Re: [netmod] ?==?utf-8?q? YANG tree diagram uses

Hi Benoit,
the way I understand it is that the draft tries to mention only the core Tree Diagram characteristics and formatting leaving the details implementation-specific. I was simply asking for another opinion on how to handle an implementation detail, I think the draft fulfils the mentioned purpose just fine.

Regards,
Michal

On Sunday, February 4, 2018 10:24 CET, Benoit Claise <bclaise@cisco.com> wrote: 

> Martin, Michal,
> 
> Do we need any clarification in the draft?
> 
> Regards, B.
> > Michal Vaško <mvasko@cesnet.cz> wrote:
> >> Hi,
> >>
> >> we have encountered some problem while implementing a feature from
> >> draft-ietf-netmod-yang-tree-diagrams-05, specifically not resolving
> >> groupings and printing uses names instead (Section 2.2).
> >>
> >> We have 2 example models, A and B. A defines a container and a
> >> grouping. B defines an augment that adds uses into the container from
> >> A and resolves to the grouping from model A.
> >>
> >> grouping A:g;
> >> A:c {
> >>    B:uses A:g;
> >> }
> >>
> >> Now, if printing model A with the augment not resolving uses we
> >> currently print
> >>
> >> +--rw c
> >>     +---u B:A:g;
> > pyang prints this as well, but it is more "by accident".   It looks
> > quite odd.
> >
> > It wouldn't be correct to write
> >
> >      +---u B:g;
> >
> > since 'g' isn't defined in B.
> >     
> > OTOH,
> >
> >      +---u A:g;
> >
> > is correct in the sense that "A:g" is the "name of the grouping", and
> > that is what the current document says should be printed.  Granted,
> > this doesn't show the whole picture, but maybe this is good enough.
> >
> > It might be wise to not print a grouping like this in order to avoid
> > confusion.
> >
> >
> > /martin
> >
> >
> >> since the uses is foreign. We could not decide what the "correct"
> >> output should be and it is likely left to various interpretations but
> >> we were wondering what some of you think. Should it perhaps be only
> >> "B:g" since the grouping becomes local? But what if the grouping would
> >> be from a third model, are 2 prefixes okay? Thanks for your opinions.
> >>
> >> Regards,
> >> Michal
> >>
> >> _______________________________________________
> >> netmod mailing list
> >> netmod@ietf.org
> >> https://www.ietf.org/mailman/listinfo/netmod
> >>
> > _______________________________________________
> > netmod mailing list
> > netmod@ietf.org
> > https://www.ietf.org/mailman/listinfo/netmod
> > .
> >
> 