From: William Kempf (sirwillard_at_[hidden])
Date: 2001-01-09 09:50:18
--- In boost_at_[hidden], "David Abrahams" <abrahams_at_m...> wrote:
> Not to my knowledge. Why would you want something like that?
Doesn't a raw
> pointer work just fine?
The main difference is that raw pointers will lead to "dangling
pointer" problems that won't occur with "weak smart pointers". I've
never had to deal with this situation (I've avoided smart pointers in
cases with circular references) so I can't tell you how bad the raw
pointer approach is or how good the weak smart pointer approach is,
but the reasoning was well presented in the posted article.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk