GNOME Bugzilla – Bug 534995
incessant failures running gtk apps without gnome desktop
Last modified: 2014-03-28 13:28:04 UTC
Steps to reproduce: Virtually every time I run any gtk app over an ssh forwarded connection to my display which is NOT running gnome desktop (I've got a minimal fvwm session up), the first time I run the app, it always crashes like so: The program 'synaptic' received an X Window System error. This probably reflects a bug in the program. The error was 'BadAccess (attempt to access private resource denied)'. (Details: serial 442 error_code 10 request_code 148 minor_code 5) (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.) I then run the app a second time, and it works fine. I have this deep suspicion that the gtk toolkit is expecting something to be already setup by the gnome desktop environment (what, I have no idea), and its fallback code does the setup, but can't keep the original app from crashing. Then when I run it a second time, the setup has been done by the failed app, and everything works (it is only a theory :-). Whatever it is only affects gtk apps (no problem running Qt apps, or simple X11 progs like xlogo and such), and the first gtk app that crashes seems to fix everything for all subsequent gtk apps. I file this report in the hope that someone will be curious enough to fix it, but for the forseeable future, I guess I'll just run all gtk apps twice the first time. This behavior spans a vast number of different linux distros. The above message comes from trying to run synaptic on an ubuntu 8.04 system, but I see similar behavior with things like gedit on fedora 8 and 9, etc. The common thread in all cases is a gtk app and a non-gnome session at the other end of the ssh forwarded connection (using -Y or -X on the ssh command makes no difference). Stack trace: Other information:
Have you done what the error message asks you to do ?
I spent quite a while trying to find the ubuntu debuginfo files so I could try and get a backtrace, but had no luck. I'm gonna try again when I have time on a fedora box (I may already have one with most of the debuginfo files already loaded). In my investigations, however, I discovered a new fact: It doesn't appear to be merely the first time I run a gtk app over ssh, it is the first time I run one after booting the remote system. Apparently whatever state changes is something that happens on the remote system, not on my X server. This makes reproducing the problem a pain, since I always have to reboot the remote system to make it happen again (of course, most of them are virtual Xen machines, but it happens with real remote machines as well).
Found debuginfo, and a new fact: This only happens when the remote machine is set to boot a runlevel 3 (no gdm), if it boots runlevel 5 (gdm starts), then the 1st time aborts don't happen. Here's a similar error from running gedit on a fedora 8 virtual machine with most of the debuginfo files loaded so I can get a backtrace. Gdk-ERROR **: The program 'gedit' received an X Window System error. This probably reflects a bug in the program. The error was 'BadAccess (attempt to access private resource denied)'. (Details: serial 184 error_code 10 request_code 148 minor_code 5) (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.) aborting... Aborted Running again after a reboot and using gdb this time: Breakpoint 2, gdk_x_error (display=0x8709600, error=0xbfa71278) at gdkmain-x11.c:614 614 if (error->error_code) (gdb) bt
+ Trace 198753
I can confirm this problem, but haven't investigated yet. One of my virtual machines is an ubuntu with gnome but no xserver or gdm, I run a few apps on it but only on remote X via ssh. The first time I try to run any gtk (gnome) program it aborts with the above error. After trying again everything is fine.
Looks like XShmCreatePixmap throws some unexpected errors, instead of returning NULL ?
This backtrace is quite old. Does anyone have a more recent one with perhaps a simple application like one of the demos?
Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for. Thanks!