IMO, if some patch managed to get to master, it's worth being released,
even if IT test fail.
So I would make the first step being a `mvn deploy` of what[s in master,
before running other tests.
I think this would be much simpler, allow community to test some "master"
snapshot even if it's bugged (so community can help in resolving), and
would cover all use-cases I have about snapshot. And -to put it another
way- making complex effort to run IT tests on all OS before publishing a
snapshot is over-engineering.

Reply via email to