Ruling out arbitrary cost copy construction?

[Andrei Alexandrescu]

On 10/29/10 13:19 CDT, Jonathan M Davis wrote:
> I'm arguing that you do _not_ define those and that you let user code fend for
> itself. If a programmer is foolish enough to write code with overly expensive
> postblit constructors, then they're just shooting themselves in the foot.

I think it would be hasty to classify them as foolish. Expensive copy 
constructors are meant to protect an abstraction, not some questionable 
practice. In brief your stance gets back to refcounting: a programmer 
must choose to use refcounting whenever they have expensive state to 
manipulate.

Andrei