Re: [yang-doctors] [Pce] Yangdoctors early review of draft-ietf-pce-pcep-yang-08

[tom petch <[e-mail]>]

----- Original Message -----
From: "Dhruv Dhody" <[e-mail]>
Sent: Tuesday, March 26, 2019 5:35 PM

Done and a new version is posted.

<tp>

The prefixes look good and the line lengths I think are ok - I make it
81 characters including the left margin spaces -  but ...

in -10 you had
"   import ietf-te-types {   prefix "te-types";  reference
       "I-D.ietf-teas-yang-te-types: Traffic Engineering Common YANG
       Types";  "
which looks ok

in -11 you have
  import ietf-te-types {     prefix te-types;     reference
      "[I-D.ietf-teas-yang-te-types]: Traffic Engineering Common YANG
       Types";
which does not look ok to me

Recall that a YANG module is plain text so there cannot be any XML/HTML
style references/links and that
     [I-D.ietf-teas-yang-te-types]:
looks to me as if it has XML underpinning it which is not ok in a YANG
module

And this has happened in 16 places.

Also, IANA Considerations does not register

   URI:  urn:ietf:params:xml:ns:yang:ietf-pcep-stats

Security Considerations talks of a YANG module when the I-D has two,
giving the flavour, as with the IANA Considerations, that the second
module, pcep-stats, has not really been considered.  Since it is
statistics, I imagine that there are no sensitive objects there, in
which case I would add a sentence to spell that out.

And something new
/Segement Routing /Segment Routing /

Tom Petch

On Tue, 26 Mar 2019 at 18:34, tom petch <[e-mail]> wrote:

> Dhruv
>
> I commented up thread that the prefixes used in this I-D were not the
> ones that appear in the imported modules and said I thought that that
> was discouraged but ok. Checking RFC8407, it says
>
>    o  The proper module prefix MUST be used for all identifiers
imported
>       from other modules.
>
> It is a MUST not a SHOULD so I believe that you must bring those
> prefixes in line for key-chain, tls-client, tls-server.  YANG allows
it,
> YANG guidelines does not.
>
> Tom Petch
>
>
> ----- Original Message -----
> From: "Dhruv Dhody" <[e-mail]>
> Sent: Tuesday, March 26, 2019 2:38 PM
>
> Hi Martin,
>
> The newer version of pyang worked! Thanks for your help!
>
> I found the full tree useful when I am searching for a leaf in the
> yang models and understand how it fits in the overall tree. Thus I see
> value in both. We can also consider if we should also update 5.2-5.6
> additionally.
>
> Thanks!
> Dhruv
>
> On Tue, Mar 26, 2019 at 3:00 PM Martin Bjorklund <[e-mail]>
wrote:
> >
> > Hi,
> >
> > Dhruv Dhody <[e-mail]> wrote:
> > > Hi Mahesh, Tom,
> > >
> > > Got it, will make the necessary change soon.
> > >
> > > Where I need help is the tree creation, even though I use
> > > '--tree-line-length' I faced the issue with overrunning the 80
> > > characters.
> > >
> > > pyang --ietf -f tree --tree-line-length=68 --tree-depth=10
> > > [e-mail] --ietf >ietf-pcep.tree
> >
> > Have you tried using pyang 1.7.8?  When I run that the tree seems to
> > fit the line lengths.
> >
> > > That made me pick a shorter prefix, but happy to learn if there is
a
> > > better way out there!
> >
> > Personally, I'm not too fond of very large tree diagrams.  I prefer
to
> > split them into smaller diagrams.  So I like your overview diagram
in
> > section 5.1.  I would then probably add a small diagram in each of
the
> > section 5.2-5.6, and remove secion 5.7 completely.  But this is just
> > my personal preference!
> >
> > /martin
> >
> > > Thanks!
> > > Dhruv
> > >
> > > On Tue, Mar 26, 2019 at 2:16 PM Martin Bjorklund <[e-mail]>
> wrote:
> > > >
> > > > Hi,
> > > >
> > > > tom petch <[e-mail]> wrote:
> > > > > On the question of prefix, where I an interested in the
opinion
> of a
> > > > > YANG
> > > > > Doctor, you use the single letter 'p' and say that a longer
> prefix gives
> > > > > you line length problems.  YANG does allow statements to span
> lines, as
> > > > > happens in almost every TEAS module so for me that is not a
very
> good
> > > > > reason; I would prefer something of two characters or more.
> > > > >
> > > > > I note that IANA Considerations says
> > > > >        Prefix:       pcep
> > > > > which would be my first choice even if I then have to span
> lines.
> > > >
> > > > I strongly agree.  Since the prefix is actually part of the IANA
> > > > registry and needs to be unique, I think you should use a longer
> > > > prefix.  "pcep" seems reasonable.  If you run into line length
> > > > problems, I'll be glad to help you fix them.
> > > >
> > > > Before this document goes to the RFC editor, I suggest you run
the
> > > > tool:
> > > >
> > > >    pyang -f yang --keep-comments --yang-line-length 69 <FILE>
> > > >
> > > > on these modules, in order to get them formatted consistently
with
> the
> > > > rest of the IETF modules.
> > > >
> > > > > You import the module key-chain but you do not use the prefix
> that it
> > > > > defines, namely key-chain; not forbidden but not recommended
> practice
> > > > >
> > > > > Likewise tls-client should be tlsc and tls-server tlss.
> > > > >
> > > > > Security and IANA Considerations deal with
> > > > >        Name:         ietf-pcep
> > > > > What about
> > > > >    module ietf-pcep-stats {
> > > > > which I think needs separate coverage, a separate section, in
> Security
> > > > > and must be covered in IANA Considerations.
> > > > >
> > > > > The problem with
> > > > > "I-D.ietf-pce-association-group: PCEP Extensions for ...
> > > > > as a reference is that when it appears in the text of the I-D,
> then it
> > > > > is as
> > > > >  [I-D.ietf-pce-association-group]
> > > > > i.e. a XML/HTML type anchor which is picked up by tools so the
> RFC
> > > > > Editor cannot miss it.
> > > > >
> > > > > When it appears in the YANG module, it must be plain text as
in
> > > > >        "I-D.ietf-pce-association-group: PCEP Extensions for
....
> > > > > so the tools cannot pick it up, it must be spotted by eye and
so
> might
> > > > > be missed.  Hence I suggest using
> > > > >
> > > > > "RFC YYYY - PCEP Extensions for
> > > > >        Establishing Relationships Between Sets of LSPs";
> > > > >
> > > > > with a note to the RFC Editor asking them to replace YYYY with
> the RFC
> > > > > number assigned to I-D.ietf-pce-association-group
> > > > >
> > > > > Likewise RFC ZZZZ for
> > > > >        "I-D.ietf-pce-segment-routing: PCEP Extensions for
> Segment
> > > > > and so on for the others (of which there are several)
> > > > >
> > > > > The RFC Editor is ok, likes even, all the notes thereon to
> appear once
> > > > > at the start of the I-D.
> > > > >
> > > > > So my previous comment was that using XXXX for multiple I-Ds
was
> > > > > confusing but I meant to use YYYY ZZZZ, with an RFC Editor
Note
> for
> > > > > each, and not to use the I-D name.
> > > > >
> > > > > HTH
> > > > >
> > > > > Tom Petch
> > > > >
> > > > > ----- Original Message -----
> > > > > From: "Dhruv Dhody" <[e-mail]>
> > > > > To: "Mahesh Jethanandani" <[e-mail]>
> > > > > Cc: <[e-mail]>;
> <[e-mail]>;
> > > > > <[e-mail]>
> > > > > Sent: Sunday, March 24, 2019 9:07 PM
> > > > > Subject: Re: [Pce] Yangdoctors early review of
> > > > > draft-ietf-pce-pcep-yang-08
> > > > >
> > > > >
> > > > > Hi Mahesh,
<snip>