Absolutely the best outcome. BSDs has always been ( to me at least ) about getting things right, take time to get it baked before committing. The old, out of fashion style of getting things done properly and not shipping for the sake of it. Which is both a good thing and a bad thing in the modern world. But it is a trade off.
I also hope FreeBSD sort of look into why it was committed in the first place.
Hopefully this also set the tone for Scott and Netgate. Even having mild scepticism of the one sided bash to them on HN had me downvoted into oblivion.
> I also hope FreeBSD sort of look into why it was committed in the first place.
It is a widely desired feature that no one else was working on and unfortunately, did not see adequate review. FreeBSD largely trusts committers to seek out review and submit quality code. There are certainly downsides to this model, sometimes more impactful than others. Unfortunately, FreeBSD does not have the large body of employer-paid maintainers Linux has available to thoroughly code review every patch that lands.
WireGuard is gone from the kernel in 13.0-RELEASE. Given the choice between "buggy" and "less than a week old", we're going with the third option of "you can ship a kernel module via the ports tree".
They are removing both implementations(the new and the broken one) in order to put more work and review on the new one, and release it properly at a later time.
