I see the rest of the VM folks must have taken the weekend off too?
I'll dig deeper into this discussion tomorrow, unless someone else
wants to answer why for David?



>Maybe I'm not understanding the issue, so please forgive me in advance
>if my comments make no sense in this context.
>
>a) Can't the VM just send a message to some Smalltalk object?
>
>b) Can't you access the VM by providing external-code entrypoints, or
>alternatively in the Squeak model, provide a new primitive?
>
>c) QKS Smalltalk VM contained asynch-message queues. Any code (including
>hardware/OS interrupt drivers) could put a message on the queues. The
>VM's Smalltalk interrupt mechanism would subsequently detect there were
>elements on the queue and generate soft-Smalltalk interrupts at well
>known points. Like between message send/returns, etc. Those "soft"
>interrupts would send a message (vm-callback) to an appropriate
>Smalltalk object to process a given asynch queue entry. This was how we
>provided almost all the real-time services for timebase code execution
>within QuickTime tracks and other multi-media, real-time timer task
>interrupts,  and MacTCP services on MacSTA circa 1991-1995.
>
>-- Dave S. [SmallScript Corp]
>
>SmallScript for the AOS & .NET Platforms
>David.Simmons at SmallScript.com | http://www.smallscript.org

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