Hello

> Ihor Radchenko <yantar92@posteo.net> writes:
>
> P.S. There going to be Emacs meetup this Saturday, July 22
> (https://emacs-apac.gitlab.io/announcements/july-2023/). I plan to be
> there and, if you need it, we can try to resolve difficulties in more
> interactive way.

I am currently travelling this weekend, otherwise I happily would
attend so that this patch could speed forward. Next month maybe?

> This is failing because of subtlety in
> `org-test-with-temp-text' that inserts text after activating
> org-mode.  PROPERTY lines are only computed when org-mode is
> activated, unless you explicitly call
> `org-set-regexps-and-options' to refresh them.

Okay, it's working as intended now.

> This fails because you did not provide LANG for source block.
> override-document-and-heading-tfile-with-yes test is also
> missing LANG.

Thanks, I thought I fixed this last time you mentioned it but
apparently only in my mind...

> Side note: You are re-implementing the already available ERT
> features for failed test reporting. Instead of controlling
> which tests failed manually, you should better follow the
> example of all other tests and simply use a sequence of
> `should'/`should-not' forms.

Okay, I've re-written this to conform to the `should' macros.  I
*really* wish these macros could be named so that I could debug
failing statements better, but for now numbering them in the
comments and using the `l' binding works well enough for
debugging.

I have two tests that I've set to "should-not" pass for now, but
once the correct merging behaviour has been implemented, I
believe that I will change them to "should" pass statements:

- 9. do-not-tangle-this-block
- 12. tangle-file-with-spaces

Attached is the new diff (travelling makes it hard to write
meaningful commit messages so I did not use git-format-patch),
which should be used against the upstream/main.

Looking forward to further feedback!

Best,
Mehmet