Add `engine.math` overloads that work with doubles

[accmltr]

Make this work:

val f: Float = 0.23413229812341
val d: Double = 0.23413229812341
println(s"float 'math.abs' result: ${engine.math.abs(f)}")
println(s"double 'math.abs' result: ${engine.math.abs(d)}")

Then users and engine devs can seamlessly switch to higher precision for operations that require it.