1;
156 }
157 }