[Summary view] [Print] [Text view]

   1  <?php
   2  
   3  /**
   4   * Validates a host according to the IPv4, IPv6 and DNS (future) specifications.
   5   */
   6  class HTMLPurifier_AttrDef_URI_Host extends HTMLPurifier_AttrDef
   7  {
   8  
   9      /**
  10       * Instance of HTMLPurifier_AttrDef_URI_IPv4 sub-validator
  11       */
  12      protected $ipv4;
  13  
  14      /**
  15       * Instance of HTMLPurifier_AttrDef_URI_IPv6 sub-validator
  16       */
  17      protected $ipv6;
  18  
  19      public function __construct() {
  20          $this->ipv4 = new HTMLPurifier_AttrDef_URI_IPv4();
  21          $this->ipv6 = new HTMLPurifier_AttrDef_URI_IPv6();
  22      }
  23  
  24      public function validate($string, $config, $context) {
  25          $length = strlen($string);
  26          if ($string === '') return '';
  27          if ($length > 1 && $string[0] === '[' && $string[$length-1] === ']') {
  28              //IPv6
  29              $ip = substr($string, 1, $length - 2);
  30              $valid = $this->ipv6->validate($ip, $config, $context);
  31              if ($valid === false) return false;
  32              return '['. $valid . ']';
  33          }
  34  
  35          // need to do checks on unusual encodings too
  36          $ipv4 = $this->ipv4->validate($string, $config, $context);
  37          if ($ipv4 !== false) return $ipv4;
  38  
  39          // A regular domain name.
  40  
  41          // This breaks I18N domain names, but we don't have proper IRI support,
  42          // so force users to insert Punycode. If there's complaining we'll
  43          // try to fix things into an international friendly form.
  44  
  45          // The productions describing this are:
  46          $a   = '[a-z]';     // alpha
  47          $an  = '[a-z0-9]';  // alphanum
  48          $and = '[a-z0-9-]'; // alphanum | "-"
  49          // domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum
  50          $domainlabel   = "$an($and*$an)?";
  51          // toplabel    = alpha | alpha *( alphanum | "-" ) alphanum
  52          $toplabel      = "$a($and*$an)?";
  53          // hostname    = *( domainlabel "." ) toplabel [ "." ]
  54          $match = preg_match("/^($domainlabel\.)*$toplabel\.?$/i", $string);
  55          if (!$match) return false;
  56  
  57          return $string;
  58      }
  59  
  60  }
  61  
  62  // vim: et sw=4 sts=4