33 replies to this topic

[christophecvr]

As a small extra I also set the location on stb for dvd.cuts file saving in case the media location is not writable,

to /home/root/.dvdcuts folder.

 

All patches are send as push request :

 

https://github.com/O...enigma2/pull/23

[Taapat]

There is a small error in assigning value to m_resume in seekTo: https://github.com/O...enigma2/pull/25

[christophecvr]

There is a small error in assigning value to m_resume in seekTo: https://github.com/O...enigma2/pull/25

Thank's I missed that one.

[christophecvr]

at dev's,

 

I noticed that in the gst-1 branch the libdreamdvd is not updated yet. In master well.

[MastaG]

yeah... merging master into gst-1 would be nice :-)

[christophecvr]

I did found out that with some dvd's there was again a problem. (it wen't ok cause I made a typo)

 

Now it is solved.

 

All dvd's do resume. (at least all mine)

 

Also by a lot off dvd's (not all) I can when they are well playing ask menu back set new audio and or sub settings and request the resume dvd(if avbl by the dvd) It will resume with my new settings on the right location.

 

Patch send as pull request :

https://github.com/O...i/enigma2/pulls

[littlesat]

Thanks pushed...

[christophecvr]

For users from dvd services on vuduo2 /solo2. The LCD4Linux does bug the stb's by sometimes asking audiotrack info in a never ending loop. Wheater plugin is on or off does not mather. As soon the plugin is installed it does that.

 

Also some iptv(or free streams) could cause sometimes the box to be locked .

 

With the LCD4linux removed also this never happens. (but here I can not confirm that the LCD4inux is the cause )

 

What happens well the box lock's You see the sandkeeper and the only option are to power off stb and repower.

or log in trough telnet/ssh and killall -9 enigma2 .

 

I found the code in LCD4Linux which cause'd this bug.

 

End users solution.

 

1) Remove the LCD4Linux plugin from stb but yess gues a lot will keep that .

 

2) use the code where only the responsible code has been commented :

branch master.

https://github.com/c...opy/tree/master

 

 

 

3) use the code where audio track info is removed out off code (I commented it only for now)

    This is the best , since by some media when there is audiotrackinfo this is passed to enigma2 by a pointer to base class audio.

    This pointer is assigned to early when the audio info from media is not known yet. By the LCD4Linux.

https://github.com/c...py/tree/testing

 

 

 

 

In :

Edited by christophecvr, 15 August 2015 - 08:41.

[joergm6]

I do not know if it is the problem, but here is a loop with no index increase. At this point, there would never be an end. Maybe try.

add idx+=1 at the end of while loop.

							L4logE("Audio %d" % idx, self.Laudiodescription)
							idx+=1

 

[joergm6]

hmm ... no later edit passible  

and audio will be tested for the presence! (therefore can not be the problem)

"if audio:"

[christophecvr]

 

I do not know if it is the problem, but here is a loop with no index increase. At this point, there would never be an end. Maybe try.

add idx+=1 at the end of while loop.

							L4logE("Audio %d" % idx, self.Laudiodescription)
							idx+=1

 

 

Yes checking this loop out there is never an increase, You're right. Whenever an condition occure that trigger this loop it will always block the stb, there is no escape out of that loop.

[christophecvr]

Here how i discovered that it was that line which caused the deadlock .

 

 

Attached Files

•  enigma2_crash_log_stuckloop.txt   17.01KB   5 downloads

[christophecvr]

It's confirmet with the last commit I just add the

idx += 1

Never  stuck loop.

 

https://github.com/c...py/tree/testing

[joergm6]

yes "he" ask only 0 ..... "getTrackInfo: id=0" . Thanks to find out