Re: [YANG] mandatory empty

Andy Bierman <ietf@andybierman.com> wrote:
> Hi,
> 
> What does this mean exactly?
> 
>    leaf foo {
>      type empty;
>      mandatory true;
>    }
> 
> Does this translate to:
> 
>    <element name='foo' minOccurs='0'/>

No, b/c the default type in XSD is anyType, which means any xml.  And
if it's mandatory, it should be minOccurs='1' (which is the default).
But it depends on what your XSD is supposed to validate - a NETCONF
PDU or the datastore.

I translate this into:

      <xs:element name="foo">
        <xs:complexType/>
      </xs:element>

> The empty data type can be modeled in XSD as an empty complexType:
> 
>    <complexType name='empty'/>

Yes.

> Can it be modeled as a simpleType?
> 
>    <simpleType name='empty'>
>      <restriction base='xs:string'>
>        <maxLength value='0'/>
>      </restriction>
>    </simpleType>

I don't know.  Maybe?

> If mandatory means that the manager must supply a value,
> then the leaf 'foo' can never be missing.
> 
> Even if it means the agent must supply a value,
> it still seems strange.  Perhaps more text is needed
> about the empty data type.
> 
> IMO, 'mandatory' and 'units' are not applicable to built-in type 'empty'.

I agree that units on a type empty seems weird.

mandatory true on type empty also seems weird, since a type empty leaf
provides a single bit of information (in its existance).  So if it
always has to be there, what is the point?

I don't know if we have to make this an error though.

/martin
_______________________________________________
YANG mailing list
YANG@ietf.org
http://www.ietf.org/mailman/listinfo/yang