That would make sense if it would be Jersey's target to be community
controlled. But actually it is Oracle developed and with so few people
that SVN is far from being exhausted. I do not see that there are so
many community contributions that a centralized repository would be
needed. As long as every contributor has to sign a contract with Oracle
and support CDDL (instead of solely GPL or LGPL) I do not see that
Jersey would really become a community driven project. Currently Jersey
is an Oracle-only project and nothing that I have experienced so far
would convince me that a change is really in progress.
-----Original Message-----
From: Cameron Heavon-Jones [mailto:cmhjones_at_gmail.com]
Sent: Dienstag, 12. April 2011 15:41
To: users_at_jersey.java.net
Subject: [Jersey] Re: SVN, GIT or MERCURIAL for Jersey 2.0?
Workflow aside, a DVCS provides an infrastructure which is capable of
being developed and extended in an non-centralized way. That isn't about
having a decentralized software architecture but non-centralized
organizational control.
A GPL project should provision for community dissension and a DVCS
provides this as infrastructure.
cam
On 12/04/2011, at 2:01 PM, Mohan KR (mkannapa) wrote:
> + 1 SVN.
>
> I would echo the same comments, if the current workflow is supported
by SVN, why on earth would one
> want to switch to a DVCS. I have used both GIT/Hg, yeah I see the
benefits of local repositories for a
> really distributed teams. But if one is going to "centralize" it
(push/pull), to synchronize the "nodes", I say
> what's the point?
>
>
> Thanks!
> Regards,
>
> Mohan KR
>
> -----Original Message-----
> From: Marek Potociar [mailto:marek.potociar_at_oracle.com]
> Sent: Monday, April 11, 2011 7:31 AM
> To: users_at_jersey.java.net
> Subject: [Jersey] SVN, GIT or MERCURIAL for Jersey 2.0?
>
> Hey Folks,
>
> For Jersey 2.x we are considering to switch our VCS from SVN into Git
or Mercurial. Sticking to SVN is still an option too. FWIW, here's my
take on the topic:
>
> - Functionality:
>
> SVN meets most of our existing needs today, I do miss the agile nature
of DVCS though and ability to fix a recent commit.
>
> I don't have a clear DVCS winner. Mercurial is compact and easier to
learn, esp. if one comes from the SVN background.
> It has superior branching and merging support compared to SVN. It's
branching concept however seems to be seriously flawed as it is
virtually impossible to delete named branches. Also, it is not as
flexible as Git and configuring it's extensions can be painful.
>
> Git is faster than Mercurial, super flexible, and "unix-like" set of
coherent tools sharing a common platform rather being one compact piece
of software. It is thus bendable to most esoteric work flows. Also it's
merging algorithm is ...wait for it... LEGENDARY! :) Learning to get
full use of Git however requires time (and practice).
>
> - Documentation:
>
> SVN and Mercurial both seem to provide superior documentation compared
to Git.
>
> - Tooling:
>
> SVN has a great tooling support and so does Git. I don't have a
significant experience with Mercurial, but I suppose it will be on par
with Git and SVN.
>
> - Adoption:
>
> SVN is very popular. Also Git appears to gain larger portion of mind
share every day. Git community is very active and visible, e.g.
github.org is especially vibrant. It has almost 10x larger community
than bitbucket.org for Mercurial.
> Mercurial community seems to be both smaller and "quieter".
>
> So, what would you, members of the community, prefer to use going
forward?
>
> [ ] SVN
> [ ] Git
> [ ] Mercurial
>
> Please cast your votes!
>
> Thanks,
> Marek
>