Am Mittwoch, den 03.02.2010, 08:44 +1000 schrieb Peter Hutterer:
> On Tue, Feb 02, 2010 at 09:31:51AM +0100, Rainer Klier wrote:
> > hi,
> > 
> > Am Dienstag, den 02.02.2010, 10:22 +1000 schrieb Peter Hutterer:
> > > On Mon, Feb 01, 2010 at 09:27:34AM +0100, Rainer Klier wrote:
> > 

> > so i don't need the tablet to work as a mouse.
> > i need it as signing-pad only.
> 
> right. From an X point-of-view, you can use XI2 to wait for an event coming
> from the tablet and then parsing the data into the signature. If you grab
> the tablet device directly once it's pressed, you "detach" it temporarily
> and thus make sure the data only goes to your app. Once you ungrab the
> device, it continues to work normally.

but how can i grab the device?
by pressing the pen down to the tablet-surface?
i tried the tablet with gimp.
it is not easy to work with.
as long as i don't press the pen down, the tablet acts like a mouse, and
therefor the area of the tablet is mapped to my x-screen.
i have a dual-screen setup, so the mouse-pointer goes over both screens,
when i move the pen over the tablet.
it is very hard to get the pen into an open gimp-window to paint.
and because the tablet is mapped to the whole screen, the area of the
gimp-window where i want to paint, ist a very little area on the tablet.
after pressing down the pen somewhere inside the gimp-window, i can
paint on the whole gimp-window.
the pen is then somehow "caught" inside the gimp-window.
but i can not "air-move" the pen on some areas of the gimp-window,
because the pen flies outside/over the gimp-window.
it is hardly to describe.
maybe/hopefully you know what i mean.

is this the way (like gimp does it) you mean with "detaching"?

> for a java app this is more tricky, I have no idea if there are decent xlib
> bindings for java and if so they certainly won't exit for XI2 yet. So I
> don't think that's a way to go for you.

yes, i also think so.

> > all i need is a c-language-layer which gets the data from the tablet and
> > talks to the java-coda via JNI.
> 
> look at wacdump, it's a simple kernel debugging tool that dumps the data

yes, i am already into this.

> from the kernel device. That may be enough for the base.

correctly.
i am also thinking to go that way.

> however, accessing the kernel device is tricky since the wacom X driver
> issues an EVIOCGRAB on the kernel device. so while the driver is
> loaded, no other application will receive data from the kernel device.

and yes, thank you, i noticed this already.... ;-)

> if you make sure the X driver isn't loaded, you'll be fine.

absolutely correct.
currently i renamed the wacom_drv.so, so that it can't be loaded.
this is enough for me right now.
but i have to find a more elegant solution to disable the x-driver for
the future and for production releases....

> since this is a rather specific use-case where the tablet isn't needed in
> the desktop itself, I'd say accessing the kernel for data would be enough.

yes, i also think this is the way to go.

> If you want the tablet to work with the desktop too, I can't recommend going
> the kernel way.

ok, understood.
thank you.
-- 
Rainer Klier
Research & Development
xyzmo SIGNificant Group  | Austria | USA  | Germany
A-4052 Ansfelden, Haiderstraße 23
Tel.: +43 7229 88060-0, E-Mail: [email protected]

------------------------------------------------------------------------------
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