Thanks, Justin! I'll look over your PR, try to replicate the failure locally, and test it against the patch to see if it still occurs. I think your analysis of the issue makes sense, with it likely being a race condition between the test prototype "MyProto" and the event loop.