The edge case where two possible paddings exist (e.g 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 xx) is not sufficiently tested; currently it can be bypassed by returning empty string for this case.

From this you know that the monarch's signature is " - HRH" at the end of a message.

Only the first test case in the example tests has the correct signature.