idle detection

[John Bailey]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Sean Egan wrote:
> When you boil down to it, it's not much more extreme than once every 5
> seconds, the previous timeout.

True.  I proposed the one-minute polling as a sort of compromise, knowing that
we did poll "excessively" in the past but not remembering the exact frequency of
that polling.

> My biggest concern is that when you're idle, that's when you'd prefer
> to have Pidgin sleep most to keep the CPU in low-power mode as long as
> possible.
> 
> -s.

I agree here, which is why to me a one minute poll interval seems a reasonable
compromise, because at most you're inaccruately shown as idle for 59 seconds or
so, compared to 9 minutes and 59 seconds.  The timers I mentioned previously
would be a benefit here because as I understood it, they grouped all the actions
performed at this frequency in a rather tight spacing to concentrate the time
that the CPU needed to be active before returning to low-power states, thus
hopefully saving more power in the process by avoiding many state transitions.

John
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFGqDJDBWJH/emdNtsRArE8AJ0V6lcw1J+6dTE4NG62fbP87k4ybACcCdkH
Cg2KTaYsnFwy0/QEbp1CpTU=
=5cW5
-----END PGP SIGNATURE-----