Couple V8AbstractEventListener's lifetime to the V8 listeners lifetime

third_party/WebKit/Source/platform/heap/ThreadState.cpp

Index: third_party/WebKit/Source/platform/heap/ThreadState.cpp
diff --git a/third_party/WebKit/Source/platform/heap/ThreadState.cpp b/third_party/WebKit/Source/platform/heap/ThreadState.cpp
index 5dade81422fb90875c853471da1f73fa91d9e779..1ea1f8ce6626646f26a27849ea0dc717a5f5278e 100644
--- a/third_party/WebKit/Source/platform/heap/ThreadState.cpp
+++ b/third_party/WebKit/Source/platform/heap/ThreadState.cpp
@@ -101,6 +101,9 @@ ThreadState::ThreadState()
     , m_shouldFlushHeapDoesNotContainCache(false)
     , m_gcState(NoGCScheduled)
     , m_traceDOMWrappers(nullptr)
+#if ENABLE(ASSERT)
+    , m_numberOfSelfRefs(0)
+#endif
 #if defined(ADDRESS_SANITIZER)
     , m_asanFakeStack(__asan_get_current_fake_stack())
 #endif
@@ -287,7 +290,7 @@ void ThreadState::cleanup()
         }
         // We should not have any persistents left when getting to this point,
         // if we have it is probably a bug so adding a debug ASSERT to catch this.
-        ASSERT(!currentCount);
+        ASSERT(currentCount == m_numberOfSelfRefs);

[sof, 2015/11/24 08:41:47]

If this holds, tell the persistentRegion() to clear them all out and run another
GC (or several, the live object set should be really quite small by now..) ?

         // All of pre-finalizers should be consumed.
         ASSERT(m_orderedPreFinalizers.isEmpty());
         RELEASE_ASSERT(gcState() == NoGCScheduled);