------- Additional Comments From giovannibajo at libero dot it 2004-10-11 09:28 ------- (In reply to comment #4)
> I disagree with the notion that it is just a diagnostic related issue; > because it comes with a semantics part. Let's not not disguise a > language extension under the name of diagnostic improvement. > It helps nobody. I am not trying to, really. In my view, the compilation unit will produce exactly the same object file, whether attribute(reimpl) is implemented or not, used or not. The only difference is that with attribute(reimpl) we might emit a warning iff the method does not override another virtual method in the base class. Why do you think there is a semantics part? Maybe I do not get it. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17920
