On 22 Feb 2018, at 06:20, Petr Špaček <petr.spa...@nic.cz> wrote: > On 22.2.2018 11:38, Joe Abley wrote: > >> A few people now have mentioned that they like zero padding. What is >> it about zero padding or fixed-size labels that makes implementation >> easier than specifying no zero padding? > > It is important to note that this 'special label trigger' is not seen > anywhere else in DNS, so this is first piece of code which has to match > DNS labels in the 'hot path'. > > I think it is a good and easy optimization to minimize use of > regexes/string matching in the hot path, and simple condidion > if (label_len == X || label_len == Y) > before the heavy-weight pattern matching will reduce frequency of regex > use significantly.
You could use the simple condition (X < label_len < Y) in the same way without zero-padding and I suspect get a similar benefit. Really, so long as the format is standardised I don't think it matters much. It's not like we're going to see a significant number of experiments of this kind; more like a small flurry of Duane/Geoff action once every five years. [*] But I think mandatory zero-padding is ugly, and it does make me a little bit sad. :-) Joe [*] perhaps try not to dwell too long on the phrase "Duane/Geoff action"
Description: Message signed with OpenPGP
_______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop