While I was trying to execute a job using spark-submit, I discover a scala.MatchError at runtime... a DriverStateChanged.FAILED message was send to an actor, and the match statement used was not taking that value into account.
When I inspected that DriverStateChange.scala file I discovered that it's a scala enumeration; if a sealed trait and case objects would have used, the compiler would have refused to compile spark until that match statement would have covered all possible values. As a scala developer I prefer to catch that kind of errors at compile time and I would like to understand why a scala enumeration has been used instead of a sealed trait + case object. If changing that to a sealed (and possible others) keeps binary compatibility, would that kind of PR be welcomed? Kind regards, Luis
