I'm using the 3.0pre2 Unix VM under Linux 2.4.0, with a 3.1 image. What I'm finding is that:
* normally things run pretty well; if I do a Message Tally I see about 80% of the time being spent in Delay>>wait. The CPU usage is below 5% for Squeak.
* sometimes, though, Squeak uses about 100% of a CPU, and if I do a similar Message Tally, I see 75% of the time being spent in EventSensor>>nextEventFromQueue.
It can go from the "good" state to the "bad" state just by being left alone with no interaction.
Saving the image, quitting, and restarting fixes this problem.
Any suggestions?