detecting tablet mode

[Martin Flöser]

Am 2018-01-05 12:48, schrieb Marco Martin:
> I would like to give it a try about it if you don't mind(ie had
> precise plans or code started)..

I don't mind :-)

> so i was thinking about a property for the current state, plus a
> signal for state change.
> as for the location, into ConnectionAdaptor would probably be the
> easiest, iff makes semantically sense it comes from
> org.kde.KWin.InputDeviceManager, which is proably arguable for both
> yes and no ;)

I wouldn't put it into ConnectionAdaptor. What I had in mind was basing 
it on top of D9521 which passes the event down into KWin core. That way 
other areas in KWin can use it.

So I would do something like a small class

class TabletModeManager : public QObject, public InputEventSpy
{
};

and monitor from there whatever the state is. That would make the 
feature independent from the libinput implementation.

Cheers
Martin

> 
> On Fri, Jan 5, 2018 at 6:23 AM, Martin Flöser <mgraesslin at kde.org> 
> wrote:
>> I want to provide a dbus service in KWin to export the state. Apps 
>> cannot use libinput directly, one needs root for that.
>> 
>> Cheers
>> Martin
>> 
>> Am 4. Januar 2018 23:22:43 MEZ schrieb Marco Martin 
>> <notmart at gmail.com>:
>>> Hi all,
>>> one thing i wanted to look into, is to detect when a trasformable
>>> laptop goes
>>> in tablet mode from applications, so that plasmashell and kirigami 
>>> apps
>>> can
>>> adapt themselves and do things(tm)
>>> KWin does that (on wayland) already, and i see it uses libinput for
>>> that, like
>>> i seem to evince from this recent commit (and the logic when to show
>>> the on
>>> screen keyboard)
>>> https://cgit.kde.org/kwin.git/commit/?
>>> id=ac2f41c86d3b5b13ba3655e6c2f17f59a4cb7f01
>>> 
>>> would a normal application use libinput too or is something reserved 
>>> to
>>> the
>>> compositor?
>>> what would be the right way(tm) there? a dbus signal from KWin?