Re: invisible

[Luc Teirlinck]

Klaus Zeitler wrote:

   there was a long discussion, but I don't think that an agreement
   came out of it, what to do.

It was not completely clear to me whether something actually was
decided or not.  I believe the main issues were more or less left up
in the air.

Stefan made a pretty radical change to the invisibility property.
Prior to Stefan's change interactive commands would just move into
invisible text (unless it was explicitly made intangible) and you
could even test your agility and memory skills by editing the text
while invisible.  Then Stefan made invisible text "pseudo-intangible",
completely changing the behavior of the invisibility property.  Prior
to Stefan's change, another change was made independently: a special
feature meant to prevent point from standing still while moving over
invisible-intangible text.  After Stefan's changes to the invisibility
property, the variable line-move-ignore-invisible no longer seems to
make much sense and Stefan's changes introduced bugs if the variable
is nil (the default) one of which is the bug you reported.

I know of five different bugs related to line-move's interaction with
the invisibility property, or with the invisible-intangible
combination.  I do not believe it makes any sense to even try to fix
any of these bugs, unless one would have a definitive answer to the
following questions:

1.  Do we want to keep Stefan's "invisible implies pseudo-intangible
    feature"?

2.  Do we want to keep the prior invisible-intangible feature meant to
    avoid point to appear to stand still (which if one answers yes to
    (1) does not happen anymore anyway).

3.  Do we want to keep supporting line-move-ignore-invisible as a user
    option?  (We probably should keep it as a non-user variable, regardless.)

Maybe some of these questions _were_ answered in the prior thread.
That is just not completely clear to me.

Three bugs of the five bugs have been reported in this thread.  Two,
including your one, could be solved just by having
line-move-ignore-invisible always set to t.  The third one could be
solved by eliminating Stefan's feature, although that would remove the
motivation for eliminating line-move-ignore-invisible as a user
option.  Eliminating Stefan's feature would also eliminate Kazu's
second bug, without any need to eliminate the user option.

There are two other bugs related to invisibility, which I might report
later.  One occurs with line-move-ignore-invisible set to nil and is
caused by Stefan's feature.  It disappears by either getting rid of
Stefan's feature _or_ by getting rid of line-move-ignore-invisible as
a user option.  The other occurs with line-move-ignore-invisible set
to t and is _not_ caused by Stefan's feature.  The fifth one needs to
be fixed regardless of the answers to the questions above, but it is
intertwined with the fourth one, so even trying to fix the fifth does
not make any sense without knowing exactly the answers to the three
questions above.

Sincerely,

Luc.