From: Antonin Houska <ah@cybertec.at>
To: Oleg Ivanov <o.ivanov@postgrespro.ru>
cc: pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] Proposal: generic WAL compression
In-reply-to: <23990.1510853483@localhost>
References: <59F90ABF.2060407@postgrespro.ru> <23990.1510853483@localhost>
Comments: In-reply-to Antonin Houska <ah@cybertec.at>
   message dated "Thu, 16 Nov 2017 18:31:23 +0100."
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Date: Mon, 20 Nov 2017 08:49:48 +0100
Message-ID: <8015.1511164188@localhost>
Precedence: bulk

Antonin Houska <ah@cybertec.at> wrote:

> Oleg Ivanov <o.ivanov@postgrespro.ru> wrote:
>=20
> > In order to overcome that issue, I would like to propose the patch, whi=
ch
> > provides possibility to use another approach of the WAL record
> > construction.
>=20
> After having spent several hours reviewing this patch I dare to send the
> following comments:

One more idea:

I think the metadata (ALIGN_GAP) should be stored separate from the actual
data so that you can use memcpy() instead of this loop:

    while (i < j)
    {
	    char		c =3D targetRegionAligned[i++];

	    writeToPtr(ptr, c);
    }


--=20
Antonin Houska
Cybertec Sch=C3=B6nig & Sch=C3=B6nig GmbH
Gr=C3=B6hrm=C3=BChlgasse 26
A-2700 Wiener Neustadt
Web: http://www.postgresql-support.de, http://www.cybertec.at