From: marsh@extendedsubset.com   
      
   On 12/11/2011 02:51 AM, Pete Becker wrote:   
   > On 2011-12-10 13:56:41 +0000, Marsh Ray said:   
   >> On 12/09/2011 11:28 AM, Peter C. Chapin wrote:   
   >>> the author of class Foo is free to change or remove Foo::f at their   
   >>> whim. Such a change should not be a breaking change since Foo::f is   
   >>> private. Yet if class Bar could name Foo::f that code would fail to   
   >>> compile after Foo::f is removed.   
   >>   
   >> Perhaps it would make sense for some future standard to downgrade   
   >> unmatched specifiers in friend declarations to a diagnostic message?   
   >   
   > Please note that the standard requires a diagnostic message for code   
   > that violates the language rules (roughly speaking). Nothing more.   
   > There's no way to downgrade anything to a diagnostic message, since   
   > that's what's currently required.   
      
   Forgive me if I didn't use the correct wording. I was attempting to   
   suggest that the language explicitly permit such code (where possible)   
   and "encourage implementations to issue a warning".   
      
   - Marsh   
      
      
   --   
         [ See http://www.gotw.ca/resources/clcm.htm for info about ]   
         [ comp.lang.c++.moderated.    First time posters: Do this! ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)