But wouldn't it be nice to have a Real(tm) beta-test cycle for once? It
appears that we still have problems with stability in 2.1, given the
amount of reboots we're seeing, and I'd like to see those resolved. I
just started seeing reboots on my 2.1 box which ran the exact same
workload under 2.0R with uptimes of 60 and 90 days, and I can't get over
a week with it in -stable. I just enabled dumps, so hopefully I can
provide more information.
Well, I'm all for fixing the bugs but I don't think that declaring a
"beta" will help much. Consider all the time we had between 2.0.5 and
2.1 to shake out the bugs and they clearly still didn't get shaken
out. People's tolerance to BETAs seems to have dwindled to the point
where I'm lucky if I get 3-4 mails during the BETA cycle. Doesn't
exactly fill me with confidence that declaring a 2-3 week BETA for
2.1.1 is going to net us anything but gratuitous delay.
It's sort of like the 4MB installation problem - the time to start
working on it is *now*, not 3 weeks before the release, and the same
goes double for any kernel pathogens.