[RFC] Cleaning and extening the diagnostics machinery

[Bill Wendling]

Also sprach Joseph S. Myers:
} On 31 May 2001, Gabriel Dos Reis wrote:
} 
} >    - "Sorry, unimplemented: " for unimplemented fonctionalities;
} 
} I'd like two versions of this, one which acts as an error and one which
} acts as a warning.
} 
Shouldn't it always act as an error? I.e., the compiler can't interpret
it, so the code it generates won't be what the programmer wanted?

-- 
|| Bill Wendling			wendling@ncsa.uiuc.edu
|| Coding Simian
|| Jaj Istenem, de fejsze van a fejemben!! (Hungarian)