On 22.12.2016 11:53, Ladislav Lhotka wrote:
Expected server/protocol behaviour for nodes of a given status is specified in
sec. 7.21.2 and IMO it is a different issue. What may need to be clarified wrt
YANG is how to figure out the status of each node from the data model. In
particular,
- what ancestor/descendant combinations of statuses are permitted
- what is the default if "status" statement is omitted.
I agree with Andy's view when a single module is concerned. In this
sense "references" == "its ability to be accessed is granted by being a
descendant of". I think we may have to have more complex rules taking
into account enclosing statement and inter-module relationships: what
harm is done by instantiating an obsolete grouping?
Unfortunately we cannot legally prohibit 'status current' from appearing
in an augmented (or refined) case from a different module, as not
breaking downstream models seems to be obsolete's raison d'etre (based
on 7950 section 11).
Hence the effective status of any 'status current' nodes with an
obsolete ancestor should be obsolete. I do not think it should be
required to issue more than a warning.
Regards,
Robert
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
