Re: Stipples and scroll optimizations

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: emacs-devel@gnu.org
> Date: Sun, 08 May 2022 18:59:52 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > But that can be different (even VERY different) from what _should_ be
> > on the glass.  We cannot rely on the current matrix for making such
> > decisions.
> 
> Yes, but the problem here is that we cannot copy what already _is_ on
> the glass if there is a stipple.  So the contents of the desired matrix
> are irrelevant for our purposes.

Not necessarily true.  update_window does this:

      /* Try reusing part of the display by copying.  */
      if (row < end && !desired_matrix->no_scrolling_p)  <<<<<<<<<<<<<
        {
          int rc = scrolling_window (w, (tab_line_row != NULL ? 1 : 0)
                                     + (header_line_row != NULL ? 1 : 0));

If the desired matrix is irrelevant, why do we test the no_scrolling_p
flag on that matrix?

Anyway, feel free to do it your way, but I'd be very weary of using
flags from the current matrix if a safer method exists.