@oliver:
as you might have seen those branches are just external. they never get
added to the main repository. instead the content is applied via a rebase.
at deltaspike we only used that if it's easier to show something based on a
refactoring or a small prototype (instead of long threads on the
mailing-list).

@anatole:
at the asf the whole code-base is open to every committer of the project
and everybody is equal (when it comes to commits).
a commit which isn't "optimal" can be fixed by a subsequent commit.
what we did at other projects is e.g. a release-review (at least in view of
the api).
being too restrictive during the development-phase will scare
community-members off.

regards,
gerhard



2014-11-27 16:53 GMT+01:00 Anatole Tresch <[email protected]>:

> My experience shows that they are only rarely used. Most of the discussions
> we had in JSR 354 were done only on the mailing list, and I will follow a
> similar discussion scheme to enable this as well here.
> Basically every EG member could commit to the API, whereas RI and TCK was
> quite open. Basically this has proven to work well, as long as every commit
> is reviewed by somebody (the project or stream lead).
>
> Static code analysis, continuous build etc is a must for me, for
> everything.
>
> I tend to say: let us start on the mailing list and keep things simple
> first so we have quick progress. I am quite sure, especially with the
> quality of committers on the list, we will not face bigger issues. But
> definitively, when growing over time, we must add some kind of more
> advanced quality gates.
>
> Nevertheless we must ensure that everybody cannot change APIs without prior
> discussion and agreement with the core team. But from my experience (JSR
> 354) this was never an issue when you have a relatively small committer
> community and the process is clearly stated (documented). Bugfixes / tests
> are generally easier to handle.
>
> Anatole
>
>
> Oliver B. Fischer <[email protected]> schrieb am Thu Nov 27 2014 at
> 16:37:37:
>
> > I still doubt that feature branches the right way to go. I don't refuse
> > them generally but they are not a good choice for ongoing development
> > which is not finished after a short period. Many feature branches lead
> > to late integration and the effort for integration is much higher then
> > without. They even prohibit refactorings if a large amount of the code
> > base is effected.
> >
> > If we worry about the code quality there are other ways open for us.
> > Better and more tests, better coverrage, static code analysis and good
> > and communicated coding rules.
> >
> > I would prefer to have a stable branch and an development branch.
> > Development branch is open for direct commits until it enters a
> > stabilizing phase.
> >
> > I still think that we should have an open commit policy until we have
> > reached a specific state for Tamaya.
> >
> > wdyt?
> >
> > Best,
> >
> > Oliver
> >
> >
> >
> >
> > Am 27.11.14 10:40, schrieb Gerhard Petracek:
> > > you don't need a formal vote (= 72h waiting time) in any case.
> > > for sure we need to discuss every part of the api (also the parts of
> the
> > > initial import).
> > > (for such cases a formal vote is just needed if there is no agreement.)
> > >
> > > @feature branches:
> > > if we need to discuss ideas which are prototyped, we can follow [1].
> > > -> you don't get tons of branches and merge-commits, but you have most
> of
> > > the benefits.
> > >
> > > @github:
> > > it used to be a read-only mirror.
> > >
> > > just fyi:
> > > history-rewrites of pushed commits are restricted
> > >
> > > regards,
> > > gerhard
> > >
> > > [1]
> > > http://deltaspike.apache.org/suggested-git-workflows.html#
> > discussion-workflow-optional
> > >
> > >
> > >
> > > 2014-11-27 10:21 GMT+01:00 Anatole Tresch <[email protected]>:
> > >
> > >> The project is also mirrored to GH, i did not check if it is writable
> > from
> > >> the mirror as well, but if so wecan benefit from the GH
> > infrstructure...?
> > >>
> > >> -
> > >> Anatole Tresch
> > >> Glärnischweg 10
> > >> 8620 Wetzikon
> > >> Tel +41 (43) 317 05 30
> > >> -
> > >> Send from Mobile
> > >>
> > >>> Am 27.11.2014 um 10:14 schrieb "Oliver B. Fischer" <
> > >> [email protected]>:
> > >>> I agree on the quality aspect and that the hard part of Tamaya will
> be
> > >> the discussion. Only my personal experience with feature branches is
> > not so
> > >> positive. I know it works for a lot of people and even for a lot of
> > >> projects at GitHub and at Bitbucket. Such plattforms provide tooling
> > >> support we don't have. Or do we?
> > >>> Bye,
> > >>>
> > >>> Oliver
> > >>>
> > >>> --
> > >>> N Oliver B. Fischer
> > >>> A Schönhauser Allee 64, 10437 Berlin, Deutschland/Germany
> > >>> P +49 30 44793251
> > >>> M +49 178 7903538
> > >>> E [email protected]
> > >>> S oliver.b.fischer
> > >>> J [email protected]
> > >>> X http://xing.to/obf
> > >>>
> >
> > --
> > N Oliver B. Fischer
> > A Schönhauser Allee 64, 10437 Berlin, Deutschland/Germany
> > P +49 30 44793251
> > M +49 178 7903538
> > E [email protected]
> > S oliver.b.fischer
> > J [email protected]
> > X http://xing.to/obf
> >
> >
>

Reply via email to