>>Someone out there informed me that they've had some stability 
>>problems with the 3.2.5 mac VM on system 8.x.
>
>Ok, it's very possible that the mac classic vm from 3.2.1 onwards 
>has an issue with some computers where after sleep it refuses to 
>recognize the keyboard. I'm still confirming this issue but I think 
>it relates to the code I added to decouple the left/right shift keys.
>
>If that is the case then that code will disappear because it's not 
>workable for most laptops, and not present in the os-x version 
>because of how it interfaces to the hardware and now it appears 
>there is a problem with it's interaction with the sleep logic.
>
>Other feedback is still welcome, especially for the os-x version 
>which is *significantly* different from previous versions of the VM.
>--

After a bit of diagnostics with some input from Dan we discovered 
that the 3.2.x series of Mac VM. most likely will stop accepting 
input every 16r418938 ticks or about 19.8841 hours after rebooting 
your mac if you use the classic VM under os-9/8/7.

This *feature* was introduced by this chuck of common code in sq.h
#define ioLowResMSecs()	((1000 * clock()) / CLOCKS_PER_SEC)

after we switched from a mac specific clock to the more *common* 
clocks in the VM. However I note the Unix port doesn't use this 
#define they fall back C routines for all their clocks, and I suspect 
the windows version does too? The OS-X version also redefines all 
it's clocks and doesn't even use the problem code anyway.

However an interesting issue to discover after all these years.

I'll build a new VM shortly.


-- 
--
===========================================================================
John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
===========================================================================