Re: [netmod] regarding draft-bierman-netmod-yang-data-ext-00

[Andy Bierman <[e-mail]>]

Hi,

The use-cases for groupings/uses and augment are not identical.

Alternative NMDA Approach:

I don't see a big difference between defining YANG for an artifact vs.
defining some YANG for a special-purpose datastore.
There is nothing about the YANG data that is different.
There are only differences in the operations on that data,
which are properties of the datastore. Perhaps an <artifact> datastore
is a cleaner long-term solution.


Andy


On Sat, Sep 2, 2017 at 9:59 AM, Kent Watsen <[e-mail]> wrote:

>
>
> >> Gotcha.  What do other people think, would a "uses-yang-data"
> >> statement be generally more useful?
> >
> > But does this mean we also do uses-yang-container, uses-yang-list,
> > uses-yang-xyz to other definitions as well? I do not think this is
> > desirable and why would yang-data be any different?
>
>
> Oh, no.  I view yang-data as a body-stmt, not a data-def-stmt,
> akin to 'rpc' and 'notification'.  But back to the draft's
> proposed solution, I understand that augment-stmt is a top-level
> body-stmt, but it's also a uses-stmt, and in that context, it is
> a sibling to the refine-stmt.  Already ANIMA has a need to both
> augment and refine a yang-data defined structure.  To me, a
> 'uses-yang-data' statement seems just one step more general than
> a 'augment-yang-data' statement.
>
> Specially, if we were updating YANG:
>
>    // just added last entry
>    body-stmts          = *(extension-stmt /
>                            feature-stmt /
>                            identity-stmt /
>                            typedef-stmt /
>                            grouping-stmt /
>                            data-def-stmt /
>                            augment-stmt /
>                            rpc-stmt /
>                            notification-stmt /
>                            deviation-stmt /
>                            yang-data-stmt)
>
>    // something like this?
>    yang-data-stmt            = yang-data-keyword sep identifier-arg-str
> optsep
>                          (";" /
>                           "{" stmtsep
>                               ;; these stmts can appear in any order
>                               [status-stmt]
>                               [description-stmt]
>                               [reference-stmt]
>                               *(typedef-stmt / yang-data-grouping-stmt)
>                               *data-def-stmt
>                               [uses-yang-data-stmt]
>                           "}") stmtsep
>
>    // like 'uses-stmt', but with a different keyword
>    uses-yang-data-stmt       = uses-yang-data-keyword sep
> identifier-ref-arg-str optsep
>                          (";" /
>                           "{" stmtsep
>                               ;; these stmts can appear in any order
>                               [when-stmt]
>                               *if-feature-stmt
>                               [status-stmt]
>                               [description-stmt]
>                               [reference-stmt]
>                               *refine-stmt
>                               *uses-augment-stmt
>                           "}") stmtsep
>
>    // like 'grouping-stmt', but with 'action' and 'notification' removed
>    yang-data-grouping-stmt       = yang-data-grouping-keyword sep
> identifier-arg-str optsep
>                          (";" /
>                           "{" stmtsep
>                               ;; these stmts can appear in any order
>                               [status-stmt]
>                               [description-stmt]
>                               [reference-stmt]
>                               *(typedef-stmt / yang-data-grouping-stmt)
>                               *data-def-stmt
>                           "}") stmtsep
>
>
>    yang-data-keyword              = %s"yang-data"
>    uses-yang-data-keyword         = %s"uses-yang-data"
>    yang-data-grouping-keyword     = %s"yang-data-grouping"
>
>
> How close to this can we get using an extension statement, so that a
> future adoption to YANG will be as seamless as possible?
>
> K.
>
>
>