On 9/3/06, IOhannes m zmoelnig <zmoelnig@iem.at> wrote:

i think i'll revert the changes.

i introduced them in trying to solve a crash on debian systems whenever
an exception was thrown.
i tried to make GemExceptions as "standard" as possible.
it turned out to be a problem of the used stdlibc++ with the
nvidia-drivers, so my changes didn't help at all.

i thought that if they don't help, they won't hurt; obviously i was not
correct here.

Apple's GCC 4 now throws a warning because the functions are virtual but the destructor is not.  I guess you can't really please everyone.