[boost] Re: smart_ptr implementation descussion topics

From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2002-05-06 15:17:52

"David B. Held" <dheld_at_[hidden]> wrote in message
>  I didn't want to change
> Andrei's ref_counted_mt policy, since it works with a generic threading
> model.
That the point: I am not sure that we have clear understanding how threading
policy should look like.

> > What do you mean as overgeneralization? Is it inconvenient or to
complex?
> > I argue that is not.
>
> It seems like your checking policy is more power than the average person
> needs.
Why? It's just provide a way to specify separate checker for every specific
checks, while proving empty defaults.

> Therefore, it seems like more work for the compiler
It's questionable, cause you force compiler to compile all you classes
always, while in my case it will compile only used specification. In any
case it's minor.

> and more work for the user to take advantage of it.
I argue that it is vice versa, especially in case of custom checks. Compare
what you need to write to implement custom checking policy that checks
dereferencing and what I would write.
BTW I can always provide typedef for the same functionality you version is
providing.

> Dave

Gennadiy.