Openpli-5 (still next master)
Re: Openpli-5 (still next master) #241
Re: Openpli-5 (still next master) #242
Posted 6 April 2016 - 11:44
tried to compile the image with glibc 2.22 to make sure this is the issue, but it breaks Samba:
Collected errors: * satisfy_dependencies_for: Cannot satisfy the following dependencies for samba-base: * libc6 (>= 2.23) * * opkg_install: Cannot install package samba-base.
T90 | 28.2E-23.5E-19.2E-13E-9E-4.8E-0.8W-4W
Zgemma H7S | OpenPLI develop branch
Re: Openpli-5 (still next master) #243
Re: Openpli-5 (still next master) #244
Posted 6 April 2016 - 11:57
yep, weird...Weird. Kernel does not use glibc. Drivers are part of kernel and hence also do not use glibc.
no idea what's been affected by glibc upgrade:
but the fact is that, image with 2.22 boots, and with 2.23 does not (and no warnings errors during build)glibc: Upgrade to 2.23
Drop kconfig and options-group support
Forward port cross-localedef support
Assume ssp support in libc when building gcc-initial
is there any way to investigate this issue further?
Re: Openpli-5 (still next master) #245
Posted 6 April 2016 - 12:06
tried to compile the image with glibc 2.22 to make sure this is the issue, but it breaks Samba:
Collected errors: * satisfy_dependencies_for: Cannot satisfy the following dependencies for samba-base: * libc6 (>= 2.23) * * opkg_install: Cannot install package samba-base.
You're compiling half your packages with 2.23 and some others with 2.22 now.
Re: Openpli-5 (still next master) #246
Posted 6 April 2016 - 20:37
I did a fresh fresh build from the master-next head with glibc 2.22, and compiled successfully and box is booting. Thanks for the hint. Hopefully someone can fix it, as this chip is not such ancient.
T90 | 28.2E-23.5E-19.2E-13E-9E-4.8E-0.8W-4W
Zgemma H7S | OpenPLI develop branch
Re: Openpli-5 (still next master) #247
Posted 7 April 2016 - 09:27
When upgrade to latest master OE core .you need to rename the python .bbappens from 2.7.9 to 2.7.11
Small issue: when flash the box and go through the install wizard. Enigma will crash when using auto-AutoDiseqc
2263.714337> File "/usr/lib/enigma2/python/Components/TuneTest.py", line 24, in tune < 2263.715164> File "/usr/lib/enigma2/python/Components/TuneTest.py", line 28, in tuneSatObj < 2263.715856> File "/usr/lib/enigma2/python/enigma.py", line 2790, in __init__ < 2263.716894> AttributeError: No constructor defined < 2263.717373> [ePyObject] (CallObject(<bound method AutoDiseqc.statusCallback of <class 'Screens.AutoDiseqc.AutoDiseqc'>>,()) failed) ]]> </enigma2crashlog> </crashlogs>No big deal. Just configure your sat settings manually.
I found the solution, you need to enigma_python.i file to add the line:
%feature("notabstract") eDVBFrontendParameters;
Edited by Alexvrs, 7 April 2016 - 09:27.
Один — стоит, Единый движется.
Re: Openpli-5 (still next master) #248
Posted 7 April 2016 - 12:31
Re: Openpli-5 (still next master) #249
Posted 7 April 2016 - 14:14
@Alexvrs or it's still needed after swig fixes?
TNX, I did not pay attention to it
I used this as an alternative solution, since iDVBFrontendParameters class is abstract and eDVBFrontendParameters does not contain all of its members, and I do not know how this will work in future versions of Python and SWIG, but it can be kept in reserve.
Official details can be found here
Один — стоит, Единый движется.
Re: Openpli-5 (still next master) #250
Posted 7 April 2016 - 16:33
Just as useful info to all who build there own image and want to be up to date.
I just build from scratch image and feeds included at once.
1) for vuduo2
2) dm8000
base clone command followed by general make (whitout box but just to have all recipes )
git clone https://github.com/OpenPLi/openpli-oe-core -b master-next make
1) I added my personalised local images recipes.
2) Corrected some packages.
- replex located at openpli-oe-core/meta-openpli/recipes-multimedia
changes replex my bundle included into tar.gz file
- adapted the openpli-oe-core/meta-openembedded/meta-oe/files/waf-cross-answers/cross-answers-mips.txt
cp cross-answers-mips.txt cross-answers-mipsel.txt
- corrected the package openpli-oe-core/meta-openembedded/meta-oe/recipes-devtools/python/python-simplejson_3.8.1.bb
by adding :
do_install_append () { install -d ${D}/usr/lib/python2.7/site-packages/simplejson-3.8.1-py2.7.egg-info mv ${WORKDIR}/image/usr/lib/python2.7/site-packages/simplejson-3.8.1-py2.7-linux-x86_64.egg/simplejson ${WORKDIR}/image/usr/lib/python2.7/site-packages/simplejson mv ${WORKDIR}/image/usr/lib/python2.7/site-packages/simplejson-3.8.1-py2.7-linux-x86_64.egg/EGG-INFO/* ${WORKDIR}/image/usr/lib/python2.7/site-packages/simplejson-3.8.1-py2.7.egg-info/ rm -rf ${WORKDIR}/image/usr/lib/python2.7/site-packages/simplejson-3.8.1-py2.7-linux-x86_64.egg }
- added the very last gstreamer head bundle included here.
- changed : /home/christophe/openpli50/openpli-oe-core/meta-openpli/recipes-openpli/enigma2/enigma2.bb
replaced line :
gstreamer1.0-plugins-bad-fragmented \
by
gstreamer1.0-plugins-bad-hls \
Build 100 % ok at once for vuduo2 and dm8000 flash ok working ok.
Attached Files
Re: Openpli-5 (still next master) #251
Re: Openpli-5 (still next master) #252
Re: Openpli-5 (still next master) #253
Posted 8 April 2016 - 06:27
Can you give some example of such a file?
I do not think that the problem is a separate ffmpeg.
Libav now also simply clone original ffmpeg in own directory, but compile it with all possible options.
Options for ffmpeg may also set on the libav compilation, but you can not remove ffmpeg libraries that are not used.
P.S.
Witch ffmpeg config you use?
Edited by Taapat, 8 April 2016 - 06:29.
Re: Openpli-5 (still next master) #254
Posted 8 April 2016 - 08:12
I was using your gst bbappends (just chnged the rev. to 1.8.0 tagged), and your ffmpeg 3.0.1 recipe.
So far I know, that with the old gst recipes from openpli tweaked for 1.8.0 it's fine. I just compiled gst with the OE's recipes for the same rev. without the libav packageconfig to make sure, not a patch from OE is breaking the AV sync - will test this later today.
In all cases I was using the latest openpli master-next head with OE current head (just changed glibc to 2.22), and Chris's multibox sink (also tried without his latest changes)
T90 | 28.2E-23.5E-19.2E-13E-9E-4.8E-0.8W-4W
Zgemma H7S | OpenPLI develop branch
Re: Openpli-5 (still next master) #255
Posted 8 April 2016 - 09:56
@Taapat
Tried your approach to use ffmpeg in libav, the size was really noticeable smaller, but had huge AV sync problems, audio 1 sec late in mkv/h264/ac3 combination
So far I know gst-libav (or ffmpeg) is not required at all to play media with ac3 audio.
Unless the vuuno does require it ??? could be. ac3 is a on board codec. Actually the gstreamer-1.0-libav is only required for media containing wma audio. All the others audio are playing fine whitout gstreamer-1.0-libav. (which actually is using ffmpeg n3.0 and not libav anymore)
for other non on board audio codecs flac,mp3,ogg, dts we are using gstreamer plugins whitout gstreamer-1.0-libav
Edited by christophecvr, 8 April 2016 - 09:58.
Re: Openpli-5 (still next master) #256
Re: Openpli-5 (still next master) #257
Re: Openpli-5 (still next master) #258
Re: Openpli-5 (still next master) #259
Re: Openpli-5 (still next master) #260
11 user(s) are reading this topic
0 members, 11 guests, 0 anonymous users