Added some specific float and double rounding implementations (to be used by FastDecimalFormat):
* FMath::RoundHalfToEven
* FMath::RoundHalfFromZero
* FMath::RoundHalfToZero
* FMath::RoundFromZero
* FMath::RoundToZero
* FMath::RoundToNegativeInfinity
* FMath::RoundToPositiveInfinity
Added a generic platform wrapper for the float and double modf function (FMath::Modf).
Added some missing double implementations (FMath::RoundToDouble, FMath::CeilToDouble).
#platformnotify Josh.Adams
[CL 2677455 by Jamie Dale in Main branch]