Re: [opensuse-kde] Bug strategy
Am Donnerstag, 17. Februar 2011, 21:41:38 schrieb Stephan Kulow:
Am Donnerstag, 17. Februar 2011, 21:36:00 schrieb Karsten König:
openSUSE usually does not take upstreams point updates after release anyways, there was an exception for KDE4.3 where an SLE service pack was going to ship with a more recent version, I actually liked how that turned out, maybe Bille can comment a bit on the extra work this caused, if possible.
Especially for 4.6.1, there's such a large number of fixes in, that basically 11.4 is _already_ outdated if we don't ship 4.6.1.
Sorry, 11.4's roadmap was known for half a year now.
Fair enough, we had our chance. If we find bugs we deem important on bko and can trigger them on opensuse will you accept such patches or not? We will document the upstream bugreport and clarify that we could reproduce on opensuse so I hope this fits your requirement of a well documentated fix. Cheers, Karsten PS: I hope you don't mind me cc'ing that to opensuse-kde, so just everyone knows what's what. -- To unsubscribe, e-mail: opensuse-kde+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kde+help@opensuse.org
Am Donnerstag, 17. Februar 2011, 22:44:09 schrieb Karsten König:
Sorry, 11.4's roadmap was known for half a year now.
Fair enough, we had our chance.
Did we? Was there ever a oS release with less Novell staff dedicated to KDE?
If we find bugs we deem important on bko and can trigger them on opensuse will you accept such patches or not? We will document the upstream bugreport and clarify that we could reproduce on opensuse so I hope this fits your requirement of a well documentated fix.
This wastes a lot of time one could put into making sure that the KR46 repo is polished and provides tested KDE 4.6.1 packages for the 11.4 release. Given that resources are limited I think that one should just look at the final result and changing resources does also imply that "that's how it always was" changes as well. So if we stick to the current scheme we do: - find out whether bugs are opensuse-specific or not - search for bugs and report them at bko - wait for patches and report those back to bnc - backport them - test them regarding regressions (who of those active on IRC uses STABLE?) - document/beg for including the fixes into 11.4 after 11.4 this continues: - bugs that are fixed in KDE 4.6.x have to be found at bko - report at bnc, document, beg for official update - of course only major bugs will get those updates, so most smaller fixes will never see 11.4 -> if you want fixes, use KR46. - in fact, not having branch updates makes testing harder since you cannot tell upstream that a fix did not work. Simply because it might be that it was not included or the backport failed. Result: 11.4 is outdated the day it gets released but continues to suck-up time for backports. And backports are always only consuming downstream resources whereas regressions are worked on upstream as well. If one puts the same resources into polishing KR46 which is not restricted by some roadmap which does not care about KDE releases, we would get the following: - find out whether bugs are opensuse-specific or not - search for bugs and report them at bko + only report opensuse-specific bugs back to bnc + no backporting + test them (who of those active on IRC uses KDF or KR46?) after the 11.4 release this continues: + bugs that are fixed in KDE 4.6.x do not have to be backported + all upstream fixes will be included - regressions may appear and have to be reported and fixed which is done upstream or downstream. Result: more fixes, less double-reporting, less double-documenting, less backporting, less time to spend on begging for updates, more time to spend on opensuse-specific bugs, more time to spend on knm etc., easier testing and reporting against upstream, MAYBE some regressions on which one would have to see whether they are more severe than the bugs that will remain in 11.4 because of not updating to KDE 4.6.1, better tumbleweed packages. So IMHO the only thing that makes sense is to get all KDE 4.6.1 fixes plus _maybe_ some regressions and then spend resources on fixing those regressions rather than only get a minimum of fixes and waste time on work that was already done. I gave up reporting to bnc because Will is too busy and most fixes will not be allowed to get into the official updates anyway, so my time is better spent reporting and testing upstream, i.e. on KR46 after 11.4 got released. Sven -- To unsubscribe, e-mail: opensuse-kde+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kde+help@opensuse.org
Am Freitag, 18. Februar 2011, 00:12:23 schrieb Sven Burmeister:
I gave up reporting to bnc because Will is too busy and most fixes will not be allowed to get into the official updates anyway, so my time is better spent reporting and testing upstream, i.e. on KR46 after 11.4 got released.
Forgot one thing: I think we should spend less time on documenting/fixing/testing backports but rather spend more time on testing/polishing KRx-repos, upstream projects like knm and oS-specific bugs. Following that we should include a KRx repo for every release, i.e. oS 11.4 users should stick to KR46 even if KR47 is created and get all upstream fixes on KDE 4.6.x release day. KDF would still be the testing repo in-between those minor releases and if there are urgent fixes they can be put into KR46 right-away. So KDF becomes the new STABLE until a KRx repo is at x.y.4. There might be some details I have not thought about but the idea should be clear. Sven -- To unsubscribe, e-mail: opensuse-kde+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kde+help@opensuse.org
participants (2)
-
Karsten König
-
Sven Burmeister