From: Douglas Gregor (gregod_at_[hidden])
Date: 2002-05-13 12:16:25
On Monday 13 May 2002 12:52 pm, you wrote:
> > I hadn't realized that bind didn't unwrap ref()/cref() for its first
> > argument, and I'd really like to have this. (Hint, hint ;->)
> Two options: special-case (again) ref(x) in bind, or make ref(f)(x, y)
> invoke f(x, y), allowing ref() to be used with standard algorithms and
> stateful function objects, as I originally envisioned.
> What do you think? Will this break existing ref() uses?
I prefer the latter, because it is useful in many more applications (and also
makes signals usable with standard algorithms).
I can't think of any existing uses of ref() that would break, but the change
probably shouldn't go into 1.28.0 to be on the safe side.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk