advancedxy commented on PR #42236: URL: https://github.com/apache/spark/pull/42236#issuecomment-1668006357
> While I'm not certain if it's reasonable, I still want to point out that relocating the content of the `spark-protobuf` module may result to a poorer user experience: In order to use this sql function, users have no choice but to relocate the content of the Java PB description files used in their business according to Spark's project rules(replcace all `com.google.protobuf.` to`org.sparkproject.spark_protobuf.protobuf.` in their generated Java files, Perhaps I misunderstood something?). Is this really user-friendly for existing data and businesses? Meanwhile, the `spark-protobuf` module is a module that won't be packaged into spark-client tar ball, is it very risky to only publish unshaded jars? @rangadi @HyukjinKwon +1 for @LuciferYang 's opinion. It's not reasonable that user has to package its protobuf jar relocated. Since the spark-protobuf jar is not included in the spark-client binary, I think a better solution would be publishing both unshaded and shaded jars. Users could simply packing their protobuf message classes into a jar and use spark-protobuf jar as needed. The shaded spark-protobuf jar is kept for compatibility and also for last resort. According to https://protobuf.dev/support/cross-version-runtime-guarantee/, I believe as long as Spark uses higher 3.x versions, user compiled java classes should be fine to run. -- 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]
