Pow support

[cdchamness]

Issue #118 asked for implementation of the Pow trait for OrderedFloat and NotNan. I have implemented the Pow trait for both in 2 ways.

1. This implements the trait for exponents that are primitive types (i8, i16, i32, u8, u16, f32, f64) that num_traits does, https://docs.rs/num-traits/latest/num_traits/pow/trait.Pow.html, for the f32 and f64 types respectfully.
2. This implements the trait for exponents and bases that both either OrderedFloat or NotNan.

I have not implemented any mixing of OrderedFloat and NotNan, as the proper return type is not obvious. I have also not implemented Pow for the base being primitive (i8, i16, i32, u8, u16, f32, f64) with an exponent of either OrderedFloat or NotNan, as again the return type for this operation is not clear to me.

I also added a few tests to ensure things are working as expected.

[mbrubeck]

Thanks! It looks like these just need to be #[cfg(feature = "std")].