Right. Being in a same group with Spark, SparkSQL .. make sense if your groovy interpreter is designed to work with Spark.
Otherwise, i think it's better we move z.load into common parents, and make it available for all interpreters. There's related discussion inside of https://issues.apache.org/jira/browse/ZEPPELIN-153, The discussion is about moving ZeppelinContext to common parents. Best, moon On Sun, Jul 5, 2015 at 12:15 PM tog <guillaume.all...@gmail.com> wrote: > Hi Moon > > So if for the groovy interpreter I am working on, I want to benefit from > the classpath set up by z.load, I have to put my interpreter in the same > group as Spark, SparkSql, ... is that correct ? > > Cheers > Guillaume > > On 5 July 2015 at 17:03, moon soo Lee <m...@apache.org> wrote: > >> Hi, >> >> When interpreters are in the same group, they can get reference each >> other. >> >> Following example is how SparkSqlInterpreter get reference to >> SparkInterpreter. >> >> https://github.com/apache/incubator-zeppelin/blob/branch-0.5/spark/src/main/java/org/apache/zeppelin/spark/SparkSqlInterpreter.java#L95 >> >> Hope this helps. >> >> Best, >> moon >> >> >> On Sun, Jul 5, 2015 at 3:40 AM tog <guillaume.all...@gmail.com> wrote: >> >>> Hi >>> >>> I am trying to get a new interpreter. Is there a internal mechanism >>> provided to share variables between interpreters? >>> >>> Cheers >>> Guillaume >>> >>> >>> -- >>> PGP KeyID: 2048R/EA31CFC9 subkeys.pgp.net >>> >> > > > -- > PGP KeyID: 2048R/EA31CFC9 subkeys.pgp.net >
