From 222b01f5169f1c7e69762e0e8904c24f78f71882 Mon Sep 17 00:00:00 2001 From: Pierre Schmitz Date: Wed, 28 Jul 2010 11:52:48 +0200 Subject: update to MediaWiki 1.16.0 --- maintenance/cleanupTable.inc | 161 ++++++++++++++++++++++++++++++++----------- 1 file changed, 119 insertions(+), 42 deletions(-) (limited to 'maintenance/cleanupTable.inc') diff --git a/maintenance/cleanupTable.inc b/maintenance/cleanupTable.inc index 75699c52..3549a9a1 100644 --- a/maintenance/cleanupTable.inc +++ b/maintenance/cleanupTable.inc @@ -1,31 +1,58 @@ targetTable = $table; - $this->maxLag = 10; # if slaves are lagged more than 10 secs, wait - $this->dryrun = $dryrun; +require_once( dirname(__FILE__) . '/Maintenance.php' ); + +class TableCleanup extends Maintenance { + protected $defaultParams = array( + 'table' => 'page', + 'conds' => array(), + 'index' => 'page_id', + 'callback' => 'processRow', + ); + + protected $dryrun = false; + protected $maxLag = 10; # if slaves are lagged more than 10 secs, wait + public $batchSize = 100; + public $reportInterval = 100; + + public function __construct() { + parent::__construct(); + $this->addOption( 'dry-run', 'Perform a dry run' ); } - function cleanup() { + public function execute() { + global $wgUser; + $wgUser->setName( 'Conversion script' ); + $this->dryrun = $this->hasOption( 'dry-run' ); if( $this->dryrun ) { - echo "Checking for bad titles...\n"; + $this->output( "Checking for bad titles...\n" ); } else { - echo "Checking and fixing bad titles...\n"; + $this->output( "Checking and fixing bad titles...\n" ); } - $this->runTable( $this->targetTable, - '', //'WHERE page_namespace=0', - array( $this, 'processPage' ) ); + $this->runTable( $this->defaultParams ); } - function init( $count, $table ) { + protected function init( $count, $table ) { $this->processed = 0; $this->updated = 0; $this->count = $count; @@ -33,10 +60,10 @@ abstract class TableCleanup extends FiveUpgrade { $this->table = $table; } - function progress( $updated ) { + protected function progress( $updated ) { $this->updated += $updated; $this->processed++; - if( $this->processed % 100 != 0 ) { + if( $this->processed % $this->reportInterval != 0 ) { return; } $portion = $this->processed / $this->count; @@ -47,39 +74,89 @@ abstract class TableCleanup extends FiveUpgrade { $estimatedTotalTime = $delta / $portion; $eta = $this->startTime + $estimatedTotalTime; - printf( "%s %s: %6.2f%% done on %s; ETA %s [%d/%d] %.2f/sec <%.2f%% updated>\n", - wfWikiID(), - wfTimestamp( TS_DB, intval( $now ) ), - $portion * 100.0, - $this->table, - wfTimestamp( TS_DB, intval( $eta ) ), - $this->processed, - $this->count, - $this->processed / $delta, - $updateRate * 100.0 ); + $this->output( + sprintf( "%s %s: %6.2f%% done on %s; ETA %s [%d/%d] %.2f/sec <%.2f%% updated>\n", + wfWikiID(), + wfTimestamp( TS_DB, intval( $now ) ), + $portion * 100.0, + $this->table, + wfTimestamp( TS_DB, intval( $eta ) ), + $this->processed, + $this->count, + $this->processed / $delta, + $updateRate * 100.0 + ) + ); flush(); } - function runTable( $table, $where, $callback ) { - $count = $this->dbw->selectField( $table, 'count(*)', '', __METHOD__ ); + public function runTable( $params ) { + $dbr = wfGetDB( DB_SLAVE ); + + if ( array_diff( array_keys( $params ), + array( 'table', 'conds', 'index', 'callback' ) ) ) + { + throw new MWException( __METHOD__.': Missing parameter ' . implode( ', ', $params ) ); + } + + $table = $params['table']; + $count = $dbr->selectField( $table, 'count(*)', $params['conds'], __METHOD__ ); $this->init( $count, $table ); - $this->log( "Processing $table..." ); + $this->output( "Processing $table...\n" ); - $tableName = $this->dbr->tableName( $table ); - $sql = "SELECT * FROM $tableName $where"; - $result = $this->dbr->query( $sql, __METHOD__ ); - foreach( $result as $row ) { - call_user_func( $callback, $row ); + $index = (array)$params['index']; + $indexConds = array(); + $options = array( + 'ORDER BY' => implode( ',', $index ), + 'LIMIT' => $this->batchSize + ); + $callback = array( $this, $params['callback'] ); + + while ( true ) { + $conds = array_merge( $params['conds'], $indexConds ); + $res = $dbr->select( $table, '*', $conds, __METHOD__, $options ); + if ( !$res->numRows() ) { + // Done + break; + } + + foreach ( $res as $row ) { + call_user_func( $callback, $row ); + } + + if ( $res->numRows() < $this->batchSize ) { + // Done + break; + } + + // Update the conditions to select the next batch. + // Construct a condition string by starting with the least significant part + // of the index, and adding more significant parts progressively to the left + // of the string. + $nextCond = ''; + foreach ( array_reverse( $index ) as $field ) { + $encValue = $dbr->addQuotes( $row->$field ); + if ( $nextCond === '' ) { + $nextCond = "$field > $encValue"; + } else { + $nextCond = "$field > $encValue OR ($field = $encValue AND ($nextCond))"; + } + } + $indexConds = array( $nextCond ); } - $this->log( "Finished $table... $this->updated of $this->processed rows updated" ); - $result->free(); + + $this->output( "Finished $table... $this->updated of $this->processed rows updated\n" ); } - function hexChar( $matches ) { + protected function hexChar( $matches ) { return sprintf( "\\x%02x", ord( $matches[1] ) ); } - - abstract function processPage( $row ); - } + +class TableCleanupTest extends TableCleanup { + function processRow( $row ) { + $this->progress( mt_rand( 0, 1 ) ); + } +} + -- cgit v1.2.2