Another person commented the same, but a "project loom" virtual threads based HTTP server implementation would be very desirable right now. A new NIO based server probably doesn't add a whole lot of value to the ecosystem. But note, I'm not discounting the effort here.
I do like the fact that this project attempts a no-dependency implementation. That's nice. And of course, there's always room for one more http server implementation, if it offers some alternative to existing solutions (throughput, api, processing model, etc.). But any "new" http server projects wanting to start from scratch today would be wise to base its implementation on virtual threads (in my opinion).
I do like the fact that this project attempts a no-dependency implementation. That's nice. And of course, there's always room for one more http server implementation, if it offers some alternative to existing solutions (throughput, api, processing model, etc.). But any "new" http server projects wanting to start from scratch today would be wise to base its implementation on virtual threads (in my opinion).