RE: [boost] On ownership policies and boost::raw_ptr

From: Yitzhak Sapir (yitzhaks_at_[hidden])
Date: 2002-06-27 07:05:50

I've had problems with the #import<> generated files defining operator& for all their smart pointers.  These, in turn, release the smart pointer.  This causes problems with stl containers, at least in stlport, but I suppose it might be a more generic problem across other stl implementations.

Anyway, the point is, #import auto-generates operator& and you have no control over this.

> -----Original Message-----
> From: Peter Dimov [mailto:pdimov_at_[hidden]]
> Sent: Thursday, June 27, 2002 12:22 PM
> To: boost_at_[hidden]
> Subject: Re: [boost] On ownership policies and boost::raw_ptr
> 
> 
> From: "Joe Gottman" <jgottman_at_[hidden]>
> > >
> > > And I still like &*p better than p.get() or mumble_ptr(p).
> > >
> >
> >    I don't.  &*p has two disadvantages.  The major one is that it is
> > undefined if p.get() happens to be null.  Also, it fails if 
> the class that
> p
> > points to has operator& overloaded.
> 
> Programmers who overload unary operator& should be sentenced 
> to writing
> libraries that need to operate properly when fed such classes.
> 
> _______________________________________________
> Unsubscribe & other changes: 
> http://lists.boost.org/mailman/listinfo.cgi/boost
>