Re: [netmod] Does defining a feature require the module be implemented?

[Kent Watsen <[e-mail]>]

> Kent Watsen <[e-mail]> wrote:
>> YANG Doctors, 
>> 
>> 
>> Does "foo" need to be "implemented", in order for its feature to be
>> define?
>> 
>> 	module foo {
>> 	  yang-version 1.1;
>> 	  namespace "https://example.net/foo";
>> 	  prefix "f";
>> 
>> 	  feature foo-feature;
>> 
>>          ...
>> 	}
>> 
>> 
>> Specifically, using the previous YANG Library (RFC 7895), would this
>> be possible:
>> 
>>      {
>>        "name": "foo",
>>        "feature": [
>>          "foo-feature"
>>        ],
>>        "namespace": "https://example.net/foo",
>>        "conformance-type": "import"
>>      },
>> 
>> 
>> Or does "foo" also need to be "implemented", in order for its feature
>> to be defined?
>> 
>> 
>> PS: the answer to this impacts the "crypto-types and friends" drafts
>> in the NETCONF WG, where it is assumed (and various tools agreed, sans
>> a recent change in `yanglint`) that the implementation-status of a
>> module is orthogonal to what features supported.
> 
> Can you show a specific example where this is a problem?

The issue is that the modules need to be implemented, and thus any protocol-accessible nodes contained therein are enabled.   The issue was detected by an "imported" module containing a protocol-accessible leaf with "mandatory true".  Validation failed.  The fix was/is to define a special feature (e.g., protocol-accessible-nodes-supported) to put on the top-level nodes contained in the module.  This way the nodes aren't enabled simply because the module is implemented, as the feature must be defined as well.


> /martin

K.