bug#30182: Update

[martin rudalics]

> Btw, if using window-at-side-p is so much more efficient than
> window-in-direction, then how come the former is not documented in the
> ELisp manual?  Should it be?

I'm not sure.  So far `window-at-side-p' is only used in
`window-in-direction' and to quickly tell whether a mode or header
line can be dragged ("quickly" because that doesn't check for whether
windows have fixed size or are already to small to get resized).  It's
not internal (not called `window--at-side-p') because it's called from
mouse.el

Also, `window-at-side-p' could check arguments better as evaluating
(window-at-side-p nil t) shows.

martin