Add set_min_ttl_v4 and set_min_hopcount_v6

[taspelund]

Adds linux/illumos Socket methods for setting IP_MINTTL and IPV6_MINHOPCOUNT along with associated unit tests.

[taspelund]

Tests pass on both platforms this new wrapper is being added for (linux and illumos):

# illumos
trey@korgano 07:06:08 PM | ~/git/socket2
‣ cargo test --all-features --test socket min_
    Updating crates.io index
     Locking 1 package to latest compatible version
    Updating libc v0.2.186 -> v0.2.189
   Compiling libc v0.2.189
   Compiling socket2 v0.6.5 (/home/trey/git/socket2)
    Finished `test` profile [unoptimized + debuginfo] target(s) in 2.66s
     Running tests/socket.rs (target/debug/deps/socket-1129f55d11df1d5a)

running 4 tests
test min_hopcount_v6 ... ok
test min_hopcount_v6_preserves_outbound_hops ... ok
test min_ttl_v4_preserves_outbound_ttl ... ok
test min_ttl_v4 ... ok

test result: ok. 4 passed; 0 failed; 0 ignored; 0 measured; 47 filtered out; finished in 0.00s


# ubuntu 24.04
treyaspelund@lima-ubuntu-24-04:~/git/socket2$ cargo test --all-features --test socket min_
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.01s
     Running tests/socket.rs (target/debug/deps/socket-8ef69f38d614db53)

running 4 tests
test min_hopcount_v6_preserves_outbound_hops ... ok
test min_hopcount_v6 ... ok
test min_ttl_v4_preserves_outbound_ttl ... ok
test min_ttl_v4 ... ok

test result: ok. 4 passed; 0 failed; 0 ignored; 0 measured; 81 filtered out; finished in 0.00s

[taspelund]

The CI failure seems unrelated to my changes. Is there something I need to do to move past it?

[Thomasdezeeuw]

It seems that both Linux and illumos only accept a value in the range 0-255, maybe we should use an u8 for this option?

[Thomasdezeeuw]

The CI failure seems unrelated to my changes. Is there something I need to do to move past it?

You can ignore that CI failure it's unrelated to this pr.

[taspelund]

It seems that both Linux and illumos only accept a value in the range 0-255, maybe we should use an u8 for this option?

I went with a u32 because set_ttl_v4()/set_hoplimit_v6() accept a u32 despite their valid range being 0-255 as well. My thought was that consistency across methods would probably be more appreciated than making the signature more constrained. That said, I'm certainly open to making it a u8. What are your thoughts?

[Thomasdezeeuw]

That makes sense.

Thanks @taspelund