On Thu, Jan 23, 2020 at 4:28 PM Stefan Monnier <address@hidden
> >> Good point: `minibuffer-force-complete-and-exit` probably needs to be
> >> changed accordingly (i.e. to just `exit-minibuffer` when `required`
> >> was nil).
> > How will it determine that? minibuffer-completion-confirm seems to be the
> > same whether require-match is t or not.
> We could introduce a new `minibuffer-require-match` variable (and mark
> `minibuffer-completion-confirm` obsolete).
+1. Not particularly pretty, but it would you join the ranks of many
other `minibuffer-*` special variables that hold these kinds of things.