> On 09 Jun 2015, at 22:23, Martin Bjorklund <[email protected]> wrote:
>
> Ladislav Lhotka <[email protected]> wrote:
>> Hi,
>>
>> RFC 6020bis says in Sec. 7.10: The "anydata" statement is used to
>> represent an unknown set of nodes that can be modelled with YANG." I'd
>> like to clarify what this means with respect to namespaces.
>>
>> 1. Is it allowed that the XML content (value) of an anydata node have no
>> namespace? That is:
>>
>> anydata foo;
>>
>> <foo>
>> <bar xmlns=""/>
>> </foo>
>>
>> I assume it is not allowed.
>
> Right, b/c this cannot be encoded with the encoding rules for anydata,
> which says that the element's namespace is the module's XML namespace
> (and this cannot be "").
>
>> 2. For a namespace URI that's used for anydata content, is it required
>> that there exist a YANG module defining this URI as its namespace?
>>
>> I assume it is not required.
>
> Yes.
>
> I don't see how it could be enforced? How can you tell if a YANG
> module exists?
For example, it could be required that the server advertises such a module so
as to be known at run time. I am not proposing this, just verifying that
everybody interprets “can be modelled” the same way.
It is also important for the JSON encoding - it means there won’t necessarily
be a way for mapping XML-encoded instance to JSON and vice versa.
Lada
>
> I do want to be able to have a generic model like:
>
> list logged-notification {
> ...
> anydata notif;
> }
>
> and implement this in a server that doesn't necessarily know the data
> model of every possible notification.
>
>
> /martin
--
Ladislav Lhotka, CZ.NIC Labs
PGP Key ID: E74E8C0C
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
