• 1
  • 65
  • 66
  • 67(current)
  • 68
  • 69
  • 495
v18 LibreELEC Testbuilds for RaspberryPi (Kodi 18.0)
(2017-02-14, 21:05)Milhouse Wrote:
(2017-02-14, 17:58)esco Wrote: Thank you Milhouse. I have tested 4.10-rc8 build #0213x build and unfortunatelly the same behavior occurs.

Can you provide log information? "journalctl -a", "dmesg" and "kodi.log"?

Is #0101 the last build that has this working mux, or just the last working build you've used? Can you identify when this mux problem started in the normal daily builds?

We switched to using the media_build package in #0108 so you might be able to confirm that is when the problem started (ie. #0107 works, #0108 doesn't).

#0107 is fine, no issues at all. With #0108 I am able to scan muxes (however some "poll timeout" errors appear) but unable to watch TV - the log is full of MPEG2AUDIO/ H264 "Invalid start code" and "Continuity counter" errors. Am I supposed to send the logs with the latest #____x build?
Thank you.
Hi popcornmix and milhouse

the problem with to fast scrolling with mouse in a folder is always on all Kodi Krypton builds for RPi and win32 present.
Nobody of the "skin - developer" have fixed them - making a little youtube-video and have postet that in the relevant thread some times ago.

New for my is, that i have now problems in Kodi-Menue to scroll in Addons:

Scrolling down and up (vertical) Video-/Musik-/Programm-Addons and Pictures will only work sometimes/temporarily with mouse.
It's to slow and choppy ant it needs a long time so to select something.

...horizontal scrolling in a selectet addon works perfect.

it aint't fun to me - I can't help myself.
Take 80% of speed from the folder-scrolling and add this to addonscrolling :-)

Please help us to resolve this or i have to buy additional two logitech k400 keyboards, because all arrow-keys will work. :-)


Best regards
Harald
(2017-02-17, 08:44)garbear Wrote:
(2017-02-15, 23:05)linkinsoldier Wrote: I think I found a bug in the controller mapping: I mapped 4 controller with the rpi3 via Bluetooth. I started with controller 1 and have been going on until the last (controller 4) was connected via Bluetooth. After this I played and at some point I shutdown the rpi. At some other day I took the old controller 3 and connected / started it (and only this controller). Funny enough the ps3 controller (Nr.3) showed player 1 (correctly!) and I could play nes and game boy games... But when I started a snes game nothing happened when I pressed the buttons start and select (exiting the game via PS button works)... Restarting it gets me the same situation, however when I reconnect the old controller (Nr. 1) I can play as player 1 (although it says player 2 on the controller). The emulator is pocketsnes.

How can I support further?

This sounds like a bug. Can you post a debug log? Kodi logs every button pressed and the resulting actions they map to, so I can usually see what goes wrong.

(2017-02-16, 21:09)linkinsoldier Wrote: Using bnes as emulator fixes this, so I guess emulator bug?

There are seven SNES emulators: http://kodi.wiki/view/Game_add-ons#Libretro_cores . Which ones work? Which ones have problems?

(2017-02-16, 21:09)linkinsoldier Wrote: does this help? if not, can i send you the zip file via email?

Yes, this is helpful. I've been trying to fix these remaining input bugs for the last year. Can you give me steps so that I can reproduce the problem?

Its pretty reproducable with the description above and the game super Mario all-stars on pocketsnes but I will give further details on weekend. The input on the bnes emulator works (but is too slow on rpi3), pocketsnes has the issue with the input but is working great otherwise!

In which one of the debug logs can I look for your information? I reproduced it twice yesterday but couldn't find anything regarding button press... Did I maybe enable the wrong debug log for this test?
#0216 broke my CEC again, have to reboot Pi after tv been shut down, with #0215 its ok.
(2017-02-17, 15:42)doldi Wrote: the problem with to fast scrolling with mouse in a folder is always on all Kodi Krypton builds for RPi and win32 present.
Nobody of the "skin - developer" have fixed them - making a little youtube-video and have postet that in the relevant thread some times ago.

If it's a general kodi issue that effect win32 then I'd suggest you open a trac ticket. It is not something we will be fixing here.
(2017-02-17, 15:53)linkinsoldier Wrote: Its pretty reproducable with the description above and the game super Mario all-stars on pocketsnes but I will give further details on weekend. The input on the bnes emulator works (but is too slow on rpi3), pocketsnes has the issue with the input but is working great otherwise!

I'll try super mario all-stars with pocketnes and see if I can get it to bug out.

(2017-02-17, 15:53)linkinsoldier Wrote: In which one of the debug logs can I look for your information? I reproduced it twice yesterday but couldn't find anything regarding button press... Did I maybe enable the wrong debug log for this test?

There's only 1 log, kodi.log, but it sounds like "debug logging" is disabled so you're only seeing INFO, WARNING and ERROR lines. You can enable debug logging from settings (see http://kodi.wiki/view/Debug_log )
RetroPlayer releases: https://github.com/garbear/xbmc/releases

Donations: eigendude.eth
(2017-02-17, 10:41)esco Wrote:
(2017-02-14, 21:05)Milhouse Wrote:
(2017-02-14, 17:58)esco Wrote: Thank you Milhouse. I have tested 4.10-rc8 build #0213x build and unfortunatelly the same behavior occurs.

Can you provide log information? "journalctl -a", "dmesg" and "kodi.log"?

Is #0101 the last build that has this working mux, or just the last working build you've used? Can you identify when this mux problem started in the normal daily builds?

We switched to using the media_build package in #0108 so you might be able to confirm that is when the problem started (ie. #0107 works, #0108 doesn't).

#0107 is fine, no issues at all. With #0108 I am able to scan muxes (however some "poll timeout" errors appear) but unable to watch TV - the log is full of MPEG2AUDIO/ H264 "Invalid start code" and "Continuity counter" errors. Am I supposed to send the logs with the latest #____x build?
Thank you.

Logs from latest build is fine, yes. In build #0108 we switched to the media_build package for DVB support so it's possible some hardware isn't fully supported in the media_build package. This will need input from a media_build expert - post your logs and I'll ping someone.
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.
(2017-02-17, 23:21)garbear Wrote:
(2017-02-17, 15:53)linkinsoldier Wrote: Its pretty reproducable with the description above and the game super Mario all-stars on pocketsnes but I will give further details on weekend. The input on the bnes emulator works (but is too slow on rpi3), pocketsnes has the issue with the input but is working great otherwise!

I'll try super mario all-stars with pocketnes and see if I can get it to bug out.

(2017-02-17, 15:53)linkinsoldier Wrote: In which one of the debug logs can I look for your information? I reproduced it twice yesterday but couldn't find anything regarding button press... Did I maybe enable the wrong debug log for this test?

There's only 1 log, kodi.log, but it sounds like "debug logging" is disabled so you're only seeing INFO, WARNING and ERROR lines. You can enable debug logging from settings (see http://kodi.wiki/view/Debug_log )
Will try tomorrow. You have to start the first game (Mario Bros. 85 or so) and then you will get 1 player or 2 player selection and you can then start it with the start button but then only the second controller input works...

Maybe fifa 97 is easier? Here no controller input seem to work with pocketsnes!

see github issue:
https://github.com/kodi-game/game.libret...2/issues/2
I've added the bugs to the Release Notes in the RetroPlayer test builds thread. I'm tracking all core-specific bugs in their github repos. Can you open Issues on the cores that don't work? If you don't have a github account, I'll open the issues myself when I get a chance to reproduce this bug.
RetroPlayer releases: https://github.com/garbear/xbmc/releases

Donations: eigendude.eth
(2017-02-16, 19:46)Milhouse Wrote:
(2017-02-16, 19:10)raptorjr Wrote: I am using that update channel, but don't see anything newer than #211.

I guess I'm doing something wrong.

We still have an issue with the LE clients pulling the list of available versions from the server. To force this you can toggle "Show Custom Channels" off/on, then set "Update Channel" to Milhouse-9.0 (again), or you can reboot.

Obviously we would rather it pulled the latest list automatically, but we currently need to implement that in a way which doesn't hammer the server - possibly a mixture of caching (background hourly/daily pulls for automatic notification) plus on demand ("Available Versions" button).

possible a "update list" button right below that lets a user manually refresh the list which should be less extensive on the server.
Wink 
(2017-02-16, 19:10)raptorjr Wrote:
(2017-02-15, 01:29)Milhouse Wrote:
(2017-02-15, 01:15)liberator Wrote: As an aside, I also see that no builds after #201 are shown in Updates - Available Versions. No problem for me because I get the updates with wget but maybe a problem for others?

You need to change the update channel to Milhouse-9.0.

I am using that update channel, but don't see anything newer than #211.

I guess I'm doing something wrong.

i just go to custom channel with the milhouse link and click ok then remove or add the trailing slash then press ok, this forces it to refresh and when you go to available versions, it will be the latest.
(2017-02-18, 01:32)garbear Wrote: I've added the bugs to the Release Notes in the RetroPlayer test builds thread. I'm tracking all core-specific bugs in their github repos. Can you open Issues on the cores that don't work? If you don't have a github account, I'll open the issues myself when I get a chance to reproduce this bug.

I will reproduce it later and try to open the bug tickets on GitHub.
New LibreELEC.tv Leia build #0217: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.9.10 #1 Sat Feb 18 07:39:59 GMT 2017 armv6l GNU/Linux

# vcgencmd version
Feb 17 2017 20:41:51
Copyright (c) 2012 Broadcom
version 88b8ca93c6eb22e38531075a6d7e4f25f900af43 (clean) (release)

# lsb_release
LibreELEC (Milhouse) - Version: devel-20170218073854-#0217-g6072304 [Build #0217]

# Kodi version
(18.0-ALPHA1 Git:4c792db). Platform: Linux ARM 32-bit

Based on tip of LibreELEC.tv master (6072304a, changelog) and tip of XBMC master (128c5ba7, changelog) with the following modifications: Build Highlights:
  1. New firmware (gpuserv)
  2. Exclude ffmpeg/gpuservice changes until build issues resolved
  3. Updated PR:1327 with Harmony fix
  4. (Clean build)
Build Details:
  1. Firmware (Feb 17):
    • firmware: Rebuild libs with gpuserv
    • gpuserv: Boost gpu frequencies when in use
  2. LibreELEC.tv:
    • busybox: add hexdump for C2 RAM OClocking script (PR:1333, 1 commit, 1 file changed)
    • packages/linux: Update Amlogic kernel to 3.10-de626d8 (PR:1362, 1 commit, 1 file changed)
  3. XBMC:
    • [depends] use ccache if available (PR:11529, 2 commits, 8 files changed)
    • Update README.raspberrypi (PR:11024, 0 commits, 0 files changed)
    • Fix cleanonupdate to clean music library (PR:11609, 1 commit, 1 file changed)
    • [Estouchy] add seekbar (PR:11642, 1 commit, 2 files changed)
    • [PVR] Timer settings dialog: Prefill recordings folder with timer title (PR:11679, 1 commit, 1 file changed)
    • [PVR] Refactoring: Remove GUI code from PVR core components, part 2. (PR:11676, 16 commits, 32 files changed)
    • [Fix]Limited number of Random songs apply sort and limit in DB (PR:11680, 1 commit, 1 file changed)
  4. inputstream.rtmp:
    • build for windows (PR:14, 4 commits, 9 files changed)
  5. pvr.mythtv:
    • Use rebased PTS for the program chain (2710285e)
    • no longer handle demuxing (18fe56e4)
  6. newclock5:
    • Commits no longer in build:
      • VideoPlayer: VideoReferenceClock to the right place (872427ce)
      • RPi: Fixup move of videosync to windowing (fe71b08a)
      • fixup! RPi: Fixup move of videosync to windowing (80f8cf4c)
      • windows: adapt to move of videosync to windowing (922e5b4c)
  7. kernel 4.9.y:
    • New commits in this build:
      • BCM270X_DT: Add SMSC ethernet controller to DT (0f4c9f7a)
      • brcmfmac: Mute expected startup 'errors' (535dec89)
      • clk-bcm2835: Mark used PLLs and dividers CRITICAL (6c453992)
      • clk-bcm2835: Add claim-clocks property (d8bc59c1)
  8. Additional commits/pull requests/changes not yet merged upstream:
    • Added: [env] PR:1365: build system: switch to absolute paths
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.
(2017-02-17, 23:21)Milhouse Wrote:
(2017-02-17, 10:41)esco Wrote:
(2017-02-14, 21:05)Milhouse Wrote: Can you provide log information? "journalctl -a", "dmesg" and "kodi.log"?

Is #0101 the last build that has this working mux, or just the last working build you've used? Can you identify when this mux problem started in the normal daily builds?

We switched to using the media_build package in #0108 so you might be able to confirm that is when the problem started (ie. #0107 works, #0108 doesn't).

#0107 is fine, no issues at all. With #0108 I am able to scan muxes (however some "poll timeout" errors appear) but unable to watch TV - the log is full of MPEG2AUDIO/ H264 "Invalid start code" and "Continuity counter" errors. Am I supposed to send the logs with the latest #____x build?
Thank you.

Logs from latest build is fine, yes. In build #0108 we switched to the media_build package for DVB support so it's possible some hardware isn't fully supported in the media_build package. This will need input from a media_build expert - post your logs and I'll ping someone.

As requested:

dmesg: http://pastebin.com/f1DCtsZa
journal: http://pastebin.com/p52rvz6t
kodi.log http://pastebin.com/91sid8Z2

Thank you Milhouse.
(2017-02-18, 01:32)garbear Wrote: I've added the bugs to the Release Notes in the RetroPlayer test builds thread. I'm tracking all core-specific bugs in their github repos. Can you open Issues on the cores that don't work? If you don't have a github account, I'll open the issues myself when I get a chance to reproduce this bug.

@garbear:
i opened the issue:
https://github.com/kodi-game/game.libret...2/issues/2
I found out the pocketsnes was renamed to snes9x 2002 - so i installed this but it has the same problem..Can I be of any more assistance?
I think i should wait for the bugfix of this before i open another ticket because there is no working input on fifa97, right?

I have two more topics i found:

1) snes9x2002 and nestoopia emulators are running too good / too fast on the rpi3 - is there anyway to sync tho "real time"? Smile
2) opening mario kart 64 on mupen64 is rebooting libreelec

Both topics are already known right? Is there anyone who can say when N64 emulator is going to work on rpi3?

Could you please do me a favor and pinpoint on how i could my games section layout? Detailed question (with pictures) are here:
http://forum.kodi.tv/showthread.php?pid=...pid2521113

@Mildhouse/garbear:
I now know, why i had those issues finding anything worth finding in the "log files". I looked in the samba share/logfiles/*.zip - it contains a 1- kodi.log, but is only has crashlog data from end of january ... i think this is a bug right? .kodi/temp/kodi.log contained the desired information!
  • 1
  • 65
  • 66
  • 67(current)
  • 68
  • 69
  • 495

Logout Mark Read Team Forum Stats Members Help
LibreELEC Testbuilds for RaspberryPi (Kodi 18.0)24