I've now logged in again. Instead of going first to firefox, I tried to deal with Kmail, and virtuoso-t went to the top of my CPU usage at over 80%. Is that normal?

Just a tip for a possible workaround as you are using openSUSE-12.3:
The still maintained kde3-repository of opensuse offers the nice feature to be fully mergeable with
kde4, e.g. you can keep your full featured kde4-desktop and use kmail from the kde3-repository
(without akonadi/nepomuk/kwallet etc.)
Similar to you - after upgrading to suse 12.3 - I spent days in the effort to get back a working
kmail and despite the help I got from vaious forums I didn't succeed. After switching back
to the former kmail all works perfect again
Hence, not a fix, but a pragmatical solution to upgrade and still keep a working kmail.

P.S.: I'm sure the problems I encountered has something to do with my backup-history.
But its certainly not the maildir-migration, because I had a pure-maildir beforehand.
So I still suspect akonadi/nepomuk and/or their kmail-integration as not optimal.
However, my major problem was that because of the super-extra-added-complexity
of those tools I have no glue any more of what is happening under the hood.
Therefore error-tracking is a pain (more a random trial&error than a detection).
In the end I simply gave up, but still hope to catch up again on a later kde.

Many thanks for the suggestion but I've looked in http://download.opensuse.org/repositori ... SUSE_12.3/ without success. Am I going blind or looking in the wrong place?

Try http://en.opensuse.org/KDE_repositories#KDE_3.5.10

Also, the following is my current list of repositories (plus priorities for a basic kde4.10 with optional kde3-apps):

Code: Select all

| Prio | Name                                | URI
+------+-------------------------------------+-----------------------------------------------------------------------------
|  70  | Packman Repository                  | http://ftp.gwdg.de/pub/linux/packman/suse/openSUSE_12.3/
|  75  | libdvdcss repository                | http://opensuse-guide.org/repo/12.3/
|  79  | KDE_410                             | http://download.opensuse.org/repositories/KDE:/Release:/410/openSUSE_12.3/
|  81  | openSUSE BuildService - KDE:KDE3    | http://download.opensuse.org/repositories/KDE:/KDE3/openSUSE_12.3/
|  85  | openSUSE BuildService - LibreOffice | http://download.opensuse.org/repositories/LibreOffice:/Stable/openSUSE_12.3/
|  90  | Main Update Repository              | http://download.opensuse.org/update/12.3/
|  91  | Update Repository (DEBUG)           | http://download.opensuse.org/debug/update/12.3/
|  92  | openSUSE-12.3-Update-Non-Oss        | http://download.opensuse.org/update/12.3-non-oss/
|  98  | nVidia Graphics Drivers             | http://download.nvidia.com/opensuse/12.3/
|  99  | Main Repository (NON-OSS)           | http://download.opensuse.org/distribution/12.3/repo/non-oss/
|  99  | Main Repository (OSS)               | http://download.opensuse.org/distribution/12.3/repo/oss/
|  99  | Main Repository (DEBUG)             | http://download.opensuse.org/debug/distribution/12.3/repo/oss/


For preferring kde3's kmail as default I just deactivated the kmail-package from kde-410 and activated kdepim3
(from kde3, plus the minimum required kde3-packages, resolved by yast/zypper).
Of course, you can also leave both installed, but then you have to use the absolute path to /opt/kde3/bin/kmail
because in case of duplicates your search-path will find kde4-apps before kde3 (which is what I want).

Both your reference to the openSUSE Wiki and your own current repositories seem to confirm that I was looking in the right place, but I still can't find version 3 of Kmail. Have you any idea where I'm going wrong?

You already did. I should have been looking for kdepim instead of Kmail. The possibility of having two versions of the other PIM components conflicting makes me nervous, but you seem to be telling me that it isn't a problem, thanks.

If you want to discuss parallel installation of KDE PIM 3.5.x and 4.x, please open a separate thread, so discussion can continue (if needed) focusing on addressing the issue.

As I see it, the only need is for me to confirm that it works. After six weeks of wait . . .ing in vain, I felt the need to press on with a solution that was known, at least by marcaurel, to be effective. Unfortunately, that means that I didn't take the time to experiment with a new user account, so can't report any results.