Fixing remaining VC++ 2015 64-bit build breaks

third_party/WebKit/Source/core/layout/TracedLayoutObject.cpp

Index: third_party/WebKit/Source/core/layout/TracedLayoutObject.cpp
diff --git a/third_party/WebKit/Source/core/layout/TracedLayoutObject.cpp b/third_party/WebKit/Source/core/layout/TracedLayoutObject.cpp
index ee60df3dcdedfef88bb996b50d8a887137531a8c..288bf003d59b5dc40765ae88e996969874e234bd 100644
--- a/third_party/WebKit/Source/core/layout/TracedLayoutObject.cpp
+++ b/third_party/WebKit/Source/core/layout/TracedLayoutObject.cpp
@@ -27,7 +27,9 @@ String TracedLayoutObject::asTraceFormat() const
 }
 
 TracedLayoutObject::TracedLayoutObject(const LayoutObject& object, bool traceGeometry)
-    : m_address((unsigned long) &object)
+    // Cast through uintptr_t and then unsigned int to make the truncation

[Will Harris, 2015/11/10 18:05:14]

looks like m_address should be a different type here, &object will not always be
< 32-bits long. it should perhaps be a uintptr_t?

[brucedawson, 2015/11/10 18:43:26]

Yes, ideally, but the setNumber function that they use takes a double so it
can't handle a 64-bit pointer. Apparently it doesn't really matter and the code
may be deleted and... Sigh...

I've suggested this to the authors. I could change it or leave it - since it's
currently unused it doesn't matter much as long as it compiles.

Note that on Linux and OSX m_address is already always pointer sized, so
uintptr_t would be more consistent.

[Will Harris, 2015/11/10 19:45:46]

this code is crazy. I agree with not complicating the craziness but there really
should be a TODO or something to remove this code if it's not being used.

[brucedawson, 2015/11/10 19:54:33]

I changed it to uintptr_t so it's at least portably weird. I'll follow-up with
the devs.

+    // explicit.
+    : m_address(static_cast<unsigned long>(reinterpret_cast<uintptr_t>(&object)))
     , m_isAnonymous(object.isAnonymous())
     , m_isPositioned(object.isOutOfFlowPositioned())
     , m_isRelPositioned(object.isRelPositioned())