Re: nil became bound to (#<save_value ptr=0x085526f0 int=0> 96536 96288)

[Eli Zaretskii]

> From: Joe Wells <address@hidden>
> Date: Mon, 30 Jan 2006 03:31:01 +0000
> Cc: address@hidden
> 
> >  > Anyway, maybe not everyone else will have the same attitude that you
> >  > do.
> >
> > Others might just ignore the message, but no-one without Gentoo can help you
> > if the bug is no longer in CVS.
> 
> I've been running this version of Emacs for 6 months, during which
> time the bug happened exactly once.  Furthermore, I can not reproduce
> it.  Therefore, it is obviously a bug that is very difficult to
> trigger.  Therefore, there is a good chance the bug still exists.
> Given that the bug involves nil becoming rebound, which cripples
> Emacs, it is clear that it is serious.  Therefore, if someone else
> were to come across the bug, they could benefit from knowing something
> about the circumstances under which I encountered it.  Therefore, I
> reported it.
> 
> I quite frankly don't care why you feel the need to be rude about my
> report.  Take it for what it is, a report of some details about an
> interesting way in which something very bad can happen to Emacs.

No one is being rude to you, the responses were polite and to the
point.  It _is_ true that we can probably do nothing with a report
like yours, for the reasons you've been told.

Now, you didn't ask, but if you want an advice how to try to catch
this problem if it happens in the future, then (assuming whoever built
your Emacs binary didn't strip it of the debugging symbols) run Emacs
under GDB and put a data-write breakpoint (a.k.a. watchpoint) on Qnil.
Then, whenever some piece of code writes to it, GDB will catch it
red-handed.

FWIW, the last time I saw such problems it was due to a faulty memory
chip, so perhaps it's a hardware problem in your case, too.