Tim Rowledge <tim at sumeru.stanford.edu> wrote:


> can't get any
> response to interrupt key.
Hmph. update 5864 is responsible here.

We need to actually agree on what is supposed to happen when the user
presses the designated interrupt key. I thought the VM was supposed to
signal the specified semaphore and NOT produce a keypress event?

This has worked for ages even though the EventSensor code actually looks for the
cmd-. keys because the signal bypassed that and went straight to the
interruptWatcher loop with a jolt straight to the cortex. Ned's change
adds another guardian to the process though (hadInterrupt) which _only_
gets set to true in the EventSensor loop. Splat - no interrupts for me.

I can see the apparent logic of connecting the InterruptSemaphore to
the no-event-checking-in-a-while-tickler loop but it really ought to be
done with a separate Semaphore, indeed a separate background process.
Something like:-
[EventTicklerSemaphore waitTimeoutMSecs: EventPollWait "because
EventPollFrequency is a stupid name".
self fetchMoreEvents] repeat



tim
--
Tim Rowledge, tim at sumeru.stanford.edu, http://sumeru.stanford.edu/tim
Useful random insult:- Supports nativist theories that man is formed from clay.