From: Carl Worth <cworth@cworth.org>
To: Robin Green <greenrd@greenrd.org>, notmuch@notmuchmail.org
Subject: Re: Drafts being tagged as inbox
In-Reply-To: <8639iu5tju.fsf@greenrd.plus.com>
References: <868vsnaqqm.fsf@greenrd.plus.com> <877h86tqez.fsf@praet.org>
	<8639iu5tju.fsf@greenrd.plus.com>
User-Agent: Notmuch/0.5 (http://notmuchmail.org) Emacs/23.3.1
	(i486-pc-linux-gnu)
Date: Tue, 28 Jun 2011 13:47:59 -0700
Message-ID: <87mxh18yyo.fsf@yoom.home.cworth.org>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-=";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Precedence: list

--=-=-=
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Tue, 28 Jun 2011 08:00:53 +0100, Robin Green <greenrd@greenrd.org> wrote:
> On Tue, 28 Jun 2011 08:34:44 +0200, Pieter Praet <pieter@praet.org> wrote:
> > notmuch tag -inbox +draft -- folder:${where_your_drafts_reside}
>=20
> My notmuch help does not mention "folder:", and indeed folder:drafts as a
> search term is just treated as the words folder and drafts. Perhaps this
> is implemented in git head.

Yes, sorry about that. We're woefully late on a release that will
include this feature. Our current release manager expects that release
within a couple of days at least, so that's good!

> However, when I try to compile git head on Dragonfly BSD, I get this erro=
r:
>
> CXX -O2 notmuch-shared
> notmuch-count.o: In function `notmuch_count_command':
> notmuch-count.c:(.text+0x63): undefined reference to `notmuch_database_op=
en'
> notmuch-count.c:(.text+0x9d): undefined reference to `notmuch_query_creat=
e'

Ouch. It looks like your linker isn't finding anything in the notmuch
library at all. I wonder what went wrong there. Some things to try would
be:

	make V=3D1

to see the complete linker command line to see if something is obviously
wrong there.

And otherwise to try a "git bisect" to see when the compilation stopped
working.

To start a bisect, you'll need to know a point where notmuch actually
compiles successfully on your system=E2=80=940.5 perhaps? Double-check that
with:

	git checkout 0.5	# Get the code from the 0.5 release
        make			# See if it compiles
        git checkout master	# Come back to master

If that went fine then you can bisect with:

	git bisect start
        git bisect good 0.5	# Indicate that 0.5 is good
        git bisect bad		# Indicate that the current state is bad

After that you'll iterate by running:

	make

Followed by one of the following:

	git bisect good		# If the compile succeeded
	git bisect bad		# If the compile failed

In the end, git should tell you a "first bad commit". Send that output
to the list and we can start pinning down what went wrong. Then you can
finalize your git-bisect run with:

	git bisect reset

Good luck!

=2DCarl

--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEARECAAYFAk4KPf8ACgkQ6JDdNq8qSWggggCeP8SznQQwytE+Bgga6yemtOlZ
yJAAnR28o+JZb8/1AT4vdPlfjJbt2OWh
=GsXX
-----END PGP SIGNATURE-----
--=-=-=--