Re: [gpm] Re: gnome-screensaver 0.0.23

[William Jon McCann <mccann jhu edu>]

Hi Jaap,

Sorry for the delay.

Jaap Haitsma wrote:
> I don't really see why gnome power manager should depend on 
> gnome-screensaver for getting the idle time, because as a user for some 
> reason I might not want to install gnome-screensaver. 
Blasphemy!  But seriously, I don't think this is a problem for a few 
reasons.
g-p-m would not necessarily depend on gnome-screensaver.  It would 
depend on something that implements a similar D-Bus interface, or more 
specifically, sends a SessionIdleChanged signal.  This would be further 
reduced if we move to a org.freedesktop D-Bus namespace, eventually.
Anyway dependencies aren't really bad - they are a side effect of proper 
software layering and modularity.  As far as dependencies go, D-Bus 
relationships are as loose as they come.
> A better solution 
> is to have one C file exporting the function
> int get_keyboard_mouse_idle_time ()
>
> If that file is part of gnome-screensaver, gnome-power-manager would 
> just copy that file and put it in a cut-n-paste directory like the way
> it's currently done with libegg files
There are a few reasons why this would be so hot.  We don't want g-s and 
g-p-m disagreeing about session idleness.  So, optimally idleness is 
detected and defined in one place.  Also, the way that this idle 
detection stuff works is that it selects for events from every window in 
the session and polls the mouse every so often.  It wouldn't be very 
efficient to have multiple applications doing this sort of thing.  We 
also don't want to be fixing bugs in this somewhat complicated code in 
more places than we need to.
Jon