llama.cpp

6f693f14 - `server`: use (new) Request::is_alive as set_content_provider called after status / headers sent