bug#36886: 26.2; (global-eldoc-mode -1) should disable eldoc completely

[Eli Zaretskii <eliz@gnu.org>]

> From: Sanel Zukan <sanelz@gmail.com>
> Cc: 36886@debbugs.gnu.org
> Date: Fri, 02 Aug 2019 14:51:04 +0200
> 
> > So this only happens in M-:.  That wasn't evident from your original
> > report.  global-eldoc-mode generally shows the signatures when point
> > is inside any defun or other def* forms, not only in the minibuffer.
> 
> Hm... eldoc-message-function is part of eldoc package and I was hoping
> that, by completely disabling eldoc-mode, it will revert whatever is
> enabled when mode is initialized from that package. I guess different
> views from user and developer points.

I don't think this is about the differences in our views: I didn't say
I considered the behavior you described correct (it isn't).

My point was that your original report lacked some details, which
caused me to think the problem was not reproducible, because my naïve
attempt to reproduce it didn't use M-:.  It is important to describe
the recipe in enough detail to avoid that; please do that in the
future.