Re: nested operation attribute interoperability

[Martin Bjorklund <mbj@tail-f.com>]

Andy Bierman <ietf@andybierman.com> wrote:
>  >...
> > 
> > 1) only 'create' and 'delete' have any kind of existence tests.
> >    Merge and replace can always work and never generate an error because
> >    of data present or missing in the target configuration.
> > 
> 
> I'm wrong again. :-(
> There is another undocumented existence test. (!)
> If the default-operation is 'none' for any of the
> start nodes in the <config> data, and those corresponding
> nodes do not exist in the target, then a data-missing
> error should be generated.

But this is documented already in section 7.2 in the text about 'none':

         none: The target datastore is unaffected by the configuration
            in the <config> parameter, unless and until the incoming
            configuration data uses the "operation" attribute to request
            a different operation.  If the configuration in the <config>
            parameter contains data for which there is not a
            corresponding level in the target datastore, an <rpc-error>
            is returned with an <error-tag> value of data-missing.


> Now I have to start an errata list :-(
> 
> 1) data-exists error is for delete only, not replace and delete
>
> 2) data-missing error is for create and none, not just create

I think you mean

  1) data-missing error is for delete and none, not delete and replace.

(data-exists is for create only, which is what the draft says).



> Note that these errors return no parameters.  I am adding
> proprietary elements to identify what is or isn't missing.

You can always use the standard error-path, even if there's no
error-info, right?  That's what I do, and I also send an error-info
with a bad-element, although it's not required by the description in
the appendix.   (Of course this depends on how you interpret
"Error-info:  none" - does it mean "you MUST NOT send an error-info"
or "you don't have to send an error-info")


/martin



--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>