Re: deferred deallocation of local objects

[Richard Frith-Macdonald]

On Wednesday, October 15, 2003, at 06:14 AM, Richard Frith-Macdonald 
wrote:

> On Tuesday, October 14, 2003, at 10:17 PM, Alexander Malmberg wrote:
>
> > Figuring
> > out a way of fixing it properly has been on my TODO list for a while,
> > but it has been blocked by other things.
>
> Me too... I'm still stumped.

Oops ... I think there is a tendency to take me too literally ...
What I mean is, I've considered and dismissed a lot of ideas (some of 
which
I've mentioned) and can't see any easy solution.

Ideally I think we need to redesign how retain/release works across DO
with documentation etc.  I don't think it's currently clear exactly how 
this
operates and should operate.

We need a simple, clear retain/release scheme, which is still reasonably
efficient ... ie more sophisticated than sending individual 
retain/release
messages between processes.

This would be the ideal both for clarity and possibly minor performance
improvements, but doing this *and* maintaining backward compatibility
would be difficult.