From: Nikolay Shaplov <n.shaplov@postgrespro.ru>
To: Alvaro Herrera <alvherre@2ndquadrant.com>, pgsql-hackers@postgresql.org
Subject: Re: [PROPOSAL] Move all am-related reloption code into
 src/backend/access/[am-name] and get rid of relopt_kind
Date: Thu, 26 May 2016 15:29:44 +0300
Message-ID: <1756891.5znqRt8KMB@nataraj-amd64>
User-Agent: KMail/4.14.1 (Linux/3.16.0-4-amd64; KDE/4.14.2; x86_64; ; )
In-Reply-To: <20160525180317.GA598084@alvherre.pgsql>
References: <20160525180317.GA598084@alvherre.pgsql>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="utf-8"
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org

=D0=92 =D0=BF=D0=B8=D1=81=D1=8C=D0=BC=D0=B5 =D0=BE=D1=82 25 =D0=BC=D0=B0=D1=
=8F 2016 14:03:17 =D0=92=D1=8B =D0=BD=D0=B0=D0=BF=D0=B8=D1=81=D0=B0=D0=BB=
=D0=B8:

> > > > >This all should me moved behind "access method" abstraction...
> > > >=20
> > > > +1 relopt_kind should be moved in am, at least. Or removed.
> > >=20
> > > Hm, but we have tablespace options too, so I'm not sure that using AM=
 as
> > > abstraction level is correct.
> >=20
> > We will use am for all indexes, and keep all the rest in relopotion.c f=
or
> > non-indexes. May be divided options catalog into sections one section f=
or
> > each kind.
> That makes sense.  I can review the patch later.
That would be great! ;-)


>=20
> > And as I also would like to use this code for dynamic attoptions later,=
 I
> > would like to remove relopt_kind at all. Because it spoils live in that
> > case.
> As I remember, Fabr=C3=ADzio (in CC) had a patch for dynamic reloptions, =
but
> there was some problem with it and we dumped it; see
> https://www.postgresql.org/message-id/CAFcNs+rqCq1H5eXW-cvdti6T-xo2STEkhj=
REx
> =3DOdmAkK5tiOOw@mail.gmail.com for previous discussion.

I've read the start of that thread. In my opinion Fabr=C3=ADzio offering qu=
ite=20
different thing. I am speaking about adding options to a new object (access=
=20
method or later operator class). You add these options when you create this=
=20
object and you have a monopoly of defining options for it.

Fabr=C3=ADzio offers adding options for relkind that already exists. This s=
eems to=20
be a complex thing. You should resolve conflicts between two extensions tha=
t=20
want to define same option for example. Somehow deal with the situation whe=
n=20
the extension is dropped. Should we remove reloptions created by that=20
extension from pg_class then?

And moreover, as far as I understand reloptions is about how does relation=
=20
operates. And the external extension would not affect this at all. So we ar=
e=20
speaking not about options of a relation, but about extension that want to=
=20
store some info about some relation. I think in general this extension shou=
ld=20
store it's info into it's own data structures.=20

May be there is cases when you will really need such behavior. But it is qu=
ite=20
different task, have almost nothing in common with things I am going to do =
:-)

--=20
Nikolay Shaplov
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company


--=20
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers