> You don't need this check in the method any more, because it is caught by
> rbegin() == rend().
>
> 171 + if (client_list_stacking.empty())
> 172 + return nullptr;
True. Line removed.
> Once you push, you don't need to resubmit. Launchpad will just update the
> diff.
Alright. I've pushed the updated version. Let's see.
> Do we have any tests for this?
No.
> Is there any way to "fake" utouch events?
Only if you do some mocking in a unity test.
But you can record and then replay multitouch events straight out of a /dev/input/eventX file using utouch-evemu-tools (for the recording) and libutouch-evemu1 in your test. But it would be a functional or integration test since several layers would be involved (events coming out of /dev/input/eventX into xserver and then sent as XInput2 events to compiz)
> You don't need this check in the method any more, because it is caught by list_stacking. empty() )
> rbegin() == rend().
>
> 171 + if (client_
> 172 + return nullptr;
True. Line removed.
> Once you push, you don't need to resubmit. Launchpad will just update the
> diff.
Alright. I've pushed the updated version. Let's see.
> Do we have any tests for this?
No.
> Is there any way to "fake" utouch events?
Only if you do some mocking in a unity test.
But you can record and then replay multitouch events straight out of a /dev/input/eventX file using utouch-evemu-tools (for the recording) and libutouch-evemu1 in your test. But it would be a functional or integration test since several layers would be involved (events coming out of /dev/input/eventX into xserver and then sent as XInput2 events to compiz)