bruns added a comment.
In D22296#492147 <https://phabricator.kde.org/D22296#492147>, @sitter wrote: > > So, in order to port to the API, it'd first need finishing up really. The question is - do we really want this call here, or not. As is, it is just an expensive noop. **Iff** we want the inhibition, I see 3 possibilities: 1. Teach powerdevil to behave correctly: - do the inhibition immediately - keep running until all proxied inhibitions are released (maybe this is inversed - do not kill powerdevil from ksmserver). - somehow interact with powerdevil from ksmserver 2. Take an inhibitor lock via DBus (i.e. call `org.freedesktop.login1.Inhibit`) directly 3. Take an inhibitor lock using `Solid::Power::inhibit` (3.) is nothing more than a thin wrapper around (2.) see https://github.com/KDE/solid/blob/master/src/solid/power/backends/freedesktop/fdinhibition.cpp Although Solid::Power would give some abstraction, the reality is: - only blocking inhibitions are supported - only implementation is `org.freedesktop.login1` - Solid::Power is very incomplete, e.g. the `statesJob` is just a stub. My preference is (0.) or (2.). REPOSITORY R120 Plasma Workspace REVISION DETAIL https://phabricator.kde.org/D22296 To: bruns, #plasma, broulik, davidedmundson Cc: sitter, apol, plasma-devel, LeGast00n, jraleigh, fbampaloukas, GB_2, ragreen, Pitel, ZrenBot, himcesjf, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, mart