[Test-Announce] 2011-07-15 @ 17:00 UTC - F16 Alpha blocker bug review #1

[Adam Williamson]

On Wed, 2011-07-13 at 13:22 -0500, Bruno Wolff III wrote:
> On Wed, Jul 13, 2011 at 14:28:33 -0400,
>   James Laska <jlaska at redhat.com> wrote:
> > == Suggested Meeting Preparation ==
> 
> Is there a plan to steamline dealing with all of the systemd create a native
> service bugs?

Here's a plan: we reject them all.

Since we started 'modernizing' the blocker process we've been consistent
in saying that incomplete features are not release blockers, and I'm
increasingly confident this is the correct way to handle it.

Rationale:

# The release blocker system is intended to ensure the quality of the
release, not the completeness of its features

# Fedora works on a time-based release system: we do not consider it a
requirement that features be complete in order to ship a release

# There is a feature process which is separate from the release
validation process, and managed by FESCo

taking all of these together, I think the correct approach is that
incomplete features are not considered to block releases, and that it's
FESCo's responsibility to deal with incomplete / late features under the
feature process, in whatever way they deem best so long as it's
compatible with all freezes. This usually means that they either give
the feature a pass and let the developer complete it a bit late, or drop
the feature; they've never said 'hey, let's hold up the release for
this' except in very specific circumstances (like systemd last time
around). Since this seems to come up at least once a cycle, we should
probably explicitly codify it into the release criteria.

Thoughts welcome :)
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw
http://www.happyassassin.net