bug#12872: 24.2; Provide a feature to trigger mode-line redisplay

[Eli Zaretskii <eliz@gnu.org>]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: 12872@debbugs.gnu.org
> Date: Mon, 06 Dec 2021 07:00:39 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > How would that be different from what force-mode-line-update does now?
> > It sets the update_mode_line flag for the window's buffer, not just
> > for the window, and it also sets a flag to prevent redisplay
> > optimizations that could get in the way of the mode-line update.  What
> > do you suggest to do instead, and why would that be useful?
> 
> Somebody said earlier:
> 
> > force-mode-line-update is a blunt weapon, and causes a much more
> > thorough redisplay than its name says.
> 
> If this isn't accurate, then I guess there's nothing to do here.

Somebody else explained earlier the "blunt" part:

> > The problem with the update_mode_line flags is that they are
> > indiscriminate: the mode line shows a lot of data, each one of it
> > changes at different frequencies and due to different triggers.  If we
> > want a finer resolution there, we need to make these flags not just
> > simple booleans, but enumerations with several values, or maybe
> > bitmaps.  Then redisplay_window could be smarter about redrawing parts
> > of the mode line.
> 
> I think redrawing parts of the mode line would be more work than it's
> worth.

Then I guess we should close this as wontfix.