Is a warning for a==a possible?

[Joe Buck]

On Thu, Dec 04, 2003 at 05:51:33AM -0800, George Garvey wrote:
> Just made a stupid mistake, and wondered if GCC could have warned about it?
> 
> 	struct S {
> 		int A;
> 
> 		bool f(const int a) {
> 			const bool r(A == A);
> 			a = 3;
> 		};
> 	};
> 
> The first line in f() should have been r = A == a. No doubt there is a
> valid reason to compare something to itself that I can't see right now.

You're forgetting macros and machine-generated code, which might produce
something like A == A.