On Tue, 27 Nov 2007 08:03:35 +0100, Frank Ellermann said:
I'd say that in *practical* terms you're just *looking* for
trouble if you publish an e-mail address that doesnt have a
'.' anywhere in it.
BTW, an interpretation of RFC 2142 about abuse@ addresses for
any given FQDN takes the zone cut adding abuse(_at_)(_dot_) For the IDN
test domains that ends up with no-dot abuse@ addresses, doesn't
it ? Should that go into an erratum for RFC 2142 ? Of course
RFC 2142 couldn't foresee what RFC 2821 says, but it's still
an interesting detail for folks trying to implement RFC 2142,
or even trying to promote it - the abuse@ part is important,
the other mailbox names in RFC 2142 are less convincing.
The example test names:
Picking the Hebrew one at random:
http://xn--fdbk5d8ap9b8a8d.xn--deba0ad/ is the actual target.
I see a dot between the 8d and the xn--, and my Firefox even renders it
with a '.' (see attached .png). The other languages *also* both have
a '.' between the two parts of the URL *and* render a dot in the middle
of the non-ascii string.
Description: PGP signature