[bocki]

Hey there,

from what I remember, I used to be able to put a movie.nfo file in one of my folders, containing an URL to a different scraper.

For example:

My movie library is set to use "TMDB.org". I own a few Austrian movies that are barely popular outside our own borders, so there's no useful info in either TMDB or IMDB. For these cases I want to use OFDB.
I put a link to "http://www.ofdb.de/film/8845,M%C3%BCllers-B%C3%BCro" in the movie.nfo and let the scan do its work.

Expected behaviour: Scanning picks up the ofdb URL and gets all the information from ofdb.
Actual behaviour: .nfo is ignored. (according to VideoInfoScanner.cpp output should at least be "VideoInfoScanner: Found matching URL NFO file: %s")

If, however, I use an URL matching to the NFOUrl of the scraper of the folder, ie http://www.themoviedb.org/movie/15480, it picks the movie up just fine and also acknowledges the existence of an nfo file in the folder ("Found matching URL NFO")

http://pastebin.com/5UiQM2MM - Debug log. Watch the sections around line 323 and 359, while movie #1 gets picked up fine using tmdb, #2's movie.nfo is completely ignored.

Any suggestions would be greatly appreciated

Thanks,
-- bocki

[mkortstiege]

Looks to me as if you missed to check "Use folder names for lookups" while setting up the content.

[bocki]

Actually I just found the solution: Manually editing the language in the addon.xml for OFDB to "en", my .nfo gets picked up fine.

Is this expected behaviour? I can personally live with multiple languages in my library, especially if I explicitly override the default

[mkortstiege]

Jep, this is kind-off the expected behavior. The intention was to ensure that you do not end up with multiple languages when using url nfos. I think this logic has to be re-factored, patches welcome

[bocki]

Btw, my fix works only if the fallback setting in guisettings.xml is enabled.

Well, my C++ skills are very limited - but the magic very likely happens somewhere around line 117 in NfoFile.cpp.
I set a bounty of 1 giant cookie on a pretty solution to this small issue.

[bocki]

i put together a quick patch (3 lines is all my c++ knowledge combined!):
http://pastebin.com/pwtVKz0L

with this applied, there's a new option allowdifflanguages in guisettings.xml.

this only applies to .nfo files, and basically allows _any_ installed scraper to fetch infos from the given url in the nfo file.

issues:
the order in which scrapers are tried can't be set, so if you got 2 scrapers with an nfourl allowing imdb.com for example, there's a chance the result won't be what you expected.
genres get parsed in the language of the scraper, so you'll have a weird mix of languages - setting the genre in the nfo just adds (instead of replacing) it to the list of scraped genres, but at least allows for smart playlists...
(these 2 need _actual_ coding skills to be fixed )