maintenance/updateCollation.php

   1 <?php
   2 /**
   3  * @file
   4  * @ingroup Maintenance
   5  * @author Aryeh Gregor (Simetrical)
   6  */
   7
   8 #$optionsWithArgs = array( 'begin', 'max-slave-lag' );
   9
  10 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  11
  12 class UpdateCollation extends Maintenance {
  13         const BATCH_SIZE = 1000;
  14
  15         public function __construct() {
  16                 parent::__construct();
  17
  18                 global $wgCollationVersion;
  19                 $this->mDescription = <<<TEXT
  20 This script will find all rows in the categorylinks table whose collation is
  21 out-of-date (cl_collation != $wgCollationVersion) and repopulate cl_sortkey
  22 using the page title and cl_sortkey_prefix.  If everything's collation is
  23 up-to-date, it will do nothing.
  24 TEXT;
  25
  26                 #$this->addOption( 'force', 'Run on all rows, even if the collation is supposed to be up-to-date.' );
  27         }
  28
  29         public function execute() {
  30                 global $wgCollationVersion, $wgContLang;
  31
  32                 $dbw = wfGetDB( DB_MASTER );
  33                 $count = $dbw->estimateRowCount(
  34                         'categorylinks',
  35                         array( 'cl_from', 'cl_to', 'cl_sortkey_prefix' ),
  36                         'cl_collation != ' . $dbw->addQuotes( $wgCollationVersion ),
  37                         __METHOD__
  38                 );
  39
  40                 $this->output( "Fixing around $count rows (estimate might be wrong).\n" );
  41
  42                 $count = 0;
  43                 do {
  44                         $res = $dbw->select(
  45                                 array( 'categorylinks', 'page' ),
  46                                 array( 'cl_from', 'cl_to', 'cl_sortkey_prefix', 'cl_collation',
  47                                         'cl_sortkey', 'page_namespace', 'page_title'
  48                                 ),
  49                                 array(
  50                                         'cl_collation != ' . $dbw->addQuotes( $wgCollationVersion ),
  51                                         'cl_from = page_id'
  52                                 ),
  53                                 __METHOD__,
  54                                 array( 'LIMIT' => self::BATCH_SIZE )
  55                         );
  56
  57                         $dbw->begin();
  58                         foreach ( $res as $row ) {
  59                                 $title = Title::newFromRow( $row );
  60                                 if ( $row->cl_collation == 0 ) {
  61                                         # This is an old-style row, so the sortkey needs to be
  62                                         # converted.
  63                                         if ( $row->cl_sortkey == $title->getCategorySortkey() ) {
  64                                                 $prefix = '';
  65                                         } else {
  66                                                 # Custom sortkey, use it as a prefix
  67                                                 $prefix = $row->cl_sortkey;
  68                                         }
  69                                 } else {
  70                                         $prefix = $row->cl_sortkey_prefix;
  71                                 }
  72                                 # cl_type will be wrong for lots of pages if cl_collation is 0,
  73                                 # so let's update it while we're here.
  74                                 if ( $title->getNamespace() == NS_CATEGORY ) {
  75                                         $type = 'subcat';
  76                                 } elseif ( $title->getNamespace() == NS_FILE ) {
  77                                         $type = 'file';
  78                                 } else {
  79                                         $type = 'page';
  80                                 }
  81                                 $dbw->update(
  82                                         'categorylinks',
  83                                         array(
  84                                                 'cl_sortkey' => $wgContLang->convertToSortkey(
  85                                                         $title->getCategorySortkey( $prefix ) ),
  86                                                 'cl_sortkey_prefix' => $prefix,
  87                                                 'cl_collation' => $wgCollationVersion,
  88                                                 'cl_type' => $type,
  89                                         ),
  90                                         array( 'cl_from' => $row->cl_from, 'cl_to' => $row->cl_to ),
  91                                         __METHOD__
  92                                 );
  93                         }
  94                         $dbw->commit();
  95
  96                         $count += $res->numRows();
  97                         $this->output( "$count done.\n" );
  98                 } while ( $res->numRows() == self::BATCH_SIZE );
  99         }
 100 }
 101
 102 $maintClass = "UpdateCollation";
 103 require_once( DO_MAINTENANCE );