On 1/15/26 2:14 AM, Pawel Kowalik wrote: > Hi Andy > > On 14.01.26 19:33, Andy Newton wrote: >> >> >> On 14-01-2026 12:04 PM, Pawel Kowalik wrote: >>> The clients MUST however accept responses containing other DNS record >>> types. >> >> IMO, that just means clients need to be aware of all RR types. There >> are, by my count, 47 different RR types. Requiring clients to support >> all 47 when they are only likely to encounter 3 seems a bit much. >> Would you be satisfied if this were a SHOULD? > > The intention was to use "accept" instead of "process" or alike to > indicate that the client shall at least be able to ignore and not break > in the presence of such values. > It seems that I failed to find right formulation, so how about more > explicit version: "The clients MUST however accept responses containing > other DNS record types at minimum by ignoring their presence and not > breaking processing".
That works. Thanks, Pawel. -andy _______________________________________________ regext mailing list -- [email protected] To unsubscribe send an email to [email protected]
