Re: ODL yang models


Luis Gomez
 



On Oct 23, 2019, at 2:11 AM, Robert Varga <nite@...> wrote:

On 22/10/2019 21:08, Luis Gomez wrote:
Hi everyone,

You may know already, we have an ODL intern, Rohan Julka, working on
yang model pushing to Yang github (https://github.com/YangModels/yang).

As part of this exercise, we added a test in the distribution merge job
to collect and verify ODL generated yang models using pyang, see the
results here:

https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/distribution-merge-managed-magnesium/657/opendaylight-models/

The test shows we have issues in many models (see error file under each
project). Some of them like "import" not used, are easy to fix and have
no impact, but other may require further analysis and the fix may impact
the API.

So how do we want to go about this? should we open bugs to every project
to fix this? any other idea to get the models repaired?

Well, we need attribution to where the model is packaged -- the current
approach of matching namespace is nice (in that it points out wrong
namespaces), but really those 'external' models need to be charged to
the project packaging them (most of which is mdsal, as designed).

For this check we are using the same script that will collect and push ODL modules to Yang github. Because of this, all external (non ODL models) are organized in a separate folder.

But let me ask: what is the value-add of having the external modules in the corresponding projects? If there is some, we can have custom script for just verifying the models in CI.


As for the individual failures ... yeah, they need to be classified and
analyzed, as for example we are more liberal around keyless lists in
config (because they are not a problem for our internal workings).

Regards,
Robert

Join TSC@lists.opendaylight.org to automatically receive all group messages.