> An external symbol in a COMDAT should be identical in all COMDAT groups that > it could be in. If you can't find issue with it, I'll add it to the > langref.
I don't think that is correct. The group itself has to be equivalent in some sense, but each symbol in it can be different. The example example of the variable + guard in the same group is probably a good one. Cheers, Rafael _______________________________________________ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits