Collect type feedback for fields.

runtime/vm/raw_object.h

Index: runtime/vm/raw_object.h
diff --git a/runtime/vm/raw_object.h b/runtime/vm/raw_object.h
index 124f341a20a41d0968e3e3e000979501661c3c00..77e76d92f4c92aff3245067fe8bafa1d77670a55 100644
--- a/runtime/vm/raw_object.h
+++ b/runtime/vm/raw_object.h
@@ -664,9 +664,15 @@ class RawField : public RawObject {
                       // where this field is defined.
   RawAbstractType* type_;
   RawInstance* value_;  // Offset in words for instance and value for static.
-  RawObject** to() { return reinterpret_cast<RawObject**>(&ptr()->value_); }
+  RawArray* dependent_code_;
+  RawObject** to() {
+    return reinterpret_cast<RawObject**>(&ptr()->dependent_code_);
+  }
 
   intptr_t token_pos_;
+  intptr_t guarded_cid_;
+  intptr_t is_nullable_;  // kNullCid if field can contain null value and

[Ivan Posva, 2013/03/19 04:42:05]

As far as I understand this is a boolean value. Why not roll that into the
kind_bits_?

[Vyacheslav Egorov (Google), 2013/03/19 06:58:59]

Having a boolean complicates checking and initialization in the generated code,
for the case when field does not have a feedback.

With a boolean I need to have additional control flow that decodes/encodes. With
cid here I just compare is_nullable_ to the cid of the value when comparison
with guarded_cid_ has failed. (and I just store value's cid when initializing it
initially). So the comment is somewhat incorrect. It can contain any other cid
if it is not null.

[Ivan Posva, 2013/03/19 15:48:51]

In that case please update the comment.

[Vyacheslav Egorov (Google), 2013/03/19 19:57:50]

Done.

+                          // kIllegalCid otherwise.
   uint8_t kind_bits_;  // static, final, const, has initializer.
 };