Re: [BUG] ox-odt fails for org-id links (e.g., from org-roam v2) [9.5.2 (9.5.2-gfbff08 @ /home/moritz/.emacs.d/elpa/27.2/develop/org-9.5.2/)]

[Max Nikulin]

On 21/10/2022 10:16, Ihor Radchenko wrote:
> Max Nikulin writes:
>
> > I am apologizing if my comments make no sense. I remember a thread on
> > improper handling on id links in ox-html, so I am a bit surprised that a
> > link to the whole file (if I got the goal of the patch correctly, of
> > course) is the only problem with ODT.
>
> It is likely not the only problem. But this thread is about one
> particular problem that should also be fixed regardless of others.
> If you find other problems, please report them.

I am sorry. I confused this thread with some other bug report. I agree 
that the original post describes namely file links. When Moritz provided 
an example in response to your request for more details I decided that 
more similar problems are coming. Perhaps particular use case does not 
include links to specific headings in files and I was wrong. On the 
other hand if ox-odt code works with id links better than ox-html then 
the former may be a source of inspiration how to fix the latter.

> Here, we at least fix the export error. So that people can at least get
> some exported .odt file.

I have tried main HEAD (your fix should be applied) and I did not get an 
error, but the result is quite strange. Consider a single file with an 
internal link:

:PROPERTIES:
:ID:       3c885843-13ad-4dd8-a57f-de8271403b85
:END:
#+TITLE: Note A

* H 1
Some text

* H 2
[[id:3c885843-13ad-4dd8-a57f-de8271403b85][Link to Note A]]

Link in the exported document:

<text:a xlink:type="simple" xlink:href="test-odt.org">Link to Note 
A</text:a>

I would expect a target withing the same .odt file.

Properties as displayed in LibreOffice UI are even more peculiar:
Document path
file:///tmp/d/test-odt.odt/test-odt.org
Target in document is empty.

My expectation for path is file:///tmp/d/test-odt.odt or at least 
file:///tmp/d/test-odt.org, but certainly not their combination. I am 
unsure if you consider this observation as relevant to this thread.