Issue-77 Add support for shorhand/implicit case in augment statement …

[davidmat50]

Add support for shorhand/implicit case in augment statement when augment target is a choice

[dhubler]

https://datatracker.ietf.org/doc/html/rfc7950#section-7.9.2

As a shorthand, the "case" statement can be omitted if the branch contains a single "anydata", "anyxml", "choice", "container", "leaf", "list", or "leaf-list" statement. In this case, the case node still exists in the schema tree, and its identifier is the same as the identifier of the child node. Schema node identifiers (Section 6.5) MUST always explicitly include case node identifiers.

So it would be only .../y/q/q according to spec if case statement was missing.

[davidmat50]

Infact the fix i have provided in this pull request is for alllowing a shorhand case ( or implicit case) under a choice if the augment target was a choice. It was not about the validity of a augment target string which represents a case.

As i understand https://datatracker.ietf.org/doc/html/rfc7950#section-7.9.2, in relation with our context, it may apply as augment target should be explicitly have additional case node identifier (if the target was a case) (.../y/q/q) , since augment targets are represented as Schema node identifiers ( as per section-6.5 and section-14).

i hope you are also in sync with my above undertanding, Do you have an any suggestion/fix on how we can restrict/reject (.../y/q) for augment target if the target is a case, and only allow .../y/q/q .

[dhubler]

I see now, your solution looks good. It can be simplified a bit

_, targetIsChoice := target.(*Choice)

and then 2nd part:

if _, isCase := d.(*ChoiceCase); !isCase && targetIsChoice {
... rest is the same as you have it

Normally I would request a unit test as well but I'll add one after but i'll add one to TestParseSamples after this gets into master

[davidmat50]

Thank you for the response and the suggestions. I have done suggested changes and pushed. Please check.