[yang-tooling] Re: SID file issues

[Vojtech Vilimek <[e-mail]>]

Hi Andy, Micheal and all others!

Andy Bierman wrote:
 > Also for pyang, since the mainline 2.7.1 follows the RFC-defined syntax.
 > I guess the core-wg fork does not.

I think there is already some branch that is rebased on the currect 
upstream master.

 > The YANG typedef clearly identifies the absolute-schema-nodeid, which 
is the syntax used
 > by the augment-stmt argument.  The example in Appendix A shows path 
strings without choice and case nodes.

I wrote this on the list already. See link to the list archive [1]. I 
died off partly because I had an examination period.


Michael Richardson wrote:
 >> Is there a regression/unit test case that expresses the case?

Yes, see test 33 of the sid pyang plugin, files Makefile, 
choice-case.yang [2], test-33-expected-choice-case.sid.


 >>    > IMO this is a high priority issue for RFC 9595.  Also for 
pyang, since
 >>    > the mainline 2.7.1 follows the RFC-defined syntax.  I guess the 
core-wg
 >>    > fork does not.
 >>
 >> I am unaware of any issue/regression with .sid file parsing in 
cbor-wg/pyang

This is connected the the fact that the outputs of pyang before version 
2.7(.0) produced something which did not conform the the ietf-sid-file 
module, the code for identifiers was weird and rewrote it. Andy already 
found new bug for submodules. When I have time I'll fix it.


Michael Richardson wrote:
 > such that the the SID number stayed the same.
 > Current tooling (Laurent Toutain branch) will re-assign fun/a.

You are right it being a bug. The tooling should reuse the SID because 
the SID allocation is permanent. I think we will have A LOT of 
interesting debates because the YANG (semver) versioning is not 
compatible with the RFC 9595 SIDs.



Best regards,

Vojtech Vilimek
CZ.NIC z.s.p.o.

[1]: https://mailarchive.ietf.org/arch/msg/core/L5md7N1QtwrCdV5QUK_FOznpQVE/
[2]: 
https://github.com/mbj4668/pyang/blob/master/test/test_sid/choice-case.yang