XBMC shutdown error [not the same as recently closed bug]
#46
defender013 Wrote:why is this particular error harmless? I mean just based on the log, it almost looks like xbmc is trying to control the graphics to do...something? So when it fails, how come its not a big deal?

XBMC just retries after the error, and it obviously succeeds otherwise it wouldn't work :-) I think the error happens the first time XBMC tries to render the screen, and it seems to happen only once.

JR
Reply
#47
ah that makes so much sense...

i was wondering how i could possibly see anything on screen if rendering failed...
i will try out the fixed patch tonight

JR, just wanted to thank you again for seeing this through, i mean we're on page 5 of a thread where we are the only 2 posting ..lol.

thanks a lot for your help
Reply
#48
hi JR!

looks like the problem may have been fixed! see the logging below...

20:56:31 T:3496 M:942567424 DEBUG: CWinEventsWin32::WndProcWindow lost focus
20:56:31 T:3496 M:942567424 DEBUG: D3DERR_DEVICELOST
20:56:31 T:3496 M:942546944 DEBUG: CApplication::Render: g_Windowing.BeginRender() returned false, calling unlock
20:56:31 T:3496 M:942542848 DEBUG: CAnnouncementManager - Announcement: ApplicationStop from xbmc
20:56:31 T:3904 M:942505984 NOTICE: CThread:ConfusedtaticThread: pThread->Process() finished
20:56:31 T:3904 M:942505984 NOTICE: Previous line repeats 1 times.
20:56:31 T:3904 M:942505984 NOTICE: CThread:ConfusedtaticThread: Calling pThread->OnExit()
20:56:31 T:3904 M:942501888 NOTICE: Previous line repeats 1 times.
20:56:31 T:3904 M:942501888 NOTICE: CThread:ConfusedtaticThread: pThread->OnExit() finished
20:56:31 T:3904 M:942501888 NOTICE: Previous line repeats 1 times.
20:56:31 T:3904 M:942501888 NOTICE: CThread:ConfusedtaticThread: Calling pThread->IsAutoDelete()
20:56:31 T:3904 M:942497792 NOTICE: Previous line repeats 1 times.
20:56:31 T:3904 M:942497792 DEBUG: Thread 3904 terminating (autodelete)
20:56:31 T:3900 M:942485504 DEBUG: Thread 3900 terminating (autodelete)
20:56:31 T:3904 M:942485504 NOTICE: CThread:ConfusedtaticThread: Calling _endthreadex(123)
20:56:31 T:2588 M:942592000 NOTICE: Previous line repeats 1 times.
20:56:31 T:2588 M:942592000 DEBUG: FileCurl::Close(04C2FCE4) http://feeds.feedburner.com/xbmc
20:56:31 T:2588 M:942592000 DEBUG: Got rss feed: http://feeds.feedburner.com/xbmc
20:56:31 T:2588 M:942632960 DEBUG: RSS feed encoding: UTF-8
20:56:31 T:2588 M:942628864 DEBUG: Parsed rss feed: http://feeds.feedburner.com/xbmc
20:56:31 T:2588 M:942624768 NOTICE: CRssReader:Tonguerocess: 1036: calling UpdateObserver
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: entered UpdateObserver
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: about to call getfeed
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: feed.size = 461
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: about to call g_graphicsContext.Lock
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: m_pObserver = 44343000
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: about to call OnFeedUpdate
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: about to call g_graphicsContext.Unlock
20:56:31 T:2588 M:942624768 NOTICE: CRssReader::UpdateObserver: finished
20:56:31 T:2588 M:942624768 NOTICE: CRssReader:Tonguerocess: 1036: finished


it has not failed my usual power on/power off testing. I'll now use it normally and see if anything pops up. I"ll let you know

thanks a bunch!

btw... do you think the contextLock debug messages should just be included by default? Maybe even with a argument passed in to show which thread called it? Seems like it would save a lot of work in the future if similar problems pop up again.
Reply
#49
At last! :-)

Annoyingly I've just missed getting the fix into v10.1, but I can do a custom build of v10.1 for anyone affected by the bug.

Deciding what to log and what not to log is always a tricky issue. I think the logging in XBMC is OK as it is, apart possibly from adding some extra logging to identify which thread is which. Yes, I did have to add lots of extra logging to identify this bug, but that's only because I can't reproduce the bug on my system. Normally I'd just run XBMC in a debugger and this makes it easy to find bugs.

Anyhow, thanks again for all the effort you've put in. It's great that we've got a result :-)

JR
Reply
#50
hi JR,

a custom build would be really nice! since i have no idea when 10.2 is going to come out...

btw, do i need to file this as a bug? or are you going to take care of that already?
Reply
#51
jhsrennie Wrote:At last! :-)

Annoyingly I've just missed getting the fix into v10.1, but I can do a custom build of v10.1 for anyone affected by the bug.

Deciding what to log and what not to log is always a tricky issue. I think the logging in XBMC is OK as it is, apart possibly from adding some extra logging to identify which thread is which. Yes, I did have to add lots of extra logging to identify this bug, but that's only because I can't reproduce the bug on my system. Normally I'd just run XBMC in a debugger and this makes it easy to find bugs.

Anyhow, thanks again for all the effort you've put in. It's great that we've got a result :-)

JR

This is great, I use an external player and I had to switch the settings on my full-screen mode to get around the XBMC failure to close problem. That works okay but somtimes XBMC loses focus. Looks like I will be able to go back to real full screen mode now.

If I disable the RSS feed add-on until 10.2 comes out will that solve my shutdown problem?

Thanks again.
Reply
#52
i think that really depends. the crash on exit i think can happen for a great variety of reasons, it just happened that in my case it was the Rss that failed out.

doesnt hurt to give it a try though. OR, if you're using dharma 10.0, i would suggest giving JR's custom *.exe a try also to see if your problem goes away.
Reply
#53
@Tharagleb: I'll put a bug fixed version (just the XBMC.exe) on my web site in a day or two.

I've filed a bug report (http://trac.xbmc.org/ticket/11302) and this will be fixed in the trunk build over the weekend. From Monday the fix will be in the nightly builds, though note that the nightly builds are unsupported and you use them at your peril.

JR
Reply
#54
I've compiled a version of the XBMC v10.1 XBMC.exe with the bug fix. You can download it from:

http://swarchive.ratsauce.co.uk/xbmc/XBM...wnHang.zip

JR
Reply
#55
I have exactly the same problem. Deactivating the RSS-Feed does not work for me.

Would it be possible, that you compile a fix included version also for linux? Would be awesome, when I can solve the problem. After every use I have to reset my HTPC and thats not the best way to restart your system Wink

Edit:
I just noticed, that the fix will be in the nightly builds over the weekend. That should work for me, too Smile
Thanks!
Reply
#56
psuch Wrote:I just noticed, that the fix will be in the nightly builds over the weekend. That should work for me, too Smile
Thanks!

Ironically, after all that work, the nightly builds didn't have the bug because the locking mechanism has changed since the Dharma release.

I didn't think the bug applied to the Linux/Live builds as it was being caused by a DirectX error and obviously that doesn't apply to Linux. Anyhow try the nightly builds as they definitely don't suffer from this bug, though there may of course be other bugs :-)

JR
Reply
#57
Hmm, OK, that is surely possible. I'll try it out. Thank you very much Nod
Reply
#58
jhsrennie Wrote:I've compiled a version of the XBMC v10.1 XBMC.exe with the bug fix. You can download it from:

http://swarchive.ratsauce.co.uk/xbmc/XBM...wnHang.zip

JR

Thanks!
Reply

Logout Mark Read Team Forum Stats Members Help
XBMC shutdown error [not the same as recently closed bug]0