[Desktop_architects] The experience of an ISV
jg at freedesktop.org
Tue Feb 14 11:50:06 PST 2006
Xtst is a local only "solution". As such, it doesn't help LTSP.
The fundamental issue is a desktop communications issue: knowing that at
least one application is currently active that has good reason to
believe the screen saver should be suppressed. Examples: video player
(at least when full screen), presentation program (when full screen).
I claim this is outside the scope of the X protocol (explicitly), though
it can be solved by conventions implemented on top of X.
There are (at least) two ways to solve this problem, that can/should be
discussed on the xdg list:
o use of a X property somewhere to communicate with a session
management daemon (or window manager) of this desire.
o use of an out of band mechanism like DBUS.
Either is satisfactory from my perspective. Note that a proper solution
using DBUS will beg cross system issues (and making dbus secure across
the network); this ought to be done in any case (I know some work was
done to add security to dbus in the past; I don't know its current
status). Folks like LTSP (and others) really care about running over
the network, and this should be possible.
On Tue, 2006-02-14 at 20:27 +0100, Matthias Hopf wrote:
> On Feb 09, 06 13:50:03 +0100, Christian Fredrik Kalager Schaller wrote:
> > Hi,
> > Well in regards to the screensaver issue I thought that already where a
> > solution, but it could be its GNOME specific atm. In GNOME 2.14 Totem
> > for instance is able to tell gnome-screensaver to not activate using
> > dbus. No idea if any other screensaver package listens to dbus for such
> > messages, but that should be an easy fix I assume.
> This is typically done by intruding fake key release events into the X
> events stream, using libXtst.
> Certainly not a nice solution, and even with this one there had been
> issues in xine.
> Desktop_architects mailing list
> Desktop_architects at lists.osdl.org
More information about the Desktop_architects