Bug 21624 - keyboard mapping of win/alt to control_L fails
Summary: keyboard mapping of win/alt to control_L fails
Alias: None
Product: xkeyboard-config
Classification: Unclassified
Component: General (show other bugs)
Version: unspecified
Hardware: Other Linux (All)
: medium normal
Assignee: xkb
QA Contact:
Depends on:
Reported: 2009-05-07 16:54 UTC by brian smith
Modified: 2009-06-24 01:28 UTC (History)
0 users

See Also:
i915 platform:
i915 features:

xkbcomp :0 -xkb failed_out.xkb (44.02 KB, application/octet-stream)
2009-05-13 16:15 UTC, brian smith
xkbcomp :0 -xkb working_out.xkb (53.84 KB, application/octet-stream)
2009-05-13 16:17 UTC, brian smith

Description brian smith 2009-05-07 16:54:12 UTC
I filed this with Gnome but they sent me to you, hope you can help:

I use gnome-keyboard-properties to change my Alt/Win key behavior to be the
same as my control key. This works initially but fails over time. When I'm not
using the machine it doesn't fail.

Steps to reproduce:
1. gnome-keyboard-properties
2. Layouts tab
3. Keyboard model "Apple" (unnecessary I think)
4. Layout USA
5. Layout Options...
6. Alt/Win key behavior
7. Control is mapped to the Win-keys (and the usual Ctrl keys).
8. Use your computer for a few days, it will work initially then fail.

Actual results:
I'll notice the alt/win key no longer works and then I run "xmodmap -pke | grep
133" which shows:
keycode 133 = Super_L NoSymbol Super_L NoSymbol Super_L

Expected results:
When the keys are mapped correctly I can run "xmodmap -pke | grep 133" which
keycode 133 = Control_L NoSymbol Control_L NoSymbol Control_L

Does this happen every time?
Yes, in any given 72 hour period, it will happen at least once. If I'm not
using my computer, it doesn't happen.

Other information:
I'm using an apple A1048 keyboard, I don't think that's related.
I've been having this problem for over a month, I haven't been able to diagnose
a cause yet.
The workaround is to check change the Alt/Win key behavior back to Default,
then change it back again.
Comment 1 Sergey V. Udaltsov 2009-05-08 13:54:48 UTC
xmodmap is not interesting. Would you attach results of "xkbcomp :0 -xkb out.xkb" before and after the failure
Comment 2 brian smith 2009-05-13 16:15:46 UTC
Created attachment 25842 [details]
xkbcomp :0 -xkb failed_out.xkb

ran command after failure
Comment 3 brian smith 2009-05-13 16:17:41 UTC
Created attachment 25843 [details]
xkbcomp :0 -xkb working_out.xkb

ran command while mapping was working
Comment 4 Sergey V. Udaltsov 2009-05-14 01:39:53 UTC
The second file shows that your XKB is reset (and even reset incorrectly - the second file lacks the geometry part). Do you do some suspend/resume? Or perhaps plug the keyboard out, then plug in again?

BTW, what is the version of GNOME you're using? What distro?
Comment 5 brian smith 2009-06-24 01:19:29 UTC
I'm not using any suspend/resume, however I am using a USB kvm (IOGEAR). The KVM switching doesn't seem to correspond with the failure, so I never considered that to be a possible cause.

Upon the release of Fedora Core 11, I've reinstalled and no longer experience this issue (I waited a couple of weeks, no problems).
Regardless, thanks for your help with this.
I have the old install on another drive so if you need me to pull files off it I can do so.
Comment 6 brian smith 2009-06-24 01:21:50 UTC
in response to comment#4, I was running Fedora Core 10 with all updates installed (except kernel updates).
Comment 7 Sergey V. Udaltsov 2009-06-24 01:28:02 UTC
It was because of KVM/USB. The keyboard (from the X and OS POV) was plugged out/plugged in. The earlier versions of GNOME keyboard stack did not process these events. Now they should (if built properly).

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.