root/afridex/plugins/Flutter/purifier_lib/HTMLPurifier/Lexer/DOMLex.php @ 21

Revision 21, 8.4 kB (checked in by admin, 18 years ago)
Line 
1<?php
2
3/**
4 * Parser that uses PHP 5's DOM extension (part of the core).
5 *
6 * In PHP 5, the DOM XML extension was revamped into DOM and added to the core.
7 * It gives us a forgiving HTML parser, which we use to transform the HTML
8 * into a DOM, and then into the tokens.  It is blazingly fast (for large
9 * documents, it performs twenty times faster than
10 * HTMLPurifier_Lexer_DirectLex,and is the default choice for PHP 5.
11 *
12 * @note Any empty elements will have empty tokens associated with them, even if
13 * this is prohibited by the spec. This is cannot be fixed until the spec
14 * comes into play.
15 *
16 * @note PHP's DOM extension does not actually parse any entities, we use
17 *       our own function to do that.
18 *
19 * @warning DOM tends to drop whitespace, which may wreak havoc on indenting.
20 *          If this is a huge problem, due to the fact that HTML is hand
21 *          edited and you are unable to get a parser cache that caches the
22 *          the output of HTML Purifier while keeping the original HTML lying
23 *          around, you may want to run Tidy on the resulting output or use
24 *          HTMLPurifier_DirectLex
25 */
26
27class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
28{
29   
30    private $factory;
31   
32    public function __construct() {
33        // setup the factory
34        parent::__construct();
35        $this->factory = new HTMLPurifier_TokenFactory();
36    }
37   
38    public function tokenizeHTML($html, $config, $context) {
39       
40        $html = $this->normalize($html, $config, $context);
41       
42        // attempt to armor stray angled brackets that cannot possibly
43        // form tags and thus are probably being used as emoticons
44        if ($config->get('Core', 'AggressivelyFixLt')) {
45            $char = '[^a-z!\/]';
46            $comment = "/<!--(.*?)(-->|\z)/is";
47            $html = preg_replace_callback($comment, array($this, 'callbackArmorCommentEntities'), $html);
48            $html = preg_replace("/<($char)/i", '&lt;\\1', $html);
49            $html = preg_replace_callback($comment, array($this, 'callbackUndoCommentSubst'), $html); // fix comments
50        }
51       
52        // preprocess html, essential for UTF-8
53        $html = $this->wrapHTML($html, $config, $context);
54       
55        $doc = new DOMDocument();
56        $doc->encoding = 'UTF-8'; // theoretically, the above has this covered
57       
58        set_error_handler(array($this, 'muteErrorHandler'));
59        $doc->loadHTML($html);
60        restore_error_handler();
61       
62        $tokens = array();
63        $this->tokenizeDOM(
64            $doc->getElementsByTagName('html')->item(0)-> // <html>
65                  getElementsByTagName('body')->item(0)-> //   <body>
66                  getElementsByTagName('div')->item(0)    //     <div>
67            , $tokens);
68        return $tokens;
69    }
70   
71    /**
72     * Recursive function that tokenizes a node, putting it into an accumulator.
73     *
74     * @param $node     DOMNode to be tokenized.
75     * @param $tokens   Array-list of already tokenized tokens.
76     * @param $collect  Says whether or start and close are collected, set to
77     *                  false at first recursion because it's the implicit DIV
78     *                  tag you're dealing with.
79     * @returns Tokens of node appended to previously passed tokens.
80     */
81    protected function tokenizeDOM($node, &$tokens, $collect = false) {
82       
83        // intercept non element nodes. WE MUST catch all of them,
84        // but we're not getting the character reference nodes because
85        // those should have been preprocessed
86        if ($node->nodeType === XML_TEXT_NODE) {
87            $tokens[] = $this->factory->createText($node->data);
88            return;
89        } elseif ($node->nodeType === XML_CDATA_SECTION_NODE) {
90            // undo libxml's special treatment of <script> and <style> tags
91            $last = end($tokens);
92            $data = $node->data;
93            // (note $node->tagname is already normalized)
94            if ($last instanceof HTMLPurifier_Token_Start && ($last->name == 'script' || $last->name == 'style')) {
95                $new_data = trim($data);
96                if (substr($new_data, 0, 4) === '<!--') {
97                    $data = substr($new_data, 4);
98                    if (substr($data, -3) === '-->') {
99                        $data = substr($data, 0, -3);
100                    } else {
101                        // Highly suspicious! Not sure what to do...
102                    }
103                }
104            }
105            $tokens[] = $this->factory->createText($this->parseData($data));
106            return;
107        } elseif ($node->nodeType === XML_COMMENT_NODE) {
108            // this is code is only invoked for comments in script/style in versions
109            // of libxml pre-2.6.28 (regular comments, of course, are still
110            // handled regularly)
111            $tokens[] = $this->factory->createComment($node->data);
112            return;
113        } elseif (
114            // not-well tested: there may be other nodes we have to grab
115            $node->nodeType !== XML_ELEMENT_NODE
116        ) {
117            return;
118        }
119       
120        $attr = $node->hasAttributes() ?
121            $this->transformAttrToAssoc($node->attributes) :
122            array();
123       
124        // We still have to make sure that the element actually IS empty
125        if (!$node->childNodes->length) {
126            if ($collect) {
127                $tokens[] = $this->factory->createEmpty($node->tagName, $attr);
128            }
129        } else {
130            if ($collect) { // don't wrap on first iteration
131                $tokens[] = $this->factory->createStart(
132                    $tag_name = $node->tagName, // somehow, it get's dropped
133                    $attr
134                );
135            }
136            foreach ($node->childNodes as $node) {
137                // remember, it's an accumulator. Otherwise, we'd have
138                // to use array_merge
139                $this->tokenizeDOM($node, $tokens, true);
140            }
141            if ($collect) {
142                $tokens[] = $this->factory->createEnd($tag_name);
143            }
144        }
145       
146    }
147   
148    /**
149     * Converts a DOMNamedNodeMap of DOMAttr objects into an assoc array.
150     *
151     * @param $attribute_list DOMNamedNodeMap of DOMAttr objects.
152     * @returns Associative array of attributes.
153     */
154    protected function transformAttrToAssoc($node_map) {
155        // NamedNodeMap is documented very well, so we're using undocumented
156        // features, namely, the fact that it implements Iterator and
157        // has a ->length attribute
158        if ($node_map->length === 0) return array();
159        $array = array();
160        foreach ($node_map as $attr) {
161            $array[$attr->name] = $attr->value;
162        }
163        return $array;
164    }
165   
166    /**
167     * An error handler that mutes all errors
168     */
169    public function muteErrorHandler($errno, $errstr) {}
170   
171    /**
172     * Callback function for undoing escaping of stray angled brackets
173     * in comments
174     */
175    public function callbackUndoCommentSubst($matches) {
176        return '<!--' . strtr($matches[1], array('&amp;'=>'&','&lt;'=>'<')) . $matches[2];
177    }
178   
179    /**
180     * Callback function that entity-izes ampersands in comments so that
181     * callbackUndoCommentSubst doesn't clobber them
182     */
183    public function callbackArmorCommentEntities($matches) {
184        return '<!--' . str_replace('&', '&amp;', $matches[1]) . $matches[2];
185    }
186   
187    /**
188     * Wraps an HTML fragment in the necessary HTML
189     */
190    protected function wrapHTML($html, $config, $context) {
191        $def = $config->getDefinition('HTML');
192        $ret = '';
193       
194        if (!empty($def->doctype->dtdPublic) || !empty($def->doctype->dtdSystem)) {
195            $ret .= '<!DOCTYPE html ';
196            if (!empty($def->doctype->dtdPublic)) $ret .= 'PUBLIC "' . $def->doctype->dtdPublic . '" ';
197            if (!empty($def->doctype->dtdSystem)) $ret .= '"' . $def->doctype->dtdSystem . '" ';
198            $ret .= '>';
199        }
200       
201        $ret .= '<html><head>';
202        $ret .= '<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />';
203        // No protection if $html contains a stray </div>!
204        $ret .= '</head><body><div>'.$html.'</div></body></html>';
205        return $ret;
206    }
207   
208}
209
Note: See TracBrowser for help on using the browser.