I've configured the screensaver to start after 5 minutes. However this does not happen. In Gnome it works ok.
I'm suspecting that it might have the same cause of these bugreports:
Which maybe be caused by this change in gnome-screensaver:
Is gnome-screensaver actually running?
I have no intention of adding private API to xfce4-session for getting idle time. The X server provides this. If gnome-screensaver can't figure out idle time on Xfce, then gnome-screensaver is buggy.
Your problem is caused by gnome-screensaver now requiring gnome-session to run to get it's idle time. There doesn't seem to be any reason for this change in gnome-screensaver (except, maybe, causing more trouble to non-GNOME users who want to use gnome-screensaver).
Thanks for the comments. This pretty much sucks as I like to run gnome-screensaver and not xscreensaver.
Yes, it definitely suck, but there's not much one can do, exept asking for GNOME people to revert the change and stay with the X idle counter.
(In reply to comment #4)
> Yes, it definitely suck, but there's not much one can do, exept asking for
> GNOME people to revert the change and stay with the X idle counter.
D-Bus is a very important message system, but in many places (like the case of gnome-screen-saver), it is abused, causing lot of problems and hidden runtime dependencies.
Another trivial stuff was moved from gnome-screen-saver (and gpm also) to the gnome session is the ability of inhibiting automatic DPMS activation on idle (or automatic sleeps for gpm), so no way for a media player to prevent gnome screen saver from switching off the screen if gnome session isn't running, this definitely sucks, i think is time to have our own screen saver application.
What about a GTK+ front-end to Xscreensaver's unlock dialog? It's not gonna be easy but it will be possibly more secure than gnome-screensaver and kscreensaver. See http://www.jwz.org/xscreensaver/toolkits.html
(In reply to comment #6)
> What about a GTK+ front-end to Xscreensaver's unlock dialog? It's not gonna be
> easy but it will be possibly more secure than gnome-screensaver and
> kscreensaver. See http://www.jwz.org/xscreensaver/toolkits.html
Unless you're an X11 security expert, I don't think it's a good idea (exactly for the stuff explained in the above page)
Why not simple fork gnome-screensaver and in this fork revert to the old behavior?
(In reply to comment #8)
> Why not simple fork gnome-screensaver and in this fork revert to the old
By studying the source for gnome-screensaver it seems to be these 3 functions in gs-watcher-x11.c which would need a rewrite:
on_presence_status_changed (DBusGProxy *presence_proxy,
on_presence_status_text_changed (DBusGProxy *presence_proxy,
const char *status_text,
connect_presence_watcher (GSWatcher *watcher)
Eg. rewrite connect_presence_wather to optain idle signal from another source than gnome-session and adjust the two callbacks to reflect this change.
This is getting a bit out of scope for this bug. If you want to complain about gnome-screensaver's deficiencies, you'll need to file a bug on bugzilla.gnome.org. Nothing is going to get solved here.
As mentionned in the Gnome Bugzilla  (and other related bugs), I've had the same problem with my simple Fvwm-based environment and ended up writing a Python script which impersonates the Gnome SessionManager, and does the XIDLE detection for Gnome ScreenSaver. It's available from , and the code is documented at .