Hi,
Andy Bierman <[email protected]> wrote:
> I have a issue request for next week.
> IMO the text for what an arbitrary "YANG tool" MUST, SHOULD, or MAY
> do with an extension-stmt is not clear and no consensus exists yet either.
>
> The specific use-case is the addition of an "ephemeral-stmt" as a
> sub-statement
> of the individual data-def-stmt types.
>
> I think Martin has suggested a solution that I want to expand and get into
> YANG 1.1.
>
> - A YANG tool MUST support the deviate-stmt and refine-stmt for external
> statements
>
> grouping shared-parms {
> // plain config parms w/ no ephemeral-stmt
> }
>
> container foo-config {
> uses shared-parms {
> refine parm1 {
> i2rs:ephemeral true;
> }
> refine parm3 {
> i2rs:ephemeral true;
> }
> }
> }
>
>
> I do not think the current ABNF and normative text supports external-stmts.
I think the grammar is ok, but I agree normative text is missing. The
grammar allows the example with refine i2rs:ephemeral true above.
In our data models, we explicitly specify if an external statement is
supported in 'refine' or not.
[side node: we do this with a statement like
this:
extension actionpoint {
...
tailf:use-in "rpc";
tailf:use-in "tailf:action";
tailf:use-in "refine";
tailf:substatement "tailf:opaque";
...
}
]
> A similar example can be made for an object not in a grouping, using
> "deviate add".
I don't understand. Can you examplify?
> However deviation-stmt is not allowed in any standard modules, so this
> is also a problem.
>
> IMO this is required to add new keywords to data nodes like 'ephemeral'
> or 'related-state'. Currently YANG 1.1 is not good enough to support
> the use of external statements in standards.
/martin
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
