Boost logo

Boost :

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2001-09-14 13:59:27


----- Original Message -----
From: Peter Dimov <pdimov_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, September 14, 2001 1:09 PM
Subject: Re: [boost] More on the optional class.

> From: "Fernando Cacciola" <fcacciola_at_[hidden]>
> > From: Peter Dimov <pdimov_at_[hidden]>
> > > I apologize if this has already been discussed, but did you consider
the
> > > idiom (often used with dynamic_cast)
> > >
> > > optional<T> opt;
> > >
> > > if(T * pt = opt.get())
> > > {
> > > // initialized, use pt
> > > }
> > > else
> > > {
> > > // not initialized, no pt in scope
> > > }
> > >
> > This is already supported.
> > I've added .get() precisely in order to support this idiom, but I've
been
> > kind of reluctant to talk about it.
> > It usage is *barely* mentioned in the documentation included in the
> header.
> > I'm not sure if I would encourage this idiom since it uses a member
> > function, and I'm not very fond of using member functions for general
> > classes -which could be replaced for something similar-
>
> Well, the emphasis is not on the 'get' member function. You could easily
> provide a free function instead.
>
> The important thing about the idiom is that it checks for initialized
> optional<T> _and_ retrieves a pointer to T at the same time, so - assuming
> that you follow the idiom - you can't access an uninitialized optional by
> mistake (since the pointer is not in scope in the 'else' clause.)
>
I see. Good idea!

Do you think that the functions:

  friend template<class T> T* get ( optional<T>& opt ) ;
  friend template<class T> T const* get ( optional<T> const& opt ) ;

used as:

if(T * pt = get(opt) )
{
 // initialized, use pt
}
else
{
 // not initialized, no pt in scope
}

could present ambiguity with any other 'get'?.

I think that koenig lookup should always resolve this right, but I am not
sure.

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk