> To: mycroft@gnu.ai.mit.edu
> Subject: Re: hp300 X 
> Date: Fri, 17 Jun 1994 09:30:02 -0700
> From: David Carrel <carrel@cisco.com>
> 
> It's at my house, of course.  I have been adding to it slowly, and was
> going to put out a complete set (or maybe just diffs) when I thought it was
> complete.  Recently I have tweaked the makefiles so that shared X libs are
> created and I added support for the shared memory extensions.
> 
What are the shared memory extensions?

> The only things remaining at this point is to get the cursor to redraw and
> to find out why the shared memory extensions cause my machine to panic.  If
> a window scrolls under the cursor, the cursor disappears.  It reappears
> only when you move the mouse.  The shared memory thing is hard to track
> because when the machine panics under X, the ITE doesn't get keyboard input
> since X11 had opened /dev/hil?.
> 
Mark Davies fixed the disappearing cursor problem.  I have included a diff
at the end of this message.  The diff also has a change I made for the
catseye which is supposed to allow you to poll the HW without affecting
the block mover.  Never noticed a difference however.  You can take it or
leave it.  If you really feel like fixing a known problem :-) text drawing
is really slow compared to the old R4 server.  Mark can give you details,
I think he knows how to fix it but just doesn't have time. 

I'm not sure I understand the last part.  If the machine paniced, what input
are you trying to give it?  Are you trying to use a debugger?  I have always
prefered taking a core and post-mortem debugging.  The gdb 4.7 in the Lite
/usr/src/contrib directory supports post-mortem debugging (and remote
interactive debugging).

Anyway, on a panic it is supposed to reset the ITE so that you get output
messages.  However, I seem to recall that not working right under 4.4
(of course, my machine never panics so I don't recall...yeah, right!)

------------------------------------------------------------------------------