Re: [yang-doctors] [netmod] Clarification needed for YANG 1.1 XPATH context

[Radek Krejčí <[e-mail]>]

Hi,
I'm refreshing an old thread to clarify specific use case, see below...

Dne 25.2.2016 v 16:31 Martin Bjorklund napsal(a):
> William Ivory <[e-mail]> wrote:
>>
>> -----Original Message-----
>> From: Martin Bjorklund [mailto:[e-mail]] 
>> Sent: 25 February 2016 15:17
>> To: William Ivory <[e-mail]>
>> Cc: [e-mail]
>> Subject: Re: [netmod] Clarification needed for YANG 1.1 XPATH context
>>
>> William Ivory <[e-mail]> wrote:
>>> Hi,
>>>
>>> I'm looking for clarification on the meaning of the following 
>>> paragraph in section 6.4.1 (XPATH context) in RFC6020bis:
>>>
>>>    'If a node that exists in the accessible tree has a non-presence
>>>    container as a child, then the non-presence container also exists in
>>>    the tree.'
>>>
>>> It's unclear to me what this is trying to say, and why - for example, 
>>> does this mean that I need to validate any 'must' and 'when'
>>> statements on the child container even when nothing within that child 
>>> container is configured?
>> must expressions are always evaluated if the node where the must
>> expression is defined exists, regardless of the number of children
>> this node has.
>>
>> [wivory] So in my example where the child container (non-presence) has
>> NO children, then it doesn't exist, and any must statement on it
>> should not be run.  Only when a non-presence container has a non-zero
>> number of children should any 'must' statements on that container be
>> run.
>>
>> [wivory] If that's the case, then would it be correct to say that the
>> intention of this paragraph is as a reminder that one must evaluate
>> 'must' statements on nodes that have no inherent meaning and exist
>> only because they contain child nodes?
> No; section 7.5.3 says:
>
>    When a datastore is validated, all "must" constraints are
>    conceptually evaluated once for each node in the accessible tree (see
>    Section 6.4.1).
>
> And the quoted paragraph of 6.4.1 says that the NP-container
> (conceptually) exists if its parent exists - regardless of number of
> children.
>
> So if the parent exists, any must expressions in the NP-container are
> evaluated.
>

what about top-level NP-container with must constraint? Is a root node something which is always present in accessible tree (even in an empty tree)? Intuitively, I believe that it is, so even constraints on top-level NP-containers are supposed to be evaluated, but I cannot find something about it in RFC.

Regards,
Radek