| Wouter Verhelst 2004-03-18, 9:41 am |
| On Wed, Mar 17, 2004 at 11:47:46PM +0100, Arnaud Vandyck wrote:
> Martin Michlmayr <tbm@cyrius.com> writes:
>
>
> <snip/>
>
>
> The problem is that there is no buildd for experimental packages.
>
> IT'd be good to have kaffe snapshots to experimental but it's useless
> without the buildd's.
>
> Is it possible to setup buildd for experimental?
This has been asked before. The answer is: not easily, if at all.
There are a few issues with setting up a buildd for experimental. One
problem is that experimental is for software that is _expected_ to
break. As a result, a buildd chroot for experimental will break a lot
more than an unstable one, and will therefore require a lot more work
from the buildd maintainer. A second problem is the infrastructure
problem: we're having problems keeping up with unstable as it is from
time to time already, for some architectures; adding experimental to
that will result in a situation where we cannot build everything on some
architectures, until our capacity is increased. Finally there's the
issue that experimental is incomplete, and I'm not entirely sure whether
wanna-build can cope with having to get information from multiple
quinn-diff runs. Not sure about that one, though; but if it's
impossible, that's a showstopper, since that would make dep-waits on a
build-dependency in unstable impossible, resulting in a lot of wasted
effort.
If the wanna-build problem turns out to be not an issue, I personally
would not mind running an experimental buildd, but only on one
condition: that it is clear to everyone that experimental will be
autobuilt on a "best-effort" basis, and that packages in experimental
might not get built, ever. Unstable will remain our primary focus.
--
Wouter Verhelst
Debian GNU/Linux -- http://www.debian.org
Nederlandstalige Linux-documentatie -- http://nl.linux.org
"Stop breathing down my neck." "My breathing is merely a simulation."
"So is my neck, stop it anyway!"
-- Voyager's EMH versus the Prometheus' EMH, stardate 51462.
|