Re: feature/icomplete-vertical

[Gregory Heytings]

> > I did not see objections from Stefan.  And the objection of Eli is 
> > only, AFAIU, that in a longer term a better solution should be 
> > implemented.
>
> No, my objections are that your proposal is not safe wrt recursive 
> editing, and will affect uses of mini-windows other than prompting the 
> user, where these effects aren't intended.
>
> AFAIU, Stefan agreed that this is a problem in your proposal.

I don't know which proposal you refer to in that sentence, but I guess 
it's the (second) patch I sent on bug#43572 a week ago, against Emacs C 
core.

The solution in this thread works at the Lisp level, and will not affect 
other uses of the miniwindow, because it relies on a minibuffer-local 
variable.  AFAICS it is also safe wrt recursive editing.