call for votes: KDE4 uploads to universe

[Luka Renko]

On Wednesday 14. of March 2007 05:09:34 Jordan Mantha wrote:
> Sarah Hobbs wrote:
> > The MOTU is not expected to know how KDE, or XFCE, or even gnome works -
> > it's desktop agnostic, and that's the way it should be.  So, why is it
> > making decisions on kde4 packages anyway, which is an approved spec?

Sarah, I am sure that every MOTU cannot know the details about KDE, but at the 
end it is MOTU who has general responsibility for universe. This is why we 
have to discuss it, as it is not a simple decision and it may also set 
grounds for future decisions in this direction.

> First, I disagree with Sarah that the issue is for the Kubuntu Council
> to decide. I really respect the KC but the packages in question are to
> be in Universe so it is the MOTU's domain. If the KC wants to make the
> ultimate decision they should put the packages in Main. That said, I 

I do not see that Kubuntu Council can make such decision for main either. ;-) 
Kubuntu Coucnil has only discussed this point and has got to proposal that it 
has put to the MOTU.

> don't particularly see this as something the MOTU Council should be
> deciding either. I have to say I'm a little disturbed that the MC has
> been making decisions that I really feel MOTU should be making (this and
> SRU policy are the particular ones I'm thinking of). I thought the MC
> was supposed to approve members and MOTUs and provide some leadership,
> not take on every decision that comes up. I think it's both unfair to
> put the burden on the MC and I think it's also demotivating to MOTUs at
> a time when we are trying to excite and build our community.

I think this is where we need a bit better defintion of what is charter of 
such councils. I think that work on MOTU Coincil charter already started and 
I think similar would be benefitial for other Councils (Kubuntu's 
included ;-)).

> Back to the issue at hand, I see 3 issues:
>
> 1) Supportability.  We are trying to do better with Universe QA/QC and
> it would be somewhat of a bad example to have KDE (I'm not sure everyone
> will know that it's development only) packages that we have no intention
> of supporting. Jonathan has said that they don't plan on supporting the
> packages so it'll be up to the MOTU. I also think these package will be
> inherently difficult to support after release. We normally do not
> include new upstream versions in -updates so I see them as basically
> unsupportable in a stable release beyond basic packaging issues.

I think Jonathan statement was a bit strong: kde4 snapshots are technology 
preview code, they do not have support by upstream in terms of backporting 
fixes to particular snapshot and that is what Jonathan also do not plan to 
support. But we can support releasign new snapshots on regular basis. This is 
why Jonathan asked for general exception, so that we could include latest & 
greatest that is available at the time of feisty release.

> 2) Usefulness. I question the usefulness of the packages. I was trying
> to set up some KDE4 development stuff the other day for Kalzium and I
> was told that I shouldn't use the snapshot packages because they will be
> outdated so quickly. If this is the case and the packages will be
> virtually useless at or shortly after the release of Feisty then I see
> little benefit. In fact, it would seem more confusing to people because
> they won't know whether to get the Feisty packages or from some other
> source. If this is not the case, please chime in. I think this is most
> of the issue, we are seeing little gain for the work.

They are getting more and more useful as library freeze is getting near. Less 
and less things will change from then on.
Regarding outdating: I think that -backports is the right solution for this. 
We would continue to have latest snapshots in feisty+1 and they should be 
backportable to feisty withou major problems.

>
> 3) Timing. We a long past NewPackagesFreeze (Feb. 22nd) so this is an
> exception request. Given the items above I just don't see a lot of gain.

I agree with that, but this is why exceptions are for. ;-) 

> On one hand, we've already got a lot of crack in Universe and I'm
> guessing a few KDE4 packages wouldn't add a ton to it. On the other
> hand, the request seems mostly a workaround for Jonathan to get more
> archs. I think perhaps a better solution would be to get him some access
> to other archs. I think Brandon and Andrew are working getting our build
> farm more suitable for this sort of thing. I also wonder if Canonical
> would have anything for Jonathan to build on. Seems sort of odd he'd be
> stuck with i386 only.

I think what we are missing is something similar to debian experimental, where 
such preview versions could be centrally built, hosted and mirrored. Such 
repository could have more relax policies in terms of freezes and inclusion 
process. Maybe you are right and the build farm + some hosting can step in 
here.

> Overall, I think maybe we just needed better/earlier communication. This
> is not a KDE-specific issue, at least for me, and I hope the KC,
> Jonathan, and the Kubuntu MOTUs don't take it that way. In the end I
> don't have a strong opinion either way. I'd rather we just decide and
> move on.

I agree fully here and this kind of discussion would come sooner than later. 

Regards,
Luka