This bug has been reported to launchpad as: https://bugs.launchpad.net/bugs/354830 xfce4-settings-helper crashed with SIGSEGV in g_type_create_instance() Purely starting X. Actually I'm starting X using vncserver and it comes up with that error message literally as X loads and the window manger is loaded. Architecture: i386 DistroRelease: Ubuntu 9.04 ExecutablePath: /usr/bin/xfce4-settings-helper Package: xfce4-settings 4.6.0-1ubuntu2 ProcCmdline: xfce4-settings-helper ProcEnviron: LANG=en_US.UTF-8 SHELL=/bin/bash Signal: 11 SourcePackage: xfce4-settings StacktraceTop: ?? () g_type_create_instance () from /usr/lib/libgobject-2.0.so.0 ?? () from /usr/lib/libgobject-2.0.so.0 g_object_newv () from /usr/lib/libgobject-2.0.so.0 g_object_new_valist () from /usr/lib/libgobject-2.0.so.0 Title: xfce4-settings-helper crashed with SIGSEGV in g_type_create_instance() Uname: Linux 2.6.28-11-generic i686 UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare Stacktrace: http://launchpadlibrarian.net/24783142/Stacktrace.txt ThreadStacktrace: http://launchpadlibrarian.net/24783143/ThreadStacktrace.txt
I see this crash on Solaris with 4.6.2 and 4.8. Again, VNC appears to be the problem. On VNC, XInputExtension isn't present. The check for XI doesn't work: Core was generated by `xfce4-settings-helper'. Program terminated with signal 11, Segmentation fault. #0 0x08056a35 in xfce_pointers_helper_init (helper=0x808ef70) at pointers.c:139 139 if (version == NULL || !version->present) A little further checking indicates that version is not NULL, so it is hitting a SEGV on the !version->present check. My guess is that if the extension isn't present, the version struct returned is nonsense.
I think this was fixed in 01aa823b.