465 replies to this topic

[twol]

Hi,

 

I have built a build environment on my Windows 10 PC running the Ubuntu shell.  The build process is failing with the following:

root@windows:~/openpli-oe-core/build# MACHINE="beyonwiz t3" bitbake openpli-enigma2-image
ERROR: /home/IanSav/openpli-oe-core/meta-local/recipes-local/images/my-image.bb: Unable to determine endianness for architecture 'INVALID'
ERROR: /home/IanSav/openpli-oe-core/meta-local/recipes-local/images/my-image.bb: Please add your architecture to siteinfo.bbclass
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hdx_3.6.0.bb: cannot map 'INVALID' to a linux kernel architecture
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hdp_3.14.2.bb: cannot map 'INVALID' to a linux kernel architecture
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hde_3.12.1.bb: cannot map 'INVALID' to a linux kernel architecture
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hde2_3.14.2.bb: cannot map 'INVALID' to a linux kernel architecture
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-drivers/rt8723bt_4.1.6.bb: cannot map 'INVALID' to a linux kernel architecture
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hdx_3.6.0.bb: linux-ini-hdx: LIC_FILES_CHKSUM contains an invalid URL: t3-oe-linux/linux-ini-hdx/3.6.0-.1/linux-3.6.0/COPYING;md5=d7810fab7487fb0aad327b76f1be7cd7
ERROR: Failed to parse recipe: /home/IanSav/openpli-oe-core/meta-local/recipes-local/images/my-image.bb
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hdp_3.14.2.bb: linux-ini-hdp: LIC_FILES_CHKSUM contains an invalid URL: t3-oe-linux/linux-ini-hdp/3.14.2-.1/linux-3.14.2/COPYING;md5=d7810fab7487fb0aad327b76f1be7cd7
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hde_3.12.1.bb: linux-ini-hde: LIC_FILES_CHKSUM contains an invalid URL: t3-oe-linux/linux-ini-hde/3.12.1-.1/linux-3.12.1/COPYING;md5=d7810fab7487fb0aad327b76f1be7cd7
ERROR: /home/IanSav/openpli-oe-core/pli-extras/meta-ini/recipes-linux/linux-ini-hde2_3.14.2.bb: linux-ini-hde2: LIC_FILES_CHKSUM contains an invalid URL: t3-oe-linux/linux-ini-hde2/3.14.2-.1/linux-3.14.2/COPYING;md5=d7810fab7487fb0aad327b76f1be7cd7
 
Summary: There were 12 ERROR messages shown, returning a non-zero exit code.

This was the first sign of any issue with the build process.

 

Can anyone suggest a resolution?

 

Regards,

Ian.

MACHINE=beyonwizt3   ?? runs for me (no quotes or space) ... but also have you installed all the required tools into Ubuntu?

[IanSav]

Hi,

 

The above error was due to using an incorrect build name.  That is fixed but now I get the following failure:

root@windows:~/openpli-oe-core/build# MACHINE=beyonwizt3 bitbake openpli-enigma2-image
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-drivers/fulan-dvb-modules-spark7162.bb: Unable to get checksum for fulan-dvb-modules-spark7162 SRC_URI entry modules.conf: file could not be found
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-drivers/fulan-dvb-modules-spark7162.bb: Unable to get checksum for fulan-dvb-modules-spark7162 SRC_URI entry modules-conf.conf: file could not be found
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-drivers/fulan-dvb-modules-spark.bb: Unable to get checksum for fulan-dvb-modules-spark SRC_URI entry modules.conf: file could not be found
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-drivers/fulan-dvb-modules-spark.bb: Unable to get checksum for fulan-dvb-modules-spark SRC_URI entry modules-conf.conf: file could not be found
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-bsp/stslave-fw_0.1.bb: Unable to get checksum for stslave-fw SRC_URI entry audio.elf: file could not be found
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-bsp/stslave-fw_0.1.bb: Unable to get checksum for stslave-fw SRC_URI entry video.elf: file could not be found
WARNING: /home/IanSav/openpli-oe-core/pli-extras/meta-fulan/recipes-bsp/stlirc_0.9.0.bb: Unable to get checksum for stlirc SRC_URI entry lircd_beyonwizt3.conf: file could not be found
ERROR: ExpansionError during parsing /home/IanSav/openpli-oe-core/meta-openpli/recipes-openpli/openmultiboot/enigma2-plugin-extensions-openmultiboot.bb
Traceback (most recent call last):
  File "/home/IanSav/openpli-oe-core/bitbake/lib/bb/data_smart.py", line 412, in DataSmart.expandWithRefs(s='1+git${SRCPV}', varname='PV'):
                 try:
    >                s = __expand_var_regexp__.sub(varparse.var_sub, s)
                     try:
  File "/home/IanSav/openpli-oe-core/bitbake/lib/bb/data_smart.py", line 111, in VariableParse.var_sub(match=<_sre.SRE_Match object; span=(5, 13), match='${SRCPV}'>):
                 else:
    >                var = self.d.getVarFlag(key, "_content")
                 self.references.add(key)
  File "/home/IanSav/openpli-oe-core/bitbake/lib/bb/data_smart.py", line 794, in DataSmart.getVarFlag(var='SRCPV', flag='_content', expand=True, noweakdefault=False, parsing=False):
                     cachename = var + "[" + flag + "]"
    >            value = self.expand(value, cachename)
 
  File "/home/IanSav/openpli-oe-core/bitbake/lib/bb/data_smart.py", line 436, in DataSmart.expand(s='${@bb.fetch2.get_srcrev(d)}', varname='SRCPV'):
         def expand(self, s, varname = None):
    >        return self.expandWithRefs(s, varname).value
 
  File "/home/IanSav/openpli-oe-core/bitbake/lib/bb/data_smart.py", line 426, in DataSmart.expandWithRefs(s='${@bb.fetch2.get_srcrev(d)}', varname='SRCPV'):
                 except Exception as exc:
    >                raise ExpansionError(varname, s, exc) from exc
 
bb.data_smart.ExpansionError: Failure expanding variable SRCPV, expression was ${@bb.fetch2.get_srcrev(d)} which triggered exception OperationalError: disk I/O error
 
 
Summary: There were 7 WARNING messages shown.
Summary: There was 1 ERROR message shown, returning a non-zero exit code.

Any suggestions?

 

Regards,

Ian.

 

 

[IanSav]

Hi Twol,

 

When I first tried to run the code I was given a list of missing modules.  I installed them all.

 

Regards,

Ian.

[Huevos]

Ian, you are wasting your time doing anything with release. You have to commit to the develop branch so you need to work in that branch.

[IanSav]

Hi Huevos,
 
I thought I *was* trying to build the develop branch.  Are you saying that there is no build kit and/or feed for the develop branch?
 
Gee this is so much effort just to contribute code.  This is *so* much easier on OpenViX and Beyonwiz.  I have tested my code on Beyonwiz and OpenViX but there are enough differences in OpenPLi that I was to test it here.  If I can't get a reliable build and update feed would anyone like to test my proposed code on OpenPLi for me?
 
Regards,
Ian.

[Abu Baniaz]

Ian, you are wasting your time doing anything with release. You have to commit to the develop branch so you need to work in that branch.

The pli-metas use the Dev branch

 

Maybe the build related posts should go in the other thread?

Edited by Abu Baniaz, 28 March 2018 - 00:48.

[nautilus7]

I tried to build openpli-dev with the windows subsystem for linux (ubuntu on windows 10), but I didn't succeed. I didn't get the same errors as above, but something else (I can't remember, it's been some time since I tried). I think it's something else that prevents building, and not a user error, or a code error.

[IanSav]

Hi Nautilus7,

 

There are lots of modules missing from Windows 10 that have to be fetched and installed.  I did all of that.  I also installed the Beyonwiz image builder on the same system as the OpenPLi builder and it runs to completion.  Lots of warnings are generated (and I am told that this is normal) but it does finish.

 

But even if the image does finish being built the problem remains that the image can't be updated online.  Theoretically the image that was built for me should be enough to move forward.  Unfortunately since that image was made OpenPLi has changed a lot and the current source code does not match that image. 

 

Is there a way to update a running development image to the current development code?  I don't want to keep wiping the system and starting from scratch every time OpenPLi make a change to the develop branch.

 

Regards,

Ian.

[nautilus7]

What windows version do you use? I have a 16299.309 build that comes with beta windows subsystem for linux. Perhaps they have fixed some things because in my case I can't build at all.

 

I don't want to advertise other images, but you can switch to satdreamgr experimental image (if there's support for your box) and get every openpli-dev update right on your box, without any hassle.

[IanSav]

Hi Nautilus7,

 

I am running Windows 10 Pro, version 1709, OS Build 16299.334.  I have also applied all available Ubuntu updates.

 

Surely there has to be an easier way to actively develop for OpenPLi using a deveop image?  Isn't there *any* sort of central feed to keep a develop image updated?

 

Regards,

Ian.

[IanSav]

Hi,
 
It looks like my IP address is blocked from at least one of the OpenPLi source servers (http://www.streamboard.tv/svn/oscam/):

502 - BAD GATEWAY

Anti-DDoS Filter cannot connect to protected host.[/size]
Most likely the firewall (CSF?) on protected host blocks filter IP addresses due to a large number of requests.[/size]
If you are the administrator, please disable the firewall or whitelist the filter IP's.

Regards,
Ian.

Edited by IanSav, 29 March 2018 - 17:17.

[WanWizard]

We don't have anything to do with Streamboard and/or Oscam development.

 

Streamboard is down very often, it's a bloody nightmare. There used to be a clone online (oscam.cc), but that is down too. There are a few git mirrors, but that require a bit of hacking in your local.conf to switch from svn to git.

 

If you're up for it, here's one: https://github.com/o.../tree/oscam-svn

[nautilus7]

Hi Nautilus7,

 

I am running Windows 10 Pro, version 1709, OS Build 16299.334.  I have also applied all available Ubuntu updates.

 

Surely there has to be an easier way to actively develop for OpenPLi using a deveop image?  Isn't there *any* sort of central feed to keep a develop image updated?

 

Regards,

Ian.

 

Weird. I might try again with WSL at some point in the future.

[prl]

It was never quite clear to me what the "correct" display in Abu Baniaz's ButtonTexts(2).zip plugin should be.

 

I think the current Beyonwiz code should work with most doubled

<widget name=.../> <widget source=.../>

skin definitions for the button hints:

 Screen Shot 2018-06-09 at 11.45.07.png   93.09KB   4 downloads

 

 Screen Shot 2018-06-09 at 11.48.27.png   133.29KB   4 downloads

 

Is that basically what people want for this? If so, I'm now in a position to test the Beyonwiz changes against the OpenViX code, and submit a pull request. OpenPli folk can cherry-pick it from there if it's accepted.

Edited by prl, 9 June 2018 - 02:52.

[IanSav]

Hi Prl,

 

This is a copy of a post I made to you on the Beyonwiz forum.  I felt the post may be helpful here as well.

 

The issue is that neither OpenPLi or OpenViX have accepted your Beyonwiz based fix for the issue.  They called it a "hack" and refused to merge the code.  (I submitted a pull request using your Beyonwiz based fix.)

 

On those systems the test plugin shows that for the Button() and Label() objects the "source=" widget gets this initial text but does not get any of the .setText() method updates triggered by pressing the GREEN button.

 

The first part of my proposal was to eliminate the Button() object and simply make it a synonym for the Label() object.  In almost all regards it is already mostly just a synonym.

 

The next part of the proposal was to separate the knowledge link between the Python code and the skin.  By this I mean that a skin developer should not need to know how the code was written to know how to write a skin.  At the moment a skin developer needs to know if an object to be displayed was created via a Button(), Label() or StaticText() object.  The Python objects map directly to "Name=", "Name=" and "source=" skin widgets respectively.  It would be better and easier if all three Python objects get folded into a single object that can be used EQUALLY for a "name=" or "source=" skin widget.  All Python coders would be freed from having to decide which of the three current objects should be used.  All current objects would become synonyms so that all existing / legacy code would remain fully functional.  This would also mean that all existing / legacy skins should also remain fully functional.  Moving forward a skin author would then be able to use a "name=" widget for simple unconditional display of the object value.  They could equally select a "source=" style widget if the object value needed conditional display or the value needs to be passed through one of the many skin based converters.

 

Over time the legacy object names could be cleaned up to make the source code easier to read and be free of the old names.

 

This is a very concise description of the proposal.  If anyone needs the concept expanded then please ask.

 

Regards,

Ian.

Edited by IanSav, 9 June 2018 - 04:40.

[Abu Baniaz]

It was never quite clear to me what the "correct" display in Abu Baniaz's ButtonTexts(2).zip plugin should be.

That was a repost of Huevos' test plugin to show the errors, not mine.

Is that basically what people want for this? If so, I'm now in a position to test the Beyonwiz changes against the OpenViX code, and submit a pull request. OpenPli folk can cherry-pick it from there if it's accepted.

PLi seldom cherry pick

[littlesat]

I cintributed a lot to make this right... what beyonwiz change makes statictext an synoniem of button and label...?.
Button differs a bit from label as behind button you can add a Function, but that was never used... and never worked yet!

Edited by littlesat, 9 June 2018 - 06:19.

[prl]

The Beyonwiz code takes the existing DummySource hack and makes it work in the way it appears it was intended to work.

[littlesat]

Where is it?
Note a source object is havier then a ‘fixed’ label (or button) so when make something like this it might be better to consider something else..

Edited by littlesat, 9 June 2018 - 06:34.

[prl]

PLi seldom cherry pick

 

 

 

I have neither the time nor inclination to do Beyonwiz builds for two non-Beyonwiz versions of enigma2, especially after the amount of time it took to get an OE-Alliance 4.1 build environment to build for a Beyonwiz.