On Tue, Apr 15, 2008 at 09:46:26AM +0200, Johannes Wiedersich wrote: > Joey Hess wrote: Both of you, thanks! > > Andrew Sackville-West wrote: > >> In your opinion, am I right in my assessment that testing is more > >> likely to be in an unusable state for longer than sid? (at least at > >> the package, not system, level)? > > > > No, I don't think so. If a package has a bug that makes it unusable, > > then > > > > a) Someone will generally notice a bug in the two weeks before that buggy > > package gets into testing, and file a RC bug to keep it out. > > b) If a bug that makes a package unusable does get into testing, it > > can be fixed in 2 days in most cases. > > c) The graph of release critical bugs[1] currently shows 1750 in unstable, > > and only 571 of those affect testing. (658 of them affect *stable*). > > http://bugs.debian.org/release-critical/ > Interesting to see that Etch has more RC bugs than Lenny at this point. > I second the experience that there are not too many, if any serious > problems with testing. I've been using testing on 'newer' computers -- > especially laptops -- that wouldn't run well with stable for a total of > years and never had any serious problem. > > Testing has the benefit that -- unlike unstable -- it will eventually > become stable. I enjoy the 'quiescence' that sets in after testing > becomes stable and there is nothing happening to my system, except for > trivial security fixes. If you are like me and generally prefer 'stable' > software, than 'testing' is your route. Well, then I'll adjust my view accordingly ;) I seem to recall *something* (who knows what at this point) that slipped into testing early after the sarge release and because of a series of unfortunate events, the poor testing users were stuck with serious problems while those in sid blithely moved along. I'll chalk it up to either a one-off situation or (more likely) corrupted bits in the ol' wetware. A
Attachment:
signature.asc
Description: Digital signature