Re: Markdown export is non-deterministic?

[Ihor Radchenko]

Dominique Dumont  writes:

> The function shown in this blog helps a lot:
>
> https://amitp.blogspot.com/2021/04/automatically-generate-ids-for-emacs.html
>
> This works quite well for my org files.

Feel free to help with https://orgmode.org/list/87fsfutwin.fsf@localhost

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at .
Support Org development at ,
or support my work at 

Re: Markdown export is non-deterministic?

[Dominique Dumont]

On Saturday, 8 October 2022 00:31:00 CEST Matias Eyzaguirre wrote:
> I’m using org-mode to generate a markdown README for a project that is
> under version control. I’ve noticed that the anchor tags used for the
> table of contents seem to be annoyingly random and change from export
> to export. Is it possible to make the generated ids deterministic?

The function shown in this blog helps a lot:

https://amitp.blogspot.com/2021/04/automatically-generate-ids-for-emacs.html

This works quite well for my org files.

HTH

Dod

Re: Markdown export is non-deterministic?

[Ihor Radchenko]

Matias Eyzaguirre  writes:

> I’m using org-mode to generate a markdown README for a project that is
> under version control. I’ve noticed that the anchor tags used for the
> table of contents seem to be annoyingly random and change from export
> to export. Is it possible to make the generated ids deterministic?

In addition to the detailed explanation by Max, you can use ox-publish.
Publishing will preserve the ids as much as possible.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at .
Support Org development at ,
or support my work at 

Re: Markdown export is non-deterministic?

[Max Nikulin]

On 08/10/2022 05:31, Matias Eyzaguirre wrote:


#+begin_example
1.  [Introduction](#org775f005)
2.  [Roadmap](#orgccee397)
#+end_example

While these identifiers are generally not visible to a reader, they
add noise to the version control diffs. I'm wondering if it's possible
to force the relevantrandom number generator to use a fixed seed or,
more ideally, have the ids be based on the section title (or its
hash)?


You may try to add CUSTOM_ID property to section. Current behavior is 
intentional, see e.g.


Nicolas Goaziou… Re: HTML export uses anchor ids which change on every 
export. Sat, 29 May 2021 21:50:24 +0200. 
https://list.orgmode.org/87bl8tz5dr@nicolasgoaziou.fr


A long discussion:

Samuel Wales. stability of toc links. Tue, 8 Dec 2020 16:28:46 -0700. 
https://list.orgmode.org/cajcao8txvw39h18fuxltwodynkqsfhqcp+ydiywflee29b1...@mail.gmail.com


Some people managed to implement anchors based on heading titles

TEC. [Interest] Determanistic Org IDs. Sun, 19 Jul 2020 22:27:31 +0800. 
https://list.orgmode.org/e1jxajg-0004dk...@lists.gnu.org

Markdown export is non-deterministic?

[Matias Eyzaguirre]

I’m using org-mode to generate a markdown README for a project that is
under version control. I’ve noticed that the anchor tags used for the
table of contents seem to be annoyingly random and change from export
to export. Is it possible to make the generated ids deterministic?

For example one invocation of ~org-md-export-to-markdown~ will give me
this:

#+begin_example
1.  [Introduction](#org8f0a2e0)
2.  [Roadmap](#orga3f4ca7)
#+end_example

While exporting a second time, even if there are no changes to the
file, will produce something different such as:

#+begin_example
1.  [Introduction](#org775f005)
2.  [Roadmap](#orgccee397)
#+end_example

While these identifiers are generally not visible to a reader, they
add noise to the version control diffs. I'm wondering if it's possible
to force the relevantrandom number generator to use a fixed seed or,
more ideally, have the ids be based on the section title (or its
hash)?

Thanks,
  Matias