Re: [i2rs] Kathleen Moriarty's No Objection on draft-ietf-i2rs-yang-l3-topology-08: (with COMMENT)

[Xufeng Liu <[e-mail]>]

Hi,

> -----Original Message-----
> From: Martin Bjorklund [mailto:[e-mail]]
> Sent: Friday, February 3, 2017 10:32 AM
> To: Xufeng Liu <[e-mail]>
> Cc: [e-mail]; [e-mail];
> [e-mail]
> Subject: Re: [i2rs] Kathleen Moriarty's No Objection on draft-ietf-i2rs-yang-l3-
> topology-08: (with COMMENT)
> 
> Hi,
> 
> 
> Xufeng Liu <[e-mail]> wrote:
> > Hi Kent,
> >
> > > -----Original Message-----
> > > From: Kent Watsen [mailto:[e-mail]]
> > > Sent: Thursday, February 2, 2017 8:39 PM
> > > To: Xufeng Liu <[e-mail]>
> > > Cc: [e-mail]; [e-mail]
> > > Subject: Re: [i2rs] Kathleen Moriarty's No Objection on
> > > draft-ietf-i2rs-yang-l3-
> > > topology-08: (with COMMENT)
> > >
> > > [reducing distribution]
> > >
> > >
> > > Hi Xufeng,
> > >
> > > > Assume the following model:
> > > >
> > > > +--rw nodes
> > > >   +--rw node [id]
> > > >      +--rw id   string
> > > >      +--rw under-lay-attribute-a ???
> > > > +---ro nodes-state
> > > >   +--ro node [id]
> > > >      +--ro id   string
> > > >      +--ro attribute-a string
> > > >
> > > > I cannot define the under-lay-attribute-a to reference attribute-a as:
> > > >               type leafref {
> > > >                 path "../node/attribute-a"'
> > > >               }
> > >
> > >
> > > True, but maybe it could be:
> > >
> > >    type leafref {
> > >       path "../node/attribute-a"
> > >       require-instance false;
> > >       description
> > >         "In the case when the referenced instance is not a configured
> > >          object, the system may resolve it by looking for it under the
> > >          /nodes-state node.  As the referenced operational state data
> > >          may have a lifecycle independent of configuration, this results
> > >          in an effect much like pre-provisioning interfaces in RFC 7223.";
> > >    }
> > [Xufeng] I think that "require-instance false" does not help here. The
> > validation for path "../node/attribute-a" still fails because
> > "attribute-a" does not exist under /nodes/node/.
> 
> No, "require-instance false" essentially turns off validation.  Section
> 9.9.3 in RFC 7950 says:
> 
>    If "require-instance" is "false", it means that the instance being
>    referred to MAY exist in valid data.
> 
[Xufeng] My understanding is that this section talks about instance "data" validation, but I was talking about schema validation. The path statement is pointing to a non-existing schema node. At least, pyang, yangdump, and yangvalidator all fail on this currently.
> 
> /martin