[cpp-threads] Re: Weak vs Strong Atomics

[Herb Sutter]

> I meant something different.  You concern, as I understand it, is that
> you don't want people to accidently use the weak operations.  If the
> weak operations were syntactically distinct and greppable, so that one
> had simple rules and static tools to enforce those rules, would you
> be satisfied?

I want an easy way to tell people in a short sound bite "don't even think about reaching into 'feature bag X.'" I would be happy if I can make 'feature bag X' be "named member functions of weak_atomic<T>" or "named nonmember functions on atomic<T>" (preferably in a subnamespace).

So the answer is it depends on the proposal? If I have to give a slide/page-long list of functions not to use, that's much less helpful. What would the proposal look like? (Sorry, I've been swamped in the deluge and haven't been able to track the discussions on this and other topics in detail.)

Herb