Re: proposing bash-3.0 upload to unstable

[Martin Michlmayr <tbm@cyrius.com>]

* Steve Langasek <vorlon@debian.org> [2004-07-27 23:31]:
> If you feel it only justifies an urgency=medium upload, then that's
> what I would recommend.  This means it probably won't make the
> freeze, since medium is 5 days and the 31st is 4 days from now, but
> those are the breaks.

I think the 31st should apply to uploads to unstable rather than
testing.  You basically came with your freeze announcement completely
out of the blue and said testing will freeze in 6 days, not giving
maintainers any chance at all to make normal uploads which take 10
days to propogate to testing.  This could lead to a) packages being
uploaded in a rush even tough they are not well tested and b) them
being uploaded at high severity to make the freeze, therefore getting
even less testing.  I think this is a risk you should not promote and
therefore the 31st should be a rough cut-off date for unstable,
freezing testing about a week or 10 days later.  (Of course, I have no
problems with and think it's a good idea to freeze testing early and
hand-approve uploads, but the policy of "RC only" shouldn't go into
effect on 31st because of the reasons above.)

In future announcements, I'd like to see the point stressed that
freezing testing != freezing unstable and that people have to make
sure to upload well in advance.  I'd explicitly mention "we're
freezing testing on day 15 which means the last upload to unstable
that will go in _in the best cirumstance_ (not considering delays and
dependencies) has to be done on the 5th (4th?)".

-- 
Martin Michlmayr
tbm@cyrius.com