On Wed, 26 Nov 2014 10:54:19 -0500 Stefan Monnier <monnier@IRO.UMontreal.CA> wrote:

>> Oops, you commented on the last version of my patch, which, however, I
>> had already considered superseded by the last version of the command
>> subsequently posted by Paul Rankin, which in effect already addressed
>> your concerns.  Rather than repost the diff for the latter, I'm
>> appending the two corrected versions of outline-move-subtree-down to
>> facilitate comparison and deciding which to use.
>
> Please just send the diff that (you think) should be applied.
> Or two diffs to apply in sequence if you want to distinguish your
> changes from Paul's.

Ok, I've done the latter, appended below.  The first change just fixes
the bug, incorporating your simplification of my patch for when to move
forward and when to add a newline.  The second change is Paul's
refactoring of the code to avoid setq's of let-bound variables, which
I think make the code cleaner and more elegant (the version he posted
also fixed the bug, but I prefer your simplification).

>> The only thing I added to both is a condition for triggering the
>> user-error, since the message doesn't seem appropriate when you try to
>> move a subtree down at eob or up at bob (if you don't think it's worth
>> avoiding the message there, I'll remove the condition).
>
> I don't see why we should avoid the error in those cases: AFAICT, we
> can't do what the user asked, so we should signal an error.

Ok, I removed that part.

Steve Berman