Re: PG 19 release notes and authors - Mailing list pgsql-hackers

On Mon, Apr 6, 2026 at 1:20 PM Bruce Momjian <bruce@momjian.us> wrote:
> I am confused how this was not clear in the January 2025 discussion and
> why people didn't mention they didn't like it then.  I can't quote
> anything from anyone but myself from a private email, so here is some
> text by me to you:
>
>         I started using Co-Author as a way to indicate that the Co-Author
>         wrote some of the patch, but I modified it enough that I don't want
>         to attribute/blame the work entirely on the Co-Author.  Are you
>         saying when that happens, I should name myself also as a Co-Author?

I just don't understand what that has to do with the present
situation. The answer to the question you quote here was "yes", but it
doesn't touch on the present question, which is what happens when
there are both 1 or more Author: tags and also 1 or more
Co-authored-by: tags. The unanimous answer from everyone here, except
you, is that all those names should be listed as authors for release
notes purposes. I still don't understand how or why the previous
discussion led you to any other conclusion. I think we were really
clear that the committer must list themselves as an author or
co-author if they wish to be so credited, and if they don't, the
authors are exactly as named. The only thing we're adding to that now
is that if there's a mix of author and co-author tags, that
distinction is to be ignored for release note purposes. I think the
reason that wasn't discussed previously is because people just assumed
that was the only reasonable outcome. I mean, if commit #1 says
Co-authored-by: Fred Co-authored-by: Bob and is credited as (Fred,
Bob), and commit #2 says Author: Fred, Co-authored-by: Bob, how would
anyone justify crediting the second one as just (Fred)? If
Co-authored-by was good enough to justify mentioning the name in the
first case, it must be in the second case as well.

To be fair, I don't think this is a perfect answer. I mean, I have had
situations in the past where I (let's say) push 40 commits consisting
of 20,000 lines of code to implement some feature. You (rightly)
bundle all those into a single release note entry. Out of those 40
commits, one small commit (say, 150 lines) was written jointly by me
and another person. When the release notes come out, the authors of
the overall feature are me and that other person. Somehow, that feels
like it grossly overstates the contribution of that other person, and
I have been known to be a little miffed about it. However, as others
have already pointed out, it's better to be generous in crediting
other people than not, so I think it is the right answer for the
project despite my occasional pique. At the end of the day, the
chances that people know that Robert Haas contributes a bunch of stuff
to PostgreSQL are pretty good; the chances that they know that the
other person also contributes are not as good. Therefore, it's more
important not to understate that person's contribution than it is to
not understate mine. If somebody really wants to know what happened,
they can click through to the commits, and from there to the mailing
list discussions.

--
Robert Haas
EDB: http://www.enterprisedb.com