Casting an Int32 to a Float32 might lose precision?

Help & Support
1

Please check following example

def foo(x : Int32) : Int32
  x
end

def bar(x : Float32) : Float32
  x
end

bar(foo 1)     # Fails, casting an Int32 to a Float32 might lose precision

Why i consider, casting a Float32 to Int32 will lose precision?

2

First of all, the cast is Int32 to Float32, not the other way around. Second, not all Int32 can be represented exactly by a Float32. Which is not hard to see, considering they have the same amount of bits but Float32 covers a bigger range of numbers. That means there needs to be holes somewhere where two integers map to the same float.

3

Thanks for explain.

And, current compiler no this error message for now, it just output:

Error: no overload matches 'bar' with type Int32

Overloads are:
 - bar(x : Float32)

I copy some code from my note with comment, my fault.