Re: Japanese keyboard auto-detection

[Takuma Murakami <murakami at ipl dot t dot u-tokyo dot ac dot jp>]

On Mon, 22 Sep 2003 11:48:19 +0200 (MEST)
Alexander Gottwald <alexander.gottwald@s1999.tu-chemnitz.de> wrote:

> It seems the jp106 layout definition in the xkb sources is broken.
> Does "setxkbmap -v 10 jp -model jp" work for you?

It works well.  I get the correct keyboard except the
behaviour of Caps_Lock.  At first I press Caps_Lock, it
works correctly.  But the second time I press it, it does
not cancel the shift state.  So once I press Caps_Lock
accidentally, I can never get lower-case letters again.

I think the reason of this behaviour is that in Japanese
keyboards Caps_Lock is assigned to the combination of
"Shift + Eisu_toggle".  However, when I change the
definition in /usr/X11R6/lib/X11/xkb/symbols/jp from
key <CAPS> { [ Caps_Lock,Eisu_toggle]	};
to
key <CAPS> { [ Eisu_toggle,Caps_Lock]	};
I can't get the desired result.  Pressing Eisu_toggle
without Shift still works as Lock modifier.

The best situation is "it works as Caps_Lock only when
pressed with shift keys", but I don't know whether the
"cascade of modifiers" can be done.  If not, I prefer it
acts only as Eisu_toggle to prevent the accidental entry
into Lock state.  I have used my own .Xmodmap to achieve
this.  The attached is the patch for
/usr/X11R6/lib/X11/xkb/symbols/jp to get the same effect.

I'm not sure the topic is suitable for cygwin-xfree because
it seems not specific to cygwin but for XFree86 itself.
At least, the keyboard obtained by the setxkbmap is a big
progress.

Takuma Murakami (murakami@ipl.t.u-tokyo.ac.jp)

Attachment: jp.patch
Description: Binary data