Update on j3
Hi, The level of activity on the j3 sandbox has been pretty nice in the past few weeks. Would someone working on the code care to give a short update on the latest developments? The commit messages have been rather terse for now... BR, Jukka Zitting
Re: Update on j3
hi jukka On Thu, Sep 1, 2011 at 2:11 PM, Jukka Zitting wrote: > Hi, > > The level of activity on the j3 sandbox has been pretty nice in the > past few weeks. Would someone working on the code care to give a short > update on the latest developments? The commit messages have been > rather terse for now... we're still at a very early stage and the code base is still far from stabilizing. we're experimenting with different ideas/approaches which sometimes get discarded soon after (much like in a proper sandbox ;). that's why commit msgs might be terse at times. the goal is to have a working spi-based stack with the microkernel underneath (jcr2spi->spi2mk). such a stack would serve us as a POC for the technical viability of a MVCC-based microkernel approach. furthermore it would allow to benchmark the mk-based stack against jackrabbit-core and jcr2spi->spi2jcr. one thing we've learned so far is that the spi abstraction layer probably needs to be largely re-written/re-designed in order to fully take advantage of a MVCC-based microkernel. cheers stefan > > BR, > > Jukka Zitting >
Re: Update on j3
Hi, The level of activity on the j3 sandbox has been pretty nice in the past few weeks. Would someone working on the code care to give a short update on the latest developments? The commit messages have been rather terse for now... the goal is to have a working spi-based stack with the microkernel underneath (jcr2spi->spi2mk). such a stack would serve us as a POC for the technical viability of a MVCC-based microkernel approach. furthermore it would allow to benchmark the mk-based stack against jackrabbit-core and jcr2spi->spi2jcr. To get an idea on how far this has been progressed, have a look at the pom [1]. I just enabled the TCK test suite for the default build. The pom lists the known issues and their reasons. All known issues which have a fixme tag in the comment should be fixed. The other ones are currently 'as designed' i.e. limitations which we deem acceptable for the prototype at its current state. one thing we've learned so far is that the spi abstraction layer probably needs to be largely re-written/re-designed in order to fully take advantage of a MVCC-based microkernel. That's why I branched jcr2spi, spi and spi-commons. The idea is to evolve these into the new architecture where possible and rewrite where necessary. There has no real work been done in this area. My large commit [2] from today was just syntactical cleanup. Michael [1] http://svn.apache.org/viewvc/jackrabbit/sandbox/jackrabbit-mk/jackrabbit-spi2microkernel/pom.xml?view=markup [2] http://svn.apache.org/viewvc?view=revision&revision=1164026 cheers stefan BR, Jukka Zitting