Re: Another "testing" vs "unstable" question
On 2004-06-24, John Summerfield penned:
>
> Its only since its IPO that RH has become money-hungry. I am
> comfortable with the notion of paid-for support in the way of
> security advisories and bug-fixes: the only matter for debate is cost.
Well, if I understood you earlier, you have paying clients. I guess
having a paid support contract is a nice CYA maneuver in that kind of
situation.
(I like debian better, but then, I've never tried the paid version of
linux support; maybe it's just fantabulous.)
> Indeed. While I disagree with much of the Debian project (before you
> jump in, I'd point ot that many of the Debian Developers disagree with
> each other too), I do admire their endevour and commitment to the
> project.
Gd, do they ever disagree!
I don't disagree with much of the project, but I'm right there with you.
I think it's a lot like a quote I heard about the ACLU -- "If you agree
with half of what we do, you should contribute. If you agree with 75%
of what we do, you should be on our Board of Directors!" Something like
that.
>>No, what's missing is the testing infrastructure. *System* testing,
>>not just the individual package.
>>
> Better, I think to seek ways towards that ideal. Some cliches come to
> mind - the person who makes no mistakes does nothing, a journey of a
> thousand leagues begins with a single step...
Right. The question is whether the product can realistically be
improved/sped up or not. I'm reminded of that whole "nine women can't
make a baby in one month" business.
> I haven't yet seen a Debian beta process, so I don't know what
> happens, but if (as I've read) the DDs are mostly running testing or
> unstable, then there has to be something wrong in _their_ estimation
> with Woody.
Er. They *have* to run testing or unstable in order to test their
packages! Not all of them have multiple boxes (or even permanent
network connections); many of them may not be running mission-critical
systems at home; and they're all experienced enough not to have to run
stable to avoid the fear of accidentally doing a Bad Thing.
I'm pretty sure all the debian servers run stable, although it would be
interesting to hear if they don't.
> The recent move to subversion has had the effect of officially cutting
> Woody users off from the latest source - there is no offical Woody
> build of subversion.
Eh? Whose recent move to subversion? I've been distracted by
non-computer things recently; have I missed something?
>>And now a lot of people who aren't motivated enough to do a google
>>search or ask on d-u are installing packages that haven't been fully
>>tested with the system. The status quo at least ensures that the
>>people who are using backports have at a minimum the ability to
>>research questions.
>>
> Do you think the current situration is perfect? If not, how do _you_
> think it may be improved.
There's a difference between "imperfect" and "needs to be fixed." I
stand by my belief that adding packages after the official release
introduces risk. Now, would releasing a new version of stable more
often be a good thing? I guess it depends on if it's deemed truly
stable.
Okay, I'm way too tired for rational thought right now. Must go beddy
bye ...
--
monique
Reply to: