gdm, RELENG_6 update

[Jürgen Dankoweit]

On Sun, 01 Jan 2006 15:56:02 +0100
Mark Nowiasz <buckaroo at gmx.de> wrote:

> Jürgen Dankoweit wrote:
> 
> 
> > Yes. I Have fixed the problem for by doing the following things:
> > (1) Disable starting gdm from /etc/rc.conf
> > (2) Rewriting the start script into the old startup style
> > (3) renaming /usr/X11R6/etc/rc.d/gdm.sh to
> > /usr/X11R6/etc/rc.d/xxxxxserver.sh
> > 
> > This solved the problem with gdm that I have since FBSD 5.3 stable
> 
> Hmm - you mean changing /etc/ttys to
> 
>         ttyv8   "/usr/X11R6/bin/gdm -nodaemon"  xterm   on secure
> 
> ?

no, no, no. I mean renaming an d rewqriting the start script in
/usr/X11R6/etc/rc.d! 
Like:

PREFIX=/usr/X11R6

case "$1" in
        start)
                export LANG=de
                export LANGUAGE=de
                export LC_TYPE=de_DE.8859-15
                export LC_ALL=de_DE.8859-15
                ${PREFIX}/bin/gdm
                ;;
        stop)
                /usr/bin/killall -m gdm 2>/dev/null
                ;;
        *)
                echo "Usage: `basename $0` start | stop"
                exit 64
                ;;
esac

exit 0

> 
> I still don't understand why gdm is racing against initd *after*
> updating the world (gdm worked fine before the update) and why the
> changed gdm.conf (or the factory conf) doesn't work - could it be that
> there's a slight difference between the amd64 world and the i386
> world?

The scripts are started in a special order (rcorder). They wait on
special conditions a certain time. If this time is over the script runs
without checking the conditions. That's my interpretation of the whole
problem.

> 
> Happy new year everyone,
> Mark

Same to you

> -- 
> Be careful of reading health books, you might die of a misprint. 
> 		-- Mark Twain
> 

That's true :-)