•   
  • 1
  • 121
  • 122
  • 123(current)
  • 124
  • 125
  • 156
  •   
  Thread Closed
OpenELEC Testbuilds for RaspberryPi Part 3 (Kodi 14.0)
(2014-10-28, 15:34)bastos Wrote: Sorry for the noobish question but I can't find "Update library on startup" option under Video > Library in this build.

I'm trying to disable it because it's consuming too much CPU on my rpi.

Change your settings level to "Standard" or above.
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.
(2014-10-28, 15:52)Milhouse Wrote:
(2014-10-28, 15:34)bastos Wrote: Sorry for the noobish question but I can't find "Update library on startup" option under Video > Library in this build.

I'm trying to disable it because it's consuming too much CPU on my rpi.

Change your settings level to "Standard" or above.

It's already in Expert mode. Maybe it's some related to my language settings (Brazilian Portuguese). I can't test it now am at work but I will test as soon as I get home and let you know.

Thanks
(2014-10-28, 16:52)bastos Wrote: It's already in Expert mode. Maybe it's some related to my language settings (Brazilian Portuguese). I can't test it now am at work but I will test as soon as I get home and let you know.

Also make sure you're testing with stock Confluence.
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.
(2014-10-28, 02:05)Milhouse Wrote: New OpenELEC Helix build: #1027

...

Build Highlights:
  1. ...
  2. Fix configuration bug in libcec which may improve double-key suppression. Previous double-key suppression patch now dropped. Please test if this build is better/worse/same.
  3. ...
  4. ...
Build Details:
  1. ...
  2. ...
  3. newclock4:
    • ...
    • fixed: don't try to stick a value > 255 in a byte
    • Increase double tap timeout to 300ms
    • ...
  4. ...
Works like a charm! I wasn't able to reproduce the double key issue. Great to see that it has been properly fixed instead of using a work-around.

One thing I did notice is that responding becomes relatively slow when you press a key really, and I mean almost abnormally, fast. It actually goes faster when you press slightly less fast. Could this be because of the time-out increase?
I used the shutdown menu as a low action list instead of e.g. the movie library with heavier loads on switching entries for testing.
I do want to be able to navigate relatively fast every once in a while and the delay is on the fringe of being possibly annoying right now.

I'm happy with this being fixed in it's current state nevertheless.
(2014-10-28, 20:01)Forage Wrote: Works like a charm! I wasn't able to reproduce the double key issue. Great to see that it has been properly fixed instead of using a work-around.

One thing I did notice is that responding becomes relatively slow when you press a key really, and I mean almost abnormally, fast. It actually goes faster when you press slightly less fast. Could this be because of the time-out increase?
I used the shutdown menu as a low action list instead of e.g. the movie library with heavier loads on switching entries for testing.
I do want to be able to navigate relatively fast every once in a while and the delay is on the fringe of being possibly annoying right now.

I'm happy with this being fixed in it's current state nevertheless.

Yes. The Panasonic TVs generate two button press events if you hold the button for more than about 250ms (which is a quite normal duration for a button press).
The workaround is to reject a second button if it occurs within ~300ms of the first.
Unfortunately two rapid presses produces the same result as the single leisurely press which means they also get filtered out.

I've been working on improving the behaviour of CEC remotes.
Currently when you hold down a button it repeats at a rate chosen by the TV. Typically that is between 2 and 5 times per second (varies with TV manufacturer).
This is a much slower repeat than a keyboard (repeats 12 times per second) which makes scrolling long lists slow with CEC.

My plan is to (optionally) ignore the CEC presses and generate a different stream of presses with a configurable repeat rate, (e.g. allowing 12 repeats per second).
Unfortunately CEC devices are pretty inconsistent in behaviour. It can be impossible to distinguish a few quick presses from a single longer press.
Some devices generate button released events which are useful, but the Sony in front of me doesn't always...
So I'll allow a few settings to tweak how it behaves and we'll see if people prefer it (the default will behave how it always has).

Not ready for tonight, but maybe tomorrow night.
My bad just updated ..
New OpenELEC Helix build: #1028
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 3.17.1 #1 PREEMPT Tue Oct 28 22:45:32 GMT 2014 armv6l GNU/Linux

# vcgencmd version
Oct 17 2014 17:56:05
Copyright (c) 2012 Broadcom
version 845092531e360acd37f2f1964157f8079c77edfb (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20141028224420-r19478-g4b97acf [Build #1028]

Based on tip of OpenELEC master (4b97acfb, changelog) and tip of XBMC master (662c5dc7, changelog) with the following modifications:
  • Includes newclock4 patches
  • Excludes the OpenELEC fernetmenta patches due to conflicts with newclock4
  • Excludes the OpenELEC linux-01-RPi_support patch in favour of sourcing these and possibly more recent patches directly from kernel branch rpi-3.17.y
  • Excludes the OpenELEC xbmc-001-newclock4 patch in favour of sourcing these and possibly more recent patches directly from newclock4 branch
  • Default setting for "Show RSS Feed" changed to disabled (new installs only) [patch details]
  • Disabled "Total Duration" in Confluence (see build #0221 for details)
  • Adapted service.openelec.settings to take advantage of PR:5217 [patch details]
  • Includes latest libnfs master (ab1af695)
  • Includes latest libcec master (9be2206f), with latest testing commits from popcornmix
  • Includes latest xbmc-pvr-addons master (067befe6)
  • Includes latest xbmc-addon-xvdr master (2bf2563c)
  • Includes CONFIG_COREDUMP=y to allow creation of coredumps (see here)
  • Includes additional ffmpeg codecs/muxers enabled for testing/benchmarking purposes. See patch
  • Includes commits from libcec-2.2.0 (popcornmix)
  • Includes PR:5312: dvdplayer: do a seek after having changed subtitle stream
  • Includes PR:5416: added: POC mp4 movie/episode tag support
  • Includes PR:5573: webserver: improved caching control (see discussion)
  • Includes PR:5584: HTTP/Webdav encoding fixes - Fixes ticket #15365
  • Includes PR:5587: Invalidate path hash on remove so the infoscanner can pick it up again
  • Includes PR:5593: fix script.library type missing in runscript/runaddon checks.
  • Reverts e502b099: Segfault accessing DVD ISO/IFO over SMB (see PR5534)
  • Reverts 5d7960cb: Temporary conflict with popcornmix's branch
Build Highlights:
  1. Update to latest Pulse-Eight libcec master v2.2.0
  2. Add extra libcec de-bounce commits from popcornmix - new delay/repeat/release options in Settings -> System -> Input devices -> Peripherals -> CEC Adapter
  3. Revert only the commit causing DVD/SMB problems, not the whole PR:5534
Build Details:
  1. XBMC:
    • [win32] [vfs] Win32SMBDirectory: Do not try to connect with not required. (PR:5571, 1 commit, 1 file changed)
    • [smb] remove logging of url on open (PR:5362, 1 commit, 1 file changed)
    • Respect parent path decision when cleaning paths. (PR:5524, 1 commit, 1 file changed)
    • [confluence] fix default focus in ProfileSettings.xml (fixes #15507) (PR:5591, 1 commit, 1 file changed)
  2. newclock4:
    • [cec] Add settings for configuring button repeats
  3. libcec-2.2.0 (popcornmix):
    • Make released key polling wait for exact time until key gets released
    • Keep track of time since initial button press and last button update
    • Support repeating button presses with configurable repeat rate
    • Skip double press removal. It is handled through other means.
  4. Custom:
    • Added PR:5587: Invalidate path hash on remove so the infoscanner can pick it up again
    • Added PR:5593: fix script.library type missing in runscript/runaddon checks.
    • Reverted e502b099: Segfault accessing DVD ISO/IFO over SMB (see PR5534)
    • Reverted 5d7960cb: Temporary conflict with popcornmix's branch
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.
(2014-10-29, 01:44)Milhouse Wrote: Build Highlights:
[*]Add extra libcec de-bounce commits from popcornmix - new delay/repeat/release options in Settings -> System -> Input devices -> Peripherals -> CEC Adapter

If you use CEC then please test this.

New settings in system/input/peripherals/cec.

delay: duration before repeating (ms)
repeat: rate the button will repeat at when held (ms)
release: duration after less pressed message until button is considered released (ms)

delay should be set as low as possible, without unwanted double presses
increase rate for faster scrolling of lists when held down. The keyboard in xbmc uses 80ms.
if release is zero, we just rely on the released messages through CEC. Otherwise released messages are ignored and the release will occur after this duration.

I'd suggest trying delay=300, repeat=80, release=0 as a starting point, which works well on my Panasonic.
Even though kodi doesn't respond with either remotes I was still able to access it through putty and was able to access directories on pi using explorer. I have now enabled debug on pi and will post it the next time it doesn't respond.

Thank you
(2014-10-29, 01:59)denz Wrote: Even though kodi doesn't respond with either remotes I was still able to access it through putty and was able to access directories on pi using explorer. I have now enabled debug on pi and will post it the next time it doesn't respond.
Please see my comment at 1823332 (post), which may help.
(2014-10-17, 12:21)popcornmix Wrote: Try with a lower number - e.g. 10 or 20?
Thanks, it seems 10 is the magic number. I've been testing this for ~10 days with a bunch of 25 fps videos with passthrough enabled and I'm unable to reproduce (fingers crossed). (I was getting this.)

As it took a few hours to reproduce with 50, I have skipped intermediate values (50->10).
(2014-10-15, 04:52)gendo Wrote:
(2014-10-15, 00:30)popcornmix Wrote:
(2014-10-12, 15:49)gendo Wrote: Popcornmix did you have time to look at the "hack: revert squash: don't update originaldts when marked as invalid" I have with live tv? Could you if not a big problem revert it for the time being so that i can try out latest builds again?

BTW, this is (temporarily) reverted again.

Thanks Smile

Popcornmix could you please revert again so that i can test latest builds. Or tell me if i can assist in fixing this bug?
On the latest build I'm seeing what I would consider high cpu usage while Internet streaming a low quality sd video (%72+-), streaming a 1080p video over UPnP nets around %35+- usage which is great. I'm just wondering if this is normal since I haven't checked CPU usage in a good while. If not normal I will provide logs and more pertinent information.

Pi is overclocked to:

arm_freq=900
core_freq=330
sdram_freq=400
over_voltage=2

*Edit*
It was yatse (the android app) causing the CPU usage spike, I dont know what all it queries but it causes %15 to %20 additional CPU usage when running.
Pi has stopped responding I have uploaded the log hopefully it is of some help


http://xbmclogs.com/show.php?id=333597
(2014-10-29, 12:51)denz Wrote: Pi has stopped responding I have uploaded the log hopefully it is of some help

http://xbmclogs.com/show.php?id=333597

It doesn't look like any button presses came through, so my changes shouldn't have an effect.
Can you confirm if the night before's build has the same issue?
That also had some new libCEC updates (but not the repeat button changes).
  •   
  • 1
  • 121
  • 122
  • 123(current)
  • 124
  • 125
  • 156
  •   
  Thread Closed
 
Thread Rating:
  • 8 Vote(s) - 4.88 Average



Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi Part 3 (Kodi 14.0)4.888