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

Re: Bits from the Release Team - Kicking off Wheezy

On Sun, May 01, 2011 at 10:41:07PM +0200, Stefano Zacchiroli wrote:
> On Sun, May 01, 2011 at 10:08:46PM +0200, Pierre Habouzit wrote:
> > Those are real users from real life. I'm not saying "we"-re
> > representative of a majority of Debian Users, but unlike all the
> > handwaived users we've read about in this thread, those are real.
> First of all I think you should concede that the exercise you're asking
> us to do cannot be done as easily as you did yours. You've been able to
> mention real users of an existing distro, you are asking others to
> provide evidence of users of a thing that does not exist yet; quite
> challenging... So you do have an inherent advantage, but let me try
> nonetheless.

I don't concede that. I've read your mail, and to sum up you say:

"I've been in contact with lots of people interested into testing", Okay
fine, I buy that, I won't call you a liar.

So the next question is "why" your mail doesn't answer that. I still
think that rolling is a bad idea, until you've proven me that it's the
sole way to address a real life issue/need/itch.

So why do Aptosid/Sidux/... exist, why do people think we should sell
testing more? You don't answer that and it's critical to know it.

IOW what does having CUT/rolling fixes please.

FWIW I see one problem with the current testing, tied to the freeze,
which is that a 6-months freeze means that we skip a release for all
those projects that release twice a year, and those are many. Which is
annoying for some of our users (gnome releases when it's not gnome3 are
pretty sound e.g. it's not really bleeding edge that harms), and even
more for packagers because it's harder not to package all upstream
versions, it means that the upgrade is harder to deal with. But *that* I
think can be adressed with a better experimental (using PPA, but we
discussed that already and I think it's a consensus).

And for that I think that rolling isn't the solution because of the
"entry-point" issue I talked elsewhere: You still need to update testing
during the freeze.  And the larger the project, the higher the
probability is that:
  - users will want the last version available;
  - you will have updates to send during the freeze.

I surely miss other uses cases, so please enlighten me.

·O·  Pierre Habouzit
··O                                                madcoder@debian.org
OOO                                                http://www.madism.org

Reply to: