At 7:47 AM +0200 6/24/10, Patrik Fältström wrote:
>Sure, but, support for unknown RR Types is said to be needed since long time 
>back. And what API do not handle the ability to request an RR with a specific 
>RRTYPE?
>
>int res_query(const char *dname, int class, int type, u_char *answer, int 
>anslen);
>
>Anyway...this discussion has been held in the IETF I do not know how many 
>times. Instead of writing another 10 lines of code (or whatever is needed) 
>people fall back to existing RR Types, and not only that, define future 
>protocols because of lack of #define for new RRTYPES.
>
>I know people have different views here, and I have one specific view ;-)

As someone who normally has that "different view", I support a new RRTYPE in 
this case because the option of reusing SRV is not sufficient: it requires 
DNS-SRV-followed-by-HTTP. I think a new RRTYPE that keeps the DNS lookup 
entirely in the DNS protocol far outweighs reusing SRV but requiring HTTP on 
both sides.

--Paul Hoffman, Director
--VPN Consortium
_______________________________________________
Ietf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ietf

Reply via email to