On 18/07/2022 17:55, Max Nikulin wrote:

> I am sorry if I am wrong, but I do not see you among known Org, 
> contributors. You patch is likely greater than it is allowed for
> TINYCHANGE, so before you patch can be committed, copyright assignment
> should be signed, see
> <https://orgmode.org/worg/org-contribute.html#copyright> for details.

I've already assigned copyright to the FSF for work on Emacs (see
attached certificate). Do I need to do anything else for org-mode in
particular?

> The cases might be improved by using different values, so when
> particular `should' form fail it is easier to find it in the code

Nice idea. Done.

> "link text" might be a bit ambiguous here. I would consider "link
> location", "string containing link type and target", or something
> else.

Clarified, and added some examples in the docstring as well. You're
right about this---I've found such language ambiguous in the past.

> Emacs-26.3:
> make test-dirty BTEST_RE=test-ol/insert-link-insert-description
> [failing test]

Hmmmm. I ran:

,----
| make test-dirty BTEST_RE=test-ol/insert-link-insert-description
`----

at the top-level of the org dir on my machine, and I got:

,----
| Running 1 tests (2022-07-21 12:55:40+0100, selector ‘"test-ol/insert-link-insert-description"’)
|    passed  1/1  test-ol/insert-link-insert-description (0.016860 sec)
| 
| Ran 1 tests, 1 results as expected, 0 unexpected (2022-07-21 12:55:40+0100, 0.017086 sec)
| 
| /usr/bin/make cleantest
| make[1]: Entering directory '/home/hugo/.config/emacs/straight/repos/org'
| rm -fr /tmp/tmp-orgtest || { \
|   find /tmp/tmp-orgtest -type d -exec chmod u+w {} + && \
|   rm -fr /tmp/tmp-orgtest ; \
| }
| make[1]: Leaving directory '/home/hugo/.config/emacs/straight/repos/org'
`----

So the test passes on my machine.

Are you sure you're using the most recent version of the patch? The
quote from me you used is dated from the 16th, which was at least one
version ago. Could you retest with the version attached to this email?

If it still fails, could you try isolating whether it's a specific
`should' clause which is the problem, if it fails with any of them?

Thanks,

Hugo