In early new implementation of the JcrDeliveryEndpoint, depth is set to 0 by default, while node-type configuration only specifies which node-types are "servable".
But setting depth = 2 for a page might return sub-pages as well.

In particular for the query, collection-returning endpoint method: if a page and one of its sub-pages match the given query, we don't want the sub-page to be part of "result" of the parent one, thus avoiding duplicate info. One approach for this is:

Once a "top-result" item is resolved from the JCR query, exclude its sub-nodes of similar type, but include others. e.g. a page would be returned with all its areas and components, but without its sub-pages.

Additional notes:

• With this, we could even afford to set the depth default value as "unlimited".
• Maybe folders need to be further excluded from results (content delivery should not serve a whole tree, and dump the heavy-lifting onto the client)
• Proposal to indicate in the response that some "recursive" child-nodes are hidden, is welcome
• WorkspaceParameters can still be changed