Test cases for new InstanceNode.merge() method

[kwatsen]

Hi Lada,

There are two test cases:

• no-op: merges self into self, asserts no diff.
• simple: merges a tweaked-self into self, asserts expected diff.

Additionally, I integrated into my project and it seems to be working okay...albeit, the project test-cases just uses the Jukebox data-model and data-set from RFC 8040. :laugh:

It is still my opinion that:

• leaf-lists should be merged
• ordered-by user [leaf-]lists should be collated
• bits should be merged

That is, ConfD got it right, libyang got it half-right, YumaPro and now Yangson got it wrong.

If you're willing to update to match ConfD's approach, I'm happy to provide another PR. Otherwise, I concede that there is no definition, it's your project, and it doesn't seem to matter in practice (Andy said)...

Thanks for helping to make this update happen!

Best regards,
Kent

[llhotka]

Merged, thanks.

• leaf-lists should be merged

Yes, this is what sec. 7.7.9 in RFC 7950 says, so I changed the method behaviour for leaf-lists.

• ordered-by user [leaf-]lists should be collated

I think they are: remaining entries from value argument are appended at the end. Or do you mean something else by "collated"?

• bits should be merged

We could change this too, but sec. 7.6.7 says quite clearly:

• If the operation is "merge" or "replace", the node is created if it does not exist, and its value is set to the value found in the XML RPC data.