Hi, Sven Burmeister schrieb: > He is using the least risky way in order to be able to test new KDE versions > for the next SuSE release and hence helps fixing bugs and thus contributes to > the quality of a SuSE product, i.e. something those that do not test newer > versions of KDE profit from.
I'm not sure that using these packages really helps finding and fixing bugs for the next release. The best and maybe even the only way to do that is using factory. 10.1 + build service is not what will become the next release - that is factory and nothing else. There might be artifacts because the way the base system and the KDE packages are combined is different from what will become the next release. There might be even wrong bug reports because of that, which might actually consume additional time and effort instead of helping. > Further he is using a bugfix-release, hardly any bugfixes are backported and > supplied via YOU, so this is the only choice apart from more risky (betas and > alphas) or more inconvenient methods (compiling). Of course bugfixes will be supplied via YOU. Actually there was already a qt3 + kdebase3 bugfix YOU. The packages in the build service are not bugfix releases. Don't mix up the upstream development with the packaging work: From the stand point of upstream development, KDE 3.5.3 is clearly a bugfix release, but packaging software is much more than downloading it and making it compile. There is integration work to do, and a version upgrade to something that is published as bugfix release from upstream does not necesserily have to behave like a bugfix package because upstream development and packaging are very different things. > If Novell would think this is too risky for users, it would not call > it "genuine added value" but "genuine added risk". Added value == added risk. There is no contradiction here. Sometimes I seriously dislike the way people expect these extra packages to work. Some users expect them to have the same or even higher quality than the originally distributed ones because the difference between the version-release number and the work that has to be done to make packages behave as expected is not clear. And claiming that using these packages helps improving the quality of the next release is a very problematic thing. This might be true in many cases, but there might be exceptions, caused by artifacts and possibly invalid bug reports as described above. Sometimes I just don't understand why these extra packages are so popular among consumers. The packagers are spending huge amounts of time for doing integration work and bugfixes only, everyone who read opensuse-commit in the last months is able to know that, and then consumers just throw everything away because there's something available that looks like a bugfix update because it has a higher version-release number. By the way, splitting KDE itself and the most popular KDE applications into separate repositories is a great step forward compared to the old supplementary FTP. It can prevent some common problems and misunderstandings - if people appreciate and follow the split instead of quickly adding both repositories to their $PACKAGEMANAGER configuration just because they can... Andreas Hanke --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
