Re: Crash in show_busy_cursor.

[Eli Zaretskii]

On Wed, 29 Nov 2000, Jason Rumney wrote:

> Hmm, perhaps gdb is reporting a false value for f due to it being kept
> in a register?

It's possible.  This is easy to find out: say "p/x &f" at GDB's prompt and 
observe.  If f sits in a register, GDB will tell you that it cannot take 
its address.

You should be able to do this even in a session that didn't crash, of 
course, just set a breakpoint inside show_busy_cursor.

(A variable that sits in a register usually has a totally garbled value, 
not something nice like zero, but who knows?)

> Unfortunately ddd has dumped core over the top of this core file now,
> and I cannot reproduce this (I can't even get this show_busy_cursor to
> be called when running under the debugger), so I cannot get any more
> information (like register values etc).

Well, next time, then ;-)

> > Unless XCreateWindow (or other X* functions called for previous frames in 
> > the list) somehow corrupted the stack, that is.
> 
> They haven't been called yet, so this (which is what I was thinking of
> before) is not the problem.

Yes, this seems to be clear from the disassembly.