Date: Sun, 6 Oct 2013 21:03:36 -0700
From: Dave Johansen <davejohansen@gmail.com>
To: Development discussions related to Fedora
        <devel@lists.fedoraproject.org>
Subject: Re: boost141 and stability of Boost API?
Message-ID:
        <CAAcYxUd9ng9_Q2m=WpA=WYnE8tgs4M-ZR8CC1hs45Bg9PmN51w@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

On Sat, Oct 5, 2013 at 2:12 PM, Denis Arnaud <denis.arnaud_fedora@m4x.org>wrote:

> Hi Dave,
>
> note that Boost-1.48 has been packaged for EPEL 5 and 6, but not yet
> approved: http://bugzilla.redhat.com/show_bug.cgi?id=921134
> In case it is useful to anyone, do not hesitate to approve it :)
> And if there is more love, we could even embark on the way to package
> Boost-1.54 for EPEL... But that is another story.
>

That's good to hear because it's always good to have options, but newer
versions just bring up the question of a "moving target". Just like how you
mentioned 1.54, if we're shooting for newer versions, then why not go with
that one, or 1.55 after it? The point of RHEL is a stable platform and
development environment and chasing the newest version of a library
(especially one as volatile as Boost) just doesn't seem to fit the RHEL
mentality.

Note that those packages (boost141, boost148) are intended only for EPEL, IN ADDITION to the version officially supported by RedHat (Boost-1.31 on RHEL 5 and Boost-1.41 on RHEL 6). On Fedora, we strive to deliver the latest possible version of Boost every six months, not in between.

Hence, EPEL packagers have the choice of the version of Boost they build upon, and are not forced to perform complex patch retrofits for their own packages (having dependencies on Boost). That does not impact AT ALL the EPEL packages built on top of the officially supported version of Boost.

Kind regards

Denis