Wouldn't it be better if forward/backward-char kept the old behavior, and
the rest of the commands did something similar to this in the end:

(setq disable-point-adjustment
      (preceding-or-following-character-visible-p))

I'm not entirely sure whether it would be better, but at the moment, I
can't think of a case where it wouldn't.

Evgeni


On Sat, May 4, 2013 at 1:16 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> > Date: Sat, 4 May 2013 12:17:35 +0100
> > From: E Sabof <esabof@gmail.com>
> > Cc: 14346@debbugs.gnu.org
> >
> > I see what you mean. But it still looks like a bug - whether I follow the
> > above recipe, or press C-e C-a, the point will (should?) go to the same
> > position, but the behavior is different.
>
> The behavior depends on the direction point was moving before ending
> up in the invisible text.  It's a heuristic, and as every heuristic,
> it sometimes fails.
>