Fix not throwing error when redefine eval or arguments in strict mode.

src/parsing/parser-base.h

Index: src/parsing/parser-base.h
diff --git a/src/parsing/parser-base.h b/src/parsing/parser-base.h
index 6058d25c6f887069b59e777fb3815366e9f0c79b..bd9e77c4a35a0711dd2850bbf0535943e7d45eb4 100644
--- a/src/parsing/parser-base.h
+++ b/src/parsing/parser-base.h
@@ -1909,6 +1909,13 @@ typename ParserBase<Traits>::ExpressionT ParserBase<Traits>::ParsePropertyName(
       scanner()->IsGetOrSet(is_get, is_set);
       if (this->IsAwait(*name)) {

[adamk, 2016/08/05 22:29:29]

I actually don't know why we handle this here instead of in the caller...see
below.

[lpy, 2016/08/09 01:01:47]

Done.

         *is_await = true;
+      } else if (this->IsEval(*name)) {

[adamk, 2016/08/05 22:29:29]

What about "arguments"?

Is there a reason check has to be done here, and not in the caller? It seems
like we handle most of the other similar errors in ParsePropertyDefinition, in
the block that handles the PropertyDefinition production.

[lpy, 2016/08/09 01:01:47]

My understanding is when parsing the property name, when a specific word
encountered, it will record (for example, set is_await).

But since `name` is an out parameter, I think we can do it in caller.

Not sure if we should do this in a separate CL for await, I move the error
report for eval and arguments to caller as suggested.

+        classifier->RecordStrictModeFormalParameterError(

[adamk, 2016/08/05 22:29:29]

Not clear on why this is an error in sloppy mode.

Can you add the test-parsing.cc tests we discussed? That'd let us more easily
verify that we get the right behavior in strict vs sloppy.

+            scanner()->location(), MessageTemplate::kStrictEvalArguments);
+        if (is_strict(language_mode())) {
+          classifier->RecordBindingPatternError(
+              scanner()->location(), MessageTemplate::kStrictEvalArguments);
+        }
       }
       break;
   }