Feedback-ID: id4a34324:Fastmail
Date: Fri, 20 Mar 2026 12:27:49 -0400
From: Andres Freund <andres@anarazel.de>
To: Michael Paquier <michael@paquier.xyz>
Cc: Nathan Bossart <nathandbossart@gmail.com>,
	shihao zhong <zhong950419@gmail.com>,
 PostgreSQL-development <pgsql-hackers@postgresql.org>
Subject: Re: Fixes inconsistent behavior in vacuum when it processes multiple
 relations
Message-ID: <rzxpxod4c4la62yvutyrvgoyilrl2fx55djaf2suidy7np5m6c@3l2ln476eadh>
References: <aFRzYhOTZcRgKPLu@nathan>
 <aFS6m1tNH2sGMQ61@paquier.xyz>
 <aFWznvZtYZ3VJxgi@nathan>
 <aFiWUzW8VChL-hGg@paquier.xyz>
 <aFl1AB9ofnQrrckV@nathan>
 <aFn82kXgvwKHncs-@paquier.xyz>
 <aFrSlSz9fi47j5Sr@nathan>
 <aFtRd1Q4QuCWcg5k@paquier.xyz>
 <aFwT31Rw-Oj-bgO9@nathan>
 <aGIzjYsRrMCEADc5@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <aGIzjYsRrMCEADc5@paquier.xyz>
Archived-At: 
 <https://www.postgresql.org/message-id/rzxpxod4c4la62yvutyrvgoyilrl2fx55djaf2suidy7np5m6c%403l2ln476eadh>
Precedence: bulk

Hi,

On 2025-06-30 15:49:49 +0900, Michael Paquier wrote:
> On Wed, Jun 25, 2025 at 10:21:03AM -0500, Nathan Bossart wrote:
> > On Wed, Jun 25, 2025 at 10:31:35AM +0900, Michael Paquier wrote:
> >> Attached is the remaining patch for HEAD, planned once v19 opens, and
> >> the tests I have used on the back-branches as a txt to not confuse the
> >> CI, for reference.
> >
> > Looks reasonable to me.
>
> Thanks.  Applied on HEAD, then.

I was just looking at the includes for tableam.h - a quite widely included
header - and got rather sad seeing vacuum.h in there, as it has a quite large
dependency tree and tableam.h is obviously widely included.

I was surprised that I would have added such a wide include when working on
tableam. Turns out I hadn't - it was 2252fcd4276c.


I think this change was quite ill considered. Look at the dependency tree of
vacuum.h:

https://doxygen.postgresql.org/vacuum_8h.html


That's a *lot* of headers to pull into something as widely included as
tableam.h


Why wasn't it enough to add const markers and keep passing by pointer?


Greetings,

Andres