bug#38181: Actual height of mode-line not taken into account

[Eli Zaretskii]

> Cc: jonas@bernoul.li, 38181@debbugs.gnu.org
> From: martin rudalics <rudalics@gmx.at>
> Date: Tue, 5 May 2020 18:57:06 +0200
> 
> With Bug#40639 and Bug#41071 the situation is simple: The internal
> border changed face but when we enter init_iterator
> inhibit_free_realized_faces is true and we don't handle it.

Why do we need to enter init_iterator when the border face changes?
And why do we need init_iterator to handle the face change?

Thanks for the rest, I will look at that soon.