[PlanetCCRMA] RE: [Rosegarden-user] RG make issue:undefined symbol: __ti15QGPluginManager

[Fernando Pablo Lopez-Lezcano]

>    Here's the problem. I have 3 machines going right now. All of them
> are RH 7.3, and all use Qt via the distribution. (as per your info.)
> 
> If you execute the command:
> 
> /usr/lib/qt3/bin/uic
> 
> on two of the machines this command executes correctly. On the 3rd
> machine the command fails with the error message in the title of this
> email.(undefined symbol....) This same error happens during the make
> step of RG on that machine.
> 
>    I have not been able to find what I might have done to make this one
> machine work incorrectly. That major differences are:
> 
> 1) The one that fails is running fluxbox. The two that pass are running
> KDE. I will change the failing machine back to Gnome/KDE tomorrow and
> see what happens.

I would try two things. First I would see if there are differences in
the rpms that are installed in both machines. You can do a "rpm -q -a
>somefile" in one machine, do the same on the other ("rpm -q -a
>somefile2") and then compare the two files as in "diff -uNr somefile
somefile2". Just to see if there is something obviously missing in the
computer that is not behaving. 

I would also suspect something you installed manually (fluxbox?)
overwrote a file or library in the qt3 installation. You could check the
integrity of the installed files by doing "rpm -V somepackage" and
replacing somepackage with the name of the package to check. You can get
a list of the qt 3 packages by saying "rpm -q -a | grep qt | grep 3.0"

> 2) The machine that is failing has audio hardware in it. The two that
> are passing do not. (Yet...)
> 
>    All three have your kernel. Only the failing machine has been
> configured as per the audio config step on the Planet.
> 
>    Others that have responded to this thread have said the error looks
> like a problem that occurs when the Qt RPM is compiled with a different
> C compiler than is installed on the machine using the qt3 library. This
> is why I asked if you knew what compiler was used. From your info I
> infer that it was the compiler that came with the distribution. (And
> again, if it was the wrong compiler, then I think that all of my
> machines would fail identically.)

I agree. Let's see what happens if you try the aforementioned tests...
-- Fernando