1 |
Hi |
2 |
|
3 |
This release fix the "dead keys with xdm/gdm/kdm" bug that we |
4 |
have been having. It replaces /etc/init.d/xdm and |
5 |
/etc/inittab, so do NOT replace these with older versions! |
6 |
|
7 |
The cause of this bug, is not baselayout-1.7.2, but rather |
8 |
the new order it starts scripts in, enhanced this, as |
9 |
xdm/gdm/kdm had more time to startup than before. The full |
10 |
reason for this bug is (sorry for paste): |
11 |
|
12 |
----------------cut here------------------- |
13 |
Any Display manager (gdm,kdm,xdm) have the following problem: |
14 |
if it is started before any getty, and no vt is specified, it |
15 |
will usually run on vt2. When the getty on vt2 then starts, |
16 |
and the DM is already started, the getty will take control of |
17 |
the keyboard, leaving us with a "dead" keyboard. |
18 |
|
19 |
Resolution: add the following line to /etc/inittab |
20 |
|
21 |
x:3:respawn:/etc/X11/startDM.sh |
22 |
|
23 |
and have /etc/X11/startDM.sh start the DM in -nodaemon mode |
24 |
if a lock is present, else just fall through (with a big |
25 |
enouth delay so that init will not "detect" a respawning |
26 |
process and halt if for some odd time). |
27 |
|
28 |
We then use /etc/init.d/xdm to control the presence of this |
29 |
lock, and this should ensure that the getty's is always |
30 |
started *before* the DM. |
31 |
--------------------cut here------------------------------- |
32 |
|
33 |
Support for the new sys-apps/kbd package has been added as |
34 |
well. |
35 |
|
36 |
Another issue that I still need to adress, is getting proper |
37 |
docs up for the initsystem. I have started, but is not |
38 |
finished yet, and will try to get it out as soon as time |
39 |
permits. |
40 |
|
41 |
As usual, check the ChangeLog for details. |
42 |
|
43 |
|
44 |
Greetings, |
45 |
|
46 |
-- |
47 |
|
48 |
Martin Schlemmer |
49 |
Gentoo Linux Developer, Desktop Team Developer |
50 |
Cape Town, South Africa |