Just to add my ยข2 on this point: In the last version of the draft that I wrote (and the one used for most implementations), there was no MUST/SHOULD terminology, so it was a little vague.
Reason we've added the truncation at the time is to save space. This is DNS, where saving bytes is sometimes still considered important. With IPv4 the saving is ridiculously low, but with IPv6 with truncation you can easily cut off 8 \0-bytes. The original wording was kind of intended as a MUST, for example in case of later versions of the draft having additional bits of info after the address. But likely for a new version a new option code would have to be used anyway so that's not really important. I have no strong feelings about this. Slight preference for MUST, but nothing's going to fall apart if this becomes a SHOULD. I've just checked the parser in our codebase, it will just ignore extra bytes at the end. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
