Re: pg_dump bug for extension owned tables

From: Fabrízio de Royes Mello <fabriziomello(at)gmail(dot)com>
To: Andrew Dunstan <andrew(dot)dunstan(at)2ndquadrant(dot)com>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: pg_dump bug for extension owned tables
Date: 2020-07-13 18:46:18
Message-ID: CAFcNs+okXey4PdnENWZ_GkrUDCgG=ArTvbCq0TuQNOXebTpzug@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Jul 13, 2020 at 3:29 PM Andrew Dunstan <
andrew(dot)dunstan(at)2ndquadrant(dot)com> wrote:
>
>
> On 7/13/20 10:52 AM, Fabrízio de Royes Mello wrote:
> >
> > On Sat, Jul 11, 2020 at 8:07 PM Andrew Dunstan
> > <andrew(dot)dunstan(at)2ndquadrant(dot)com
> > <mailto:andrew(dot)dunstan(at)2ndquadrant(dot)com>> wrote:
> > >
> > >
> > > On 6/26/20 2:10 PM, Fabrízio de Royes Mello wrote:
> > > >
> > > > On Fri, Jun 26, 2020 at 11:55 AM Fabrízio de Royes Mello
> > > > <fabriziomello(at)gmail(dot)com <mailto:fabriziomello(at)gmail(dot)com>
> > <mailto:fabriziomello(at)gmail(dot)com <mailto:fabriziomello(at)gmail(dot)com>>>
wrote:
> > > > >
> > > > >
> > > > > On Fri, Jun 26, 2020 at 11:24 AM Andrew Dunstan
> > > > <andrew(dot)dunstan(at)2ndquadrant(dot)com
> > <mailto:andrew(dot)dunstan(at)2ndquadrant(dot)com>
> > > > <mailto:andrew(dot)dunstan(at)2ndquadrant(dot)com
> > <mailto:andrew(dot)dunstan(at)2ndquadrant(dot)com>>> wrote:
> > > > > >
> > > > > >
> > > > > > On 6/26/20 9:57 AM, Andrew Dunstan wrote:
> > > > > > > It appears that for extension owned tables
> > tbinfo.attgenerated isn't
> > > > > > > being properly populated, so line 2050 in REL_12_STABLE, which
> > > > is line
> > > > > > > 2109 in git tip, is failing.
> > > > > > >
> > > > > > >
> > > > > >
> > > > > > Should have mentioned this is in src/bin/pg_dump/pg_dump.c
> > > > > >
> > > > >
> > > > > Having a look on it.
> > > > >
> > > >
> > > > Seems when qualify the schemaname the the "tbinfo->interesting"
field
> > > > is not setted for extensions objects, so the getTableAttrs can't
fill
> > > > the attgenerated field properly.
> > > >
> > > > I'm not 100% sure it's the correct way but the attached patch works
> > > > for me and all tests passed. Maybe we should add more TAP tests?
> > > >
> > > >
> > >
> > >
> > > I just tried this patch out on master, with the test case I gave
> > > upthread. It's not working, still getting a segfault.
> > >
> >
> > Ohh really sorry about it... my bad... i completely forgot about it!!!
> >
> > Due to my rush I ended up adding the wrong patch version. Attached the
> > correct version.
> >
> > Will add TAP tests to src/test/modules/test_pg_dump
>
>
> yeah, that's the fix I came up with too. The only thing I added was
> "Assert(tbinfo->attgenerated);" at about line 2097.
>

Cool.

>
> Will wait for your TAP tests.
>

Actually I've sent it already but it seems to have gone to the moderation
queue.

Anyway attached with your assertion and TAP tests.

Regards,

--
Fabrízio de Royes Mello Timbira - http://www.timbira.com.br/
PostgreSQL: Consultoria, Desenvolvimento, Suporte 24x7 e Treinamento

Attachment Content-Type Size
pg_dump_segfault_on_extension_and_schema_v4.patch text/x-patch 3.0 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Geoghegan 2020-07-13 18:50:10 Re: Default setting for enable_hashagg_disk
Previous Message Justin Pryzby 2020-07-13 18:43:16 Re: Default setting for enable_hashagg_disk