On Sun, Nov 02 2025, Ihor Radchenko wrote:

> Kristoffer Balintona <[email protected]> writes:
>
>> I believe I have encountered a bug with `org-element-interpret-data`.
>> Below is a simple case that demonstrates the misbehavior I observe.
>>
>> Given an org buffer will just the following plain list:
>> + foo
>>   1) bar
>> when I call `(org-element-interpret-data (org-element-parse-buffer))`, I
>> expect the returned string to be identical to the contents of the
>> buffer, since to my understanding `org-element-interpret-data` is the
>> reciprocal of `org-element-parse-buffer`.
>
> This expectation is not correct.
> `org-element-interpret-data' loses certain information along the way
> and normalizes keywords. For example, indentation is not retained and
> keywords are downcased. In your case, lists are renumbered and
> normalized to use consistent standard bullets.
>
>> Instead, the returned string is this:
>> - foo
>>   1. bar
>> As is seen, the bullet points in the plain list are changed. It seems
>> that all non-numbered bullet points are set to the "-" bullet point and
>> all numbered bullet points are changed to "1."-style bullet points. This
>> is true regardless of how large and how many levels the plain list is.
>
> This is expected and intentional.
>
> Not a bug.
> Canceled.

Thank you for the clarification.

> We can discuss whether parsing/printing loop should truly retain the
> buffer contents byte-to-byte, but such feature would require major
> changes in the parser and interpreter. It is not just about lists.

Makes sense, and I don't think the necessary changes you describe would
be worth pursuing.

-- 
Kind regards,
Kristoffer

Reply via email to