> I know the accompanying text. I am trying to fix the ABNF by restricting the
> given parameter-name to occur once.
>
> I don't see problem you mentioned in my proposal.
>
> to-additional-param = ( 0*1(SEMI tag-param)) / ( *(SEMI generic-param))

> This doesn't forces the tag-param to be the Ist parameter. It says the
> to-additional-param nonterminal can have either tag-param or
> generic-param

Actually, the biggest problem with your ABNF is that it specifies you can have a 
tag-param OR any number of generic-params, but not both a tag-param AND generic-params.

ABNF is not perfect for describing cases like this one, where the order of parameters 
is optional but the number of parameters of a particular type is not.  Hence the extra 
textual information in the specification.

In more than one case within RFC-3261, the ABNF does not fully limit how parsing 
should occur.  Given that, I think just leaving the ABNF and accompaning text alone as 
it is today will be adequate.  The implementor is responsible for understanding both 
the ABNF and it's context.

- rob

_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors

Reply via email to