bug#16874: 24.3.50; Regression in active window detection

[Eli Zaretskii <eliz@gnu.org>]

> From: Damien Cassou <damien.cassou@gmail.com>
> Date: Tue, 25 Feb 2014 17:28:57 +0100
> Cc: 16874@debbugs.gnu.org
> 
> On Tue, Feb 25, 2014 at 5:18 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> > It's a deliberate change in Emacs that breaks functions which assume
> > that the window whose mode line is being redrawn is selected at that
> > time.
> 
> is there anything we can do about that?

What kind of thing did you have in mind?

FWIW, I think code that relies on implementation details of the
display engine runs a high risk of becoming broken.  I'd rather have
application programmers request features that will allow them to do
what they want using some documented public interfaces, instead of
looking high and low for how to trick Emacs into revealing some of its
internals.

So perhaps you could explain what feature/API/hook would you want to
be added to Emacs so that powerline, whatever it is, could do its job
reliably?  Who knows, it might even be that such a feature already
exists, or is easy to implement.

Thanks.