Re: [netmod] Y34: remove anyxml

[Martin Bjorklund <[e-mail]>]

Phil Shafer <[e-mail]> wrote:
> Martin Bjorklund writes:
> >> I'm not following why anyxml is being considered for removal.
> >> I consider it an important feature and strongly object.
> >The problem is that it is encoding-specifc.  In most case where it is
> >used, what we really mean is "any data modelled with YANG".  Hence,
> >the proposal Y34-02 suggests we add 'anydata' for this purpose.  Such
> >data can be encoded in XML or JSON.
> 
> I take is as a given that there's no realistic say to round trip
> between arbitrary XML and arbitrary JSON (or whatever else the
> future may hold for other potential non-NETCONF uses of YANG).
> So the choices are to keep anyxml or replace it with a string.
> I would be strongly against the latter.  anyxml allows one to
> carry opaque chunks of XML in a YANG data model.
> 
> Consider the recent proxy feature as spec'd in the "peer mount"
> draft and how easily anyxml allows the encapsulation of subhierarchies
> of data within a parent hierarchy.  This is just the sort of problem
> anyxml was meant to address.
> 
> Or consider the need for undocumented content.  In JUNOS, we put
> undocumented configuration under an "anyxml undocumented" member
> so we can transmit our secrets without documenting them.  Clients
> can validate the known parts of their config without tripping on
> the unknown parts.
> 
> anyxml is a useful feature and should not be removed.

The proposal is (Y34-02):

  Keep 'anyxml'.  Introduce 'anydata'

  'anyxml' would still be used to represent unrestricted XML, as is
  done in NETCONF.

  'anydata' would be an unknown chunk of data that can be modelled
  with YANG.  Can be encoded as xml or json.



/martin