[01:36] <mup> PR snapd#12578 opened: interfaces/apparmor: Add read of /proc/PID/cpuset to base template <Created by alexmurray> <https://github.com/snapcore/snapd/pull/12578>
[10:48] <mup> PR snapd#12579 opened: cmd/snap,image: refactor seed manifest code in preparation of validation-sets <Created by Meulengracht> <https://github.com/snapcore/snapd/pull/12579>
[11:09] <stuartlangridge> hey, all. Firefox (as a snap) tries to inhibit my screensaver when it plays videos (correctly), but the D-Bus call is being denied by apparmor. How can I allow that? Is this a bug in the firefox snap metadata that I need to file, or should it be creating a snap connection somehow itself?
[11:10] <stuartlangridge> the denied message, shown by snappy-debug, is:
[11:10] <stuartlangridge> Log: apparmor="DENIED" operation="dbus_method_call"  bus="session" path="/org/freedesktop/PowerManagement/Inhibit" interface="org.freedesktop.PowerManagement.Inhibit" member="Inhibit" mask="send" name="org.freedesktop.PowerManagement" pid=288754 label="snap.firefox.firefox" peer_pid=2736 peer_label="unconfined"
[11:12] <lissyx> stuartlangridge, bandali can help you on that but I suggest you file a bug on bugzilla and/or on launchpad for that
[11:13] <stuartlangridge> lissyx, I can certainly do that; what I wasn't sure about is whether it's a local config issue, or an actual problem with the snap. (I mean, I think this ought to work out of the box, of course!) If filing a bug is the best course I'll definitely do that
[11:28] <stuartlangridge> ah. there already is a bug, which nobody is working on :(  https://bugzilla.mozilla.org/show_bug.cgi?id=1785799
[11:33] <ogra> stuartlangridge, TBH this looks more like a snapd bug to me, i ave seen that message with things like the snapped vlc before 
[11:35] <stuartlangridge> I think the actual issue is: firefox isn't allowed to send dbus Inhibit messages to org.freedesktop.PowerManager. What I don't know is whose bug that is; does it need an extra setting in firefox's snapcraft.yaml? Does snapd need to allow these by default? Is there some way a user can create that connection at runtime, and so that should be an
[11:35] <stuartlangridge> automatically made connection on install?
[11:42] <amurray> thanks for the heads up stuartlangridge - there is already a screen-inhibit-control interface in snapd that allows various dbus methods to inhibit the screen saver and firefox plugs this - but the interface doesn't include this particular dbus interface - I just submitted https://github.com/snapcore/snapd/pull/12580 to resolve this
[11:42] <mup> PR #12580: interfaces/screen-inhibit-control: Add support for xfce-power-manager <Created by alexmurray> <https://github.com/snapcore/snapd/pull/12580>
[11:43] <mup> PR snapd#12580 opened: interfaces/screen-inhibit-control: Add support for xfce-power-manager <Created by alexmurray> <https://github.com/snapcore/snapd/pull/12580>
[11:51] <stuartlangridge> nice one alexmurray!
[11:53] <jamesh> Doesn't firefox inhibit the screensaver via xdg-desktop-portal? The blocked upower call would be more about inhibiting automatic suspend
[11:57] <stuartlangridge> jamesh: it may be
[11:58] <stuartlangridge> I'm only seeing apparmor blocks for org.freedesktop.PowerManagement.Inhibit
[11:58] <stuartlangridge> firefox thinks it's trying to inhibit screensavers too according to MOZ_LOG
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock WakeLockListener video-playing state locked-foreground
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock shouldLock 1
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock SendInhibit(): FreeDesktopScreensaver
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock SendInhibit(): GNOME
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock SendInhibit(): FreeDesktopPower
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock WakeLockListener video-playing state unlocked
[11:58] <stuartlangridge> [Parent 288754: Main Thread]: D/LinuxWakeLock shouldLock 0
[11:58] <stuartlangridge> but... the screensaver is not inhibited.
[11:59] <stuartlangridge> and I don't see any of those calls in dbus-monitor, although I might be using dbus-monitor wrong
[14:49] <mup> PR snapd#12580 closed: interfaces/screen-inhibit-control: Add support for xfce-power-manager <cherry-picked> <Created by alexmurray> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/12580>
[14:49] <stuartlangridge> ok, having tested this stuff a bunch more, I think that what's supposed to happen is that firefox calls org.freedesktop.powermanager over dbus to inhibit suspend; on xfce that's provided by xfce-power-manager; xfce-power-manager then calls org.xfce.screensaver by dbus to inhibit that. But apparmor blocks the initial call. That will hopefully be
[14:49] <stuartlangridge> fixed by alexmurray's patch, which should hpefully arrive on people's desktops in relatively short order because snap core updates outside a release... so that may fix everything! Hooray!
[14:50] <stuartlangridge> https://bugzilla.mozilla.org/show_bug.cgi?id=1785799 has a detailed update on what might be happening, but hopefully this fixes everything when it arrives.
[14:50] <stuartlangridge> And I see it's just been merged too, nice one mvo. So this is perhaps all great :)
[15:34] <mup> PR snapd#12489 closed: image/preseed: re-execute snap-preseed --reset <Run nested> <Created by valentindavid> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/12489>
[16:19] <mup> PR snapd#12581 opened: release: 2.58.3 <Simple 😃> <Created by mvo5> <https://github.com/snapcore/snapd/pull/12581>
[16:33] <lissyx> stuartlangridge, sorry but I thought I could be working today but I can't really, saw your comment on https://bugzilla.mozilla.org/show_bug.cgi?id=1785799#c2
[18:24] <stuartlangridge> lissyx: I'm quietly hopeful that alexmurray's fix above may sort it, but I wanted to get as much stuff written down that was relevant as I could, so someone could use it to look into the problem more later if the snapcore fix isn't the thing needed.
[18:24] <stuartlangridge> but this certainly seems to be an issue with the snap sandbox, since the binary does it right when it's not sandboxed