Code review required for commits - formal Team vote.

[Andrew Bartlett]

On Fri, 2012-10-12 at 22:01 -0700, Matthieu Patou wrote:
> On 10/12/2012 04:13 PM, Michael Adam wrote:
> > my vote: +1
> >
> > question:
> > what kind of vote is this intended to be?
> > unanimous? majority? ...
> >
> > comments:
> > * I don't yet understand the reasons for "-1"s
> > * it makes no sense to require review but add a free pass after a
> >    timeout
> > * Productivity need not drop when not pushing ones own
> >    patches, because one can work on a private branch until
> >    patches hit master.
> >
> > More later..
> I'll try to repeat the reasons that made me vote -1.

G'day Matthieu,

Thank you so much for expressing that so well.  

One of the big challenges we have (and why I'm so uncomfortable with
this, particularly the 'let's vote it out' attitude) is that we are all
speaking past each other.

That is (in summery, and far from covering the full views here):

I don't understand why this is urgent to be decided before Jeremy's trip
on Monday, what problem we are trying to solve and where I'll find 1300
reviews a year and someone with the time to put in to a very
heavy-weight process. 

You don't understand how you could possibly continue contributing given
the additional overheads proposed.

Micheal doesn't understand why I feel this is a significant extra
burden.

Jeremy doesn't understand why we are hesitant to pick up an industry
best practice.

Kai doesn't understand why this 'trial' is being phrased in terms place
the burden on those opposing the policy to reverse it, rather than the
burden on those supporting it to show that it really did work well
against all the fears. 

We are all valued and respected members of the Samba Team here, with a
desire to improve Samba.  While we do not even seem to appreciate or
understand the viewpoints of our fellow team is not the time to be
calling a vote on a policy here.  (And if I've misrepresented anyone
here, then I'm sorry, but that's actually my point). 

Instead, it is a time to figure out something that we can all live with,
reasonable terms for a trial and a way to measure the improvement it
hopefully makes during that trial. 

Andrew Bartlett

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org