[mossywell]

12.0

Hi all, I've read this: http://wiki.xbmc.org/index.php?title=NFO_files/movies and it says "This is useful for instance, if you want the entry in your database to have another name than the one officially provided by the scraper". So, if I have an NFO with this (using a Star Trek movie as an example):
<movie>
<title>TEST</title>
</movie>
http://www.imdb.co.uk/title/tt0079945/

I think that it should use the URL, including all the movie metadata and override just the title. But what actually happens is that it ignores the URL, so all I get is a movie called "TEST" with no metadata. Is this intended behaviour?

[artrafael]

What did you name the NFO file?
Is your movie file placed in its own folder with the same name?

[mossywell]

movie.nfo (it's a DVD scrape --> VOB and IFO files). The nfo is of course being read because the movie is called TEST.(I tried moving it into the VIDEO_TS and calling it VIDEO_TS.nfo) but in general I had less success with that than using a movie.nfo file.

[jmarshall]

Yes, you definitely want it in the root, not inside the VIDEO_TS folder.

The Debug Log while refreshing the movie should show that a combined nfo file was found, and you should see it querying themoviedb. Pastebin it if it doesn't make things clear.

Cheers,
Jonathan

[mossywell]

Cheers guys for rapid responses. Debug says "Found matching Full NFO file" and points to the correct NFO. So, it looks like it doesn't think it is a combined NFO which explains the behaviour. I'll have a play with the file format (ANSI, UNICODE etc etc) and see if I can get it to find a combined NFO.

[mossywell]

No luck, so I had a quick look at the code and can't see how it'll ever return a combined NFO (though it was only a quick look over lunch):
CheckForNFOFile calls m_nfoReader.Create, and the return result says it is a FULL_NFO enum. m_nfoReader is an instantiation of CNfoFile. The Create method says return m_scurl.m_url.empty() ? FULL_NFO : COMBINED_NFO where m_scurl is a member variable. Without finding out what happens to m_nfoReader prior to this, it's difficult to tell what is going on, but I'm guessing that for some reason, m_scurl.m_url is empty. If I get lunch tomorrow, I might dig a bit deeper.

[jmarshall]

My guess is it's the .uk link that isn't being picked up by any of the scrapers as a valid URL that it can use.

[mossywell]

Thanks for the reply. Changing to .com didn't fix it. However, having spent a bit more time looking at the code I think a) it's a, er, flaw in the logic that handles the nfo files (well, possibly 2 actually) and b) I should be able to take a branch and fix it. My only consideration is finding the time. :-\
First things first just for completeness, I've upgraded to 12.1. No difference as expected. I'll document the debug log outputs depending on the nfo inputs to kick things off.

[kaffekask]

I know it worked in 12.0 since this was one of the things i tested when it was released.

I wasn't sure if it worked in 12.1 so I had to test it and it works for me, I get a movie named TEST and poster and metadata from "Star Trek: The Motion Picture" with your example (only change is imdb.com)

Code:

<movie>
  <title>TEST</title>
</movie>
http://www.imdb.com/title/tt0079945/

Then I tried with another movie, and with co.uk and then I just got a movie called TEST2 with no metadata, so co.uk is the reason.

Code:

<movie>
  <title>TEST2</title>
</movie>
http://www.imdb.co.uk/title/tt0079946/

Then I did this:
Moved the movie to another directory (not scanned by XBMC)
Clean Library
Move back the movie
Changed to imdb.com in movie.nfo
Update Library
Success, a movie named TEST2 with poster and metadata for the movie Starcrash.

Edit:
Just tried again ith forcing an error in a movie by using imdb.co.uk, changing to imdb.com and and doing a refresh on the movie and choosing to use local data will do trick also.

[mossywell]

I'm not sure why changing the URL from a .co.uk to a .com did not work first time round because I finally got around to a full set of tests. Looks like .com does work after all. Cheers all. (I still think the NFO categorisation could do with some work. )

0. XBMC debug logging enabled.
1. Remove previous sucessful tests from library.
2. Clean library.
3. Stop XBMC.
4. Delete old log files.
5. Make changes to file system (to allow detection to work).
6. Set up movie.nfo file.
7. Start XBMC.
8. Scan movies.
9. Check if movies added.
10. Stop XBMC.
11. Check log file.



1.
<movie>
<title>TEST123</title>
</movie>
Movie added with title TEST123 only.
VideoInfoScanner: Found matching Full NFO file

2.
http://www.imdb.com/title/tt0079945/
Movie added will all details.
scraper: NfoUrl returned
VideoInfoScanner: Found matching URL NFO file

3.
<movie>
<title>TEST123</title>
</movie>
http://www.imdb.com/title/tt0079945/
Movie added will all details but with override on title.
scraper: NfoUrl returned
VideoInfoScanner: Found matching Mixed NFO file

4.
http://www.imdb.co.uk/title/tt0079945/
No movie added.
ADDON::CScraper::FindMovie
No VideoInfoScanner

5.
<movie>
<title>TEST123</title>
</movie>
http://www.imdb.co.uk/title/tt0079945/
Movie added with title TEST123 only.
VideoInfoScanner: Found matching Full NFO file

[jmarshall]

Check the regexps that the scrapers are using, and feel free to recommend better ones.

[mossywell]

Yup - will do. And some stuff about the CNfoFile::NFOResult enumeration (and logging thereof) as well.
Cheers for the help.