Adeodato Simó wrote:
* Riku Voipio [Mon, 22 Nov 2004 13:05:00 +0200]:On Mon, Nov 22, 2004 at 03:30:03PM +1000, Anthony Towns wrote:
Yay, confusing quote styles.
I agree. In this way, the kde packages have allways been broken. We have been so far reluctant to enforce all of kde being the same version, since it would seem make testing maigration even harder
The point isn't to try making all of KDE be the exact same version, it's to ensure all the different versions of KDE you've got installed at the same time be compatible.
That's tricky and difficult. If the solution you've come up seems easy, it's almost certainly wrong.
Worse, there are different levels of incompatability: there's the "doesn't work at all" variety, and there's the "looks ugly and substandard" variety.
That is the cleanest idea so far. But we would still need to add something like: Package: kdebase Conflicts: kfoo ( << 3.3.0 ), kfoo2 ( << ... Provides: kde-api-13
No, you specifically don't need those Conflicts: lines. You might need a couple of different "kde-api-N" virtual packages though; eg "Provides: kde-foo-api-12, kde-bar-api-9". Anything that depends on "kde-bar-api-8" or "kde-foo-api-10" will not be installable with that version of kdebase.
But before we even _think_ about that, we need to be absolutely clear on what combinations of packages are "acceptable". What problems happens with which sorts of mismatches? What do we think about those problems, and when do we want to prevent them?
Since in sid (and sarge) kde packages do not depend on a kde-api virtualpackage. Adding lots of versioned Conflicts was something that was strongly discouraged by apt maintainers[1] and the policy...
As long as it's not a regression, it's not a major problem -- KDE upgrades are already broken, if that's solved for the future, and not made worse right now, that's still a win.
but this is going (I guess) off-topic for -release, though I'd be very pleased to have aj or any RM help us with the issue!
I don't think this is majorly off topic for -release, though Colin or Steve might want to jump in.
Anthony, I didn't mean to blame the testing scripts at all.
*shrug* The question is where the fix needs to go; it's easy to assume that it needs to go into the testing scripts if the problem only happens when they run. But the underlying issue is the dependencies in the packages.
If you and the Release maintainers believe, that this in issue that must be solved for kde3.3 to enter sarge
I'm pleased to say it doesn't matter what I believe :) That said, I just think it has to be solved, whether before or after kde3.3 and/or sarge doesn't really bother me one way or another. But the sooner it's solved, the sooner you'll be able to consistently get simple, speedy and positive answers when you ask about uploading KDE N.M to unstable (and then testing) even though the release is only k days away.
I'm pretty sure I did mention that I didn't think this was something that'd make sarge in my previous mail?
Cheers, aj
Attachment:
signature.asc
Description: OpenPGP digital signature