Shawn Walker wrote: > On Wed, Mar 19, 2008 at 12:50 PM, Jim Walker <James.Walker at sun.com> wrote: >> Alan Coopersmith wrote: >> >> Yes core on ASSERT should be disabled before any product release. The >> >> intention is to uncover these subtle bugs during development so that >> >> they won't lead to subtle errors in the product. >> > >> > So that sounds like they should always be disabled in the Nevada builds >> > of JDS and only enabled in the Vermilion builds, since we're making >> > "products" like SXDE & Indiana out of many Nevada builds (and once the >> > Indiana repos are going, basically all of them). >> >> I agree. The current OpenSolaris releases should be considered product >> releases and desktop stability overrides uncovering minor bugs >> especially when thousands of people are impacted. >> >> The JDS team could push out there own "beta test versions" if they want, >> but the main path stuff should be product release quality all the time. > > The only problem I see with this is that it reduces the ability to get > good test feedback. > > While FCS quality all the time is a great thing to have, the > OpenSolaris releases are test releases for all intents and purposes > from my understanding.
I for one am very interested in good test feedback. The problem here is this crosses the line and will hurt us more than help us. First time "users" aren't going to put up with multiple desktop crashes. Maybe we need an even odd release approach like gnome, or other method to get the right test feedback verses stable product balance. Cheers, Jim
