Re: Truncating scroll runs that copy to where we copied to

[Eli Zaretskii <eliz@gnu.org>]

> Date: Wed, 23 Nov 2011 09:41:18 +0900
> From: YAMAMOTO Mitsuharu <mituharu@math.s.chiba-u.ac.jp>
> Cc: david.reitter@gmail.com,
> 	emacs-devel@gnu.org
> 
> >>>>> On Tue, 22 Nov 2011 04:54:28 -0500, Eli Zaretskii <eliz@gnu.org> said:
> 
> > If truncation indirectly prevents copying of disabled rows, then I
> > think at least a comment to that effect should be in the loop which
> > "assigns the rows".  Just looking at the loop, it is not at all
> > apparent that only enabled rows are being assigned.
> 
> Maybe we can add an assertion as well as such a comment.

Yes, let's do that.

> > The question was whether the current code is right when it
> > unconditionally sets that flag.  If the `from' row is disabled, why
> > should its assignee `to' row be enabled?  If, after your changes, a
> > disabled row is never assigned, then `to' will already have its
> > enabled_p flag set, by virtue of the assignment.  Either way,
> > setting this flag unconditionally after the call to assign_row looks
> > bogus to me.  (By contrast, resetting the flag in `from' looks like
> > TRT, at least as long as we garble it.)
> 
> Ah, I misunderstood that you were proposing setting to->enabled_p to 1
> unconditionally.  Yes, maybe we can replace this assignment with an
> assertion.

Let's do that as well.

Otherwise, I think your changes are fine, thanks.  At least I cannot
find anything wrong with them.  Please go ahead and install them.