Erik Auerswald <auers...@unix-ag.uni-kl.de> writes: > On Mon, May 12, 2025 at 09:13:14AM +0200, Simon Josefsson via Bug reports for > the GNU Internet utilities wrote: >> Collin Funk <collin.fu...@gmail.com> writes: >> > [...] >> > I wouldn't mind having 'ping' support both IPv4 and IPv6 with arguments >> > to chose though. I think that was even but in the TODO file long ago. >> > But I am in no rush to do so. >> [...] >> I disagree with Anna and think that IPv4 should still be the default, >> although I think we should be open to revisit this in a few years if >> IPv4 availability lowers (which I'm not seeing any signs of). > > The iputils ping program (the default ping on Debian GNU/Linux and derived > distributions) combines IPv4 and IPv6 support in a single binary for > many years now. As commonly expected from dual-stack applications, > it prefers IPv6 if both IPv4 and IPv6 addresses are returned by > getaddrinfo(). Another free software dual-stack ping program would be > fping, also preferring IPv6, also using getaddrinfo(). The dual-stack > applications that are currently part of GNU Inetutils also prefer IPv6. > The proprietary dual-stack ping implementation on Windows also prefers > IPv6. It seems to me that diverging from this common behavior in a > dual-stack ping implementation would be rather surprising. > > As such I'd prefer a combined ping+ping6 implementation, i.e., a possible > future dual-stack GNU Inetutils ping, to prefer IPv6.
Hmm, you make a good case for this, and I'm changing my position to neutral. What I think is more important is to align ping and ping6 more to reduce code duplication. If that happens to lead to using normal getaddrinfo() selection defaults for IPv4 vs IPv6, that is probably a good thing for system-wide consistency. /Simon
signature.asc
Description: PGP signature
