Mailing-List: contact dev-help@subversion.apache.org; run by ezmlm
Precedence: bulk
Received-SPF: pass (athena.apache.org: local policy)
Date: Wed, 8 Feb 2012 13:55:33 +0100
From: Matthijs Kooijman <matthijs@stdin.nl>
To: dev@subversion.apache.org
Subject: Re: [PATCH] More bindings for gnome_keyring /
	platform_specific_client_providers
Message-ID: <20120208125533.GO16542@login.drsnuggles.stderr.nl>
Mail-Followup-To: Matthijs Kooijman <matthijs@stdin.nl>,
	dev@subversion.apache.org
References: <20120105200619.GZ17548@login.drsnuggles.stderr.nl>
 <20120208122848.GL16542@login.drsnuggles.stderr.nl>
 <20120208124553.GD19105@ted.stsp.name>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="Gi2vSq8QtPYbuTNa"
Content-Disposition: inline
In-Reply-To: <20120208124553.GD19105@ted.stsp.name>
User-Agent: Mutt/1.5.18 (2008-05-17)


--Gi2vSq8QtPYbuTNa
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Hey Stefan,

> For example, the addition of new functions prevents people from rolling
> back to an earlier 1.7 patch release without breaking scripts which were
> modified to use the new functions after updating to 1.7.3. That would
> be a violation of our release compatibility guidelines.
Hmm, that's a pity, since the addition of this new unlock_prompt_func
function would really be the one commit I'd like to see backported (I
only did the other patches for consistency and because I now better
understand swig than a year ago when I did the perl patch).

Also, if I understand your policy correctly, backporting the below
patches isn't exactly useful either, since any script that will be
modified to use the svn_auth_get_platform_specific_client_providers will
break as well when downgrading to 1.7.2 (since there this function
(binding) exists, but will always error out). I guess the distinction
between "add a new function" and "fix an unusable function" is a bit
vague here.

Or does the policy concern the C API? Since all of the patches
(including the unlock_prompt_func one) only touch the bindings, not the
actual C api.

Gr.

Matthijs

> Changes which don't add new API symbols can be backported.
> Currently, the following changes are nominated for backport to 1.7:
>=20
>  * r1241530, r1241713, r1241726
>    Fix the python bindings for svn_auth_get_platform_specific_client_prov=
iders.
>    Justification:
>      The bindings should see passwords cached in the platform-specific pr=
oviders.
>      Already fixed for the Perl bindings.  Ruby fix is nominated separate=
ly.
>    Notes:
>      r1241713 and r1241726 tweak the unit test.
>    Votes:
>      +1: danielsh
>=20
>  * r1241553
>    Fix the ruby bindings for svn_auth_get_platform_specific_client_provid=
ers.
>    Justification:
>      The bindings should see passwords cached in the platform-specific pr=
oviders.
>      Already fixed for the Perl bindings.  Python fix is nominated separa=
tely.
>    Votes:
>      +1: stsp
>=20
>=20

--Gi2vSq8QtPYbuTNa
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAk8ycMUACgkQz0nQ5oovr7wxyQCg3PUe7LPVgK/RPQs/anB1x1dl
KWYAn2zinkWio75wVk+Iv43i8S0kUH8E
=6ce8
-----END PGP SIGNATURE-----

--Gi2vSq8QtPYbuTNa--