MIME-Version: 1.0
References: <CAN-LCVMV3m+pzSDqOLyzgVRXA2zJ94n_T7kyAc2Z5JJpH018sA@mail.gmail.com>
 <20230206104917.sipa7nzue5lw2e6z@alvherre.pgsql> <CAN-LCVOv5is9dUn1U3KYy4hAhVK7tx6a7goOFmgbrGM7ZSAxRA@mail.gmail.com>
In-Reply-To: <CAN-LCVOv5is9dUn1U3KYy4hAhVK7tx6a7goOFmgbrGM7ZSAxRA@mail.gmail.com>
From: Aleksander Alekseev <aleksander@timescale.com>
Date: Mon, 6 Feb 2023 17:38:34 +0300
Message-ID: <CAJ7c6TMZ7xySLzhygvzezPysyM5aZ6OjO8zHSUTO+GMBeL1a3g@mail.gmail.com>
Subject: Re: Pluggable toaster
To: pgsql-hackers <pgsql-hackers@postgresql.org>
Cc: Alvaro Herrera <alvherre@alvh.no-ip.org>, Nikita Malakhov <hukutoc@gmail.com>, 
	vignesh C <vignesh21@gmail.com>, Jacob Champion <jchampion@timescale.com>, 
	Andres Freund <andres@anarazel.de>, Matthias van de Meent <boekewurm+postgres@gmail.com>, 
	Robert Haas <robertmhaas@gmail.com>, Greg Stark <stark@mit.edu>, Teodor Sigaev <teodor@sigaev.ru>
Content-Type: text/plain; charset="UTF-8"
Archived-At: <https://www.postgresql.org/message-id/CAJ7c6TMZ7xySLzhygvzezPysyM5aZ6OjO8zHSUTO%2BGMBeL1a3g%40mail.gmail.com>
Precedence: bulk

Hi,

> The main reason behind this decision is that keeping the first implementation
> on the side of the vanilla (I mean rebasing it) over time is very difficult due
> to the very invasive nature of this solution.
>
> So we decided to reduce changes in the core to the minimum necessary
> to make it available through the hooks, because the hooks part is very
> lightweight and simple to keep rebasing onto the vanilla core. We plan
> to keep this extension free with the PostgreSQL license, so any PostgreSQL
> user could benefit from the TOAST on steroids, and sometimes in the
> future it will be a much simpler task to integrate the Pluggable TOAST into
> the vanilla, along with our advanced TOAST implementations which
> we plan to keep under Open Source licenses too.

That's great to hear. I'm looking forward to the link to the
corresponding GitHub repository. Please let us know when this effort
will be available for testing and benchmarking!

I would like to point out however that there were several other pieces
of feedback that could have been missed:

* No one wants to see this as an extension. This was my original
proposal (adding ZSON to /contrib/) and it was rejected. The community
explicitly wants this to be a core feature with its syntax,
autocompletion, documentation and stuff.
* The community wants the feature to have a simple implementation. You
said yourself that the idea of type-aware TOASTers is very invasive,
and I completely agree.
* People also want this to be simple from the user perspective, as
simple as just CREATE COMPRESSED TABLE ... [USING lz4|zstd];

At least this is my personal summary/impression from following the mailing list.

Anyhow since we are back to the stage where we discuss the RFC I
suggest continuing it in the compression dictionaries thread [1] since
we made noticeable progress there already.

[1]: https://postgr.es/m/CAJ7c6TOtAB0z1UrksvGTStNE-herK-43bj22%3D5xVBg7S4vr5rQ%40mail.gmail.com

-- 
Best regards,
Aleksander Alekseev