[Vm-dev] rearchitecting the FFI implementation for reentrancy

[Andreas Raab]

David -

My recommendation would be not to get held up by whatever Eliot is 
working towards. There is no telling how long it is going to take us 
until any of it is released since we always go through a full 
discussion-review-decision cycle before making such choices (incl. 
approval by the board). In fact, at this point there isn't even any 
telling whether we're going to release any of it or not ;-)

So what I'd recommend doing is to move forward with whatever has been on 
the list of things to do. Eliot is pretty good in folding these things 
into his work where it makes sense, and 64bit stuff is important for us 
too; in particular with server deployments.

Cheers,
   - Andreas

David T. Lewis wrote:
>  
> On Wed, Aug 05, 2009 at 05:56:02PM -0700, Eliot Miranda wrote:
>>  
>> Hi All,
>>     I'm looking at making the Squeak FFI reentrant to support nested calls
>> and possibly threading.  The current FFI has a couple of issues which render
>> it non-reentrant.
> 
> Your approach sounds fairly straightforward. Slightly off of your intended
> topic, but I want to mention also that we have some changes queued up for
> FFI to address 64 bit pointer issues. The changes are fairly extensive, so
> I'd like to lobby for getting these done before too much bit rot sets in.
> 
> For a description of the status of the 64-bit FFI updates:
>   http://lists.squeakfoundation.org/pipermail/vm-dev/2008-April/001900.html
> 
> Patches on Mantis:
>   http://bugs.squeak.org/view.php?id=7237
> 
> Eliot, I don't know what your time frame is for doing the reentrant FFI work,
> but if we could do the 64-bit updates first that would be a very good thing
> from my point of view.
> 
> Is there any interest in taking this on right now? The affected folks are
> Andreas, Ian, and John. Opinions?
> 
> Dave
>