While previous version worked, it didn't behave nicely during concurrent requests to the same peer due to lack of awareness of pending requests, resulting in this:

2024-10-17T16:55:57.396088Z WARN subspace_networking::node_runner: Failed to dial disconnected peer on generic request error=Dial error: dial condition was configured to only happen when both disconnected (PeerCondition::Disconnected) and there is currently no ongoing dialing attempt (PeerCondition::NotDialing), but node is already connected or dial is in progress, thus cancelling new dial.

While I could fix that with more hacks, I decided to simply patch libp2p while we're discussing it in libp2p/rust-libp2p#5634, here is the patch (subspace-v8 branch): autonomys/rust-libp2p@ae75274

Patch adds addresses argument to send_request method, fixing this issue.

Code contributor checklist:

• I have read, understood and followed contributing guide