On Sep 12, 2013, at 12:18 PM, Shai Erera <[email protected]> wrote: > I never spoke about random immediate releases. > Yeah, sorry - I responded to you and then to the thread. Didn't mean to imply that you argued for that - my reply to you was in terms of what can make it in.
> All I said is that we should branch early, then give some time for people > before the RC is cut. Branching early, I think, puts less risk that something > random will get in. If you need to port something to the branch, you know > it's the release branch. Maybe you'd decide not to do it and let it bake in > 4x a while longer. > I have no problem with branching - I also didn't mean to reject your entire email in my reply - just the part I quoted. > And maybe this is all just bullshit philosophy, I don't know. I do think that > we're spending too much time on these silly arguments. We've never had two > releases in a week's difference, there's no point threatening to do it. Plus, > if it will happen, I'd probably vote for it too. > Because we never had it does not mean it's not entirely possible. And yes, it only takes 3 votes - both releases would likely go out. It's a 'threat' in the weakest sense of the term - who does it harm? But it would let us put out a release that we felt better about (those that don't want immediate releases). I don't see that it would be meant to 'punish' or 'harm' those doing the release before that. > And this time Adrien already said he won't be able to work on it until next > Wednesday. So what are we arguing about? Rules? Principles? > Yup. This is how the culture and best practices evolve - these discussions. Arguing about rules and principles - we have done it since the start, and that is how our community has evolved. I fully believe these discussions help plot the course of the project - if I did not, I would not waste my time on them. > As far as I'm concerned, it should be the RM decision. Maybe he has time now > but won't in a week? Fine, then another RM can come a week later and do > another release... that's fine too. > Perhaps it's a bad time to release for that RM then. Not a big deal these days, because releasing is easier and it tends to happen more anyway. I know if 4.5 was not picked up by Adrien, I was planning on jumping on it shortly. - Mark > Shai > > On Sep 12, 2013 6:53 PM, "Mark Miller" <[email protected]> wrote: > > On Sep 12, 2013, at 11:45 AM, Robert Muir <[email protected]> wrote: > > > I guess its ok that we disagree: fortunately you cannot veto releases, > > so its up to the individual release manager. > > > > I can throw up a RC right now and if 2 other people think its ok, its > > out the door. > > > > If i start to see the release process dragging ass for weeks and > > taking forever I might do such a thing. its just a few unix comands > > and an email. > > > > Your right - it only takes 3 PMC members to get a no warning release out the > door. > > However, fighting with the rules is fairly anti productive. > > Yes, a group could start doing random releases with no warning. > > Another group could start doing releases a week or two after those - taking > their time to wrap some things up first. > > Another guy could say, I'm going to start doing a release the last day of > every month, and then you will always be able to plan and we can work > together. > > Yes, if we can't get consensus, we can fall back to nuclear rule fallout. > > It seems more productive to simply give a weeks notice. > > If you are frustrated that a release has taken a long time and you want to > put up an RC? More power to you. That's not a random, no warning release. I'd > just as likely vote for it. > > - Mark > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] >
