|
Boost : |
From: Jonathan Turkanis (technews_at_[hidden])
Date: 2004-10-13 11:08:14
"Howard Hinnant" <hinnant_at_[hidden]> wrote in message
news:197E7641-1CA8-11D9-92A6-003065D18932_at_twcny.rr.com...
> On Oct 12, 2004, at 6:56 PM, Jonathan Turkanis wrote:
>
> >> Buy Metrowerks Pro 9 and you will see it. :-)
> >
> > For some reason I thought it used parts of the Boost code base. I
> > gather I was
> > wrong?
>
> I implemented it from the TR1 spec.
Don't get me wrong -- I knew you could do it yourself ;-)
> Peter Dimov was a tremendous help
> though (jeeze that guy is smart! :-) ). And as Peter recently pointed
> out:
>
> > But legalese aside, the most important part of the implementation are
> > the tests, anyway. ;-)
>
> I'd agree 100%, but without the smiley. The shared_ptr/weak_ptr spec
> implies some of the most subtle code I've ever tried to implement.
I'm surprised -- but I probably shouldn't be. To be honest, I haven't used weak
pointers since I programmed in Java. Each time I've thought I needed weak_ptr
I've always ended up adopting some other solution. But now you have me very
intrigued.
> -Howard
Jonathan
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk