Re: [gtkmm] using RefPtr out of Gtkmm

[Chris Vine <chris cvine freeserve co uk>]

On Wednesday 11 February 2004 19:12, Murray Cumming wrote:
> On Wed, 2004-02-11 at 00:23, Chris Vine wrote:
> > On Tuesday 10 February 2004 14:30, Murray Cumming wrote:
> > > On Fri, 2004-02-06 at 18:34, Jose Felix Hernandez Barrio wrote:
> > > > Hi,
> > > >
> > > > I'm developing an application that must use a smart pointer, and i
> > > > have a few questions:
> > > >
> > > > 1- Is there any tutorial for refptr?
> > > > 2- Can i use refptr out of gtkmm class hierarchy? must i derive my
> > > > classes from any glib o gtk classes and add any functionality ???
> > >
> > > RetPtr is documented in the reference documentation and the book. It is
> > > not a general purpose smartpointer. But it's not difficult to write a
> > > general purpose shared reference-counting smartpointer.
> >
> > However, Glib::RefPtr appears to be a fairly typical intrusive smart
> > pointer. According to the gtkmm/glibmm documentation it only requires the
> > referenced object to have reference() and dereference() methods (in other
> > words, to increment and decrement its own reference count and to delete
> > itself when this drops to 0).  If these are home-made objects, this is
> > very easy to do, and if you are already using glibmm is probably worth
> > doing as it saves having to add another library or dependency.
>
> I generally find that it is better and easier to implement a generic
> _non-intrusive_ reference-counting shared smartpointer. You can see
> examples of me doing that in mysqlcppapi and dbus-cpp, and probably
> other places.

It is certainly more aesthetically pleasing, I agree, and puts the reference 
counting where intuitively one thinks it ought to be.  Intrusive pointers are 
fine for languages which require all objects to be derived from a common base 
object class (which can do the reference counting), but that isn't the model 
that C++ follows.

Chris.