On Mon Sep 1 19:29:58 2025 +0000, Rémi Bernon wrote:
Well most of the other variants of the helpers have an extra parameter to check that an interface is *not* supposed to be there, therefore they have to handle the failure case. It's useful in such case, and it'd be useful too if the helper were called within todo_wine but it's not the case here. But I was also talking more generally, we usually don't handle error cases in tests unless they actually happen under some configuration.
I agree with Rémi. In general checking seems a waste of time. If a test fails it needs to be addressed, whether that failure is just an incremented failure count or a crash.