Re: [gentoo-project] My manifesto for 2015/2016

[William Hubbs]

On Sat, Jul 11, 2015 at 01:43:06AM -0700, Daniel Campbell (zlg) wrote:
> On 07/10/2015 09:40 AM, William Hubbs wrote:
> >
> > I'm not exactly sure what you mean by "Gentoo's stack" -- can you
> > elaborate a little more on this?
>
> Sure. By "the Gentoo stack", I'm talking primarily about the software
> stack -- kernel, init, early userland, and to a lesser extent, @system
> - -- that Gentoo systems generally all have in common, that won't change
> without good reason. In effect, I'm asking if you favor the current
> situation where people have a pretty minimal, workable base that will
> bend to their needs (be they systemd, openrc, some other init; udev,
> eudev, mdev, vdev, etc), or if you'd rather see a base that is treated
> as mostly immutable like many other distros are going in the direction
> of (or have already settled on).
 
 I'm fine with a minimal @system set and allowing users to build their
 systems on top of that. In fact, the movement has been toward removing
 more things from the @system set at some point.

> I think a big part of why we use Gentoo is its flexibility and
> insistence on *not* forcing decisions on users or developers, so I
> wanted a little clarification on that in contrast to the "keeping up
> with change" part of your manifesto, which I think I agree with in
> spirit but would rather not see us become a distro that just follows,
> but instead innovates in our own way while still shipping mostly
> vanilla packages.

I'm not saying anything against innovation or flexability; in my
experience, Gentoo is the most flexable Linux distro out there, and I
want to keep it that way.

When I was brought on board in Gentoo, the idea about patches was
that they are waiting to be pushed upstream. I think unless there is a
very good technical reason for the patch, we should avoid custom patches
we do not intend to push upstream.

If a patch is rejected by upstream, we should dialog with them
to find out why it was rejected and clean it up on our side. If we are
patching software it should be so that everyone can benefit, not just
Gentoo.

Being innovative doesn't mean we keep supporting legasy software
when there is new software that is clearly compatible with it and legasy
upstream has told everyone to switch to the new software. This was the
exact debate that was going on for a while when kmod hit the tree. Some
were saying that since Gentoo is about choice we should keep m-i-t in
the tree and let people use it even though m-i-t upstream made it very
clear that it was dead and we should move to kmod going forward.

I realize that a lot of this is on a case-by-case basis, but sometimes I
feel that vocal minorities use the "gentoo is about choice" mantra to
try to force custom patches on our maintainers or force them to maintain
old dead software or keep legasy practices working just for the sake of it.

William