Horton

Re: BUG #15310: pg_upgrade dissociates event triggers from extensions

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Michael Paquier <michael@paquier.xyz>
Cc: Haribabu Kommi <kommi.haribabu@gmail.com>, nickbarnes01@gmail.com, pgsql-bugs@lists.postgresql.org
Date: 2018-08-07T20:38:03Z
Lists: pgsql-bugs 
Michael Paquier <michael@paquier.xyz> writes:
> That sounds like a bug to me.  Could you add more tests to
> src/test/modules/test_pg_dump with an event trigger part of an extension
> with a binary upgrade test?

It doesn't seem especially useful to me to lock this particular barn
door behind the horse.  What would be great is a way of cross-checking
that the grammar, pg_dump, and the various callers of
recordDependencyOnCurrentExtension are all on the same page about
which objects can be extension members.  I see no way to automate that
unfortunately :-(

			regards, tom lane

Commits

  1. Don't record FDW user mappings as members of extensions.

  2. Fix pg_upgrade to handle event triggers in extensions correctly.