[SunRay-Users] Ubuntu 9.10 - 2 days to go

Jens Langner J.Langner at fzd.de
Wed Oct 28 18:36:43 EET 2009


Hi Bob,

Bob Doolittle schrieb:

[...]
> Actually given that we now know that this version of Ubuntu is using the
> completely-rewritten GDM 2.28, we know the main problem the fact that
> the rewrite doesn't have gdmdynamic any more, or for that matter *any*
> way to dynamically create a new display. There's no solution to this
> that can be offered within SRSS.
> 
> As Joerg said, Sun engineers have been contributing to the GDM source
> base to recreate a method for dynamic display management, but that will
> take time to integrate into the main GDM trunk (obviously we'll branch
> temporarily if necessary in OpenSolaris until the trunk catches up), and
> it won't be backwards compatible so SRSS will also have to change to
> integrate the new GDM. That'll happen in our next release. The simplest
> workaround for intrepid adventurers will be to create a wrapper, call it
> "gdmdynamic", and recreate the old interface while wrapping the new one,
> then install the gdmdynamic wrapper in the location where the old GDM
> provided it. That's what OpenSolaris plans to do when integrating the
> new GDM, in order to provide a smooth transition period while SRSS
> adapts to the new native interface. Particularly since the new interface
> keeps getting revised so SRSS has nothing stable we can count on yet (at
> least not so far, maybe this has finally changed). .

That's exactly how my colleagues and I are currently trying to get SRS5
running with Ubuntu 9.10. In fact, we almost have it running after a few
challanges with the latest consolekit and gdm patches that are available
under the following URLs:

https://bugzilla.gnome.org/show_bug.cgi?id=536355
https://bugs.freedesktop.org/show_bug.cgi?id=19333
http://lists.freedesktop.org/archives/consolekit/2009-October/000060.html

One of the problems we have still left here is that after terminating a
session (logout or Ctrl+Backspace+Backspace) the session is not properly
cleaned up and can still be listed with "utsession -l" and Xnewt will
not be restarted as expected. Thus the sunray ends up in 26D waiting for
Xnewt to reappear. But we are certain that we will sort out these last
bits during the rest of the week.

BTW: During our work we have also found a small but important bug in the
"/opt/SUNWut/lib/utxsun" script:

Near line 76 you will find the following statement:

-- cut here --
        if [ x$FONTPATH != x ] ; then
                XMOREOPTS=" -fp $FONTPATH"
        fi
-- cut here --

However, looking at line 55 where XMOREOPTS is already set to
XMOREOPTS="$XMOREOPTS +bs -terminate" the above addition of the fontpath
is incorrect. Line 76 have to look like that:

-- cut here --
        if [ x$FONTPATH != x ] ; then
                XMOREOPTS="$XMOREOPTS -fp $FONTPATH"
        fi
-- cut here --


Or otherwise Xnewt will never be terminated properly because the
"-terminate" and "+bs" options will be wiped by the fontpath addition.

Perhaps you can manage to get that fixed before SRS5 will be finally
released.

regards,
jens
-- 
Jens Langner                                         Ph: +49-351-2602757
Forschungszentrum Dresden-Rossendorf e.V.
Institute of Radiopharmacy - PET Center                 J.Langner at fzd.de
Germany                                               http://www.fzd.de/


More information about the SunRay-Users mailing list