!!!!
This can NOT currently be marked as UTF-8 encoded. It uses invalid UTF-8 sequences for testing. Tests will fail under 1.9.x+ if this file is marked as UTF-8 encoded. Tests will fail under 2.0.0 if this file is NOT marked as US-ASCII.