On 10/02/2016 15:46, Kent Watsen wrote:
If there was a way that YANG patch (or equivalent) were able to return
both old and new values (in the same tree) then I think that would be
better. I don't think that such as solution would be specific to the
opstate requirements and may be useful more generally.
Already the <get-diff> RPC accepts an enumeration specifying the response
format. It seems that all that is needed is to define another enumeration for
something like “yang-diff” that can do all you say. We don’t have yang-diff today,
but someone can write a draft to define it and then it can be added or used instead.
Alternatively, the <get-diff> RPC could be moved into a diff-draft that defines both
the RPC and the format together. Something like this might make sense as <get-diff>
is generally useful outside of this opstate application.
Yes, OK.
I can easily see how a <get-diff> RPC message could provide two separate
trees, i.e. one with the values before, and a separate tree with the
values after, but I'm not sure that such a message format would really
be any better than the current YANG patch format.
However, the bit that I'm stuck on is that I can't see how is it
possible to provide the data in a single tree unless you end up with an
encoding that is broadly similar to what I have suggested in my draft,
and several folks have strong views against any such encoding. AFAIK,
no alternatives have been proposed.
Rob
Kent // contributor
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
