Ray Bellis <r...@bellis.me.uk> wrote: > > Would it be feasible to reserve a standard RCODE value in the header > that just means "see extended error"?
That would require client-to-server signalling, and the server would be unable to simply unconditionally include the extended error in the OPT record. > It has always kinda surprised me that the EDNS RCODE didn't work that > way, instead of the current situation where if you only read the bottom > 4 bits of the extended 12-bit code you could completely misinterpret the > status (e.g. treat BADVER[16] as NOERROR[0], since the bottom four bits > are all zeros for both). There can only be an EDNS RCODE if the client sent an EDNS OPT; if the client forgot it sent the OPT then it's too broken to worry about. Tony. -- f.anthony.n.finch <d...@dotat.at> http://dotat.at/ - I xn--zr8h punycode South Utsire: Northwesterly, backing southerly for a time, 5 to 7, decreasing 4 for a time. Moderate or rough. Rain for a time. Good, occasionally poor. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop