Re: Cleaning up historical portability baggage

On Mon, Aug 15, 2022 at 5:53 PM Thomas Munro <thomas(dot)munro(at)gmail(dot)com> wrote:
> Remove configure probe for IPv6.
> Remove dead ifaddrs.c fallback code.
> Remove configure probe for net/if.h.
> Fix macro problem with gai_strerror on Windows.
> Remove configure probe for netinet/tcp.h.
> mstcpip.h is not missing on MinGW.

I pushed these except one, plus one more about <sys/sockio.h> which
turned out to be not needed after a bit of archeology.

Here's a slightly better AF_INET6 one. I'm planning to push it
tomorrow if there are no objections. It does something a little more
aggressive than the preceding stuff, because SUSv3 says that IPv6 is
an "option". I don't see that as an issue: it also says that various
other ubiquitous stuff we're using is optional. Of course, it would
be absurd for a new socket implementation to appear today that can't
talk to a decent chunk of the internet, and all we require here is the
headers. That optionality was relevant for the transition period a
couple of decades ago.