Login at Kodi Home

blubserl · 2014-05-06, 11:23

hi!

i originally updated from frodo to gotham rc1 (openelec).

in general everything works very smooth, but i always get a crash & automatic xbmc restart after a few seconds when switching to fullscreen, when listening to music. as long as i don't switch to fullscreen mode i have no problem with music at all. i also have no problems with fullscreen video playback.

what i have tried so far:

- disable addons: cu-lrc lyrics, artist slideshow
- switching to default confluence skin
- changing audio device
- update to gotham final (openelec 4.0 final)

the crash ALWAYS happens after some seconds.

i gathered full debug logs, which are available here (please notice, that these logs are from the rc1 build).
unfortunately i didn't find the time to post until now, but yesterday i updated to the final and still have the exactly same behaviour. team xbmc is too lightning fast for me Big Grin

anyway: i looked through the logs, but couldn't see anything obvious.

maybe someone can help?

thanks and greetings!

FernetMenta · 2014-05-06, 11:38

please post your logs to xbmclogs.com or pastebin. nobody wants to download zip files to access a log.

blubserl · 2014-05-06, 11:49

(2014-05-06, 11:38)FernetMenta Wrote: please post your logs to xbmclogs.com or pastebin. nobody wants to download zip files to access a log.

ok...sorry. i pasted the content of all 10 generated log files here:

http://xbmclogs.com/show.php?id=194671

blubserl · (This post was last modified: 2014-05-08, 08:16 by blubserl.)

in the meantime i narrowed the problem down to the xbmc lcd proc plugin. as soon as i disable this plugin i get no more crashes.

in case it is important i pasted my lcd.xml configuration to http://xbmclogs.com/show.php?id=196055
im using the imon vfd in the origenae s10v htpc case.

FernetMenta · 2014-05-08, 08:38

thanks, I notified the dev in charge of this.

blubserl · 2014-05-08, 08:47

from the xbmc.log:

18:41:49 T:2972621632 WARNING: ### [XBMC LCDproc] - Empty Mode 6, consider checking LCD.xml
18:41:49 T:2972621632 WARNING: ### [XBMC LCDproc] - Empty Mode 7, consider checking LCD.xml
18:41:49 T:2972621632 WARNING: ### [XBMC LCDproc] - Empty Mode 8, consider checking LCD.xml

it seems that my lcd.xml which was working fine in frodo is missing some important entries for gotham - will have a look at it and compare it to the default one and report back afterwards.

**herrnst** · (This post was last modified: 2014-05-08, 08:59 by herrnst.)

(2014-05-08, 08:14)blubserl Wrote: in the meantime i narrowed the problem down to the xbmc lcd proc plugin. as soon as i disable this plugin i get no more crashes.

Did I get that right that the only thing required to trigger this is to play some music, toggle fullscreen (e.g. by hitting the TAB key) and then wait a few moments?

(2014-05-08, 08:47)blubserl Wrote: from the xbmc.log:

18:41:49 T:2972621632 WARNING: ### [XBMC LCDproc] - Empty Mode 6, consider checking LCD.xml
18:41:49 T:2972621632 WARNING: ### [XBMC LCDproc] - Empty Mode 7, consider checking LCD.xml
18:41:49 T:2972621632 WARNING: ### [XBMC LCDproc] - Empty Mode 8, consider checking LCD.xml

That means you're missing configuration blocks for xbelaunch, pvrtv and pvrradio, but that shouldn't matter for your particular music playback crash case.

nst

blubserl · 2014-05-08, 09:10

(2014-05-08, 08:55)herrnst Wrote: Did I get that right that the only thing required to trigger this is to play some music, toggle fullscreen (e.g. by hitting the TAB key) and then wait a few moments?

yes, exactly. it takes a maximum of ~40 seconds.

(2014-05-08, 08:55)herrnst Wrote: That means you're missing configuration blocks for xbelaunch, pvrtv and pvrradio, but that shouldn't matter for your particular music playback crash case.

i added the blocks and also renamed my lcd.xml to lcd.xml.backup, so that only default values are used --> the crash still occurs.

**herrnst** · 2014-05-08, 09:18

(2014-05-08, 09:10)blubserl Wrote:
(2014-05-08, 08:55)herrnst Wrote: Did I get that right that the only thing required to trigger this is to play some music, toggle fullscreen (e.g. by hitting the TAB key) and then wait a few moments?
yes, exactly. it takes a maximum of ~40 seconds.

Ok, will try to reproduce when I get home in the evening (while this doesn't help ATM, I never had anything crashy related to Python, addons/script.lcdproc or playback in a long time though, especially on my HTPC box with the iMon in it).

blubserl · 2014-05-08, 09:48

thanks for looking into this. especially strange is, that it only happens in music/fullscreen and not in any other mode. let me know if i can help in any way...

**herrnst** · (This post was last modified: 2014-05-08, 22:32 by herrnst.)

(2014-05-08, 09:48)blubserl Wrote: thanks for looking into this. especially strange is, that it only happens in music/fullscreen and not in any other mode. let me know if i can help in any way...

Well, for now, not-so-good-news at least for you I think...

No matter how hard I try, I'm unable to reproduce any sort of crash caused by any action of script.xbmc.lcdproc. Tried the following:

- Start XBMC
- Go to music, any source, any file, start playback
- Upon start, press TAB to activate fullscreen and vis
- Watch LCDproc writing stuff onto any configured LCDd driver, for as long as I liked and as many files as I wanted in a row

With two self-compiled builds (Gentoo, Gotham, 64bit, GCC 4.8.2 - the other on Ubuntu 12.04, 64bit, GCC 4.6.3 which is my HTPC box), one official build (OSX 32bit, Gotham, on 10.7), and a freshly+vanilla OpenELEC 4.0.0 (2014-05-05) 32bit Live USB Stick. LCDd was served locally as well as remotely via SSH forwarding, tried xosd and my "hardware" SoundGraph iMON LCD in the HTPC. Tested all different ways, music plays perfectly on fullscreen/vis enabled.

EDIT: Even tried on Win32 (by tunneling TCP 13666 to the LCD box), plays like a champ in fullscreen.

Sorry, really need any occasion to reproduce this so I can look at stack traces and all that stuff... The only thing I currently can think of you can try is to also make an OpenELEC live stick with a clean storage/userdata tree and check if that also crashes for you (instructions to create such a stick can be found on the OpenELEC wiki).

Regards,
nst

blubserl · 2014-05-09, 11:48

(2014-05-08, 22:11)herrnst Wrote: The only thing I currently can think of you can try is to also make an OpenELEC live stick with a clean storage/userdata tree and check if that also crashes for you (instructions to create such a stick can be found on the OpenELEC wiki).

good suggestion...thank you for investing your time in this. i will setup a clean OpenELEC usb stick and check if it behaves the same and report back!

blubserl · 2014-05-09, 13:21

also happens on a clean openelec 4.0 live stick (OpenELEC Stable - Generic i386 Version:4.0.0).
the current logs: http://xbmclogs.com/show.php?id=196872

as a next step i will install a clean openelec 3.2.4 live stick and check there again.

**herrnst** · 2014-05-09, 13:55

(2014-05-09, 13:21)blubserl Wrote: also happens on a clean openelec 4.0 live stick (OpenELEC Stable - Generic i386 Version:4.0.0).
the current logs: http://xbmclogs.com/show.php?id=196872

Code:
52:42 T:3062933888  NOTICE: Using Release XBMC x32 build, compiled May  5 2014 by GCC 4.9.0 for x86 Linux 32-bit 3.14.2

52:42 T:3062933888  NOTICE: Running on Linux 32-bit (OpenELEC (official) - Version: 4.0.0, 3.14.2 i686)

52:42 T:3062933888  NOTICE: Host CPU: AMD Athlon(tm) II X2 250 Processor, 2 cores available

That's the same build I tried yesterday, but on a Sandy Bridge i7 processor, worked like a charm (booted, skipped the initial setup dialogs, went to music, there to a folder containing audio on one of the local disks, played back and pressed TAB, and let it sit there for almost half an hour). Out of curiosity, why don't you use a 64bit build on that processor? In other words, can you try with a 64bit build?

nst

blubserl · 2014-05-09, 16:01

(2014-05-09, 13:55)herrnst Wrote: Out of curiosity, why don't you use a 64bit build on that processor? In other words, can you try with a 64bit build?

to be honest - i have no idea why i didn't install the 64bit build in the first place. Blush

but i did so now and it also crashes. it's worth noting that it took a significantly longer time to do so (~20 minutes).
i already thought the problem was solved...