[NinjaMonkeyUK]

I thought I had found a bug in the way Kodi handles importing of NFO data, but now I think I just misread the information on this page:

http://kodi.wiki/view/NFO_files (section 2.3): 

By default the xml entry is added to the scraped entry in the database and may result in some duplicates. If you want to completely override the scraped result you need to set a clear attribute for the relevant xml tag in the .nfo file.

The first quoted sentence implies that the NFO entry is added to an already-scraped database entry but this surely is the wrong way around.  Usually an NFO is created to aid in scraping, not add to the database after the initial scrape.  The issue I am having is that if I set a custom <genre> tag in my NFO, by default I would expect Kodi to add this custom genre to the movie in the database, then go out to TMDB and pull in all other genres listed against that movie.  I use a combination metadata NFO with link to TMDB to pull in the rest of the metadata.

The second quoted sentence above, implies this should be the case but unfortunately if I add a single <genre> tag to the NFO, that gets added to the database but the genres on TMDB are ignored completely.  If this is not a bug and is by design, then what is the point of the <genre clear="true"> tag?

The way I see it should work is this:

1. The <genre>Custom</genre> in the NFO should be added to the database.
   
2. The scraper then adds any other genres it finds to the database alongside the custom genre.
   
3. If the user wants to override the above functionality and use only custom genre tags (i.e. not scrape any from TMDB) then they should use the clear="true" attribute.
   

Maybe I'm just reading the wiki wrong or not seeing the bigger picture, but I can't see the point of the clear attribute if it is not to be used in this way.