Hi,
"Xufeng Liu" <[email protected]> wrote:
> While realizing that the mechanism specified in the two drafts is very
> useful for modeling, I have a few questions that are common to both
> approaches:
>
>
>
> - How to mount one tree (but not other trees) from a module? For
> example, I may want to mount interfaces to one mounting point, but
> interfaces-state to a different mounting point.
Currently this is not possible. If you want to use such a mechanism
to mount selected subtrees from a module, it is not clear that this
will work - there might be references from subtree A to subtree B, so
if subtree B is not mounted, it is unclear how such references should
be handled.
> - When the mounting module is augmented by another module, what will
> happen to these augmentations after mounting? What will be the XPath to
> refer a node in the mounting module (from mounting module, and from mounted
> module)?
If the augmenting module it not mounted, nothing happens - it is just
not there. If the augmenting module is also mounted, it will be
augment the mounted module. For example, suppose ietf-interfaces and
ietf-ip are both mounted at /foo. The resulting tree would be:
+--rw foo
+--rw if:interfaces
+--rw if:interface* [name]
...
+--rw ip:ipv4!
> - When a mounting module is used to mount to a mounting-point in
> my-module, how can the system also expose the mounting model in the original
> form, i.e. at the root level?
The system exposes all "top-level" modules as usual - just list them
in the top-level YANG library.
/martin
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
