ronjeffries at acm.org
Wed Apr 13 15:33:29 UTC 2005
If I recall correctly, Smalltalk forks only yield when explicitly
told to. So if there had been a "Processor yield" in each fork, I
think you'd have seen more what you were looking for.
Will Turner: This is either madness or brilliance.
Captain Jack Sparrow: It's remarkable how often those two traits coincide.
On Wednesday, April 13, 2005, at 10:36:20 AM, Daniel Salama wrote:
> I ran the following:
> queue := SharedQueue new.
> [10000 timesRepeat: [queue nextPut: 1]] fork.
> [10000 timesRepeat: [queue nextPut: 2]] fork.
> queue inspect.
> and when I inspected "queue", it shows all 1s and then all 2s. If I'm
> forking these two processes, shouldn't they run concurrently? I should
> be able to see 1s and 2s mixed up.
> Then looking into this, I came across someone that said I should
> ProcessorScheduler startTimeSlicing.
> before anything. However, I couldn't find this method in Squeak's
> ProcessorScheduler (it may be a VW thing).
More information about the Squeak-dev