Re: [yang-doctors] Yangdoctors early review of draft-ietf-idr-bgp-model-07

Thank you very much Andy for this great effort!

Cheers,
Mehmet

> -----Original Message-----
> From: yang-doctors <yang-doctors-bounces@ietf.org> On Behalf Of Andy
> Bierman via Datatracker
> Sent: Monday, December 30, 2019 8:49 PM
> To: yang-doctors@ietf.org
> Cc: idr@ietf.org; draft-ietf-idr-bgp-model.all@ietf.org
> Subject: [yang-doctors] Yangdoctors early review of draft-ietf-idr-bgp-
> model-07
> 
> Reviewer: Andy Bierman
> Review result: On the Right Track
> 
> 
> Major issues:
> 
>   - The ietf-bgp module uses (legal) YANG 1.1 constructs that are
>     not supported by available tools so it cannot be validated.
>     The solution options are
>     (A) rewrite the module so it does not place any definitions
>         in the main module.  Instead place all definitions in
>         submodules. Add YANG 1.0 include-stmts as needed so pyang
>         can validate the module
>     (B) wait until the opensource tools properly support this
>         YANG 1.1 usage and resubmit the module at that time
> 
>   - The modules import ietf-routing-policy. Version used was
>     2019-03-04.  This module has a fatal error caused because
>     ietf-interface-common has apparently been replaced with
>     ietf-if-extensions@2019-11-04
> 
>       leaf subinterface {
>         type leafref {
>           path "/if:interfaces/if:interface/if-cmn:encapsulation"
>              + "/if-l3-vlan:dot1q-vlan"
>              + "/if-l3-vlan:outer-tag/if-l3-vlan:vlan-id";
>         }
> 
>     The path expression is wrong. if-cmn:encapsulation is now
>     if-ext:encapsulation.  You need to check the XPath everywhere
>     when you refactor YANG modules.
> 
>   - Full review of these modules is not possible at this time without
>     proper opensource tools. Processing the YANG statements
>     by hand is extremely difficult, given the large amount
>     of groupings used which are spread across many modules and
>     submodules.
> 
> 
> Minor Issues
> 
>   - The "clear" actions need to be specific about the exact set
>     of objects that are affected. The value of each object once it
>     is cleared needs to be specified in every object that is affected
>     by the action.  Impact on operations should be explained in
>     each action-stmt
> 
>   - Some TODO items still remain indicating the YANG definitions
>     are temporary and will be replaced by more correct statements.
> 
>   - Additional reference-stmts would be useful to implementors
>     who are familar with the routing RFCs that the YANG objects
>     are based on
> 
> Editorial Comments
> 
>   - The implementation complexity appears to be very high
>     for both server and client developers.  The interactions between
>     subtrees is non-trivial and probably requires more clarifications.
>     This needs to be done by implementors, not document reviewers.
> 
> 
> _______________________________________________
> yang-doctors mailing list
> yang-doctors@ietf.org
> https://www.ietf.org/mailman/listinfo/yang-doctors