Use pen number and fix validation error

[JeanQuilbeufHuawei]

Validation errors were due to the yang-packages model that has errors on its own. I commented it out and now the validation passes in my local environment.

[JeanQuilbeufHuawei]

Still a compilation issue due to "uses" mechanism. In the yang-push model, one of the augment targets a choice for the filter, and the added case has a node with the same name as one in the other case of the augmented module but in a different namespace.

In our case, since we use a "uses" to include the augmented part, both the original case and the augmented case are in the same namespace and it is forbidden by YANG to have two nodes with the same name in different cases of a choice.

Current hack is to modify the yang-push module to add a different name (included in the repo and at the end of the draft so that datatracker accepts our module). However we need to find a solution.

[JeanQuilbeufHuawei]

Thanks Jean. With the move to YANG Push, I have one doubt: is there any reason for skipping some parts of YANG Push like filters and receivers? Why not importing the whole YANG Push model? Actually, I think the current data-collection-manifest model might be broken. The model now imports the target choice, which has references to nodes from the filters, though this container is not present in the model.

Yes there is an issue with the platform-level filters, maybe we should include them in the platform manifest? I will try to add a fix in the meantime.

For the receivers, I am not against adding them, I am not sure it is needed to later exploit the data.