OpenACS vs Postgres

From: Andrew Dunstan <andrew(at)dunslane(dot)net>
To: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: OpenACS vs Postgres
Date: 2009-10-20 16:24:42
Message-ID: 4ADDE44A.3040309@dunslane.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers


OpenACS uses three compatibility settings in its recommended config, two
of which it has been proposed to remove.

I have been investigating all three.

1. default_with_oids= true
While it's not proposed to remove this option, its use seems
quite unnecessary in OpenACS. AFAICS, the code uses oids on
exactly two tables, so these can be created with OIDS, and the
flag turned off. The use it makes of OIDS is kludgey anyway,
but that need not concern us.
2. regex_flavor = 'extended'
I have not found anywhere in the current release that actually
uses POSIX regex matching, so I suspect this was recommended
from an over abundance of caution, or else it was used in code
that is no longer present.
3. add_missing_from = on
In a particularly hacky gadget, OpenACS creates views that
look like this:

create view foo_seq as select nextval('real_foo_seq') as
nextval;

which then lets them use Oracle-like syntax such as:

select foo_seq.nextval from dual;

relying on add_missing_from to do the right thing. I have
discussed this with my client that uses OpenACS, and they will
patch OpenACS as necessary when they upgrade to Postgres 8.5
to remove this kludge.

So, bottom line, I withdraw my objection to the proposed changes.
Personally, I'll be quite glad to see the back of add_missing_from, and
so will my client.

cheers

andrew

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2009-10-20 16:27:11 Re: Could postgres be much cleaner if a future release skipped backward compatibility?
Previous Message Magnus Hagander 2009-10-20 16:18:25 Re: Application name patch - v2