The old SEH / gcc question came up again on the gcc mailing list; see http://gcc.gnu.org/ml/gcc/2005-05/msg00511.html and followups. If anyone has anything constructive to add to that thread, this is a good chance to get the story right before it shows up on ZDNet.