feat(identify): implement signedPeerRecord

[drHuangMHT]

Description

May close #4017.

Notes & open questions

pb-rs now uses Cow<'_,T> for the compiled Rust structs. But with borrowed type in the struct, FramedRead can no longer process frames correctly(trait bound not statisfied).

Change checklist

• I have performed a self-review of my own code
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• A changelog entry has been made in the appropriate crates

[drHuangMHT]

Oh CI will compile the protobuf code, didn't see that.
EDIT: Wait how does this work?

[dariusc93]

Left some comments :)

[getong]

you might rebase to main branch, as zlib license has added to ci， #5769

[getong]

Files in /home/runner/work/rust-libp2p/rust-libp2p/protocols/identify have changed, please write a changelog entry in /home/runner/work/rust-libp2p/rust-libp2p/protocols/identify/CHANGELOG.md

I think you might write some changelog text into this file, just like,

implement signedPeerRecord

the pr might be okay now.

[drHuangMHT]

the pr might be okay now.

Thanks for the review! Working on tests right now.

[drHuangMHT]

version 0.45.1 of libp2p-identify hasn't been released on crates.io yet, I'm not sure whether I should add a changelog entry right now.

[elenaf9]

version 0.45.1 of libp2p-identify hasn't been released on crates.io yet, I'm not sure whether I should add a changelog entry right now.

CHANGELOG will be fixed with #5803, you can just add the new entry to 0.46.0.

[elenaf9]

Thank you @drHuangMHT! Looks like a great start.

I found a pending libp2p spec for signed peer records: libp2p/specs#630. Most of it matches this PR, just one thing is missing:

If the signedPeerRecord is present the implementation MUST use the data contained within it and ignore duplicated fields present in the main identify message

IMO that makes sense and should be added to this PR. So instead of copying the received record as it is in handle_incoming_info, I think we should instead (when an signed envelope is present):

1. Verify that the signature matches the remote's public key
2. Deserialize the record, store the resulting addresses as listen_addrs, and ignore the original listen_addrs.

Wdyt?

[elenaf9]

LGTM. Thanks @drHuangMHT for the follow-ups!

I'd still like to wait for a review from @jxs before merging this.

[jxs]

Hi Dr Huang, overall looks good to me, thanks fr your work on this!
let some comments

[jxs]

LGTM! Thanks Dr Huang 🚀