From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
cc: Michael Paquier <michael@paquier.xyz>, pgsql-hackers@lists.postgresql.org
Subject: Re: Gripes about walsender command processing
In-reply-to: <20200916163633.GA9907@alvherre.pgsql>
References: <20200916163633.GA9907@alvherre.pgsql>
Comments: In-reply-to Alvaro Herrera <alvherre@2ndquadrant.com>
	message dated "Wed, 16 Sep 2020 13:36:33 -0300"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1633416.1600274669.1@sss.pgh.pa.us>
Date: Wed, 16 Sep 2020 12:44:29 -0400
Message-ID: <1633417.1600274669@sss.pgh.pa.us>
Precedence: bulk

Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> On 2020-Sep-15, Alvaro Herrera wrote:
>> I overlooked this in 2f9661311b83.  From this perspective, I agree it
>> looks wrong.  We still have to send *some* completion tag (the 'C'
>> message), but maybe we can invent a separate entry point in dest.c for
>> that -- EndReplicationCommand() or some such -- that takes values from a
>> separate enum?

> It seems simpler than that actually; we don't need to build a lot of
> infrastructure.

This looks moderately reasonable to me.  However, with the process
title reporting I want to add, we're going to end up with a switch
that looks like

 		case T_IdentifySystemCmd:
+			set_ps_display("IDENTIFY_SYSTEM");
 			IdentifySystem();
+			EndReplicationCommand("IDENTIFY_SYSTEM");
 			break;
 
 		case T_BaseBackupCmd:
+			set_ps_display("BASE_BACKUP");
 			PreventInTransactionBlock(true, "BASE_BACKUP");
 			SendBaseBackup((BaseBackupCmd *) cmd_node);
+			EndReplicationCommand("BASE_BACKUP");
 			break;

which is starting to look a bit repetitive and copy-pasteo-prone.
I don't see an easy way to improve on it though.  The only obvious
alternative would be to put another switch before the main one that
just fills a "const char *cmdtag" variable, but that seems ugly.

			regards, tom lane