Am Montag, 24. Januar 2022, 01:22:05 CET schrieb Albert Astals Cid:
> Are you going to propose the same for Linux and FreeBSD where we also have
> long running tests that don't succeed and no one bothers to fix them?

Yes, if a test is known to fail, and no solution is known, it should be tagged 
as EXPECT_TO_FAIL. Is there would be projects that do not do that (e.g. if the 
maintainer only cares for Windows, to turn things around), for those platforms 
where tests keep failing eternally unhandled we should drop the claim of 
support (and wasting CI resources).

Cheers
Friedrich


Reply via email to