No crashes with July 10 sources

[Richard Stallman]

It appears that the C-g stack clobbering bug does not exist in the
July 10 sources.  I have been using them now for almost 2 days
and they have not crashed yet.

When those crashes occurred, chances are Emacs was waiting in one of
the sit_for calls in keyboard.c.  That's where Emacs normally waits,
unless you have been idle for a minute or more.  So I suspect that the
recent sit_for changes (which were installed on July 10) made the
crashes start happening.

However, I see nothing even slightly suspicious in the code for
sit_for.  This leads me to suspect that the bug existed before, and
that this change only made it happen much more often.  If that is
true, just redoing the Lisp function sit-for won't fix it (keyboard.c
doesn't use the Lisp function sit-for anyway).

So I think we really need to track this down.
Does it happen to anyone else?