bug#38992: 27.0.60; when enabled, fido-mode seems to break vc-git-grep

[Stefan Monnier]

>> It's arguably breaking compatibility.  But you might be right that maybe
>> it's a non issue.  A quick `grep` shows that outside Emacs itself, at
>> least Helm might be affected.
> I don't disagree. But I'm not sure how to keep the backward
> compatibility either.
>
> Unless the old variable is kept as-is (both assigned and referred to) and
> the new one is looked up solely in one place
> (minibuffer-force-complete-and-exit). This way, we can't mark the old 
> one as obsolete, though.

Inded.

> (BTW, at least one reference to minibuffer-completion-confirm in Emacs binds
> it to t already; not sure what's the intended effect:
> lisp/calc/calc-store.el:197).

Maybe it used to be that `minibuffer-completion-confirm` used to be set
to something like (not (memq require-match '(nil t)))?

In any case, I do know it used to be that
`minibuffer-completion-confirm` was only boolean.


        Stefan