Date: Mon, 20 Aug 2018 22:19:49 -0300
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: David Rowley <david.rowley@2ndquadrant.com>
Cc: abelisto@gmail.com, pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #15343: Segmentation fault using pg_dump with
 --exclude-table if table contains identity column
Message-ID: <20180821011949.r5ogh32qluf56na6@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: 
 <CAKJS1f_RzF8s5Osi2ae3fMMnBOv=zZSw51OJ1iPur=07d0Sw-w@mail.gmail.com>
User-Agent: NeoMutt/20170306-137-4415bd-dirty (1.8.0)
Precedence: bulk

On 2018-Aug-21, David Rowley wrote:

> > To reproduce:
> > $ createdb foo
> > $ psql foo -c "create table bar(i int generated by default as identity
> > primary key);"
> > $ pg_dump foo --exclude-table=bar
> > Segmentation fault (core dumped)

> I think the fix depends exactly on what we want the behaviour to be
> here.  Should dumpSequence() still dump out a CREATE SEQUENCE
> statement for this, or should we just ignore the sequence?  I imagine
> we should just ignore identity sequences when we're not going to dump
> the table that owns them. It looks like that would require a small
> modification in getOwnedSeqs().

Yeah, we should omit the sequence too.  I don't see any principled
reason to think that the sequence is interesting on its own.

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services