Due to tuning issues, I've accumulated a few zero length recordings. Trying to play back these results in VDR crashing.
Does anyone know a solution to this? It doesn't seem to happen to vanilla VDR, only the yavdr patched VDR.
Due to tuning issues, I've accumulated a few zero length recordings. Trying to play back these results in VDR crashing.
Does anyone know a solution to this? It doesn't seem to happen to vanilla VDR, only the yavdr patched VDR.
does that happen with extrecmenu or with original menu or with both ? Default is extrecmenu in yavdr - you can change the replacement of the original one in System -> Settings -> Plugins -> extrecmenu. In any case a backtrace would be nice - Installing package vdr-dbg and setting DAEMON=/usr/bin/vdr-dbg in /etc/default/vdr should generate a crashlog in /var/log/vdr/
Tried disabling all extension in /etc/vdr/plugns/order.conf (except for the xine plugin), and it still craches. Attached crashlog.
Can you disable the extrecmenu plugin and try it again ? This plugin is definitely active.
P.S. Can you tar this empty recording directory and attach it here ?
Here's a run without extrecmenu. Will create a tar file as well.
I send your crashlog to our fellow "mini73" he's aware of the problem and this thread.
But as first guess seems a pointer in dvbplayer becomes invalid. So, I quickly checked which patches do effect "dvbplayer.c":
All these patches have already on board of the previous VDR version 1.7.21.
So, please stay tuned.
Regards
fnu
menu.c
http://paste.ubuntu.com/883027/
5165
5131/5134
cReplayControl::~cReplayControl()
{
Hide();
cStatus::MsgReplaying(this, NULL, fileName, false);
Stop();
if (currentReplayControl == this)
currentReplayControl = NULL;
}
dvbplayer.c
http://paste.ubuntu.com/883030/
323/329
cDvbPlayer::~cDvbPlayer()
{
Save();
Detach();
delete readFrame; // might not have been stored in the buffer in Action()
#ifdef USE_LIVEBUFFER
delete indexFile;
#else
delete index;
#endif /*USE_LIVEBUFFER*/
delete fileName;
delete ringBuffer;
}
Alles anzeigen
963
player.c
http://paste.ubuntu.com/883031/
106 -
void cControl::Shutdown(void)
{
cMutexLock MutexLock(&mutex);
cControl *c = control; // avoids recursions
control = NULL;
delete c;
}
vdr.c
http://paste.ubuntu.com/883033/
1241
alles gepatcht
keine ahnung ob das hilft .....
ist das hier wichtig ?:
Core was generated by `/usr/bin/vdr-dbg --lirc=/var/run/lirc/lircd -v /srv/vdr/video.00 -c /var/lib/vd'.
Core was generated by `/usr/bin/vdr-dbg --lirc=/var/run/lirc/lircd -v /srv/vdr/video.00 -c /var/lib/vdr'.
Before trying anything other I would remove the LIVEBUFFER patch ...
Yes, I would not deny the possibilty, that there's an interaction between that exterecmenu & livebuffer, if enabled. Both are my best friends ...
But, again the version has been proven for 2 month and nobody did realize an issue like that.
Something came into my mind, do you use the double XX plugin, everybody is using and nobody does talk about? If yes, do you have proper version running and could you disable that for a test?
Cheers
fnu
What is the double XX plugin? I don't have any custom plugins running.
I will dig more into this when the VDR system is available for debugging, later today.
Yes, I would not deny the possibilty, that there's an interaction between that exterecmenu & livebuffer, if enabled. Both are my best friends ...
But, again the version has been proven for 2 month and nobody did realize an issue like that.
Usually no one will prove a version with empty recordings
From a programmers point of view I would guess that an empty recording causes the destructor to be called before all initialization routines are finished, and this can result in deleting non existing objects ...
Usually no one will prove a version with empty recording
Indeed. Funny, recently I did ask one of the fellows when to hack this problem might happen ... I didn't got it ...
Cheers
fnu
The empty recordings are caused by tuning issues with nova-t 500 cards. There shouldn't be empty recordings in a perfect world of course, but then vanilla VDR doesn't choke on them either.
The empty recordings are caused by tuning issues with nova-t 500 cards. There shouldn't be empty recordings in a perfect world of course, but then vanilla VDR doesn't choke on them either.
Guess what, I could proof this, since I'm pretty much the last yaVDR fellow using exactly this VDR version. How can I simulate it? Start a recording and stopping it immediately?
cheers
fnu
Hi,
Start a recording an stopping it immediately?
Replace the TS-files of a recording with an empty 00001.ts and remove the index?
Lars.
Here's an empty recording tar file
Ok, I have some good news and some not so good.
The good, I could reproduce that issue in my installation, the bad, some fellows could also reproduce it with VDR 1.7.26 w/o Live-Buffer Patch, some not, still with other patches.
Since it is not 100% clear if LB is the root cause, I'll change nothing. Since zero lenght recordings itself is more or less a fault for itself, we mark this as a known error for now, until we'll find the real cause.
Since I already have something new w/o LiveBuffer in planning/testing, I would rather push 0.4 & 0.3.2 to this version with a fix.
Regards
fnu
Ok, is there a yavdr package that provides plain vanilla unpatched VDR?
Ok, is there a yavdr package that provides plain vanilla unpatched VDR?
No, what sence would make a package nobody from the team wants to use?
Gerald
Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!