[SPARK-18055][SQL] Use correct mirror in ExpresionEncoder

Previously, we were using the mirror of passed in `TypeTag` when reflecting to build an encoder.  This fails when the outer class is built in (i.e. `Seq`'s default mirror is based on root classloader) but inner classes (i.e. `A` in `Seq[A]`) are defined in the REPL or a library.

This patch changes us to always reflect based on a mirror created using the context classloader.

Author: Michael Armbrust <michael@databricks.com>

Closes #17201 from marmbrus/replSeqEncoder.

1 files changed, 11 insertions, 0 deletions

diff --git a/repl/scala-2.11/src/test/scala/org/apache/spark/repl/ReplSuite.scala b/repl/scala-2.11/src/test/scala/org/apache/spark/repl/ReplSuite.scala
index 55c91675ed..121a02a9be 100644
--- a/repl/scala-2.11/src/test/scala/org/apache/spark/repl/ReplSuite.scala
+++ b/repl/scala-2.11/src/test/scala/org/apache/spark/repl/ReplSuite.scala
@@ -473,4 +473,15 @@ class ReplSuite extends SparkFunSuite {
     assertDoesNotContain("AssertionError", output)
     assertDoesNotContain("Exception", output)
   }
+
+  test("newProductSeqEncoder with REPL defined class") {
+    val output = runInterpreterInPasteMode("local-cluster[1,4,4096]",
+      """
+      |case class Click(id: Int)
+      |spark.implicits.newProductSeqEncoder[Click]
+    """.stripMargin)
+
+    assertDoesNotContain("error:", output)
+    assertDoesNotContain("Exception", output)
+  }
 }