[GLUTEN-11088][VL] Add config in GlutenCastSuite to cast to char/varchar

gluten-ut/spark40/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala

@@ -102,8 +102,6 @@ class VeloxTestSettings extends BackendTestSettings {
     .exclude("cast from timestamp II")
     .exclude("SPARK-36286: invalid string cast to timestamp")
     .exclude("SPARK-39749: cast Decimal to string")
-    // TODO: fix in Spark-4.0
-    .exclude("Casting to char/varchar")
   enableSuite[GlutenTryCastSuite]
     .exclude(
       "Process Infinity, -Infinity, NaN in case insensitive manner" // +inf not supported in folly.
@@ -119,8 +117,6 @@ class VeloxTestSettings extends BackendTestSettings {
     .exclude("data type casting")
     // Revised by setting timezone through config and commented unsupported cases.
     .exclude("cast string to timestamp")
-    // TODO: fix in Spark-4.0
-    .exclude("Casting to char/varchar")
   enableSuite[GlutenCollectionExpressionsSuite]
     // Rewrite in Gluten to replace Seq with Array
     .exclude("Shuffle")

gluten-ut/spark40/src/test/scala/org/apache/spark/sql/catalyst/expressions/GlutenCastSuite.scala

@@ -27,6 +27,14 @@ import java.sql.{Date, Timestamp}
 import java.util.{Calendar, TimeZone}
 
 class GlutenCastSuite extends CastWithAnsiOffSuite with GlutenTestsTrait {
+  override def beforeAll(): Unit = {
+    super.beforeAll()
+    // Need to explicitly set spark.sql.preserveCharVarcharTypeInfo=true for gluten's test
+    // framework. In Gluten, it overrides the checkEvaluation that invokes Spark's RowEncoder,
+    // which requires this configuration to be set.
+    // In Vanilla spark, the checkEvaluation method doesn't invoke RowEncoder.
+    conf.setConf(SQLConf.PRESERVE_CHAR_VARCHAR_TYPE_INFO, true)
+  }
 
   override def cast(v: Any, targetType: DataType, timeZoneId: Option[String] = None): Cast = {
     v match {
@@ -287,5 +295,4 @@ class GlutenCastSuite extends CastWithAnsiOffSuite with GlutenTestsTrait {
     val d = Decimal(c.getTimeInMillis.toDouble / 1000)
     checkEvaluation(cast(d, TimestampType), new Timestamp(c.getTimeInMillis))
   }
-
 }

gluten-ut/spark40/src/test/scala/org/apache/spark/sql/catalyst/expressions/GlutenTryCastSuite.scala

@@ -20,13 +20,21 @@ import org.apache.spark.sql.GlutenTestsTrait
 import org.apache.spark.sql.catalyst.util.DateTimeTestUtils.{withDefaultTimeZone, ALL_TIMEZONES, UTC, UTC_OPT}
 import org.apache.spark.sql.catalyst.util.DateTimeUtils.{fromJavaTimestamp, millisToMicros, TimeZoneUTC}
 import org.apache.spark.sql.internal.SQLConf
-import org.apache.spark.sql.types.{BinaryType, ByteType, DateType, Decimal, DecimalType, DoubleType, FloatType, IntegerType, LongType, ShortType, StringType, TimestampType}
+import org.apache.spark.sql.types._
 import org.apache.spark.util.DebuggableThreadUtils
 
 import java.sql.{Date, Timestamp}
 import java.util.{Calendar, TimeZone}
 
 class GlutenTryCastSuite extends TryCastSuite with GlutenTestsTrait {
+  override def beforeAll(): Unit = {
+    super.beforeAll()
+    // Need to explicitly set spark.sql.preserveCharVarcharTypeInfo=true for gluten's test
+    // framework. In Gluten, it overrides the checkEvaluation that invokes Spark's RowEncoder,
+    // which requires this configuration to be set.
+    // In Vanilla spark, the checkEvaluation method doesn't invoke RowEncoder.
+    conf.setConf(SQLConf.PRESERVE_CHAR_VARCHAR_TYPE_INFO, true)
+  }
 
   testGluten("data type casting") {
     val sd = "1970-01-01"