Suse 7.1 kernel 2.4.17 kde 2.2.2 XF86-4.2.0 I have had frequent similar problems. I find myself continually logging out and back in to correct this. Restarting klauncher has no effect. I cannot find any running processes that may be causing a problem. Can anyone explain how DCOP and klauncher work. Is there anything I can do to make them solid ? Brian Marr On Tuesday 16 April 2002 09:02, you wrote:

Hello SuSE folkz, When I try to launch any of the programs on KDE desctop and the panel, I'm getting a persistent error message: "KLauncher can not be reached via DCOP" Applications just won't start. I already did dcop and kde related directories and files removal from /tmp, /root, /root/.kde2 and restarted KDE. It didn't solve the problem. Could somebody tell me please how to fix it or where to find a cure for this problem. Thank you in advance. Alex

Hi With sux it works. Jaska Viestissä Tiistai 16. Huhtikuuta 2002 05:04, Jon Doe kirjoitti:

On Tue, 16 Apr 2002 11:52:31 +0930

Brian Marr wrote:

...

Suse 7.1 kernel 2.4.17 kde 2.2.2 XF86-4.2.0

I have had frequent similar problems. I find myself continually logging out and back in to correct this. Restarting klauncher has no effect. I cannot find any running processes that may be causing a problem. Can anyone explain how DCOP and klauncher work. Is there anything I can do to make them solid ?

Brian Marr

I have noticed that this happens everytime I run a KDE app as su.

On Tue, 16 Apr 2002 09:03:54 +0300 jaakko tamminen wrote:

Hi

With sux it works.

Jaska

Actually sux is what I use, then after once using it, the rest of the session I can't open any KDE app as user or as root till I log out then back in.

::Get the drivers, install them with "rpm -Uhv --nodeps --force", and if the ::links in /usr/lib ( /usr/lib/libGL* ) don't point to the nvidia glx ::sharedlibs, then run switch2nvidia_glx. Make sure again that the libs aren't ::pointing to some mesa lib... I don't have much of an issue with the instructions below, but if you want to make sure everything is kosher then I would do this. Get the 2880 tar.gz files from http://www.nvidia.com/drivers then put them in /usr/src/packages/SOURCES. You should then just untar them like this... tar -zxvf <tar file name> cd into each directory that is made from untarring the files and just type make. The GLX make script just moves the binary files to the places they should be and makes the symlinks. You not actually compiling anything. If you do this the GLX binaries will be put in the correct place and you should avoid using the "switch2nvidia_glx" because with the latest drivers (2880) it screws up everything horribly. The switch2 script hasn't been needed for 2 revisions of the nVidia driver and has just been a source of issue. As for the kernel module if you just type make in the untarred directory then it will link the kernel module with your current running kernel. You may be running a different kernel then the default and I believe they are made for certain kernels when you get the RPM. Note that the driver will end up in /lib/modules/2.4.16-4GB/kernel/drivers/video/ It will be called NVdriver. I rename it each time to NVdriver.o and put it in my initrd which seems to work pretty well. You would have to put it in /etc/rc.config and rerun mk_initrd and then lilo. -- INITRD_MODULES="reiserfs NVdriver" -- The above is what I have in /etc/rc.config. Cheers! ::Did you install the latest SaX2 patches ? No ? Do so. You want to :-) :: ::Run SaX2. Don't let it tell you what to do. Select your mouse ( you may have ::to use the cursor keys if it manages to screw the mouse ). Do so, and select ::everything up to the graphics card. :: ::When you are at the graphics card, select the card, and choose 'properties'. :: ::You want the 'nvidia' driver, not the 'nv' driver. :: ::Also you may want glx. ( found under 3d ) :: ::'dri' does not work on nvidia cards. :: ::Make sure you don't just select a monitor. Make quite sure it's frequencies ::are correct. You'll find it's min/max frequencies in its manual. :: ::When SaX2 is done, save. JUST SAVE. Do NOT let SaX2 start a test X server. :: ::SaX2 somehow manages *sometimes* to trash the XF86config file if it's not all ::right. :: ::copy the /etc/X11/XF86config file. Copy it to your home directory. Be safe. :: ::do 'startx mwm' :: ::if you see a gray background, a "X" and a pale-blue menu when the left ::mouse-button is pressed, you successfully started the mwm windowmanager. ( ::any windowmanager does ok for testing purposes. mwm or twm are the lightest ::ones ) :: ::ctrl-alt-backspace kills the wm. :: ::Now you can continue as normal :) :: ::'gears' ( while in a wm/xterm ) gives you the glx throughput. :: ::If gears don't run, the libs are still wrong. Check them again. -=Ben --=====-----=====-- mailto:ben@whack.org --=====-- "I've never been quarantined. But the more I look around the more I think it might not be a bad thing." -GC --=====-----=====--

* John (jfmurphy@charter.net) [020415 21:09]: :: ::Guys thanks for specific instructions on how to approach this. With ::Suse 8.0 in another week is it likely that Suse will recognize this ::card. Or should I get use to this procedure because of the nVidia ::license issue. I guess I'm trying to put off the inevitable, trouble ::is my gray hair count is getting high and was hoping to save some. SuSE 8.0 really has nothing to do with this nVidia issue. SuSE will most likely not ship the binary only drivers with 8.0 there will be a redirector in YOU to go get them. They don't want the responsiblity of providing support for something that they can't control because they don't get source to them..no one does. I would give the instructions a try. With the instructions so far..you shouldn't have much of an issue. If you do. Please email the list again and we can help you. -=Ben --=====-----=====-- mailto:ben@whack.org --=====-- "I've never been quarantined. But the more I look around the more I think it might not be a bad thing." -GC --=====-----=====--