documentation on ForkJoinPool should mention shutdown()

[ktegan]

The threads in the ForkJoinPool are not released unless you call some flavor of shutdown().
The code sample in https://docs.scala-lang.org/overviews/parallel-collections/configuration.html is essentially leaking resources. This issue was discussed here as well: scala/bug#11097 .

[ktegan]

When I try to push a branch to the github repo I get "The requested URL returned error: 403" so I'm not sure that I have permission. In case I don't have permission here are my diffs:

diff --git a/_overviews/parallel-collections/configuration.md b/_overviews/parallel-collections/configuration.md
index 80f07788..9dc9d4df 100644
--- a/_overviews/parallel-collections/configuration.md
+++ b/_overviews/parallel-collections/configuration.md
@@ -43,7 +43,10 @@ Here is a way to change the task support of a parallel collection:
     scala> val pc = mutable.ParArray(1, 2, 3)
     pc: scala.collection.parallel.mutable.ParArray[Int] = ParArray(1, 2, 3)

-    scala> pc.tasksupport = new ForkJoinTaskSupport(new scala.concurrent.forkjoin.ForkJoinPool(2))
+    scala> val forkJoinPool = new java.util.concurrent.ForkJoinPool(2)
+    forkJoinPool: java.util.concurrent.ForkJoinPool = java.util.concurrent.ForkJoinPool@6436e181[Running, parallelism = 2, size = 0, active = 0, running = 0, steals = 0, tasks = 0, submissions = 0]
+
+    scala> pc.tasksupport = new ForkJoinTaskSupport(forkJoinPool)
     pc.tasksupport: scala.collection.parallel.TaskSupport = scala.collection.parallel.ForkJoinTaskSupport@4a5d484a

     scala> pc map { _ + 1 }
@@ -59,6 +62,14 @@ executor:
     scala> pc map { _ + 1 }
     res1: scala.collection.parallel.mutable.ParArray[Int] = ParArray(2, 3, 4)

+    scala> forkJoinPool.shutdown()
+
+Note that if you are making your own `ForkJoinPool` instance you should call
+`ForkJoinPool.shutdown()` when you no longer need the thread pool.  If you do
+not call `ForkJoinPool.shutdown()` and continue to create new instances of
+ForkJoinPool the JVM may eventually run out of available threads and throw a
+`java.lang.OutOfMemoryError`.
+
 When a parallel collection is serialized, the task support field is omitted
 from serialization. When deserializing a parallel collection, the task support
 field is set to the default value-- the execution context task support.

And here is the commit comment:

Call ForkJoinPool.shutdown(), use new package location, PR for #1137

 - If you want to repeatedly create new instances of ForkJoinPool
   you must call some flavor of shutdown() on the pool.  This
   adds the command to the example and gives a note after the
   example.

 - Use java.util.concurrent.ForkJoinPool as the package for
   ForkJoinPool as scala.concurrent.forkjoin.ForkJoinPool is
   deprecated

[SethTisue]

the way this works on GitHub is, you fork the repo, you push the change you want to a branch in your fork, and then you open a pull request from that branch

(if you'd rather not bother with the usual process, I can apply your diffs for you this time, just let me know.)

[SethTisue]

PR: #1138