From: Florian Friesdorf <flo@chaoflow.net>
To: Carl Worth <cworth@cworth.org>, notmuch@notmuchmail.org
Subject: patch submission guide (was Re: [PATCH] create and set temporary home
	directory)
In-Reply-To: <87pqnvk08h.fsf@yoom.home.cworth.org>
References: <877har5uek.fsf@eve.chaoflow.net>
	<87pqnvk08h.fsf@yoom.home.cworth.org>
User-Agent: Notmuch/0.6 (http://notmuchmail.org) Emacs/23.3.1
	(x86_64-unknown-linux-gnu)
Date: Sat, 07 May 2011 01:30:10 +0200
Message-ID: <87k4e376gt.fsf@eve.chaoflow.net>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-=";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Precedence: list

--=-=-=
Content-Transfer-Encoding: quoted-printable


Hi Carl,

On Fri, 06 May 2011 14:06:22 -0700, Carl Worth <cworth@cworth.org> wrote:
> On Mon, 18 Apr 2011 19:41:39 +0200, Florian Friesdorf <flo@chaoflow.net> =
wrote:
> > My first patch send to the list, not sure whether done properly.
>=20
> Just fine, Florian. Thanks for the contribution!
>=20
> One small thing you might do differently in the future is to tweak the
> email message to read exactly like a commit message. For example, a
> sentence like the above "not sure whether done properly" is fine in an
> email message, but doesn't make sense in the commit message.
>=20
> > I think the tests should not touch the build user's home directory. The
> > patch creates a directory in the temporary test directory and sets home
> > accordingly.
>=20
> Similarly, everything in a commit message is known to be your opinion,
> so you should omit phrases like "I think". Instead, you should describe
> what the commit actually does, and then describe why it does that.

That text was intended as a comment (cover letter according to what I
just learned) and the commit message itself was intended to be more to
the point.

> Finally, this little separator with three dashes:
>=20
> > ---
> >  test/test-lib.sh |    7 +++++++
> >  1 files changed, 7 insertions(+), 0 deletions(-)
>=20
> Is what lets git know where your commit message ends. Anything you
> include after that (And before the patch itself) will be ignored by
> git. So that's the perfect place to put a sentence like "My first
> patch---not sure whether done properly".

I now sent a patch with a cover letter using only git (search
=2D-format=3Dsanitized_text):

git format-patch --cover-letter <COMMIT>^
git send-email 000*

If everything is correct, I would conclude this into a short howto on
notmuchmail.org. So please, if there is still room for improvement, let
me know.

Do we prefer sendemail.chainreplyto or the new shallow format for patch
series [1]?

Should people who have a public git repository use it to publish their
patches (in addition to sending them here / instead of)?

I have seen (amdragon I think) the concept of
for-review/... branches. Would that be a best practice?

> I applied this patch last week, and would have pushed it, except that
> just after applying it, I also tried cleaning up some of this part of
> the code. And in the process it seems I managed to get the test suite to
> run "rm -rf ${HOME}" with my actual home directory (oops!).

I got a bit afraid, shouldnt my patch prevent that? I cannot imagine it
caused it.

[1] https://felipec.wordpress.com/2009/10/25/git-send-email-tricks/

regards
florian
=2D-=20
Florian Friesdorf <flo@chaoflow.net>
  GPG FPR: 7A13 5EEE 1421 9FC2 108D  BAAF 38F8 99A3 0C45 F083
Jabber/XMPP: flo@chaoflow.net
IRC: chaoflow on freenode,ircnet,blafasel,OFTC

--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (GNU/Linux)
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=Lb4y
-----END PGP SIGNATURE-----
--=-=-=--