commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Thomas <>
Subject Re: [VOTE] Move Apache Commons to Git for SCM...
Date Wed, 16 Oct 2013 21:14:52 GMT
On 16/10/2013 21:34, Christian Grobmeier wrote:
> On 14 Oct 2013, at 9:13, Mark Thomas wrote:
>> On 13/10/2013 23:59, sebb wrote:
>>> On 13 October 2013 20:47, Phil Steitz <> wrote:
>>>> On 10/13/13 8:09 AM, James Carman wrote:
>>>>> Well, it has been 72 hours, so let's tally up the votes.  As I see it
>>>>> (counting votes on both lists):
>>>>> +1s
>>>>> James Carman
>>>>> Romain Manni-Bucau
>>>>> Matt Benson
>>>>> Benedikt Ritter
>>>>> Bruno Kinoshita
>>>>> Gary Gregory
>>>>> Luc Maisonobe
>>>>> Oliver Heger
>>>>> Christian Grobmeier
>>>>> Torsten Curdt
>>>>> -1s
>>>>> Mark Thomas
>>>>> Thomas Vandahl
>>>>> Damjan Jovanovic
>>>>> Gilles Sadowski
>>>>> Jorg Schaible
>>>>> +0.5
>>>>> Olivier Lamy
>>>>> +0
>>>>> Ralph Goers
>>>>> -0
>>>>> Emmanuel Bourg
>>>>> The vote passes, so Apache Commons will be moving to Git for SCM.  We
>>>>> should begin working on a plan.  I propose we set up a wiki page for
>>>>> that.
>>>> I protest.  It is fine for some components to experiment, but if we
>>>> are going to force all to move, we really need consensus and that is
>>>> clearly not the case here.  I did not vote as I frankly saw the VOTE
>>>> as premature.  We should use VOTEs as a last resort, not a first
>>>> step or way to avoid getting to consensus on non-release issues.
>>> I agree entirely with Phil.
>>> I would have voted -1 earlier, but was off-line for a few days.
>>> This is a huge change, and should not be bulldozed through.
>> I too challenge the assertion that there is consensus for this change.
>> I also agree with Sebb's characterisation of this being "bulldozed
>> through".
> I disagree.
> We have discussed it, we had a vote. We have not voted to push a red
> button on friday
> and to work with git alone on saturday. This was a vote for a general
> decision and
> it is clear (or should be) that changes like that are not made in a
> single day.
> Now what are you folks expecting? A full-fleshed out plan how to move? I
> think we should
> first decide IF we move and that was was happening here.

What I was expecting was decisions to be made on the basis of consensus.

The vote was not for a trial with a single component nor was it for a
gradual move to git as components decided that they wanted to move. The
vote was for a very black and white proposal to move the entire of
Commons from svn to git.

The vote did not get consensus - far from it with around a third of
those voting against the proposal. Therefore my objection was to the
statement in the vote result that "Apache Commons will be moving to Git
for SCM".

> It was also pretty clear to start with a small step first and move a
> single component.
> If that would went wrong we could either go back without bigger loss or
> discuss what needs to be improved.

That is not what was stated in the vote. If it had been, I would have
voted +1. I indicated as much when I voted.

> We are not using experimental bleeding edge technology here. We just
> wanted to decide if we will follow the git path or not.
> I really can't see anything bulldozed here.

The bulldozing was the statement "Apache Commons will be moving to Git
for SCM" when a significant proportion of the committers voted against
such a move.

>> I have no objection to a switch to git for those components where there
>> is consensus to do so amongst the active developers.
>> I continue to strongly recommend that a single component volunteers to
>> be the svn->git guinea pig for Commons and that we allow that component
>> to work out any issues that crop up before any mass switch starts. If
>> there are no issues, great. If there are issues, better to have to deal
>> with one set of them rather than 40+ sets.
> I have not understood it otherwise.
> Why did you start to believe we move all components at once?

The text of the vote, the text of the vote result and the context in
which the vote was conducted. At no point did the James (who was driving
this issue) make any statement that suggested (to me at least) anything
other than a wholesale migration from svn to git.

>> Further, if the consensus amongst the active developers on a component
>> is that they wish to stick to svn, I see no why that component should be
>> forced to switch to git.
> I had the idea too and support it.

At this point I am unclear what support there is for what since folks
appear to have very different interpretations of exactly what was being
voted on.

I think that there is consensus for a single component to trial the svn
to git migration to see how it goes. That approach certainly has my
support although I won't be volunteering any of the components I'm
working on - while I can see the advantages of git, the git mirrors give
me most of the advantages with none of the migration pain. I'm sure that
balance will change over time but personally I'm not there yet.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message