JoshRosen commented on code in PR #42599:
URL: https://github.com/apache/spark/pull/42599#discussion_r1302282764
##########
sql/hive/src/main/scala/org/apache/spark/sql/hive/client/IsolatedClientLoader.scala:
##########
@@ -130,8 +130,7 @@ private[hive] object IsolatedClientLoader extends Logging {
}
val hiveArtifacts = version.extraDeps ++
Seq("hive-metastore", "hive-exec", "hive-common", "hive-serde")
- .map(a => s"org.apache.hive:$a:${version.fullVersion}") ++
- Seq("com.google.guava:guava:14.0.1") ++ hadoopJarNames
+ .map(a => s"org.apache.hive:$a:${version.fullVersion}") ++
hadoopJarNames
Review Comment:
Here, I think the Guava version needs to be a function of `version`, rather
than matching Spark's Guava version or omitting Guava: if we remove Guava from
shared classes but also changed the now-non-shared metastore Guava version then
it's the same net effect and breaks older Hive versions.
I believe that newer versions of Hive shade Guava (in which case they're
insensitive to whatever value we set here). I think we could either (a)
continue to unconditionally use Guava 14.0.1 here, or (b) conditionally use it
only for older Hive versions that predated the Guava shading.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
