--FBZZyUSgjpMOo1m2
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Jul 18, 2007 at 11:08:28PM +0200, degroote@netbsd.org wrote:
> On Wed, Jul 18, 2007 at 10:23:14PM +0200, Quentin Garnier wrote:
> >=20
> > On Wed, Jul 18, 2007 at 09:33:18PM +0200, Joerg Sonnenberger wrote:
> > > On Wed, Jul 18, 2007 at 09:31:03PM +0200, Quentin Garnier wrote:
> > > > The way we handle firmwares in pkgsrc is just wrong.  We should have
> > > > only one package, that installs all the versions the kernel could
> > > > possibly ask for (which means the kernel asks for a specific versio=
n).
> > >=20
> > > Keep in mind that the firmware images are not only used by NetBSD.
> > > Please carefully consider this before changing any policy...
> >=20
> > There's a policy?  I'd actually like to set one...
>=20
> I understand what you mean and I agree. The current way we deal with
> firmware version is a pain, particulary when constructors decided to chan=
ge
> the firmware layout every two minors revision of firmware ...=20
>=20
> On other side, I don't think it is a good idea that the kernel ask for a
> specific version. In most of the case, new version of the firmware solves
> internal issues and doesn't need change in the driver. It may not be a
> problem in current but in the stable branche, we can't ask for people to
> patch by hand their kernel in order to use the new firmware.

Why not use a sysctl node then?  Or a symlink, or whatever, there are
many ways to solve that.

> On pkgsrc side, what I understand from your proposal is that you want to
> rename the firmware so we can retrieve the different versions of firmware.
> What joerg says I think is that other Os may not find the firmware file
> because it doesn't have the canonical name found in the default install.

Sure, but what matters is that we agree on something and stick to it.
When you update the package like you did, in a way that requires changes
to the kernel, you're not helping other OSes that might not have the
necessary kernel changes yet either...

--=20
Quentin Garnier - cube@cubidou.net - cube@NetBSD.org
"You could have made it, spitting out benchmarks
Owe it to yourself not to fail"
Amplifico, Spitting Out Benchmarks, Hometakes Vol. 2, 2005.

--FBZZyUSgjpMOo1m2
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (NetBSD)

iQEVAwUBRp6H89goQloHrPnoAQIJigf/YBFtVDzPDu0gy6BDzW3XvQn+7LblsGjm
JXLVawRCmk0XAnYwiSKl31eSETKeFBIy2x1kvHpQYcwt8ZLElfcAv3Afq2MrI1i8
kQDeZ7bpeSyd63pkw/d234iW7hEAPn0ZDSn42Pz2EHlgUiq/nKttRaAZ8O7eVZCs
lzKgzHg2LJ5U79wxj8kEBCSm8Ke0x5jwF/DjrG/ZGqPxUcbqJ615M1WWHz8JroRu
5JfX9Shq5kPY2V/Qy2AkgU6vdIS+58843esoePmvZy2ivkwwhvkmZI6MsA0T61Mh
ukbv/wy0j5bBCPB8wqdl6c3cJfdS/sBkjA8Jigs0stbUHnxTHjjivQ==
=nY8d
-----END PGP SIGNATURE-----

--FBZZyUSgjpMOo1m2--