X11 screensaver inhibition

[RJVB]

When I watch long enough videos, series episodes or movies on my Linux rig, I'm obliged to move the mouse periodically because my screensaver will have kicked in.

I see that an attempt is made to prevent this by calling an X11 primitive at a regular interval (fixed, rather than tunable to the user's screensaver/locker timeout) but apparently this call has no effect on xscreensaver which I happen to use.

Xscreensaver doesn't have a "pause" option but its deactivate command has the effect of resetting the timer even if the screensaver is not currently active.

IOW, if xscreensaver-command is found on the path, it could be invoked periodically (and at the start of playback) with the -deactivate argument. It doesn't seem rocket science either to do reimplement this command; it involves resetting DPMS (if enabled) and sending a synthetic ClientMessage to the xscreensaver window.

[zaps166]

Check if disabling DBUS_PM helps.

[RJVB]

Why do you think it would? I just confirmed that xscreensaver does somehow activate in reaction to org.FreeDesktop.ScreenSaver.SetActive . But that's in fact the screen locker (and fortunately I do not have to unlock the screen when it triggers in the middle of a movie).

I think what's happening here is that xscreensaver is a standalone utility that (probably) works best when KDE's idle screen locking is deactivated.

Do you know if either of the 2 mechanisms you already hook into prevent DPMS from turning the display off, if the user has that configured?
EDIT: yes, it does seem so, and thus all I'd need to do is reimplement the code to find the xscreensaver window and send it a regular event when found.

Or maybe sending an event to the X11 root window is seen as user activity too?

[zaps166]

When I used X11 desktops (Mate, Xfce4), I never had issues with this and IIRC I rather used DPMS-only, never real screen saver. And I don't remember details, this code is very old (10+ years). If XForceScreenSaver() doesn't work for you, you have to fix it yourself locally.

Executing xscreensaver-command periodically is not a good idea - better use API without spawning process every x seconds.

All modern desktops use D-Bus, so I hope it's working fine (no issues for me).

[RJVB]

On Monday May 11 2026 14:49:58 Błażej Szczygieł wrote: If `XForceScreenSaver()` doesn't work for you, you have to fix it yourself locally.

I'd have to ask why that doesn't work, but would you consider a fix for a PR?

Executing `xscreensaver-command` periodically is not a good idea - better use API without spawning process every x seconds.

In principle you're right of course, but this doesn't have to be done every few seconds of course - who sets a screensaver to kick in that fast? ;)

All modern desktops use D-Bus, so I hope it's working fine (no issues for me).

xscreensaver isn't exactly modern, but it *is* very effective. And it's still the only way I know to set only a screensaver that doesn't lock automatically if you don't want that.

[zaps166]

In principle you're right of course, but this doesn't have to be done every few seconds of course - who sets a screensaver to kick in that fast? ;)

IIRC it was possible to set < 1 sec for X11 😄 Currently QMPlay2 resets X11 screen saver every 30 seconds.

xscreensaver isn't exactly modern, but it is very effective. And it's still the only way I know to set only a screensaver that doesn't lock automatically if you don't want that.

AFAIK desktop environments have option whether to lock on screen saver or not. I think Xfce4 have this option 🤔

I'd have to ask why that doesn't work, but would you consider a fix for a PR?

Why not, but it should not interfere with current implementation. IIRC Long time ago QMPlay2 had different screen saver inhibition code 🤔 Maybe the screen saver reset timer has been called every video frame? Or it used different method for this. I don't remember, please look at git history 😄