romangg added inline comments. INLINE COMMENTS
> davidedmundson wrote in pointerconstraints_interface.h:126 > do we not want a getter method? > > One of kwayland's main roles is to be a property cache The hint is used by the compositor after unlock again. This unlock can also happen on destruction of the LockedPointerInterface. So one need to to connect to a signal announcing this and then set the pointer position. Resource has a `aboutToBeUnbound` signal. Maybe use this? > davidedmundson wrote in pointerconstraints_interface_v1.cpp:175 > Only non-negative x and y values are allowed > > is it worth enforcing that with a check here? I wanted to allow negative values, such that a client can decide to opt-out of hinting again. Sadly the specs does not say anything about invalid values. But my assumption is that a client should only be allowed to shift the cursor in the dimensions of its surface. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D14175 To: romangg, #plasma, davidedmundson Cc: davidedmundson, kde-frameworks-devel, michaelh, ngraham, bruns