Quoting Thomas Schmitt (scdbac...@gmx.net): > > But worse, the keyboard has gone american: > > Everybody should have a QWERTY and only use 7-bit ASCII. > This would promote international understanding, world peace, > and C programming.
:) > > The output from dumpkeys -f is very different in the two > > keycode 1 = Escape Escape VoidSymbol VoidSymbol VoidSymbol Meta_Escape > > keycode 1 = Escape Escape Escape Escape Escape Escape Escape Escape > > You must have deeply confused it. > Where did you get the syntax proposal of > ctrlr keycode 106 = 'Hello' Yes, that's incorrect. I have discovered that any errors in /etc/console-setup/remap.inc are severely punished with the bizarre behaviour. I can only assume that the perl script ckbcomp does not validate its input, and quite possibly has some mistakes. If the 'errors' are slight, then it does nothing, leaving dumpkeys -f unchanged. So, lines like these work: ctrll keycode 105 = F55 Control keycode 105 = F55 control keycode 105 = F55 but any one of these silently fails: Ctrll keycode 105 = F55 Ctrlr keycode 105 = F55 ctrlr keycode 105 = F55 which is ridiculous. > > > udevadm trigger --subsystem-match=input --action=change > > No. I reboot, > Ouch uptime. Once I got *something* to change, I found it was unnecessary to reboot all the time. I circulate the F55 numbers each time to ascertain whether anything changes or not. Rebooting gives you a few minutes break (to walk to another computer!) except when I got a recurrence of my binfmt_misc problem: I now have binfmt_misc in /etc/modules to make sure it gets loaded. This is the first time since adding the line that [ Cylon eye ] A start job is running for Enable support for additional executable binary formats (9min 52s / no limit) appeared and, as usual, I could ssh into the box, but ls -lR /proc/sys/fs hangs and even halt -f does nothing (after saying 'Halting.'). I do now remember to umount /home before I hard reset. Cheers, David.