Re: more problems with line-move

[Richard Stallman]

    I can't remember exactly which case, but I think that it can happen in some
    cases where the invisible area has neither end sticky in which case there
    is no clear preference in which of the two ends to go to.

Shouldn't this always be decided based on continuing the direction in
which point was moving?  If it is done this way, why is there a problem
when there's no stickiness?