Mandag den 30. maj 2011 16:39:25 skrev Will Stephenson:
* 12.1 features: scheduling for stability (cb400f)
Sorry for the long post. I'd like to put on my "wannabe project manager" hat, and start this discussion early, before the IRC meeting. First let's have a look at some things that could have been handled better for 11.4 in my not so humble opinion. 11.4 EVALUATION FOR PACKAGERS/DEVELOPERS ========================================= 11.4 featured a lot of major changes, that could each potentially have a very significant impact on overall user experience, e.g.: * PulseAudio running by default * KUpdateApplet replaced by KPackageKit * Phonon-backend-gstreamer new default instead of Xine * Plasmoid-networkmanagement replacing KNM * Bluedevil replacing KBluetooth * Inclusion of rekonq * Oxygen-gtk replacing oxygen-molecule * Removing HAL from default install I think all of these changes were submitted within the last 3½ months of 11.4 development - with christmas holidays, new years and an overloaded legal review team slowing things up further, meaning a lot of the things were barely tested, with very little time to fix anything. And much of it ended up being shipped in a sub-optimal state - to put it mildly. Most of these changes could easily have been done several months earlier to ensure more testing, and a less hectic final couple of months of development (of course some couldn't have been done early, like e.g. HAL removal or oxygen-gtk inclusion). There were some special circumstances making the 11.4 process difficult. Lubos had left the team and his replacement was being held up by work permit issues. The schedule was very tight with 11.4 release and KDE SC 4.6.0 release being realtively close. There were unusually many, significant upstream developments around the 11.4 feature freeze (e.g. with phonon and oxygen-gtk). 11.4 EVALUATION FOR TESTERS ========================== This is one of the areas where I'm able to contribute and I'm particularly embarrassed that I didn't test the DVD install in the RC phase, and therefore didn't notice that the DVD installs KNetworkManager instead of plasmoid- networkmanagement which is a horrible mess and support nightmare. And a bug which could probably easily have been fixed. Another lesson learnt is that testing KDF is not always "enough". There have been a number of occurences where stuff is in KDF, but hasn't been submitted to openSUSE:Factory - like currently rekonq which is at 0.6.1 in o:F while KDF has had 0.7 for weeks and weeks. So we, testers, should also do better. 12.1 PLANNING ============= The good news is that 11.4 seems to be doing quite well, despite a number of pretty major oopses. So with a bit better planning all around 12.1 could be really great. For 12.1, it doesn't look like there'll be a need for many major changes (from KDE point of view), and the schedule wrt. to KDE SC 4.7.x is not as tight. But there are other challenges, like Will Stephenson leaving on parental leave all summer, and only coming back part-time in the fall. Though there might be a replacement. So for 12.1 we should try to add new packages/features as early as possible, for maximum testing, and so that we can focus on critical bugs towards the end of the cycle. E.g. it seems like a no-brainer to me that we would want to include kamoso in 12.1, so why wait until the eve of version/feature freeze to submit it? So it's time to plan what we can realistically do for 12.1, and when. I've updated the Ideas wiki page and cleaned it a bit: http://en.opensuse.org/openSUSE:KDE_ideas Other ideas for improving the working process? -- To unsubscribe, e-mail: opensuse-kde+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kde+help@opensuse.org