Andy Bierman <[email protected]> wrote:
> On Sun, Oct 25, 2015 at 6:58 AM, Martin Bjorklund <[email protected]> wrote:
> 
> > Andy Bierman <[email protected]> wrote:
> > > On Sat, Oct 24, 2015 at 6:07 AM, Martin Bjorklund <[email protected]>
> > wrote:
> > >
> > > > Juergen Schoenwaelder <[email protected]> wrote:
> > > > > On Fri, Oct 23, 2015 at 10:35:48AM +0200, Ladislav Lhotka wrote:
> > > > > > Martin Bjorklund <[email protected]> writes:
> > > > > >
> > > > > > > auto-deletion in choice/when should be described as a property
> > of the
> > > > > > > data model for the datastore.  Parts of the text from Section
> > 8.2.2
> > > > > > > should be made more generic and moved, probably to a new section
> > > > > > > 8.1.1.   I will have a look at this.
> > > > >
> > > > > [...]
> > > > >
> > > > > > IMO YANG spec should tell what's valid and what isn't, and stop
> > there.
> > > > >
> > > > > As technical contributor, I tend to agree. The purpose of validation
> > > > > should be to return a boolean - datastore is valid or invalid.
> > > >
> > > > Right.  This is what "must" does.  "when" is different.  If a node's
> > > > "when" expression becomes false, that node is deleted, and its other
> > > > constraints are no longer used (e.g. must, mandatory etc).  These are
> > > > two different use cases, two different tools available to the data
> > > > model designer.
> > > >
> > > > If we put "when" to the side for a moment, do you also think that
> > > > there should be no auto-deletion of cases in a choice?
> > > >
> > > > If this discussion had started from implementation/deployment
> > > > experience that said that "when" could not be implemented or that it
> > > > made it difficult to write NMS system or something else, things would
> > > > be different.  But now we have a feature that has been in use for 5+
> > > > years, and there are several implementations of it out there, and now
> > > > we say that it should be removed?  Or worse, keep the syntax but
> > > > radically change the semantics.
> > > >
> > > >
> > >
> > > The one special case that has been identified is data provided
> > > as part of an edit that is silently deleted and <ok/> returned.
> > > This could be considered different than providing an edit that
> > > deletes existing data nodes.
> >
> > Right.  I think an error should be returned to the client in this
> > case.  This is nicer to the client.  Otherwise, the client might set
> > something, get <ok/> back and the might be surprised when the changes
> > did not actually happen.
> >
> > (FWIW, this is how we interpreted the spec and thus have implemented)
> >
> >
> 
> What text says return an error here?

I think we already had this discussion.  I thought that 8.3.1 was
supposed to cover this, but I can see how it doesn't.  So I agree
there is no such clear text in 6020, even though that was the
intention.  I am open to an discussion about the best behavior in this
case.  It is trivial to change the code to not return the error.



/martin

_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod

Reply via email to