vinothchandar edited a comment on pull request #2625: URL: https://github.com/apache/hudi/pull/2625#issuecomment-798850321
@nsivabalan thanks for this. I think generating thrice is okay. Don't see any other way, to avoid this. Whenever we can, we should avoid any breaking changes . So can we not change name of the existing bundle name for spark2. i.e `hudi-spark2-bundle_2.12-0.8.0-SNAPSHOT.jar` ? We can just generate one for spark3 called `hudi-spark3-bundle_2.12-0.8.0-SNAPSHOT.jar` Having sparkbundle.version is okay. can we set this through the spark3 maven profile though? That way if you run a run for spark3, the bundle is called `hudi-spark3... ` ? Keeping this in-sync with `spark.version` inside pom is a better approach. less variables to remember at the command line . wdyt? >Do we need to stop uploading hudi-spark_2.12-0.7.0.jar So this is what users who write spark jobs depend on today. We should keep this as-is? They will ultimately compile their application against the right spark version anyway. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
