[xinput] coordinate system of root_x/y

[Thomas Jaeger]

Thanks, I wasn't aware of that.  The the xserver package that comes with
ubuntu hardy (they call it 2:1.4.1~git20080131-1ubuntu9.2, fwiw)
contains a backport of the scaling bug fix, but root_x/y are reported in
screen coordinates.

Anyway, the master behavior is definitely the right one since it seems
very difficult to reliably transform device coordinates into screen
coordinates.  My issue here is that the wacom driver doesn't generate
ChangeDeviceNotify or DevicePresence events when the device is rotated
(btw, which would be the right one?), so it's hard to detect a change in
the valuators' min_value/max_value.

Thanks,
Tom

Peter Hutterer wrote:
> IIRC, the specs are a ambiguous about that, so we have to be careful here.
>
> master reports in screen coordinates because that is a deliberate design
> decision of XI 2. Ignore master for this issue.
>
> server 1.4.0 reported in screen coordinates because of the input coordinate
> scaling bug, but this is fixed in 1.4.1.
> server 1.5 has the same code, so should report the same as 1.4.1.
>
> I don't actually know what 1.3 and before did. Anyone has a 1.3 server
> running?
>
> Cheers,
>   Peter
>