Ironically, a few days after my post explaining why you shouldn’t write tests based on method call verification, I ended up finding a case where it’s not only useful but also the Right Thing To Do™.
This came up after I reported SPR-13615 and saw the patch that was submitted as a result. The patch uses method call verification in the tests correctly because the intended behavior is that the framework calls that specific method.
This is something that I didn’t think of because I don’t usually write real framework code but, in hindsight, definitely makes sense.
Never say never, I guess.
