Track JS allocations as they arrive with no affection on performance when tracking is switched off

test/cctest/test-heap-profiler.cc

Index: test/cctest/test-heap-profiler.cc
diff --git a/test/cctest/test-heap-profiler.cc b/test/cctest/test-heap-profiler.cc
index 59f7a258b1094fa5a0ddb2b2e42a2170f2560f55..9ad043a257488bb8451046db26a8a0d7fb1827e7 100644
--- a/test/cctest/test-heap-profiler.cc
+++ b/test/cctest/test-heap-profiler.cc
@@ -2005,3 +2005,20 @@ TEST(JSFunctionHasCodeLink) {
       GetProperty(foo_func, v8::HeapGraphEdge::kInternal, "code");
   CHECK_NE(NULL, code);
 }
+
+
+// This is an example of using checking of JS allocations tracking in a test.

[Hannes Payer (out of office), 2013/10/02 18:00:29]

Either remove the test if it is not testing anything or add a proper test. I
would prefer adding a proper test.

[Alexandra Mikhaylova, 2013/10/03 16:27:55]

This test checks that we don't miss some simplest allocations and
HeapObjectsTracker works correctly. In the HeapObjectsTracker constructor,
allocations tracking is switched on. In the destructor, allocations tracking is
checked and untracked objects are found (if any).

[Hannes Payer (out of office), 2013/10/10 18:54:40]

Sure, but wouldn't it make sense to check which objects are found?

+TEST(HeapObjectsTracker) {
+  LocalContext env;
+  v8::HandleScope scope(env->GetIsolate());
+  HeapObjectsTracker tracker;
+  CompileRun("var a = 1.2");
+  CompileRun("var a = 1.2; var b = 1.0; var c = 1.0;");
+  CompileRun(
+    "var a = [];"
+    "for (var i = 0; i < 5; ++i)"
+    "    a[i] = i;\n"
+    "for (var i = 0; i < 3; ++i)"
+    "    a.shift();\n"
+    );
+}