Ensure that Heap.h (not Visitor.h) is #included from Oilpan user code

Source/platform/heap/Visitor.h

 /*
  * Copyright (C) 2013 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 13 matching lines @@
  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
 #ifndef Visitor_h
 #define Visitor_h
 
+#if !defined(Heap_h) && !defined(Visitor_cpp)
+# error Do not include platform/heap/Visitor.h directly. Include platform/heap/Heap.h instead.

[haraken, 2015/02/16 04:50:45]

Heap.h => Handle.h

I don't fully understand a benefit of disallowing the include. It would be just
an issue of coding convention.

If we want to force some convention, we should force Handle.h (in almost all
cases) and Heap.h (in a limited case where we cannot include Handle.h for some
reason). But still I don't see much benefit in the enforcement.

+#endif
+
 #include "platform/PlatformExport.h"
 #include "platform/heap/StackFrameDepth.h"
 #include "platform/heap/ThreadState.h"
 #include "wtf/Assertions.h"
 #include "wtf/Atomics.h"
 #include "wtf/Deque.h"
 #include "wtf/Forward.h"
 #include "wtf/HashMap.h"
 #include "wtf/HashTraits.h"
 #include "wtf/InstanceCounter.h"
@@ skipping 955 matching lines @@
 struct GCInfoTrait {
     static size_t index()
     {
         return GCInfoAtBase<typename GetGarbageCollectedBase<T>::type>::index();
     }
 };
 
 }
 
 #endif