BUG: Using transport reuse results in an extra listener on top of the original listener instead of replacing it

Nobody has claimed this yet.

When using quic reuse to use your own udp listener and transport to intercept raw quic connections, libp2p still creates its own quic transport and doesn't replace it with the one from reuse. This results in having either errors with quic port reported as already being used by the original transport, or it results in having two separate quic addresses with different ports... And that's not what we are trying to achieve with reuse I assume. At least for our use case which inspired the whole quic reuse thing, we should end up with a single quic address that is intercepting raw connections and giving non-raw ones back to libp2p.

You'd think that in order to result in one quic transport, you'd skip giving transport option to the new libp2p instance, but that causes the transport from reuse to not be used and we end up with no quic listening addresses.

Start by reproducing the issue with quic reuse, a custom UDP listener, and a transport, then trace the libp2p transport setup described in the report. Done means the reused transport replaces the original, producing one QUIC listening address that intercepts raw connections while returning non-raw connections to libp2p.

Written by the indexing model from the issue text.