Re: outline-mode : cursor and point mismatch ?

[Bastien <bzg@gnu.org>]

Hi Nicolas,

Nicolas Richard <theonewiththeevillook@yahoo.fr> writes:

> I have GNU Emacs 24.1.1 (i686-pc-linux-gnu) of 2012-06-14.
> I experience a strange behaviour with outline-mode (and also in
> org-mode) :
>
> (emacs is started with emacs -Q)
>
> Say I have a buffer with the two lines:
> * A
> Some text
>
> and I hit C-c C-d (hide-subtree) and M-<. Now I have:
> * A...
> ^
> where '...' indicates hidden text and '^' indicates the
> position of the cursor (on the title line).
>
> Then hitting C-f a few times will advance the cursor successively :
> * A...
>  ^
> * A...
>   ^
> * A...
>      ^ <- I did not expect this.
> * A...
>       ^
>
> The behaviour that I expected was :
> * A...
>  ^
> * A...
>   ^
> * A...
>    ^   <- different
> * A...
>       ^

The three dots here are not characters, they are a single graphical
entity (a "glyph").  Moving the point after the "A" character leaves 
the cursor one "visual character length" before the end of the glyph,
hence what you observe.

> And in fact, adding 'text' when in the situation
> * A...
>      ^
> gives the following situation:
> * Atext...
>          ^

That's because, even though the point visually appears right before the
end of the glyph, it *is* right before the whole glyph.  I agree this
looks wrong.

Note that in org-mode, when `org-catch-invisible-edits' is non-nil, it
will help you catch edits in such invisible areas.

> I am a little bit worried that I do not find references to this
> behaviour on the net ; am I the only one experiencing this ?

HTH,

-- 
 Bastien