Re: gtk_timeout_change()?

[Paul Davis <pbd Op Net>]

>If I have a timeout pulsing every 0.3 seconds, then change that to
>something else, I have removed the 0.3 second pulse.  So of course it
>won't be smooth.  While I'm updating my timer pulse via a slider or any
>other mechanism, unless the delay between the update is LESS than the
>pulse interval, of course I won't hear anything.  Once I move the
>slider, I'm creating a new timer interval, which starts at time 0.  A
>continuous slider move creates dozens of them, all waiting for their
>first beat.
>
> This is why a gtk_timeouot_update() would be so nice.  All I really
>need to do is change the value of the timer interval, not reset it each
>time.  There's a big difference.
>
>Am I making sense?  It's a GTK design issue, not a latency / scheduling
>one.

yes, this makes sense, but its still a bad practice to use the GUI
event loop to deliver precisely timed events. you'd be much better off
using sigitimer(), although alas the obvious method is to use
sigitimer and threads, and thats a problematic area (pthreads + POSIX
signals = headache)

--p