Re: pg_receivexlog and replication slots

On 2014-10-01 11:21:12 -0300, Alvaro Herrera wrote:
> Andres Freund wrote:
> 
> > From d667f7a63cd62733d88ec5b7228dfd5d7136b9d7 Mon Sep 17 00:00:00 2001
> > From: Michael Paquier <michael@otacoo.com>
> > Date: Mon, 1 Sep 2014 20:48:43 +0900
> > Subject: [PATCH 3/4] Refactoring of pg_basebackup utilities
> > 
> > Code duplication is reduced with the introduction of new APIs for each
> > individual replication command:
> > - IDENTIFY_SYSTEM
> > - CREATE_REPLICATION_SLOT
> > - DROP_REPLICATION_SLOT
> > A couple of variables used to identify a timeline ID are changed as well
> > to be more consistent with core code.
> > ---
> >  src/bin/pg_basebackup/pg_basebackup.c  |  21 +----
> >  src/bin/pg_basebackup/pg_receivexlog.c |  38 ++------
> >  src/bin/pg_basebackup/pg_recvlogical.c | 116 ++++++------------------
> >  src/bin/pg_basebackup/streamutil.c     | 159 +++++++++++++++++++++++++++++++++
> >  src/bin/pg_basebackup/streamutil.h     |  10 +++
> >  5 files changed, 207 insertions(+), 137 deletions(-)
> 
> Not objecting to any of this, but isn't it a bit funny that a patch that
> aims to reduce duplication ends up with more lines than there were
> originally?

The reduction is there in combination with a later patch - which needs
most of the code moved to streamutil.c.

Greetings,

Andres Freund

-- Andres Freund                       http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services