-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 2 Dec 2002 03:55, [EMAIL PROTECTED] wrote:
> Almost working! I think there is more to this than just the z-axis
> however. Putting the recommended changes into XF86Config for Z-Axis makes
> scrolling random and unpredictable, i.e. often nothing happens, and when
> it does scrolling jumps about in unpredictable ways.
> This is quite strange...
<snip>
The mouse appears to have dodgy sensors, which are giving you correlated 
readings for motion that is purely in one axis.
The "random and unpredictable" bit is because the normal wheel sensor gives 
you +/-1 for a small amount of motion (and wheels often have physical 
feedback, so you can feel it "clicking over"). You may be able to change the 
window manager response to this, depending on what you are running (KDE can, 
but probably not enough - you can only get down to a single line per click, 
and you need something like 0.05 lines per click).
I'm actually starting to think that the button should have been two abolute 
axes. No wonder they were giving them away....

> ...then suddenly the mouse dies!!!!......Had to restart the machine (i.e.
> not just the X-server). Couldn't see which process had died because I
> didn't have a before and after :(
Be real! This isn't helping anyone. You haven't told us which kernel you are 
using, which X, what you were doing at the time, haven't provided a oops 
dump. nothing.

> So, it seems there IS a correlation between left and right i.e. (-ve and
> +ve) but that these readings overlap with up and down (there don't seem to
> be any threshold values do there?).I wonder if evtest.c picking
> *everything* up?
evtest could possibly do a bit more error checking, but I think that the mouse 
is reporting things that are being correctly displayed. The report you 
provided shows that the Wheel and Z reports correspond directly with motion 
in the vertical and horizontal directions. There _will_ be a bit of 
inaccuracy (go back and look at your original reports for for vertical and 
horizontal motion of the mouse itself - note that you get some X and Y in 
both.

> An interesting puzzle?
Not really. As far as the USB subsystem is concerned, the mouse works, and the 
right values are being sent to userspace. If you want to use those values, 
that is a matter for userspace policy and policy implementation.

There is obviously a problem when the mouse stops responding, but that could 
be anywhere (based on the "its broken" level of detail that you've provided).

Brad
- -- 
http://linux.conf.au. 22-25Jan2003. Perth, Aust. I'm registered. Are you?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE96m+7W6pHgIdAuOMRAjPiAJwPMf308yGiXjt5zL0G0V6adQkfZACeJ6ua
e8uhCzOlY8t27SFmcg+VVDM=
=CdRK
-----END PGP SIGNATURE-----



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users

Reply via email to