lyd_new_path or lyd_new_path2

[choppsv1]

I want to get back the node that a path refers to when using one of the lyd_new_path* functions. I believe that normally this is actually lyd_new_path2 as it returns the "last node created" as opposed ot the "first node created" which if a bunch of parents are created to get to a leaf or presence container or list_entry woudl be the top-most container node not the target.

So next the question is will the "last node created" actually be the target node? What happens if I am creating a list entry. The path will have the key as a predicate, technically I think the "last node created" is probably going to be the key leaf node and not the target of the path (the parent list node of that key leaf)?

i.e., say I do lyd_new_path2() on /foo/bar[name="baz"]. Will I get back the specific bar list node, or will I get back the baz key leaf node?

It would be good if the docs/API could make this more clear too.

Thanks,
Chris.

[michalvasko]

Fine, docs updated.

[choppsv1]

For others who land here, the new text is: "* @param[out] new_node Optional last node created (in case of a list, it is the list instance, not any of the key nodes)."

[choppsv1]

Are there other cases where the actual last node isn't the target (e.g., when creating a presence container with children that have default values)?

If it's the case that the returned node is always the "target of the xpath" or "the node the xpath describes" or something like that, and not "the last node created", then maybe there's a better way to say it than "last node created" and then a bunch of exceptions?

In particular as a user of this API, I specifically want the node that I'm asking for with the xpath, I don't care how libyang is going about constructing the result.

Right now my code calls lyd_find_path() immediately after calling lyd_new_path2() b/c I don't actually know whether I'm getting that node that I asked for back, so I call lyd_find_path() to make sure.

[michalvasko]

I was thinking about rewording it already so fine, changed.