Github user srowen commented on a diff in the pull request: https://github.com/apache/spark/pull/23062#discussion_r234846234 --- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/ScalaReflection.scala --- @@ -788,12 +788,37 @@ object ScalaReflection extends ScalaReflection { } /** - * Finds an accessible constructor with compatible parameters. This is a more flexible search - * than the exact matching algorithm in `Class.getConstructor`. The first assignment-compatible - * matching constructor is returned. Otherwise, it returns `None`. + * Finds an accessible constructor with compatible parameters. This is a more flexible search than + * the exact matching algorithm in `Class.getConstructor`. The first assignment-compatible + * matching constructor is returned if it exists. Otherwise, we check for additional compatible + * constructors defined in the companion object as `apply` methods. Otherwise, it returns `None`. */ - def findConstructor(cls: Class[_], paramTypes: Seq[Class[_]]): Option[Constructor[_]] = { - Option(ConstructorUtils.getMatchingAccessibleConstructor(cls, paramTypes: _*)) + def findConstructor(cls: Class[_], paramTypes: Seq[Class[_]]): Option[Seq[AnyRef] => Any] = { --- End diff -- Just cast with: `method.apply(args: _*).asInstanceOf[T]`. That's valid because that is actually the assumption this code is making. I get no compile errors with no further changes beyond that. `cls: Class[T]` knows it produces a T from `newInstance`; did you make that change?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org