Any hope of KDE 3.5 in F10? I want it too !

Kevin Kofler kevin.kofler at chello.at
Sat Jun 21 00:33:44 UTC 2008


Rex Dieter <rdieter <at> math.unl.edu> writes:
> Time commitment would likely be proportional to your expections of quality
> and continuing support.  To do well, and if *I* were doing it, my estimations
> would be at least ~6 hours up front, including pkg builds, hosting,
> deployment, plus who know how much extra time/effort for onging support and
> bugfixes.

I think that's very optimistic. I'd expect the successful builds alone to take 
at least 2 days; adding all the specfile editing and fixing, that would be at 
least a week of work. The edits are nontrivial, for example one has to bump the 
Epoch (or change the Name) everywhere so the packages don't get replaced by the 
KDE 4 version in the F9 repo, and then adjust the other packages to refer to 
the correct Name or Epoch. It's easier to backport a new version to an old 
distro (like you are doing for kde-redhat unstable) than to forward-port an old 
version to a new distro, due to how RPM works.

To do the work properly (and here we are at your "expectations of quality" 
point), one also has to track down the stuff which depends on KDE (not just 
kdelibs, but e.g. kdebase(-workspace)) and rebuild it for KDE 3, possibly 
undoing the changes done in F9 to make it use KDE 4. For example compiz. That's 
potentially weeks of work, also counting the time needed to track down all the 
stuff which has to be reverted. If that is not done, then some packages (like 
compiz) will not be usable with those KDE 3 packages.

All in all it would be a colossal effort, probably ending up with an 
unsupportable mix of Fedora 8 and 9, and with a target audience much less 
tolerant of problems than the clientele of kde-redhat unstable (because 
problems are what drives people to ask for KDE 3 in the first place).

        Kevin Kofler




More information about the fedora-list mailing list