inetaf/netaddr

Audit Go CLA compliance

bradfitz opened this issue ยท 15 comments

We'd like to start exploring getting netaddr into the Go standard library.

As part of that, I started looking whose code might not be able to get upstream into Go.

Almost all the code is authored by Go contributors, but there are a few people who would need to submit the Google CLA if they want their code included: (perhaps some of these 6 are also Go contributors, but it's not easy for me to check)

Less of a concern:

  • @nwt for e5237cd but perhaps not needed to be included upstream, as error handling would likely change to conform to Go style anyway(?)
  • @jawnsy for 21243a5 (worst case we could remove those test cases for upstream)
  • @moreati for d57edf1 (but worst case we remove those comments)

Not a concern:

I think @terinjokes is fine, due to working at CloudFlare and "CloudFlare Inc." is in https://golang.org/AUTHORS
...

You beat me replying. :)

So just @majst01 remains, but worst case the Go standard library doesn't necessarily need (*IPSet).RemoveFreePrefix, at least in the first round.

I might have signed it for Kubernetes-related contributions when I was at Red Hat some years ago, but I have also just signed it as an individual. Please let me know what else I need to do, if anything!

I just signed the individual CLA

nwt commented

I've signed the individual CLA.

Hi,
i signed the google/cadvisor CLA back in 2018, google/cadvisor#1780, don't know if this counts. But i am happy to sign the CLA for go as well, any pointers where this can be done ?

x-way commented

FYI: I've signed a CLA back in 2018 (for contributions to google/gopacket), happy to sign a new one if needed.

@majst01, @x-way, thank you! We're all good.

Is there a proposal open for upstreaming netaddr?

Nope. Brad, Dave, and I had a live chat yesterday about some of the outstanding API questions and inconsistencies. Dave is doing a bit of legwork to decide whether our tentative decisions make sense.

(IIRC, those decisions were roughly: Only support zones for IP and IPPort. For all other types: If the method/func return errors, return an error when zones are present. If the method/func doesn't return errors, silently strip input zones. When comparing against IPs/IPPorts with zones, e.g. IPPrefix.Contains, if the IP/IPPort has a zone, fail closed. Add an error return to IPSetBuilder.IPSet, but not other methods.)

Once we're all (including you) happy with all the API edges, then we'll put together a proposal for upstreaming. Hopefully very, very soon.

And for the record, @AlexanderYastrebov's commit 900fa71 (from #92) has a typo in the email address. yastebov.alex@gmail.com is missing an r. The typo exists in the original git commit.

The correct email address is listed at https://github.com/AlexanderYastrebov

@bradfitz Yes, correct, sorry for the confusion.