!!! Sawfish/Sawfish-gnome 

  

 If you do not get a drop-down menu when you click on a window's menu button, can't change the number of workspaces, and/or gnome-control-center says "the applet encountered an error" when you try to configure sawfish, it means that you have versions of sawfish and rep (and/or rep-gtk[[-gnome]?) that don't like each other. For example, you have sawfish from debian woody (stable) but a rep from debian testing or unstable, even though these packages don't have an official conflict. Other symptoms include lines like "No such file or directory, sawfish/client" from the command line or in your .xsession-errors file. 

  

  

 !!! GNOME 2.x - various sound applications 

  

  GStreamer-ERROR **: No default scheduler name - do you have a registry ? 

  

 Note - this was for the gnome2.2 back port for [Debian] 3.0 Woody. Other versions/distributions probably have different dependencies/package names. 

  

 This happened again (much later) while running Debian Testing. The answer is to 

 on package installation. 

  

  

 !!! tar(1) 

 After doing these two steps, gdm behaved properly when started from /etc/init.d/gdm. If you determine how to fix it, please edit this page! 

  

 __2004-05-14 9:43 EST__ The correct way to deal with this problem is: 

 * Enable "debug" in your /etc/gdm/gdm.conf file under: 

  

 !# This will enable debugging into the syslog, usually not neccessary and it creates a LOT of spew of random stuff to the syslog. However it can be useful in determining when something is going very wrong. 

  

  Enable=true 

  

 Then your messages will be logged to /var/log/syslog. 

  

 * GDM has to always restart X with an entry like the following in gdm.conf: 

  

 * and finally, make sure that you have the following under the same daemon section: 

 !# Automatic VT allocation. Right now only works on Linux. This way we force X to use specific vts. turn VTAllocation to false if this is causing problems. 

  

  VTAllocation=false 

  

 !# Should double login be treated with a warning (and possibility to change vts on linux systems for console logins) 

  

  

 That should be enough to at least know what the source of the problem is. Please note that the latest Debian gdm as of this writing, has issues with [UDev]. So if you have udev generating __/dev/vc*__ block files for you (and removing them when you close X), gdm can't get a hold on running X servers and it let's them running, trying to launch another one in a different vts! 

 Therefore, you will have a few servers running in vt6, vt7, vt8, ..., vtN. The way to "kill them" without having to login in a console (vty1) is to "__CTRL+ALT+FN__", where N is the number of the vt. Say, for vt8: __CTRL+ALT+F8__. Once you are there, if you see an X server running (you will be able to move the mouse, but no other application might be running), hit __CTRL+ALT+BACKSPACE__ to kill it. Do the same with the others you may find. 

  

 Hopefully somebody else might be able to write a better solution for this ;-). Or I'll come back once I find a more concrete answer. 

 If you are the kind who avoids typing, you might want to make a copy of the __factory-gdm.conf__ file, found under __/etc/gdm__. 

  

  

 ---- 

 CategoryErrors