On Sep 11, 2012, at 3:04 PM, jahanian <[email protected]> wrote: >>> -(void) Name1:(id) Name2:(id)Arg2; // expected-warning {{no parameter name >>> in the middle of a selector may result in incomplete selector name}} \ >>> + // expected-warning {{selector has only >>> bare colons in its name}} \ >>> // expected-note {{did you mean Name1:: >>> as the selector name}} >> >> Fariborz, >> >> What do you think is better? The original warning, or the new warning? >> Both of them really aren't that great, and don't really tell us what the >> problem is. The new warning is very clinical, but doesn't actually say why >> this is bad? >> >> Also, the suggestion in the note is completely wrong. It's likely the case >> that they meant to do "Name1:Name2:", not "Name1::". It's fine if the the >> warning itself is very clinical as long as the note itself provides a useful >> suggestion (which it doesn't). > > Yes, note is wrong. It wasn't meant for the new warning. >
I see. I'd expect to see a note *like this* for the new warning, just with the correct suggestion. Otherwise the new warning is fairly useless.
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
