The removal of SchemaNode.getPath() leaves a nasty gap in our integration with yang.data.api: there is no alternative to initialize a SchemaInferenceStack when have a YangInstanceIdentifier. DataSchemaContextTree can be used to locate the corresponding DataSchemaNode, but there is not sufficient information exposed to tie that lookup to a SchemaNodeIdentifier.Absolute path.

There is a secondary issue, which directly impacts NETCONF-818, which is that there is no avenue to build up a YangInstanceIdentifier and also build up the corresponding SchemaInferenceStack state.

Both of these facilities are available indirectly by means of DataSchemaContextNode.getDataSchemaNode().getPath() – an avenue not available since 8.0.0.

Provide extend DataSchemaContextNode to provide

DataSchemaContextNode<?> enterChild(SchemaInferenceStack stack, PathArgument child);

DataSchemaContextNode<?> enterChild(SchemaInferenceStack stack, QName child);

both of which work just as getChild(), except it also push the appropriate instruction into provided stack.

Also provide a DataSchemaContextTree.enterPath(YangInstanceIdentifier), which will provide a both a context node and an initialized SchemaInferenceStack.