any IPv6 network should stringify as ::/0

[abraxxa]

currently ::0/0 which doesn't conform to the IPv6 standard.

[oschwald]

I agree this is weird, but it is not wrong. Both representations are valid. It looks like this was intentionally introduced in this commit.

[abraxxa]

@autarch can you please comment if there is a reason to stringify like this?

[autarch]

I really can't remember why this was changed. The commit that did this is from nearly seven years ago.

[abraxxa]

I finally found the RfC I was looking for: https://tools.ietf.org/html/rfc5952
Section 4.2 describes the shortening rules but doesn't explicitly mention this case.
All operating systems I've checked show :::* for sockets listening for any IPv6 address (the third : is the delimiter between address and port, really a great choice by the IPv6 designers to use the same as the delimiter between the fields).
Are there any objections against changing the stringification to ::/0?

[oschwald]

I am not opposed to a PR that does this and the RFC on the recommended formatting is a compelling argument. I would suspect that it was done as there was some variance between platforms on what inet_ntop would return.