Re: cataloguing NOT NULL constraints

On Wed, May 15, 2024 at 09:50:36AM +0200, Álvaro Herrera wrote:
> On 2024-May-14, Bruce Momjian wrote:
>
> > Turns out these commits generated a single release note item, which I
> > have now removed with the attached committed patch.
>
> Hmm, but the commits about not-null constraints for domains were not
> reverted, only the ones for constraints on relations. I think the
> release notes don't properly address the ones on domains. I think it's
> at least these two commits:
>
> > -Author: Peter Eisentraut <peter(at)eisentraut(dot)org>
> > -2024-03-20 [e5da0fe3c] Catalog domain not-null constraints
> > -Author: Peter Eisentraut <peter(at)eisentraut(dot)org>
> > -2024-04-15 [9895b35cb] Fix ALTER DOMAIN NOT NULL syntax
>
> It may still be a good idea to make a note about those, at least to
> point out that information_schema now lists them. For example, pg11
> release notes had this item

Let me explain what I did to adjust the release notes. I took your
commit hashes, which were longer than mine, and got the commit subject
text from them. I then searched the release notes to see which commit
subjects existed in the document. Only the first three did, and the
release note item has five commits.

The then tested if the last two patches could be reverted, and 'patch'
thought they could be, so that confirmed they were not reverted.

However, there was no text in the release note item that corresponded to
the commits, so I just removed the entire item.

What I now think happened is that the last two commits were considered
part of the larger NOT NULL change, and not worth mentioning separately,
but now that the NOT NULL part is reverted, we might need to mention
them.

I rarely handle such complex cases so I don't think I was totally
correct in my handling. Let's get a reply to Peter Eisentraut's
question and we can figure out what to do.

--
Bruce Momjian <bruce(at)momjian(dot)us> https://momjian.us
EDB https://enterprisedb.com

Only you can decide what is important to you.