Thanks for the info Tobias. Good to know the issue is noted. In our case we'll just stick with the (working) sjlj compilers. We're more interested in working exception handling functionality than a better performing implementation thereof (although we'll gladly move to seh when ready). Martin. On 03/26/2013 05:06 PM, Tobias Burnus wrote: > Hi, > > Martin Corino wrote: >> What we found is that with the win32-seh versions of the compiler we >> were not able to catch an exception thrown from within one of the >> framework DLLs in the application code using that DLL. I.e. the >> following did not work: > > That's a known problem - but, unfortunately, no solution has been found, > yet. > > See also the following problem report > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56742 > > (If you wonder why you didn't see that PR before: Kai has just filled > it, motivated by your email, to make tracking of the issue simpler.) > > At least for the test case in the PR, the program works when compiled > with -O1 and only fails with higher optimization levels. I don't know > whether that would be an option for you. > > Tobias >
