First, thanks a lot for your interest.
Regarding to Kodi-metadata requests.The info I'm able to give is that Kodi, when working as DLNA/upnp server is able to answer to other Kodi clients (accessng Kodi server just like a upnp/DLNA client, not using mysql or file sharing) lots of metadata. That's totally logical, obviously, nothing special.
But JRiver when playing the role as DLNA server is also able to answer certain metadata to Kodi without issues. At least: Title, Genre(s) and Director. That's which makes me think that it's just a matter of some work to admit a few more formated requests and adding new metadata structures to the already answered metadata by JRiver the DLNA server, since, as I've said some already work.
Since Kodi is open source you can check this file, where the routines to get that inf from XML structures are written. You can find that genre(s) is there, also director, hence, I think with a little more debug info It shouldn't be very difficult to add premiered, studio, country, actors, and maybe more metadata in the way Kodi could parse without issues.
https://github.com/xbmc/xbmc/blob/master/xbmc/video/VideoInfoTag.cppTo be honest I don't know the low level details, I mean I don't know if DLNA metadata request is just one XML answer/file for each movie/file, or Kodi launches one DLNA request to the server for each kind of "possible" metadata available. I guess JRives accomplishes this document (or higher version
http://upnp.org/resources/documents/UPnP_AV_tutorial_July2014.pdf) in which standards are defined, and high chances Kodi has added/changed some of those, but "respecting" the protocol basis.
This seems the core Kodi file with the upnp routines to populate metadata:
https://github.com/xbmc/xbmc/blob/master/xbmc/network/upnp/UPnPInternal.cppAs far as I'm able to understand it seems that tag info is loaded from one xml file through URL (but I could be totally wrong since my cpp is long forgotten), in such case, maybe adding a few more xml tags to the standard xml tag answser could work, just a guess.
And regarding to the dummy DLNA control entries.Just add elements (like view/branches) with no real access to any media item or folder, just to launch/trigger some basic maintenance utils like: shutdown, restart, run-autoimport now,... and maybe more. When a dlna client opens or renders those items (which could be shown as folders, not single items) the server would trigger that actions and would show in t he renderer the content of the "false folder", showing the results "items" like...."Ok" ... "Shutting down in 5 seconds" or messages like those. It could be more than enought to perform "easy" tasks withouth having to use a browser. In my opinion it could be really useful from TV's, cell phones, etc...
Thanks again.