Use now apparently maintained if-addrs

[eskimor]

instead of unmaintained get-if-addrs.

[eskimor]

This PR depends on a new release of if-addrs.

[eskimor]

As far as I can see, all the build errors are already fixed on master of if-addrs, just waiting to be released.

[bltavares]

@eskimor I'm looking at the crate, and it does not seem to have any meaningful improvements over the fork. The last commits so far have been only renames of the crate from a few weeks ago.

I don't see any reason to change over to that dependency given it has no code change so far. I'd rather stay on a stable release until fixes/improvements appear on the fork.

[eskimor]

Fair enough :-) I was planning on integrating your reverse patch there, but I don't see a problem waiting until that actually happens ;-) One minor thing though worth mentioning: The fork already bumped the winapi dependency, so that could be useful already.

[bltavares]

Oh, adding the reverse there would be great! It seems to have some new activity there. I would consider moving to the fork to remove the hotfix on mips 👍

[mon]

Bit of a necro, but I need support for link-local addresses, which the new if-addrs exposes as a feature. Can this PR be resurrected?

[mon]

There's actually some more code required as I have 3 link-local addresses on my machine, and two of them are actually disconnected and fail with ErrorKind::AddrNotAvailable. Worth just making my own PR with the new if_addrs and the other required handling for this case?

[bltavares]

Sure thing, I'd consider replacing the lib if the MIPS support have been merged, which was the blocker last time.

[mon]

Truthfully, I can't tell if they fixed the MIPS issue yet - do you have a MIPS system to test it on? I've never touched that arch.

[bltavares]

Yes, I do have some mips routers to test, and the CI should check for compilation at least.

[mon]

I looked a little more intently - the old MIPS bug is caused by a manual sockaddr parse:
https://github.com/maidsafe-archive/get_if_addrs/blob/master/src/sockaddr.rs#L37-L43

The new one uses to_ne_bytes, which will be big-endian aware:
https://github.com/messense/if-addrs/blob/master/src/sockaddr.rs#L42

I'll start work on the re-port :)

[bltavares]

Thanks for the interest @mon :) Let me know how it goes, and I can test on a device later

[bltavares]

Closed by #19