Re: Extension Templates S03E11

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Simon Riggs <simon(at)2ndQuadrant(dot)com>
Cc: Stephen Frost <sfrost(at)snowman(dot)net>, Jeff Davis <pgsql(at)j-davis(dot)com>, Dimitri Fontaine <dimitri(at)2ndquadrant(dot)fr>, Robert Haas <robertmhaas(at)gmail(dot)com>, Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>, Peter Eisentraut <peter_e(at)gmx(dot)net>, Alvaro Herrera <alvherre(at)2ndQuadrant(dot)com>, Boszormenyi Zoltan <zb(at)cybertec(dot)at>, Thom Brown <thom(at)linux(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Extension Templates S03E11
Date: 2013-12-17 23:42:51
Message-ID: 444.1387323771@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Simon Riggs <simon(at)2ndQuadrant(dot)com> writes:
> On 17 December 2013 18:32, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
>> I have no idea where you're going with this, but I *do* object to
>> sticking an SQL script which defines a bunch of objects into a catalog
>> table *right next to where they are properly defined*. There's just no
>> sense in it that I can see, except that it happens to mimic what we do
>> today- to no particular purpose.

> The purpose is clear: so it is part of the database backup. It's a
> fairly boring purpose, not fancy at all. But it is a purpose, indeed
> *the* purpose.

The point Stephen is making is that it's just as easy, and far more
reliable, to dump the package-or-whatever-you-call-it by dumping the
definitions of the contained objects, as to dump it by dumping the text
blob it was originally created from. So I don't see a lot of merit
to claiming that we need to keep the text blob for this purpose.

We did it differently for extensions in part because you can't dump a .so
as a SQL command, so dump-the-contained-objects wasn't going to be a
complete backup strategy in any case. But for a package containing only
SQL objects, that's not a problem.

> We aim to have the simplest implementation that meets the stated need
> and reasonable extrapolations of that. Text in a catalog table is the
> simplest implementation. That is not a reason to reject it, especially
> when we aren't suggesting a viable alternative.

The first part of this assertion is debatable, and the claim that no
viable alternative has been suggested is outright wrong.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Cédric Villemain 2013-12-17 23:59:14 Re: pg_rewarm status
Previous Message Simon Riggs 2013-12-17 23:12:29 Re: Extension Templates S03E11