On Thu, 17 Jul 2025 at 15:59, Tomasz Kaminski wrote: > The standard just says that this is UB, so we are testing something that is > not > required to pass by the standard. This is why I am a bit hesitant on death > tests like that. > Jonathan, do you have an opinion on this?
If we have an assertion there, it's good to test that it does actually catch the bug. I wouldn't insist on a test if Luc hadn't already written it, but since he's done it, I think we might as well keep it! They're only _GLIBCXX_DEBUG_ASSERT which is not expected to be fast, so if it catches more bugs that's good.
