use Math::BigRat;  # Useless
use bignum;        # Uses M::BigInt and M::BitFloat

##</code><code>##

use bignum upgrade => "Math::BigRat";

##</code><code>##

use bigrat;