RFC: Refactoring SystemABI

[Maurizio Cimadamore]

On 22/11/2018 21:30, Jorn Vernee wrote:
> 3.) There is a shortcut being taken when native code gives us a 
> Pointer back to one of our upcall stubs, and we want to call it from 
> Java. In that case we don't really need to call into native, to call 
> our own upcall stub to then call back into Java to call our target 
> method. So I think the ABI should also have a method like `Object 
> getFunctorForStub(Pointer<?> ptr);` which takes the role of what 
> `UpcallHandler.getUpcallHandler` is doing right now; It retrieves the 
> UpcallHandler instance that's embedded in the stub, and asks it for 
> the functor object that it stores, or otherwise returns null. I think 
> for this last one to work all UpcallHandler implementations should 
> have a common super type to expose this receiver object. 

What you propose is analogous to what I proposed. Here you have a method 
(getFunctor) which gives you the receiver directly. In my proposed 
approach, the UpcallHandle interface has a getReceiver method. So you 
ask the handle to the ABI, and then you ask the receiver from the handle.

What is it that bothers you about this approach?

Maurizio