https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61587
--- Comment #2 from Marc van Leeuwen <Marc.van-Leeuwen at math dot univ-poitiers.fr> --- But the error in this program is so much not with the implicitly declared copy constructor of A itself (I suppose it is "A::A (const &A)" that the compiler is trying to generate here) but with the fact that the code is (accidentally) _using_ that copy constructor. That usage in done in the indicated line 18, and without that line the program compiles correctly. I agree that it might have been better to explicitly mark that implicitly-declared copy constructor as deleted in the class definition of A, but one does not _have to_ do that, as long as such a copy constructor is not being used anywhere. So the compiler must have been reading at least up to line 18 before it decided to emit this error message; my complaint is just that _nothing_ in the list of error message gies any indication of that. I would think C++ compilation is basically a single-pass process, and that it would be rather easy for a compiler to start (or end) each error message with an indication of exactly where it is during this pass; in my experience with gcc such an indication sometimes comes in the middle of a long sequence of messages, which is fine with me too, but in this particular case such indication seems to be entirely absent. In practice such illegal copy-constructing of an A object could happen in code very far from the class definition; how is a user going to trace the place where she did so?
