Re: [Netconf] [yang-doctors] YANG Doctor question: empty mandatory choice?

["Eric Voit (evoit)" <[e-mail]>]

From: Andy Bierman, August 1, 2018 9:42 PM


On Wed, Aug 1, 2018 at 6:34 PM, Kent Watsen <[e-mail]<mailto:[e-mail]>> wrote:

> It just seems strange that the model designers are positive the solution belongs in a
> case-stmt inside this choice.  So positive that MUST is in order, which means harm
> to the Internet will happen if this choice is not provided, with no possibility of any
>  exception ever.

Disagree.  As Robert pointed out, the yang update rules allow mandatory true to
become mandatory false.  In my view, choosing to do the "mandatory true" now
is the conservative choice.

PS: I do not believe it is possible to have a receiver that does not define a transport.
       If that were allowed, then it would be vendor magic.



OK -- I do think it is a huge burden for a vendor to provide an augment,
even if it is just to indicate "my real transport config is over there
or it is not in YANG".

<Eric>  Instead of “augment” for now we will simply going to deviate away the “empty mandatory choice” until call home information is placed into vendor native or IETF YANG models.

That deviate approach is easier than augmentation.  And it matches what we need to do when something isn’t available in YANG anyway.

Does anyone have an issue if we try a consensus call on Kent’s proposal of “empty mandatory choice”?

Eric

Kent // contributor


Andy