PlasmaZones 3.0.14

Fixed

• DPMS-wake autotile orphan reassignment still triggered intermittently (#527, #536): 3.0.13 closed the dropped-monitor case but missed the dual-monitor wake-up where the second output coming back simply shifts the first output’s x-offset. With no output actually removed, oldScreenStillConnected stayed true, and isScreenChangeInProgress() hadn’t flipped on yet because KWin emits the per-window outputChanged before the virtualScreenGeometryChanged that the screen-change debounce listens for, so the orphan reached the autotile-delegation guard with both legs of the check false. screenAdded and screenRemoved are now also wired into the screen-change handler, latching the pending-change flag at the earliest point KWin tells us the output set is changing. The settle path that runs once virtualScreenGeometryChanged catches up is unchanged.