On Sat, Mar 20, 2010 at 5:52 PM, Noah Slater <[email protected]> wrote: > Historically, test suite failures on Futon were release blockers. I'd like > to get some consensus on if this is still the case. It's imperative we have > a formal rule here, to help me, and to help in-between development. Once > we've cleared that up, I suggest that we make this part of the documented > release procedure. That way, asking for a release signifies that the test > suites pass acceptably in the environments we have agreed the must pass in. > I think this will reduce a significant amount of friction for future > releases. If I get blocked, or stalled, in any way - that should indicate a > process failure. Thoughts?
The changes failure may be more important than a simple browser thing. I noticed this temporary failure in couchdbkit tests too. I think that we should make sure all tests pass in ff at least to declare this ok. - benoit
