Peter Eisentraut <[EMAIL PROTECTED]> writes: > Bug #218971 describes some unclear interaction between licq (which uses > Qt as user interface), Qt, and the kernel.
I can't reproduce that bug, FWIW. > I don't understand a word of what it says there. The previous > maintainer of licq solved this bug by providing a wrapper script that > does > > LD_PRELOAD=/usr/lib/libqt-mt.so.3 /usr/bin/licq.real > > But it now turns out that this causes a set of other problems (for > example, the KDE session management cannot know the right program to > restart), so I'd prefer to work on getting rid of this wrapper. Can > any of the Qt experts here comment on this? The bug says that it is > caused by improper dlopen handling (licq loads a pluging at run time > that is linked with libqt). What is the proper handling? So it mixes libqt-mt and libqt? That's bad. If that's the case, it would be a licq bug--nothing in Debian should be linked with libqt. > Here is a Red Hat bug that seems related: > https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=102584 > > They talk about installing an updated kernel. What about the Debian > kernels? No idea what that stuff is about. In any case, it's not a Qt problem AFAICT. -- You win again, gravity!