> > I don't want to go down this road too far right now, but I think all of the > changes we've put in this past week indicate that 4.2 is not stable enough > to be released. > > It is true that more issues have probably been fixed than opened; however, > I don't think we've had enough time to play with the code to be confident > enough that recent fixes have not broken anything critical. > > Perhaps in the future, we might at least consider a seven-day period after > a release candidate has been built where we wait and see if people > executing regression tests on the code find any additional Blocker or > Critical issues. If we do find some, they are fixed and a new RC is built, > and another seven-day period begins. > > I know that kind of a process messes with time-based releases, but I think > quality is more important than getting these out every four months. > > Honesty, with the current state of our regression tests, I don't see how we > can seriously entertain a four-month release cycle yet, but that has been > discussed before. > > My fear with 4.2 is that if we would have taken a week to regression test > its current RC build that we'd have found more Blocker and Critical bugs, > but - we won't find them until after the release with the current process. >
IMO it's a Catch-22. Most people don't test until we get an RC out. We don't get an RC out until we are ready to release. We haven't been successful in getting any large scale testing done outside of folks on Citrix's QA team - and FWIU they seem relatively happy with the current state of what they have tested. --David
