Description of problem: Virt-manager crashs when I try to open the console window of a VM by double clicking or selecting "open". Using virt-manager's --no-fork option I get the following error message: The program 'virt-manager.py' received an X Window System error. This probably reflects a bug in the program. The error was 'BadLength (poly request too large or internal Xlib length erro'. (Details: serial 4317 error_code 16 request_code 144 minor_code 17) (Note to programmers: normally, X errors are reported asynchronously; that is, you will receive the error a while after causing it. To debug your program, run it with the --sync command line option to change this behavior. You can then get a meaningful backtrace from your debugger if you break on the gdk_x_error() function.) Starting virt-manager with --sync as recommended in this message doesn't work. This is what happens in a German localized desktop environment. While investigating the problem I tried to start virt-manager with LANG=C set and it didn't crash! But I got different error messages: (virt-manager.py:3352): GtkGLExt-CRITICAL **: gtk_widget_set_gl_capability: assertion `GDK_IS_GL_CONFIG (glconfig)' failed (virt-manager.py:3352): gtk-vnc-WARNING **: Could not enable OpenGL /usr/share/virt-manager/virtManager/console.py:77: Warning: g_object_unref: assertion `G_IS_OBJECT (object)' failed self.vncViewer = gtkvnc.Display() Should I file another bug for gtk-vnc? Nonetheless virt-manager should not crash with the German localization. Version-Release number of selected component (if applicable): 0.5.4-4.fc9.x86_64 How reproducible: Always with LANG=de_DE.UTF-8 Steps to Reproduce: 1. Start virt-manager using "LANG=de_DE.UTF-8 virt-manager --no-fork" 2. Double click on a VM 3. See virt-manager crashing Actual results: Virt-manager crashs if "LANG" is not set to "C" Expected results: No crashs for your international friends ;-)
I'm seeing this too, with LANG en_GB.utf8. Seems to be more of a gtk-vnc problem, as I can connect with the old VNC client.
(vinagre:4624): GtkGLExt-CRITICAL **: gtk_widget_get_gl_window: assertion `GTK_WIDGET_REALIZED (widget)' failed (vinagre:4624): GtkGLExt-CRITICAL **: gtk_widget_get_gl_context: assertion `GTK_WIDGET_REALIZED (widget)' failed (vinagre:4624): GdkGLExt-CRITICAL **: gdk_window_is_gl_capable: assertion `GDK_IS_WINDOW (window)' failed (vinagre:4624): GdkGLExt-CRITICAL **: gdk_window_set_gl_capability: assertion `GDK_IS_WINDOW (window)' failed (vinagre:4624): GtkGLExt-WARNING **: cannot set OpenGL-capability to widget->window ** Gtk:ERROR:gtkwidget.c:7934:gtk_widget_real_map: assertion failed: (GTK_WIDGET_REALIZED (widget)) Could be a gtkglext problem, rather than gtk-vnc?
(In reply to comment #2) > Could be a gtkglext problem, rather than gtk-vnc? Could well be. We are seeing two bugs here. This bug report is primarily about virt-manager being unable to handle that (other) error gracefully in a localized environment. I don't know against which package the other bug should be filed. I just suspected gtk-vnc and hoped the maintainer could help, as he seems to be the maintainer and an upstream developer of both...
Just stick with one bug report for now, until we can definitely say they are unrelated - its quite possible one error leads to another. I've had a couple of reports of the problem in comment #2, but they are fairly non-reproducable. With latest rawhide install the original reporter was no longer able to see those assertion failures.
*** Bug 452640 has been marked as a duplicate of this bug. ***
Okay, I was seeing a similar issue. As root, over ssh -X, open virt-manager and try to view the console of a running VM: # rpm -q gtk-vnc gtk-vnc-0.3.6-4.fc10.x86_64 # virt-manager --no-fork (virt-manager.py:2365): GtkGLExt-CRITICAL **: gtk_widget_get_gl_window: assertion `GTK_WIDGET_REALIZED (widget)' failed (virt-manager.py:2365): GtkGLExt-CRITICAL **: gtk_widget_get_gl_context: assertion `GTK_WIDGET_REALIZED (widget)' failed (virt-manager.py:2365): GdkGLExt-CRITICAL **: gdk_window_is_gl_capable: assertion `GDK_IS_WINDOW (window)' failed (virt-manager.py:2365): GdkGLExt-CRITICAL **: gdk_window_set_gl_capability: assertion `GDK_IS_WINDOW (window)' failed (virt-manager.py:2365): GtkGLExt-WARNING **: cannot set OpenGL-capability to widget->window /usr/share/virt-manager/virt-manager.py:304: GtkWarning: gdk_window_set_cursor: assertion `GDK_IS_WINDOW (window)' failed gtk.main() /usr/share/virt-manager/virt-manager.py:304: GtkWarning: gdk_gc_new: assertion `drawable != NULL' failed gtk.main() /usr/share/virt-manager/virt-manager.py:304: GtkWarning: gdk_drawable_get_visual: assertion `GDK_IS_DRAWABLE (drawable)' failed gtk.main() /usr/share/virt-manager/virt-manager.py:304: GtkWarning: gdk_visual_get_screen: assertion `GDK_IS_VISUAL (visual)' failed gtk.main() /usr/share/virt-manager/virt-manager.py:304: GtkWarning: _gdk_image_new_for_depth: assertion `visual || depth != -1' failed gtk.main() Segmentation fault I then pulled down gtk-vnc 0.3.7 via koji and this problem was fixed: I could connect to the VM no problem. If there was a gtk-vnc bug tracking this issue that someone knows of, please close this as a dup of that.
I can confirm that this crash doesn't occur with gtk-vnc-0.3.7. I fetched the SRPM from koji and rebuilt it for F9 using mock. I'll request an update of gtk-vnc for F9 and let it block this bug. Both bugs can be closed when the update hits bodhi.
gtk-vnc-0.3.7-1.fc9 has been pushed to the Fedora 9 stable repository and fixes this bug.