Mailing-List: contact dev-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@couchdb.apache.org
Received-SPF: pass (nike.apache.org: domain of adam.kocoloski@gmail.com
 designates 209.85.212.52 as permitted sender)
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Apple Message framework v1082)
Subject: Re: Tree::term() is really a tree()
From: Adam Kocoloski <kocolosk@apache.org>
In-Reply-To: <AANLkTi=PfxgHprGzAVyPxqtu0N1FkUW5XPW1RVx7yduL@mail.gmail.com>
Date: Wed, 9 Mar 2011 10:03:28 -0500
Content-Transfer-Encoding: quoted-printable
Message-Id: <A7C3DB14-A4F1-45CB-9788-8551E4FA66C0@apache.org>
References: <AANLkTi=PfxgHprGzAVyPxqtu0N1FkUW5XPW1RVx7yduL@mail.gmail.com>
To: dev@couchdb.apache.org

On Mar 9, 2011, at 9:43 AM, Andrey Somov wrote:

> Hi all,
> I have found the following comment in the source (couch_db.hrl):
>=20
> % Tree::term() is really a tree(), but we don't want to require R13B04 =
yet
> -type branch() :: {Key::term(), Value::term(), Tree::term()}.
>=20
> Does it mean that CouchDB includes its own copy of gb_trees.erl ?
> Does it mean that CouchDB implements its own tree but the interface is
> identical to gb_trees.erl and it will be changed once R13B04 is =
required ?
>=20
>=20
> -
> Andrey

Hi Andrey, neither.  It simply means that the third element of a =
branch() is of type tree().  The reason for comment is that recursive =
type specifications are not supported in Erlang/OTP releases prior to =
R13B04, so I fell back to calling it a term() instead.

The tree() data structure being described is a custom data structure =
implemented entirely in couch_key_tree.erl.  It's used to store document =
revision histories.  Best,

Adam=