Glenn McGrath wrote:
I dont agree with Joel's answers but i think its wrong and counterproductive to discourage people from trying to improve debians release _process_
Just a minor thought/interjection: mightn't the entire processes be speeded up if developers were _encouraged_ to use 'testing' as their distribution? And I mean "encouraged to" like "publicly flamed and beaten soundly for not".
The only reason I say that is that testing simply gives a more stable, in general, way to keep on top of new developments, but does little to nothing to getting a new release out. On the other hand, if developers aren't even able to update to their own latest packages until they filter into testing, the distribution as a whole remains more suitable for release, because things like library dependencies are kept better in sync and safe updates should filter in more quickly.
The issue of unresolved release critical bugs will not go away, except for the fact that they shouldn't be pushed into testing at all, given that they are caught in time. The current crop of release critical bugs then are just a manifestation of the fact this this release of testing is no different that a traditional freeze of unstable, since Debian never had a clean testing distribution to start with because testing was created from unstable in the beginning, not from stable [iirc].
Either way, the delay is the typical glacial Debain release process, hopefully for the last time (then we have a stable testing). Now if only all developers developed against testing, we'd be one step closer to the endgame come new release time.
$0.02, Christopher