public static byte[] Decode(string input)
{
var bytes = DecodeToBigInteger(input).ToByteArray();
// We may have got one more byte than we wanted, if the high bit of the next-to-last byte was not zero. This
// is because BigIntegers are represented with twos-compliment notation, thus if the high bit of the last
// byte happens to be 1 another 8 zero bits will be added to ensure the number parses as positive. Detect
// that case here and chop it off.
var stripSignByte = bytes.Length > 1 && bytes[0] == 0 && bytes[1] >= 0x80;
// Count the leading zeros, if any.
var leadingZeros = 0;
for (var i = 0; input[i] == _alphabet[0]; i++)
{
leadingZeros++;
}
// Now cut/pad correctly. Java 6 has a convenience for this, but Android can't use it.
var tmp = new byte[bytes.Length - (stripSignByte ? 1 : 0) + leadingZeros];
Array.Copy(bytes, stripSignByte ? 1 : 0, tmp, leadingZeros, tmp.Length - leadingZeros);
return tmp;
}
