A simple answer could be: use quotes

As in
Some-Header: moot; t1="a,b"; t2="x,y"

Regards,
jeroen

Nap wrote:
> Jeroen, Paul
> 
> Thanks for the replies. You are right, my stack (though I wrote it)
> also depended on this behavior :-)
> Question is, if there is a situation where a certain header is
> required to carry some tuples of information then how should it be
> done?
> 
> P-Charging-Function-Addresses is a good example, I have a similar
> requirement. If one were to also be semantically correct then how
> would the BNF 
> look like for P-Charging-Function-Address?
> 
> Now I cannot have multi-valued headers because the header value is an
> anchor and the params hold tuples related to the anchor eg. (the
> header name is just a crude example)
> 
> P-My-Header: Value;p=v1,v2;p=w1,w2
> 
> Here v1,v2 and w1,w2 are all related to "Value" (the anchor) and are
> an ordered pair amongst themselves.
> 
> In fact there will be multiple headers of P-My-Header (just an
> example) with different "Value"
> 
> I also do not need the named parameter "p" and could have had -
> 
> "P-My-Header" HCOLON my-value *(SEMI tuple)
> my-value = Method
> tuple  = value1 COMMA value2
> value1  = token
> value2  = token
> 
> Appreciate your help in this.
> 
> Thanks
> 
> 
> 
> On 3/20/07, Paul Kyzivat <[EMAIL PROTECTED]> wrote:
>> 
>> 
>> Jeroen van Bemmel wrote:
>>> see RFC3455 http://www.ietf.org/rfc/rfc3455.txt section 4.5.2.3
>>> Examples of Usage message F2 (P-Charging-Function-Addresses)
>>> 
>>> Not recommended though - it violates RFC3261
>> 
>> Right. There is semantic, not syntactic (its not in ABNF)
>> requirement in 3261 that parameters be unique. And I know of stack
>> implementations that depend on this.
>> 
>>         Paul
>> 
>>> Regards,
>>> Jeroen
>>> 
>>> 
>>> Nap wrote:
>>>> Is there a precedence somewhere of a SIP header with a multivalued
>>>> parameter (not header).
>>>> 
>>>> An example could be -
>>>> 
>>>> Some-Header: Value;p1=pvalue1;p1=pvalue2;p1=pvalue3 etc  [Note
>>>> name of param is p1 only]
>>>> 
>>>> Alternatively I could think of -
>>>> 
>>>> Some-Header: Value;p1=pvalue1|pvalue2|pvlaue3  with a special
>>>> separator like a PIPE.
>>>> 
>>>> If someone has seen this kind of BNF in any draft or RFC then it
>>>> will immensely help.
>>>> 
>>>> Thanks
>>>> _______________________________________________
>>>> Sip-implementors mailing list
>>>> [email protected]
>>>> https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
>>> 
>>> _______________________________________________
>>> Sip-implementors mailing list
>>> [email protected]
>>> https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
>>> 
>> 
> _______________________________________________
> Sip-implementors mailing list
> [email protected]
> https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

Reply via email to