Re: [YANG] new pyang errors

Martin Bjorklund writes:
>Andy has already defined such a capability in one of his drafts (I
>forgot which one...).

http://www3.ietf.org/internet-drafts/draft-bierman-ncx-ext-00.txt

This defines a set of protocol extensions:

     2.1.  test-only Enumeration  . . . . . . . . . . . . . . . . . .  5
     2.2.  with-defaults Attribute  . . . . . . . . . . . . . . . . .  7
     2.3.  with-metadata Attribute  . . . . . . . . . . . . . . . . .  9
     2.4.  last-modified Attribute  . . . . . . . . . . . . . . . . . 11

IMHO these should be distinct capabilities, since this is the way
NETCONF allows for such extensions.

One odd line (on first read) is:

   The 'with-defaults' attribute is defined in the namespace specified
   as the 'targetNamespace' in Section 3.  However, an agent should
   accept it even if no namespace is used.

This seems to play fast and loose with XML namespaces, which isn't
a good idea.

Another issue is that the draft puts these as attributes on the
<rpc>, instead of putting them inside the proper operations:

   If the 'with-defaults' attribute is
   present in the <rpc> element, for any of the protocol operations
   which return the contents of a configuration database, the agent will
   use the 'with-defaults' value to control default suppression.

I'm also missing where the metadata returned with with-metadata is
defined.

Thanks,
 Phil

_______________________________________________
YANG mailing list
YANG@ietf.org
https://www1.ietf.org/mailman/listinfo/yang