[Q] Has SharedQueue a very subtle bug?

Martin McClure martin at hand2mouse.com
Sun Apr 14 18:37:16 UTC 2002


At 4:15 PM -0700 4/12/02, Tim Rowledge wrote:
>
>The 'spec' is that process switching is permitted between any two
>bytecodes but the practise for a long time now has been the process
>switching occurs when :-
>a) backwards branch (to catch loops)
>b) real message send (ie a new context is created and control passed on)
>c) long running primitive completes

Ah, that's good to know.

It does seem that we should write code as though a process switch can 
happen between any two bytecodes. If we don't, we make any future 
changes to process switching that much more painful.

-Martin



More information about the Squeak-dev mailing list