TitleParser.php

Go to the documentation of this file.

<?php
namespace MediaWiki\Title;
 
use InvalidArgumentException;
use LogicException;
use MediaWiki\Interwiki\InterwikiLookup;
use MediaWiki\Language\Language;
use MediaWiki\Message\Message;
use MediaWiki\Parser\Sanitizer;
use Wikimedia\IPUtils;
 
class TitleParser {
    private Language $language;
    private InterwikiLookup $interwikiLookup;
    private NamespaceInfo $nsInfo;
 
    private array $localInterwikis;
 
    private $createMalformedTitleException;
 
    public function __construct(
        Language $language,
        InterwikiLookup $interwikiLookup,
        NamespaceInfo $nsInfo,
        $localInterwikis
    ) {
        $this->language = $language;
        $this->interwikiLookup = $interwikiLookup;
        $this->nsInfo = $nsInfo;
        $this->localInterwikis = $localInterwikis;
 
        // Default callback is to return a real MalformedTitleException,
        // callback signature matches constructor
        $this->createMalformedTitleException = static function (
            $errorMessage,
            $titleText = null,
            $errorMessageParameters = []
        ): MalformedTitleException {
            return new MalformedTitleException( $errorMessage, $titleText, $errorMessageParameters );
        };
    }
 
    public function overrideCreateMalformedTitleExceptionCallback( callable $callback ) {
        // @codeCoverageIgnoreStart
        if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
            throw new LogicException( __METHOD__ . ' can only be used in tests' );
        }
        // @codeCoverageIgnoreEnd
        $this->createMalformedTitleException = $callback;
    }
 
    public function parseTitle( $text, $defaultNamespace = NS_MAIN ) {
        // Convert things like &eacute; &#257; or &#x3017; into normalized (T16952) text
        $filteredText = Sanitizer::decodeCharReferencesAndNormalize( $text );
 
        // NOTE: this is an ugly kludge that allows this class to share the
        // code for parsing with the old Title class. The parser code should
        // be refactored to avoid this.
        $parts = $this->splitTitleString( $filteredText, $defaultNamespace );
 
        return new TitleValue(
            $parts['namespace'],
            $parts['dbkey'],
            $parts['fragment'],
            $parts['interwiki']
        );
    }
 
    public function makeTitleValueSafe( $namespace, $text, $fragment = '', $interwiki = '' ) {
        if ( !$this->nsInfo->exists( $namespace ) ) {
            return null;
        }
 
        $canonicalNs = $this->nsInfo->getCanonicalName( $namespace );
        $fullText = $canonicalNs == '' ? $text : "$canonicalNs:$text";
        if ( strval( $interwiki ) != '' ) {
            $fullText = "$interwiki:$fullText";
        }
        if ( strval( $fragment ) != '' ) {
            $fullText .= '#' . $fragment;
        }
 
        try {
            $parts = $this->splitTitleString( $fullText );
        } catch ( MalformedTitleException ) {
            return null;
        }
 
        return new TitleValue(
            $parts['namespace'], $parts['dbkey'], $parts['fragment'], $parts['interwiki'] );
    }
 
    public function splitTitleString( $text, $defaultNamespace = NS_MAIN ) {
        $dbkey = str_replace( ' ', '_', $text );
 
        # Initialisation
        $parts = [
            'interwiki' => '',
            'local_interwiki' => false,
            'fragment' => '',
            'namespace' => (int)$defaultNamespace,
            'dbkey' => $dbkey,
        ];
 
        # Strip Unicode bidi override characters.
        # Sometimes they slip into cut-n-pasted page titles, where the
        # override chars get included in list displays.
        $dbkey = preg_replace( '/[\x{200E}\x{200F}\x{202A}-\x{202E}]+/u', '', $dbkey );
 
        if ( $dbkey === null ) {
            # Regex had an error. Most likely this is caused by invalid UTF-8
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-utf8', $text );
            throw $exception;
        }
 
        # Clean up whitespace
        $dbkey = preg_replace(
            '/[ _\xA0\x{1680}\x{180E}\x{2000}-\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}]+/u',
            '_',
            $dbkey
        );
        $dbkey = trim( $dbkey, '_' );
 
        if ( str_contains( $dbkey, \UtfNormal\Constants::UTF8_REPLACEMENT ) ) {
            # Contained illegal UTF-8 sequences or forbidden Unicode chars.
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-utf8', $text );
            throw $exception;
        }
 
        $parts['dbkey'] = $dbkey;
 
        # Initial colon indicates main namespace rather than specified default
        # but should not create invalid {ns,title} pairs such as {0,Project:Foo}
        if ( $dbkey !== '' && $dbkey[0] == ':' ) {
            $parts['namespace'] = NS_MAIN;
            $dbkey = substr( $dbkey, 1 ); # remove the colon but continue processing
            $dbkey = trim( $dbkey, '_' ); # remove any subsequent whitespace
        }
 
        if ( $dbkey == '' ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-empty', $text );
            throw $exception;
        }
 
        # Namespace or interwiki prefix
        $prefixRegexp = "/^(.+?)_*:_*(.*)$/S";
        do {
            $m = [];
            if ( preg_match( $prefixRegexp, $dbkey, $m ) ) {
                $p = $m[1];
                $ns = $this->language->getNsIndex( $p );
                if ( $ns !== false ) {
                    # Ordinary namespace
                    $dbkey = $m[2];
                    $parts['namespace'] = $ns;
                    # For Talk:X pages, check if X has a "namespace" prefix
                    if ( $ns === NS_TALK && preg_match( $prefixRegexp, $dbkey, $x ) ) {
                        if ( $this->language->getNsIndex( $x[1] ) ) {
                            # Disallow Talk:File:x type titles...
                            $exception = ( $this->createMalformedTitleException )(
                                'title-invalid-talk-namespace',
                                $text
                            );
                            throw $exception;
                        } elseif ( $this->interwikiLookup->isValidInterwiki( $x[1] ) ) {
                            # Disallow Talk:Interwiki:x type titles...
                            $exception = ( $this->createMalformedTitleException )(
                                'title-invalid-talk-interwiki',
                                $text
                            );
                            throw $exception;
                        }
                    }
                } elseif ( $this->interwikiLookup->isValidInterwiki( $p ) ) {
                    # Interwiki link
                    $dbkey = $m[2];
                    $parts['interwiki'] = $this->language->lc( $p );
 
                    # Redundant interwiki prefix to the local wiki
                    foreach ( $this->localInterwikis as $localIW ) {
                        if ( strcasecmp( $parts['interwiki'], $localIW ) == 0 ) {
                            if ( $dbkey == '' ) {
                                # Empty self-links should point to the Main Page, to ensure
                                # compatibility with cross-wiki transclusions and the like.
                                $mainPage = Title::newMainPage();
                                return [
                                    'interwiki' => $mainPage->getInterwiki(),
                                    'local_interwiki' => true,
                                    'fragment' => $mainPage->getFragment(),
                                    'namespace' => $mainPage->getNamespace(),
                                    'dbkey' => $mainPage->getDBkey(),
                                ];
                            }
                            $parts['interwiki'] = '';
                            # local interwikis should behave like initial-colon links
                            $parts['local_interwiki'] = true;
 
                            # Do another namespace split...
                            continue 2;
                        }
                    }
 
                    # If there's an initial colon after the interwiki, that also
                    # resets the default namespace
                    if ( $dbkey !== '' && $dbkey[0] == ':' ) {
                        $parts['namespace'] = NS_MAIN;
                        $dbkey = substr( $dbkey, 1 );
                        $dbkey = trim( $dbkey, '_' );
                    }
                }
                # If there's no recognized interwiki or namespace,
                # then let the colon expression be part of the title.
            }
            break;
        } while ( true );
 
        $fragment = strstr( $dbkey, '#' );
        if ( $fragment !== false ) {
            $parts['fragment'] = str_replace( '_', ' ', substr( $fragment, 1 ) );
            $dbkey = substr( $dbkey, 0, strlen( $dbkey ) - strlen( $fragment ) );
            # remove whitespace again: prevents "Foo_bar_#"
            # becoming "Foo_bar_"
            $dbkey = rtrim( $dbkey, "_" );
        }
 
        # Reject illegal characters.
        $rxTc = self::getTitleInvalidRegex();
        $matches = [];
        if ( preg_match( $rxTc, $dbkey, $matches ) ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-characters', $text, [ $matches[0] ] );
            throw $exception;
        }
 
        # Pages with "/./" or "/../" appearing in the URLs will often be un-
        # reachable due to the way web browsers deal with 'relative' URLs.
        # Also, they conflict with subpage syntax.  Forbid them explicitly.
        if (
            str_contains( $dbkey, '.' ) &&
            (
                $dbkey === '.' || $dbkey === '..' ||
                str_starts_with( $dbkey, './' ) ||
                str_starts_with( $dbkey, '../' ) ||
                str_contains( $dbkey, '/./' ) ||
                str_contains( $dbkey, '/../' ) ||
                str_ends_with( $dbkey, '/.' ) ||
                str_ends_with( $dbkey, '/..' )
            )
        ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-relative', $text );
            throw $exception;
        }
 
        # Magic tilde sequences? Nu-uh!
        if ( str_contains( $dbkey, '~~~' ) ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-magic-tilde', $text );
            throw $exception;
        }
 
        # Limit the size of titles to 255 bytes. This is typically the size of the
        # underlying database field. We make an exception for special pages, which
        # don't need to be stored in the database, and may edge over 255 bytes due
        # to subpage syntax for long titles, e.g. [[Special:Block/Long name]]
        $maxLength = ( $parts['namespace'] !== NS_SPECIAL ) ? 255 : 512;
        if ( strlen( $dbkey ) > $maxLength ) {
            $exception = ( $this->createMalformedTitleException )(
                'title-invalid-too-long',
                $text,
                [ Message::numParam( $maxLength ), Message::numParam( strlen( $dbkey ) ) ]
            );
            throw $exception;
        }
 
        # Normally, all wiki links are forced to have an initial capital letter so [[foo]]
        # and [[Foo]] point to the same place.  Don't force it for interwikis, since the
        # other site might be case-sensitive.
        if ( $parts['interwiki'] === '' && $this->nsInfo->isCapitalized( $parts['namespace'] ) ) {
            $dbkey = $this->language->ucfirst( $dbkey );
        }
 
        # Can't make a link to a namespace alone... "empty" local links can only be
        # self-links with a fragment identifier.
        if ( $dbkey == '' && $parts['interwiki'] === '' && $parts['namespace'] !== NS_MAIN ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-empty', $text );
            throw $exception;
        }
 
        // Allow IPv6 usernames to start with '::' by canonicalizing IPv6 titles.
        // IP names are not allowed for accounts, and can only be referring to
        // edits from the IP. Given '::' abbreviations and caps/lowercaps,
        // there are numerous ways to present the same IP. Having sp:contribs scan
        // them all is silly and having some show the edits and others not is
        // inconsistent. Same for talk/userpages. Keep them normalized instead.
        if ( $dbkey !== '' && ( $parts['namespace'] === NS_USER || $parts['namespace'] === NS_USER_TALK ) ) {
            $dbkey = IPUtils::sanitizeIP( $dbkey );
            // IPUtils::sanitizeIP return null only for bad input
            '@phan-var string $dbkey';
        }
 
        // Any remaining initial :s are illegal.
        if ( $dbkey !== '' && $dbkey[0] == ':' ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid-leading-colon', $text );
            throw $exception;
        }
 
        // Fill fields
        $parts['dbkey'] = $dbkey;
 
        // Check to ensure that the return value can be used to construct a TitleValue.
        // All issues should in theory be caught above, this is here to enforce consistency.
        try {
            TitleValue::assertValidSpec(
                $parts['namespace'],
                $parts['dbkey'],
                $parts['fragment'],
                $parts['interwiki']
            );
        } catch ( InvalidArgumentException $ex ) {
            $exception = ( $this->createMalformedTitleException )( 'title-invalid', $text, [ $ex->getMessage() ] );
            throw $exception;
        }
 
        return $parts;
    }
 
    public static function getTitleInvalidRegex() {
        static $rxTc = false;
        if ( !$rxTc ) {
            # Matching titles will be held as illegal.
            $rxTc = '/' .
                # Any character not allowed is forbidden...
                '[^' . Title::legalChars() . ']' .
                # URL percent encoding sequences interfere with the ability
                # to round-trip titles -- you can't link to them consistently.
                '|%[0-9A-Fa-f]{2}' .
                # XML/HTML character references produce similar issues.
                '|&[A-Za-z0-9\x80-\xff]+;' .
                '/S';
        }
 
        return $rxTc;
    }
}
 
class_alias( TitleParser::class, 'TitleParser' );