library/HTMLPurifier/Lexer/PEARSax3.php

   1 <?php
   2
   3 /**
   4  * Proof-of-concept lexer that uses the PEAR package XML_HTMLSax3 to parse HTML.
   5  *
   6  * PEAR, not suprisingly, also has a SAX parser for HTML.  I don't know
   7  * very much about implementation, but it's fairly well written.  However, that
   8  * abstraction comes at a price: performance. You need to have it installed,
   9  * and if the API changes, it might break our adapter. Not sure whether or not
  10  * it's UTF-8 aware, but it has some entity parsing trouble (in all areas,
  11  * text and attributes).
  12  *
  13  * Quite personally, I don't recommend using the PEAR class, and the defaults
  14  * don't use it. The unit tests do perform the tests on the SAX parser too, but
  15  * whatever it does for poorly formed HTML is up to it.
  16  *
  17  * @todo Generalize so that XML_HTMLSax is also supported.
  18  *
  19  * @warning Entity-resolution inside attributes is broken.
  20  */
  21
  22 class HTMLPurifier_Lexer_PEARSax3 extends HTMLPurifier_Lexer
  23 {
  24
  25     /**
  26      * Internal accumulator array for SAX parsers.
  27      */
  28     protected $tokens = array();
  29
  30     private $parent_handler;
  31     private $stack = array();
  32
  33     public function tokenizeHTML($string, $config, $context) {
  34
  35         $this->tokens = array();
  36
  37         $string = $this->normalize($string, $config, $context);
  38
  39         $this->parent_handler = set_error_handler(array($this, 'muteStrictErrorHandler'));
  40
  41         $parser = new XML_HTMLSax3();
  42         $parser->set_object($this);
  43         $parser->set_element_handler('openHandler','closeHandler');
  44         $parser->set_data_handler('dataHandler');
  45         $parser->set_escape_handler('escapeHandler');
  46
  47         // doesn't seem to work correctly for attributes
  48         $parser->set_option('XML_OPTION_ENTITIES_PARSED', 1);
  49
  50         $parser->parse($string);
  51
  52         restore_error_handler();
  53
  54         return $this->tokens;
  55
  56     }
  57
  58     /**
  59      * Open tag event handler, interface is defined by PEAR package.
  60      */
  61     public function openHandler(&$parser, $name, $attrs, $closed) {
  62         // entities are not resolved in attrs
  63         foreach ($attrs as $key => $attr) {
  64             $attrs[$key] = $this->parseData($attr);
  65         }
  66         if ($closed) {
  67             $this->tokens[] = new HTMLPurifier_Token_Empty($name, $attrs);
  68         } else {
  69             $this->tokens[] = new HTMLPurifier_Token_Start($name, $attrs);
  70         }
  71         $this->stack[] = $name;
  72         return true;
  73     }
  74
  75     /**
  76      * Close tag event handler, interface is defined by PEAR package.
  77      */
  78     public function closeHandler(&$parser, $name) {
  79         // HTMLSax3 seems to always send empty tags an extra close tag
  80         // check and ignore if you see it:
  81         // [TESTME] to make sure it doesn't overreach
  82         if ($this->tokens[count($this->tokens)-1] instanceof HTMLPurifier_Token_Empty) {
  83             return true;
  84         }
  85         $this->tokens[] = new HTMLPurifier_Token_End($name);
  86         if (!empty($this->stack)) array_pop($this->stack);
  87         return true;
  88     }
  89
  90     /**
  91      * Data event handler, interface is defined by PEAR package.
  92      */
  93     public function dataHandler(&$parser, $data) {
  94         $this->tokens[] = new HTMLPurifier_Token_Text($data);
  95         return true;
  96     }
  97
  98     /**
  99      * Escaped text handler, interface is defined by PEAR package.
 100      */
 101     public function escapeHandler(&$parser, $data) {
 102         if (strpos($data, '--') === 0) {
 103             // remove trailing and leading double-dashes
 104             $data = substr($data, 2);
 105             if (strlen($data) >= 2 && substr($data, -2) == "--") {
 106                 $data = substr($data, 0, -2);
 107             }
 108             if (isset($this->stack[sizeof($this->stack) - 1]) &&
 109                 $this->stack[sizeof($this->stack) - 1] == "style") {
 110                 $this->tokens[] = new HTMLPurifier_Token_Text($data);
 111             } else {
 112                 $this->tokens[] = new HTMLPurifier_Token_Comment($data);
 113             }
 114         }
 115         // CDATA is handled elsewhere, but if it was handled here:
 116         //if (strpos($data, '[CDATA[') === 0) {
 117         //    $this->tokens[] = new HTMLPurifier_Token_Text(
 118         //        substr($data, 7, strlen($data) - 9) );
 119         //}
 120         return true;
 121     }
 122
 123     /**
 124      * An error handler that mutes strict errors
 125      */
 126     public function muteStrictErrorHandler($errno, $errstr, $errfile=null, $errline=null, $errcontext=null) {
 127         if ($errno == E_STRICT) return;
 128         return call_user_func($this->parent_handler, $errno, $errstr, $errfile, $errline, $errcontext);
 129     }
 130
 131 }
 132
 133 // vim: et sw=4 sts=4