Re: [BUG] Source block indentation does not work properly for yaml-mode [9.6.6 ( @ /home/user/.emacs.d/elpa/org-9.6.6/)]

[Sébastien Miquel]

Ihor Radchenko writes:
> > What happens is: in the org buffer, the line is not empty, because it
> > has the org indentation (which was possibly just added by
> > org-indent-line), but in the edit buffer, the line is empty, because
> > the common org indentation was removed. In that case, we want to add
> > back the org indentation.
>
> May you please provide an example when it is necessary?
> `org-indent-line' will run `org-babel-do-key-sequence-in-edit-buffer', so
> it should still use `org-src--contents-for-write-back' and will not
> modify the org buffer text directly.

You're at the end of a line, you press =org-return-and-indent=.
 1. It adds a newline character.
 2. =org-indent-line= adds the org indentation, _before_ calling
    =org-babel-do-key-sequence-in-edit-buffer=
 3. The native edit call removes the common indentation, before
    calling tab in the native buffer.
 4. Calling tab in the native buffer possibly does nothing.
 5. =org-src--contents-for-write-back= sees the current line is empty,
    but it should indent it (with org indentation) nonetheless.


> > --- a/lisp/org-macs.el
> > +++ b/lisp/org-macs.el
> > @@ -483,9 +483,12 @@ line.  Return nil if it fails."
> >           (when skip-fl (forward-line))
> >         (while (not (eobp))
> >           (let ((ind (progn (skip-chars-forward " \t") (current-column))))
> > -           (cond ((eolp) (delete-region (line-beginning-position) (point)))
> > -                 ((< ind n) (throw :exit nil))
> > -                 (t (indent-line-to (- ind n))))
> > +           (cond ((< ind n)
> > +                   (if (eolp) (delete-region (line-beginning-position) (point))
> > +                     (throw :exit nil)))
>
> This function is actually confusing both before and after the change.
> According to the docstring:
>
>      When optional argument N is a positive integer, remove exactly
>      that much characters from indentation, if possible.
>
> But the function can actually remove less than N characters.
>
> Before your change, all the blank non-empty lines were unconditionally
> removed. After your change, the first such line is removed and the
> function returns nil without continuing.

I don't understand. With this change, the function only stops if it
finds a non blank line with less than n indentation (same as before).
When a blank line with less than n indentation is found, it is emptied
(same as before), and execution continues.

> > * lisp/org-macs.el (org-do-remove-indentation): Preserve
> > indentation (spaces vs tabs) past the common indentation to remove.
> > Do not empty blank lines.
>
> Since not removing blank lines is intentional after the change, why
> doing it on a single line that is indented less than N?

Are you advocating for computing N using blank lines as well ?

 1. It isn't consistent with the previous behaviour.
 2. If I mistakenly add a space to an empty line in a src block, an
    edit-special round trip will add indentation to every line.

Is there any benefit ?

--
Sébastien Miquel