I am playing around with the experimental flow control functionality in ServerCallStreamObserver.
I've got a couple questions on how flow control works.
From what I can see, the request-side flow control is actually managed through the passed in responseObserver object.
In order to take control, I should do something along the lines of:
((CallStreamObserver) responseObserver).disableAutoInboundFlowControl();
Then, I can request additional messages through:
((CallStreamObserver) responseObserver).request(1);
Is this the recommended way to do this?
Is there a way, or plans to provide a way, to manage flow control through the request StreamObserver returned by the application code handler?
This is a bit awkward for my use case as my request StreamObserver will need to keep a reference to the response StreamObserver in order to manage the request-side flow control.