Re: [Gnash] Xorg signall 11

[John Gilmore]

> >> Hello, I'm having a few problems with gnash killing my X server.

I hope you've filed a bug against the X server.  Other programs should
never be able to make the X server crash.  So there's a bug in X, and
possibly a second bug in gnash that triggers the X bug.  (Or perhaps
gnash is doing things right, and all will be well when X is fixed.)

The hard part will be narrowing down a test case that the X folks can
reproduce on their own AMD64 (without your particular version of
Mozilla and all its support libraries, gnash and all its support
libraries, etc).

Does it help to use -synchronous as a way to get xlib to avoid
buffering its X operations?  I think this would have to be passed to
the plugin when Mozilla executes it (or otherwise be forced by a 
code change in the plugin).

Does this happen when running Mozilla/gnash on one machine, displaying
them on the X server on another machine?  If you have an AMD64 and
a non-AMD64, can you try displaying on the non-AMD64 and see if
*its* X server crashes?  And then flip around and run Mozilla/gnash
on the non-AMD64, displaying on the AMD64's X server.  We may discover
that the bug is not AMD64-related at all...

        John