[CORE][MINOR] Add minor code change to TaskState and Task

## What changes were proposed in this pull request?
- TaskState and ExecutorState expose isFailed and isFinished functions. It can be useful to add test coverage for different states. Currently, Other enums do not expose any functions so this PR aims just these two enums.
- `private` access modifier is added for Finished Task States Set
- A minor doc change is added.

## How was this patch tested?
New Unit tests are added and run locally.

Author: erenavsarogullari <erenavsarogullari@gmail.com>

Closes #15143 from erenavsarogullari/SPARK-17584.

core/src/main/scala/org/apache/spark/TaskState.scala

@@ -21,7 +21,7 @@ private[spark] object TaskState extends Enumeration {
 
   val LAUNCHING, RUNNING, FINISHED, FAILED, KILLED, LOST = Value
 
-  val FINISHED_STATES = Set(FINISHED, FAILED, KILLED, LOST)
+  private val FINISHED_STATES = Set(FINISHED, FAILED, KILLED, LOST)
 
   type TaskState = Value
 

core/src/main/scala/org/apache/spark/scheduler/Task.scala

@@ -239,7 +239,7 @@ private[spark] object Task {
    * and return the task itself as a serialized ByteBuffer. The caller can then update its
    * ClassLoaders and deserialize the task.
    *
-   * @return (taskFiles, taskJars, taskBytes)
+   * @return (taskFiles, taskJars, taskProps, taskBytes)
    */
   def deserializeWithDependencies(serializedTask: ByteBuffer)
     : (HashMap[String, Long], HashMap[String, Long], Properties, ByteBuffer) = {