On Tue Nov 19 19:58:27 2024 +0000, Jacek Caban wrote:
I guess it depends on the actual implementation, but I'd expect a single way of handling todos (that is replacing your current logic instead of having two mechanisms for the same thing) to be nicer.
Well I can't see how it can work without two mechanisms (meaning, separated) because one tests lack of prop on wine (but found on Windows), and one tests extra prop on wine (but not found on Windows). Just think from Windows' perspective: the former must be tested (because it exists on Windows), but the latter either must not be tested, or tested for non-existence (IMO it should just not be tested, considering it's wine's fault and eventually fixed).
Basically, they're different types of todos.
Also unrelated but, idk why the test fails on specific test machine with that `deviceSessionId` prop. I suppose I should mark that prop as flaky or something? Considering it's not available on other test machines.