Re: [art] draft-ietf-dnsop-attrleaf

[Adam Roach <[e-mail]>]

On 8/9/17 6:36 PM, John R Levine wrote:
> See rfc3588 which has some _sctp names for NAPTR records. 


I presume you're referring to the example in Appendix B? I can see how 
it might look like that on first glance, but a careful parsing of what's 
there shows that the only reason it looks like that is because there's 
been some formatting damage. The literal text that's there is:

    ;;          order pref flags service           regexp  replacement
       IN NAPTR 50   50  "s"  "AAA+D2S"           ""
       _diameter._sctp.example.com IN NAPTR 100  50  "s"  "AAA+D2T"
       ""  _aaa._tcp.example.com

This is accompanied with an explanation that there are two NAPTR records 
present (one for AAA+D2S and one for AAA+D2T). I suppose you're reading 
the second line of this text to say "the host 
_diameter._sctp.example.com has an NAPTR record for 'AAA+D2T', with an 
empty regexp and a replacement of '_aaa._tcp.example.com"

Aside from that being kind of nonsense, I'll note that the *first* 
record (if this "_diameter._sctp.example.com" string is part of the 
second record) has no replacement value -- which clearly doesn't work.

The only way this makes sense is if you read the first record as:

   IN NAPTR 50   50  "s"  "AAA+D2S"  ""  _diameter._sctp.example.com

...and the second record as:

   IN NAPTR 100  50  "s"  "AAA+D2T"  ""  _aaa._tcp.example.com

...at which point it becomes clear that the underscores are only in the 
"replacement" field, and not used as input for NAPTR.

I'll file an erratum.

/a