That sounds fine, but also like a detour from the actual goal. One way or
another, the in-tree packaging that is there for Ubuntu won't be at stake
when making it work on Debian. Also note that the in-tree debian/ directory
doesn't work even for Ubuntu, strictly speaking. We already have the 14.04
bits in a separate branch.

I'd focus on making it properly test more distributions first (not just
Debian either) and later worry about whether we should get rid of debian/
or not.

On Mon, Jan 16, 2017 at 12:02 PM, Zygmunt Krynicki <
zygmunt.kryni...@canonical.com> wrote:

> I'd like to work on enabling Debian in the CI loop and I was thinking
> that it would be somewhat easier we switched to non-native packaging
> in the upstream tree and similarly switched to quilt in the Debian
> tree (we could have separate packaging trees for sid / stretch if that
> would help). Since my view may be simplistic I would like to ask the
> current most active Debian maintainers of snapd for opinion.
>
> Right now almost all of the CI in the tree is performed on the
> packaging that is in the tree as well. The notable exception is 14.04
> which has a separate packaging branch. This is unrealistic as the
> Debian packaging tree is widely different and even if we built a
> package from the in-tree debian directory and tested it on a real
> Debian machine the result would not be representative of what a
> subsequent upstream release would look like in Debian.
>
> I'd like to propose that we remove the debian directory from the
> upstream repository (no special casing) and work on ensuring that
> Ubuntu and subsequently Debian are tested equally well whenever we
> make a pull request.
>
> Best regards
> ZK
>
> --
> Snapcraft mailing list
> Snapcraft@lists.snapcraft.io
> Modify settings or unsubscribe at: https://lists.ubuntu.com/
> mailman/listinfo/snapcraft
>



-- 
gustavo @ http://niemeyer.net
-- 
Snapcraft mailing list
Snapcraft@lists.snapcraft.io
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/snapcraft

Reply via email to