[Bug c++/18470] [4.0 regression] array bound rejected as non-constant in template

["mark at codesourcery dot com" <gcc-bugzilla at gcc dot gnu dot org>]

------- Additional Comments From mark at codesourcery dot com  2004-11-24 17:38 -------
Subject: Re:  [4.0 regression] array bound rejected as non-constant
 in template

giovannibajo at libero dot it wrote:
> ------- Additional Comments From giovannibajo at libero dot it  2004-11-24 10:05 -------
> Yes, Mark, but it used to work just a few weeks ago, and it is a rejects-valid.

I understand.  That's why I left the target for this PR set at 4.0.

We have an ever-growing list of these, with respect to using-declarations.

None of these have ever worked correctly; they've all worked because we 
made various mistakes in name lookup.  Every time we fix name lookup, we 
introduce more problems with using-declarations, because our 
using-declarations are not really using-declarations, but just access 
specifiers.  For example, we broke things in this area when we 
introduced two-phase name lookup.

The correct fix -- and, I suspect, the only correct fix -- is to 
implement using-declarations for classes correctly.  I'd love to do 
that, but depending on how big of a job it is, it may not happen for 4.0.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18470