Jumpy Plugin (26 October 2016)

[infidel]

Adding resolve_at_playback=false in jumpy.conf completely fixed the playback problem

See if this changes anything when resolve_at_playback=true:

xbmc-test-130813.zip

unzip to plugins/jumpy/xbmc - for testing only

(19.62 KiB) Downloaded 856 times

[Casper]

Thanks for the fixes, I've installed update2-a3, commented out youtube from ffmpeg.webfilters and applied the Jumpy patch 130813. FFmpeg is working great.

After one test I couldn't get resolve_at_playback=true to stick (videos were still behind a folder) and it wasn't until I did a clean reinstall that I caught update.jar was still in use. Processes javaw.exe and python.exe still running seems to indicate something had gone wrong but I didn't notice an explicit error making this a rather nasty error to catch. Once that had been addressed and everything reinstalled unfortunately the playback problem remains.

I've been sitting on another unusual bug I hope you can help me with. Jumpy is causing an old PMS listing bug to reappear for my Panasonic TV (XBMC & Samsung AllShare are unaffected) which is detailed on the PMS forum along with solution. The bug does not allow scrolling past 20 items, on the 21st item the list goes blank and the log file explodes, only going back a level restores the list. To address this they added a new renderer variable 'AnalizeFolderAllFiles'/ 'AnalyzeAllBeforeFolderShown' with a default value of true for maximum compatibility, it now seems to have been changed to 'CreateDLNATreeFaster' with a default value of false. When I set this to true there is no listing at all with UMS which is in accordance with the original bug report. When set to false, UMS file and web feed lists work perfectly indicating they fixed the problem but Jumpy causes it to return.


This happens with all the versions of UMS and Jumpy I have tried.

It looks like PANASONIC require real TotalMatches value.

This quote from the PMS forum seems to be the key, is Jumpy reporting correct TotalMatches?

[infidel]

After one test... [adventures with javaw process termination and file locking] ...the playback problem remains.

I'm taking all this to mean "basically, test-130813 made no difference"?

Jumpy is causing an old PMS listing bug to reappear for my Panasonic TV ... which is detailed on the PMS forum along with solution. The bug does not allow scrolling past 20 items, on the 21st item the list goes blank and the log file explodes,

Thx for the as usual superbly documented bug report . Does it happen if you disable bookmarks in jumpy.conf:

Code: Select all

bookmarks=false

[Casper]

Thx for the as usual superbly documented bug report

After struggling with PlayOn, TVersity, Plex & Serviio, doing a little troubleshooting for the beast that is UMS+Jumpy+XBMC is well worth the reward!

Further testing indicates the 130813 patch has fixed the playback issue with resolve_at_playback=true on Panasonic renderer. Testing with XBMC renderer on the same machine still has the playback issue like with Jumpy 0.2.7.

Turning Bookmarks off removes the 'Add Bookmark' entry in video listings but the 20 list limit bug still persists. All folders are affected including the root folder, folders without 'Add Bookmarks' entry and the Bookmarks folder itself.

[infidel]

the beast that is UMS+Jumpy+XBMC

Sounds a tad hellish .

the 130813 patch has fixed the playback issue with resolve_at_playback=true on Panasonic renderer. Testing with XBMC renderer on the same machine still has the playback issue

Good, any advice on how to recreate the error in xbmc?

All folders are affected including the root folder, folders without 'Add Bookmarks' entry and the Bookmarks folder itself

Can you please generate some debug info in a test install of a3/0.2.8:

• remove/rename update.jar and plugins/jumpy-0.2.8.jar.
• unzip these to ums top folder (note: for testing only):

  jumpy-a3-testpan2.zip

  (82.59 KiB) Downloaded 1035 times

  ums-a3-update-testpan2.zip

  (71.55 KiB) Downloaded 945 times

• send Panasonic debug.log(s) for WEB.conf vs. jumpy (on TED again would be good).

EDIT: replaced attachments: testpan2.

[Luxfero]

Trying to install jumpy for the sole purpose of running a few scripts. When I try the quick setup (jumpy-0.2.8) on windows (pms-1.90.1) from "C:\Program Files\PS3 Media Server\plugins", I get the following massage

Error: can't find debug.log.

Please verify that this installer is located in your PMS 'plugins' folder.

Press any key to continue . . .

Is that the wrong folder? Any help is appreciated.

[infidel]

When I try the quick setup (jumpy-0.2.8) on windows (pms-1.90.1) from "C:\Program Files\PS3 Media Server\plugins", I get the following massage

Error: can't find debug.log.

The location of debug.log has changed in PMS as of 1.90.0. If you temporarily put a copy of C:\ProgramData\PMS\debug.log in C:\Program Files\PS3 Media Server the installer will be happy .

See also the requirements section in the docs if all you need is a minimal setup.

[Luxfero]

The location of debug.log has changed in PMS as of 1.90.0. If you temporarily put a copy of C:\ProgramData\PMS\debug.log in C:\Program Files\PS3 Media Server the installer will be happy .

See also the requirements section in the docs if all you need is a minimal setup.

Thank you, that did it. Earlier I tried moving the Debug.log to the plugins directory, forgot the Debug.log original location. I also got the winp.dll load problem but I read the previous posts. One last question how do I skip packages for a minimal setup, I only saw the phython and xbmc external installer windows. For now I'll play around with the full install, Thanks.

[infidel]

how do I skip packages for a minimal setup, I only saw the phython and xbmc external installer windows

The installer is really a (ridiculously long) batch file with no provision to skip the rtmpdump, phantomjs, imagemagick convert, and youtube-dl parts, but

• you can purge these from the win32 folder afterwards (NOTE: don't purge win32/convert.exe if your scripts are using the jumpy API).
• Since python and py4j are now in place, you can update jumpy in the future by just unpacking the zip version.
• You'll probably want to remove/rename plugins/jumpy/xbmc.py and also disable any unused scripts in jumpy-scripts.ini.

[Casper]

any advice on how to recreate the error in xbmc?

There's nothing unusual about my setup, it's just a clean install of XBMC 12.2 Frodo on Windows. Tested on a Windows 7 64bit local machine (AMD CPU/GPU) and a Windows XP 32bit networked machine (Intel CPU/GPU) with the same result.

I did a quick test of different XBMC versions (all with resolve_at_playback=true) XBMC 11.0, 12.0, 12.2, 13.0 Alpha 3, 4, 5 & 6 and only XBCM 11.0 works. It is notably different in not showing the video duration (both for Jumpy and UMS feeds) while all the others show duration which is only a few seconds and either doesn't update or updates slowly resulting in playback catching up and halting. In several versions it crashes at this point or when attempting to stop. When resolve_at_playback=false the duration does update and playback works. So maybe there is some incompatibility in the way duration is being reported?

XBMC apparently had (has?) a problem with getting the duration of mpeg2 files.
http://forum.xbmc.org/showthread.php?tid=70183&page=4

A separate issue I mentioned before with XBMC always listing alphabetically with resolve_at_playback=false was because XBMC does 'Default' sorting (no sorting) for directories but does 'Title' sorting (alphabetical) for files. Default sorting is unavailable for files so this is strictly an XBMC issue.

I only use XBMC for testing of course, such a closed loop is quite an abomination.

Can you please generate some debug info in a test install of a3/0.2.8:

I noticed you added 'getDLNAResources' debug output. With WEB the start value changes correctly 0-20-40-60-80-100 as the list is scrolled through. With Jumpy, scrolling down changes it to 20 only momentarily and then back to 0 which is repeated continuously (list is blank). When scrolling up the list is meant to wrap around to the bottom, the start value maintains a value of 100 which is also repeated continuously (list is blank). Hopefully this is some help identifying the cause.