On Wed Aug 16 08:25:48 2023 +0000, Jacek Caban wrote:
I think that ieframe doesn't need this mechanism, it would be better to just adjust the test. I created !3577 for that. I also hope that we could do better in other tests. For example, ntdll tests could have a loop in the beginning of `test_file_full_size_information` that would keep querying until sizes match. I'm generally skeptical that we need such a global mechanism.
* I don't think MR!3577 really fixes the ieframe:webbrowser race condition. * Looping at the beginning of test_file_full_size_information() sounds like manually reimplementing this mechanism. I don't see how it is better to have a dozen different implementations instead of a single one that everyone understands.
Also there are quite a few tests that need this mechanism so I think it's better to have a general solution rather than ad-hoc ones everywhere.