davidedmundson added inline comments.
INLINE COMMENTS
> romangg wrote in orientation_sensor.cpp:29
> Yea, we need to start it in the beginning to see if a sensor is available.
> But we could cache the result and then shut it off again in case auto
> rotation or screen is off. Needs some more logic though.
From what I can tell, we can call
if (!sensor->connectToBackend()) {
m_available = false;
}
in the constructor
and then we have the option to call start/stop whenever.
> romangg wrote in orientation_sensor.h:38
> Looking at it again I don't think it is advisable to remove these two values
> from the class interface. Instead having them in there for later and ignoring
> them in the meantime in the implementation in the KScreenDaemon is fine. See
> the comment there:
>
> > We currently don't do anything with FaceUp/FaceDown, but in the future we
> > could use them to shut off and switch on again a display when display is
> > facing downwards/upwards.
I don't really understand.
The only reason to wrap QOrientationSensor in a wrapper class is to try and
encapsulate the details of the sensor into something domain specific.
If we just forward everything 1:1, what does this wrapper provide over just
having the other code use QOrietnationSensor directly.
(but whatever this isn't a topic I'm particularly passionate about, so whatever)
REPOSITORY
R104 KScreen
REVISION DETAIL
https://phabricator.kde.org/D26037
To: romangg, #plasma
Cc: plasma-devel, davidedmundson, LeGast00n, The-Feren-OS-Dev, jraleigh,
zachus, fbampaloukas, GB_2, ragreen, ZrenBot, ngraham, alexeymin, himcesjf,
lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, ahiemstra, mart
