On 04/06/2009 07:22 AM, Harald Hanche-Olsen wrote:

+ Christof Wolf <ch.wolf@hfph.mwn.de>:

What is the output of 'xinput test 2' during this experiment?

not sure, what do you mean? on a Console?


I think he means this:

Open two terminal windows (or two xterms). In one window, run
xinput test 2   and let it run. In the other, run xev.

If I do this, put the cursor in the xev window and press and hold the
down arrow key for a while, the output from  xinput test 2  is only

key press   133 
key release 133 

(the second line appears when I let go of the key) while in the xev
window I see numerous keypress and keyrelease events for keycode
133. In other words, the xev output shows the key repeating, the
xinput output does not. (Hence I am not sure what useful info the test
will reveal in your case, but I expect Jeremy has a notion.)

Yes, that's the behavior I see as well (and expect). The reason for the xinput output is to eliminate the possibility that the system is somehow seeing the key "really" come up before he releases it (yes, a long shot).

I'm confused whyyour keys aren't repeating... I'm also confused as to why you have the 100% CPU issue while X11.app is in the background (currently worked around by #if 0ing out some code). That 100% CPU issue happens during the mouse event processing. I'm curious, do you have any non-standard input devices or drivers. I seem to recall you have a wacom tablet.

Can you please tell me the output of:
xset q

I'll put together a patch for you to try via MacPorts that will provide more helpful debugging information.