BadFileLookup.php

Go to the documentation of this file.

<?php
 
namespace MediaWiki;
 
use BagOStuff;
use Hooks;
use MalformedTitleException;
use MediaWiki\Linker\LinkTarget;
use RepoGroup;
use TitleParser;
 
class BadFileLookup {
    private $blacklistCallback;
 
    private $cache;
 
    private $repoGroup;
 
    private $titleParser;
 
    private $badFiles;
 
    public function __construct(
        callable $blacklistCallback,
        BagOStuff $cache,
        RepoGroup $repoGroup,
        TitleParser $titleParser
    ) {
        $this->blacklistCallback = $blacklistCallback;
        $this->cache = $cache;
        $this->repoGroup = $repoGroup;
        $this->titleParser = $titleParser;
    }
 
    public function isBadFile( $name, LinkTarget $contextTitle = null ) {
        // Handle redirects; callers almost always hit wfFindFile() anyway, so just use that method
        // because it has a fast process cache.
        $file = $this->repoGroup->findFile( $name );
        // XXX If we don't find the file we also don't replace spaces by underscores or otherwise
        // validate or normalize the title, is this right?
        if ( $file ) {
            $name = $file->getTitle()->getDBkey();
        }
 
        // Run the extension hook
        $bad = false;
        if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
            return (bool)$bad;
        }
 
        if ( $this->badFiles === null ) {
            // Not used before in this request, try the cache
            $blacklist = ( $this->blacklistCallback )();
            $key = $this->cache->makeKey( 'bad-image-list', sha1( $blacklist ) );
            $this->badFiles = $this->cache->get( $key ) ?: null;
        }
 
        if ( $this->badFiles === null ) {
            // Cache miss, build the list now
            $this->badFiles = [];
            $lines = explode( "\n", $blacklist );
            foreach ( $lines as $line ) {
                // List items only
                if ( substr( $line, 0, 1 ) !== '*' ) {
                    continue;
                }
 
                // Find all links
                $m = [];
                // XXX What is the ':?' doing in the regex? Why not let the TitleParser strip it?
                if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
                    continue;
                }
 
                $fileDBkey = null;
                $exceptions = [];
                foreach ( $m[1] as $i => $titleText ) {
                    try {
                        $title = $this->titleParser->parseTitle( $titleText );
                    } catch ( MalformedTitleException $e ) {
                        continue;
                    }
                    if ( $i == 0 ) {
                        $fileDBkey = $title->getDBkey();
                    } else {
                        $exceptions[$title->getNamespace()][$title->getDBkey()] = true;
                    }
                }
 
                if ( $fileDBkey !== null ) {
                    $this->badFiles[$fileDBkey] = $exceptions;
                }
            }
            $this->cache->set( $key, $this->badFiles, 24 * 60 * 60 );
        }
 
        return isset( $this->badFiles[$name] ) && ( !$contextTitle ||
            !isset( $this->badFiles[$name][$contextTitle->getNamespace()]
                [$contextTitle->getDBkey()] ) );
    }
}