Re: Bug: shiftmeta[left|right] on multi line items [9.1.2 (release_9.1.2-40-g6ca906 @ /usr/local/share/emacs/27.0.50/lisp/org/)]

[Nicolas Goaziou <mail@nicolasgoaziou.fr>]

Hello,

Nathan Aclander <nathan.aclander@gmail.com> writes:

> Using ^ as point like you did, what I am expecting to happen is:
>
> Case 1:
>  
> * Heading
>
> - foo
>   - ^bar
>     second bar line
>
> M-S-left/right moves
>
> - bar
>   second bar line 
>
> left and right.
>
> Case 2:
>
> - foo
>   - bar
>     ^second bar line
>
> M-S-left/right moves
>
> - bar
>   second bar line 
>
> left and right.
>
> Case 3:
>   ...
>   some very long text
>   ^some very long text
>   some very long text
>   ...
>
> Whatever sub list that "some very long text" is part of should be moved
> left and right.
>
> None of these scenarios seem very confusing, and ( at least to me ) seem
> more consistent than the current behavior. Could you explain why you
> find adding this behavior would make this confusing?

This is confusing because it is ambiguous. The same point could
correspond to a table, multiple lists, and multiple headings, all
reclaiming S-M-left binding.

Of course, we could apply the change to the closest, i.e., the inner,
structure, but, in some cases, e.g. the third one, the context is not
obvious. You may end up ignoring if the whole section was shifted, or
just some random sub-list. 

I don't find this behaviour particularly satisfactory. OTOH, only
reacting when point is on item bullet's line is totally unambiguous.

> As a current workaround, what I have to do is:
>
> 1. save point
> 2. Go to the parent sub-list
> 3. Shift left or right
> 4. move point back to its previous location. 

You can also define a function that does this and add it to
`org-shiftmetaleft-hook' or `org-shiftmetaright-hook'.

I suggest to do that if you think that's better.

Regards,

-- 
Nicolas Goaziou                                                0x80A93738