headers
Elwell, John | 3 Feb 2010 17:07

Re: Question on draft-ietf-sipping-v6-transition-07

Vijay,

A follow-up question on the statement:
"For this, IPv6 implementations MUST use a domain
       name within the .invalid DNS top-level domain instead of using
       the IPv6 unspecified address (i.e., ::)."
Presumably just "invalid" (alone, as opposed to "xxxx.invalid" would be legal?

John 

> -----Original Message-----
> From: Vijay K. Gurbani [mailto:vkg <at> alcatel-lucent.com] 
> Sent: 28 January 2010 20:10
> To: Elwell, John
> Cc: sipping <at> ietf.org
> Subject: Re: [Sipping] Question on draft-ietf-sipping-v6-transition-07
> 
> Elwell, John wrote:
> > [JRE] Thanks, Vijay. However, RFC 3264 specifies only 
> 0.0.0.0 for the
> > case where the address is not known in the initial offer (I am not
> > talking about the deprecated use for hold). It does not specify
> > .invalid, so I don't know what you mean by two alternative 
> solutions.
> 
> Correct; rfc3264 does not specify .invalid.  sipping-v6-transition
> is supposed to update rfc3264 to do so.
> 
> The two alternative solutions are supporting "::" and ".invalid";
> since at the time of writing of sipping-v6-transition, there
(Continue reading)

Permalink | Reply | 11 comments |
headers
Dale Worley | 4 Feb 2010 17:44
Favicon

Re: Question on draft-ietf-sipping-v6-transition-07

On Wed, 2010-02-03 at 17:07 +0100, Elwell, John wrote:
> A follow-up question on the statement:
> "For this, IPv6 implementations MUST use a domain
>        name within the .invalid DNS top-level domain instead of using
>        the IPv6 unspecified address (i.e., ::)."
> Presumably just "invalid" (alone, as opposed to "xxxx.invalid" would be legal?

That would be the implication of the usual meaning of "within".  Looking
at RFC 1034 (DOMAIN NAMES - CONCEPTS AND FACILITIES), it doesn't appear
that "within" has a special, defined meaning in regard to domain names.

Dale

_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP
Permalink | Reply | 0 comments |
headers
Elwell, John | 5 Feb 2010 12:02

Re: Question on draft-ietf-sipping-v6-transition-07

Vijay,

Thanks, but I was concerned that a domain name beginning with "." might be wrong; Taking the SIP ABNF as an
example (SDP ABNF isn't so precise), we have:

hostname         =  *( domainlabel "." ) toplabel [ "." ] 

John

> -----Original Message-----
> From: Vijay K. Gurbani [mailto:vkg <at> bell-labs.com] 
> Sent: 04 February 2010 20:37
> To: Elwell, John
> Cc: sipping <at> ietf.org
> Subject: Re: [Sipping] Question on draft-ietf-sipping-v6-transition-07
> 
> On 02/03/2010 10:07 AM, Elwell, John wrote:
> > A follow-up question on the statement:
> > "For this, IPv6 implementations MUST use a domain
> >         name within the .invalid DNS top-level domain 
> instead of using
> >         the IPv6 unspecified address (i.e., ::)."
> > Presumably just "invalid" (alone, as opposed to 
> "xxxx.invalid" would be legal?
> 
> John: I think ".invalid" itself suffices, but I will defer to 
> Gonzalo's
> view on this.
> 
> Gonzalo: do you have an opinion.
(Continue reading)

Permalink | Reply | 6 comments |
headers
Kevin P. Fleming | 5 Feb 2010 16:03
Favicon
Gravatar

Re: Question on draft-ietf-sipping-v6-transition-07

Elwell, John wrote:
> Vijay,
> 
> Thanks, but I was concerned that a domain name beginning with "." might be wrong; Taking the SIP ABNF as an
example (SDP ABNF isn't so precise), we have:
> 
> hostname         =  *( domainlabel "." ) toplabel [ "." ] 

That would in fact be wrong, domain names cannot begin with '.'. In
addition, I'd recommend against allowing an 'invalid' name that does not
contain '.', because many systems (including my laptop) are configured
with default DNS search domains that are searched when names that
contain only one component are resolved. 'invalid' in my case would
search in at least two domains for 'invalid.xxx.yyy', which is not
helpful in the context of this draft.

Explicitly specifying 'this.is.invalid' or something similar seems to be
prudent in this case.

--

-- 
Kevin P. Fleming
Digium, Inc. | Director of Software Technologies
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
skype: kpfleming | jabber: kpfleming <at> digium.com
Check us out at www.digium.com & www.asterisk.org
_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP
(Continue reading)

Permalink | Reply | 2 comments |
headers
Elwell, John | 5 Feb 2010 16:45

Re: Question on draft-ietf-sipping-v6-transition-07

Thanks, that seems to make sense to me.

John 

> -----Original Message-----
> From: Kevin P. Fleming [mailto:kpfleming <at> digium.com] 
> Sent: 05 February 2010 15:03
> To: Elwell, John
> Cc: Vijay K. Gurbani; sipping <at> ietf.org
> Subject: Re: [Sipping] Question on draft-ietf-sipping-v6-transition-07
> 
> Elwell, John wrote:
> > Vijay,
> > 
> > Thanks, but I was concerned that a domain name beginning 
> with "." might be wrong; Taking the SIP ABNF as an example 
> (SDP ABNF isn't so precise), we have:
> > 
> > hostname         =  *( domainlabel "." ) toplabel [ "." ] 
> 
> That would in fact be wrong, domain names cannot begin with '.'. In
> addition, I'd recommend against allowing an 'invalid' name 
> that does not
> contain '.', because many systems (including my laptop) are configured
> with default DNS search domains that are searched when names that
> contain only one component are resolved. 'invalid' in my case would
> search in at least two domains for 'invalid.xxx.yyy', which is not
> helpful in the context of this draft.
> 
> Explicitly specifying 'this.is.invalid' or something similar
(Continue reading)

Permalink | Reply | 0 comments |
headers
Vijay K. Gurbani | 5 Feb 2010 17:31
Favicon

Re: Question on draft-ietf-sipping-v6-transition-07

Elwell, John wrote:
> Thanks, but I was concerned that a domain name beginning with "."
> might be wrong; Taking the SIP ABNF as an example (SDP ABNF isn't so
> precise), we have:
> 
> hostname         =  *( domainlabel "." ) toplabel [ "." ]

John: Quite true.  However, in the case we are talking about,
the ".invalid" appears in the SDP, which is governed by the
SDP ABNF.  The ".invalid" is not appearing in the URI, but
in the "c=" line in SDP.

The SDP ABNF in rfc4566 has this grammar for the
"c=" line:

connection-field =    [%x63 "=" nettype SP addrtype SP
                          connection-address CRLF]
connection-address =  multicast-address / unicast-address
unicast-address =     IP4-address / IP6-address / FQDN / extn-addr
extn-addr =           non-ws-string
non-ws-string =       1*(VCHAR/%x80-FF)
                       ;string of visible characters

Does not ".invalid" in the SDP result from the extn-addr
production rule above?  Consequently, should it not be legal?
Or am I missing something here?

Thanks,

- vijay
(Continue reading)

Permalink | Reply | 2 comments |
headers
Vijay K. Gurbani | 4 Feb 2010 21:37
Favicon

Re: Question on draft-ietf-sipping-v6-transition-07

On 02/03/2010 10:07 AM, Elwell, John wrote:
> A follow-up question on the statement:
> "For this, IPv6 implementations MUST use a domain
>         name within the .invalid DNS top-level domain instead of using
>         the IPv6 unspecified address (i.e., ::)."
> Presumably just "invalid" (alone, as opposed to "xxxx.invalid" would be legal?

John: I think ".invalid" itself suffices, but I will defer to Gonzalo's
view on this.

Gonzalo: do you have an opinion.

Thanks,

- vijay
--

-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg <at> {alcatel-lucent.com,bell-labs.com,acm.org}
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP
Permalink | Reply | 9 comments |
headers
Elwell, John | 5 Feb 2010 17:44

Re: Question on draft-ietf-sipping-v6-transition-07

Vijay,

Yes, I know we are talking SDP, but the SDP ABNF was so imprecise that I looked elsewhere. I am not sure where
the definitive source is, but I just took SIP as an example, which seemed to suggest that the people who
wrote RFC 3261 thought that a single element (without any dot) was wrong. This could mean that if you take a
single element and use it as input to the DNS mechanism, you get the wrong outcome, and hence including a
single element in SDP would be equally wrong. Kevin's message seems to support that belief.

John

> -----Original Message-----
> From: Vijay K. Gurbani [mailto:vkg <at> alcatel-lucent.com] 
> Sent: 05 February 2010 16:32
> To: Elwell, John
> Cc: sipping <at> ietf.org
> Subject: Re: [Sipping] Question on draft-ietf-sipping-v6-transition-07
> 
> Elwell, John wrote:
> > Thanks, but I was concerned that a domain name beginning with "."
> > might be wrong; Taking the SIP ABNF as an example (SDP ABNF isn't so
> > precise), we have:
> > 
> > hostname         =  *( domainlabel "." ) toplabel [ "." ]
> 
> John: Quite true.  However, in the case we are talking about,
> the ".invalid" appears in the SDP, which is governed by the
> SDP ABNF.  The ".invalid" is not appearing in the URI, but
> in the "c=" line in SDP.
> 
> The SDP ABNF in rfc4566 has this grammar for the
(Continue reading)

Permalink | Reply | 1 comment |
headers
Vijay K. Gurbani | 5 Feb 2010 18:15
Favicon

Re: Question on draft-ietf-sipping-v6-transition-07

Elwell, John wrote:
> Yes, I know we are talking SDP, but the SDP ABNF was so imprecise
> that I looked elsewhere. I am not sure where the definitive source
> is, but I just took SIP as an example, which seemed to suggest that
> the people who wrote RFC 3261 thought that a single element (without
> any dot) was wrong. 

John: I would believe the authors of rfc3261 focused on
the signaling itself and relegated the nuances of body handling
to appropriate mechanisms and documents.  We should probably
seek their opinion.

That said, I am not refuting your or Kevin's stance.  Clearly,
implementers opt for the path of least resistance, and this
dictates that they parse the token from SDP "c=" line and feed
it directly into the DNS routines (which will certainly hiccup
when given ".invalid").  However, pedantically speaking, the SDP
ABNF seems to indicate that ".invalid" is legal (at least to my
reading.)

Thus, when a developer armed with the pedantic interpretation
squares off against a developer armed with an implementation
that took the path of least resistance, who is right?  I am
sure you have been part of countless arguments where standards
say one thing and what is implemented is a bit ... well ...
different ;-)

Thanks,

- vijay
(Continue reading)

Permalink | Reply | 0 comments |
headers
Dale Worley | 5 Feb 2010 20:29
Favicon

Re: Question on draft-ietf-sipping-v6-transition-07

On Thu, 2010-02-04 at 14:37 -0600, Vijay K. Gurbani wrote:
> On 02/03/2010 10:07 AM, Elwell, John wrote:
> > Presumably just "invalid" (alone, as opposed to "xxxx.invalid" would be legal?
> 
> John: I think ".invalid" itself suffices, but I will defer to Gonzalo's
> view on this.

Why have we put a leading "." on "invalid"?  We do not write, e.g.,
".bell-labs.com".  I know of no other place where a domain name is
written with a leading period.

Dale

_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP
Permalink | Reply | 1 comment |

Gmane