Brandon Cheng

I PR'ed against neli but the library hasn't released an update yet. I'd like to remove `libc::IFLA_LINKINFO` in favor of `neli::consts::IFLA_LINKINFO`.

I've had a bit of difficulty deciding whether the types returned from this library should be easy to use or reflective of the raw Netlink responses. Although a map is...

Hey @xtexChooser, what happens on `rc/peer_conf/mod.rs:48:62`? Is that line adding a new device?

## Mismatched Paths It looks like this command applies `CAP_NET_ADMIN` to all files under `./target/debug/deps` ``` ➜ peerd git:(main) ✗ find ./target/debug/deps -maxdepth 1 -type f -executable | xargs -n...

It looks like the test binaries haven't been built by the time `setcap` is called in the output above. Could I have you try rerunning the `find ./target/debug/deps -maxdepth 1...

Apologies that it took me a bit to understand your bug report. I see that we get the `ENODEV` error. ``` 5\x00\x01\x00\x00\x00\x24\x00\x03\x00\x50\x52\x53\xa9"...], 248, 0, NULL, 0) = 248 recvfrom(3, [{nlmsg_len=268,...

@xtexChooser Does this block you from your work? I think improving the error message requires upgrading neli, which is something in progress.

The test setup creates and deletes WireGuard devices through `NETLINK_ROUTE`. As you've noticed `set_device` fails on `No ack received` when the device doesn't exist yet, which happens as tests running...

Yeah, the error is very confusing. 😞 Adding a test to make sure this gets revisited. https://github.com/gluxon/wireguard-uapi-rs/commit/33d2db3d499977c7d9868ff35a2c367f15d6b619

That's a good question. Unfortunately not without significant caveats to performance and usability. The main blockers are: ## 1. Other Tooling **The TypeScript output directory isn't looked at exclusively by...