fix up Wconversion issues in float_math/double_math

We are fixing all Wconversion as part of issue (TODO), however fixing the ones in float_math/double_math is much more risky, and these functions pass exhaustive tests as is, so I have disabled Wconversion files for these for now