Oilpan: Disabling lazy sweeping shouldn't deactivate EAGERLY_FINALIZE.

Source/platform/heap/Heap.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 1163 matching lines @@
     if (size < 128)
         return NormalPage3HeapIndex;
     return NormalPage4HeapIndex;
 }
 
 inline bool Heap::isNormalHeapIndex(int index)
 {
     return index >= NormalPage1HeapIndex && index <= NormalPage4HeapIndex;
 }
 
-#if ENABLE(LAZY_SWEEPING)
 #define DECLARE_EAGER_FINALIZATION_OPERATOR_NEW() \
 public:                                           \
     GC_PLUGIN_IGNORE("491488")                    \
     void* operator new(size_t size)               \
     {                                             \
         return allocateObject(size, true);        \
     }
 #define EAGERLY_FINALIZE_WILL_BE_REMOVED()
 #if ENABLE(ASSERT) && ENABLE(OILPAN)
 class VerifyEagerFinalization {
@@ skipping 12 matching lines @@
     }
 };
 #define EAGERLY_FINALIZE()                             \
 private:                                               \
     VerifyEagerFinalization m_verifyEagerFinalization; \
 public:                                                \
     typedef int IsEagerlyFinalizedMarker
 #else
 #define EAGERLY_FINALIZE() typedef int IsEagerlyFinalizedMarker
 #endif
-#else
-#define DECLARE_EAGER_FINALIZATION_OPERATOR_NEW()
-#define EAGERLY_FINALIZE()
-// TODO(Oilpan): define in terms of Oilpan's EAGERLY_FINALIZE() once lazy
-// sweeping is enabled non-Oilpan.
-#define EAGERLY_FINALIZE_WILL_BE_REMOVED()

[sof, 2015/06/09 13:58:51]

This strikes me as wrong; we need to eagerly finalize some mixins with
!ENABLE(OILPAN) when lazy sweeping is enabled, but aren't. That's a problem with
trunk, not this CL per se.

[sof, 2015/06/09 14:48:10]

I think the analysis from 


https://groups.google.com/a/chromium.org/d/msg/oilpan-reviews/u3dax9mf9Go/a1o...

is still valid & not addressed on trunk in other ways, non-Oilpan.

Trying to reproduce the referred-to ASan failure locally; will see where that
goes before following up.

-#endif
 
 NO_SANITIZE_ADDRESS inline
 size_t HeapObjectHeader::size() const
 {
     size_t result = m_encoded & headerSizeMask;
     // Large objects should not refer to header->size().
     // The actual size of a large object is stored in
     // LargeObjectPage::m_payloadSize.
     ASSERT(result != largeObjectSizeInHeader);
     ASSERT(!pageFromObject(this)->isLargeObjectPage());
@@ skipping 156 matching lines @@
     size_t copySize = previousHeader->payloadSize();
     if (copySize > size)
         copySize = size;
     memcpy(address, previous, copySize);
     return address;
 }
 
 } // namespace blink
 
 #endif // Heap_h