RemexRemoveTagHandler.php

Go to the documentation of this file.

<?php
 
namespace MediaWiki\Parser;
 
use Wikimedia\RemexHtml\Tokenizer\Attributes;
use Wikimedia\RemexHtml\Tokenizer\PlainAttributes;
use Wikimedia\RemexHtml\Tokenizer\RelayTokenHandler;
use Wikimedia\RemexHtml\Tokenizer\TokenHandler;
 
class RemexRemoveTagHandler extends RelayTokenHandler {
    private $source;
 
    private $htmlsingle;
 
    private $htmlsingleonly;
 
    private $htmlelements;
 
    private $commentRegex;
 
    private $attrCallback;
 
    private $callbackArgs;
 
    public function __construct(
        TokenHandler $nextHandler,
        string $source,
        array $tagData,
        ?callable $attrCallback,
        ?array $callbackArgs,
        array $options
    ) {
        parent::__construct( $nextHandler );
        $this->source = $source;
        $this->htmlsingle = $tagData['htmlsingle'];
        $this->htmlsingleonly = $tagData['htmlsingleonly'];
        $this->htmlelements = $tagData['htmlelements'];
        $this->attrCallback = $attrCallback;
        $this->callbackArgs = $callbackArgs ?? [];
        $this->commentRegex = $options['commentRegex'] ?? null;
    }
 
    public function comment( $text, $sourceStart, $sourceLength ) {
        if ( $this->commentRegex !== null && preg_match( $this->commentRegex, $text ) ) {
            $this->nextHandler->comment( $text, $sourceStart, $sourceLength );
        }
    }
 
    private static function validateTag( string $element, Attributes $attrs ): bool {
        if ( $element == 'meta' || $element == 'link' ) {
            $params = $attrs->getValues();
            if ( !isset( $params['itemprop'] ) ) {
                // <meta> and <link> must have an itemprop="" otherwise they are not valid or safe in content
                return false;
            }
            if ( $element == 'meta' && !isset( $params['content'] ) ) {
                // <meta> must have a content="" for the itemprop
                return false;
            }
            if ( $element == 'link' && !isset( $params['href'] ) ) {
                // <link> must have an associated href=""
                return false;
            }
        }
 
        return true;
    }
 
    public function startTag( $name, Attributes $attrs, $selfClose, $sourceStart, $sourceLength ) {
        // Handle a start tag from the tokenizer: either relay it to the
        // next stage, or re-emit it as raw text.
 
        $badtag = false;
        $t = strtolower( $name );
        if ( isset( $this->htmlelements[$t] ) ) {
            if ( $this->attrCallback ) {
                $attrs = ( $this->attrCallback )( $attrs, ...$this->callbackArgs );
            }
            if ( $selfClose && !( isset( $this->htmlsingle[$t] ) || isset( $this->htmlsingleonly[$t] ) ) ) {
                // Remove the self-closing slash, to be consistent with
                // HTML5 semantics. T134423
                $selfClose = false;
            }
            if ( !self::validateTag( $t, $attrs ) ) {
                $badtag = true;
            }
            $fixedAttrs = Sanitizer::validateTagAttributes( $attrs->getValues(), $t );
            $attrs = new PlainAttributes( $fixedAttrs );
            if ( !$badtag ) {
                if ( $selfClose && !isset( $this->htmlsingleonly[$t] ) ) {
                    // Interpret self-closing tags as empty tags even when
                    // HTML5 would interpret them as start tags.  Such input
                    // is commonly seen on Wikimedia wikis with this intention.
                    $this->nextHandler->startTag( $name, $attrs, false, $sourceStart, $sourceLength );
                    $this->nextHandler->endTag( $name, $sourceStart + $sourceLength, 0 );
                } else {
                    $this->nextHandler->startTag( $name, $attrs, $selfClose, $sourceStart, $sourceLength );
                }
                return;
            }
        }
        // Emit this as a text node instead.
        $this->nextHandler->characters( $this->source, $sourceStart, $sourceLength, $sourceStart, $sourceLength );
    }
 
    public function endTag( $name, $sourceStart, $sourceLength ) {
        // Handle an end tag from the tokenizer: either relay it to the
        // next stage, or re-emit it as raw text.
 
        $t = strtolower( $name );
        if ( isset( $this->htmlelements[$t] ) ) {
            // This is a good tag, relay it.
            $this->nextHandler->endTag( $name, $sourceStart, $sourceLength );
        } else {
            // Emit this as a text node instead.
            $this->nextHandler->characters( $this->source, $sourceStart, $sourceLength, $sourceStart, $sourceLength );
        }
    }
 
}