21836314048

Committed 09 Feb 2026 06:24PM UTC coverage: 98.75% (-0.5%) from 99.277%

Build # 21836314048

Build Type

push

github

Committed by

xemlock

Commit Message

Fix catastrophic backtracking in Core.AggressivelyFixLt

Run Details

2 of 3 new or added lines in 1 file covered. (66.67%)

3 existing lines in 1 file now uncovered.

1501 of 1520 relevant lines covered (98.75%)

3547.56 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

61.9

/library/HTMLPurifier/Lexer/HTML5.php

<?php

/**
 * Experimental HTML5-compliant parser using masterminds/html5 library.
 */
class HTMLPurifier_Lexer_HTML5 extends HTMLPurifier_Lexer_DOMLex
{
    /**
     * @throws HTMLPurifier_Exception
     * @codeCoverageIgnore
     */
    public function __construct()
    {
        if (!class_exists('\Masterminds\HTML5')) {
            throw new HTMLPurifier_Exception('Cannot instantiate HTML5 lexer. \Masterminds\HTML5 class is not available');
        }
        parent::__construct();
    }

    /**
     * Transforms an HTML string into tokens.
     *
     * @param  string                $html
     * @param  HTMLPurifier_Config   $config
     * @param  HTMLPurifier_Context  $context
     * @return HTMLPurifier_Token[]
     */
    public function tokenizeHTML($html, $config, $context)
    {
        $html = $this->normalize($html, $config, $context);
        $html = $this->armor($html, $config);

        // masterminds/html5 requires <html>, <head> and <body> tags
        $html = $this->wrapHTML($html, $config, $context, false);

        // Parse the document. $dom is a DOMDocument.
        $html5 = new \Masterminds\HTML5(array('disable_html_ns' => true));
        $doc = $html5->loadHTML($html);

        $body = $doc->getElementsByTagName('html')->item(0)  // <html>
                    ->getElementsByTagName('body')->item(0); // <body>

        $tokens = array();
        $this->tokenizeDOM($body, $tokens, $config);

        return $tokens;
    }

    /**
     * Attempt to armor stray angled brackets that cannot possibly
     * form tags and thus are probably being used as emoticons
     *
     * @param  string               $html
     * @param  HTMLPurifier_Config  $config
     * @return string
     */
    protected function armor($html, HTMLPurifier_Config $config)
    {
        if ($config->get('Core.AggressivelyFixLt')) {
            // Since HTMLPurifier 4.19.0, see: https://github.com/ezyang/htmlpurifier/pull/440
            if (method_exists(get_parent_class($this), 'aggressivelyFixLt')) {
                return $this->aggressivelyFixLt($html);
            }

            $char = '[^a-z!\/]';
            $comment = "/<!--([^>]*)(-->|\z)/is";
            $html = preg_replace_callback($comment, array($this, 'callbackArmorCommentEntities'), $html);

            do {
                $old = $html;
                $html = preg_replace("/<($char)/i", '&lt;\\1', $html);
            } while ($html !== $old);

            $html = preg_replace_callback($comment, array($this, 'callbackUndoCommentSubst'), $html); // fix comments
        }

        return $html;
    }
}

1	<?php
2
3	/**
4	* Experimental HTML5-compliant parser using masterminds/html5 library.
5	*/
6	class HTMLPurifier_Lexer_HTML5 extends HTMLPurifier_Lexer_DOMLex
7	{
8	/**
9	* @throws HTMLPurifier_Exception
10	* @codeCoverageIgnore
11	*/
12	public function __construct()
13	{
14	if (!class_exists('\Masterminds\HTML5')) {
15	throw new HTMLPurifier_Exception('Cannot instantiate HTML5 lexer. \Masterminds\HTML5 class is not available');
16	}
17	parent::__construct();
18	}
19
20	/**
21	* Transforms an HTML string into tokens.
22	*
23	* @param string $html
24	* @param HTMLPurifier_Config $config
25	* @param HTMLPurifier_Context $context
26	* @return HTMLPurifier_Token[]
27	*/
28	public function tokenizeHTML($html, $config, $context)
29	{
30	$html = $this->normalize($html, $config, $context);	495✔
31	$html = $this->armor($html, $config);	495✔
32
33	// masterminds/html5 requires <html>, <head> and <body> tags
34	$html = $this->wrapHTML($html, $config, $context, false);	495✔
35
36	// Parse the document. $dom is a DOMDocument.
37	$html5 = new \Masterminds\HTML5(array('disable_html_ns' => true));	495✔
38	$doc = $html5->loadHTML($html);	495✔
39
40	$body = $doc->getElementsByTagName('html')->item(0) // <html>	495✔
41	->getElementsByTagName('body')->item(0); // <body>	495✔
42
43	$tokens = array();	495✔
44	$this->tokenizeDOM($body, $tokens, $config);	495✔
45
46	return $tokens;	495✔
47	}
48
49	/**
50	* Attempt to armor stray angled brackets that cannot possibly
51	* form tags and thus are probably being used as emoticons
52	*
53	* @param string $html
54	* @param HTMLPurifier_Config $config
55	* @return string
56	*/
57	protected function armor($html, HTMLPurifier_Config $config)
58	{
59	if ($config->get('Core.AggressivelyFixLt')) {	495✔
60	// Since HTMLPurifier 4.19.0, see: https://github.com/ezyang/htmlpurifier/pull/440
61	if (method_exists(get_parent_class($this), 'aggressivelyFixLt')) {	495✔
62	return $this->aggressivelyFixLt($html);	495✔
63	}
64
UNCOV 65	$char = '[^a-z!\/]';	×
NEW 66	$comment = "/<!--([^>]*)(-->\|\z)/is";	×
UNCOV 67	$html = preg_replace_callback($comment, array($this, 'callbackArmorCommentEntities'), $html);	×
68
69	do {
70	$old = $html;	×
71	$html = preg_replace("/<($char)/i", '<\\1', $html);	×
72	} while ($html !== $old);	×
73
UNCOV 74	$html = preg_replace_callback($comment, array($this, 'callbackUndoCommentSubst'), $html); // fix comments	×
75	}
76
77	return $html;	×
78	}
79	}

xemlock / htmlpurifier-html5 / 21836314048

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous