Re: attndims, typndims still not enforced, but make the value within a sane threshold

On Mon, Dec 23, 2024 at 01:56:24PM +0900, Michael Paquier wrote:
> On Thu, Dec 12, 2024 at 03:40:32PM +0800, jian he wrote:
> > remove pg_type typndims column patch attached.
> 
> FWIW, I have been paying more attention to applications that may use
> this attribute and bumped into quite a few cases that use quals based
> on (t.typndims > 0 AND t.typbasetype > 0) to check that they're
> dealing with domains over array types.  So even removing this switch
> would be unwise, I am afraid..

Well, I would be more excited about keeping the fields if they actually
were reliable in recording information.  This email from November 2023
explains the limitations of attndims:

   https://www.postgresql.org/message-id/flat/20150707072942.1186.98151@wrigleys.postgresql.org

*  dimensions not dumped by pg_dump
*  dimensions not propagated by CREATE TABLE ... (LIKE)
*  dimensions not propagated by CREATE TABLE AS
*  dimensions not displayed by psql

So, if users are referencing attndims and the values are not accurate,
how useful are they?  I was suggesting removal so people would stop
relying on inaccurate information, or we correct attndims to be
accurate.  Allowing people to continue relying on inaccurate information
seems like the worst of all options.

-- 
  Bruce Momjian  <bruce@momjian.us>        https://momjian.us
  EDB                                      https://enterprisedb.com

  Do not let urgent matters crowd out time for investment in the future.