Valentin Lab <valentin.lab@kalysto.org> writes:

>>> @@ -1216,6 +1259,13 @@
>>>               (org-test-with-temp-text "* H\n:PROPERTIES:\n:key:\n:END:"
>>>                 (org-indent-region (point-min) (point-max))
>>>                 (buffer-string)))))
>>> +  ;; ;; Indent property drawers according to `org-adapt-indentation'.
>>> +  ;; (let ((org-adapt-indentation 'headline-data))
>>> +  ;;   (should
>>> +  ;;    (equal "* H\n  :PROPERTIES:\n  :key:\n  :END:\n\ncontent2"
>>> +  ;;           (org-test-with-temp-text "* H\n:PROPERTIES:\n:key:\n:END:\n\ncontent"
>>> +  ;;             (org-indent-region (point-min) (point-max))
>>> +  ;;             (buffer-string)))))
>> 
>> This test is commented. Is it intentional?
>
> My bad ! and an interesting talking point. I'm removing these commented 
> line in the upcoming patch. They were here (and inadvertently committed) 
> because while trying to test that my addition would not indent beyond 
> the headline data, I noticed that actually `org-adapt-indentation' set 
> to `headline-data' was actually indenting beyond headline data ! As I 
> don't want to break anything, I was left quite puzzled with what to do:
> - I can fix this, but fixing this is for me subject to another 
> submission, and will touch behaviors that might be wanted,
> - Not fixing this make me submitting a feature that carries what I see 
> like a "bug".
>
> But, is that a bug ? Here is the case:
>
> --8<---------------cut here---------------start------------->8---
> * H
> :PROPERTIES:
> :key:
> :END:
>
> content
> --8<---------------cut here---------------start------------->8---
>
> Using `org-indent-region' on all the content, with 
> `org-adapt-indentation' set to `headline-data', will result to:
>
> --8<---------------cut here---------------start------------->8---
> * H
>    :PROPERTIES:
>    :key:
>    :END:
>
>    content
> --8<---------------cut here---------------start------------->8---
>
> My issue is with the treatment of the 'content' line that is not 
> headline-data for me, and should not have been indented. Am I right in 
> my expectation ?

Yes are right and we got a bug here.

I am attaching the tentative fix.

Bastien, I feel that the current implementation of
`org--get-expected-indentation' is wrong since we introduced
`org-adapt-indentation' 'headline-data value. AFAIU, it was done by you
in e3b79ad2b. See the commit message in the patch below for the
explanation why I think that something is going wrong.

Am I missing something?

Best,
Ihor