Hi Craig,
 
> 	Currently, "smallest" and "working" will both crash 
> unless there is a server for them to speak to (e.g., an 
> instance of Sorcerer in the "control" snapshot). "smallest" 
> speaks a different (older) object serialization protocol than 
> "control" does, so the only way to do anything with it is in 
> the virtual machine simulator. I mention this on the Squat 
> release site.

	Can the control.image talk to working.image asking it
	to shrink to something like smallest.image ?

	What's involved ? That would make a nice tutorial
	for using Squat.
 
> > BTW, I hope you can find some time to put a Squat package 
> on SM. This 
> > would be very helpful for those who want to review Squat.
> 
> 	People might think it's something they can install and 
> run with a mainstream release; it's not. The most helpful 
> thing for me would be for people to review what I've provided 
> on the Squat website. Is there a precedent for SqueakMap 
> packages which make inherently fundamental changes to the 
> virtual machine (as opposed to just a plugin)?

	You already gave the cs (and interp.c) for changes 
	in the VM. I think that's good enough. 

	What I meant are the cs (or dare I say mcz ;-) for 
	the Squat related classes.

	I thought it might help TFNR to have Squeak beheaded ;-). 
	Anyway, it's your call.

> > PPS: Latest news, latest control.image on Feb 19 is broken. 
> No crash 
> > but output console was open with many messages for a little 
> while then 
> > Squeak self terminated.
> 
> 	I cannot duplicate this (on win98, I don't have winXP). 
> I suggest rebooting, deleting all Squat bits, and starting over. :)
> 
> 
	I can confirm that it's the image that's causing the problem 
	because the old image run OK with the new flow plugin.

	Maybe it's a combination of new control.image and new flow.

	Hope this helps.

	Cheers,

	PhiHo.