L. Kärkkäinen

Was something changed? I recall myself implementing support for full URL SERVER_NAME a few years back, although IIRC it wouldn't (yet) support path but at least you could say https...

Did you check how this affects ASGI, or how are connections lost with that anyway? The catching/handling logic in Sanic is extremely complicated but I see that you did a...

@ahopkins Do you mean for https vs. wss? Other than that I cannot really think of reasons to leave the scheme open. In particular, nothing should have external URLs without...

You are right, the create server mode should respect the existing loop. Sanic does not recommend the use of that mode because the implementation cannot quite accomplish that. In general,...

~~Still preferring a solution that only uses `__aiter__`.~~ Don't care which way it is done.

Thanks for the PR. This is a very useful feature and it helps to have parity with the websockets module. Receive on a closed socket should probably follow that as...

Really need to reimplement the server startup so that it can work without forking. IIRC, it involves two main milestones: - Restructure things so that you can start new processes...

If that can be implemented, I don't think we need fork at all anymore. Any performance advantages of forking vs. separate processes are miniscule because we only run a few...

REUSEPORT/REUSEADDR absolutely cannot be used for this. On Mac only one process will then receive all requests and the other only acts as fallback, and I believe while it is...

I tried it a couple of years back, specifically to fix this problem in Sanic. Found out that this was the case and then looked it up on Stackoverflow or...