• 1
  • 17
  • 18
  • 19(current)
  • 20
  • 21
  • 84
Linux OpenELEC Testbuilds for x86 (Kodi 17)
Just purely for WAF factor we tend to use OpenELEC... if it breaks switch it off and on Smile

Thanks for the builds latest Jarvis Beta 3 is playing very nicely especially with the tvheadend pvr and predictive channel switching!

It actually changes channel quicker than my 2012 not so smart anymore samsung tv...

Anyway... back to the game plan just wanted to say thank you for taking the time to do this Smile
(2015-11-25, 20:40)kenofs Wrote: Just to let you know,I have also found that I am unable to log in to my Pandora account with your builds, "cannot authenticate" message from Pandora while I can log in to Pandora with the latest builds from Her Fritsch.

Thanks for letting us know, but without a debug log (wiki) there's nothing any can do.

Have any of these builds worked? If so, which was the last that worked?
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2015-11-26, 00:42)C-Quel Wrote: Anyway... back to the game plan just wanted to say thank you for taking the time to do this Smile

Much appreciated. Smile
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2015-11-26, 05:50)Milhouse Wrote:
(2015-11-25, 20:40)kenofs Wrote: Just to let you know,I have also found that I am unable to log in to my Pandora account with your builds, "cannot authenticate" message from Pandora while I can log in to Pandora with the latest builds from Her Fritsch.

Thanks for letting us know, but without a debug log (wiki) there's nothing any can do.

Have any of these builds worked? If so, which was the last that worked?

All builds up to and including 1119 worked without connection issues of any kind, after that is when I had to go to the openelec settings and set the wireless network to active and that is when I couldn't log into Pandora. About that debug log.... I might have to get back to you later on that.Big Grin
New OpenELEC Jarvis build #1127: Generic / Nvidia_Legacy
(Supercedes previous build)

Code:
# uname -a
Linux OpenELEC 4.3.0 #1 SMP Fri Nov 27 23:06:04 GMT 2015 x86_64 GNU/Linux

# lsb_release
OpenELEC (Milhouse) - Version: devel-20151127230454-#1127-g3514983 [Build #1127]

Based on tip of OpenELEC master (35149839, changelog) and tip of XBMC master (3a87b76e, changelog) with the following modifications: Build Highlights:
  1. Fix TrueHD passthrough
  2. CNetwork - implement IPv6
  3. Web traversal security fix
  4. Add pvr.hdhomerun client
  5. MySQL 5.7.x support workarounds
  6. dcadec v0.1.0; libcec updates
Build Details:
  1. OpenELEC:
    • taglib: update bytevector.patch (PR:4467, 1 commit, 1 file changed)
    • Add pvr hdhomerun (PR:4469, 2 commits, 2 files changed)
  2. XBMC:
    • [PVR] Ensure that newly added channel groups are persisted, fixes trac 16365 (PR:8439, 1 commit, 1 file changed)
    • [Confluence] fix button alignment - fixes trac #16382 (PR:8423, 1 commit, 1 file changed)
    • [dbwrapers] Include database name in error message (PR:8421, 1 commit, 2 files changed)
    • Fix ticket 15727, SMS input not updating CGUIKeyboardGeneric (PR:7279, 1 commit, 1 file changed)
    • FIX: [droid] always handle media keys... (fixes #16388) (PR:8438, 1 commit, 1 file changed)
    • [PVR] Context Menu Crash Fix (trac #16229) (PR:8413, 1 commit, 5 files changed)
    • [addons] workaround for pausing repository updates during playback (PR:8407, 2 commits, 3 files changed)
    • Time sometimes moves backwards coming up from base libraries (PR:8426, 1 commit, 1 file changed)
    • [MySQL] Workaround MySQL 5.7.x query optimiser changes (PR:8393, 2 commits, 2 files changed)
    • FIX: allow to delete a default key binding (PR:8430, 1 commit, 4 files changed)
    • [PVR] Mark Created ChannnelGroups as Loaded on Persist (extra fix #16365) (PR:8445, 1 commit, 1 file changed)
  3. dcadec:
    • Rewrite multiple assignments as a single statement. (859fed50)
    • Restrict excessive core frame size workaround. (112ab42e)
    • Avoid using reserved suffix for typedefs. (7a152965)
    • Set SONAME for shared library, install a symlink. (95e447ab)
    • Rename version defines and stabilize the API. (feb6ceb4)
    • Release version 0.1.0. (2449e5d0)
    • Add link to lib with the lib*.so.[0-9] pattern (5bc00958)
  4. libcec:
    • Add support for configurable button repeats with libcec (PR:111, 5 commits, 7 files changed)
    • Fix for holding buttons on remotes of Philips TVs (PR:125, 1 commit, 1 file changed)
    • updated README.md with even more detailed compilation instructions (2e880768)
    • fixed: segfault when an empty EDID blob is passed to GetPhysicalAddressFromEDID() (b8218fdc)
    • changed: moved the buffer in CDRMEdidParser::GetPhysicalAddress() from the stack to heap (3561d5fc)
    • changed: added #include to make the IDE happy. not actually used when compiling (f816e4bb)
    • fixed: combo key (stop by default) in the EventGhost plugin. fixes #128 (a39a5599)
    • fixed: unused var (0eb74da4)
    • fixed: silence compiler warnings (d8795318)
    • changed: clean up device detection code a bit and replace winapi calls with CM_* calls. issue #130 (de14cd45)
    • changed: replaced lockdev by flock. closes #141 (ff86c07a)
  5. platform:
    • Simplify pkconfig file creation (PR:16, 1 commit, 3 files changed)
    • fix libdir for gentoo (PR:18, 1 commit, 1 file changed)
  6. pvr.dvblink:
    • Native Kodi timers support and some fixes (PR:31, 5 commits, 30 files changed)
  7. xbmc/master (FernetMenta):
    • New commits in this build:
      • AE: fix passthrough for truehd (10362ce2)
  8. Additional commits/pull requests/changes not yet merged upstream:
    • Exclude [env] libcec-01-add-repeating-keypress.patch: Already merged upstream
    • Updated: [env] patch: Enable pvr addons, disable pvr updates
    • Added: [pkg] PR:7030: CNetwork - implement IPv6
    • Added: [pkg] PR:8196: jsonrpc: expose profile directory in Profiles.GetCurrentProfile/GetProfiles.
    • Added: [pkg] PR:8446: [webserver] - fixed possible directory traversal bug due to insufficient url checking
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
@fritsch: I've noticed with this build (and also with #1125), that if a video is paused, when playback is resumed the audio is out of sync (sometimes quite significantly) - are you able to reproduce with your builds?

debug log:

02:26:46: Playback starts, audio is synchronised
02:30:01: Playback is paused for one minute
02:31:01: Playback is un-paused (resumed) - audio is no longer synchronised, there is now a noticeable lip-sync issue
02:31:57: Playback ends

If I stop the video, then resume playback from the resume point, audio will be perfectly synchronised.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
@Milhouse, I've noticed the same behavior with 1125.
(2015-11-24, 01:04)Karnis Wrote: I rolled back to 1119 but the post-reboot-OE message saying "waiting for network" for about 2 seconds issue persisted. Again, my network connection is unaffected.

Yes, it's coming from the network-online.service.

In a worst case you may see the following, which isn't very nice, though not a show-stopper (I only ever reboot when updating, as I'm able to suspend/resume reliably).

Image

It started with #1121 (which bumped systemd), and the reason it's also there in #1119 is because #1121 is adding a whole bunch of new persistent system.d services in /storage/.config/system.d:

/storage/.config/system.d in build #1119 (before booting #1121):
Code:
OpenELEC (unofficial) Version: devel-20151119035205-#1119-g785752e
OpenELEC git: 785752e5df9c5cf284e915d5916ebe30c290d515
OpenELEC:~ # ls -la /storage/.config/system.d
total 40
    32 -rw-r--r--    1 root     root         28816 Nov 28 04:03 README
     4 -rw-r--r--    1 root     root          1772 Nov 28 04:03 cifs.mount.sample
     4 -rw-r--r--    1 root     root          1729 Nov 28 04:03 nfs.mount.sample

/storage/.config/system.d in build #1121:
Code:
OpenELEC (unofficial) Version: devel-20151121220629-#1121-g297af74
OpenELEC git: 297af74bdf741f12076a4a503dd1acf1baa57399
OpenELEC:~ # ls -la /storage/.config/system.d
total 108
drwxr-xr-x   17 root     root          4096 Nov 28 04:07 .
drwxr-xr-x   17 root     root          4096 Nov 28 04:03 ..
-rw-r--r--    1 root     root         28816 Nov 28 04:03 README
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 basic.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 bluetooth.target.wants
-rw-r--r--    1 root     root          1772 Nov 28 04:03 cifs.mount.sample
lrwxrwxrwx    1 root     root            37 Nov 28 04:07 ctrl-alt-del.target -> /usr/lib/systemd/system/reboot.target
lrwxrwxrwx    1 root     root            46 Nov 28 04:07 dbus-fi.epitest.hostap.WPASupplicant.service -> /usr/lib/systemd/system/wpa_supplicant.service
lrwxrwxrwx    1 root     root            35 Nov 28 04:07 default.target -> /usr/lib/systemd/system/kodi.target
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 graphical.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 halt.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 installer.target.wants
lrwxrwxrwx    1 root     root            37 Nov 28 04:07 kbrequest.target -> /usr/lib/systemd/system/rescue.target
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 kodi.service.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 kodi.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 local-fs.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 multi-user.target.wants
-rw-r--r--    1 root     root          1729 Nov 28 04:03 nfs.mount.sample
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 poweroff.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 reboot.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 sleep.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 sockets.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 sysinit.target.wants
drwxr-xr-x    2 root     root          4096 Nov 28 04:07 textmode.target.wants

Consequently, rolling back to #1119 (or *any* systemd-based OE release) from #1121 means you still have all these extra system.d services starting, which appears to delay start up long enough for the "waiting on Network to come online ..." mesage to appear.

I'll raise it as an issue on OpenELEC github.

Edit: Issue #4470 on OpenELEC github. Wonder if this might also be responsible for the WiFi issues, as it would explain why the WiFi issues continue after rolling back to a version prior to #1121.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2015-11-23, 15:38)kenofs Wrote:
(2015-11-23, 08:25)Milhouse Wrote:
(2015-11-23, 06:13)kenofs Wrote: Thanks, but I found the answer in the "openelec settings". I have to manually reset the wireless network to active, never had to before.

Hmm, good to know - presumably you only had to do this once and not every boot, so all is good now?

At a guess, maybe connman-1.30 is seeing your network hardware with a slightly different label than before, so "forgot" the old settings created with connman-1.23.

Even switching back to the latest Fritsch build (20151118) I have to reset every time now.

The issue in the previous comment would neatly explain your new WiFi behaviour after booting build #1121.

If you:
Code:
cd /storage/.config/system.d && rm -fr *.wants *.target dbus-fi*
then upgrade to @fritsch build 20151118 (or my build #1119) I'd be pretty confident your WiFi will start connecting automatically again - can you confirm?
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Okay, with build 1121, I have tried a mp4 video with mp3 audio, either over the network or with the file in local storage on the PC.

In both cases, the video plays but there is no sound. Logs here:

https://www.dropbox.com/s/07bga1ll8y2b1q...6.zip?dl=0

Hopefully this excludes network as the problem.

Updated to 1127 and the problem remains. Logs from playing the file from local storage, not over the network.

https://www.dropbox.com/s/srx4ipu9axjovi...0.zip?dl=0

Link to file. Please let me know once you have downloaded and tested this, as I am only providing this for test purposes. I would never wish to infringe copyright, but better to provide an actual file I have a problem with.

https://www.dropbox.com/s/k83zwp9495xu38...s.mp4?dl=0

Note that this problem is the same with ALL files encoded in a similar way, and they play fine in Windows, Kodi 15.2 and everything else I have tried them in.

Best wishes,

Mark
(2015-11-28, 12:59)mcelliott Wrote: Okay, with build 1121, I have tried a mp4 video with mp3 audio, either over the network or with the file in local storage on the PC.

In both cases, the video plays but there is no sound. Logs here:

https://www.dropbox.com/s/07bga1ll8y2b1q...6.zip?dl=0

Hopefully this excludes network as the problem.

Updated to 1127 and the problem remains. Logs from playing the file from local storage, not over the network.

https://www.dropbox.com/s/srx4ipu9axjovi...0.zip?dl=0

Link to file. Please let me know once you have downloaded and tested this, as I am only providing this for test purposes. I would never wish to infringe copyright, but better to provide an actual file I have a problem with.

https://www.dropbox.com/s/k83zwp9495xu38...s.mp4?dl=0

Note that this problem is the same with ALL files encoded in a similar way, and they play fine in Windows, Kodi 15.2 and everything else I have tried them in.

Best wishes,

Mark

To follow up on my own post, it seems that older videos with the same settings play fine on Kodi 16 nightly builds on Windows (easier for me to test without annoying the family) but newer ones are really pixellated and have no sound.

It could be that the problem files were encoded with Handbrake 0.10.2 (newest version) and the ones that play fine were encoded with an earlier version. I don't know how easy it is to check this, as I can't remember exactly when I updated.

However, they all play perfectly in Kodi 15.2, so whether or not there are problems in the files, this is significant regression since 15.2.

Mark
It's not the first time that handbrake breaks mp3 ... but - the file is playing perfectly in mplayer / mpv and so on ... so we should look why kodi fails. I asked our lead developer (videoplayer) to take a look if time becomes available.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
(2015-11-28, 14:31)fritsch Wrote: It's not the first time that handbrake breaks mp3 ... but - the file is playing perfectly in mplayer / mpv and so on ... so we should look why kodi fails. I asked our lead developer (videoplayer) to take a look if time becomes available.

Thanks. To be fair, with the windows test builds (latest as of today) the picture is so pixellated that even if there was sound, the video would be completely unwatchable.
(2015-11-28, 15:48)mcelliott Wrote:
(2015-11-28, 14:31)fritsch Wrote: It's not the first time that handbrake breaks mp3 ... but - the file is playing perfectly in mplayer / mpv and so on ... so we should look why kodi fails. I asked our lead developer (videoplayer) to take a look if time becomes available.

Thanks. To be fair, with the windows test builds (latest as of today) the picture is so pixellated that even if there was sound, the video would be completely unwatchable.


I know - I always test what millhouse ships and then check with current upstream master :-) - never the less the complete sample works perfectly in other players - so - curious. Every such sample improves our player - so thanks much for it. Though I cannot promise any timeline.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
(2015-11-28, 16:30)fritsch Wrote:
(2015-11-28, 15:48)mcelliott Wrote:
(2015-11-28, 14:31)fritsch Wrote: It's not the first time that handbrake breaks mp3 ... but - the file is playing perfectly in mplayer / mpv and so on ... so we should look why kodi fails. I asked our lead developer (videoplayer) to take a look if time becomes available.

Thanks. To be fair, with the windows test builds (latest as of today) the picture is so pixellated that even if there was sound, the video would be completely unwatchable.


I know - I always test what millhouse ships and then check with current upstream master :-) - never the less the complete sample works perfectly in other players - so - curious. Every such sample improves our player - so thanks much for it. Though I cannot promise any timeline.

No worries. I can keep with 15.2 for regular usage and test new builds as and when for improvement.

I appreciate your help!
  • 1
  • 17
  • 18
  • 19(current)
  • 20
  • 21
  • 84

Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for x86 (Kodi 17)4