Draft and Discussion to figure out custom namecall self type definition

[karl-police]

The RFC: https://github.com/karl-police/rfcs/blob/patch-1/docs/user_defineable_self_namecall_type.md

It's basically just all about how self isn't controllable compared to Class.TestMethod(self: insert_type_here)

function Class:TestMethod()
        -- Default Compile Optimization thanksfully seems to skip this, which is good.
        -- Otherwise this alternative would have a optimization impact.
	self = self :: ClassObject
end

and how : takes away the first argument, but I am not sure if this would cause issues for the typesolver. It seems to work fine when using --!strict when doing self = self :: ClassObject

and when looking at the Bytecode

I don't see a instruction in the code where it is processing self = self which makes the current alternative within pure : functions possible, but annoying.

But if Bytecode Compilation ever gets changed by Luau where self = self would count as a MOVE instruction, then this would be bad.

[andyfriesen]

Some of this is going to get a lot better when we implement https://github.com/luau-lang/rfcs/blob/master/docs/shared-self-types.md