On Wed, 21 May 2008 16:20:39 -0300
Alexandre Oliva <aoliva(a)redhat.com> wrote:
On May 21, 2008, Brian Pepple <bpepple(a)fedoraproject.org>
wrote:
> You want something to be discussed? Send a note to the list in reply to
> this mail and I'll add it to the schedule.
Given that FreedomĀ² is a major fedora feature, I'd like to discuss
enabling the creation of Fedora spins containing exclusively Free
Software. These are related sub-topics:
. Permission to distribute under the mark 'Fedora' spins containing
kernel-libre packages, whose sole difference from identically-numbered
Fedora kernel builds is the removal of a few pieces of non-Free
Software.
All spins must be composed of packages that are contained within the
Fedora repositories. kernel-libre does not fit that category (today).
. Inclusion in Fedora (future and recent past releases) of the
kernel-libre package, a 100% Free Software variant of the kernel
Linux, that I've been maintaining tracking Fedora kernel builds at
http://www.fsfla.org/~lxoliva/fsfla/linux-libre/
We've had this discussion. We aren't going to allow a forked kernel
package. Please work with the kernel team to integrate this into the
main kernel package.
. Inclusion in Fedora (future and recent past releases) of a
fedora-freedom "virtual" package, that Requires: linux-libre and
Conflicts: with any Fedora package known to contain software (firmware
included) that does not respect the 4 freedoms established in the Free
Software definition. AFAIK these would pretty much amount to the
standard non-Free kernel and a bunch of *-firmware packages, but there
could be sub-packages to cover other debatable packages with obscure
source code, dubious licensing policies, etc.
You don't need a package. Make a comps group.
I realize these packages should probably be submitted for inclusion
through the regular package submission process, but I was advised to
discuss linux-libre in FESCo first, and the second is closely related
and has no upstream.
I'm a bit hesitant, for these appear to be more of policy than
engineering issues, and my understanding is that the board is in
charge of such decisions. Anyhow, it (hopefully :-) wouldn't hurt for
the board to get recommendations from engineering in this regard,
assuming my understanding as to how policy decisions are made is
correct.
Please let me know whether this is a suitable topic for discussion in
tomorrow's meeting, and I'll do my best to be there, i.e., save for
unforeseeable issues or ISP LoQoS I've been subject to recently :-/
the L in LoQoS is for Lack, in case it's not obvious :-)
I think we can certainly discuss it. However I believe the biggest
hurdle to what you propose is the extra kernel-libre package. Your
overall proposal hinges on that, and the way you've stated you would
like to provide it has been frowned upon quite a bit.
josh