Re: invisible

[Luc Teirlinck]

Miles Bader wrote:

   Perhaps Stefan's change caused problems with line-move, but that's
   kind of to be expected because the line-move implementation is a
   confusing bloated piece of crap, which _always_ had annoying
   problems with invisiblity and intangibility (even before the
   change).  I don't know if it got better or worse -- probably the
   bugs just changed slighly -- but I think the real solution is to
   rewrite line-move.  [I expect it won't be done soon, but the
   problems are minor enough that I don't think there's all that much
   hurry.]

I was not necessarily arguing for undoing Stefan's changes.  I just
want to know exactly where we stand on this issue.  Stefan's changes
seem to clash with the prior invisible-intangible feature I
mentioned.  So if we decide to keep Stefan's feature, an obvious
question is whether we want to keep that one.  The majority (but not
all) of the problems caused by Stefan's changes concern the case where
line-move-ignore-invisible is nil.  Does a nil value still make sense
after Stefan's changes _as a user option_ (one might still want to
bind it to nil from Lisp code occasionally)?  Any brave person
volunteering to rewrite the entire line-move complex (_not_ me), is
going to have to know the answers to these questions.  Anybody wanting
to make less radical changes to fix these and similar, yet to be
discovered, bugs is going to have to know the answers to.  Depending
on the answers, the majority (but not all) of the bugs would
automatically disappear anyway.  This includes three out of the five I
know of.

Sincerely,

Luc.