Having in all cases dozen of skipped test just makes the results readability worse. So it makes little sense unless we can figure out way to test both implementations. Signed-off-by: Michal Čihař <michal@cihar.com>