I prefer the solution with restart of the ServiceMix code base and build
it from scratch based on Karaf 3.x (we could use the new plugins and the
new configuration files from Karaf). But I saw the current version of
ServiceMix 5 has been updated to the newest version of Karaf 2.3.x and
the problems I had when I was trying to perform the upgrade to 2.3.3 was
only caused by the sclala library.
I haven't fully tested ServiceMix 5, but the samples I have tested make
no problems. When ServiceMix 5 is almost stable perhaps we could fix
the existing issues and release it in this form (based on Karaf 2.3.x,
probably upgraded later to Karaf 2.4.x) and start ServiceMix 6 from
scratch based on Karaf 3.x and Karaf features. I think people have made
some good work to make ServiceMix5 and perhaps we shouldn't send it into
/dev/null :) What do you think about this?
Regards
Krzysztof
On 14.02.2014 10:33, Gert Vanthienen wrote:
Hi Krzystof,
Up to now, the codebase is using Karaf 2.3.x, but there's nothing
preventing us from upgrading. If we're going for a new start, we
might as well start with the latest and greatest, I think. Would you
fancy taking a stab at that upgrade?
Regards,
Gert Vanthienen
On Fri, Feb 14, 2014 at 10:14 AM, Krzysztof Sobkowiak
<[email protected]> wrote:
Hi
Due to the training for my company I couldn't follow the whole discussion
about future of ServiceMix5. Are we going to upgrade it to Karaf 3.x or will
it be released on basis of Karaf 2.x?
Regards
Krzysztof
On 13.02.2014 09:53, Gert Vanthienen wrote:
L.S.,
Moving the discussion from the users@ to dev@ mailing list, it would
be good to figure out how to go about things.
For now, I would propose we avoid adding any extra complexity and just
focus on getting a simple assembly project in place that combines the
latest and greatest bits of ActiveMQ, CXF, Camel, Karaf, ...
How about we start with the ServiceMix 5 codebase, remove the new
Camel interceptor bits for now and just focus on getting the assembly
itself, the examples and the integration tests working? That way, we
should have a fairly manageable amount of work to perform before we
can get to a new release.
We have a board report coming up in March, so we can convey this new
direction in that. Perhaps we should use the next board report after
that as a checkpoint to see if we are making any progress: by then, I
think we should have been able to do at least one or two releases and
we should have found a few people that are willing/able to help out
with the work. If we have not made any progress on the community
aspect by then, it may be a good time to reflect on how to gracefully
move people over to a plain Karaf/CXF/Camel/... solution instead.
Wdyt?
Gert Vanthienen
--
Krzysztof Sobkowiak
JEE & OSS Architect | Technical Architect @ Capgemini
Capgemini <http://www.pl.capgemini.com/> | Software Solutions Center
<http://www.pl.capgemini-sdm.com/> | Wroclaw
e-mail: [email protected] <mailto:[email protected]> |
Twitter: @KSobkowiak
--
Krzysztof Sobkowiak
JEE & OSS Architect | Technical Architect @ Capgemini
Capgemini <http://www.pl.capgemini.com/> | Software Solutions Center
<http://www.pl.capgemini-sdm.com/> | Wroclaw
e-mail: [email protected] <mailto:[email protected]> |
Twitter: @KSobkowiak
