Hi Patrick, On Tue, 25 Apr 2023 18:41:26 -0500 Patrick Mevzek wrote:
> Two (+1) quick points on the content for draft version 4. > > Section 8.1 uses twice the same XML namespace for both cases. I suspect > `:schema:` is missing in the second one. Thanks, this is fixed. > As for TTL values: > <restriction base="nonNegativeInteger"> > <minInclusive value="1"/> > <maxInclusive value="4294967295"/> > </restriction> > > The highest bound should be instead 2147483647 seconds. Also fixed. > Alternatively, personally, I would prefer the value to be given as XML type > of "duration". > This allows to still pass it as seconds for those that prefer that, but also > allows to pass it in more human friendly formats such as number of minutes, > hours, days, etc. > It does make however stating the maximum value possible more complicated I > guess, so maybe not worth doing? It looks like duration types support the minInclusive and maxInclusive attributes, but there is an issue that a duration such as "P1M" can be between 2,419,200 seconds and 2,678,400 seconds, depending on when the duration is converted into an integer. Durations are always relative to a particular date and time and don't represent absolute values. Using an non-negative integer seems safer to me. > Also, not sure if a single TTL value for everything is enough. > Are we sure that all registries will be fine using the same TTL for > both NS/A/AAAA and DS? > If I take `.com` right now, NS has 2 days of TTL, where DS only one day. This is one of the open issues that I would welcome WG input on. Should we allow multiple <ttl:secs/> elements, with an attribute containing a list of DNS record types to disambiguate? <ttl:infData> <ttl:secs for="NS">172800</ttl:secs> <ttl:secs for="DS">86400</ttl:secs> <ttl:secs for="A,AAAA">3600</ttl:secs> </ttl:infData> It's not inconceivable that some future extension to the DNS will allow additional records to be published in parent zones for child zones, so can we hard-code the list of record types, or allow any type? G. _______________________________________________ regext mailing list regext@ietf.org https://www.ietf.org/mailman/listinfo/regext