On Fri, Nov 30, 2001 at 07:37:48AM +0000, Jules Bean wrote:
> The fact is that the current procedure is both more complex and
> cleverer than the freeze procedure we are used to. As far as I know,
> it's also undocumented (correct me if I'm wrong). Hence there are
> lots of questions: people want to understand what's going to happen to
> their packages; they want to know the justification for certain
> choices.
Well, that's all well and good, but it's not really reasonable to
try to do an analysis of an experiment in the middle of it (one of the
justifications for testing -- and the one under consideration here -- is
to try to imporve release management). The basic justifications of all
this stuff have been stated when they've been announced, by and large.
What hasn't been justified in many cases -- including this one -- is why
this solution's been chosen rather than some other possible one. For
example, "why don't we have a `frozen' dist in between testing and
unstable?" [0], and "why don't we change katie and britney to handle
the BTS in such-n-such a way?" [1], and "why don't the autobuilders build
against libraries in testing rather than unstable?" [2].
That's not to say these aren't good questions to be asking: they are. But
the right time to ask them isn't now, it's after the release is over
when we have some experience in how all this new stuff actually works,
and can afford to take some time and develop new software to make the
whole process work better.
Cheers,
aj
[0] Archive bloat, incorrectly targetted uploads causing the same problems
Adrian's worried about, difficulty autobuilding.
[1] Difficulty of implementation is a showstopper, general complexity is
another problem. That it doesn't actually help with RC bugs that
were closed in May in unstable, but haven't been fixed in testing yet
makes it relatively uninteresting too. (For reference, britney is
the dak-name of the testing scripts)
[2] Because that would stop new libraries from ever getting into testing,
or ever getting tested.
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
"Security here. Yes, maam. Yes. Groucho glasses. Yes, we're on it.
C'mon, guys. Somebody gave an aardvark a nose-cut: somebody who
can't deal with deconstructionist humor. Code Blue."
-- Mike Hoye,
see http://azure.humbug.org.au/~aj/armadillos.txt
Attachment:
pgpKxw8FggYL1.pgp
Description: PGP signature