Re: pgsql: Use CompactAttribute more often, when possible

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: David Rowley <drowley(at)postgresql(dot)org>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: pgsql: Use CompactAttribute more often, when possible
Date: 2025-10-23 12:21:22
Message-ID: CA+TgmoYUA9Prs-_-+wgnsLAAzGAKYUxzkUp6m2hg882VY6ziPw@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-committers pgsql-hackers

On Tue, Oct 21, 2025 at 6:36 PM David Rowley <drowley(at)postgresql(dot)org> wrote:
> There are some locations where I've left the code using
> FormData_pg_attribute. These are mostly in the ALTER TABLE code. Using
> CompactAttribute here seems more risky as often the TupleDesc is being
> changed and those changes may not have been flushed to the
> CompactAttribute yet.

David, I've noticed several times lately you've documented in commit
messages why you thought going further with some change would be
risky, and I really like that. I think sometimes we (as committers)
are sometimes reluctant to document cases where we weren't sure about
something or didn't know something, but IMHO this sort of
documentation of perceived risk is a huge help to anyone who is
thinking of working in this area later. Whether or not they agree with
your conclusions, they'll find it helpful to know your thought
process.

Thanks,

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

In response to

Responses

Browse pgsql-committers by date

  From Date Subject
Next Message Tom Lane 2025-10-23 15:48:02 pgsql: Fix resource leaks in PL/Python error reporting, redux.
Previous Message Amit Kapila 2025-10-23 08:43:54 pgsql: Introduce "REFRESH SEQUENCES" for subscriptions.

Browse pgsql-hackers by date

  From Date Subject
Next Message Sergey Prokhorenko 2025-10-23 12:34:43 Add uuid_to_base32hex() and base32hex_to_uuid() built-in functions
Previous Message Arseniy Mukhin 2025-10-23 11:39:56 Re: amcheck: support for GiST