Login at Kodi Home

locomot1f · 2014-09-19, 17:41

@richardk,

yup... you are correct. The Daily Show did not finish all the way through. Oddly enough, The Colbert Report did finish, again.

richardk · 2014-09-20, 04:27

(2014-09-19, 17:41)locomot1f Wrote: @richardk,

yup... you are correct. The Daily Show did not finish all the way through. Oddly enough, The Colbert Report did finish, again.

Just watched the September 18 episode of The Colbert Report, and it stopped after the second segment. (I was able to resume it.) So I'm still having the problem with both shows.

In case it matters, I'm in the US, and I'm using the 2500 bitrate.

slices · (This post was last modified: 2014-09-20, 13:51 by slices.)

I have a complete rewrite of the way we handle the local server and subtitles that should fix this, just checked in to github.

richardk · 2014-09-20, 19:34

(2014-09-20, 11:20)slices Wrote: I have a complete rewrite of the way we handle the local server and subtitles that should fix this, just checked in to github.

Sorry to be the bearer of bad news, but....the Github version seems to have made the segment transition problem worse.

We just tried to watch the September 18 episode of The Daily Show. Playback stopped after each segment. (Before, the transition between the first and second segments seemed to work.) Resuming the programs caused playback to start beyond the start of the segment, and it was not possible to resume to the short final segment at all.

slices · 2014-09-20, 19:47

No idea what's going on. Can you confirm that line 25 of _main_viacom.py is 'class XBMCPlayer( xbmc.Player ):', and then post a log of the issue.

richardk · 2014-09-20, 20:10

(2014-09-20, 19:47)slices Wrote: No idea what's going on. Can you confirm that line 25 of _main_viacom.py is 'class XBMCPlayer( xbmc.Player ):', and then post a log of the issue.

Line 25 of that file is what it should be, so I think I'm using the correct version.

A logfile is here:

XBMC logfile for segment transition glitch

Thanks, slices, for all the work on this great addon!

Let me know if there's anything else I can do to help.

slices · 2014-09-20, 22:29

Goodish news is that I was able to reproduce this, although only once. I've added some additional logging around the stop/end events. Also try a different bandwidth it may be there is an issue with the stream you're using. The select quality option from the context menu is probably the easiest way of doing this since it will show you all the streams.

From what I saw in my own testing there is an interesting 'feature' here, in that normally with a stacked url if one segment fails if moves onto the next. It looks like a failed playback is a stop not an end so that kills the server and stops all playback. I don't think this is a problem since we're treating all parts of the stack as a single stream, but will result in some different behavior.

richardk · 2014-09-20, 23:27

(2014-09-20, 22:29)slices Wrote: Goodish news is that I was able to reproduce this, although only once. I've added some additional logging around the stop/end events. Also try a different bandwidth it may be there is an issue with the stream you're using. The select quality option from the context menu is probably the easiest way of doing this since it will show you all the streams.

Trying that now, and will report results.

I forgot to mention that I'm getting a "script failed" on _proxy.py with the current Git version. That should show up in the logfile I posted earlier.

richardk · 2014-09-20, 23:46

(2014-09-20, 23:27)richardk Wrote:
(2014-09-20, 22:29)slices Wrote: Goodish news is that I was able to reproduce this, although only once. I've added some additional logging around the stop/end events. Also try a different bandwidth it may be there is an issue with the stream you're using. The select quality option from the context menu is probably the easiest way of doing this since it will show you all the streams.

Trying that now, and will report results.

I forgot to mention that I'm getting a "script failed" on _proxy.py with the current Git version. That should show up in the logfile I posted earlier.

Ok, I've tried all of the bandwidth settings, down to the lowest one, and nothing changed. Playback still stops after the first segment in all cases.

In case it matters, I'm using OpenELEC 4.6.1.

Hope this info helps.

slices · (This post was last modified: 2014-09-21, 19:02 by slices.)

Try the version in github with the extra logging. Need to see which event is triggering the stop server call. I think what is happening is a difference between xbmc builds. On my windows system, I only see stop events when I push the stop button, and end playback events at the end of segments.

Need to see what else is going on so I can confirm this, and work out when a stop isn't really a stop.

richardk · 2014-09-21, 22:04

(2014-09-21, 08:57)slices Wrote: Try the version in github with the extra logging. Need to see which event is triggering the stop server call. I think what is happening is a difference between xbmc builds. On my windows system, I only see stop events when I push the stop button, and end playback events at the end of segments.

Need to see what else is going on so I can confirm this, and work out when a stop isn't really a stop.

Here you go:

XBMC log -- with extra logging

Playback stopped after the second segment of The Daily Show, September 4. Bandwidth setting = 2500, OpenELEC 4.6.1 Generic (on ION hardware platform).

slices · 2014-09-21, 22:51

Not sure what's going on in that log. I can see the call to stop the server _connection :: getURL :: url = http://localhost:12345/stop
but no event logging message before it.

There may be an issue where it's holding on to a cached version of the class, although I would expect multiple messages. The latest from github should fix this, will also need to restart xbmc,

richardk · 2014-09-21, 22:57

(2014-09-21, 22:51)slices Wrote: Not sure what's going on in that log. I can see the call to stop the server _connection :: getURL :: url = http://localhost:12345/stop
but no event logging message before it.

There may be an issue where it's holding on to a cached version of the class, although I would expect multiple messages. The latest from github should fix this, will also need to restart xbmc,

I'm getting "installation failed" on the current github version.

richardk · 2014-09-22, 20:15

(2014-09-21, 22:57)richardk Wrote:
(2014-09-21, 22:51)slices Wrote: Not sure what's going on in that log. I can see the call to stop the server _connection :: getURL :: url = http://localhost:12345/stop
but no event logging message before it.

There may be an issue where it's holding on to a cached version of the class, although I would expect multiple messages. The latest from github should fix this, will also need to restart xbmc,

I'm getting "installation failed" on the current github version.

Re-downloaded this morning, and got the github version to install.

On the September 18 episode of Colbert, playback stopped after the second segment, at 11:08. It resumed properly, and the other transitions worked properly.

Probably not enough data to conclude much....

slices · 2014-09-22, 20:22

Are there any event messages in the log