Github user szyszy commented on the issue:
https://github.com/apache/spark/pull/20761
About the ResourceTypeValidator: The reason why this class seems complex is
that it is capable of printing "smart" error messages like they are described
in the scaladoc.
When you define memory with the new style configuration, the error message
points you to the right direction, explicitly printing the right config to use
instead.
See excerpt from scaladoc:
* Then the following two error messages will be printed:
* - "memory cannot be requested with config
spark.yarn.driver.resource.memory,
* please use config spark.driver.memory instead!
* - "cores cannot be requested with config
spark.yarn.executor.resource.cores,
* please use config spark.executor.cores instead!
If you say that this is unnecessary, I can live with that in order to move
this PR forward quicker, but please advise how should I refactor the validator!
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
