Re: [DIGImend-devel] [PATCH v2] Use kernel-reported tilt range and resolution

[Nikolai K. <sp...@gm...>]

On 08/28/2012 09:16 PM, Jason Gerecke wrote:
> On Mon, Aug 27, 2012 at 12:23 PM, Nikolai Kondrashov<sp...@gm...>  wrote:
>> +                       /*
>> +                        * TODO: set to 0 once Wacom kernel driver is
>> +                        *       updated to report zero-centered values.
>> +                        */
> I don't think this TODO is necessary.

> Ideally, when we change the kernel driver to report zero-centered values,
> we'd also set the resolution and be handled by the above case. There would
> be a slight accuracy penalty (the tilt factor would be 57.3/57 = 1.005
> instead of 1.000), but at 64 degrees we'd only be off by a third of a
> degree.  Even if we came out with hardware that somehow senses a full +-90
> degrees of tilt, the calculation would be off by just under half a degree
> at the extremes. Nothing to worry about.

Sure.

Otherwise, we can specify tilt resolution to be, say, 573 and multiply
values by 10 in the kernel to raise the precision. Or 917 and shift by 4
for extra neatness :)

> Instead, I'd change the comment to note that this is for
> backwards-compatibility with kernel drivers which don't follow use the
> proper semantics. It can be removed (along with wcmTiltXOff) in the
> future when these kernels are no longer in use.

OK, will fix shortly.

>> +                       /*
>> +                        * TODO: set to 0 once Wacom kernel driver is
>> +                        *       updated to report zero-centered values.
>> +                        */
> Ditto above.

Sure.

> Great work! :) I think this patch hits all the issues squarely on the
> head. We still need to decide when to get the kernel fix in place
> (3.8?)

3.8 seems fine.

> and poke the Qt/GTK developers to fix their frameworks, but
> this patch should facilitate a nice smooth transition.

Yes, that was the intention.

> Reviewed-by: Jason Gerecke<kil...@gm...>

Thanks, Jason!

Sincerely,
Nick