[SoWErA]

I also prefered the old method it allowed me to keep track of which movies/tv episodes hadn't had it's thumbnail added yet now I am gonna have to write each of them down manually in a document and keep track of them that way

[Montellese]

See PR1034

[kiboy6]

Hi Montellese and everyone.

First of all....thanks so much for this feature. For me, it's a massive improvement.

Sorry if I'm being thick, but while the new sort (I'm running Frodo) sorts the movies correctly on initial import of the whole source, any movie for which i refresh the info still jumps to the top of the list.

Is there any way to stop this happening?

Thanks

(Sorry if i missed the solution earlier in the thread - looked but couldn't find it and from what I can understand of the pull request, the advanced settings wouldn't help with this either)

[Montellese]

Hm sounds like it uses the current date when refreshing an item instead of using mtime or ctime. I guess you don't have <dateadded>0</dateadded> in your advancedsettings.xml. The only other time when XBMC falls back to the current date is when mtime and/or ctime is invalid (i.e. 0 or in the future). Does it happen with every movie in your library or just a specific one (which might have a bad/invalid date)?

[kiboy6]

It definitely happens with every movie in my library.

No, I don't have <dateadded>0</dateadded> in my advancedsettings.xml . Should I?

Thanks so much

[Montellese]

No <dateadded>0</dateadded> would force XBMC to use the current date as the value for dateadded when adding a video to the library. Sounds very odd, this is the first report of such a behaviour that came to my attention so far.

[kiboy6]

Hmmm strange...should I be running a newer nightly version than the May 2nd frodo release?

If that's not the reason here's a debug log.

In this log i refresh the info for movie "Footloose" and sure enough it moves right to the top my date-added sorted list after refresh.

Thanks a lot

[lastimp]

No <dateadded>0</dateadded> would force XBMC to use the current date as the value for dateadded when adding a video to the library. Sounds very odd, this is the first report of such a behaviour that came to my attention so far.

I guess you didn't see my post in the linux support forum http://forum.xbmc.org/showthread.php?tid=134014 I wasn't sure where to post it. <dateadded> doesn't work properly since the merge the way I understand that it is supposed to.

[Montellese]

Hmmm strange...should I be running a newer nightly version than the May 2nd frodo release?

If that's not the reason here's a debug log.

In this log i refresh the info for movie "Footloose" and sure enough it moves right to the top my date-added sorted list after refresh.

May 2nd should be more than ok. Unfortunately the log won't say anything about the dateadded being used.

I guess you didn't see my post in the linux support forum http://forum.xbmc.org/showthread.php?tid=134014 I wasn't sure where to post it. <dateadded> doesn't work properly since the merge the way I understand that it is supposed to.

I have to admit I rarely look into the linux (and never into the osx) support forum because I mainly code on win32. How do you know the "creation date" of that movie you mentioned? Linux doesn't keep track of a file's creation date, only of "access date", "modification date" and "change date". But with <dateadded>1</dateadded> it should fall back to using ctime (change date) if mtime is newer than the current date.

I think I should add some logs in the case XBMC runs into a date that is newer than the current date (which obviously is a problem with the filesystem).

[Montellese]

Can either of you guys compile XBMC yourself? I've written a patch (see http://pastebin.com/K6idH1ty) which will write a log message to XBMC's log for every file with an invalid mtime/ctime. This should also make it easier for users to determine if they have files with "bad" information. If not I'll just commit it to master as it will benefit other users as well.

[kiboy6]

Sorry, not able to compile XBMC myself. Let me know if i can do anything else to help troubleshoot.

[lastimp]

Can either of you guys compile XBMC yourself? I've written a patch (see http://pastebin.com/K6idH1ty) which will write a log message to XBMC's log for every file with an invalid mtime/ctime. This should also make it easier for users to determine if they have files with "bad" information. If not I'll just commit it to master as it will benefit other users as well.

I compile my own bulid every few days but have never done a build with a patch so I am not sure how to, I am looking into it now and will report back. Also I didn't realize thats the changed date, if I do a stat on the files in linux it says changed date but if I do properties on the same file from windows it says creation date (My files reside on my Freenas server.)

[Montellese]

After you pulled the latest commits from git you can run "git am path/to/the/patch" and the patch should be applied on top of all the other ones. To get rid of it (after you've tested it) you can run "git reset --hard HEAD~1". Thanks for giving it a try.

Yeah like I said linux and windows interprete ctime differently.

[lastimp]

After you pulled the latest commits from git you can run "git am path/to/the/patch" and the patch should be applied on top of all the other ones. To get rid of it (after you've tested it) you can run "git reset --hard HEAD~1". Thanks for giving it a try.

Yeah like I said linux and windows interprete ctime differently.

Just posted the relevant info of one failed movie on http://pastebin.com/U5LCKNzv if you need more let me know. Looks to me like it tried to use the ctime and then fellback to the current time. Also here is the stat for that file:

File: `The Book of Eli.mkv'
Size: 7041004012 Blocks: 13744408 IO Block: 16384 regular file
Device: 13h/19d Inode: 291259 Links: 1
Access: (0644/-rw-r--r--) Uid: ( 1000/ lastimp) Gid: ( 1000/ lastimp)
Access: 2012-06-16 10:57:55.522864100 -0700
Modify: 2016-11-09 15:33:11.000000000 -0800
Change: 2010-05-19 04:30:33.326882600 -0700

[Martijn]

Modify time in 2016?