Re: [netmod] ?==?utf-8?q? augment target statements

[Michal Vaško <[e-mail]>]

 
 
 
On Tuesday, January 24, 2017 10:22 CET, Martin Bjorklund <[e-mail]> wrote: 
 
> Hi,
> 
> Michal Vaško <[e-mail]> wrote:
> > Hi,
> > we have noticed that RFC 7950 page 120, when mentioning augment
> > substatements targeting container, list, case, input, output, and
> > notification, skips "anydata" and "anyxml".
> 
> This is correct; the *target* of an "augment" cannot be "anydata" or
> "anyxml".

Okay, I was ambiguous and actually wanted to ask, whether "anydata" or "anyxml" can be a substatement of an augment targeting container, list, case, input, output, or notification. If not, then I wonder why.

> > Based on the next section
> > 7.17.1. and the grammar, it seems not intentional and we wanted to
> > submit errata.
> 
> 7.17.1 lists substatemets to the "augement" statement, i.e., valid
> nodes to *add to* the target node.
> 
> > Also, in the same paragraph, the target notification is allowed and
> > mentioned, but action not. Again, is that intentional or just an
> > oversight? Thanks.
> 
> Intentional.  The target of an augment cannot be an "action" (or
> "rpc").  Note that the target can be "input" or "output".  (And note
> that the "input" and "output" nodes always exist as children to
> rpc/action)
 
Right, of course.

Regards,
Michal Vasko