Emitted during replay before each logical step. Arguments: step name and content (e.g. raw text being parsed; empty string when none). Listeners can show content and wait for user (e.g. Enter) before returning.