Thanks for the feedback @aeros

1. You could store the version as a global constant in conf.py, I don't think we need a dedicated text file just to store the latest unicode version.

I waffled on this a bit, but I went back and took another crack at it. Mostly I wanted to avoid having makeunicodedata.py accidentally mangle conf.py. I think I got it working reliably, though.

2. UNICODE_VERSION or UCS_VERSION would be a far more explicit name for the constant. As far as I'm aware, "UCD" is not a commonly known acronym for Unicode.

Kinda embarrassed that I didn't clean that up before submitting the PR. I should have slept first. :) I renamed everything to 'UNIDATA_VERSION' to match the original variable in makeunicodedata.py. This could just as easily be changed to 'UNICODE_VERSION', but I erred on the side of staying consistent to what was already in place.

3. IANAL, but I believe there are legal issues associated with copyrighting your own file additions. Our primary means of providing credit/acknowledgement to contributors these days is by adding a Contributed by <name>. sentence at the end of the news entry for the PR. You can add this in yourself.

I wasn't sure of the convention. I copied the heading format from what was in a few of the other sphinx extensions in the same directory. I inserted the bpo link instead (which was also in one of the other extensions), and updated the news entry.

I think the comments in the file constitute a change log, not a copyright notice. The copyright situation is already handled by the CLA check, which ensures that the contribution has a suitable license.

It seems OK to me to add a line there, even though there is external history in VCS metadata.

I feel this would be easier by using rst_prolog like:

rst_prolog = f"""
.. |UNIDATA_VERSION| replace:: {UNIDATA_VERSION}
"""

it would not need an extension.

Closing per #66783 (comment)

A