KDE status

[Marc Espie <espie@nerim.net>]

I'm currently importing most of kde3 (probably all of it) into ports/.

You may consider the kde3 ports of beta quality at this point. They have
seen very little testing on OpenBSD (partly because of me not having
the time to test them around RC1) -> stick with kde 2.2.2 if you want
stability.

There should be snapshots available in a few days, time permitting.

Note that kde3 is insanely expensive to build with the current tools, on
i386 at least, because of a pending issue with our linker: you need 600MB
to build kdelibs3, I'm not kidding ! 
I have some tweaks to ld that ought to improve the situation drastically, 
but they need some careful testing, so they won't be in OpenBSD 3.1.

kde3 should bring a few massive improvements over kde2:
- much better font support,
- correct asian language support (thanks to qt3 being much better at coping
with environments with incomplete locale support)
- improved speed (very noticeable in konqueror).

One interesting improvement that applies to kde 2.2.2 as well is the plist
generation. It's been much improved, and kde packages should no longer
yield errors in @dirrm commands.

The ports still have no explicit maintainer, intentionally so.  Because
of the size, I don't consider it realistic to have one single maintainer
for them.  They need testing, and fixing.

Here are some known issues:
- port patches that were done for kde2 to kde3. This includes
making ksysguardd compile, letting konsole set tty permissions correctly,
- find out why kdm config initial creation crashes at exit,
- explain sudo to kdesu,
- fix kdeprint paths and make it work,
- explain mandoc to kioslave man,
- provide OpenBSD manpages in src format in the distribution, so that
kioslave man knows about them,
- explain that we have two info dirs to kioslave info,
- make audiocd: mp3 work again,
- fix kpat background rendering,
- figure how to compile koffice 1.1.1 with qt3,
- where are all the missing manuals (no manual for konqueror, for
instance),
- improve the plist generation tools some more, to be able to resurrect
subpackages in kde3, as this is insane to do manually,
- fix big font support in XFree (this is an xshm issue, not a kde issue),
- investigate null socket security issues,
- find out why java applets show up in separate windows,
- find out if --disable-dynamic-ssl works, and fix it if it doesn't...


All in all, considering the sheer size of kde3, I'm suitably impressed
with the quality of the work done by the kde developers.  It's reasonably
clean from annoying linuxisms for a project of this size.