Re: Gripes about walsender command processing

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Michael Paquier <michael(at)paquier(dot)xyz>
Cc: pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: Gripes about walsender command processing
Date: 2020-09-15 03:34:44
Message-ID: 1236185.1600140884@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Michael Paquier <michael(at)paquier(dot)xyz> writes:
> On Mon, Sep 14, 2020 at 12:51:39PM -0400, Tom Lane wrote:
>> (Is there a better way to get the name of a replication command?
>> I didn't look very hard for one.)

> Wouldn't that just be in cmdtaglist.h, but extended for nodes used for
> replication commands? Then you could just call CreateCommandTag() to
> get the command string to print as postgres.c does. There is already
> one for repslot drop, in some way. This would have the advantage to
> just call once set_ps_display() before the switch split.

Mmm, not sure whether having CreateCommandTag know about replication
commands is a good thing or not. We certainly could do it like that,
since there's only one namespace of NodeTag values, but conceptually
it feels a bit weird to me. There's a lot of other infrastructure
for SQL command nodes that we're surely never going to build out for
replication commands, so should we do it in CreateCommandTag?

Anybody else have an opinion about it?

(I don't quite follow your comment about repslot drop, btw.)

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Geoghegan 2020-09-15 03:48:05 Re: logtape.c stats don't account for unused "prefetched" block numbers
Previous Message Tom Lane 2020-09-15 03:24:55 Re: Fix for parallel BTree initialization bug