querystring.unescape

// a safe fast alternative to decodeURIComponent
QueryString.unescapeBuffer = function(s, decodeSpaces) {
  var out = new Buffer(s.length);
  var state = 'CHAR'; // states: CHAR, HEX0, HEX1
  var n, m, hexchar;
  for (var inIndex = 0, outIndex = 0; inIndex <= s.length; inIndex++) {
    var c = s.charCodeAt(inIndex);
    switch (state) {
      case 'CHAR':
        switch (c) {
          case charCode('%'):
            n = 0;
            m = 0;
            state = 'HEX0';
            break;
          case charCode('+'):
            if (decodeSpaces) c = charCode(' ');
            // pass thru
          default:
            out[outIndex++] = c;
            break;
        }
        break;
      case 'HEX0':
        state = 'HEX1';
        hexchar = c;
        if (charCode('0') <= c && c <= charCode('9')) {
          n = c - charCode('0');
        } else if (charCode('a') <= c && c <= charCode('f')) {
          n = c - charCode('a') + 10;
        } else if (charCode('A') <= c && c <= charCode('F')) {
          n = c - charCode('A') + 10;
        } else {
          out[outIndex++] = charCode('%');
          out[outIndex++] = c;
          state = 'CHAR';
          break;
        }
        break;
      case 'HEX1':
        state = 'CHAR';
        if (charCode('0') <= c && c <= charCode('9')) {
          m = c - charCode('0');
        } else if (charCode('a') <= c && c <= charCode('f')) {
          m = c - charCode('a') + 10;
        } else if (charCode('A') <= c && c <= charCode('F')) {
          m = c - charCode('A') + 10;
        } else {
          out[outIndex++] = charCode('%');
          out[outIndex++] = hexchar;
          out[outIndex++] = c;
          break;
        }
        out[outIndex++] = 16 * n + m;
        break;
    }
  }
  // TODO support returning arbitrary buffers.
  return out.slice(0, outIndex - 1);
};