Investigate early warnings for callback API mock misses

I'm currently working with testdouble.js with a callback-based API, and if I miss a stubbing, I only know that when the test times out (or when I see it hang). Looking through the API again, I don't see a way to throw an early warning (or error) for missed mocks, which would be really helpful with these callback-based APIs (i.e. request).

Commence thinking out loud:

Stubbings are currently unordered, right? One solution would be to order them, so that stubbing A has to occur before stubbing B for any given double. If that were true, then we could throw on the first miss—your invocations were out of order. Not elegant, but it could be a valid solution.

Another would be to throw on any miss. Why don't we do this currently?

I'm currently running v2.0.2, which is otherwise working really well. I hooked it into a dependable container with zero issue to replace request proper, and it's perfect. 😍

Start by reviewing the stubbing and invocation behavior for callback-based APIs, especially the missed-mock case described for request. Compare ordered stubbings, early warnings, and throwing on any miss, then define the expected behavior and coverage needed for the chosen approach.