On Wed, 01 Jun 2011 10:15:54 +0200, Sanne Grinovero <sanne.grinovero(a)gmail.com&gt; wrote: > Some time ago I experienced a similar issue with Hibernate Core's > repository, > and solved it by renaming my master, checking out a fresh copy and > rebasing in my changes from my local copy. Right. That was the second option. Keep what was on master on GitHub and rebasing the local changes on top of it. We did it the other way around, because we thought that most people would have the state we had locally. So restoring this would create less issues. > From what I understood, (when it happened before) it seemed that > somebody had renamed the master from another branch, but I could have > messed up differently. Well, I stay away from renaming master :-) In the long run I would like us to move Core development to the same development style we have in Search and Valiator. Everything is done via pull requests. Not only would this prevent situation like this one, but it also increases code awareness, since you see what is happening across the whole code base. Maybe Search and Validator are a little easier to handle, because we are less people working on it, but I think this should not stop us to try a similar approach on Core. Maybe not a good time to start right away with this due to the amount of changes atm, but maybe once the code settles a little more ...

--Hardy _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

I did the force, but to be honest I am still not understanding why this was a problem. Perhaps I am just misunderstanding what a forced push does.

i guess before you force push, you didn't pull so, the commits after your last pull just losted after you force push ----------- Strong Liu <stliu(a)hibernate.org&gt; http://hibernate.org http://github.com/stliu On Jun 1, 2011, at 5:12 PM, Steve Ebersole wrote: > I did the force, but to be honest I am still not understanding why this

> a problem. Perhaps I am just misunderstanding what a forced push does. > On Jun 1, 2011 3:43 AM, "Sanne Grinovero" <sanne(a)hibernate.org&gt; wrote: >> 2011/6/1 Hardy Ferentschik <hardy(a)hibernate.org&gt;: >>> On Wed, 01 Jun 2011 10:15:54 +0200, Sanne Grinovero >>> <sanne.grinovero(a)gmail.com&gt; wrote: >>> >>>> Some time ago I experienced a similar issue with Hibernate Core's >>>> repository, >>>> and solved it by renaming my master, checking out a fresh copy and >>>> rebasing in my changes from my local copy. >>> >>> Right. That was the second option. Keep what was on master on GitHub

>>> rebasing the local changes on top of it. >>> We did it the other way around, because we thought that most people >>> would have the state we had locally. So restoring this would create

>>> issues. >>> >>>> From what I understood, (when it happened before) it seemed that >>>> somebody had renamed the master from another branch, but I could have >>>> messed up differently. >>> >>> Well, I stay away from renaming master :-) >>> In the long run I would like us to move Core development to the same >>> development >>> style we have in Search and Valiator. Everything is done via pull > requests. >>> Not only would this prevent situation like this one, but it also > increases >>> code awareness, since you see what is happening across the whole code > base. >>> Maybe Search and Validator are a little easier to handle, because we

>>> less >>> people working on it, but I think this should not stop us to try a > similar >>> approach on Core. >>> Maybe not a good time to start right away with this due to the amount

>>> changes >>> atm, but maybe once the code settles a little more ... >> >> It's working pretty well on Infinispan, same model as Search but with >> a fairly larger team. Sometimes when there's less people involved pull >> requests tend to stack up a bit and we need to ping each other for >> who's going to volunteer to take X but we never lack volunteers. >> I don't think you need it, especially when you need all possible speed >> and are releasing alpha versions of a new mayor version. >> >> Sanne >> >>> >>> --Hardy >>> >>> _______________________________________________ >>> hibernate-dev mailing list >>> hibernate-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >>> >> _______________________________________________ >> hibernate-dev mailing list >> hibernate-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/hibernate-dev > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev

I did a pull+rebase. The rebase had conflicts. But I never found a way to continue the rebase initiated from pull. I thought `git rebase --continue`was the incantation, but it was not working. So I guess conceivably there were additional commits still to process which would explain the missing commits. Is there a way to revert my forced push and I can try again?

On Jun 1, 2011 4:46 AM, "Steve Ebersole" <steve(a)hibernate.org <mailto:steve@hibernate.org>> wrote: > I did pull. I pulled immediately before the push > On Jun 1, 2011 4:28 AM, "Strong Liu" <stliu(a)hibernate.org <mailto:stliu@hibernate.org>> wrote: >> i guess before you force push, you didn't pull >> so, the commits after your last pull just losted after you force push >> >> ----------- >> Strong Liu <stliu(a)hibernate.org <mailto:stliu@hibernate.org>> >> http://hibernate.org >> http://github.com/stliu >> >> On Jun 1, 2011, at 5:12 PM, Steve Ebersole wrote: >> >>> I did the force, but to be honest I am still not understanding why this > was >>> a problem. Perhaps I am just misunderstanding what a forced push does. >>> On Jun 1, 2011 3:43 AM, "Sanne Grinovero" <sanne(a)hibernate.org <mailto:sanne@hibernate.org>> wrote: >>>> 2011/6/1 Hardy Ferentschik <hardy(a)hibernate.org <mailto:hardy@hibernate.org>>: >>>>> On Wed, 01 Jun 2011 10:15:54 +0200, Sanne Grinovero >>>>> <sanne.grinovero(a)gmail.com <mailto:sanne.grinovero@gmail.com>> wrote: >>>>> >>>>>> Some time ago I experienced a similar issue with Hibernate Core's >>>>>> repository, >>>>>> and solved it by renaming my master, checking out a fresh copy and >>>>>> rebasing in my changes from my local copy. >>>>> >>>>> Right. That was the second option. Keep what was on master on GitHub > and >>>>> rebasing the local changes on top of it. >>>>> We did it the other way around, because we thought that most people >>>>> would have the state we had locally. So restoring this would create > less >>>>> issues. >>>>> >>>>>> From what I understood, (when it happened before) it seemed that >>>>>> somebody had renamed the master from another branch, but I could have >>>>>> messed up differently. >>>>> >>>>> Well, I stay away from renaming master :-) >>>>> In the long run I would like us to move Core development to the same >>>>> development >>>>> style we have in Search and Valiator. Everything is done via pull >>> requests. >>>>> Not only would this prevent situation like this one, but it also >>> increases >>>>> code awareness, since you see what is happening across the whole code >>> base. >>>>> Maybe Search and Validator are a little easier to handle, because we > are >>>>> less >>>>> people working on it, but I think this should not stop us to try a >>> similar >>>>> approach on Core. >>>>> Maybe not a good time to start right away with this due to the amount > of >>>>> changes >>>>> atm, but maybe once the code settles a little more ... >>>> >>>> It's working pretty well on Infinispan, same model as Search but with >>>> a fairly larger team. Sometimes when there's less people involved pull >>>> requests tend to stack up a bit and we need to ping each other for >>>> who's going to volunteer to take X but we never lack volunteers. >>>> I don't think you need it, especially when you need all possible speed >>>> and are releasing alpha versions of a new mayor version. >>>> >>>> Sanne >>>> >>>>> >>>>> --Hardy >>>>> >>>>> _______________________________________________ >>>>> hibernate-dev mailing list >>>>> hibernate-dev(a)lists.jboss.org <mailto:hibernate-dev@lists.jboss.org> >>>>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >>>>> >>>> _______________________________________________ >>>> hibernate-dev mailing list >>>> hibernate-dev(a)lists.jboss.org <mailto:hibernate-dev@lists.jboss.org> >>>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >>> _______________________________________________ >>> hibernate-dev mailing list >>> hibernate-dev(a)lists.jboss.org <mailto:hibernate-dev@lists.jboss.org> >>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >>

On Wed, 01 Jun 2011 13:13:09 +0200, Steve Ebersole <steve(a)hibernate.org&gt; wrote: > On 06/01/2011 05:14 AM, Steve Ebersole wrote: >> I did a pull+rebase. The rebase had conflicts. But I never found a way >> to continue the rebase initiated from pull. I thought `git rebase >> --continue`was the incantation, but it was not working. So I guess >> conceivably there were additional commits still to process which would >> explain the missing commits. >> >> Is there a way to revert my forced push and I can try again? > > I guess that is what you did Strong? > https://github.com/hibernate/hibernate-core/commit/7a187648932f33b0b8138b... We pushed our version of the repository (forced) and then applied the changes you committed and we didn't have yet on top of it. Master should contain now the work from everyone :-) No need for you to re-commit anything. Just make sure that you sync properly w/ master now.

--Hardy _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev