On Thu, Oct 17, 2013 at 12:11 AM, Benedikt Ritter <[hidden email]>wrote:
> Hi Hen,
> Send from my mobile device
> > Am 17.10.2013 um 08:24 schrieb Henri Yandell <[hidden email]>:
> > Wooo! I won on my first post, and by being on the fence. Be afraid when I
> > have a strong opinion, be wery, wery afraid :) Not allowed to drink
> > Hacking along tonight, I'm reminded of one reason why I would like to try
> > Git in Commons. It's the only place I tend to be working on parallel
> > at the same time and I would like to stash (if that's the right verb) a
> > patch that's part ready but waiting on feedback online. I started to
> > the site with reports based on the uncommitted code and had to abort and
> > restart.
> With git you can stash changes AND work in local branches (or push local
> branches with history to your remote). Stashing is btw supported by some
> IDE without SCM at all (I think idea can do it). Nevertheless I agree with
> you, that this is a big + for git.
> I'd say we push out lang 3.2 and use lang as a test project, if all of
> lang's developers can agree on this.
> <[hidden email]>
I'm up for that. I'll play the role of a clueless newbie, just for the sake
of testing you understand. :)
On 16/10/2013 22:39, Gary Gregory wrote:
> Why don't we side-step the consensus vs. majority and so on issue, and let
> whomever wants git propose to move one component and see how that goes?
I think ignoring the consensus issue and why the discussion and vote
failed to gain consensus would be to lose an opportunity to learn some
valuable lessons that might help the next time there is a proposal to
change something common to all the components.
I have no objection to one component testing out the git to svn move
providing that there is consensus to do so amongst the active developers
of that component. Further, based on what I have seen to far in this
thread, I don't think anyone else would object either.
To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
On Wed, Oct 16, 2013 at 11:56 PM, Henri Yandell <[hidden email]> wrote:
> There's no veto notion here - if we're abiding by the lowest denominator of
> the base Apache voting rules, vetoes are only for code votes. While this is
> to do with code, it's not code itself.
> I see it settled in that an understanding is reached.
> The majority of those voting have indicated that they have a preference for
> git over svn and would like Commons to move in that direction.
> I'm definitely confused by the proposal. Being selfish - what's this going
> to change?
The best thing to change for me will be:
- I'll pay attention to Git pull requests. Right now, I do not because I
cannot simply download a patch from JIRA and use my IDE to apply it. I just
do not want to be bothered with reformatting the pull request or fiddling
with the git command line (or GitHub site) until I get git to create a diff
file SVN will digest.
- I'll be able to stash work in progress to address for urgent tasks. No
more creating a patch, saving it some place, getting a clean sandbox,
applying another patch and so on.
The other stuff will be the same but done differently (Maven magic,
> The discussion implied code review would be used (are we moving
> to RTC?). It implied that there would be issues in checking all of Commons
> out (which has always been very important to me, though I'll admit not
> right now as I've not been supporting cross-Commons features the way
> others, noticeably Sebb, are). If we break the ability for someone to fix
> issues across all components, we increase the likelihood that central
> changes won't be pushed out. Will GitHub pull requests get better? Because
> they're currently a mess. Will we lose existing contributors due to putting
> a hurdle in their way? Will the development workflow change? While I use
> git at the moment, I'm aware I use it in an svn way because I'm always
> hitting pains where git's support for my workflow involves doing odd items
> (acknowledging the issue is me for not developing in a git way). If we move
> a component to git, will I still be able to commit to it via some form of
> svn2git bridge, or will each partial migration mean a component vanishing
> from trunks-proper?
> Browsing the git discuss thread, it was surprisingly light on details. To
> be excited by this and not feel frustrated, I suspect I'll need more
> support (explanations before hand, answers to dumb questions). However this
> seems much like the moves to maven1 and maven2. A difference to the
> maven1/maven2 moves is that they were done with overlap. Components were
> not unusual to have Ant, Maven 1 and Maven 2 build systems.
> Summary: I won't add my vote because I don't understand the question. We're
> not voting on moving to Git, we're voting on something bigger and only
> those voting +1 know what that is :) I'm not against it, but I know there
> will be pain, someone else is going to do all the work [hey, I served my
> time on jira and svn] and I'll slowly catch up and hopefully not get lost
> along the way :)
> An aside: I'm not convinced btw that another thread entitled "[VOTE] Stay
> on Subversion" wouldn't also be passed. To conjecture culturally, those
> fastest to respond are most likely to want to move to Git, while those
> slower are most likely to want to stay on Subversion. Mobilization of the
> SVN vote would probably exceed the Git vote, however I believe there is a
> level of those interacting more often with the scm having a greater voice
> in the choice of system being interacted with.
> On Wednesday, October 16, 2013, James Ring wrote:
> > So did any committer want to exercise a veto? Otherwise the matter is
> > settled right?
> > On Oct 16, 2013 6:38 PM, "sebb" <[hidden email]> wrote:
> > > On 17 October 2013 02:10, Ralph Goers <[hidden email]>
> > wrote:
> > > >
> > > > On Oct 16, 2013, at 2:46 PM, James Ring wrote:
> > > >
> > > >> Do Apache by-laws require a quorum? Was there a quorum for this
> > > >>
> > > >
> > > > Apache voting rules are documented at
> > > http://www.apache.org/foundation/voting.html. However, that page
> > > define "consensus" which is where some of the disagreement came from.
> > >
> > > It's defined in the glossary:
> > >
> > > http://www.apache.org/foundation/glossary.html#ConsensusApproval > > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [hidden email] > > > For additional commands, e-mail: [hidden email] > > >
> > >
> Those who wanted to move to Git have given up several days ago, leaving
> this thread to be 'argued' by
> those who successfully squashed the action. James has already canceled the
> test project request in INFRA, and
> so it seems pointless for this thread to continue. You won, go off and
> have a beer, and enjoy.
Well said. This thread perfectly feeds back into the "what's wrong with