Re: [MiNT] [XAAES] Does not generate MU_TIMER events when asked

[Evan Langlois <Evan@CoolRunningConcepts.com>]

On Thu, 2005-07-07 at 17:48 +0200, Odd Skancke wrote:
>  Under 'other AES', the 10ms delay is not reset, so even when you tell
> evnt_multi() you wanted 10ms, you would get what remains of the FIRST
> MU_TIMER you set. So, if each evnt_multi() in this case took 1ms, you
> would get a MU_TIMER each 10th evnt_multi(), even tho this is not
> logical since you do tell it to wait 10ms. So what is the logical
> purpose of MU_TIMER?

Just my 2 cents ...

Anyone that simply wants to do a "poll" of evnt_multi() with a
short-timeout can do so.  If you are looking for longer delay times and
want to be able to service AES messages at the same time but still be
notified of the timeout, the programmer should be able to get a clock
value before each call to event_multi() and set the new timeout by
recalculating when the new expire event should occur.  This should give
the desired behavior under all circumstances.

Of course, you should also be able to set yourself an alarm signal which
should interrupt evnt_multi() as well.

>  However, XaAES is now changed to the 'other AES' method for
> compatibility.

Compatibility is always good.