[mythtv] Tracing Channel Change

[Edward Wildgoose]

> > -                usleep(500);
> > +                usleep(100);
>
> > However, this merely pushes the wait time into the =
> > ringbuffer->WaitForPause();  So first of all, why usleep(500) anyway? =
> > And also, does anyone have any ideas on how to trace where this second =
> > call is spending all it's time?
>
> Keep in mind these are microseconds, and a single frame at 29.97fps is
> 33000 microseconds, so this pause isn't that long.  I think it is actually
> longer than 600 microseconds though because the scheduler can't give back
> control to the process that quick, but it shouldn't be anywhere near 33000
> microseconds (33 milliseconds, 0.033 seconds).

Oh, durr.  I assumed they were millisecs...

Thanks