ParserBase should accept ESCAPED_STRICT_RESERVED_WORD as an identifier

test/cctest/test-parsing.cc

Index: test/cctest/test-parsing.cc
diff --git a/test/cctest/test-parsing.cc b/test/cctest/test-parsing.cc
index 287b61b02c27b9fbbc206fff4b74f5d6b2dce74a..59ee11896f3334f1ece84a62b6558bf1a9247a52 100644
--- a/test/cctest/test-parsing.cc
+++ b/test/cctest/test-parsing.cc
@@ -9449,3 +9449,24 @@ TEST(NoPessimisticContextAllocation) {
     }
   }
 }
+
+TEST(EscapedStrictReservedWord) {
+  // Test that identifiers which are both escaped and only reserved in the
+  // strict mode are accepted in non-strict mode.
+  v8::V8::Initialize();
+  v8::HandleScope scope(CcTest::isolate());
+  v8::Context::Scope context_scope(v8::Context::New(CcTest::isolate()));
+  const char* scripts[] = {
+      "if (true) l\u0065t: ;",           "function l\u0065t() { }",
+      "(function l\u0065t() { })",       "async function l\u0065t() { }",
+      "(async function l\u0065t() { })", "async l\u0065t => 42"};
+  for (const char* script : scripts) {
+    v8::TryCatch try_catch(CcTest::isolate());
+    v8_compile(script);

[vabr (Chromium), 2017/02/14 23:58:53]

I'm not at all sure that I'm doing the right thing here.

If I place some obvious failure to |scripts|, e.g., "var var -", the test fails
by reaching the CHECK below and printing the message.

However, it passes for all |scripts| above, despite d8 showing errors:
d8> function l\u0065t() { }
(d8):1: SyntaxError: Keyword must not contain escaped characters
function l\u0065t() { }
         ^^^^^^^^
SyntaxError: Keyword must not contain escaped characters


So while my fix is obviously not complete yet, I'm not even sure how to test it
properly. Any advice?

+    if (try_catch.HasCaught()) {
+      v8::String::Utf8Value exception(try_catch.Exception());
+      v8::base::OS::Print("Compiling %s failed with %s.", script, *exception);
+      CHECK(false);
+    }
+  }
+}