[Interpreter] Move jump processing to bytecode array writer.

src/interpreter/bytecode-register-optimizer.h

Index: src/interpreter/bytecode-register-optimizer.h
diff --git a/src/interpreter/bytecode-register-optimizer.h b/src/interpreter/bytecode-register-optimizer.h
index d0962eca5f03eca5f8928e4e56c75bd87b6762d8..f1de6fb1c37e562410ade99fba8c1d7f3934264a 100644
--- a/src/interpreter/bytecode-register-optimizer.h
+++ b/src/interpreter/bytecode-register-optimizer.h
@@ -26,9 +26,13 @@ class BytecodeRegisterOptimizer final : public BytecodePipelineStage,
   virtual ~BytecodeRegisterOptimizer() {}
 
   // BytecodePipelineStage interface.
-  size_t FlushForOffset() override;
-  void FlushBasicBlock() override;
   void Write(BytecodeNode* node) override;
+  void WriteJump(BytecodeNode* node, BytecodeLabel* label) override;
+  void BindLabel(BytecodeLabel* label) override;
+  void BindLabel(const BytecodeLabel& target, BytecodeLabel* label) override;
+  Handle<BytecodeArray> ToBytecodeArray(
+      int fixed_register_count, int parameter_count,
+      Handle<FixedArray> handler_table) override;
 
  private:
   static const uint32_t kInvalidEquivalenceId = kMaxUInt32;