<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body bgcolor="#ffffff" text="#000000">
I think I created a better rig. Leaving Cog out of the picture seems
to work better. <br>
<br>
A standard image from squeak.org with VMMaker loaded in from
MetacelloRepository. <br>
<br>
I create a fullscreen stimulator.image with the code in the
workspace:<br>
<br>
<i>(InterpreterSimulatorLSB new openOn: 'source.image') test</i><br>
<br>
Then open the stimulator image and execute. I get the satisfying
"Object Pointer Relocating" message and it's on its way. <br>
<br>
I then get a ByteSymbol>>coerceTo:sim: error stemming from
Interpreter>>addNewMethodToCache<br>
<br>
While this could work, I don't get the impression that this is how
people hack the vm: activating the simulator; getting the
satisfactory image; generate C files to compile. <br>
<br>
I bet people hack the methods in the Interpreter class directly. I
bet that's faster than getting the simulator to work and then doing
things in slow motion with the slow simulator. <br>
<br>
The simulator is more historical than practical, I'm thinking. <br>
<br>
Chris <br>
</body>
</html>