[C++] PATCH c++/17542

Gabriel Dos Reis gdr@cs.tamu.edu
Fri Oct 29 17:03:00 GMT 2004


"Giovanni Bajo" <giovannibajo@libero.it> writes:

| Matt Austern wrote:
| 
| > +   if (declspecs->attributes)
| > +     {
| > +       cp_warning_at ("attribute ignored in declaration of %q#T", t);
| > +       cp_warning_at ("attribute for %q#T must follow the class key",
| > t);
| > +     }
| 
| I would think that 'class key' is a legal term which is not common in C++
| development. Probably saying "enum/class/struct keyword" is a little easier to
| follow.

I disagree.  If we really want to be anal precise about it, we should
be displaying the real class-key written by user.

| Also, I reckon the second line should be an inform() but do we have
| cp_inform_at?

No, we don't.  I'm not sure we want to increate the number of xxx_at
functions.  I was under the impression that Zack wanted to come up
with something different.  Is my memory failing?

-- Gaby



More information about the Gcc-patches mailing list