Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 2 Mar 2011 18:12:40 +0100
From:      "Pietro Cerutti" <gahr@gahr.ch>
To:        Michal Varga <varga.michal@gmail.com>, freebsd-ports@freebsd.org
Subject:   Re: gegl
Message-ID:  <20110302171239.GG1550@gahrfit.gahr.ch>
In-Reply-To: <1299080893.52738.28.camel@xenon>
References:  <201103011706.45759.lumiwa@gmail.com> <20110302075219.GA2182@reindeer.exwg.net> <20110302100645.GB2182@reindeer.exwg.net> <4D6E5691.9060906@yandex.ru> <1299077867.52738.5.camel@xenon> <20110302152634.GF1550@gahrfit.gahr.ch> <1299080893.52738.28.camel@xenon>

next in thread | previous in thread | raw e-mail | index | archive | help

--jkO+KyKz7TfD21mV
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On 2011-Mar-02, 16:48, Michal Varga wrote:
> On Wed, 2011-03-02 at 16:26 +0100, Pietro Cerutti wrote:
> > On 2011-Mar-02, 15:57, Michal Varga wrote:
> > > Out of curiosity - how is it even possible that this update made it as
> > > far as into ports tree, when it - at least as I understand from the P=
R -
> > > clearly doesn't work on *both* *STABLE* *supported* branches and only
> > > goes well with the very latest CURRENT?
> >=20
> > The original update doesn't cause any problems on CURRENT,
> > 7.4-RELEASE, or 8.2-RELEASE. It does cause problems on STABLE branches.
> >=20
>=20
> Is this guaranteed?
>=20
> As here with...
>=20
> > uname -a
> FreeBSD 7.4-STABLE FreeBSD 7.4-STABLE #0: Sun Feb 27 22:30:14 CET 2011 =
=20
>=20
> ...gegl (unpatched) still fails:
>=20
>=20
> exp_combine.cpp: In function 'gfloat expcombine_get_file_ev(const gchar*)=
':
> exp_combine.cpp:99: error: 'log2f' was not declared in this scope
> gmake[2]: *** [exp_combine-exp_combine.o] Error 1
>=20
>=20
> Note that 7.4-RELEASE was back in Feb 21. Or was there a change specific
> to 7.4-RELEASE that for some reason didn't make it into STABLE?

OK, I have nailed down the source of the problem. GEGL has an optional
dependency on graphics/exiv2 which is not tracked by the port, i.e.
exiv2 is not built as a dependency of the port, and GEGL doesn't rely on
it if it doesn't find it. It turns out that the failing C module
(exp_combine.c) is only compiled when EXIV is found.

Thus, on a fresh system (i.e., tinderbox) the error won't appear. It
will, however, on a system where EXIV is installed.
Please check whether you have graphics/exiv2 installed in your system to
confirm this hypothesis.

I have still not decided whether to include exiv2 as a mandatory or
optional dependency, cause configure doesn't handle --without-exiv2
correctly when exiv2 is installed (i.e., GEGL will be built with EXIV
support in that case).

Thanks for the hints, and sorry for the noise.

--=20
Pietro Cerutti
The FreeBSD Project
gahr@FreeBSD.org

PGP Public Key:
http://gahr.ch/pgp

--jkO+KyKz7TfD21mV
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (FreeBSD)

iEYEARECAAYFAk1ueocACgkQwMJqmJVx9441DQCfYchp4W/R+TT6KPx8wvHJQ6vf
koMAniQFM3jwdQHtE6jzK0mOg3A9v4sZ
=O/s2
-----END PGP SIGNATURE-----

--jkO+KyKz7TfD21mV--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110302171239.GG1550>