On Thu, Jun 17, 2004 at 05:41:33PM +0100, Thomas Leonard wrote:
> On Sun, Jun 13, 2004 at 06:33:14PM +0200, "Daniel P=C3=B6rsch" wrote:
> > Hi devs,
> > since upgrading to new gtk-version (2.4.0) rox sometimes crashes when=
> > start to drag a file, e.g. for setting a new icon. It doesn't happen =
> > time. The error log tells the following:
> > Sun Jun 13 14:37:30 2004
> > The program 'ROX-Filer' received an X Window System error.
> > This probably reflects a bug in the program.
> > The error was 'BadWindow (invalid Window parameter)'.
> > (Details: serial 73382 error_code 3 request_code 3 minor_code 0)
Request code 3 is XGetWindowAttributes. ROX-Filer only calls that in one
place (tasklist.c), and it's protected by error_trap_push/pop calls. So,
the error is probably in GTK.
> > (Note to programmers: normally, X errors are reported asynchronousl=
> > that is, you will receive the error a while after causing it.
However, XGetWindowAttributes is synchronous, so the backtrace should
still be useful here even if running without --sync. If GTK is compiled
with debugging enabled, it should abort/coredump here rather than
exiting, which would provide more useful information...
> > 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() fun=
> Several reports of crashes like this, all with GTK 2.4. Could someone f=
> whom it happens reliably try the suggestion in the error message? Eg:
> $ rox --debug
> (gdb) run -n --sync
> (press Ctrl-C once ROX-Filer has started)
> (gdb) break gdk_x_error
> (if that doesn't work, try this:
> (gdb) p XSetErrorHandler(abort)
> Resume (continue) the filer with:
> (gdb) c
> Then do something to make it crash. When you get back to the (gdb) prom=
> ask for a backtrace (and send it to us!):
> (gdb) bt
> If you get "Program received signal SIG35, Real-time event 35.", this i=
> due to a directory changing rather than a bug. Use 'c' to continue
> These instructions are somewhat untested, as it won't crash reliably fo=
> me... let us know if you have problems.
Thomas Leonard http://rox.sourceforge.net
tal00r at ecs.soton.ac.uk tal197 at users.sourceforge.net
GPG: 9242 9807 C985 3C07 44A6 8B9A AE07 8280 59A5 3CC1