[esnext] prototype runtime implementation for async functions

src/frames.cc

Index: src/frames.cc
diff --git a/src/frames.cc b/src/frames.cc
index 4f2f0086a794243001b82db0f5a41bec354ebbd5..630d8f96e3093af7bb8fb11d3422ce814d035e61 100644
--- a/src/frames.cc
+++ b/src/frames.cc
@@ -863,6 +863,14 @@ void JavaScriptFrame::GetFunctions(List<JSFunction*>* functions) const {
 void JavaScriptFrame::Summarize(List<FrameSummary>* functions) const {
   DCHECK(functions->length() == 0);
   Code* code = LookupCode();
+
+  // TODO(caitp): these builtins cause crashes here --- that's not good!
+  // Should probably be a better way to keep functions from being included in
+  // stack trace.
+  if (code == *function()->GetIsolate()->builtins()->AsyncFunctionNext() ||

[caitp (gmail), 2016/04/16 20:09:30]

other than the off-by-one bug mentioned and fixed above, this is the remaining
issue with throw resumption. 

Ideally, these functions wouldn't show up in a stack trace at all.

Currently, the stack trace for exceptions thrown within an async function are a
bit weird, since they include the "AsyncFunctionStart" stack entry, which is
serialized as the async function itself (so you can end up with duplicate stack
trace entries, which is not a good look).

+      code == *function()->GetIsolate()->builtins()->AsyncFunctionThrow())
+    return;
+
   int offset = static_cast<int>(pc() - code->instruction_start());
   AbstractCode* abstract_code = AbstractCode::cast(code);
   FrameSummary summary(receiver(), function(), abstract_code, offset,