The documentation on the website mentions patches should be made for trunk first (http://kafka.apache.org/contributing.html)
However 0.8 branch seems to be the active one and there are multiple 0.8 merges into trunk, but "lagging" behind 0.8.
It seems hard to keep merging 0.8 into trunk and resolving conflicts with every merge.
Is there a plan to have a branching model similar to other Apache projects (I'm thinking about Hadoop and HBase)? I think that having all the latest patches in master/trunk would be helpful.