Hi,
acme_pjz {l Wrote}:mclarge: Maybe you disabled C++ exception handling when compile STK? In Visual Studio in Windows, C++ exception handling can be diasbled ...
unlikely, since I've asked mclarge to do a simple exception handling test just at the start of main, and just before the crash happens, and that worked fine. Also I am 'reasonable' sure that other exceptions are being used in stk - before and after, and they don't appear to cause a problem. valgrind doesn't show a problem (some issues in the graphics drivers, but that's more or less normal).
I admit that I have no good idea. One thing I am not sure about is that mclarge's system has two graphics card - one internal and an ATM (iirc) which is used on demand. Can anyone with a similar issue confirm that it works for them?
It might be a mix up of some system libraries perhaps (afaik libgcc can be around as static or shared library, which can cause problems when catching exceptions from one lib in another - then again in this case it's actually all within STK, so it shouldn't matter).
The only hope I see is either to find a patient support person on a Mac forum, or to first significantly reduce the test case (STK): easy things to test: remove karts and tracks (but of course let one in that raises the exception); remove music, sfx, 3d models, ... - then it gets hard: remove irrlicht, and use irrxml to do the xml reading, and see if you can still reproduce the problem (if not, go to the irrlicht forum

). Then remove opengl (if the problem goes away, ask apple/update graphics drivers, ...). If this works, try to remove irrxml, ... etc.
Now my impression is that mclarge does not have the skills to do that (since it requires a lof of stk and irrlicht knowledge), and I certainly can't do it either (way too much work, even if I had a test case). So I am running out of ideas ... except for finding a patient Mac insider

Cheers,
Joerg