Mailing-List: contact dev-help@subversion.apache.org; run by ezmlm
Precedence: bulk
Received-SPF: pass (nike.apache.org: local policy)
Date: Fri, 18 Jul 2014 11:32:19 +0200
From: Matthias Gerstner <Matthias.Gerstner@ncp-e.com>
CC: <dev@subversion.apache.org>
Subject: Re: [PATCH] optionally disable normalization of working copy files
 in diff invocations
Message-ID: <20140718093219.GB7615@mgpc.ncp.de>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="pvezYHf7grwyp3Bc"
Content-Disposition: inline
User-Agent: Mutt/1.5.22 (2013-10-16)

--pvezYHf7grwyp3Bc
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

> I'd hope we could address this without public API changes and
> without adding yet another command line option.

Yes the API change is pretty tough.

> How about we make this the default if a third party diff tool is used?
> This way, third party diff tools will always display differences in
> keywords, and possibly EOLs.

Would be fine for the common case I guess. But it still might break some
more exotic uses. At least reversing the default might be a good idea,
so the normalization only takes place if explicitly requested when using
external tools.

Another more complex approach would be to keep passing temporary,
normalized files to the differ but then afterwards check whether it has
been modified an incorporate the changes back into the working copy. But
that's beyond what I can do here...

> In terms of coding style, I'd use a boolean that switches normal
> form on, rather than off. I find it easier to keep track of this way.

That's true.

> The interaction between use_text_base and your new no_normalization
> flag isn't made clear. You can't have both!
>=20
> Also, your change only addresses BASE->WORKING diffs, as far as I can tel=
l.
> What about REPOS->WORKING or WORKING->REPOS diffs?

I've actually just been annoyed with the inability to edit my diffs so I
poked around in the code and that's the result. It fixed my particular
use case. I had quite a hard time getting so far, not knowing the
internal workings of subversion very well.

> The diff below (compile-tested but otherwise untested) shows how
> I would try to address this without any public API changes.
> Does this do what you want?

I've tested it but didn't get the desired result. Using --diff-cmd I
still get passed temporary files.

--=20
Matthias Gerstner, Dipl.-Wirtsch.-Inf. (FH)
Entwicklung
=20
NCP engineering GmbH
Domb=FChler Stra=DFe 2, D-90449, N=FCrnberg
Gesch=E4ftsf=FChrer Peter S=F6ll, HRB-Nr: 77 86 N=FCrnberg
=20
Telefon: +49 911 9968-153, Fax: +49 911 9968-229
E-Mail: Matthias.Gerstner@ncp-e.com
Internet: http://www.ncp-e.com

--pvezYHf7grwyp3Bc
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=FLfn
-----END PGP SIGNATURE-----

--pvezYHf7grwyp3Bc--