bug#35161: 26.1; follow-mode spurious motion

[Eli Zaretskii <eliz@gnu.org>]

> From: Devon Sean McCullough <Emacs-Hacker2018@jovi.net>
> Date: Fri, 5 Apr 2019 11:58:23 -0400
> 
> bash$ /Applications/Emacs.app/Contents/MacOS/Emacs -Q -eval '(progn (insert-file-contents-literally "/dev/stdin") (eval-buffer))' <<.
> (xref-find-definitions 'set-frame-font)
> (toggle-frame-maximized)
> (follow-mode)
> (delete-other-windows)
> (mark-defun)
> (split-window-right)
> (setq s1 0
>       t1 (run-at-time 1 1 (lambda (&rest args)
> 			    (other-window 1)
> 			    (sit-for 0)
> 			    (when (> (setq s1 (1+ s1)) 3)
> 			      (cancel-timer t1)
> 			      (message "Now when you type C-x o the unselected cursor will spuriously move to the bottom line.")))))
> .
> bash$ ## Both windows' cursors should stay put, but instead one of them moves.
> 
> 		Peace
> 			--Devon
> 
> P.S. I'm curious why programmatic (other-window 1) (sit-for 0)
> doesn't exhibit this bug.
> Perhaps ELisp manual chapter 28.21 Windows and Point
> should mention relevant pitfalls?

It's not a bug, not in redisplay anyway.  The last line of the left
window is not fully visible, at least on my system, and that causes
the effects you see (if I understand correctly the issue you find
unexpected).

Btw, calling sit-for from a timer function is not a good idea, but
this is a tangent.

Thanks.