[Squeak-VMdev] re: Squat VM changes for v4 (was "version 4 changes")

[Craig L. <cr...@ne...>]

Hi Andreas--

> > Incorporate method-lookup change to support remote message-sending,
> > in anticipation of Squat's support for minimal snapshots and
> > inter-system communication.
> 
> What kind of method-lookup change? What is needed here...

	The current Squat implementation checks if the receiver's class is a
special "proxy" class (currently called "Other"), which is in the
special objects array. If it is, it kicks the send back to image, using
the same mechanism as for >>doesNotUnderstand:, so that the message may
be forwarded by the proxy.

> and what are the implications for both speed and potential impact on
> JIT?

	I have found the performance impact negligible; the usual case is one
additional identity check against a special object (in C, "class ==
(longAt(((((char *) specialObjectsOop)) + someInteger"). I see no impact
on JIT (assuming it can deals acceptably with the way method lookup
happens now). Hopefully Ian will chime in about that.

> > Allocate header format five for Squat's method dictionary
> > [activation] marking, and version bits in method trailers for
> > Squat's module system.
> 
> I'm slightly hesitent to reserve an object format for something like
> method dictionary marking. Can you elaborate on why this would be
> needed?

	The VM is the thing that does the activation marking. I'm using that
mark to tell when any method from a method dictionary has been run; it's
useful for, e.g., calculating which classes can be swapped out of an
image at some point in time. (There are also primitives for reading and
clearing the marks.)

> ...we only have so many object format types.

	Well, the one I want to use (5) has gone unused for Squeak's entire
history, as far as I can tell. :)  And it's not the only one (format 7
is also unused). And in eight years I've yet to hear of anyone else
wanting one of them. :)  And this is a very good cause. :)  And I'm not
asking for any additional header bits. ;)  (oops, smilie overload...)

> Version bits in method trailers shouldn't be needed if I understand the
> CM changes correctly - you might just add another iVar to CMs.

	True, it just seemed like a logical thing to put in the trailer, and I
was already adding a known-to-the-VM bit there. Oops, I forgot to
mention that in my previous message. That's for *method* activation
marking.

> > The Squat homepage is http://netjam.org/squat/. The release page has
> > links to all the relevant code.
> 
> Hm... something a little more specific would be nice ;-)

	'Sorry, I was pressed for time (still am :). See
http://www.netjam.org/squat/releases/current/vmChanges.zip . Pick a VM
and snapshot from http://www.netjam.org/squat/releases/current/#theBits
to see the image side of things.


	thanks again,

-C

--
Craig Latta
improvisational musical informaticist
cr...@ne...
www.netjam.org
[|] Proceed for Truth!