Heh.

I asked the question partially in jest, but also to highlight and get some discussion going on the differences between what's out there and of course what /should/ be out there.  It's so easy to make code ipv4 specific without necessarily realising it. How many people have used gethostbyname in their code, for example ?

And agreed, it's a good start to see companies thinking about this sort of thing.


On Thu, Feb 12, 2009 at 2:52 PM, Dean Pemberton <nznog@deanpemberton.com> wrote:

Not entirely.

They SHOULD be agnostic, but programmers have been assuming that IP == IPv4 for so long that anything can happen when you try and put in addresses with :s in them.
As Skeeve points out, it's great to see that people are thinking about this and adding "IPv6 compliance" to regression tests.

Dean