[Summary view]

   1  <?php
   2  
   3  /**
   4   * Class that handles operations involving percent-encoding in URIs.
   5   *
   6   * @warning
   7   *      Be careful when reusing instances of PercentEncoder. The object
   8   *      you use for normalize() SHOULD NOT be used for encode(), or
   9   *      vice-versa.
  10   */
  11  class HTMLPurifier_PercentEncoder
  12  {
  13  
  14      /**
  15       * Reserved characters to preserve when using encode().
  16       * @type array
  17       */
  18      protected $preserve = array();
  19  
  20      /**
  21       * String of characters that should be preserved while using encode().
  22       * @param bool $preserve
  23       */
  24      public function __construct($preserve = false)
  25      {
  26          // unreserved letters, ought to const-ify
  27          for ($i = 48; $i <= 57; $i++) { // digits
  28              $this->preserve[$i] = true;
  29          }
  30          for ($i = 65; $i <= 90; $i++) { // upper-case
  31              $this->preserve[$i] = true;
  32          }
  33          for ($i = 97; $i <= 122; $i++) { // lower-case
  34              $this->preserve[$i] = true;
  35          }
  36          $this->preserve[45] = true; // Dash         -
  37          $this->preserve[46] = true; // Period       .
  38          $this->preserve[95] = true; // Underscore   _
  39          $this->preserve[126]= true; // Tilde        ~
  40  
  41          // extra letters not to escape
  42          if ($preserve !== false) {
  43              for ($i = 0, $c = strlen($preserve); $i < $c; $i++) {
  44                  $this->preserve[ord($preserve[$i])] = true;
  45              }
  46          }
  47      }
  48  
  49      /**
  50       * Our replacement for urlencode, it encodes all non-reserved characters,
  51       * as well as any extra characters that were instructed to be preserved.
  52       * @note
  53       *      Assumes that the string has already been normalized, making any
  54       *      and all percent escape sequences valid. Percents will not be
  55       *      re-escaped, regardless of their status in $preserve
  56       * @param string $string String to be encoded
  57       * @return string Encoded string.
  58       */
  59      public function encode($string)
  60      {
  61          $ret = '';
  62          for ($i = 0, $c = strlen($string); $i < $c; $i++) {
  63              if ($string[$i] !== '%' && !isset($this->preserve[$int = ord($string[$i])])) {
  64                  $ret .= '%' . sprintf('%02X', $int);
  65              } else {
  66                  $ret .= $string[$i];
  67              }
  68          }
  69          return $ret;
  70      }
  71  
  72      /**
  73       * Fix up percent-encoding by decoding unreserved characters and normalizing.
  74       * @warning This function is affected by $preserve, even though the
  75       *          usual desired behavior is for this not to preserve those
  76       *          characters. Be careful when reusing instances of PercentEncoder!
  77       * @param string $string String to normalize
  78       * @return string
  79       */
  80      public function normalize($string)
  81      {
  82          if ($string == '') {
  83              return '';
  84          }
  85          $parts = explode('%', $string);
  86          $ret = array_shift($parts);
  87          foreach ($parts as $part) {
  88              $length = strlen($part);
  89              if ($length < 2) {
  90                  $ret .= '%25' . $part;
  91                  continue;
  92              }
  93              $encoding = substr($part, 0, 2);
  94              $text     = substr($part, 2);
  95              if (!ctype_xdigit($encoding)) {
  96                  $ret .= '%25' . $part;
  97                  continue;
  98              }
  99              $int = hexdec($encoding);
 100              if (isset($this->preserve[$int])) {
 101                  $ret .= chr($int) . $text;
 102                  continue;
 103              }
 104              $encoding = strtoupper($encoding);
 105              $ret .= '%' . $encoding . $text;
 106          }
 107          return $ret;
 108      }
 109  }
 110  
 111  // vim: et sw=4 sts=4