Thanks. What you describe in the document certainly makes sense, but I think is sort of orthogonal to this. If I were to state the fundamental problem I am trying to address here, it is simply that the RFCs do not authoritatively say what QDCOUNT > 1 means, nor how to do it, nor do they authoritatively say that it is invalid. As a consequence of this, an implementation started using QDCOUNT > 1. So we should do something about that.
I think what you propose in your draft is exactly what should have been done, but since we don't say "don't send multiple questions" anywhere, it wasn't done that way. :) On Mon, Feb 13, 2023 at 9:46 AM Kazunori Fujiwara <fujiw...@wide.ad.jp> wrote: > > From: Ted Lemon <mel...@fugue.com> > > Ohta-san, thanks for pointing out that text about the motivation for MX > in > > RFC1035―I hadn't noticed that. > > "DNS SRV RR" [RFC 2782] also mentions adding SRV RR Target A/AAAA to > the additional section. > > # draft-ietf-dnsop-svcb-https also mentions adding SVCB/HTTPS target A/AAAA > # to the additional section. > > In 2017, 2018, QDCOUNT>1 or multiple qtypes were discussed in dnsop WG. > At the time, I proposed draft-fujiwara-dnsop-additional-answers-01 > and compared 5 proposals. > > At the time, all proposals avoided QDCOUNT>1. > > -- > Kazunori Fujiwara ( = fujiw...@jprs.co.jp ) >
_______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
