From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
cc: pgsql-hackers@lists.postgresql.org,
        Manuel Rigger <rigger.manuel@gmail.com>
Subject: Re: Broken defenses against dropping a partitioning column
In-reply-to: <26099.1563560043@sss.pgh.pa.us>
References: <20190708143153.GA5929@alvherre.pgsql>
 <30041.1562597936@sss.pgh.pa.us> <26099.1563560043@sss.pgh.pa.us>
Comments: In-reply-to Tom Lane <tgl@sss.pgh.pa.us>
	message dated "Fri, 19 Jul 2019 14:14:03 -0400"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <11606.1563822144.1@sss.pgh.pa.us>
Date: Mon, 22 Jul 2019 15:02:24 -0400
Message-ID: <11607.1563822144@sss.pgh.pa.us>
Precedence: bulk

I wrote:
> Here's a proposed patch for that.  It's mostly pretty straightforward,
> except I had to add some recursion defenses in findDependentObjects that
> weren't there before.  But those seem like a good idea anyway to prevent
> infinite recursion in case of bogus entries in pg_depend.
> Per above, I'm envisioning applying this to HEAD and v12 with a catversion
> bump, and to v11 and v10 with no catversion bump.

Pushed.  Back-patching turned up one thing I hadn't expected: pre-v12
pg_dump bleated about circular dependencies.  It turned out that Peter
had already installed a hack in pg_dump to suppress that complaint in
connection with generated columns, so I improved the comment and
back-patched that too.

I nearly missed the need for that because of all the noise that
check-world emits in pre-v12 branches.  We'd discussed back-patching
eb9812f27 at the time, and I think now it's tested enough that doing
so is low risk (or at least, lower risk than the risk of not seeing
a failure).  So I think I'll go do that now.

			regards, tom lane