RFR: 8220503: Move ShenandoahTerminatorTerminator::should_exit_termination out of header

[Aditya Mandaleeka]

Addressing another silly active bug... this fixes 8220503 by moving the definition of
ShenandoahTerminatorTerminator::should_exit_termination into shenandoahTaskqueue.cpp, matching how
it is in shenandoah/jdk8u.

I removed the comment next to the function. Looking at the history, it was originally next to the
virtual "should_force_termination" function back when that function existed and its default
implementation returned false. I don't know if there's value in keeping the comment now, especially
because it might accidentally be misunderstood to mean that the function returns true. If someone
feels strongly about keeping it, I'm happy to add it back (and perhaps reword it for clarity).

Bug: https://bugs.openjdk.java.net/browse/JDK-8220503
Patch: See below

Thanks,
Aditya

===

diff -r e44b68e5bdaf src/hotspot/share/gc/shenandoah/shenandoahTaskqueue.cpp
--- a/src/hotspot/share/gc/shenandoah/shenandoahTaskqueue.cpp   Sun Mar 08 15:15:38 2020 +0900
+++ b/src/hotspot/share/gc/shenandoah/shenandoahTaskqueue.cpp   Sun Mar 08 21:03:46 2020 -0700
@@ -88,3 +88,7 @@
   }
 }
 #endif // TASKQUEUE_STATS
+
+bool ShenandoahTerminatorTerminator::should_exit_termination() {
+  return _heap->cancelled_gc();
+}
diff -r e44b68e5bdaf src/hotspot/share/gc/shenandoah/shenandoahTaskqueue.hpp
--- a/src/hotspot/share/gc/shenandoah/shenandoahTaskqueue.hpp   Sun Mar 08 15:15:38 2020 +0900
+++ b/src/hotspot/share/gc/shenandoah/shenandoahTaskqueue.hpp   Sun Mar 08 21:03:46 2020 -0700
@@ -335,8 +335,7 @@
   ShenandoahHeap* _heap;
 public:
   ShenandoahTerminatorTerminator(ShenandoahHeap* const heap) : _heap(heap) { }
-  // return true, terminates immediately, even if there's remaining work left
-  virtual bool should_exit_termination() { return _heap->cancelled_gc(); }
+  virtual bool should_exit_termination();
 };

 #endif // SHARE_GC_SHENANDOAH_SHENANDOAHTASKQUEUE_HPP