Patch to detect invalid and missing ATTRIBUTE const/pure [take 2]

[Kaveh R. Ghazi]

 > From: John Wehle <john@feith.com>
 > 
 > >> 2.  If a function is marked pure, and gcc detects the function is
 > >>     really const, should we proceed to apply the stricter attribute?
 > >
 > >I think so. It is same as marking unmarked function as pure :)
 > 
 > I disagree.  The programmer is explicitly stating a design decision
 > by marking a function as pure.  It's fair to suggestion in a "warning"
 > message that the function could be marked const, however I'm not wild
 > about the compiler overriding the programmer's decision on its own.
 > -- John

I don't feel strongly about it so I'll go with your recommendation in
my next patch.

--
Kaveh R. Ghazi			Director of Systems Architecture
ghazi@caip.rutgers.edu		Qwest Global Services