incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <gst...@gmail.com>
Subject Re: Insanity (of the release process)
Date Tue, 10 Nov 2009 17:34:20 GMT
On Tue, Nov 10, 2009 at 11:22, Leo Simons <mail@leosimons.com> wrote:
> On Tue, Nov 10, 2009 at 4:05 PM, Greg Stein <gstein@gmail.com> wrote:
>...
>> The IPMC is in charge of its operation. It can redefine the rules of
>> releases as it pleases. The three +1 rule was developed to show that
>> the PMC is "in charge" of the release, and is therefore legally liable
>> for it. The IPMC can do whatever it likes around releases, as long as
>> that process specifically claims or disclaims liability.
>
> Ok, that is interesting (and probably more workable than a big reorg).
> I still think we should claim liability.
>
> Could we, for example, have a release process that is lazy-by-default
> from the IPMC side and still claim that the ASF gets liability?

Unfortunately, no. The IPMC has to be *active* in some way, in order
to show oversight and responsibility. So the "lazy-by-default" won't
work. But your suggestion below might.

> for example, to release:
>
> 1) PPMC must vote for the release according to their rules (which
> should at least match the 3 +1 / majority rule requirements)
> 2) at least one PMC member must vote +1 (usually the mentor)

This basically states, "The PPMC has followed our guidelines and
processes, has been conducted under the purview of the IPMC, and at
our direction. The IPMC hereby directs the PPMC to continue with their
release."

> 3) if there are no -1 votes, the PPMC sends the general@ list a
> request for a release ACK, after they get that ACK from a PMC member,
> they wait for 72 hours, and if there are still no -1s, the release is
> approved.
> 4) if there are any -1 votes, then the rule becomes the normal 3 +1s
> from PMC members / majority

Any -1 votes within the PPMC or from the IPMC should be a trigger.

> Downside:
> * more complex
> * increased dependency on single person to teach the "basics"
>
> Upside:
> * better reflects relationship between incubator and PPMC
> * more responsibility for project
> * hopefully fewer stalled releases

Well.. let's call this the "expedited" form of release. It leaves the
PPMC a bit more self-sufficient.

I'd think that any first release would follow the "standard" release
mechanic. After that, the expedited can be used unless a -1 arises. At
that point, they have to follow the standard process (even if it all
restarts). After that release concludes, they can switch back to
expedited.

I'd really want Roy to review some of this thinking. The real question
is just how far the IPMC can delegate their oversight and authority.

Cheers,
-g

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message