View Full Version : Refresh entire library?
midgetspy
2009-02-23, 22:50
Is there a way to get XBMC to pick up changes made to thumbnails & nfos that are stored locally (as tbn/nfo with my movies)? My library was scanned from those local files originally but I've since modified many of them. Do I have a choice other than deleting my library and re-scanning it or manually refreshing every single movie?
No, there's no other way then those 2.
If you find another way let me know. ;) I've been just tinkering and refreshing Movies as I fix them. Slow and painful, but so far its been worth it.
jmarshall
2009-02-24, 00:50
If you modify a .nfo file then it should be picked up via Update Library. Is it not?
There should be an option to refresh compleet library.
Miss that in XBMC. Is needed if you use an external scraper,
midgetspy
2009-02-24, 01:10
If you modify a .nfo file then it should be picked up via Update Library. Is it not?
No, updating library does nothing if I modify an .nfo.
From experience it appears that when you hit Update Library it looks to see if a file is there and if it is it continues on, but if it doesn't it says "Oh Look something new lets add it."
Yeah, it might be a huge hassle to to check every nfo, tbn and such for changes within the file itself. I can't imagine that this would be done during a standard "Scan for new Content." Maybe during a "Verify Library Content," but not the regular "Update Library."
midgetspy
2009-02-24, 04:26
Well I think it's supposed to just hash the directory's contents and see if it's changed, but probably the hash just checks names not sizes/dates.
A function to refresh library would be nice, I'm tired of deleting mine over and over, hehe.
Well I think it's supposed to just hash the directory's contents and see if it's changed, but probably the hash just checks names not sizes/dates.
A function to refresh library would be nice, I'm tired of deleting mine over and over, hehe.
I was actually playing around with it last night, it only hashes the path to the media file or the folder! not the file content.
and if the scrap fails for a file it will not even try it again since the hash for that folder path is already in the db.
that's what i was able to figure out based on the debug logs.
I would defiantly welcome a refresh db. or re scrap folder.
I agree it shouldn't be part of the background update, maybe something you call from the context menu or another option beside clean library.
jmarshall
2009-02-24, 04:45
It currently hashes:
1. The files in the folder.
2. The size of the files in the folder.
3. The date of the files in the folder.
If the hash changes, this triggers a rescan of that folder.
The main problem is we don't currently know (from the db side of things) where the info came from, and we don't want to hit the web just because the hash changed. Think about those who have 1000 movies in one folder for instance. Ideally the scanner should take any _local_ information and update the db with that, but I suspect it's just not that easy to do.
And yes, there are many problems with the scanner, such as it not correctly detecting when a fetch fails and updating the path hash anyway, in addition to being quite inefficient in some places. I'd prefer to fix it via rewrite than try and workaround certain issues in the meantime. Unfortunately I have no time to commit to that at this point.
Cheers,
Jonathan
For reference:
http://xbmc.org/trac/ticket/4632
http://xbmc.org/wiki/?title=New_Video_Scanner_Ideas