Re: BUG #18545: \dt breaks transaction, calling error when executed in SET SESSION AUTHORIZATION

Robert Haas <robertmhaas@gmail.com>

From: Robert Haas <robertmhaas@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Андрей Рачицкий <therealgofman@mail.ru>, pgsql-bugs@lists.postgresql.org
Date: 2024-08-06T18:48:06Z
Lists: pgsql-bugs, pgsql-hackers
On Sun, Aug 4, 2024 at 6:08 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> However, what the farm found is that assign_client_encoding
> is flat-out broken.  It's ignoring the first commandment
> for GUC hooks, which is "Thy assign hooks shalt not fail".
> (If we didn't need that, there wouldn't be a separation
> between check hooks and assign hooks in the first place.)

Interesting. Looks like my mistake, dating to
10c0558ffefcd12bf1d3dc35587eba41d1ce4571. I'm honestly kind of
surprised that nobody discovered this problem for 8 years. I would
have expected it to cause more problems.

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



Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →
  1. Fix parallel worker tracking of new catalog relfilenumbers.

  2. Restore relmapper state early enough in parallel workers.

  3. Allow parallel workers to cope with a newly-created session user ID.

  4. Clean up handling of client_encoding GUC in parallel workers.

  5. Fix rd_firstRelfilenodeSubid for nailed relations, in parallel workers.

  6. Fix several mistakes around parallel workers and client_encoding.