From: Tom Lane <tgl@sss.pgh.pa.us>
To: Justin Pryzby <pryzby@telsasoft.com>
cc: pgsql-hackers@lists.postgresql.org
Subject: Re: optimizer/clauses.h needn't include access/htup.h
In-reply-to: <20201123205505.GJ24052@telsasoft.com>
References: <20201123205505.GJ24052@telsasoft.com>
Comments: In-reply-to Justin Pryzby <pryzby@telsasoft.com>
	message dated "Mon, 23 Nov 2020 14:55:05 -0600"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <166714.1606168841.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Mon, 23 Nov 2020 17:00:41 -0500
Message-ID: <166715.1606168841@sss.pgh.pa.us>
Precedence: bulk

Justin Pryzby <pryzby@telsasoft.com> writes:
> It was only needed between these:
> commit a8677e3ff6bb8ef78a9ba676faa647bba237b1c4
> commit f09346a9c6218dd239fdf3a79a729716c0d305bd

Hm, you're right.  Removed.

> I noticed while looking at "what includes what" and wondered if some of =
these
> are kind of "modularity violations". =


Yeah.  I've ranted before that we ought to have some clearer idea of
module layering within the backend, and avoid cross-header inclusions
that would break the layering.  This particular case didn't really
do so, I suppose, since htup.h would surely be on a lower level than
the optimizer.  But it still seems nicer to not have that inclusion.

Anyway, if you're feeling motivated to explore a more wide-ranging
refactoring, by all means have a go at it.

			regards, tom lane