On Wed, May 20, 2015 at 11:27 PM, Juergen Schoenwaelder
<[email protected]> wrote:
> On Wed, May 20, 2015 at 10:51:28AM -0700, Andy Bierman wrote:
>
>> The solution is not intuitive at all.  The implications of cherry-picking
>> YANG statements from various revisions of a YANG module are
>> not well understood.
>
> If there are issues that have not been brought up before that we do
> not understand, please put them on the table. But things need to be
> concrete, not abstract.
>
>> The concept of writing a YANG module so it works with all possible
>> combinations of all possible back-revisions of imported modules is
>> not well understood.
>
> This is an example of an abstract claim that is difficult to work
> with.

According to Y45-04, if I use import-by-revision it only applies
to some statements.  The vendor can use any revision for
the rest of the statements.  This may be abstract to you, but
it sounds like a tech-support nightmare to me.

>
>> The question "why can't the module being updated use the latest revision
>> of an import" is quite legitimate.  So is the question "Why can't I determine
>> what a server implementation is required to support for a given YANG module?"
>>
>> The issue title "Better YANG conformance" is misleading because YANG
>> conformance specificity is actually much worse with this solution.
>> It's great for vendors who want to interpret for themselves
>> what conformance means.
>
> This is what I wrote:
>
>   I believe we have reached rough consensus to adopt Y45-04 in order
>   to resolve import ambiguities (aka typedef drift and grouping drift)
>   and we will leave it to YANG extensions (to be worked on in the
>   future) to provide means to define explicit conformance requirements
>   (instead of trying to derive conformance requirements from import
>   relationships alone).
>
> I think we concluded from ~1 year of discussions that deriving
> conformance requirements from import statements does not work.
>


Then we agree that this issue does nothing for conformance
(except make it worse).


> /js
>

Andy

> --
> Juergen Schoenwaelder           Jacobs University Bremen gGmbH
> Phone: +49 421 200 3587         Campus Ring 1 | 28759 Bremen | Germany
> Fax:   +49 421 200 3103         <http://www.jacobs-university.de/>

_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod

Reply via email to