Random.rand(max : Float32) should return a Float32, not Float64

[j8r]

Random.rand takes any Float and returns a Float64. to_f32 is needed to cast back to a Float32.
To be consistent with other methods taking integers, like Random.rand(max : Int) : Int, which returns the same type as the given integer, I suggest to have the method returning a Float32 if one was given.

Example of the current behavior:
p typeof(Random.rand(1_f32)) #=> Float64

[straight-shoota]

Sounds very reasonable to return the same type that was given.