Hello Paul, On Thu, 2021-04-15 at 23:15 +0000, Paul Hoffman wrote: > Everyone (but particularly resolver developers): please review the document > carefully, particularly the algorithm in Section 3, to see if it matches what > you expect.
PowerDNS implementation notes: * we combined step 5 and step 6d (and call it step 5); we fall back to 'no minimisation' on anything that's not a NoError response (but the fallback might be handled by the RFC8020 code, especially if that non-NoError response was NXDOMAIN) (the default for RFC8020 is 'on signed domains only') * we use 1-1-1-3-3-.. label steps, which is not exactly what section 2.3 describes (you can find some more details at https://doc.powerdns.com/recursor/appendices/internals.html#qname-minimization ) Kind regards, -- Peter van Dijk PowerDNS.COM BV - https://www.powerdns.com/ _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
