Date: Fri, 17 Oct 2025 14:35:12 -0500
From: Nathan Bossart <nathandbossart@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@postgresql.org, david@justatheory.com,
	Andrew Dunstan <andrew@dunslane.net>
Subject: Re: abi-compliance-check failure due to recent changes to
 pg_{clear,restore}_{attribute,relation}_stats()
Message-ID: <aPKacE_vakRY17FV@nathan>
References: <aPJ03E2itovDBcKX@nathan>
 <1713509.1760721320@sss.pgh.pa.us>
 <aPKFmL8VjOKNVLcw@nathan>
 <1723302.1760726712@sss.pgh.pa.us>
 <aPKUzk2xFhZBbWmz@nathan>
 <1728139.1760729230@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1728139.1760729230@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/aPKacE_vakRY17FV%40nathan>
Precedence: bulk

On Fri, Oct 17, 2025 at 03:27:10PM -0400, Tom Lane wrote:
> Nathan Bossart <nathandbossart@gmail.com> writes:
>> I've attached a first try.  You'll notice that I have borrowed heavily from
>> .git-blame-ignore-revs.  Some other things that might be worthwhile:
> 
> There would need to be an initial entry at the time the file is
> created, which would presumably point to some commit shortly before
> the .0 version stamp is applied (or maybe we'd choose to do it around
> rc1).  The mockup should include that.

Sure, makes sense.

> I'd be slightly inclined to have just one non-comment line, which
> is the active reference hash value, and all the rest be comments.
> The way you have it here requires the reading code to be smart
> about end-of-line comments, which is code complexity we don't need
> and doesn't seem amazingly legible either.  OTOH, the precedent of
> .git-blame-ignore-revs may be worth following regardless of our
> personal druthers.

That crossed my mind, too.  I'm personally not too concerned about small
deviations from .git-blame-ignore-revs, especially if it improves
machine/human readability.

-- 
nathan