Max Inden

Thanks @jellllly420. Unfortunately I don't see additional insights based on the new logs. The connection is timing out. That can be due to many reasons, e.g. the host itself, but...

Friendly ping @jellllly420.

> At the moment, I don't think we support bootnodes as such but it would be fairly trivial to create a `NetworkBehaviour` whose only job it is to maintain connections...

Could you test with the patch below? ``` diff diff --git a/protocols/relay/examples/relay_v2.rs b/protocols/relay/examples/relay_v2.rs index 0e3fde7c..2794405b 100644 --- a/protocols/relay/examples/relay_v2.rs +++ b/protocols/relay/examples/relay_v2.rs @@ -57,8 +57,8 @@ fn main() -> Result { let...

Thanks for debugging this @demfabris! Just for the sake of completeness, this is what is happening: - With `OrTransport(relay_transport, dns_tcp_transport)` (The error case) 1. `OrTransport` forwards the dial to `dns_transport`...

> Why can’t the OrTransport try with the second option after having seen the error from the first? Each `Transport::dial` implementation can be thought of as two phases. 1. Synchronous...

> Also, since we are proposing a fix specifically to `relay` module, how confident are you about other possible problematic transport combinations? @demfabris I am not aware of any other...

> Your suggested diff would cause it to be required to have an existing transport to construct a `v2::Client`, even tho I don't see the use case for a single...

> I totally agree, I'm asking this because one of the tests is implemented with `v2::Client` alone and it got me wondering Oh, mind pointing me to it @demfabris?

@fishseabowl you are correct, though I am not sure how to act on your report. - Are you suggesting to move away from `prost`? - Do you want this to...