Problem: Unescaping of '\' characters didn't work when processing N'xxx'.
Fix: using get_text() instead of get_token() when scanning nationa strings.