bug#35464: [PATCH] Refactor update_window_begin and update_window_end hooks

[Eli Zaretskii]

> From: Alex Gramiak <agrambot@gmail.com>
> Cc: 35464@debbugs.gnu.org
> Date: Thu, 02 May 2019 23:10:41 -0600
> 
> I have another change that I don't think is worth a new bug report,
> which is refactoring scroll_run_hook in a similar way. Is it okay to
> apply it?

I'm not sure this is a change for the best.  scroll_run is already a
frame-specific method; just because its 3 implementations are similar
in some of their parts doesn't yet mean _all_ possible implementations
will be that way.  Suppose the NS implementation needs to change for
some reason, or the TTY case needs to use this -- what do we do then?
move the code back to *term.[cm]?

The change also splits the code's logic into 2 parts in 2 different
places, another minus in my book.

This also adds an ifdef where currently there is none, which is an
additional small disadvantage.

Bottom line: I think this does not have enough significant advantages
to change the current code.

Thanks.