On Sun, Dec 05, 2010 at 10:18:44PM +0100, Levente Uzonyi wrote:
> On Sat, 4 Dec 2010, Chris Muller wrote:
> 
> >It's December, and time to consider freezing the 4.2 code base.  If we
> >can push any final enhancements and image documentation desired for
> >4.2 in over the next week, we can freeze officially on the 11th.  Is
> >this ok?
> 
> There were plans for releasing 4.2 with SqueakVM and CogVM too. What's the 
> status of the current VMs? Will they be ready for the release?
>

Ian, Andreas, John and Eliot are standing by for a VM build mid-December,
and Esteban is working with John on transition of the Mac responsibilities.
So knock wood, yes the VMs will be ready.

In the interest of managing expectations, I should say that the Cog
and classic VM code bases are not yet integrated (and will not be
in the Squeak 4.2 time frame), so there may be some minor incompatibilities.
But at this point in the process, that's a small price to pay for a
high performance VM. There will probably also be some packaging
issues on Linux (which VM gets run by default, and how?) which will
probably get sorted out after the the new VMs are built.

What you can expect for the 4.2 release is that Cog will be available
as the performance VM on the major 32-bit intel platforms, and that
new versions of the classic interpreter VM will be able to run images
saved from Cog. So there will be a high degree of interoperability,
excellent performance on the mainstream platforms that support Cog,
and the traditional portability and availability of the interpreter VM.

In addition, the weak collections improvements in Squeak trunk are
a major performance improvement for Squeak 4.2, and the necessary
VM support will be present in all of the new VMs.

Dave