Github user shimamoto commented on the issue:
https://github.com/apache/incubator-predictionio/pull/436
It is not supposed to need the new Guava dependency because Spark has Guava
dependency.
But if this dependency isn't included, compile error occurs at data project.
When I investigated the cause, I discovered Spark 2.1.1 and Spark 2.2.0
were different Guava dependencies.
- Spark 2.1.1 : com.google.guava:guava:14.0.1
- Spark 2.2.0 : com.google.guava:guava:11.0.2
It is expected guava:14.0.1 in Spark 2.2 intrinsically.
The root cause is probably this:
https://github.com/sbt/sbt/issues/2861
The Apache Curator version referenced in Spark has changed in 2.2.0.
This has been thought to have some effect.
https://github.com/apache/spark/blob/v2.1.1/pom.xml#L130
https://github.com/apache/spark/blob/v2.2.0/pom.xml#L126
https://github.com/apache/curator/blob/2.4.0/pom.xml#L307
https://github.com/apache/curator/blob/apache-curator-2.6.0/pom.xml#L424
---
