Jump to content


Photo

Openpli-5 (still next master)


  • Please log in to reply
1177 replies to this topic

Re: Openpli-5 (still next master) #421 MastaG

  • Senior Member
  • 1,531 posts

+118
Excellent

Posted 24 August 2016 - 19:25

So oe-a is also already at gcc6 :-)
Great.

Re: Openpli-5 (still next master) #422 athoik

  • PLi® Core member
  • 8,458 posts

+327
Excellent

Posted 24 August 2016 - 20:09

So oe-a is also already at gcc6 :-)
Great.

 

I think OpenATV 5.4 uses next branch of OE-A, so they are still testing.


Wavefield T90: 0.8W - 1.9E - 4.8E - 13E - 16E - 19.2E - 23.5E - 26E - 33E - 39E - 42E - 45E on EMP Centauri DiseqC 16/1
Unamed: 13E Quattro - 9E Quattro on IKUSI MS-0916

Re: Openpli-5 (still next master) #423 MiLo

  • PLi® Core member
  • 14,045 posts

+298
Excellent

Posted 25 August 2016 - 07:59

So aint it easier to just revert back the oe-snapshot to gcc5 and create bbappends for the stuff that dont build when the host has gcc6.


"Easier" would be to never change anything.

Purpose of the merge was not just to allow building on your specific host. It was to finally force to move things forward after months of inactivity. The BSP problems were holding OpenPLi back, better to have that in the open.
Real musicians never die - they just decompose

Re: Openpli-5 (still next master) #424 MiLo

  • PLi® Core member
  • 14,045 posts

+298
Excellent

Posted 25 August 2016 - 08:04

So oe-a is also already at gcc6 :-)


There are two distinct problems with GCC6:

-1- A few native recipes will not build correctly with GCC6. This means you cannot build it on a host system that has GCC v6 as its default compiler.
-2- OE itself moved to using GCC v6 for cross-compiling. This causes kernels before 4.2 to fail to build.

Problem -1- can be solved using some bbappends.
Problem -2- requires all BSPs to be upgraded, and also solves -1-
Real musicians never die - they just decompose

Re: Openpli-5 (still next master) #425 babsy98

  • Senior Member
  • 166 posts

+18
Neutral

Posted 25 August 2016 - 08:36

yes all kernel in oe-a ready with gcc 6.1.1 , i have update for 3 month all builds, but test we found more and more small issues and not ready yet. so atv 5.4 only close beta for a small user group

 

public beta need more time we have to many issue with vu drivers and wait for fixes but no response from vu



Re: Openpli-5 (still next master) #426 MastaG

  • Senior Member
  • 1,531 posts

+118
Excellent

Posted 25 August 2016 - 11:52

Thanks for explaining MiLo.

I'll merge the latest changes and see what I can do.



Re: Openpli-5 (still next master) #427 athoik

  • PLi® Core member
  • 8,458 posts

+327
Excellent

Posted 25 August 2016 - 11:59

Please note that error on split_kernel_module_packages still exists.

 

diff --git a/meta/classes/kernel-module-split.bbclass b/meta/classes/kernel-module-split.bbclass
index ed42d2b..513396a 100644
--- a/meta/classes/kernel-module-split.bbclass
+++ b/meta/classes/kernel-module-split.bbclass
@@ -50,7 +50,7 @@ python split_kernel_module_packages () {
         tmpfile = tf[1]
         cmd = "%sobjcopy -j .modinfo -O binary %s %s" % (d.getVar("HOST_PREFIX", True) or "", file, tmpfile)
         subprocess.call(cmd, shell=True)
-        f = open(tmpfile)
+        f = open(tmpfile, encoding="latin-1")
         l = f.read().split("\000")
         f.close()
         os.close(tf[0])

 

Also dm800se is too fat... (dm800se.rootfs.jffs2 is too big. This doesn't work. Sorry.Invalid data partition!).

 

So a diet program required..


Wavefield T90: 0.8W - 1.9E - 4.8E - 13E - 16E - 19.2E - 23.5E - 26E - 33E - 39E - 42E - 45E on EMP Centauri DiseqC 16/1
Unamed: 13E Quattro - 9E Quattro on IKUSI MS-0916

Re: Openpli-5 (still next master) #428 peti

  • Senior Member
  • 115 posts

+1
Neutral

Posted 25 August 2016 - 17:33

I was able to compile the latest head for Vu+ Uno. So far working well. 


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) #429 MiLo

  • PLi® Core member
  • 14,045 posts

+298
Excellent

Posted 25 August 2016 - 17:59

Please note that error on split_kernel_module_packages still exists.
 
 

diff --git a/meta/classes/kernel-module-split.bbclass b/meta/classes/kernel-module-split.bbclass
index ed42d2b..513396a 100644
--- a/meta/classes/kernel-module-split.bbclass
+++ b/meta/classes/kernel-module-split.bbclass
@@ -50,7 +50,7 @@ python split_kernel_module_packages () {
         tmpfile = tf[1]
         cmd = "%sobjcopy -j .modinfo -O binary %s %s" % (d.getVar("HOST_PREFIX", True) or "", file, tmpfile)
         subprocess.call(cmd, shell=True)
-        f = open(tmpfile)
+        f = open(tmpfile, encoding="latin-1")
         l = f.read().split("\000")
         f.close()
         os.close(tf[0])


Shouldn't that be "utf-8" encoding or so? latin-1 is rare on posix...

Also dm800se is too fat... (dm800se.rootfs.jffs2 is too big. This doesn't work. Sorry.Invalid data partition!).
So a diet program required..


Maybe as simple as setting "IMAGE_INSTALL_remove_dm800se"? I think you could even do that in the machine.conf.
Real musicians never die - they just decompose

Re: Openpli-5 (still next master) #430 athoik

  • PLi® Core member
  • 8,458 posts

+327
Excellent

Posted 25 August 2016 - 18:06

Shouldn't that be "utf-8" encoding or so? latin-1 is rare on posix...


It is utf8 now with python3, the latin1 emulates python2 behaviour.

The error was mentioned before here: http://forums.openpl...es-split-fails/
 

Also dm800se is too fat... (dm800se.rootfs.jffs2 is too big. This doesn't work. Sorry.Invalid data partition!).
So a diet program required..




Maybe as simple as setting "IMAGE_INSTALL_remove_dm800se"? I think you could even do that in the machine.conf.


I guess the IMAGE_INSTALL_remove_dm800se = "samba-base" will do it.

Edited by athoik, 25 August 2016 - 18:07.

Wavefield T90: 0.8W - 1.9E - 4.8E - 13E - 16E - 19.2E - 23.5E - 26E - 33E - 39E - 42E - 45E on EMP Centauri DiseqC 16/1
Unamed: 13E Quattro - 9E Quattro on IKUSI MS-0916

Re: Openpli-5 (still next master) #431 MiLo

  • PLi® Core member
  • 14,045 posts

+298
Excellent

Posted 25 August 2016 - 18:19

binary:

        with open(tmpfile, "rb") as f:
            l = f.read().split(b"\000")
but that makes "l" a list of binary objects, causing problems later on in the code of course, which will expect strings.
Real musicians never die - they just decompose

Re: Openpli-5 (still next master) #432 Taapat

  • PLi® Core member
  • 2,343 posts

+120
Excellent

Posted 25 August 2016 - 19:08

I was able to compile the latest head for Vu+ Uno. So far working well. 

 

I think so too, until I tried ffmpeg.
At the library is load, I got an error:

error while loading shared libraries: /usr/lib/libavformat.so.57: ELF file ABI version invalid

But I see that on the openembedded-core has been added glibc patch that would fix it and so I'll try to update openembedded-core again.

I wanted to add that immediately after gcc 6.1 update on openembedded-core, I did not have such problems.

 

P.S.

Or any of openpli watched pull requests and comments on github?


Edited by Taapat, 25 August 2016 - 19:09.


Re: Openpli-5 (still next master) #433 Brievenbus

  • Member
  • 22 posts

0
Neutral

Posted 25 August 2016 - 21:02


Also dm800se is too fat... (dm800se.rootfs.jffs2 is too big. This doesn't work. Sorry.Invalid data partition!).

 

So a diet program required..

 

DM?

Will Pli5 be for Dreambox again????



Re: Openpli-5 (still next master) #434 WanWizard

  • PLi® Core member
  • 68,625 posts

+1,739
Excellent

Posted 26 August 2016 - 10:09

What do you mean by again? There is a current image for the DM800SE?


Currently in use: VU+ Duo 4K (2xFBC S2), VU+ Solo 4K (1xFBC S2), uClan Usytm 4K Pro (S2+T2), Octagon SF8008 (S2+T2), Zgemma H9.2H (S2+T2)

Due to my bad health, I will not be very active at times and may be slow to respond. I will not read the forum or PM on a regular basis.

Many answers to your question can be found in our new and improved wiki.


Re: Openpli-5 (still next master) #435 Brievenbus

  • Member
  • 22 posts

0
Neutral

Posted 26 August 2016 - 11:00

Ah, you are right of course, it's just that Dream Multimedia is not mentioned on the home page, but, it is on the downloadpage.



Re: Openpli-5 (still next master) #436 WanWizard

  • PLi® Core member
  • 68,625 posts

+1,739
Excellent

Posted 26 August 2016 - 12:31

Everything on the homepage is supported.

 

Manufacturers we have a mutual agreement with are in the center of the hompage with a logo. Manufacturers we don't have an agreement with (but we can still make an image for) are still on the homepage, but in the sidebar, and those are supported until we can no longer support them (due to lack of manufacturer support). DMM is listed in the sidebar. On the download page, you see all brands we currently support.


Currently in use: VU+ Duo 4K (2xFBC S2), VU+ Solo 4K (1xFBC S2), uClan Usytm 4K Pro (S2+T2), Octagon SF8008 (S2+T2), Zgemma H9.2H (S2+T2)

Due to my bad health, I will not be very active at times and may be slow to respond. I will not read the forum or PM on a regular basis.

Many answers to your question can be found in our new and improved wiki.


Re: Openpli-5 (still next master) #437 peti

  • Senior Member
  • 115 posts

+1
Neutral

Posted 26 August 2016 - 21:48

to decrease the image size we can append some gst-libav arguments suggested by Chris earlier. As far as I know the whole package is needed just because of the wma codecs. 

 

LIBAV_INCLUDED_EXLUCED = "--disable-everything \
--enable-decoder=wmalossless \
 --enable-decoder=wmapro \
 --enable-decoder=wmav1 \
 --enable-decoder=wmav2 \
 --enable-decoder=wmavoice"

LIBAV_EXTRA_CONFIGURE_COMMON_ARG += "${LIBAV_INCLUDED_EXLUCED}"

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) #438 MastaG

  • Senior Member
  • 1,531 posts

+118
Excellent

Posted 26 August 2016 - 21:59

I dont even install gst1.0-libav on my dm800se.

Re: Openpli-5 (still next master) #439 christophecvr

  • Senior Member
  • 3,131 posts

+140
Excellent

Posted 1 September 2016 - 08:06

I dont even install gst1.0-libav on my dm800se.

gst-libav is only needed for vu+ series as far I know and only for the wma (audio codecs) If You just compile for those it's a very small library . The dm8000 and dm800se do have wma audiocodecs on board.

 

There could be other boxes who require gst-libav that I don't.



Re: Openpli-5 (still next master) #440 MastaG

  • Senior Member
  • 1,531 posts

+118
Excellent

Posted 2 September 2016 - 09:54

Hi Chris!

 

Good to see you around again :D




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users