Hello, today I had to write some "Merry Christmas" messages to my Esperantists friends and.. I found out that my Multi_key didn't work anymore! It is correctly set in /etc/default/keyboard (compose:rwin) -- and in fact it worked, I hadn't changed anything -- but xev revealed that rwin actually triggered: state 0x0, keycode 133 (keysym 0xffeb, Super_L), same_screen YES, instead of the expected: state 0x40, keycode 133 (keysym 0xff20, Multi_key), same_screen YES, After putting "keycode 133 = Multi_key" in ~/.xmodmap, and loading it, it started working again, and xev shows the correct key being input. I'm not sure whether this is the correct package to file this bug against, please reassign as appropriate. Dankegon, David
I don't think it's a bug in keyboard-configuration that its settings get overridden by your desktop's. Which is what I think happened here. Cheers, Julien
I'm using XFCE, and it doesn't have a setting for Multi_key. It just uses X's, AFAIK. Kindly, David
David Paleino <dapal@debian.org> (25/12/2012): What's your display manager, what's the output of “setxkbmap -print” without the xmodmap thing in Xfce, and what's the output of the same command in a bare X session (no X, no display manager, e.g. using: startx /usr/bin/xterm -- :1)? Mraw, KiBi.
reassign 696651 gdm3
found 696651 gdm3/3.6.1-2
thanks
Ah, there you go!
I used gdm3, and here's the output of setxkbmap -print:
xkb_keymap {
xkb_keycodes { include "evdev+aliases(qwerty)" };
xkb_types { include "complete" };
xkb_compat { include "complete" };
xkb_symbols { include "pc+it+us:2+inet(evdev)" };
xkb_geometry { include "pc(pc105)" };
};
Then I switched to lightdm, and there you go:
xkb_keymap {
xkb_keycodes { include "evdev+aliases(qwerty)" };
xkb_types { include "complete" };
xkb_compat { include "complete" };
xkb_symbols { include
"pc+it+inet(evdev)+level3(ralt_switch)+compose(lwin)+terminate(ctrl_alt_bksp)" };
xkb_geometry { include "pc(pc105)" };
};
A bare X session, using xterm, also gives this last one output.
I guess the culprit is gdm3/experimental (which I was using) here -- bug
reassigned.
Thanks for your help,
David