Open Stack

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 07/16/2015 03:06 PM, Thomas Goirand wrote:
> What we discussed was that distributions would run their own CI
> and validate patches away from upstream CI, then we would agree on
> a patch and share it.
> 

That's fine. I think a good start would be to get them voting on
existing *supported* branches. It will only give stable maintainers
another way to assure backports make sense.

But I am not convinced that those CI systems are in place everywhere:
I don't know anything that works on patch granularity in RDO (bulk rpm
updates are indeed validated).

It's my assumption that setting all those per-distro CIs working is a
tough nut to crack. It would be easier for everyone to provide
resources to keep the gate passing for existing stable branches. If
the gate is passing for a stable branch, and the project has enough
human resources to expect that it will still work in the next month, I
don't see particular reason to kill the branch if there is indeed
demand to keep it.

Working on upstream gate stability obviously does not invalidate any
effort to introduce distribution CI votes in gerrit, and I would be
happy to see RDO or Ubuntu meaningfully voting on backports. It's my
belief though that distribution CI votes cannot serve as a replacement
for upstream gate.

Ihar
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJVp7HTAAoJEC5aWaUY1u57Dt8IAIY+T99uFmGbSBS3MhekxaJ3
Ry3XD96JOORCvGJ/N8c69GhQrY4h/qa1aK+Ua0cZ4BDo+VwUYvjhVO35xJo+mWXB
Ir8A60iAARpaUXtEZ8E9R8dtEk7mQXAIl/igeq6M/DlTSUG+QSJG1Eh1Au4RI8PH
gyqFuIxNk8UU8hDhBMQLtm+ZB7jIoF2GdAZKnGR2o7ZdwE1Y6PJbzeTkr274XcNM
WLMY/JEv2kkxiZBfKbIUjO8PEab+wN5OOxsEzapx77Kulb1fYAKSxR33jHPm0S/w
zZt2ev6fhkcnTQQGsLfeIfczzYHYmdT7D/85R2lGKR1S/5HKyQxj5omOeycihTQ=
=wbqg
-----END PGP SIGNATURE-----