All, As someone who has to help organizations migrate and define adoption plans I much prefer the plan outlined by Andreas. Creating a break in backwards compatibility in a minor release causes a loss of faith in enterprise organizations. That faith is hard to rebuild later on without a compelling reason.
I also believe having a long term roadmap to adopt OSGi r.5 as part of the Karaf 4.0 baseline reinforces good faith. It allows technical leads and architects the time to create a migration plan. It also shows maturity in the leadership of Karaf which makes my life much easier. Just my two cents. Scott ES Scott England-Sullivan http://FuseSource.com On Aug 19, 2012, at 6:45 AM, Andreas Pieber <anpie...@gmail.com> wrote: > IMHO we don't need to include the same amount of features for 4.0 as > we did for 3.0 and I neither see any problem releasing Karaf 4 in 6-10 > month... > > Kind regards, > Andreas > > On Sun, Aug 19, 2012 at 1:01 PM, Łukasz Dywicki <l...@code-house.org> wrote: >> We spent one year before releasing 3.x. With this velocity we'll support >> OSGi 5 in same time as Java will support Jigsaw. >> >> IMHO Karaf 4 is too far to plan anything in it, especially that Equinox and >> Felix will be OSGi 5 compatible in more or less half year, I guess. >> >> Best regards, >> Lukasz >> >> Wiadomość napisana przez Jean-Baptiste Onofré w dniu 18 sie 2012, o godz. >> 23:49: >> >>> The first solution sounds good to me. >>> >>> Regards >>> JB >>> >>> On 08/18/2012 07:19 AM, Andreas Pieber wrote: >>>> TBH I personally would rather prefer: >>>> >>>> - Karaf 2.2.x => OSGi core 4.2.0 >>>> - Karaf 2.3.x => OSGi core 4.3.0 >>>> - Karaf 3.x.y => OSGi core 4.3.0 >>>> - Karaf 4.x.y => OSGi core 5.0.0 >>>> >>>> or >>>> >>>> - Karaf 2.2.x => OSGi core 4.2.0 >>>> - Karaf 2.3.x => OSGi core 4.3.0 >>>> - Karaf 3.x.y => OSGi core 5.0.0 >>>> >>>> I know that both options are not really optimal, but I don't think >>>> that we really want to upgrade the inner core of the entire thing to >>>> the next major version in a minor of Karaf. There where some API >>>> changes from a user lvl of view which disallow to drop the same bundle >>>> into a 4.2/4.3 environment and a 5.0. I do not really fancy the idea >>>> of breaking compatibility between minor releases. If ppls setup break >>>> from one minor to the next they lost trust and upgrade ways too late >>>> although there should be no reason. >>>> >>>> WDYT? >>>> >>>> Kind regards, >>>> Andreas >>>> >>>> On Sat, Aug 18, 2012 at 12:47 AM, Jean-Baptiste Onofré <j...@nanthrax.net> >>>> wrote: >>>>> Hi team, >>>>> >>>>> As you know, OSGi core 5.0.0 has been released. >>>>> >>>>> I would like to discuss about which version of Karaf should support which >>>>> version of the OSGi core (and so the corresponding OSGi frameworks). >>>>> >>>>> I have the following proposal: >>>>> >>>>> - Karaf 2.2.x => OSGi core 4.2.0 >>>>> - Karaf 2.3.x => OSGi core 4.3.0 >>>>> - Karaf 3.0.x => OSGi core 4.3.0 >>>>> - Karaf 3.1.x => OSGi core 5.0.0 >>>>> >>>>> I would wait Karaf 3.1.x to update to OSGi core 5.0.0 as the OSGi >>>>> frameworks >>>>> are not yet fully ready on this specification. >>>>> >>>>> WDYT ? >>>>> >>>>> Regards >>>>> JB >>>>> -- >>>>> Jean-Baptiste Onofré >>>>> jbono...@apache.org >>>>> http://blog.nanthrax.net >>>>> Talend - http://www.talend.com >>> >>> -- >>> Jean-Baptiste Onofré >>> jbono...@apache.org >>> http://blog.nanthrax.net >>> Talend - http://www.talend.com >>
