Horton

Re: Binary support for pgoutput plugin

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Dave Cramer <davecramer@gmail.com>
Cc: Daniel Gustafsson <daniel@yesql.se>, Petr Jelinek <petr@2ndquadrant.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, Michael Paquier <michael@paquier.xyz>, Dmitry Dolgov <9erthalion6@gmail.com>, Thomas Munro <thomas.munro@gmail.com>, Tomas Vondra <tomas.vondra@2ndquadrant.com>, Petr Jelinek <petr.jelinek@2ndquadrant.com>, Andres Freund <andres@anarazel.de>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2020-07-14T18:36:37Z
Lists: pgsql-hackers 
Dave Cramer <davecramer@gmail.com> writes:
> On Tue, 14 Jul 2020 at 12:59, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> So I started looking through this seriously, and my first question
>> is why do the docs and code keep saying that "base types" are sent
>> in binary?  Why not just "data"?  Are there any cases where we
>> don't use binary format, if the subscription requests it?

> Well for some reason I thought there were some types that did not have send
> and receive functions.

There are, but they're all base types, so this terminology is still
unhelpful ;-).

It'd be possible for the sender to send binary for columns it has a
typsend function for, and otherwise send text.  However, this only helps
if the receiver has receive functions for all those types; in
cross-version cases they might disagree about which types can be sent
in binary.  (Hm ... maybe we could have the receiver verify that it has
typreceive for every column included in its version of the table, before
asking for binary mode?)

			regards, tom lane

Commits

  1. Weaken type-OID-matching checks in array_recv and record_recv.

  2. Fix construction of updated-columns bitmap in logical replication.

  3. Allow logical replication to transfer data in binary format.

  4. Phase 2 pgindent run for v12.