Re: [boost] Re: Traits-based design and ODR

From: Phil Nash (phil.nash.lists_at_[hidden])
Date: 2002-05-01 19:56:51

> Admittedly my originally suggestion was a bit "pie-in-the-sky". I hope I
> have fleshed it out a little. A more usable example is something that will
> have to go on my weekend to-do list

Hmm, having thought about this before, and seeing some of the other comments
elsewhere about boost::shared_ptr and runtime policies, I am beginning to
think that what I had in mind is pretty close to what we already have in
shared_ptr anyway. So perhaps it is not so pie-in-the-sky after all :-)

In which case, FWIW, I would add my voice of agreement that something like
shared_ptr could be the "standard" way to pass a shared pointer across a
published interface, as long as that version of shared_ptr could absorb the
compile-time policies of a Loki-like smart_ptr (that is a better tool for
internal implementation) into run-time policies held internally. It already
does that in some ways (such as the intrustive ref-count detection).

Of course, as has been stated several times now, static policies provided to
Loki::SmartPtr could be written that defer their decisions to run-time
anyway - and so, if done carefully, the boost::loki::smart_ptr<T> (default
policies) could replace shared_ptr altogether (name change intended there).

This too I am in hearty agreement with as we gain greater scope for
inter-operability between compatible policies, as Andrei has reminded us of
so many times.

Andrei, put in that light does what am saying make more sense and sound
reasonable now (I think it means I was agreeing with you more than I thought
right from the start)?

Regards,

[)o
IhIL..