On Tue, Feb 24, 2004 at 09:51:10AM +0100, Falk Hueffner wrote: > Branden Robinson <branden@debian.org> writes: > > > It may be necessary to attach gdb to the XFree86-debug server from a > > remote terminal to locate the source of the problem. > > > > Unless you can ulimit -c unlimited and get a good backtrace off the > > core from XFree86-debug? > > Given that it kills the kernel, both methods don't seem very viable... > the best I could try is redirect the log to the networks somehow and > see what occurs last. Do you have any recommendation how to do so? In > case the console output of the X process contains everything > /var/log/XFree.log contains, I could just use ssh. Falk and I discussed this on IRC. I suggested he pass "-verbose 3" to the X server so that everything that would go to the log file goes to stdout as well. That may identify the module that causes the problem, and an appropriate breakpoint can be set within one of that module's functions to trace the server (remotely) until it dies. -- G. Branden Robinson | The basic test of freedom is Debian GNU/Linux | perhaps less in what we are free to branden@debian.org | do than in what we are free not to http://people.debian.org/~branden/ | do. -- Eric Hoffer
Attachment:
signature.asc
Description: Digital signature