On Sat, Jan 28, 2023 at 10:12:37AM -0800, Rob Sayre wrote:
> So far we've found Chrome and Postfix, just to name a few examples, using
> UTS-46. I think the draft is overstating the influence of IDNA2008. After
> all, it's 14-15 years old, and things have changed. But, as Orie notes,
> there are still bugs.
Note that Postfix isn't using UTS-46 because Wietse decided that it is
preferable to IDNA2008, rather Postfix uses UTS-46 because there's a
ubiquitous library that supports the UTS-46 superset of IDNA2008, we'd
equally use IDNA2008, if it came with a suitably ubiquitous library.
Finally, it is still unclear why any of this is an issue for this work.
A given U-label either is or isn't accepted by the application, which it
converts to A-labels using whatever library performs that job. The name
in the certificate is an A-label name.
The one significant departure from A-label DNS names in certificates is
SmtpUTF8Mailbox
https://datatracker.ietf.org/doc/html/rfc8398
which is outside the scope of this work:
https://www.ietf.org/archive/id/draft-ietf-uta-rfc6125bis-10.html#section-1.4.2
--
Viktor.
_______________________________________________
Uta mailing list
Uta@ietf.org
https://www.ietf.org/mailman/listinfo/uta
