Bug 15706

This defect is considered MUST-FIX for Winston Release-Candidate #1

Fixed in RC1. Thanks for noticing this one.

Sorry, wrong again. This time it becomes de-latin1 instead of
de-latin1-nodeadkeays.
Please, do check this before final release. Seems to be a common point of
mistake ...

I think I can do this myself, saving Glen some keystrokes ... ;)

This defect is considered MUST-FIX for Winston Gold.

I can't reproduce this. I've tried de, de-latin1, and de-latin1-nodeadkeys and
it seems to work perfectly.

Chose it and got it-ibm both on upgrade 6.2->RC1 and on install RC1

*** Bug 16678 has been marked as a duplicate of this bug. ***

What kind of install is this?  NFS?  CD-ROM?  Did you do keyboard selection from
the text part of the install?  Or GUI?

Still there in RC2. GUI CD Installation. 102 PC keyboard - german - disable dead
keys.
(text NFS installation to follow today.)

Are you selecting the keymap in a text window or a GUI one?

*dazed and confused*
Hm, above I wrote that it's a GUI install. Don't really understand that question
as I think this has already been answered. *dazed and confused*

Text CD install works, de-latin1-nodeadkeys is configured as desired.

Where are you picking the keyboard it doesn't work? X-Windows or not? If in X,
what
keyboard are you choosing? If not, what keyboard are you choosing? Do you ever
ask answer any installations questions in X?

I know this is a long list :-(

Ahh, the problem is when you pick de and activate dead keys in GUI that the
wrong keymap is selected in text. Matt's problem :-)

GUI CDROM install; select "German" keyboard and choose "Activaeren dead-keys"
(sic), post install the /etc/sysconfig/keyboard looks like:

KEYBOARDTYPE="pc"
KEYTABLE="de-latin1"

instead of having:

KEYTABLE="de-latin1nodeadkeys"

Fixing this problem would require a good bit of extra code -- it would
completely revamp the mapping of X keymaps back to kernel keymaps.  Remind me
next release.

Does this mean that those with a german keyboard have to accept that dead keys
are active even if they choose nodeadkeys? Well, I have to accept this, but this
is a _major_ bug. Using german keyboards this is _very_ annoying (well, I'm
using an english SGI keyboard, so no problem with me ... ). Hm, if I remember
correctly this problem was not present in RHAT 6.2. So I wonder why a major
rewrite is necessary now (again, it doesn't affect me - not even my users as my
kickstart disks can easily catch this problem).

(And yes, I will definitely remind you ... :) ).

I can't see how this problem didn't exist in 6.2

This is fixed in CVS.  That is, selecting a German keyboard with dead keys
disabled now generates a correct /etc/sysconfig/keyboard file.  However, there
is a related bug, 16678, that has the following attributes:  In text mode,
selecting de-latin1-nodeadkeys generates a correct /etc/sysconfig/keyboard file
but an incorrect XF86Config file.

using qa1211.0 test tree:

verified /etc/sysconfig/keyboard contents:

GUI install, stage1 TUI keyboard selection:

KEYBOARDTYPE="pc"
KEYTABLE="de-latin1-nodeadkeys"

GUI install, stage1 GUI keyboard selection:

post-install traceback:

Traceback (innermost last):
  File "/var/tmp/anaconda-7.1//usr/lib/anaconda/iw/progress_gui.py", line 19, in
run
    rc = self.todo.doInstall ()
  File "/var/tmp/anaconda-7.1//usr/lib/anaconda/todo.py", line 1693, in
doInstall
    self.x.write (self.instPath + "/etc/X11")
  File "/var/tmp/anaconda-7.1//usr/lib/anaconda/xf86config.py", line 959, in
write
    config4 = self.Version4Config ()
  File "/var/tmp/anaconda-7.1//usr/lib/anaconda/xf86config.py", line 1241, in
Version4Config
    info["enableVariant"] = ""
NameError: info

*** Bug 18385 has been marked as a duplicate of this bug. ***

note that the traceback above holds up this issue being fixed, as post-install
the /etc/sysconfig/keyboard values are:

KEYBOARDTYPE="pc"
KEYTABLE="de-latin1-nodeadkeys"

So Brock, should this bug be closed now or what?