On Wed, Jul 15, 2009 at 07:26:59AM -0700, Alan Coopersmith wrote:
> Dan McDonald wrote:
> > Background:  I don't use gdm.  I type "xinit" after I get a console shell.  
> > I have a Toshiba Portege R600, as sold with OpenSolaris built-in.  I use 
> > IPS.  I have an external keyboard  (old MacAlly iKey, circa 1999) and use 
> > an external 1920x1200 display for working at home.
> > 
> > In build 111b, everything worked.  I had to insert a symlink for this file:
> > 
> > /usr/X11/lib/X11/xkb/rules/base
> > 
> > but after that, things worked.
> > 
> > Now that we're up to 118, things broke again.  My external USB keyboard had 
> > its alphabetic characters mapped to Arabic, according to xev.  I managed to 
> > work around this by replacing the new version of "base" above with the one  
> > from build 111b (which is actually "xorg" from 111b).
> 
> Strange - what does the /var/log/Xorg.0.log show the keyboard
> layout is detected as and trying to auto-load?

I'll have to revert the "base" file and try again.  Stay tuned.

> > So the first breakage (remapping after entering X) is worked around by 
> > using 111b's "base" file.
> > 
> > The SECOND breakage is more alarming.  After I exit X, all keyboards start 
> > generating complete and utter garbage.  This includes the built-in laptop 
> > keyboard, for crying out loud.
> 
> Try running kbd_mode -a to reset to the input mode the kernel
> console expects.   The X server kbd driver should be doing
> this on exit - normally the manual kbd_mode is only needed if
> X crashes, and only if you run xinit directly instead of the
> more user-friendly "startx" wrapper script that makes sure
> kbd_mode is called after X exits.

Hmmm... startx is a wrapper to xinit, so I take it the XINITRC environment
variable would be honored.  This is good to know.

I'll be back with that Xorg.0.log in a bit.

Thanks!
Dan

Reply via email to