[Dev] Re: refcounted

[Katie Capps Parlante]

We also agreed in the IRC discussion that we have a pretty clear path 
ahead -- we won't use the pinning apis (SetPinning), and should we move 
to a new python version in the future that doesn't use refcounting, 
we'll implement the repository api that we *are* using in some other way 
(as Andi suggested, making use of weak references).

Cheers,
Katie

Heikki Toivonen wrote:

> Andi Vajda wrote:
> 
>> Pedronis, on IRC, just added this to the debate. If a faster python VM 
>> were to be made, would Chandler not try to take advantage of it ?
> 
> 
> To which I replied a little later in the IRC discussion that I don't see 
> such a new Python version being usable within the next three years or 
> even longer. Chandler must be usable much earlier than that. Spending 
> resources on a hypotethical gain in so distant future that Chandler is 
> either usable without that advantage or won't be around seems wasteful. 
> Maybe like this discussion, so I'll stop now :)
> 
> 
> ------------------------------------------------------------------------
> 
> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
> 
> Open Source Applications Foundation "Dev" mailing list
> http://lists.osafoundation.org/mailman/listinfo/dev