generatePhpCharToUpperMappings.php

Go to the documentation of this file.

<?php

use MediaWiki\Shell\Shell;

require_once __DIR__ . '/../Maintenance.php';

class GeneratePhpCharToUpperMappings extends Maintenance {

    public function __construct() {
        parent::__construct();
        $this->addDescription( 'Update list of upper case differences between JS and PHP.' );
    }

    public function execute() {
        global $wgContLang, $IP;

        $data = [];

        $result = Shell::command(
                [ 'node', $IP . '/maintenance/mediawiki.Title/generateJsToUpperCaseList.js' ]
            )
            // Node allocates lots of memory
            ->limits( [ 'memory' => 1024 * 1024 ] )
            ->execute();

        if ( $result->getExitcode() !== 0 ) {
            $this->output( $result->getStderr() );
            return;
        }

        $jsUpperChars = json_decode( $result->getStdout() );

        for ( $i = 0; $i <= 0x10ffff; $i++ ) {
            if ( $i >= 0xd800 && $i <= 0xdfff ) {
                // Skip surrogate pairs
                continue;
            }
            $char = \UtfNormal\Utils::codepointToUtf8( $i );
            $phpUpper = $wgContLang->ucfirst( $char );
            $jsUpper = $jsUpperChars[$i];
            if ( $jsUpper !== $phpUpper ) {
                $data[$char] = $phpUpper;
            }
        }

        $mappingJson = str_replace( '    ', "\t",
            json_encode( $data, JSON_PRETTY_PRINT | JSON_UNESCAPED_UNICODE )
        ) . "\n";
        $outputPath = '/resources/src/mediawiki.Title/phpCharToUpper.json';
        $file = fopen( $IP . $outputPath, 'w' );
        fwrite( $file, $mappingJson );

        $this->output( count( $data ) . " differences found.\n" );
        $this->output( "Written to $outputPath\n" );
    }
}

$maintClass = GeneratePhpCharToUpperMappings::class;
require_once RUN_MAINTENANCE_IF_MAIN;