Hi!

I think, I found the problem - I didn't notice that there's now a
InterpreterSimulatorLSB class which should be probably used instead of the
InterpreterSimulator class on Windows machines.  :-(

BTW, the interpreter simulator needs CObjectAccessor (from Squeak-Plugins)
and CArrayAccessor (from Squeak-PluginTest), both categories of which I
thought they could be savely removed.  At least the required classes should
be removed from a test category then IMHO.

Now I finally can test my delegation changes...


At 17:08 17.11.99 +0100, you wrote:

>> (InterpreterSimulator new openOn: 'sma2.6.image') test
>
>1.   start up and then save+quit "sma2.6" on your local machine.
>2.   (InterpreterSimulator new openOn: 'sma2.6.image') test
>
>If that doesn't work, 
>
>2b.  (InterpreterSimulator new openOn: 'sma2.6.image') test

Ian,

Is it just me or are the two lines and my original expression identical?  I
think, they are.  So what's the difference between your suggestions?  I
tried both :-) and I always get the same problem - an out of bounds
exception after the interpreter tried to swap bytes of a "foreign image".
I don't know why my image is "foreign" as I prepared and saved in multiple
times on my local machine.

>PS: You might have a bug anyway.  I recently spotted what I think
>might cause problems in some of the "internal" loop code.

I'd be glad if I'd reach that point ;-)

InterpreterSimulator>>openOn:extraMemory: determines after reading the
image version number, that my image needs the bytes to be swaped.  This
seems to be a correct assumption as headSize and endOfMemory values look
very good.  Then it swaps the whole image in #reverseBytesInImage.  This
still looks good.

Thereafter, #byteSwapByteObjects is called to undo the swapping for byte
objects, compiled methods and so on.  It happily starts with the first
object at oop=8 which has a boring type of 0, but an amazing size of
2705716480 bytes!  I think, this is the problem.  

bye
--
Stefan Matthias Aust  //  Bevor wir fallen, fallen wir lieber auf.