Content-Type: text/plain;
	charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3776.700.51\))
Subject: Re: PostgreSql and VMS operating System
From: Christophe Pettus <christophe.pettus@pgexperts.com>
In-Reply-To: <Z5O7HX4QQPQebEgo@momjian.us>
Date: Fri, 24 Jan 2025 08:20:30 -0800
Cc: Daniel Gustafsson <daniel@yesql.se>,
 pgsql-advocacy@lists.postgresql.org,
 Bruce Momjian <bruce@momjian.us>
Content-Transfer-Encoding: quoted-printable
Message-Id: <3CD15C4F-5050-47AB-A2A2-FD491381CC53@pgexperts.com>
References: 
 <CAF14C7Cu-Zjyd59EEnXPLHpUkGoea9AmB9vAn-U2Y1QHDsjCOQ@mail.gmail.com>
 <E0D44021-B723-41C9-B38F-237B7A48D7D4@yesql.se>
 <CAF14C7DZG8mUcaED=jO092b-USQ6sre-LSuaF1Fp3c3iAM314Q@mail.gmail.com>
 <CFD8CDAD-4456-48A9-A582-616ADBFAED3B@pgexperts.com>
 <Z5O7HX4QQPQebEgo@momjian.us>
To: Vish Penmetsa <vish.penmetsa@gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/3CD15C4F-5050-47AB-A2A2-FD491381CC53%40pgexperts.com>
Precedence: bulk


> On Jan 24, 2025, at 08:09, Bruce Momjian <bruce@momjian.us> wrote:
> Our normal development flow is:
>=20
> https://wiki.postgresql.org/wiki/Todo#Development_Process
> Desirability -> Design -> Implement -> Test -> Review -> Commit
>=20
> so I would focus on Desirability at this point.  The next question is
> whether the demand justifies the code changes.

To expand on this just a bit, you do not need anyone's permission to =
just take the existing PostgreSQL code base and get it running on VMS.  =
PostgreSQL is much more truly open-source than a lot of "open source" =
projects.  However, what you then have is a fork, rather than the =
mainline PostgreSQL codebase.  Since it is extremely unlikely that you =
can just compile the current code on VMS and have it work, some patches =
will be required to do conditional compilation and provide support code =
for VMS, and perhaps some additions to and modifications of the =
documentation and test suite.

The part where you need to negotiate with the community is whether or =
not to accept those changes back into the mainline.  To be extreme, if =
the difference is one small conditionally-compiled primitive and the =
code otherwise runs just as it did before (including performance) on the =
currently-supported platforms, the chance is pretty good; if there is =
extensive surgery that requires a lot of changes, or if those changes =
have performance or maintenance impact on the mainline, the chance is =
much lower.
--
Christophe Pettus / christophe.pettus@pgexperts.com
Chief Executive Officer / PGX Inc. / 24x7 Support, Consulting, =
Development / pgexperts.com
See us at: SCaLE, March 6-9, Los Angeles / Nordic PgDay, Copenhagen, =
March 18