Arch Linux

Another thing: start playing something on Rhythmbox, manually lock the screen, and click on the 'next' or 'previous' button that appears. Rhythmbox crashes immediatly.

Attached is what gdb spits when I do this.

gobject-introspection (which contains was updated on July 16, so I'll try to downgrade it.

edit: downgrade the package to the previous repo version, gdb spits this instead:

'Program received signal SIGABRT, Aborted.
0x00007ffff6c7a1c9 in raise () from /usr/lib/libc.so.6'

Rebuilding the package solves the more serious bug that causes rhythmbox to crash upon commands in the lock screen, but it still disables auto-locking preferences.

I DISABLED one plugin which is enabled by default:

- MPRIS D-Bus interface

And then I got my screen locked as usual with rhythmbox playing.

I have dbus 1.6.12-1 installed.

BUT, if I restart Rhythmbox, even with this plugin disabled, the screen does not lock again.

This behaviour is caused by the 'power-manager plugin'. This plugin, for some reason is supplied as 'built in' with the source, and so the plugin cannot be disabled from the plugin menu, as it is not visible. See this bug report (https://bugs.launchpad.net/ubuntu/+source/rhythmbox/+bug/912912), not the same exactly, but similar (visibility) issue.

You can make the plugin visible easily by opening, as root:
/usr/lib/rhythmbox/plugins/power-manager/power-manager.plugin

and change the line:
Builtin=true
to
Builtin=false

I'm unsure if this plugin being hidden is intentional?, but I am certain most users would like the plugin visible so that they can toggle enabled/disabled as desired, I still consider this a bug, as the solution is virtually effortless, and the plugin is functional.

The description of the plugin doesn't match what it's doing. The plugin is supposed to inhibit suspend, which would be done by passing GTK_APPLICATION_INHIBIT_SUSPEND. What it does however is passing GTK_APPLICATION_INHIBIT_IDLE, which means the session manager will never kick in the screensaver.

Ah that's why, I recompiled with GTK_APPLICATION_INHIBIT_SUSPEND, and it no longer effects the screen lock. So it seems like it could be a mistake or something overlooked in the code, as already mentioned its action does not match the description.

@Mark H,

How did you recompile? Could you please provide a patch? I could send it upstream (https://bugzilla.gnome.org/show_bug.cgi?id=705105) then.

I recompiled with a couple of SED lines added to the PKGBUILD: http://pastebin.com/szFhNUnf. The differences(patch) are just two lines shown here: http://pastebin.com/3w2nifQZ. I wasn't sure if the hidden nature of the plugin was intended or not. As the fix is so simple I'm guessing the intended behaviour is to hide it, although doing this presents more problems, such as this bug report.

Also I have not fully tested 'GTK_APPLICATION_INHIBIT_SUSPEND', all I know is that it allows the screensaver to work. Personally, I prefer to disable the plugin, its a little buggy.

@Mark H,

Thank you very much. I took the liberty to propose the patch upstream, let's see what the developers can do.

ps: probably the hidden nature of the plugin was indeed intended, but as it is it's passing the wrong thing to the gnome shell.

status?

https://mail.gnome.org/archives/rhythmbox-devel/2013-October/msg00005.html

"It looks like GTK_APPLICATION_INHIBIT_SUSPEND prevents laptops from
suspending when the lid is closed, so it's not the right flag to use
here."