V1.1.14 Added support for (-x)^y

The fp64_pow()-function implements xy. The implementation now supports x being a negative number and y being a positive integer, e.g. (-2)5 will now correctly return -32 instead of NaN.