https://bugs.kde.org/show_bug.cgi?id=354250
--- Comment #21 from Oliver Henshaw <oliver.hens...@gmail.com> --- > @Oliver: Thanks for the investigation. Indeed, this looks like the cause for > the issue. When PowerDevil gets stuck because X blocks it never notices it's > no longer in the active session and shit hits the fan. > Actually if you set QT_MESSAGE_PATTERN so you see timestamps, e.g. following https://woboq.com/blog/nice-debug-output-with-qt.html , and compare it to your watch: then you see that the suspend session calls only happen after you've switched to the other user. The "Current session is now active" lines only happen later because (a) that's how the timing of dbus-vs-xcb events works out, and (b) powerdevil does a roundtrip back to the logind dbus service to see what the currently active session is. So it gets events fired when it wasn't active, but not until it's active again. Then the policy code checks whether it's active and of course it is, so the suspend goes ahead. Then the active-tracking parts of powerdevil realise it's active just after it resumes. -- You are receiving this mail because: You are watching all bug changes.
