Detecting LLM-Generated Text in Computing Education: A Comparative Study for ChatGPT Cases