Missing path in scope['server'], if run with UNIX socket

[nggit]

Not sure why this doesn't appear to be supported, although the underlying socket.getsockname() supports it. Perhaps because this method is too bizarre to handle?

Steps to reproduce:

1. uvicorn --uds /tmp/uvicorn.sock scopedump:app
2. On another window, curl --unix-socket /tmp/uvicorn.sock http://localhost/

It shows None instead of ('/tmp/uvicorn.sock', None).

The spec:

server (Iterable[Unicode string, Optional[int]]) – Either a two-item iterable of [host, port], where host is the listening address for this server, and port is the integer listening port, or [path, None] where path is that of the unix socket. Optional; if missing defaults to None.

https://asgi.readthedocs.io/en/stable/specs/www.html#http-connection-scope

Example code:

# scopedump.py

async def app(scope, receive, send):
    assert scope['type'] == 'http'

    await send({
        'type': 'http.response.start',
        'status': 200,
    })
    await send({
        'type': 'http.response.body',
        'body': str(scope.get('server', None)).encode(),
    })

[Kludex]

Was that always in the specification? I didn't see it before and I don't recall people complaining.

[nggit]

It was introduced here django/asgiref#216 .

[Kludex]

PR is welcome to add it.

[Kludex]

Can you check #2561 ? @nggit