From: Antonin Houska <ah@cybertec.at>
To: Michael Paquier <michael@paquier.xyz>
cc: Bruce Momjian <bruce@momjian.us>,
    Alvaro Herrera <alvherre@2ndquadrant.com>,
    pgsql-hackers@postgresql.org
Subject: Re: Incorrect visibility test function assigned to snapshot
In-reply-to: <20190214071313.GF2366@paquier.xyz>
References: <23215.1527665193@localhost>
 <20180530132854.tkv37qiamepu3slh@alvherre.pgsql>
 <20180623151856.GG21575@momjian.us> <23440.1549623545@localhost>
 <20190214071313.GF2366@paquier.xyz>
Comments: In-reply-to Michael Paquier <michael@paquier.xyz>
   message dated "Thu, 14 Feb 2019 16:13:13 +0900."
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Date: Fri, 15 Feb 2019 13:01:29 +0100
Message-ID: <16034.1550232089@localhost>
Precedence: bulk

Michael Paquier <michael@paquier.xyz> wrote:

> On Fri, Feb 08, 2019 at 11:59:05AM +0100, Antonin Houska wrote:
> > Sorry, I forgot. Patch is below and I'm going to add an entry to the
> > next CF.
>=20
> > @@ -615,6 +615,8 @@ SnapBuildInitialSnapshot(SnapBuild *builder)
> >=20=20
> >  		TransactionIdAdvance(xid);
> >  	}
> > +	/* And of course, adjust snapshot type accordingly. */
> > +	snap->snapshot_type =3D SNAPSHOT_MVCC;
>=20
> Wouldn't it be cleaner to have an additional argument to
> SnapBuildBuildSnapshot() for the snapshot type?  It looks confusing to
> me to overwrite the snapshot type after initializing it once.

I'm not sure. SnapBuildBuildSnapshot() only creates snapshot for the logical
replication purposes and the "snapshot_type" argument would make the functi=
on
look like it can handle all snapshot types. If adding an argument, I'd pref=
er
a boolean variable (e.g. "historic") telling whether user wants
SNAPSHOT_HISTORIC_MVCC or SNAPSHOT_MVCC. But that may deserve a separate
patch.

As for the bug fix, I think the additional assignment does not make things
worse because SnapBuildInitialSnapshot() already does overwrite some fields:
"xip" and "xnt".

> > @@ -1502,6 +1502,13 @@ ImportSnapshot(const char *idstr)
> >  	 */
> >  	memset(&snapshot, 0, sizeof(snapshot));
> >=20=20
> > +	/*
> > +	 * Do not rely on the fact that SNAPSHOT_MVCC is zero. (The core code
> > +	 * currently does not use this field of imported snapshot, but let's =
keep
> > +	 * things consistent.)
> > +	 */
> > +	snapshot.snapshot_type =3D SNAPSHOT_MVCC;
>=20
> Okay for this one, however the comment does not add much value.

o.k. I don't insist on using this comment.

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