Red Hat Bugzilla – Bug 100686
keyboard not configured in firstboot
Last modified: 2007-11-30 17:10:31 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 Galeon/1.2.7 (X11; Linux i686; U;) Gecko/20030131
Description of problem:
The keyboard is not configured in firstboot.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Do a fresh install of RedHat Linux and select an azerty keyboard for example.
Actual Results: Querty keyboard in firstboot.
Expected Results: keyboard configured as selected in anaconda.
Does the keyboard work at all or is it just the wrong keymap?
When i hit the 'a' key i see a 'q' char on the screen.
The keyboard work. It should be a wrong keymap.
If you boot with 'nogui', and firstboot is run *that* way, does it work?
Fresh installation of severn. Append "nogui" to grub with anaconda installer. At
the first boot, firstboot program is launched and there is no keyboard problem.
I expected as much. When rhgb is running, firstboot inherits rhgb's X server;
howerver, XKB doesn't work right with rhgb's X server, due to the read-only file
system, or somesuch.
Changing component to rhgb.
*** Bug 101220 has been marked as a duplicate of this bug. ***
*** Bug 105659 has been marked as a duplicate of this bug. ***
*** Bug 107187 has been marked as a duplicate of this bug. ***
*** Bug 107219 has been marked as a duplicate of this bug. ***
*** Bug 107209 has been marked as a duplicate of this bug. ***
*** Bug 107232 has been marked as a duplicate of this bug. ***
Also happens with Fedora Core test 3.
*** Bug 107277 has been marked as a duplicate of this bug. ***
I have a tentative fix for this. Can people try:
scratch that. That's not going to work quite correctly. I'll get another test
I am pretty sure I have this fixed in the latest rhgb/initscripts.
*** Bug 122302 has been marked as a duplicate of this bug. ***
Reopen the bug.
appear again in FC2T3.
I've the same problem with an azerty keyboard. When you add some user
to the system, the keyboard is in qwerty. version FC2 Test3.
The problem also appears in FC2 final.
This bug is especially problematic on the Add User screen. Full names
with non-ASCII characters cannot be entered correctly, and passwords
set may turn out to be wrong.
Because of the latter problem, which may cause an user to essentially
lock himself or herself out of the system, since the password set
won't be accepted later, I'm raising the severity to High.
I guess this bug give me the following behaviour (FC2 final with rhgb) :
I can't log within GDM if my user password is for instance "qwerty"
and the keyboard is supposed/known to be an "azerty" one....
In fact, you don't see what you type (*****) and I tested my keyboard
in the user/login field and this shown me that the keyboard is NOT the
"azerty" one before Gnome is fully started.
Once logged (after converting the passwords....), the "azerty"
keyboard is used : OK
So, this is definitivly "High" for users outside US.
*** Bug 121931 has been marked as a duplicate of this bug. ***
*** Bug 123940 has been marked as a duplicate of this bug. ***
The bug still exists in FC3 Test1.
Okay, confirmed, reproduced ...
I knew I shouldn't have trashed my last azerty keyboard a few years
As Bill rightly guessed in #5 the keyboard bindings set in the
X11 configuration file won't work because the keymap is generated
at stratup and with a read-only file-system well it just doesn't
work. I implemented a workaround which resets the X11 keymap
dynamically when firstboot is about to be run, and it seems to work :-)
*** Bug 121951 has been marked as a duplicate of this bug. ***
Version 0.12.6 was pushed on Rawhide and should fix this bug,
*** Bug 133952 has been marked as a duplicate of this bug. ***
Fedora Core 3 has rhgb-0.14.1-1.i386.rpm and the bug is still there.
( I selected Slovenian layout in anaconda, but firstboot used US ( or
similar ) )
... and adding myself to CC ( sorry for spam )
it works for other languages. What do you have in
Also provide the subsection of your /etc/X11/xorg.conf
defining the keyboard.
This mean the lookup for *both* methods of Slovenian layout failed !
I can't do that before tomorrow.
If it helps : At install I selected english install language,
Slovenian keyboard layout ( worked at once both in X and textual
console ) and some laguages for install ( english , Slovenian ,
Croatian, Hungarian and some more ).
After install in X the keyboard is Slovenian ( works OK ).
Option "XkbModel" "pc105"
Option "XkbLayout" "si"
How is rhgb supposed to understand that "slovene" means "si" ?
That's why it broke, for a french configured keyboard you get
something based on the ISO country name like "fr" or "fr_FR".
I think the generation of the KEYTABLE value is wrong, potentially
an anaconda bug... Can you reopen a new bug pointing to this,
addressing anaconda, with me in CarbonCopy ? I don't feel reassigning
this big bug history and list is the right way to handle that specific
I opened bug 155554 , Daniel.
So, I see that the associated bug 155554 is closed rawhide. Is this bug fixed,
(In reply to comment #38)
> So, I see that the associated bug 155554 is closed rawhide. Is this bug fixed,
Yes for me.