Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 8 Feb 2018 09:32:13 +0100
From:      Baptiste Daroussin <bapt@FreeBSD.org>
To:        Mark Linimon <linimon@lonesome.com>
Cc:        Alexey Dokuchaev <danfe@FreeBSD.org>, Ganael LAPLANCHE <martymac@FreeBSD.org>, ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org
Subject:   Re: svn commit: r460940 - in head/games: afternoonstalker burgerspace cosmosmash
Message-ID:  <20180208083213.yzrjhl4lwuqykvpg@ivaldir.net>
In-Reply-To: <20180208071931.GA13039@lonesome.com>
References:  <201802041634.w14GYHVW017651@repo.freebsd.org> <20180208071352.GA63012@FreeBSD.org> <20180208071931.GA13039@lonesome.com>

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

--w3yp6tfqobln3avm
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Feb 08, 2018 at 01:19:31AM -0600, Mark Linimon wrote:
> On Thu, Feb 08, 2018 at 07:13:52AM +0000, Alexey Dokuchaev wrote:
> > On Sun, Feb 04, 2018 at 04:34:17PM +0000, Ganael LAPLANCHE wrote:
> > > New Revision: 460940
> > > URL: https://svnweb.freebsd.org/changeset/ports/460940
> > >=20
> > > Log:
> > >   Fix build with Clang 6 (C++14 by default)
> >=20
> > ... by forcing gnu++98 without giving the details of the breakages and
> > why they were not properly fixed instead. :-)
>=20
> For once I'm going to agree with danfe (someone take a picture!) but only
> to the extent that I would like to ask that the commit messages be "work
> around the build" rather than "fix".  There have been numerous commits th=
at
> did contain actual code fixes.
>=20
> Despite the quibble, I appreciate the effort people are putting into the
> clang 6 problems -- these have especially regressed our tier-2 package
> builds.
>=20
> mcl
>=20
My point here, is variations between c++ standards may make some code incom=
patible
and sometime the version of the standard may/should be specified, because t=
he
code is clearly C++98.

In clang 6 when you build c++ instead of defaulting of c++98 like it used t=
o be
is now defaulting on c++14 meaning all the code whihch expected c++98 as de=
fault
standard are now broken and the -std flag should be supplied. The same happ=
ened
when the C compilers switched from C89 to C99 as default standard. On 143 p=
orts
we needed to make that change.

Being compatible with newer c++ standard is an upstream decision to take, n=
ot
hours. The compilers are able to specify the standard you are using, imho i=
t is
normal and expected for maintainers to use that feature :)

Bapt

--w3yp6tfqobln3avm
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=JW2W
-----END PGP SIGNATURE-----

--w3yp6tfqobln3avm--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180208083213.yzrjhl4lwuqykvpg>