Re: Implementing SQL/PSM for PG 8.2

From: Andrew Dunstan <andrew(at)dunslane(dot)net>
To: Denis Lussier <denis(at)enterprisedb(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Implementing SQL/PSM for PG 8.2
Date: 2005-06-26 20:44:13
Message-ID: 42BF139D.6090705@dunslane.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers


Is the intention here to make PSM a first class language (i.e. handled
by the main dbengine scanner/parser) of just another PL? If the latter
it seems far less worth doing. Doing this as a first class language,
however, would be great, just great.

As for pgfoundry, I think it's fair to say (from my various perspectives
as a pgfoundry admin, owner of a PL project there, and general hacker)
that experience is mixed on things that have close backend integration
requirements. In particular, I would advise you to conduct pretty much
all discussions abou the project on the -hackers list for a project like
this. That way you avoid giving anyone surprises, and you will get the
best and most wide-ranging advice and feedback.

cheers

andrew

Denis Lussier wrote:

>Hi All,
>
>My company (EnterpriseDB) is very interested in helping to make ANSI-ISO SQL Stored Procedures part of standard BSD Postgres. The SQL/PSM standard is currently used in DB2 and is being implemented in MySQL 5.0. Note that I'm NOT a big fan of adding Oracle compatibility to PL/pgSQL, but, I'm biased in this regard because EnterpriseDB's SPL (Superset Procedural Language) supports Redwood (pl/sql) and Redmond (transact-sql) style programming.
>
>For various technical and backward compatibility reasons, I don't think SQL/PSM should be a replacement for PL/pgSQL. Although I do think it should heavily leverage the solid foundation afforded by the PL/pgSQL code base. I think it should start as a separate project on PgFoundry. Once it is working and fully tested and rock solid and proven... I think it should then be considered to become part of the core & installed by default alongside plpgsql.
>
>Please note that this is all appropriate for 8.2, because changes to the server side code are necessary to support ANSI stored proc signatures and flexible out/inout parameter passing. EnterpriseDB will publish those suggested server changes for review so that work can begin on plsqlpsm sooner rather than later.
>
>What do y'all think?? I believe the first step is for us to create "plsqlpsm" as a BSD project in PgFoundry.
>
>--Luss
>
>
>
>

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2005-06-26 22:04:15 Re: [HACKERS] language handlers in public schema?
Previous Message Andrew Dunstan 2005-06-26 20:31:51 Re: language handlers in public schema?