Re: PG 19 release notes and authors

Hi,

On 2026-04-06 09:37:39 -0400, Bruce Momjian wrote:
> On Sun, Apr 5, 2026 at 11:10:52AM -0400, Tom Lane wrote:
> > Andres Freund <andres(at)anarazel(dot)de> writes:
> > > On 2026-04-05 16:09:57 +0200, Álvaro Herrera wrote:
> > >> On 2026-Apr-05, Bruce Momjian wrote:
> > >>> I just updated the wiki to handle this case because obviously
> > >>> Co-authored-by is listing more than just committers:
> >
> > > I think that is a completely unwarranted change for which there is zero
> > > concensus.
> >
> > Indeed. You exceeded your authority here.
> >
> > Even if there were consensus about making this change going forward,
> > the existing commit records were made under a different understanding.
> > You can't just say you're going to reinterpret them in a way that
> > excludes giving credit where credit is due.
>
> My email said:
>
> I need to know what to do for PG 19, and what to do for later major
> releases. I think Peter's point is why are people using Author
> and Co-authored-by in the same commits, and not just two Authors.
>
> Any changes to the wiki are going forward. While receiving emotional
> replies, I have not received answers to my specific questions.

Characterizing people disagreeing with you documenting a new, widely-disagreed
with, interpretation of Co-authored-by, while not going into all that much
detail about some questions you raised, as "emotional replies", does not seem
helpful.

Expecting detailed responses while other folks are working on getting stuff
committed before the feature freeze also seems like a bit much. And again,
that seems unrelated to the complain here that you're unilaterally making
decisions.

> What is the answer, both for PG 19, and going forward? I need an
> answer because I need rules to follow.

I think the answer is for you to roll back your changes, assume co-authorship
means co-authorship, and then, if you think we need another tag, start a
discussion about how what tag to use for "blame-but-no-credit-goes-to" going
forward. I would strongly recommend starting that discussion only once we're
well into the betas for 19, because it's just going to sow confusion if we
consider doing anything like this while still doing 19 stuff.

> I don't have a strong opinion but I do think we need a syntax for
> committers to indicate they modified a patch, might have introduced
> bugs, but don't want release note author credit, since I think several
> people have found that useful. Is that inaccurate?

I for one don't believe that's needed. Committers always are to co-blame for
stuff they commit, so when do you need to express blame-but-no-credit-goes-to?

> I updated the wiki text to now be:
>
> https://wiki.postgresql.org/wiki/Commit_Message_Guidance#Tags%3A_%22%3A%22
> Used to indicate the patch authors. "Co-authored-by:" should list
> individuals, particularly committers, who modified the patch but
> ------------------------
> should not be listed as authors in the release notes.
>
> I am updating the wiki text to try to get agreement on how to handle
> "Co-authored-by:" because no one else seems to be trying to address that
> question.

Documenting an new understanding is not getting a handle on the current
understanding. That makes absolutely no sense.

> Another question is, now that we have links to the commits, are the
> author names in the release notes only for giving credit, and not for
> knowing who was the feature author?

I don't see what the distinction you're making here is. Either the
co-authored-by person contributed substantially, or they shouldn't have been
named as a co-author.

> Is that a sufficient reason to keep the author names in the release notes?

It seems pretty crucial to me. We want people to make a living working on
postgres. For that they need to be known to have contributed to
postgres. Giving credit for nontrivial work is a huge part of that.

> Do other open source projects have names next to features?

Many do.

> I think those are the open questions.

I don't actually see any open questions here. You wanted to radically
reinterpret something for, as far as I can tell, no reason whatsoever, and
unsurprisingly got pushback. That's it.

Greetings,

Andres Freund