Rob Shakir

Thanks for the contribution -- please find some comments inline.

Taken from #95: The issue is supporting the `uses .... { when ... }` and `augment ... { when ... }` statements. When we convert to entries today, we lose...

On the latter, the reason to do this on the child entries is we know whether these are subject to `when` conditions -- we could use this information to track...

I don't think we should make `goyang` aware of the datatree. AFAICS, the approach that we've generally taken up to now is such that we make the `Entry` something that...

Absolutely, I agree that the `check_condition` function needs to be supplied by the datastore implementation. My argument was just that as long as we expose, as public functions or struct...

I think this depends on what we decide to store in the contents of the when statement. Today, you're right that the data tree implementation needs to do some additional...

@sbarguil I think that this was the PR that you mentioned where there needed to be some discussion of what we did in IS-IS, is that right? Could you remind...

w.r.t the fact that IS-IS has both `SIMPLE_KEY` within a keychain and `TEXT`, this was intentional since we noted that it wasn't possible to map some implementations to the keychain...

Sorry - meant to add a top-level comment. Are these being removed because they are harmful, or because they aren't consistently used through every enumeration? If it's the latter, if...

I am not clear I agree that there are changes that are backwards incompatible that didn't have a major revision bump for models that had a major revision version greater...