I still think it makes sense to keep it as a top level directory (would not include it in recipes). 'zookeeper-jute' is certainly a reasonable name. :)
-Brian On Thu, Aug 23, 2018 at 7:59 AM Norbert Kalmar <nkal...@cloudera.com.invalid> wrote: > Hi folks, > > So jute is a little bit the odd one out in ZooKeeper. It was pulled out of > Hadoop and it evolved independently since with ZooKeeper. And it has its > own namespace in ZK: org.apache.jute beside org.apache.zookeeper. > > For me, this is a bit odd, it suggests its a top level apache project. > The original plan in the maven migration was to move jute to simply in a > "jute" top level directory in zookeeper. Like this: > > > zookeeper | -bin | -conf | -jute | -zookeeper-client | -zookeeper-contrib | > -zookeeper-docs | -zookeeper-it (integration tests) | -zookeeper-server | > -zookeeper-recipes > > But this just seems a bit strange to me. I was thinking of possibly moving > into recipes, but then again, I ended up thinking it could still get its > own top level directory, but with the name "zookeeper-jute", as this > particular project is only used by ZooKeeper (of course I would keep the > package name intact). > > What do you think? > As this is a change that was not in the initial doc about maven migration, > I thought it's best to ask. > > Regards, > Norbert > > p.s.: The best thing would be to use a more standardized library for > serialization like protobuf or Avro. But that's just a distant dream right > now, maybe ZooKeeper v5.0 ;) >