chore(deps): bump socket2 from 0.5.10 to 0.6.4

[dependabot]

Bumps socket2 from 0.5.10 to 0.6.4.

Changelog

Sourced from socket2's changelog.

0.6.4

• Implement Send for MsgHdr(Mut) (rust-lang/socket2#650).
• Add support for horizonOS/n3ds (rust-lang/socket2#653).
• Switch QNX 8.0 and 7.1 to use TCP_KEEPIDLE instead of TCP_KEEPALIVE (rust-lang/socket2#654).
• Fixed an issue on Windows where for Socket::only_v6 sometimes booleans are 1 byte other times it's 4 bytes (rust-lang/socket2#655).

0.6.3

• Added support for wasm32-wasip2.
• Added Socket::(set_)ip_transparent_v6.
• Added Socket::set_tcp_ack_frequency.
• Support windows-sys v0.61 in addition to v0.60.

POTENTIALLY BREAKING The MSRV of windows-sys v0.61 is 1.71. To use socket2 with its MSRV of 1.70, please downgrade windows-sys to v0.60.x. This can be done using: cargo update windows-sys --precise 0.60.2

0.6.2

• MsgHdr and MsgHdrMut are marked as transparent meaning both have the same layout as msghdr on Unix and WSAMSG on Windows (rust-lang/socket2#635).
• Don't set SO_NOSIGPIPE when accepting sockets, this is inherited from the parent socket (rust-lang/socket2#632).
• Fixes Socket::tcp_notsent_lowat by using the correct argument type (rust-lang/socket2#622).

0.6.1

Added

• Added support for Windows Registered I/O (RIO) (rust-lang/socket2#604).
• Added support for TCP_NOTSENT_LOWAT on Linux via Socket::(set_)tcp_notsent_lowat (rust-lang/socket2#611).
• Added support for SO_BUSY_POLL on Linux via Socket::set_busy_poll (rust-lang/socket2#607).
• SockFilter::new is now a const function (rust-lang/socket2#609).

Changed

• Updated the windows-sys dependency to version 0.60 (rust-lang/socket2#605).

... (truncated)

Commits

• See full diff in compare view

[dependabot]

Labels

The following labels could not be found: dependencies, rust. Please create them before Dependabot can add them to a pull request.

Please fix the above issues or remove invalid values from dependabot.yml.

[github-actions]

Benchmark Results

Performance Comparison

Benchmark	Baseline	Current	Change	Status

Summary

Configuration

• Regression threshold: >10% slower
• Improvement threshold: >10% faster
• Measurements: Mean execution time

[jacderida]

Pre-merge assessment

Assessed with Claude Code before merging. All the bumps in this batch were applied together on top of current main (651a2dad) in a scratch worktree, rather than relying on the CI runs recorded here (some of which are from a base ~29 commits stale). On that combined tree:

• cargo check --all-targets — clean
• cargo clippy --all-targets -- -D warnings — clean
• cargo test --lib — 1498 passed, 0 failed

Specific to this PR:

socket2 0.6 is largely a renaming release: set_ttl → set_ttl_v4, TCP options gaining a tcp_ prefix, SockAddr now wrapping a SockAddrStorage. None of those are on paths we use. Our entire socket2 surface is Socket::new, bind, set_send_buffer_size/set_recv_buffer_size, set_only_v6, set_reuse_address, set_reuse_port, set_nonblocking and SockRef::from (in src/config/port_binding.rs and src/high_level/endpoint.rs) — all unchanged in 0.6.

Bonus: this deduplicates socket2 to a single copy in the tree.

Risk: low.

The failing linear-link / pr-template checks are the release-process gate from pr-checks.yml, not a code signal; merging with admin override.

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

@dependabot rebase

[jacderida]

Rebased onto current main — the conflict was Cargo.lock only (churn from the other dependency PRs merging ahead of this one). Dependabot's commit is preserved as-is; the lockfile was regenerated with cargo update -p socket2@0.5.10, which drops the stale 0.5.10 entry and leaves a single socket2 0.6.4 in the tree. cargo metadata --locked is clean. Assessment above is unchanged.

[dependabot]

Looks like this PR has been edited by someone other than Dependabot. That means Dependabot can't rebase it - sorry!

If you're happy for Dependabot to recreate it from scratch, overwriting any edits, you can request @dependabot recreate.