oliverhenshaw added a comment.
In https://phabricator.kde.org/D2033#50969, @broulik wrote: > So, should we go with this patch then? Powerdevil is now a separate binary, so > > IdleTime::instance()->removeAllIdleTimeouts(); > > shouldn't cause side-effects on other modules and also should fix it for non-weird X setups like mine. ;) Yep, the removeAllIdleTimeouts now looks good. The bad news is that I've been running with an evolution of https://phabricator.kde.org/D2393 with no problem for a while and still got the suspend-after-user-switch happen today. Looking through logs it seems like this time kded (I'm still using a patched stable release) was getting blocked even before onActiveSessionChanged (so before the login1 dbus signal) so I would still have had the problem with this patch. I never got the the bottom of the blocking in X issue, best guess was either that the xserver was drawing by itself and blocking, or that kded was blocked since it's a DRI client. The powerdevil binary does drawing for the fadeout effect, right? How about checking (KIdleTime::instance()->idleTime() >= msec) in Core::onKIdleTimeoutReached() before preceding with the idle action? REPOSITORY rPOWERDEVIL Powerdevil REVISION DETAIL https://phabricator.kde.org/D2033 EMAIL PREFERENCES https://phabricator.kde.org/settings/panel/emailpreferences/ To: broulik, oliverhenshaw, sebas, #plasma Cc: sebas, oliverhenshaw, graesslin, plasma-devel, lesliezhai, ali-mohamed, jensreuterberg, abetts
