Why is there a Float.X? rather than using float32, float64

[ring-c]

Why are we using custom type Float.X and not a float directly?

[Splizard]

Engine floats can be 32 or 64 bit depending on build configuration but Go doesn't have a builtin variable sized float type like it does for int. So Float.X is used to represent a float that could either be 32bits or 64bits. It's also the same number of characters to type as float32 and float64.

[Splizard]

It doesn't require an import as it is an alias of either float32 or float64 depending on build tags. If you are using float32 (which is the default) then you can use float32 interchangeably with Float.X in your code (but then it wouldn't support double precision engine builds, which would change the alias to float64).

I think it is convenient to use Float.X consistently when Godot is built with float precision (which is the default), so that you don't have to convert between all of the component fields for the math types every-time you use them, like Vector2.XY, Vector3.XYZ etc.

This means If you have a pos.X, you can pass it directly to an Engine function that accepts a float, without needing lots of conversions. If you want to pass float64 directly to an engine function without precision loss, you can use the Advanced class interface, which exposes engine types 1:1.

[Splizard]

It sounds like you may be having a bad experience with Float.X, can you share some more information on why this is?

[ring-c]

I didn't know that was possible to use float32 directly. My code is full of Float.X as arguments for PhysicsProcess, for example. So, if i build godot with double precision, I can use float64 everywhere?

[Splizard]

Yes, you will need to use -tag precision_double when building the Go shared library. Note that I've primarily been using the standard Godot engine builds, so double precision hasn't really been tested yet.