<?php
/**
* @todo document
- * @package MediaWiki
- * @subpackage Maintenance
+ * @file
+ * @ingroup Maintenance
*/
-/** */
-define( "REPORTING_INTERVAL", 100 );
-#define( "REPORTING_INTERVAL", 1 );
-
-function refreshLinks( $start, $newOnly = false, $maxLag = false, $end = 0 ) {
- global $wgUser, $wgParser, $wgUseImageResize, $wgUseTidy;
+function refreshLinks( $start, $newOnly = false, $maxLag = false, $end = 0, $redirectsOnly = false, $oldRedirectsOnly = false ) {
+ global $wgUser, $wgParser, $wgUseTidy;
+ $reportingInterval = 100;
$fname = 'refreshLinks';
- $dbr =& wfGetDB( DB_SLAVE );
- $dbw =& wfGetDB( DB_MASTER );
+ $dbr = wfGetDB( DB_SLAVE );
$start = intval( $start );
# Don't generate TeX PNGs (lack of a sensible current directory causes errors anyway)
$wgUser->setOption('math', MW_MATH_SOURCE);
# Don't generate extension images (e.g. Timeline)
- $wgParser->mTagHooks = array();
+ if( method_exists( $wgParser, "clearTagHooks" ) ) {
+ $wgParser->clearTagHooks();
+ }
- # Don't generate thumbnail images
- $wgUseImageResize = false;
+ # Don't use HTML tidy
$wgUseTidy = false;
- if ( $newOnly ) {
- print "Refreshing links from ";
+ $what = $redirectsOnly ? "redirects" : "links";
+
+ if( $oldRedirectsOnly ) {
+ # This entire code path is cut-and-pasted from below. Hurrah.
+ $res = $dbr->query(
+ "SELECT page_id ".
+ "FROM page ".
+ "LEFT JOIN redirect ON page_id=rd_from ".
+ "WHERE page_is_redirect=1 AND rd_from IS NULL AND ".
+ ($end == 0 ? "page_id >= $start"
+ : "page_id BETWEEN $start AND $end"),
+ $fname
+ );
+ $num = $dbr->numRows( $res );
+ print "Refreshing $num old redirects from $start...\n";
+
+ while( $row = $dbr->fetchObject( $res ) ) {
+ if ( !( ++$i % $reportingInterval ) ) {
+ print "$i\n";
+ wfWaitForSlaves( $maxLag );
+ }
+ fixRedirect( $row->page_id );
+ }
+ } elseif( $newOnly ) {
+ print "Refreshing $what from ";
$res = $dbr->select( 'page',
array( 'page_id' ),
array(
'page_is_new' => 1,
- "page_id > $start" ),
+ "page_id >= $start" ),
$fname
);
$num = $dbr->numRows( $res );
$i = 0;
while ( $row = $dbr->fetchObject( $res ) ) {
- if ( !( ++$i % REPORTING_INTERVAL ) ) {
+ if ( !( ++$i % $reportingInterval ) ) {
print "$i\n";
wfWaitForSlaves( $maxLag );
}
-
- fixLinksFromArticle( $row->page_id );
+ if($redirectsOnly)
+ fixRedirect( $row->page_id );
+ else
+ fixLinksFromArticle( $row->page_id );
}
} else {
- print "Refreshing link table.\n";
+ print "Refreshing $what table.\n";
if ( !$end ) {
$end = $dbr->selectField( 'page', 'max(page_id)', false );
}
for ($id = $start; $id <= $end; $id++) {
- if ( !($id % REPORTING_INTERVAL) ) {
+ if ( !($id % $reportingInterval) ) {
print "$id\n";
wfWaitForSlaves( $maxLag );
}
- fixLinksFromArticle( $id );
+ if($redirectsOnly)
+ fixRedirect( $id );
+ else
+ fixLinksFromArticle( $id );
}
+ }
+}
+function fixRedirect( $id ){
+ global $wgTitle, $wgArticle;
+
+ $wgTitle = Title::newFromID( $id );
+ $dbw = wfGetDB( DB_MASTER );
+ if ( is_null( $wgTitle ) ) {
+ return;
}
+ $wgArticle = new Article($wgTitle);
+
+ $rt = $wgArticle->followRedirect();
+
+ if($rt == false || !is_object($rt))
+ return;
+
+ $wgArticle->updateRedirectOn($dbw,$rt);
}
function fixLinksFromArticle( $id ) {
global $wgTitle, $wgParser;
-
+
$wgTitle = Title::newFromID( $id );
- $dbw =& wfGetDB( DB_MASTER );
+ $dbw = wfGetDB( DB_MASTER );
$linkCache =& LinkCache::singleton();
$linkCache->clear();
-
+
if ( is_null( $wgTitle ) ) {
return;
}
$dbw->immediateCommit();
}
-function deleteLinksFromNonexistent( $maxLag = 0 ) {
- $fname = 'deleteLinksFromNonexistent';
-
+/*
+ * Removes non-existing links from pages from pagelinks, imagelinks,
+ * categorylinks, templatelinks and externallinks tables.
+ *
+ * @param $maxLag
+ * @param $batchSize The size of deletion batches
+ *
+ * @author Merlijn van Deen <valhallasw@arctus.nl>
+ */
+function deleteLinksFromNonexistent( $maxLag = 0, $batchSize = 100 ) {
wfWaitForSlaves( $maxLag );
+
+ $dbw = wfGetDB( DB_MASTER );
- $dbw =& wfGetDB( DB_WRITE );
-
- $linksTables = array(
+ $lb = wfGetLBFactory()->newMainLB();
+ $dbr = $lb->getConnection( DB_SLAVE );
+ $dbr->bufferResults( false );
+
+ $linksTables = array( // table name => page_id field
'pagelinks' => 'pl_from',
'imagelinks' => 'il_from',
'categorylinks' => 'cl_from',
'templatelinks' => 'tl_from',
'externallinks' => 'el_from',
);
-
- $page = $dbw->tableName( 'page' );
-
-
+
foreach ( $linksTables as $table => $field ) {
- if ( !$dbw->ping() ) {
- print "DB disconnected, reconnecting...";
- while ( !$dbw->ping() ) {
- print ".";
- sleep(10);
+ print "Retrieving illegal entries from $table... ";
+
+ // SELECT DISTINCT( $field ) FROM $table LEFT JOIN page ON $field=page_id WHERE page_id IS NULL;
+ $results = $dbr->select( array( $table, 'page' ),
+ $field,
+ array('page_id' => null ),
+ __METHOD__,
+ 'DISTINCT',
+ array( 'page' => array( 'LEFT JOIN', "$field=page_id"))
+ );
+
+ $counter = 0;
+ $list = array();
+ print "0..";
+
+ foreach( $results as $row ) {
+ $counter++;
+ $list[] = $row->$field;
+ if ( ( $counter % $batchSize ) == 0 ) {
+ wfWaitForSlaves(5);
+ $dbw->delete( $table, array( $field => $list ), __METHOD__ );
+
+ print $counter . "..";
+ $list = array();
}
- print "\n";
}
-
- $pTable = $dbw->tableName( $table );
- $sql = "DELETE $pTable FROM $pTable LEFT JOIN $page ON page_id=$field WHERE page_id IS NULL";
-
- print "Deleting $table from non-existent articles...";
- $dbw->query( $sql, $fname );
- print " fixed " .$dbw->affectedRows() . " row(s)\n";
+
+ print $counter;
+ if (count($list) > 0) {
+ $dbw->delete( $table, array( $field => $list ), __METHOD__ );
+ }
+
+ print "\n";
}
+
+ $lb->closeAll();
}
-
-?>