From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
cc: Andres Freund <andres@anarazel.de>,
        Pg Hackers <pgsql-hackers@lists.postgresql.org>,
        Simon Riggs <simon.riggs@enterprisedb.com>,
        Tomas Vondra <tomas.vondra@enterprisedb.com>,
        Zhihong Yu <zyu@yugabyte.com>,
        Daniel Westermann <dwe@dbi-services.com>,
        Amit Langote <amitlangote09@gmail.com>,
        Justin Pryzby <pryzby@telsasoft.com>, Japin Li <japinli@hotmail.com>,
        Erik Rijkers <er@xs4all.nl>,
        Jaime Casanova <jcasanov@systemguards.com.ec>
Subject: Re: support for MERGE
In-reply-to: <202202271735.4o6uyat2eh7w@alvherre.pgsql>
References: <202202271735.4o6uyat2eh7w@alvherre.pgsql>
Comments: In-reply-to Alvaro Herrera <alvherre@alvh.no-ip.org>
	message dated "Sun, 27 Feb 2022 14:35:13 -0300"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2373866.1645983776.1@sss.pgh.pa.us>
Date: Sun, 27 Feb 2022 12:42:56 -0500
Message-ID: <2373867.1645983776@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2373867.1645983776%40sss.pgh.pa.us>
Precedence: bulk

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> I think we should make a decision on code arrangement here.  From my
> perspective, MERGE isn't its own executor node; rather it's just another
> "method" in ModifyTable.  Which makes sense, given that all it does is
> call parts of INSERT, UPDATE, DELETE which are the other ModifyTable
> methods.  Having a separate file doesn't strike me as great, but on the
> other hand it's true that merely moving all the execMerge.c code into
> nodeModifyTable.c makes the latter too large.  However I don't want to
> create a .h file that means exposing all those internal functions to the
> outside world.  My ideal would be to have each INSERT, UPDATE, DELETE,
> MERGE as its own separate .c file, which would be #included from
> nodeModifyTable.c.  We don't use that pattern anywhere though.  Any
> opposition to that?  (The prototypes for each file would have to live in
> nodeModifyTable.c itself.)

Yeah, I don't like that.  The point of having separate .c files is that
you know that there's no interactions of non-global symbols across files.
This pattern breaks that assumption, making it harder to see what's
connected to what; and what's it buying exactly?  I'd rather keep all the
ModifyTable code in one .c file, even if that one is bigger than our
usual practice.

			regards, tom lane