Bump bip324 to v0.10

[luisschwab]

Bumps bip324 to v0.10.

Changelog

- Bump `bip324` from v0.7.0 to v0.10
  - `AsyncProtocol` is moved to `futures::Protocol`
  - `ReadTransport::V2` and `WriteTransport::V2` now own the stream
  - `read_and_decrypt` is renamed to `read`
  - `encrypt_and_write` is renamed to `write`
  - Shutting down the V2 transport becomes a no-op, as each message is immediately flushed to the buffer

[csgui]

Left a comment about the "absence" of shutdown in the new protocol API.

[Davidson-Souza]

Could you leave a brief changelog from v0.8 in the PR description?

[luisschwab]

@Davidson-Souza done!

[luisschwab]

We could move this to v0.9.0 as well.

[Davidson-Souza]

I wonder if those changes will allow us to remove the ping hack from functional tests for V2 🤔. Something to try in a follow-up

[luisschwab]

Ping hack?

[Davidson-Souza]

ACK 5d21626; Finished IBD on signet with only V2 peers and it works fine!

[Davidson-Souza]

Ping hack?

bip324 seems to take a while to figure out a peer disconnected. So we keep sending some pings until it realizes the other side died. I wonder if the more frequent flushing will fix this, but I doubt.

[luisschwab]

Worth opening an issue there to let them know.

[luisschwab]

Aaaand shitty GH CI strikes once more!

[Davidson-Souza]

re-ACK ddbe6a5

[JoseSK999]

This means the peer doesn't support P2Pv2 no?

ERROR floresta_wire::p2p_wire::transport: Failed to establish a P2PV2 connection over SOCKS5 proxy=127.0.0.1:9050 with peer=Ipv4(95.217.198.121): Io(Custom { kind: UnexpectedEof, error: "early eof" }, RetryV1)

[Davidson-Souza]

This means the peer doesn't support P2Pv2 no?

ERROR floresta_wire::p2p_wire::transport: Failed to establish a P2PV2 connection over SOCKS5 proxy=127.0.0.1:9050 with peer=Ipv4(95.217.198.121): Io(Custom { kind: UnexpectedEof, error: "early eof" }, RetryV1)

Or killed the connection for, eg., having too many connections already

[luisschwab]

Pushed 1ebc1c1 fixing a nit in the proxy logs.

[JoseSK999]

Then replace the second with -> and

[luisschwab]

I think this is better:

Established a P2PV1 connection over SOCKS5 using proxy={proxy_addr:?} with peer={target_addr:?}

[JoseSK999]

ACK f9085be

[Davidson-Souza]

ACK f9085be

[0xB10C]

Ping hack?

bip324 seems to take a while to figure out a peer disconnected. So we keep sending some pings until it realizes the other side died. I wonder if the more frequent flushing will fix this, but I doubt.

I think the read() side should be fixed with rust-bitcoin/bip324#160 (on master) but there's still something weird going on the write() side rust-bitcoin/bip324#164 (comment).

Just don't assume something you've written has actually reached the other side. Closing the connection on read() returning an error should be fine.