Am Samstag, 1. Oktober 2011, 12:34:56 schrieb Martin Gräßlin: > There is ZERO communication with downstream about bugs. We do (at least for > KWin) get no feedback at all about important bugs for the distro and we > give no feedback to the distros. I hardly see distro people in the > bugreports and hardly any user mentions that they reported in distro > bugtracker.
What I know from e.g. the openSUSE tracker is that a lot of upstream bugs are reported. If the bug is not openSUSE specific people are told to report upstream and get back if a patch is available. That patch can then be included in the distro's update. For kwin there was e.g. llunak working for openSUSE (KDE), however he moved to libreoffice I think. So now you basically have 1-2 guys who do packaging, backporting and bug fixing. Dividing the amount of KDE components by the number of people shows how much time they would have for upstream bugs and of course they don't know the code of all KDE components, so them actually doing bug fixes is rare but they are open for including/backporting patches nevertheless. And if you wish to have a different/any communication with downstream this is certainly something one can work on. > What you describe would be an ideal solution, but the reality is that > distros don't care about bugs. Also I cannot remember that distros would > have asked to get their patches reviewes. It's hard for them to produce patches, though they do, if you have a look at who is working for them you will see that they are working on different bits within KDE. And the reason they do not seem to care is the same you do not seem to care when managing bugs, i.e. lack of time and quality of communication to the other side. Anyway, IMHO distros waste time by backporting and patching KDE instead of just shipping minor releases as bug fix releases. BUT they always point to the regression trap and thus this will not change until regressions are not an issue anymore, i.e. if they appear they are guaranteed to be fixed. IMHO this would be a win-win situation because it is easier for upstream to find and fix them and they save time on bug reports, i.e. all those old version bug reports. Downstream saves time as well and can contribute more to whatever KDE component they are working on upstream. > > You want more testers to avoid regressions since whoever is currently > > working on branch, i.e. unreleased code, does not do so well enough. > > Regressions happens - that has nothing to do with bad quality of work. True, I forgot the "test" they do not test well enough, e.g. because they are working with a different git checkout. > regressions have to be in the bugtracker and if I spot one it has highest > priority. In KWin we follow a process of backporting non-trivial fixes only > after two weeks in master. If this would be written down for all of KDE's main components and could be counted on users could persuade their distros towards spending time on releasing minor updates officially rather than selecting and backporting fixes. Sven >> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
