[async-iteration] add support for for-await-of loops in Async Functions

src/interpreter/interpreter-intrinsics.cc

 // Copyright 2015 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "src/interpreter/interpreter-intrinsics.h"
 
 #include "src/code-factory.h"
 
 namespace v8 {
 namespace internal {
@@ skipping 380 matching lines @@
 
 void IntrinsicsHelper::AbortIfArgCountMismatch(int expected, Node* actual) {
   InterpreterAssembler::Label match(assembler_);
   Node* comparison = __ Word32Equal(actual, __ Int32Constant(expected));
   __ GotoIf(comparison, &match);
   __ Abort(kWrongArgumentCountForInvokeIntrinsic);
   __ Goto(&match);
   __ Bind(&match);
 }
 
+Node* IntrinsicsHelper::CreateAsyncFromSyncIterator(Node* args_reg,
+                                                    Node* arg_count,
+                                                    Node* context) {
+  // TODO(caitp): split AsyncFromSyncIterator functionality out of
+  //              https://codereview.chromium.org/2622833002
+  __ CallRuntime(Runtime::kAbort, __ SmiConstant(kGetIterator),
+                 __ NoContextConstant());

[rmcilroy, 2017/01/20 10:34:01]

Could we avoid adding this as an intrinsic that just calls a runtime call and
just call the runtime function itself?

[neis, 2017/01/20 14:38:15]

If this will be implemented in a separate CL, please note that in the
description of this CL.

[caitp, 2017/01/20 20:17:19]

I would have thought the TODO() indicated this. I think, for now, I'll leave the
interpreter intrinsics untouched and just call the runtime directly from
bytecode-generator, as rmcilroy is asking

[neis, 2017/01/24 12:55:45]

Sure, but many people read the CL description while only a few read the code.

+  return __ UndefinedConstant();
+}
+
 }  // namespace interpreter
 }  // namespace internal
 }  // namespace v8