From: Carl Worth <cworth@cworth.org>
To: Ruben Pollan <meskio@sindominio.net>, notmuch@notmuchmail.org
In-Reply-To: <87pr67ofz2.fsf@yoom.home.cworth.org>
References: <20091126202347.GA16654@blackspot>
	<1260364206-344-1-git-send-email-meskio@sindominio.net>
	<20091209132446.GC23878@blackspot>
	<877hsv6hbo.fsf@yoom.home.cworth.org>
	<87r5qnol78.fsf@yoom.home.cworth.org>
	<87pr67ofz2.fsf@yoom.home.cworth.org>
Date: Tue, 09 Mar 2010 09:36:10 -0800
Message-ID: <873a09jt2t.fsf@yoom.home.cworth.org>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-=";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Subject: Re: [notmuch] [PATCH] Added regress option to tags iterator
Precedence: list

--=-=-=
Content-Transfer-Encoding: quoted-printable

On Mon, 21 Dec 2009 19:16:49 -0800, Carl Worth <cworth@cworth.org> wrote:
> One option is to just drop the "move_ " prefix. Then everything will be
> a two-word function. So the new proposal is:
>=20
> to_first
> has_current
> to_next
>=20
> to_last
> has_current
> to_previous
>=20
> get_current
>=20
> Better?

Looking back at this proposal now, (after a long delay), I found I
didn't like it at all. With function names like:

	notmuch_threads_to_first
	notmuch_threads_to_next

the missing verb in the name is really distracting. I ended up reading
these names as if they were conversion functions. So I've gone back to
preferring the names with the explicit verbs (even though quite long):

	notmuch_threads_move_to_first
	notmuch_threads_move_to_next

Meanwhile, I also decided that _current wasn't adding anything to the
names it showed up in, (we can never "get" anything other than the
current item, so why qualify as "get_current"?). So I'm changing from
"has_current" to "valid", and leaving "get" as it is.

So the final proposal for iteration in either direction is:

	move_to_first
        valid
        move_to_next

	move_to_last
        valid
        move_to_previous

        get

I've just pushed commits changing the existing functions (which allow
only forward iteration) to use this naming scheme. I haven't added any
of the reverse-iteration functions yet, so Ruben, if you'd like to do
those within this scheme, that would be find. (Or we could wait until we
have an actual use in mind for them.)

Thanks,

=2DCarl

--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iD8DBQFLlocL6JDdNq8qSWgRAngCAKCOC1FhXomVg/nqQhct0ChFEXj/HgCfRAoj
wJS2YBSFaiEi3Teya+Du/UI=
=l7qE
-----END PGP SIGNATURE-----
--=-=-=--