Re: invisible

[Luc Teirlinck]

Kazu Yamamoto wrote:
   
   You may notice this but I find another bug. Please consider the same
   example whose line2 is invisible.

   ----
   line1
   line3
   line4
   ----

   On the beginning of line3, end-of-line (C-e) does not move the cursor
   onto the end of the line.

Indeed.  This bug (or misfeature) did not occur in 21.3.

Note that when point appears to be at the beginning of line3, it
actually is at the beginning of line2. (C-x = gives point=7).

In Emacs 21.3, if you start at the beginning of the buffer, have
line-move-ignore-invisible set to nil (the default) and do C-n, you go
to the start of line2.  C-e carries you to the end of line2.  (But you
do not see this.) Two C-n from the beginning of the buffer carry you
to the beginning of line3.  C-e carries you to the end of line3 (which
is visible motion).

With line-move-ignore-invisible set to t, _one_ C-n at the beginning
of the buffer carries you to the beginning of line3.  The next C-e
carries you to the end of line3.

The problem with Stefan's changes is that they do not seem to work
well with _either_ value of line-move-ignore-invisible.  (Though they
work relatively better with t).

With Stefan's changes, _even_ if line-move-ignore-invisible is t, one
C-n from the beginning of the buffer still carries you to the
beginning of line2, because the beginning of line3 is forbidden by
`pseudo-intangibility'.  The behavior of the next C-e (carrying you to
the `i' of line3) is determined by that same pseudo-intangibility.

Sincerely,

Luc.