1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
|
<?php
/**
* Class to invalidate the HTML cache of all the pages linking to a given title.
* Small numbers of links will be done immediately, large numbers are pushed onto
* the job queue.
*
* This class is designed to work efficiently with small numbers of links, and
* to work reasonably well with up to ~10^5 links. Above ~10^6 links, the memory
* and time requirements of loading all backlinked IDs in doUpdate() might become
* prohibitive. The requirements measured at Wikimedia are approximately:
*
* memory: 48 bytes per row
* time: 16us per row for the query plus processing
*
* The reason this query is done is to support partitioning of the job
* by backlinked ID. The memory issue could be allieviated by doing this query in
* batches, but of course LIMIT with an offset is inefficient on the DB side.
*
* The class is nevertheless a vast improvement on the previous method of using
* Image::getLinksTo() and Title::touchArray(), which uses about 2KB of memory per
* link.
*
* @ingroup Cache
*/
class HTMLCacheUpdate
{
public $mTitle, $mTable, $mPrefix;
public $mRowsPerJob, $mRowsPerQuery;
function __construct( $titleTo, $table ) {
global $wgUpdateRowsPerJob, $wgUpdateRowsPerQuery;
$this->mTitle = $titleTo;
$this->mTable = $table;
$this->mRowsPerJob = $wgUpdateRowsPerJob;
$this->mRowsPerQuery = $wgUpdateRowsPerQuery;
$this->mCache = $this->mTitle->getBacklinkCache();
}
public function doUpdate() {
# Fetch the IDs
$numRows = $this->mCache->getNumLinks( $this->mTable );
if ( $numRows != 0 ) {
if ( $numRows > $this->mRowsPerJob ) {
$this->insertJobs();
} else {
$this->invalidate();
}
}
wfRunHooks( 'HTMLCacheUpdate::doUpdate', array($this->mTitle) );
}
protected function insertJobs() {
$batches = $this->mCache->partition( $this->mTable, $this->mRowsPerJob );
if ( !$batches ) {
return;
}
foreach ( $batches as $batch ) {
$params = array(
'table' => $this->mTable,
'start' => $batch[0],
'end' => $batch[1],
);
$jobs[] = new HTMLCacheUpdateJob( $this->mTitle, $params );
}
Job::batchInsert( $jobs );
}
/**
* Invalidate a set of pages, right now
*/
public function invalidate( $startId = false, $endId = false ) {
global $wgUseFileCache, $wgUseSquid;
$titleArray = $this->mCache->getLinks( $this->mTable, $startId, $endId );
if ( $titleArray->count() == 0 ) {
return;
}
$dbw = wfGetDB( DB_MASTER );
$timestamp = $dbw->timestamp();
# Get all IDs in this query into an array
$ids = array();
foreach ( $titleArray as $title ) {
$ids[] = $title->getArticleID();
}
# Update page_touched
$dbw->update( 'page',
array( 'page_touched' => $timestamp ),
array( 'page_id IN (' . $dbw->makeList( $ids ) . ')' ),
__METHOD__
);
# Update squid
if ( $wgUseSquid ) {
$u = SquidUpdate::newFromTitles( $titleArray );
$u->doUpdate();
}
# Update file cache
if ( $wgUseFileCache ) {
foreach ( $titleArray as $title ) {
HTMLFileCache::clearFileCache( $title );
}
}
}
}
/**
* Job wrapper for HTMLCacheUpdate. Gets run whenever a related
* job gets called from the queue.
*
* @ingroup JobQueue
*/
class HTMLCacheUpdateJob extends Job {
var $table, $start, $end;
/**
* Construct a job
* @param Title $title The title linked to
* @param array $params Job parameters (table, start and end page_ids)
* @param integer $id job_id
*/
function __construct( $title, $params, $id = 0 ) {
parent::__construct( 'htmlCacheUpdate', $title, $params, $id );
$this->table = $params['table'];
$this->start = $params['start'];
$this->end = $params['end'];
}
public function run() {
$update = new HTMLCacheUpdate( $this->title, $this->table );
$update->invalidate( $this->start, $this->end );
return true;
}
}
|
