RFC9595: Redefine the schema-node-path typedef

[vvilimek]

This change is against the YANG update rules because we change typedef's semantics.

Based on what people from the NETMOD WG are saying: The extensions are not part of the schema tree, meaning they should not appear on standard YANG "absolute-schema-nodeid" identified strings.

This definition should fix this separation.

For the rc:yang-data I propose similar solution as for the choice and case nodes.

module example {
  // ... 
  rc:yang-data never-used-name {
    container actual-yang-data {
      leaf version { type uint32; }
      leaf name { type string; }
    }
  }
}

The path for the actual-yang-data and version should be:
/example:actual-yang-data, resp.
/example:actual-yang-data/version

[vvilimek]

ML discussion

[vvilimek]

Rework: the absolute-schema-nodeid does not come with semantics binding the identifier to (classical YANG) schema nodes. That being said, we should clarify how to encode paths with sx:structure and rc:yang-data.

[vvilimek]

A related PR #171.