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      * Configuration for the generator
  31      */
  32     protected $config;
  33
  34     /**
  35      * @param $config Instance of HTMLPurifier_Config
  36      * @param $context Instance of HTMLPurifier_Context
  37      */
  38     public function __construct($config = null, $context = null) {
  39         if (!$config) $config = HTMLPurifier_Config::createDefault();
  40         $this->config = $config;
  41         $this->_scriptFix = $config->get('Output', 'CommentScriptContents');
  42         $this->_def = $config->getHTMLDefinition();
  43         $this->_xhtml = $this->_def->doctype->xml;
  44     }
  45
  46     /**
  47      * Generates HTML from an array of tokens.
  48      * @param $tokens Array of HTMLPurifier_Token
  49      * @param $config HTMLPurifier_Config object
  50      * @return Generated HTML
  51      */
  52     public function generateFromTokens($tokens) {
  53         if (!$tokens) return '';
  54
  55         // Basic algorithm
  56         $html = '';
  57         for ($i = 0, $size = count($tokens); $i < $size; $i++) {
  58             if ($this->_scriptFix && $tokens[$i]->name === 'script'
  59                 && $i + 2 < $size && $tokens[$i+2] instanceof HTMLPurifier_Token_End) {
  60                 // script special case
  61                 // the contents of the script block must be ONE token
  62                 // for this to work.
  63                 $html .= $this->generateFromToken($tokens[$i++]);
  64                 $html .= $this->generateScriptFromToken($tokens[$i++]);
  65             }
  66             $html .= $this->generateFromToken($tokens[$i]);
  67         }
  68
  69         // Tidy cleanup
  70         if (extension_loaded('tidy') && $this->config->get('Output', 'TidyFormat')) {
  71             $tidy = new Tidy;
  72             $tidy->parseString($html, array(
  73                'indent'=> true,
  74                'output-xhtml' => $this->_xhtml,
  75                'show-body-only' => true,
  76                'indent-spaces' => 2,
  77                'wrap' => 68,
  78             ), 'utf8');
  79             $tidy->cleanRepair();
  80             $html = (string) $tidy; // explicit cast necessary
  81         }
  82
  83         // Normalize newlines to system defined value
  84         $nl = $this->config->get('Output', 'Newline');
  85         if ($nl === null) $nl = PHP_EOL;
  86         if ($nl !== "\n") $html = str_replace("\n", $nl, $html);
  87         return $html;
  88     }
  89
  90     /**
  91      * Generates HTML from a single token.
  92      * @param $token HTMLPurifier_Token object.
  93      * @return Generated HTML
  94      */
  95     public function generateFromToken($token) {
  96         if (!$token instanceof HTMLPurifier_Token) {
  97             trigger_error('Cannot generate HTML from non-HTMLPurifier_Token object', E_USER_WARNING);
  98             return '';
  99
 100         } elseif ($token instanceof HTMLPurifier_Token_Start) {
 101             $attr = $this->generateAttributes($token->attr, $token->name);
 102             return '<' . $token->name . ($attr ? ' ' : '') . $attr . '>';
 103
 104         } elseif ($token instanceof HTMLPurifier_Token_End) {
 105             return '</' . $token->name . '>';
 106
 107         } elseif ($token instanceof HTMLPurifier_Token_Empty) {
 108             $attr = $this->generateAttributes($token->attr, $token->name);
 109              return '<' . $token->name . ($attr ? ' ' : '') . $attr .
 110                 ( $this->_xhtml ? ' /': '' ) // <br /> v. <br>
 111                 . '>';
 112
 113         } elseif ($token instanceof HTMLPurifier_Token_Text) {
 114             return $this->escape($token->data, ENT_NOQUOTES);
 115
 116         } elseif ($token instanceof HTMLPurifier_Token_Comment) {
 117             return '<!--' . $token->data . '-->';
 118         } else {
 119             return '';
 120
 121         }
 122     }
 123
 124     /**
 125      * Special case processor for the contents of script tags
 126      * @warning This runs into problems if there's already a literal
 127      *          --> somewhere inside the script contents.
 128      */
 129     public function generateScriptFromToken($token) {
 130         if (!$token instanceof HTMLPurifier_Token_Text) return $this->generateFromToken($token);
 131         // Thanks <http://lachy.id.au/log/2005/05/script-comments>
 132         $data = preg_replace('#//\s*$#', '', $token->data);
 133         return '<!--//--><![CDATA[//><!--' . "\n" . trim($data) . "\n" . '//--><!]]>';
 134     }
 135
 136     /**
 137      * Generates attribute declarations from attribute array.
 138      * @note This does not include the leading or trailing space.
 139      * @param $assoc_array_of_attributes Attribute array
 140      * @param $element Name of element attributes are for, used to check
 141      *        attribute minimization.
 142      * @return Generate HTML fragment for insertion.
 143      */
 144     public function generateAttributes($assoc_array_of_attributes, $element = false) {
 145         $html = '';
 146         foreach ($assoc_array_of_attributes as $key => $value) {
 147             if (!$this->_xhtml) {
 148                 // Remove namespaced attributes
 149                 if (strpos($key, ':') !== false) continue;
 150                 // Check if we should minimize the attribute: val="val" -> val
 151                 if ($element && !empty($this->_def->info[$element]->attr[$key]->minimized)) {
 152                     $html .= $key . ' ';
 153                     continue;
 154                 }
 155             }
 156             $html .= $key.'="'.$this->escape($value).'" ';
 157         }
 158         return rtrim($html);
 159     }
 160
 161     /**
 162      * Escapes raw text data.
 163      * @todo This really ought to be protected, but until we have a facility
 164      *       for properly generating HTML here w/o using tokens, it stays
 165      *       public.
 166      * @param $string String data to escape for HTML.
 167      * @param $quote Quoting style, like htmlspecialchars. ENT_NOQUOTES is
 168      *               permissible for non-attribute output.
 169      * @return String escaped data.
 170      */
 171     public function escape($string, $quote = ENT_COMPAT) {
 172         return htmlspecialchars($string, $quote, 'UTF-8');
 173     }
 174
 175 }
 176