On Wed, May 1, 2019 at 12:55 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: Jonathan Tomer <jktomer@google.com>
> > Date: Wed, 1 May 2019 12:29:42 -0700
> > Cc: Michael Albinus <michael.albinus@gmx.de>, 35497@debbugs.gnu.org
> >
> > It's true that notifications are imperfect, but IMO they are the only
> possible way to test that particular invariant,
>
> I'm not sure.  You could instead advise or hook write-region, and see
> that it is not being called to write to the file, for example.
> Wouldn't that be easier and more reliable?
>

Oh, of course -- that's quite a bit better. Thanks, will send a new patch
in a moment.