🚀 go-pugleaf

A while ago I asked what would happen if I set the interrupt level of an ST
high enough to disable the vertical blank interrupt.  The responses were
not unanymous, but I got the idea that there should be about a 5% speed
increase, against the possibility of crashing GEM.  I just tried it in a
computation-intensive GEM application.  The good news is no crash.  The bad
news is only maybe a 1% speed-up.  By the way: I tested that the vbl interrupt
was really disabled by flipping the switch that controls whether my Thomson
Ultrascan monitor is claiming to be color or mono: monitor change is detected
in the vbl routine.