On 10/11/2010 01:44 PM, Dotan Cohen wrote:
On Mon, Oct 11, 2010 at 20:37, Rex Dieter
> On 10/10/2010 03:43 PM, Dotan Cohen wrote:
>> Hi all, I haven't used Fedora since FC6, but I want to come back. I
>> have been looking for the "best" way to test KDE developments, and I
>> found he kde-fedora repo. Is it possible to install the
>> latest/greatest KDE from the repo alongside the stable released
>> version, and choose between them from KDM as one would choose between
> Currently our kde-unstable repo can track whatever we put into rawhide,
> and that can sometimes include new alpha/beta/rc builds. If we had a
> bit more manpower (and interest), we could also consider doing builds
> based on weekly upstream snapshot tarballs.
Thanks, Rex. What would that take? I might be the interested party,
but I certainly don't have the skillset currently. I'd love to learn
it, though, and if it helps the community then great! I;ve googled for
how to maintain a package but it looks like a real pain, and KDE
especially so. Is there a specific fine manual that I should be
reading? Note that I've never built KDE from source.
To manage packages in fedora, you'd need to become familiar with the
tools used, namely git (scm), koji (build system), bodhi (updates), and
fedpkg (a friendly front end to each of the prior tools).
And, then, become familiar with cmake and kde's build system and
structure, dependencies, applying and adapting patches to/from upstream.
Yeah, there's a bit of a learning curve there for each of these items,
but in the end, is well-worth the effort ... esp if you have a long-term
interest and commitment to make to fedora and/or kde.
> However, these replace what's currently installed though, and
> parallel-installable with stable/released versions.
If I were to build the package, what would need to be changed to allow
it to be a parallel install, much as KDE and Gnome are parallel
installs? Would changing the active ~/.kde directory to ~/.kde-weekly
be enough? Should I ask about this on the KDE list? What should I be
It would take a *lot* of effort to get right. IMO, not worth it really.