Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 8 Oct 2015 17:24:29 +0100
From:      David Chisnall <theraven@FreeBSD.org>
To:        Andriy Gapon <avg@FreeBSD.org>
Cc:        Craig Rodrigues <rodrigc@FreeBSD.org>, src-committers@FreeBSD.org, svn-src-all@FreeBSD.org, svn-src-head@FreeBSD.org
Subject:   Re: svn commit: r289027 - head/contrib/tzcode/stdtime
Message-ID:  <51F19592-19C7-4444-81BC-9F7DCE74F919@FreeBSD.org>
In-Reply-To: <561666CE.1050608@FreeBSD.org>
References:  <201510081142.t98BgFJd028821@repo.freebsd.org> <561666CE.1050608@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 8 Oct 2015, at 13:51, Andriy Gapon <avg@FreeBSD.org> wrote:
>=20
> What if one day github disappears but FreeBSD is still going?
> The full commit message would be lost.

That=E2=80=99s not the only thing that is bad about this commit message. =
 Why =E2=80=98Assume C89?=E2=80=99  We compile libc as C99 + GNU =
extensions and are likely to default to C11 + GNU extensions soon.  =
Reading the actual commit, it looks as if it=E2=80=99s changing K&R =
declarations to ISO C declarations. =20

It=E2=80=99s also introducing an ATTRIBUTE_PURE macro in private.h, =
which does exactly the same as __pure declared in sys/cdefs.h (which is =
included by *every single FreeBSD header*.  Why this extra spelling?  No =
idea.  Is this contrib code (it looks like it)?  If so, it should come =
in via the vendor area (not be directly committed to head), if not then =
it should have been code reviewed and not include redundant and =
confusing macro declarations.

When merging stuff from GitHub, Alfred has written some good =
documentation about how to handle them in such a way that we preserve =
the prior commit history (effectively, checkout the pr branch, rebase it =
on head, git svn dcommit the result).  Please follow this procedure.

David




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51F19592-19C7-4444-81BC-9F7DCE74F919>