library/HTMLPurifier/Generator.php

   1 <?php
   2
   3 /**
   4  * Generates HTML from tokens.
   5  * @todo Refactor interface so that configuration/context is determined
   6  *       upon instantiation, no need for messy generateFromTokens() calls
   7  * @todo Make some of the more internal functions protected, and have
   8  *       unit tests work around that
   9  */
  10 class HTMLPurifier_Generator
  11 {
  12
  13     /**
  14      * Whether or not generator should produce XML output
  15      */
  16     private $_xhtml = true;
  17
  18     /**
  19      * :HACK: Whether or not generator should comment the insides of <script> tags
  20      */
  21     private $_scriptFix = false;
  22
  23     /**
  24      * Cache of HTMLDefinition during HTML output to determine whether or
  25      * not attributes should be minimized.
  26      */
  27     private $_def;
  28
  29     /**
  30      * Cache of %Output.SortAttr
  31      */
  32     private $_sortAttr;
  33
  34     /**
  35      * Configuration for the generator
  36      */
  37     protected $config;
  38
  39     /**
  40      * @param $config Instance of HTMLPurifier_Config
  41      * @param $context Instance of HTMLPurifier_Context
  42      */
  43     public function __construct($config, $context) {
  44         $this->config = $config;
  45         $this->_scriptFix = $config->get('Output.CommentScriptContents');
  46         $this->_sortAttr = $config->get('Output.SortAttr');
  47         $this->_def = $config->getHTMLDefinition();
  48         $this->_xhtml = $this->_def->doctype->xml;
  49     }
  50
  51     /**
  52      * Generates HTML from an array of tokens.
  53      * @param $tokens Array of HTMLPurifier_Token
  54      * @param $config HTMLPurifier_Config object
  55      * @return Generated HTML
  56      */
  57     public function generateFromTokens($tokens) {
  58         if (!$tokens) return '';
  59
  60         // Basic algorithm
  61         $html = '';
  62         for ($i = 0, $size = count($tokens); $i < $size; $i++) {
  63             if ($this->_scriptFix && $tokens[$i]->name === 'script'
  64                 && $i + 2 < $size && $tokens[$i+2] instanceof HTMLPurifier_Token_End) {
  65                 // script special case
  66                 // the contents of the script block must be ONE token
  67                 // for this to work.
  68                 $html .= $this->generateFromToken($tokens[$i++]);
  69                 $html .= $this->generateScriptFromToken($tokens[$i++]);
  70             }
  71             $html .= $this->generateFromToken($tokens[$i]);
  72         }
  73
  74         // Tidy cleanup
  75         if (extension_loaded('tidy') && $this->config->get('Output.TidyFormat')) {
  76             $tidy = new Tidy;
  77             $tidy->parseString($html, array(
  78                'indent'=> true,
  79                'output-xhtml' => $this->_xhtml,
  80                'show-body-only' => true,
  81                'indent-spaces' => 2,
  82                'wrap' => 68,
  83             ), 'utf8');
  84             $tidy->cleanRepair();
  85             $html = (string) $tidy; // explicit cast necessary
  86         }
  87
  88         // Normalize newlines to system defined value
  89         $nl = $this->config->get('Output.Newline');
  90         if ($nl === null) $nl = PHP_EOL;
  91         if ($nl !== "\n") $html = str_replace("\n", $nl, $html);
  92         return $html;
  93     }
  94
  95     /**
  96      * Generates HTML from a single token.
  97      * @param $token HTMLPurifier_Token object.
  98      * @return Generated HTML
  99      */
 100     public function generateFromToken($token) {
 101         if (!$token instanceof HTMLPurifier_Token) {
 102             trigger_error('Cannot generate HTML from non-HTMLPurifier_Token object', E_USER_WARNING);
 103             return '';
 104
 105         } elseif ($token instanceof HTMLPurifier_Token_Start) {
 106             $attr = $this->generateAttributes($token->attr, $token->name);
 107             return '<' . $token->name . ($attr ? ' ' : '') . $attr . '>';
 108
 109         } elseif ($token instanceof HTMLPurifier_Token_End) {
 110             return '</' . $token->name . '>';
 111
 112         } elseif ($token instanceof HTMLPurifier_Token_Empty) {
 113             $attr = $this->generateAttributes($token->attr, $token->name);
 114              return '<' . $token->name . ($attr ? ' ' : '') . $attr .
 115                 ( $this->_xhtml ? ' /': '' ) // <br /> v. <br>
 116                 . '>';
 117
 118         } elseif ($token instanceof HTMLPurifier_Token_Text) {
 119             return $this->escape($token->data, ENT_NOQUOTES);
 120
 121         } elseif ($token instanceof HTMLPurifier_Token_Comment) {
 122             return '<!--' . $token->data . '-->';
 123         } else {
 124             return '';
 125
 126         }
 127     }
 128
 129     /**
 130      * Special case processor for the contents of script tags
 131      * @warning This runs into problems if there's already a literal
 132      *          --> somewhere inside the script contents.
 133      */
 134     public function generateScriptFromToken($token) {
 135         if (!$token instanceof HTMLPurifier_Token_Text) return $this->generateFromToken($token);
 136         // Thanks <http://lachy.id.au/log/2005/05/script-comments>
 137         $data = preg_replace('#//\s*$#', '', $token->data);
 138         return '<!--//--><![CDATA[//><!--' . "\n" . trim($data) . "\n" . '//--><!]]>';
 139     }
 140
 141     /**
 142      * Generates attribute declarations from attribute array.
 143      * @note This does not include the leading or trailing space.
 144      * @param $assoc_array_of_attributes Attribute array
 145      * @param $element Name of element attributes are for, used to check
 146      *        attribute minimization.
 147      * @return Generate HTML fragment for insertion.
 148      */
 149     public function generateAttributes($assoc_array_of_attributes, $element = false) {
 150         $html = '';
 151         if ($this->_sortAttr) ksort($assoc_array_of_attributes);
 152         foreach ($assoc_array_of_attributes as $key => $value) {
 153             if (!$this->_xhtml) {
 154                 // Remove namespaced attributes
 155                 if (strpos($key, ':') !== false) continue;
 156                 // Check if we should minimize the attribute: val="val" -> val
 157                 if ($element && !empty($this->_def->info[$element]->attr[$key]->minimized)) {
 158                     $html .= $key . ' ';
 159                     continue;
 160                 }
 161             }
 162             $html .= $key.'="'.$this->escape($value).'" ';
 163         }
 164         return rtrim($html);
 165     }
 166
 167     /**
 168      * Escapes raw text data.
 169      * @todo This really ought to be protected, but until we have a facility
 170      *       for properly generating HTML here w/o using tokens, it stays
 171      *       public.
 172      * @param $string String data to escape for HTML.
 173      * @param $quote Quoting style, like htmlspecialchars. ENT_NOQUOTES is
 174      *               permissible for non-attribute output.
 175      * @return String escaped data.
 176      */
 177     public function escape($string, $quote = ENT_COMPAT) {
 178         return htmlspecialchars($string, $quote, 'UTF-8');
 179     }
 180
 181 }
 182
 183 // vim: et sw=4 sts=4