On Mon, Feb 8, 2010 at 9:43 PM, Jason Childs
<[email protected]> wrote:
> I've looked at a lot of kernel data by putting extra buffer print
> statements in the kernel over the past few months.  I've yet to see the
> tablet report or jump to zero in the usb buffer when a finger is down
> (unless it's at 0,0 or an edge obviously).  Wherever this zero is coming
> from is definitely the cause of the jumping we've seen in relative mode,
> but it's most likely somewhere between the kernel and x.org.  The tablet
> seems pretty solid with it's data.  I have noticed that a finger event
> that contains the same data from the previous set of events sent to the
> x.org driver don't seem to show up in the x.org driver.  Is this
> expected behavior?

Thanks for feedback.  I was hoping to hear about your experience.
I've added printk's and looks good to me as well; at least it seems I
can't reproduce the same behaviors anymore.  I guess I'll switch back
to x driver for some more debugging.

BTW, in my rewrite of gesture code, I'm not really seeing jumping
cursors anymore.  Are there specific gestures your seeing it with?
Perhaps it will jog a memory I had while debuging it.  Most my jumps
were from letting channel 1 data into commonDispatchDevice() and it
getting confused by its x/y values.

As far as duplicate events not showing up, I do see xf86WcmSuppress()
in my logs and occasionally filtering stuff so I've been assuming that
duplicate events occasionally are coming.

Chris

------------------------------------------------------------------------------
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_______________________________________________
Linuxwacom-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel

Reply via email to