wypoon commented on PR #12721: URL: https://github.com/apache/iceberg/pull/12721#issuecomment-2856320947
@danielcweeks thanks for the update. In this case, it appears that `org.apache.hadoop.hive.serde2.typeinfo.StructTypeInfo.getAllStructFieldNames()` has changed its return type from `java.util.ArrayList` to `java.util.List` in Hive 4.0. Given that binary compatibility is not guaranteed if the hive-metastore module is compiled against one major version of Hive and run against a different major version, should we build and publish separate artifacts for each major version we support? I know that we bundle the hive-metastore classes in the Flink and the Spark runtime JARs. In https://github.com/apache/iceberg/pull/12693, I had verified that Flink works just fine running against hive-metastore built with Hive 4. So, if the Flink Iceberg user base is amenable, we can ship hive-metastore built with Hive 4 in the Flink runtime JARs. Spark does not currently support running against Hive 4 HMS, and still uses Hive 2.3 in its built-in Hive support, so we can ship hive-metastore built with Hive 2 in the Spark runtime JARs. What are your thoughts? And @pvary? -- 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: issues-unsubscr...@iceberg.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For additional commands, e-mail: issues-h...@iceberg.apache.org
