[xmonad] My hack to allow disabling of numlock without messing with my xmonad mod key

Thomas Løcke thomas.granvej6 at gmail.com
Wed Mar 20 15:57:14 CET 2013

On Wed, Mar 20, 2013 at 2:26 PM, Brandon Allbery <allbery.b at gmail.com>wrote:

> Again, you cannot fix this reliably with xmodmap, you must use Xkb. Your
> solution will work as long as the interaction *in the X server* between Xkb
> and xmodmap is not disturbed (luckily for you, even the xorg devs are
> terrified of disturbing it...).

Xkb is scary. I've looked at it, and I cannot even begin to grasp what I
need to do in order to accomplish what a short xmodmap one-liner does. That
said, I would of course much prefer to do this the right way, instead of my
current hackish solution.

> And again, this is not really an xmonad issue. You are fighting against
> the essential incompatibility between Xkb keymaps and xmodmap with respect
> to any modifier whose behavior is hardcoded into an Xkb keymap, a problem
> which has been known for years (the XFree86/xorg folks have tried to
> deprecate xmodmap many times since Xkb was added because of this
> incompatibility, but users won't use complex-looking Xkb mechanisms to e.g.
> swap caps lock and control keys).

I'm not saying this is an xmonad issue as such, I'm saying that when I'm
using xmonad, weird stuff happens when I use xmodmap to clear my numlock
key. The same weird stuff does not happen when I use KDE or Fluxbox. These
two WM's happily accept my .Xmodmap file and numlock stays disabled without
other keyboard shortcuts being affected.

> This *cannot* be fixed in xmonad; the problem is happening in the guts of
> the keyboard driver in the X server, and by the time xmonad gets involved
> it's far too late to deal with it.

I guess I need to find myself an Xkb wizard and bribe him/her to help me
kill the numlock key the right way. Does anybody here happen to know such a
person?  :)

Thomas Løcke
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/xmonad/attachments/20130320/e40cf4ff/attachment.htm>

More information about the xmonad mailing list