4 ms·

I often wish that there had been a way to politically make 240/4 and 0/8 commonly available. But in part I drove those projects to annoy the IPv6 crowd into act

• zamadatix 3y ago
  I think the IPv6 crowd is already decently in action, hence vast swaths of the internet and mobile network already migrated (some on the mobile even without dual stack). What's needed for more IPv6 adoption is continued annoyance felt by the IPv4 crowd to want to bother migrating to any other long term solution. That's required a lot more time than most probably expected but it has been happening, IPv4 has become more and more difficult to use. Throwing in things like "Look IPv4 crowd, there are ways to extend IPv4 for some more years again without the pain of going to multi-layered NAT" is about the exact opposite of the kind of call to action needed. Either way though, the inherent value in the technology is what has to drive the adoption and the value is already there it's just a matter of IPv4's value still holding well enough for now. As such I don't inherently mind 0/8 being available on Linux these days and wouldn't inherently mind 240/4 either but a completely separate set of "yes, you can configure that address but the problem is with getting between there and here not accepting it" is a bit disappointing for just another short term response to the problem. Unrelated but kickass work on packet scheduling btw.
• p1mrx 3y ago
  I'm still a bit miffed about the Linux kernel change that broke my unit test expecting connect(0.0.0.1) to fail fast.