To: David Bremner <david@tethera.net>, notmuch@notmuchmail.org
Subject: Re: [PATCH] emacs: Helpers for notmuch developers.
In-Reply-To: <878vllxzwt.fsf@zancas.localnet>
References: <1325685678-12710-1-git-send-email-dme@dme.org>
	<878vllxzwt.fsf@zancas.localnet>
User-Agent: Notmuch/0.10.2+151~gbf1dc2b (http://notmuchmail.org)
	Emacs/24.0.92.1 (x86_64-pc-linux-gnu)
From: David Edmondson <dme@dme.org>
Date: Fri, 06 Jan 2012 08:27:15 +0000
Message-ID: <cunhb0944nw.fsf@hotblack-desiato.hh.sledj.net>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-=";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Precedence: list

--=-=-=
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Thu, 05 Jan 2012 23:39:30 -0400, David Bremner <david@tethera.net> wrote:
> > General management (i.e. keeping up to date) of the repository it uses
> > is your responsibility, as is cleaning out old branches. You can, of
> > course, just delete the temporary repository after using it - the code
> > will re-create it next time.
>=20
> maybe the branches could be under some namespace that makes this easy?

Just a prefix? 'review/' or something?

> > +(defvar notmuch-dev-temporary-repository-name (concat "notmuch-dev-" (=
user-login-name))
> > +  "The name of the temporary repository.")
>=20
> Do we care about security at all in this context? I'm always a bit
> nervous about creating predictably named files/directories in publicly
> writable places.

I presume that you'd already have set `temporary-file-directory' if you
worry about this generally, or `notmuch-dev-temporary-directory' if just
in this instance.

> > +  ;; Causes us to switch to the magit buffer - is that unfortunate in
> > +  ;; some situations?
> > +  (magit-status notmuch-dev-temporary-repository-path))
>=20
> I found it to be a pleasant surprise.

Ending there when the whole sequence completes, I agree. The question is
really about whether displaying the magit-status buffer after simply
creating the repository is good.

> > +      (shell-command
> > +       (concat
> > +	notmuch-command " show --format=3Dmbox " (shell-quote-argument search=
-terms)
> > +	" | "
> > +	"git am --quiet"))
> > +
>=20
> Hmm. A knee jerk reaction is not to like this, like seeing system in C
> code. But I don't have a better solution off hand.

Agreed. I think that I'll split it into two parts:
  - running notmuch to create a mailbox (or perhaps directory of
    patches),
  - running git-am to apply the patches.

A bigger problem is that magit appears to be ignorant of git-am, with
the result that if applying the patches fails magit doesn't provide any
help (or even notification) that there are outstanding things to do. It
also blocks a future git-am.

Currently thinking about an alternative to using git-am that is more
magit friendly...

--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEARECAAYFAk8GsGMACgkQaezQq/BJZRaAqQCeLWxt7y5MAPU5Mnpkphg0co/G
1pcAn10Mk7G7H+DqL1RzuHmI4/teWxJ1
=Yns4
-----END PGP SIGNATURE-----
--=-=-=--