[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [hertzog@debian.org: Re: Woody retrospective and Sarge introspective]



On Wed, Jul 31, 2002 at 01:03:47AM +0200, Raphael Hertzog wrote:
> > > Now that woody is released, it's not useful anymore of course. But while
> > 
> > Then why do we even have a stable distribution, if the attitude of maintainers
> > is that work on keeping it maintained is not useful anymore?
> 
> Explain that to Joey who doesn't accept anything except security
> updates. Don't rely on me for changing that though. :-)

Bug fixes have gotten in to stable releases in the past, if they warrant it.

> candidate (supposed to be woody in that case) while having XF4.2 in
> unstable and something else in stable ...

Sure, and the same could be done with testing as it is.

> > I think what people are missing here is the existance of a working
> > testing-proposed-updates suite, which we didn't have with autobuilder
> > support until very close to release time.  I don't see any gains of
> 
> packages uploaded to t-p-u are built against which dist ?

testing.

> > as few people have it in sources, so it's just like allowing untested
> > uploads (much like your proposed "candidate" distribution).
> 
> Candidate is not for "untested uploads" even though "direct uploads" are
> possible because some changes (eg documentation) may not need a complete

Yes, but I really don't think that most developers are going to do anything
special for candidate uploads.  As other people have mentioned, otherwise
they may well forget to do so altogether and candidate will never get
anything new.  Or, people will just upload to unstable and candidate, and
it becomes equally as useless.

> approval cycle, because some uploads are needed quickly (security
> updates).

Which is what t-p-u could be used for.  What does a candidate distribution
buy us besides more space on the mirrors, and more work for the developer?
Your complaint seemed to be that no one could work on unstable, because
then they couldn't update testing if it proved needed.  This is no longer
the case with t-p-u.

> > unstable has it's name for a reason.  Many people don't seem to understand
> > that, and remain calm when everything blows up around them.  Instead they
> > send a screaming rant of a bug report with high severity, which makes the
> > maintainer not want to treat unstable as what it's description says it
> > is.
> 
> Yes and ? I was not suggesting to upload totally screwed uploads to
> unstable ... i'm just saying that the sooner the package hits unstable, the
> sooner we have a package ready (whatever is the status of the candidate dist
> ...).

Uploads to unstable are irrelevant for candidate the way you've described it,
as are packages and uploads in/to testing.  One would _hope_ that it is the
case that they hit unstable first, but you've said nothing that indicates
this is required.  The current testing setup gives us this.  Again, why
do we need another dist?

-- 
Ryan Murray, Debian Developer (rmurray@cyberhqz.com, rmurray@debian.org)
The opinions expressed here are my own.

Attachment: pgp1z1TmcVs0J.pgp
Description: PGP signature


Reply to: