Build my own Vix image

[Willo3092]

I would love to learn, are there any guides that you could point me in the direction of please?

[twol]

Hate to point to another image! .. but have a look at the bottom half of this page
https://github.com/openatv/enigma2

In the git change from OpenATV to Openvix and away you go.

[1554leslie]

I would love to learn, are there any guides that you could point me in the direction of please?Right up your street is that willo

[dodgydrains]

I would love to learn, are there any guides that you could point me in the direction of please?

If you manage to get it built please share, I am happy to test. Unfortunately my understanding of building an image is close to zero.

[Willo3092]

If you manage to get it built please share, I am happy to test. Unfortunately my understanding of building an image is close to zero.

... and mine!

[Sicilian]

It's not just image to built, plugin feeds need to be hosted too, unless for your own use locally.

I won't have time to re-test for a few weeks.

[Willo3092]

I've had a go tonight but failed unfortunately.
I found this log if it makes sense to anyone?

[twol]

I've had a go tonight but failed unfortunately.
I found this log if it makes sense to anyone?

Never seen that kind of looks like a fetch error..
Please describe what you have done ... and on what system you are running on.

[Willo3092]

Never seen that kind of looks like a fetch error..
Please describe what you have done ... and on what system you are running on.

I installed ubuntu-16.04.1-desktop-i386 and used these commands in terminal:

#1 - Install packages on your buildserver
sudo apt-get install -y autoconf automake bison bzip2 curl cvs diffstat flex g++ gawk gcc gettext git-core gzip help2man ncurses-bin ncurses-dev libc6-dev libtool make texinfo patch perl pkg-config subversion tar texi2html wget zlib1g-dev chrpath libxml2-utils xsltproc libglib2.0-dev python-setuptools zip info coreutils diffstat chrpath libproc-processtable-perl libperl4-corelibs-perl sshpass default-jre default-jre-headless java-common libserf-dev qemu quilt libssl-dev

#2 - Set your shell to /bin/bash
sudo dpkg-reconfigure dash When asked: Install dash as /bin/sh? select "NO"

#3 - Add user openvixbuilder
sudo adduser openvixbuilder

#4 - Switch to user openvixbuilder
su openvixbuilder

#5 - Switch to home of openvixbuilder
cd ~

6 - Create folder openvix
mkdir -p ~/openvix

#7 - Switch to folder openvix
cd openvix

#8 - Clone oe-alliance git
git clone git://github.com/oe-alliance/build-enviroment.git -b 4.3

#9 - Switch to folder build-enviroment
cd build-enviroment

#10 - Update build-enviroment
make update

#11 - Finally you can start building a image
MACHINE=dinobot4kplus DISTRO=openvix DISTRO_TYPE=release make image

[twol]

Try MACHINE=u52 !!

[twol]

Try MACHINE=u52 !!

If you get through this then open the build environment
there is a folder site.conf, open it and look at this line starting DL_DIR
this is where you fetch & save a LOT of files ..... so if necessary move it to somewhere safe and change the line to point to the new location.
So when you rebuild at any time you will not have to fetch anew the git libraries/build components

[Sicilian]

Try MACHINE=u52 !!

Machine is 'dinobot4kplus'.

[ronand]

I think you also need to use the 64 bit OS if I remember rightly.

[ccs]

As a matter of interest, would it be feasible to use a bootable Ubuntu USB stick, or does it need to be a full blown pc install?

[Sicilian]

As a matter of interest, would it be feasible to use a bootable Ubuntu USB stick, or does it need to be a full blown pc install?

Would build too slow, take forever.

You need a fast PC.

[Willo3092]

Try MACHINE=u52 !!

Didn't like that at all

ERROR: OE-core's config sanity checker detected a potential misconfiguration.
Either fix the cause of this error or at your own risk disable the checker (see sanity.conf).
Following is the list of potential problems / advisories:

Error, the PACKAGE_ARCHS variable contains duplicates. The following archs are listed more than once: u52


EDIT: I'm pretty sure the alpha blending issues have been addressed.
It all seems to work fine on ATV, I just don't like ATV.

[norhap]

Machine is 'dinobot4kplus'.

this is your machine

[Willo3092]

I think you also need to use the 64 bit OS if I remember rightly.

There only seems to be a 64 bit version for AMD processors and my spare box is a Dell with an i7 processor.
I could always put a spare hdd in my main box which is AMD and install it on there when I get a minute.

[abu baniaz]

There only seems to be a 64 bit version for AMD processors and my spare box is a Dell with an i7 processor.
I could always put a spare hdd in my main box which is AMD and install it on there when I get a minute.

It is just the way the linux distros are named, you can install on Intel CPUs. The build process is faster on Intel than AMD anyway.

[dodgydrains]

Sicilian is correct ... just a mental aberration!

So running builds at the moment (with alpha blending change)- will test later or tomorrow morning.

I have my fingers and toes crossed that all goes smoothly!!!!!

[Willo3092]

It is just the way the linux distros are named, you can install on Intel CPUs. The build process is faster on Intel than AMD anyway.

You're right, I've just installed the AMD 64bit version and trying again :thumbsup:

[Willo3092]

Latest error message:

NOTE: Tasks Summary: Attempted 2722 tasks of which 0 didn't need to be rerun and 2 failed.

Summary: 2 tasks failed:
/home/openvixbuilder/openvix/build-enviroment/meta-oe-alliance/meta-oe/recipes-distros/openvix/softcams/openvix-softcams-oscam-latest-arm.bb:do_fetch
/home/openvixbuilder/openvix/build-enviroment/meta-oe-alliance/meta-oe/recipes-distros/openvix/softcams/openvix-softcams-oscam-emu-arm.bb:do_fetch
Summary: There were 52 WARNING messages shown.
Summary: There were 7 ERROR messages shown, returning a non-zero exit code.
Makefile:960: recipe for target 'image' failed
make: *** [image] Error 1

[abu baniaz]

#8 - Clone oe-alliance git
git clone git://github.com/oe-alliance/build-enviroment.git -b 4.3


We are using branch 4.2 at the moment

[twol]

OK - the best log to look at is in the build-environment directory
Enter the directory, there is a build directory and in there is a log folder .... click on the lastest log.
You need to look at the log to see if its an issue with the source url or the oscam level (changed 4 days ago)
If the former then a rerun may fix it (thats why its important to save the library containing all the downloads) if the latter then go into the folder indicated in the log and revert it to the previous level (look at the OE-A web page for the info)

[Willo3092]

I think this is the log

[twol]

Fetch fails for the latest levels ... I will have a look tomorrow and see if I can provide a fix for you.

[Willo3092]

Latest failure:

I did as @twol said and copied the oscam files and .done files to the sources folder.

ERROR: Task (/home/imagebuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/binutils/binutils_2.29.bb:do_compile) failed with exit code '1'
NOTE: Tasks Summary: Attempted 14578 tasks of which 2817 didn't need to be rerun and 1 failed.

Summary: 1 task failed:
/home/imagebuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/binutils/binutils_2.29.bb:do_compile
Summary: There were 103 WARNING messages shown.
Summary: There were 2 ERROR messages shown, returning a non-zero exit code.
Makefile:960: recipe for target 'image' failed
make: *** [image] Error 1

[twol]

Latest failure:

I did as @twol said and copied the oscam files and .done files to the sources folder.

ERROR: Task (/home/imagebuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/binutils/binutils_2.29.bb:do_compile) failed with exit code '1'
NOTE: Tasks Summary: Attempted 14578 tasks of which 2817 didn't need to be rerun and 1 failed.

Summary: 1 task failed:
/home/imagebuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/binutils/binutils_2.29.bb:do_compile
Summary: There were 103 WARNING messages shown.
Summary: There were 2 ERROR messages shown, returning a non-zero exit code.
Makefile:960: recipe for target 'image' failed
make: *** [image] Error 1
Sorry forgot all about this post - really need the cooker log from builds directory
If its afetch error (which is why you should save downloads) then rerunning may just fix the error - often you find some servers are not available for a period

[Willo3092]

Maybe I need to put a bigger hdd in. I'm only running a 120gb ssd at the moment.
I tried to run an atv build but that also failed.
I've saved the ViX sources folder so it may be a case of adding stuff to it as I go.

[twol]

Obviously both images use the same core but will differ on plugins and what is included in the image. I use 225Gb ssd‘s and as I build frequently for at least 6 receivers your ssd should not initially be a restriction

Keeping the downloads can significantly save on build time - so don‘t lose it
Once you start building it will remember where it reached and not start from the beginning

Also in case there is an issue in the OE-A its worth running a make update periodically

[Willo3092]

So when it fails I just run

MACHINE=dinobot4kplus DISTRO=openvix DISTRO_TYPE=release make image

again without doing anything else?


.... and when (or if) I get it to work, I don't have to start from scratch every time, just the last command and it will only update changes?

[twol]

Basically Yes - except that after some changes in the OE-A it will go back to the start or an earlier step - it just depends on what has changed. I always run a make update before I start the build, but you probably need to get at least one clean build done (unless there is an error in the OE-A) - once you get through this and want to continue then there are easier methods than typing everything in e.g. I use execs that list distro‘s and my receivers and I just click on the appropiate distro and receiver and it does the rest

[ronand]

A 120Gb ssd is plenty (and an ssd will speed up the process). Just make sure you have at least 4Gb ram - any less and it will fail.

[Willo3092]

Yes I'm very keen to learn. I'm just running it again and hopefully will find out what the issues are now that I have an idea where to look.

The box has 12gb of ram so should be fine once I iron out the issues.

[abu baniaz]

@willo3092, I have moved all your build questions/posts to this thread

120GB HDD drive run out of space for me pretty quickly. Should you wish to use a larger HDD, I suggest you use 4.2 branch

[twol]

Yes I'm very keen to learn. I'm just running it again and hopefully will find out what the issues are now that I have an idea where to look.

The box has 12gb of ram so should be fine once I iron out the issues.

When you have a build failure it should point to the log indicating the failure and if want to see overall whats happened look at the "cooker" log.
Because you are building a non standard ViX image, once built you need to look at how you handle the ipks (plugins library) - mine sit on a NAS and I use an apache server to access........ to automatically access from the receiver I change the feeds url in the openvix file (meta-oe-alliance/meta-oe/conf/distro/openvix) so it builds with my feed url rather than the ViX standard.

Anyway - good luck!!

[Willo3092]

I left it running overnight again and came to check the error log but it seems to have worked this time?
I haven't done anything different anyway. :confused:

58832

I've only had time to flash and scan for channels so far but seems to be working okay.

Where would I get the feed ipk's from? I have a NAS at home.

Thank you for all the help so far.

This is the image file if anyone wants to try it: https://tinyurl.com/y468g375

[twol]

Congratulation!
" left it running overnight again and came to check the error log but it seems to have worked this time? " - probably because there was a fetch error - and once downloaded you don't go there again unless you delete the download library.

"Where would I get the feed ipk's from?" - next to where you found the image
e.g my images are in "media/OEA/oea43/builds/openvix/release/u52/tmp/deploy/images and ipks's in media/OEA/oea43/builds/openvix/release/u52/tmp/deployipk

If you don't change the feed url in the openvix.conf (mentioned above) then you have to manually edit all the receiver ipk conf files in /etc/opkg

[norhap]

I left it running overnight again and came to check the error log but it seems to have worked this time?
I haven't done anything different anyway. :confused:

58832



Where would I get the feed ipk's from? I have a NAS at home.



install MariaDB in your nas and you will have your own feed (change feed url in openvix.conf)

change the official url vix for your ip nas, then put all the folders that you get in tmp/deploy/ipk/ inside the indexed folder in your nas mariadb and you will have your own source.

[norhap]

(MariaDB) should also be like an installer package through its interface is simple to just run and install.

create folder indexed in your nas through the user interface of your nas ... name folder = feeds. (you give it that name).

now you only have to copy all the folders within the /tmp/deploy/ipk/* environment to your nas ... web/feeds/openvix-5.2/

[Willo3092]

I've setup some temporary feeds because I'm off to work now.

This is the image and feeds for anyone who wants to try it:

https://tinyurl.com/y468g375

Many thanks to everyone for helping me :thumbsup:

[twol]

(MariaDB) should also be like an installer package through its interface is simple to just run and install.

create folder indexed in your nas through the user interface of your nas ... name folder = feeds. (you give it that name).

now you only have to copy all the folders within the /tmp/deploy/ipk/* environment to your nas ... web/feeds/openvix-5.2/

Depends on your NAS I guess, but on my Synology NAS I just enabled Apache Server and setup a feeds folder under the WEB directory, by carefully editing the url and file names in thh Openvix conf file you can reduce the naming structure to make life easier.
Also may need to allow nfs access from the receiver(s) on your NAS.

[norhap]

I also have Synology, but I use MariaDB which are in my synology repositories, with double click install is very easy, and I agree with what you think, I simplify url route and opkg-feeds works very well.

that video is obsolete, it shows installing web station + MariaDB .. now only the installation of MariaDB leaves the web server ready.
view video

https://youtu.be/hA_gJl4PE9s

[Willo3092]

My NAS is Synology so will have a try tomorrow.
I've also ordered a 480gb SSD which is arriving tomorrow so I won't have to keep deleting stuff :)

[twol]

@willo3092 - you have done a great job in no time

[Huevos]

If you don't change the feed url in the openvix.conf (mentioned above) then you have to manually edit all the receiver ipk conf files in /etc/opkgI used to change openvix.conf but it is a pain in the neck when the build numbers change. Now I do it in site.conf (both feeds and enigma git location).

.
SCONF_VERSION = "1"
BB_NUMBER_THREADS = "32"
PARALLEL_MAKE = "-j 16"
BUILD_OPTIMIZATION = "-march=native -O2 -pipe"
DL_DIR = "/home/openvix/sources"
INHERIT += "rm_work"
ENIGMA2_URI = "git://github.com/Huevos/enigma2.git;protocol=git;branch=Dev"
DISTRO_FEED_URI = "http://*****.org/feeds/${DISTRO_NAME}/${DISTRO_TYPE}/${DISTRO_VERSION}/${MACHINE}"


Also, I just symlink to the feeds so they are available to the webserver without moving any files.

[abu baniaz]

This thread is about building images. Can you please post in the dinoboot thread about dinoboot issues?

I split the thread so things can be specific. This was so it is easy to to follow what is going on. If you want to bundle everything together, I can merge the threads again.

[twol]

I used to change openvix.conf but it is a pain in the neck when the build numbers change. Now I do it in site.conf (both feeds and enigma git location).

.
SCONF_VERSION = "1"
BB_NUMBER_THREADS = "32"
PARALLEL_MAKE = "-j 16"
BUILD_OPTIMIZATION = "-march=native -O2 -pipe"
DL_DIR = "/home/openvix/sources"
INHERIT += "rm_work"
ENIGMA2_URI = "git://github.com/Huevos/enigma2.git;protocol=git;branch=Dev"
DISTRO_FEED_URI = "http://*****.org/feeds/${DISTRO_NAME}/${DISTRO_TYPE}/${DISTRO_VERSION}/${MACHINE}"


Also, I just symlink to the feeds so they are available to the webserver without moving any files.

Thanks Huevos, never thought of sticking that info in site.conf - really a great time saver. When I was totally PC based I used the symlinks which work really well in that environment.

Really envious of the 16 threads!!!

[dodgydrains]

This thread is about building images. Can you please post in the dinoboot thread about dinoboot issues?

I split the thread so things can be specific. This was so it is easy to to follow what is going on. If you want to bundle everything together, I can merge the threads again.

Sorry buddy, I thought I was helping @Will03092 with his image.

[twol]

Sorry buddy, I thought I was helping @Will03092 with his image.

Running an image build is different from making changes to the image which is why it should be treated on a different thread.

[abu baniaz]

I have pruned this thread again. Please keep it specific to building images

For any dinobot test images,feedback and comments, please use the other existing thread for it

https://www.world-of-satellite.com/showthread.php?61269-Dinobot-4K-Any-chance-of-Openvix-Support

Thank you

[Willo3092]

I seem to able to manage the build okay now, the only issue is hosting the feeds.
I've been busy with work lately and Mrs Willo seems to be off on the same days as me, so limited time to faff with it.
I have a Synology NAS and would ideally like to host the feeds from that. I only use it as a home NAS at the minute so will have to have a look into the hosting side of things when I get time.
Thanks to everyone for their input so far:) :thumbsup:

[Huevos]

Why do you want to host the feeds on your NAS? I don't understand not hosting them on the build machine.

[abu baniaz]

PC is not on all the time.

[Willo3092]

I just like to keep all of my work on the NAS and I thought it would be easier to share the feeds from the NAS for other people to test.
Ideally I would also like to try and setup a website at some point.
I had a nasty experience with ransomware lately (I appreciate it's unlikely to affect a Linux box) and had my files in windows encrypted.
I'm a bit obsessed now with moving anything I want to keep, to my NAS and not having mapped drives on the PC.
I'm not particularly familiar with Linux to be fair but I've managed to be able to copy stuff from Linux to the NAS so that it is accessible from both boxes.

[twol]

Why do you want to host the feeds on your NAS? I don't understand not hosting them on the build machine.
Same as ABU - my NAS runs for about 18 hours a day (and in standby overnight) compared to 8 _> 10 hour on the Ubuntu PC and I also keep all my local git and build source files etc also on the NAS.

[Huevos]

Is the NAS webserver public? Is the move process automated? Personally I don't see a greater security risk on the Ubuntu machine than the NAS. BTW, I run all my websites from the NAS. Reason is the old PC based webserver used 1100 kW/h per year and the NAS only uses 85 kW/h per year.

But for me it is easier not moving the feeds from the build server. Build server is a friend's machine that was idle 24/7 and I was lucky enough to be loaned it.

[Willo3092]

No, it's not public. I have just used it for file/movie/music storage at home so far.
I've only just started exploring the possibility of hosting my own website on it.
I've also only just started using a spare PC for this build project and whilst it is plenty fast enough, I have limited storage space.

[Valiant]

I have built 5.3.001 release for my MBTwinPlus on my HP laptop over my WiFi network. It took 12 hours overnight, with quite a few warnings generated but no errors. Found the "cooker" log which said that the build had finished and found the release files under the g300 directory. I haven't had the guts to flash the receiver yet, I'll do it when I've caught up with my viewing.

Abu and I think Sicilian have both said that if you want to have the latest images on the end-of-life models, go build it yourself - so I did(!), following on from Willo's success and the great advice of twol and everyone else on this thread. I will test the build eventually, but I suppose the real choices I have are to continue with OpenVix 5.2 on the MB or upgrade the box as eventually the MB will not be able to handle the memory requirements of later releases. And since none of the testers or full-time contributors to the forum will have an MBTwinPlus set up with OpenVix 5.3 or later, they will not be able to support the configuration.

[Willo3092]

I found that after a successful build it would give about 39 warnings, if you just run it again it's much quicker and a lot of the warnings are resolved.
Best I've managed is 1 warning but the image runs fine.

[Huevos]

Don't worry about build warnings.

Just flash the image and use it.

And don't forget to set up your own feeds.

[birdman]

I found that after a successful build it would give about 39 warnings, if you just run it again it's much quicker and a lot of the warnings are resolved.The warnings are not resolved - they just belong to packages that aren't being rebuilt on a rebuild, so don't show up.

[Valiant]

And don't forget to set up your own feeds.

Thanks for the advice - it is encouraging. However I don't know anything about feeds. I expect plugin feeds to work appropriately after flash.
Are you referring to moving the sources to somewhere safe and pointing to them with the site.conf file?
Otherwise, I have seen a feeds.conf file somewhere referring to local packages but I don't know what this means to my setup.

[twol]

So in the build directory you find the receiver directory and there is an Image and IPK directory ... e.g. builds/......../tmp/deploy/images and ipk
ipk is where the feeds are situated ...
so in meta-oe-alliance/meta-oe/conf/distro/openvix I have following line which points to my NAS box and is built into the Image, so when flashed it looks at my feeds. (Huevos has another method posted in a post)

DISTRO_FEED_URI ?= "${@bb.utils.contains("DISTRO_TYPE", "release", "http://192.168.0.171/feeds/${DISTRO_NAME}/${DISTRO_TYPE}/${DISTRO_VERSION}/${MACHINE}" , "http://192.168.0.171/feeds/${DISTRO_NAME}/release/

[Valiant]

So in the build directory you find the receiver directory and there is an Image and IPK directory ... e.g. builds/......../tmp/deploy/images and ipk
ipk is where the feeds are situated ...

Thanks twol, so do I move these ipks to safety and point the feeds.conf to them?

[twol]

Thanks twol, so do I move these ipks to safety and point the feeds.conf to them?

I know Huevos keeps them where they are - just remember that when you flash you need to be able to access them.
Also every time you run a build any updates will be added - so feeds are constantly changed.

[Valiant]

Thanks twol -some things are becoming clearer now about the structure of feeds and packages, next step is to flash and test anyway and then manage my builds as per the advice given here.

Thanks for all your assistance everyone.

[birdman]

so in meta-oe-alliance/meta-oe/conf/distro/openvix I have following line which points to my NAS box and is built into the Image, so when flashed it looks at my feeds. So you do also need to have a Web server running there configured to serve the correct locations.

[twol]

So you do also need to have a Web server running there configured to serve the correct locations.
Yes, good point I have Apache running on my Synology

[Huevos]

Don't touch anything in openvix.conf.

Put your personal feeds URL in site.conf

DISTRO_FEED_URI = "http://xxxxxxx/feeds/${DISTRO_NAME}/${DISTRO_TYPE}/${DISTRO_VERSION}/${MACHINE}"

-------------------------------------------------------------------------

I just leave all the files in place on the build server and symlink them to the Apache html root folder. For me no point moving them to another machine.

mkdir -p /var/www/feeds/openvix/experimental/5.3/osninopro

ln -s /home/openvix/5.3/builds/openvix/experimental/osninopro/tmp/deploy/ipk/* /var/www/feeds/openvix/experimental/5.3/osninopro

[Willo3092]

Just started to get build failed errors.
Hopefully this is the correct log.59437

[twol]

Just started to get build failed errors.
Hopefully this is the correct log.59437

there was a change in vix-core/src/setup.xml that caused a build issue - it was fixed a few hours ago - so if you keep local copies you need to update else it should now work

[lincsat]

I've just had a go at compiling for the Unibox HDE as it's been a while since images were released. It failed on an error - gstreamer1.0-plugins-good_git.bb:do_compile I think this is the right logfile

Is this box just too old to make an image for or is it something I've done wrong?

[birdman]

I've just had a go at compiling for the Unibox HDE as it's been a while since images were released. It failed on an error - gstreamer1.0-plugins-good_git.bb:do_compile I think this is the right logfile

Is this box just too old to make an image for or is it something I've done wrong?What's really needed is the log file mentioned in the ERROR: message, namely:

/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/blackbox7405/tmp/work/mips32el-oe-linux/gstreamer1.0-plugins-good/gstreamer1.0-plugins-good-1.14.4+gitAUTOINC+6fe214e7a9-r0/temp/log.do_compile.14153

[lincsat]

OK, tried again and exited at a different location, this is the log file mentioned in the error

[lincsat]

As an update, I tried again and it failed at a different stage again, so kept restarting the make and after 4 more fails, it actually completed - just got to get around to flashing and setting up now :)

[twol]

As an update, I tried again and it failed at a different stage again, so kept restarting the make and after 4 more fails, it actually completed - just got to get around to flashing and setting up now :)

I guess my question is how much memory do you have on the box and have you done this? -
modify max_user_watches

echo fs.inotify.max_user_watches=524288 | sudo tee -a /etc/sysctl.conf
sysctl -n -w fs.inotify.max_user_watches=524288

[birdman]

I guess my question is how much memory do you have on the box and have you done this? -
modify max_user_watches

echo fs.inotify.max_user_watches=524288 | sudo tee -a /etc/sysctl.conf
sysctl -n -w fs.inotify.max_user_watches=524288If you don't do that and you need it then you get an error at the start of the build and nothing useful happens. So unlikely to be the problem here.

[lincsat]

I entered the command detailed on the ATV Git page
echo fs.inotify.max_user_watches=524288 | sudo tee -a /etc/sysctl.conf

sysctl -n -w fs.inotify.max_user_watches=524288

The machine has 16GB of RAM

I did afterwards run a build for one of the unsupported ZGemma boxes and it completed without error.

[lincsat]

Is it possible to build a Vu+ image, they don't appear to be mentioned in the makefile. If it is then what's the machine name for the Ultimo4K?

[abu baniaz]

I think it's

vuultimo4k

[lincsat]

I think it's

vuultimo4k

That appears to be building M8, thanks

[Willo3092]

I keep getting build failed errors suddenly.
It mentions libbluray in the error message.
I've tried multiple times just in case it's a site down or something but no luck.

[twol]

I keep getting build failed errors suddenly.
It mentions libbluray in the error message.
I've tried multiple times just in case it's a site down or something but no luck.

have you moved your sources folder recently? I seem to remember that there are some symbolic links formed when this library is built.... I had a similar issue way back with this library but was able to work out what was missing
You might have to delete it from your sources and get it to be re-fetched and the source files then re-built during the build..

[Willo3092]

have you moved your sources folder recently? I seem to remember that there are some symbolic links formed when this library is built.... I had a similar issue way back with this library but was able to work out what was missing
You might have to delete it from your sources and get it to be re-fetched and the source files then re-built during the build..

Spot on as usual :thumbsup: I'd created a new folder to build in and just copied the sources and site.conf over.
Back up and running now thanks :D

[lincsat]

I started to get a failure last night, similar lib errors. It mentions libs from git.videolan.org looking at their site, they are moving to code.videolan.org could that be the problem?


ERROR: libudfread-1.0.0+gitAUTOINC+131629921c-r0 do_fetch: Fetcher failure for URL: 'git://git.videolan.org/libudfread.git;branch=master;protocol=git'. Unable to fetch URL from any source.

I did try a make-update and even deleted everything and started again but the problem persists

[lincsat]

As a follow up, I reported an issue on the OE-Alliance Git and they have updated the URL. I'll try another build later

[Willo3092]

Failed build error again. I haven't changed my sources folder this time :)

[birdman]

Failed build error again. I haven't changed my sources folder this time :)But the actual error is in a different log file:


ERROR: Logfile of failure stored in: /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/temp/log.do_compile.6590

[Willo3092]

Found it. Had to zip it because it doesn't have a .log extension.

[twol]

I would run a "make update" and then try an image build again.

[Willo3092]

I'll try again although I always do that before starting the build.

[twol]

I'll try again although I always do that before starting the build.

I suppose you have checked the freespace on the device?! Its not unknown to run out of space on these repeated builds.

[Willo3092]

Failed again with the same error. 288Gb free space.

I may try starting from scratch again later on but got to go out now.

[birdman]

Found it. Had to zip it because it doesn't have a .log extension.The problem(s) are:


/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/moc_customsqlmodel.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/moc_stickman.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/xbelwriter.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/submarine.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/moc_window.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/metainfo.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/tst_qmetatype.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/tst_qdbusinterface.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/tst_qdbusmetatype.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/main.o: file not recognized: file truncated
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/u52/tmp/work/u52-oe-linux-gnueabi/qtbase/qtbase-5.9.8+gitAUTOINC+82eb6aa08e-r0/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi/../../libexec/arm-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/8.2.0/ld: .obj/tst_qimagewriter.o: file not recognized: file truncated
No idea on a solution.

[Willo3092]

Still no luck I'm afraid.
I've started from scratch and Mipsel and Arm7 images build fine but the Dinobot 4K+ build still fails with the same error message.
I also tried to build an image for the SF8008 and that also fails with the same error message.
Could it be because they are more modern processors?

[twol]

I build sf8008 regularly - although on OE-A 4.4 and I am sure Dinobot 4K would build as well (just haven‘t done it for a while)
Maybe the QT source files are corrupted - you should probably delete from your sources file and try again

[ronand]

I'd say the sources are corrupted too. I just built the SF8008 image a few minutes ago without problems.

[Willo3092]

I'll give it a go after work, thanks guys :thumbsup:

I take it that the Arm7 and Mipsel boxes don't use the qt sources?

[twol]

I'll give it a go after work, thanks guys :thumbsup:

I take it that the Arm7 and Mipsel boxes don't use the qt sources?
Some Arm boxes do - some don‘t and it maybe different library/versions when they do, compared to the HiSilicon boxes

[abu baniaz]

To exclude all of the channel settings add the attached file (after unzipping) to meta-local/recipes-local folder. Thanks to captain.

EDIT: Please note change in path from what I had posted earlier

[twol]

Neat - I just hacked it from the feeds

[abu baniaz]

To pre-install some plugins, add the attached file (after unzipping) to meta-local/recipes-local folder. Thanks to Huevos.

Note: Tabs should not be used in bb files

[Willo3092]

Can anyone help with this error?
I get the same error building on OE4.3 and 4.4 and on both ViX and ATV builds.
I've updated to Ubuntu 18.04.4 LTS and started from scratch several times but still the same.
I've also built new sources several times but no luck.
I've googled the error and some have said to omit it from the recipe but I don't know how to do that.

This is the error message and logs are attached:

ERROR: Task (/home/openvixbuilder/VixBuild/build-enviroment/meta-oe-alliance/meta-oe/recipes-extended/astra-sm/astra-sm_0.2.bb:do_compile) failed with exit code '1'

Thanks in advance :)

[Huevos]

I am running 16.04 and GCC 8.3.0.

Make sure your GCC is at least this version.



openvix@ubuntu01:~$ gcc --version
gcc (Ubuntu 8.3.0-16ubuntu3~16.04) 8.3.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

openvix@ubuntu01:~$


Also read these 2 pages...
https://askubuntu.com/questions/1028601/install-gcc-8-only-on-ubuntu-18-04
https://askubuntu.com/questions/26498/how-to-choose-the-default-gcc-and-g-version

[birdman]

I've updated to Ubuntu 18.04.4 LTS Updated to? 20.04 LTS is due out in a a week or two. I;m on 19.10, which works.


This is the error message
No, it isn't. The actual error is in the do_compile log you posted and is:


gcc: error: unrecognized command line option ‘-fmacro-prefix-map=/home/openvixbuilder.....

The "needs gcc 8" is probably correct - gcc 7.5 certainly doesn't have the option, 8.4 does.

If you wonder what it does, it's this:


-fmacro-prefix-map=old=new
When preprocessing files residing in directory old, expand the "__FILE__" and
"__BASE_FILE__" macros as if the files resided in directory new instead. This can be
used to change an absolute path to a relative path by using . for new which can result
in more reproducible builds that are location independent. This option also affects
"__builtin_FILE()" during compilation. See also -ffile-prefix-map.

[abu baniaz]

Have you checked the ATV readme file changes for 6.4 branch?

The one on 13 October may be very relevant

https://github.com/openatv/enigma2/commits/6.4/README.md

[Willo3092]

Thanks all, I'll have a go tonight hopefully :thumbsup:

[ronand]

I wiped my drive and "upgraded" to 18.04 on Sunday evening as I was having trouble building the 4.4 branch.. I followed the instructions on the ATV github and I was able to build 4.3 and 4.4 branches from scratch yesterday.

[Willo3092]

Have you checked the ATV readme file changes for 6.4 branch?

The one on 13 October may be very relevant

https://github.com/openatv/enigma2/commits/6.4/README.md


Back up and running again. Thanks to all :thumbsup:
I followed the updated instructions on ATV 6.4 which includes updating the GCC version :)

[lincsat]

Now getting build errors again, looks like something broken at git.linuxtv.org


ERROR: edid-decode-1.0+gitAUTOINC+56dd103a0c-r2 do_fetch: Fetcher failure: Fetch command export PSEUDO_DISABLED=1; export DBUS_SESSION_BUS_ADDRESS="unix:path=/run/user/1001/bus"; export PATH="/home/openvixbuilder/openvix/build-enviroment/openembedded-core/scripts:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/cortexa15hf-neon-vfpv4-oe-linux-gnueabi/edid-decode/edid-decode-1.0+gitAUTOINC+56dd103a0c-r2/recipe-sysroot-native/usr/bin/arm-oe-linux-gnueabi:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/cortexa15hf-neon-vfpv4-oe-linux-gnueabi/edid-decode/edid-decode-1.0+gitAUTOINC+56dd103a0c-r2/recipe-sysroot/usr/bin/crossscripts:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/cortexa15hf-neon-vfpv4-oe-linux-gnueabi/edid-decode/edid-decode-1.0+gitAUTOINC+56dd103a0c-r2/recipe-sysroot-native/usr/sbin:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/cortexa15hf-neon-vfpv4-oe-linux-gnueabi/edid-decode/edid-decode-1.0+gitAUTOINC+56dd103a0c-r2/recipe-sysroot-native/usr/bin:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/cortexa15hf-neon-vfpv4-oe-linux-gnueabi/edid-decode/edid-decode-1.0+gitAUTOINC+56dd103a0c-r2/recipe-sysroot-native/sbin:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/cortexa15hf-neon-vfpv4-oe-linux-gnueabi/edid-decode/edid-decode-1.0+gitAUTOINC+56dd103a0c-r2/recipe-sysroot-native/bin:/home/openvixbuilder/openvix/build-enviroment/bitbake/bin:/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/hosttools"; export HOME="/home/openvixbuilder"; LANG=C git -c core.fsyncobjectfiles=0 fetch -f --prune --progress https://git.linuxtv.org/cgit.cgi/edid-decode.git refs/*:refs/* failed with exit code 128, output:
warning: redirecting to https://git.linuxtv.org//edid-decode.git/
error: garbage at end of loose object 'ada8d54fa0dd28c18763f41632502207bda56392'
fatal: loose object ada8d54fa0dd28c18763f41632502207bda56392 (stored in ./objects/ad/a8d54fa0dd28c18763f41632502207bda56392) is corrupt
error: Unable to find 56dd103a0c20724ee956950f5bcb8cc1c8667af9 under https://git.linuxtv.org//edid-decode.git
Cannot obtain needed object 56dd103a0c20724ee956950f5bcb8cc1c8667af9
error: fetch failed.

ERROR: edid-decode-1.0+gitAUTOINC+56dd103a0c-r2 do_fetch: Fetcher failure for URL: 'git://git.linuxtv.org/cgit.cgi/edid-decode.git;protocol=https'. Unable to fetch URL from any source.
ERROR: edid-decode-1.0+gitAUTOINC+56dd103a0c-r2 do_fetch: Function failed: base_do_fetch

[twol]

As a comment, hope you don‘t throw away your sources directory after every build? Unless its been updated recently should not need to fetch

[lincsat]

As a comment, hope you don‘t throw away your sources directory after every build? Unless its been updated recently should not need to fetch

I don't think I've ever deleted the sources folder, I do a make update before every build session. Do you think it's worth deleting the sources and trying again?

[twol]

No always keep your sources - just surprised it was not already downloaded

[Sicilian]

Re-start the build, git was most likely down.

[lincsat]

Re-start the build, git was most likely down.

I've tried again today, 3rd Day now with the same error

[birdman]

error: Unable to find 56dd103a0c20724ee956950f5bcb8cc1c8667af9 under https://git.linuxtv.org//edid-decode.gitThat's a little odd.
If you clone the git repository (it's up and available) you'll find that 56dd103a0c20724ee956950f5bcb8cc1c8667af9 is the HEAD commit (from July 20).
So it is there...

[lincsat]

Well after deleting the sources and re-cloning the GIT, it's finally completed the compile. So about time to get the 4.4 build environment ready for the new images

[lincsat]

I've just tried to build the latest developer image with the 4.4 build environment and am repeatedly getting the same 2 failures, NFS-Utils and Webkit-gtk.

To get as much of the build as possible completed, I used the -k (keep going) argument


cd /home/openvixbuilder/openvix44/build-enviroment/builds/openvix/developer/vuultimo4k ; source env.source ; bitbake -k openvix-image

I have attached what I believe are the correct logs. I've not tried building a developer image before, so may have screwed something up. Are you guys succeeding in building the latest developer image?

[abu baniaz]

Which plugins are you adding to the standard build process?


Also, I do following. Adapt/remove the distrotype to suit



cd /home/openvixbuilder/openvix44/build-enviroment/
make update
MACHINE=vuultimo4k DISTRO=openvix DISTRO_TYPE=homebuild make image

[lincsat]

I'm just using the standard build, not tried adding or removing anything yet.

I'm still a novice when it comes to compiling, I assume with your homebuild, you can specify which components to compile and add

[abu baniaz]

Homebuild is just name of image. You can call it whatever you want. The component to add/remove was listed elsewhere in this thread. Post 101/103

Anyway, the ultimo4k dev build built fine. Which Ubuntu version are you using?

[ronand]

Ideally you would want to be on version 20 of Ubuntu. No problems here building the dev build of Vix 5.4 or OpenATV 6.5.

[lincsat]

Homebuild is just name of image. You can call it whatever you want. The component to add/remove was listed elsewhere in this thread. Post 101/103

Anyway, the ultimo4k dev build built fine. Which Ubuntu version are you using?

I'm on 18.04 LTE. I saw those posts to add/remove components from the image, I though you were talking about removing components from the compile.

I'll have another go at compiling from clean and if all else fails, look to upgrading to Ubuntu 20

[bbbuk]

I built successfully vix 4.4 branch but "release" version. This was on 18.04.4 LTS

I'll build the developer now and post back.

[bbbuk]

I built successfully vix 4.4 branch but "release" version. This was on 18.04.4 LTS

I'll build the developer now and post back.Successfully built developer version without any issues on 18.04.4 LTS

[birdman]

I have attached what I believe are the correct logs. I've not tried building a developer image before, so may have screwed something up. Are you guys succeeding in building the latest developer image?

nfs:

arm-oe-linux-gnueabi-gcc: error: unrecognized -march target: nativeHow did that get there? It was removed from the Makefile for 4.4.

webkit:

....arm-oe-linux-gnueabi/9.2.0/ld: ./.libs/libwebkitgtk-1.0.so: undefined reference to `WebCore::AccessibilityRenderObject::de|ach()'That de|ach looks suspiciously like it should be detach. That's a 1-bit error. Which could be a CPU or disk error?
Do you have any other issues on this system?

[ronand]

A failing hard drive will cause all sorts of problems like this. Its tempting to use an old machine/drive for these purposes especially if you are not normally using linux.

If you want to start again on another drive/clean system the steps needed to set up the build environment are listed at
github.com/openatv/enigma2 - just substitute openvix for openatv

The only errors you should really be seeing are fetch errors when sources are temporarily unavailable.

[twol]

ronand is correct - just check which OpenATV enigma2 branch you are looking at - the default for OpenATV is 6.4 and has build info for Ubuntu 18.04. ATV branch 6.5 is building on Ubuntu 20.04 and has different specs mainly because they are using it to build a python 3 image.

[lincsat]

I built the release version OK after cleaning and re-cloning the sources, will try again with developer.

The system isn't new but not old. Water cooled Intel i7, 8 core, 16 Gig RAM, 240GB SSD.

I can't imagine a processor or SSD error would cause the same error on several different attempts, something corrupted on the download sounds more likely especially as it seems to be working after a fresh download - thanks for all the suggestions

[ronand]

If it works after a fresh download that indicates possibly something in your sources folder was corrupted (usually a HDD error in my experience). It would probably be no harm to test the ram and ssd to be sure. Anyway don't let that put you off!

[BrianTheTechieSnail]

I built the release version OK after cleaning and re-cloning the sources, will try again with developer.
The system isn't new but not old. Water cooled Intel i7, 8 core, 16 Gig RAM, 240GB SSD.
I can't imagine a processor or SSD error would cause the same error on several different attempts, something corrupted on the download sounds more likely especially as it seems to be working after a fresh download - thanks for all the suggestions
Please, may I ask how long it takes to build on that system?
I'm thinking of having a go at this myself.

[ronand]

An initial build might take around 10-12 hours on such a system - further builds don't take too long as most of the downloading and compiling doesn't need to be redone.

[birdman]

An initial build might take around 10-12 hours on such a system.My 6+ years old i5 laptop only takes ~4 hours.

[lincsat]

Can I ask, if I get future errors, is it possible to just re-download a certain component? If so, would I just delete it from the sources folder and run make update again or would something else be needed?

[twol]

Never had to do it, but basically when the build downloads a component there is a also a 2nd folder stored with the same component name plus .done - so delete the 2nd folder

[birdman]

Can I ask, if I get future errors, is it possible to just re-download a certain component? If so, would I just delete it from the sources folder and run make update again or would something else be needed?It is if you set up the environment and run a suitable bitbake command.

I can post a script that lets you do this, if you're interested.

[lincsat]

It is if you set up the environment and run a suitable bitbake command.

I can post a script that lets you do this, if you're interested.

That would be useful - thanks

[birdman]

That would be useful - thanksHere it is.

60465

It will require some editing, as it defines root to be where the build goes (so also hardwires a developer build, as that's the tree I copied this script from.).
It takes the machine architecture to work on from the MACHINE environment variable (I was being quick and lazy), so you could also hard-wire that if you're only ever going to work on one.

[Huevos]

There is no necessity to update Ubuntu for branch 4.4.

Ubuntu 16.04 works fine.

[ccs]

I installed ubuntu-16.04.1-desktop-i386 and used these commands in terminal:

#1 - Install packages on your buildserver
sudo apt-get install -y autoconf automake bison bzip2 curl cvs diffstat flex g++ gawk gcc gettext git-core gzip help2man ncurses-bin ncurses-dev libc6-dev libtool make texinfo patch perl pkg-config subversion tar texi2html wget zlib1g-dev chrpath libxml2-utils xsltproc libglib2.0-dev python-setuptools zip info coreutils diffstat chrpath libproc-processtable-perl libperl4-corelibs-perl sshpass default-jre default-jre-headless java-common libserf-dev qemu quilt libssl-dev

That list has a lot of differences to the one mentioned here...


Hate to point to another image! .. but have a look at the bottom half of this page
https://github.com/openatv/enigma2

In the git change from OpenATV to Openvix and away you go.

I've just had a play, and a number (loads) of packages on both lists must have failed for one reason or another. eg gcc-8

Sorry, I've not kept any logs, I was just testing Ubuntu on a usb stick, it worked fine, but somehow corrupted windows searching when I rebooted to windows 10.
I had copied one file from the windows filesystem to Ubuntu, surely if Ubuntu lets me do it there shouldn't be a problem?
I had to restore a backup image to get it working again.

[lincsat]

I'm still getting build errors and whatever I do, I can't get past an error building qemu-native, even the do-it doesn't compile it.

I've even got a new SSD and installed Ubuntu 20 from scratch and just added the 4.4 build environment. Tried release and developer images for the vuultimo4k and the zgemmah7, all fails at the same point. I did notice several warnings for unable to get checksum file, but that's usual. I'll have another go after the Weekend in case there is a down depositary. It's more likely just me doing something stupid though.

[Willo3092]

That list has a lot of differences to the one mentioned here...

That was the cause of my issues. I then updated to Ubuntu 18.04.4 LTS and followed the insructions here:

https://github.com/openatv/enigma2/blob/ca408186ba05adc77d84933e189d6a1e636ee011/README.md

Then made sure I had gcc-8 by following the instructions here:

https://askubuntu.com/questions/1028601/install-gcc-8-only-on-ubuntu-18-04

I've managed to build 5 OpenViX 5.4.000 release images now.

So to make a developer image is it just a matter of changing 'release' to 'developer' in the script?

MACHINE=zgemmah9combo DISTRO=openatv DISTRO_TYPE=release make image

... and what difference is there?

[birdman]

That was the cause of my issues. I then updated to Ubuntu 18.04.4 LTS...Building on 20.04 LTS (which is 2-years newer....and has been out for a few months) also works fine.

[ccs]

Building on 20.04 LTS (which is 2-years newer....and has been out for a few months) also works fine.

That's what I was using, I'll try again when time/the weather permits.

[lincsat]

That's what I was using, I'll try again when time/the weather permits.

Yes, me too - even an openatv build is failing for me right now.

[ccs]

Yes, me too - even an openatv build is failing for me right now.

... but I fell flat on my face at the first hurdle. :)

[ccs]

... but I fell flat on my face at the first hurdle. :)

This is were I get to, running 20.04 from a usb stick, absolutely bog standard.


ubuntu@ubuntu:~$ sudo apt-get install -y autoconf automake bison bzip2 chrpath coreutils cpio curl cvs debianutils default-jre default-jre-headless diffstat flex g++ gawk gcc gcc-8 gettext git git-core gzip help2man info iputils-ping java-common libc6-dev libegl1-mesa libglib2.0-dev libncurses5-dev libperl4-corelibs-perl libproc-processtable-perl libsdl1.2-dev libserf-dev libtool libxml2-utils make ncurses-bin patch perl pkg-config psmisc python3 python3-git python3-jinja2 python3-pexpect python3-pip python-setuptools qemu quilt socat sshpass subversion tar texi2html texinfo unzip wget xsltproc xterm xz-utils zip zlib1g-dev
Reading package lists... Done
Building dependency tree
Reading state information... Done
Note, selecting 'git' instead of 'git-core'
Package xterm is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

Package quilt is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

Package python3-pip is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

Package subversion is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

Package texinfo is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source
However the following packages replace it:
install-info info

Package qemu is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

E: Unable to locate package chrpath
E: Unable to locate package cvs
E: Unable to locate package gcc-8
E: Unable to locate package help2man
E: Unable to locate package libegl1-mesa
E: Unable to locate package libproc-processtable-perl
E: Unable to locate package libsdl1.2-dev
E: Couldn't find any package by glob 'libsdl1.2-dev'
E: Couldn't find any package by regex 'libsdl1.2-dev'
E: Unable to locate package libserf-dev
E: Unable to locate package python3-git
E: Package 'python3-pip' has no installation candidate
E: Unable to locate package python-setuptools
E: Package 'qemu' has no installation candidate
E: Package 'quilt' has no installation candidate
E: Unable to locate package sshpass
E: Package 'subversion' has no installation candidate
E: Unable to locate package texi2html
E: Package 'texinfo' has no installation candidate
E: Package 'xterm' has no installation candidate
ubuntu@ubuntu:~$

Even had to post on here using Firefox/Ubuntu because I couldn't save the text anywhere.:smiley_yup:

[lincsat]

Looks like it can't install the required packages due to them being in use in the portable system. If you have a spare HDD or SSD, try installing on that. I've managed to get all the pre-requisites done, except for the
Use update-alternatives for having gcc redirected automatically to gcc-8 as It appears to be already set. I've also completed the step detailed in here -
https://github.com/openatv/enigma2/issues/1780

The builds (openvix and openatv, ZGemma H7 and Vu+ Ultimo4k) start but always bug out with errors, even after several re-runs, so I'm guessing a depository is offline and I'll try again after the weekend

[ccs]

Yes - the stick is mounted read-only, the internal hd is mounted as /media/ubuntu/....... and will be why I can save files in the home directory, but not on the stick.

I'd have thought that the error messages (or lack of) would have indicated that the filesystem was read-only??

[lincsat]

Yes - the stick is mounted read-only, the internal hd is mounted as /media/ubuntu/....... and will be why I can save files in the home directory, but not on the stick.

I'd have thought that the error messages (or lack of) would have indicated that the filesystem was read-only??

Yes, to stop you changing system files on the stick

[abu baniaz]

......I'm guessing a depository is offline and I'll try again after the weekend
That would have given you a fetch error.

[ccs]

... but I fell flat on my face at the first hurdle. :)

Full Ubuntu install has got past the first hurdle. :)

[lincsat]

That would have given you a fetch error.

Some of the warnings that flashed through were fetch errors and unable to locate checksum files.

[abu baniaz]

you can ignore warnings, errors stop the build.

[lincsat]

Well whatever I try and after multiple build attempts using the -k (keep going) argument, there are 3 components that don't build for the ZGemma H7


Summary: 3 tasks failed:
/home/openvixbuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu/qemu-native_4.1.0.bb:do_compile
/home/openvixbuilder/openvix/build-enviroment/openembedded-core/meta/recipes-support/boost/boost_1.71.0.bb:do_compile
/home/openvixbuilder/openvix/build-enviroment/meta-qt5/recipes-qt/qt5/qtbase_git.bb:do_compile

For the Vu+ Ultimo4k, it's just the qemu-native that fails

@CCS have you managed to complete a build or at least do qemu-native

[ccs]

... this is where I'm stuck (about 3 minutes ago)....


| /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/hosttools/ld.bfd: linux-user/syscall.o: in function `do_syscall':
| /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/x86_64-linux/qemu-native/qemu-native-3.0.0-r0/qemu-3.0.0/linux-user/syscall.c:8526: undefined reference to `stime'
| collect2: error: ld returned 1 exit status
| make[1]: *** [Makefile:199: qemu-i386] Error 1
| make: *** [Makefile:481: subdir-i386-linux-user] Error 2
| ERROR: oe_runmake failed
| WARNING: /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/x86_64-linux/qemu-native/qemu-native-3.0.0-r0/temp/run.do_compile.834121:1 exit 1 from 'exit 1'
| ERROR: Function failed: do_compile (log file is located at /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/x86_64-linux/qemu-native/qemu-native-3.0.0-r0/temp/log.do_compile.834121)
ERROR: Task (virtual:native:/home/openvixbuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu/qemu_3.0.0.bb:do_compile) failed with exit code '1'
NOTE: Tasks Summary: Attempted 1167 tasks of which 1009 didn't need to be rerun and 1 failed.

Summary: 1 task failed:
virtual:native:/home/openvixbuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu/qemu_3.0.0.bb:do_compile
Summary: There was 1 WARNING message shown.
Summary: There were 2 ERROR messages shown, returning a non-zero exit code.
make: *** [Makefile:1029: image] Error 1


Different qemu_3.0.0 to your's. ET10000

tmp/work/x86_64-linux/qemu-native/qemu-native-3.0.0-r0/qemu-3.0.0/linux-user/syscall.c:8526: undefined reference to `stime'


Selecting previously unselected package qemu.
Preparing to unpack .../152-qemu_1%3a4.2-3ubuntu6.3_amd64.deb ...
Unpacking qemu (1:4.2-3ubuntu6.3) ...

Setting up libpcrecpp0v5:amd64 (2:8.39-12build1) ...
Setting up qemu (1:4.2-3ubuntu6.3) ...
Setting up libglib2.0-dev-bin (2.64.3-1~ubuntu20.04.1) ...

[Willo3092]

I've built a release build for the H7 successfully (openvix-5.4.000.release-zgemmah7_usb)

[lincsat]

I get the same
undefined reference to `stime'

@ Willo3092 was your qemu already compiled? I managed to complete a build for the Ultimo4k a few Days ago but since doing a clean Ubuntu install and redownloading the build environment, I can't get that compiled for any image. I'm now wondering if a recent update broke that.

[ccs]

Maybe related......


Glibc 2.31 porting notes/stime removal
From the release notes:

* The obsolete function stime is no longer available to newly linked
binaries, and its declaration has been removed from <time.h>.
Programs that set the system time should use clock_settime instead.
Also note something which you may want to fix while in the area:

* We plan to remove the obsolete function ftime, and the header
<sys/timeb.h>, in a future version of glibc. In this release, the
header still exists but calling ftime will cause a compiler warning.
All programs should use gettimeofday or clock_gettime instead.

[Willo3092]

I get the same
undefined reference to `stime'

@ Willo3092 was your qemu already compiled?

Sorry, I have no idea what that means. I'm very much a bumbling poke and hope merchant :eek:
If a build fails, I try a build for something else similar. If that works, 9 times out of 10 the one that failed will now work.
I retired from work yesterday so I promise to try and learn about the stuff I'm dabbling with.

[ccs]

I'm using Ubuntu 20.04.....

Man page for stime could have the answer? .....


http://manpages.ubuntu.com/manpages/focal/man2/stime.2.html


Starting with glibc 2.31, this function is no longer available to newly linked
applications and is no longer declared in <time.h>.

[birdman]

For the Vu+ Ultimo4k, it's just the qemu-native that fails

@CCS have you managed to complete a build or at least do qemu-nativeHmmmm...there was a bug there. I may have applied a patch to my build. I reported it here...

Posts are (restricted links removed, replaced with actual text):


The build fails because of this known bug in qemu:


https://bugs.launchpad.net/qemu/+bug/1852115

Anyone with qemu already built may be OK - at least as far as compiling goes - but the qemu version needs to be bumped to fix this.




Originally Posted by birdman
It seems that I haven't submitted the patch yet though....I'll have to sort out which repository it is in.
It's in openembedded-core.

However, when I go to that on github I find that master there is configured for qemu 4.2 (which probably has the fix in already). But my fresh download of build-environment 4.4 ends up using qemu 4.1.
Any ideas as to why?


But if it were using qemu 4.2 (as it appears it should be?) this wouldn't be an issue.

I'll track down my fix....

[ccs]

.... they're all restricted access.:p

[birdman]

I'll track down my fix....

In oe-alliance/openembedded-core/meta/recipes-devtools/qemu

Edit qemu.inc to add:

file://0100-stime-call-gone.patch \
at the end of the SRC_URI (so after file://CVE-2019-15890.patch).

Add this file (unzipped) into the qemu sub-directory as 0100-stime-call-gone.patch

60475

[ccs]

I can't see any lines with file: in qemu.inc, they're in qemu_3.0.0.bb

Am I in the right directory??



openvixbuilder@Unix:~/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu$ ls -l
total 44
-rw-rw-r-- 1 openvixbuilder openvixbuilder 1211 Aug 1 12:31 nativesdk-qemu-helper_1.0.bb
drwxrwxr-x 2 openvixbuilder openvixbuilder 4096 Aug 1 12:31 qemu
-rw-rw-r-- 1 openvixbuilder openvixbuilder 2482 Aug 1 12:31 qemu_3.0.0.bb
drwxrwxr-x 2 openvixbuilder openvixbuilder 4096 Aug 1 12:31 qemu-helper
-rw-rw-r-- 1 openvixbuilder openvixbuilder 492 Aug 1 12:31 qemu-helper-native_1.0.bb
-rw-rw-r-- 1 openvixbuilder openvixbuilder 4971 Aug 1 12:31 qemu.inc
-rw-rw-r-- 1 openvixbuilder openvixbuilder 4971 Aug 1 18:32 qemu.inc.orig
-rw-rw-r-- 1 openvixbuilder openvixbuilder 914 Aug 1 12:31 qemu-targets.inc
-rw-rw-r-- 1 openvixbuilder openvixbuilder 794 Aug 1 12:31 qemuwrapper-cross_1.0.bb



cat qemu.inc

SUMMARY = "Fast open source processor emulator"
HOMEPAGE = "http://qemu.org"
LICENSE = "GPLv2 & LGPLv2.1"
DEPENDS = "glib-2.0 zlib pixman"
RDEPENDS_${PN}_class-target += "bash"

require qemu-targets.inc
inherit pkgconfig bluetooth
BBCLASSEXTEND = "native nativesdk"

# QEMU_TARGETS is overridable variable
QEMU_TARGETS ?= "arm aarch64 i386 mips mipsel mips64 mips64el ppc riscv32 riscv64 sh4 x86_64"

EXTRA_OECONF = " \
--prefix=${prefix} \
--bindir=${bindir} \
--includedir=${includedir} \
--libdir=${libdir} \
--mandir=${mandir} \
--datadir=${datadir} \
--docdir=${docdir}/${BPN} \
--sysconfdir=${sysconfdir} \
--libexecdir=${libexecdir} \
--localstatedir=${localstatedir} \
--with-confsuffix=/${BPN} \
--disable-strip \
--disable-werror \
--target-list=${@get_qemu_target_list(d)} \
--extra-cflags='${CFLAGS}' \
${PACKAGECONFIG_CONFARGS} \
"
EXTRA_OECONF_append_class-native = " --python=python2.7"

EXTRA_OEMAKE_append_class-native = " LD='${LD}' AR='${AR}' OBJCOPY='${OBJCOPY}' LDFLAGS='${LDFLAGS}'"

LDFLAGS_append_class-native = " -fuse-ld=bfd"

export LIBTOOL="${HOST_SYS}-libtool"

B = "${WORKDIR}/build"

do_configure_prepend_class-native() {
# Append build host pkg-config paths for native target since the host may provide sdlqemu_3.0.0.bb
BHOST_PKGCONFIG_PATH=$(PATH=/usr/bin:/bin pkg-config --variable pc_path pkg-config || echo "")
if [ ! -z "$BHOST_PKGCONFIG_PATH" ]; then
export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:$BHOST_PKGCONFIG_ PATH
fi
}

do_configure() {
${S}/configure ${EXTRA_OECONF}
}

do_install () {
export STRIP=""
oe_runmake 'DESTDIR=${D}' install
}

# The following fragment will create a wrapper for qemu-mips user emulation
# binary in order to work around a segmentation fault issue. Basically, by
# default, the reserved virtual address space for 32-on-64 bit is set to 4GB.
# This will trigger a MMU access fault in the virtual CPU. With this change,
# the qemu-mips works fine.
# IMPORTANT: This piece needs to be removed once the root cause is fixed!
do_install_append() {
if [ -e "${D}/${bindir}/qemu-mips" ]; then
create_wrapper ${D}/${bindir}/qemu-mips \
QEMU_RESERVED_VA=0x0
fi
}
# END of qemu-mips workaround

PACKAGECONFIG ??= " \
fdt sdl kvm \
${@bb.utils.filter('DISTRO_FEATURES', 'alsa xen', d)} \qemu_3.0.0.bb
"
PACKAGECONFIG_class-native ??= "fdt alsa kvm"
PACKAGECONFIG_class-nativesdk ??= "fdt sdl kvm"

# Handle distros such as CentOS 5 32-bit that do not have kvm support
PACKAGECONFIG_class-native_remove = "${@'kvm' if not os.path.exists('/usr/include/linux/kvm.h') else ''}"

# Disable kvm on targets that do not support it
PACKAGECONFIG_remove_darwin = "kvm"
PACKAGECONFIG_remove_mingw32 = "kvm"

PACKAGECONFIG[sdl] = "--enable-sdl --with-sdlabi=2.0,--disable-sdl,libsdl2"
PACKAGECONFIG[virtfs] = "--enable-virtfs --enable-attr,--disable-virtfs,libcap attr,"
PACKAGECONFIG[aio] = "--enable-linux-aio,--disable-linux-aio,libaio,"
PACKAGECONFIG[xfs] = "--enable-xfsctl,--disable-xfsctl,xfsprogs,"
PACKAGECONFIG[xen] = "--enable-xen,--disable-xen,xen,xen-libxenstore xen-libxenctrl xen-libxenguest"
PACKAGECONFIG[vnc-sasl] = "--enable-vnc --enable-vnc-sasl,--disable-vnc-sasl,cyrus-sasl,"
PACKAGECONFIG[vnc-jpeg] = "--enable-vnc --enable-vnc-jpeg,--disable-vnc-jpeg,jpeg,"
PACKAGECONFIG[vnc-png] = "--enable-vnc --enable-vnc-png,--disable-vnc-png,libpng,"
PACKAGECONFIG[libcurl] = "--enable-curl,--disable-curl,libcurl,"
PACKAGECONFIG[nss] = "--enable-smartcard,--disable-smartcard,nss,"
PACKAGECONFIG[curses] = "--enable-curses,--disable-curses,ncurses,"
PACKAGECONFIG[gtk+] = "--enable-gtk --with-gtkabi=3.0 --enable-vte,--disable-gtk --disable-vte,gtk+3 vte"
PACKAGECONFIG[libcap-ng] = "--enable-cap-ng,--disable-cap-ng,libcap-ng,"
PACKAGECONFIG[ssh2] = "--enable-libssh2,--disable-libssh2,libssh2,"
PACKAGECONFIG[gcrypt] = "--enable-gcrypt,--disable-gcrypt,libgcrypt,"
PACKAGECONFIG[nettle] = "--enable-nettle,--disable-nettle,nettle"
PACKAGECONFIG[libusb] = "--enable-libusb,--disable-libusb,libusb1"
PACKAGECONFIG[fdt] = "--enable-fdt,--disable-fdt,dtc"
PACKAGECONFIG[alsa] = "--audio-drv-list='oss alsa',,alsa-lib"
PACKAGECONFIG[glx] = "--enable-opengl,--disable-opengl,mesa"
PACKAGECONFIG[lzo] = "--enable-lzo,--disable-lzo,lzo"
PACKAGECONFIG[numa] = "--enable-numa,--disable-numa,numactl"
PACKAGECONFIG[gnutls] = "--enable-gnutls,--disable-gnutls,gnutls"
PACKAGECONFIG[bzip2] = "--enable-bzip2,--disable-bzip2,bzip2"
PACKAGECONFIG[bluez] = "--enable-bluez,--disable-bluez,${BLUEZ}"
PACKAGECONFIG[libiscsi] = "--enable-libiscsi,--disable-libiscsi"
PACKAGECONFIG[kvm] = "--enable-kvm,--disable-kvm"
PACKAGECONFIG[virglrenderer] = "--enable-virglrenderer,--disable-virglrenderer,virglrenderer"
# spice will be in meta-networking layer
PACKAGECONFIG[spice] = "--enable-spice,--disable-spice,spice"
# usbredir will be in meta-networking layer
PACKAGECONFIG[usb-redir] = "--enable-usb-redir,--disable-usb-redir,usbredir"

INSANE_SKIP_${PN} = "arch"

[ccs]

Trying anyway with changes made to qemu_3.0.0.bb

Failed almost immediately.....


ERROR: qemu-native-3.0.0-r0 do_patch: Command Error: 'quilt --quiltrc /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/x86_64-linux/qemu-native/qemu-native-3.0.0-r0/recipe-sysroot-native/etc/quiltrc push' exited with 0 Output:
Applying patch 0100-stime-call-gone.patch
patching file linux-user/syscall.c
Hunk #1 FAILED at 7654.
1 out of 1 hunk FAILED -- rejects in file linux-user/syscall.c
Patch 0100-stime-call-gone.patch does not apply (enforce with -f)
ERROR: qemu-native-3.0.0-r0 do_patch: Function failed: patch_do_patch
ERROR: Logfile of failure stored in: /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/x86_64-linux/qemu-native/qemu-native-3.0.0-r0/temp/log.do_patch.1199829
ERROR: Task (virtual:native:/home/openvixbuilder/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu/qemu_3.0.0.bb:do_patch) failed with exit code '1'

[lincsat]

Trying anyway with changes made to qemu_3.0.0.bb

I added it here M8


SUMMARY = "Fast open source processor emulator"
DESCRIPTION = "QEMU is a hosted virtual machine monitor: it emulates the \
machine's processor through dynamic binary translation and provides a set \
of different hardware and device models for the machine, enabling it to run \
a variety of guest operating systems"
HOMEPAGE = "http://qemu.org"
LICENSE = "GPLv2 & LGPLv2.1"

RDEPENDS_${PN}-ptest = "bash make"

require qemu-targets.inc
inherit pkgconfig ptest

LIC_FILES_CHKSUM = "file://COPYING;md5=441c28d2cf86e15a37fa47e15a72fbac \
file://COPYING.LIB;endline=24;md5=8c5efda6cf1e1b03dcfd0e6 c0d271c7f"

SRC_URI = "https://download.qemu.org/${BPN}-${PV}.tar.xz \
file://powerpc_rom.bin \
file://run-ptest \
file://0001-qemu-Add-missing-wacom-HID-descriptor.patch \
file://0002-Add-subpackage-ptest-which-runs-all-unit-test-cases-.patch \
file://0003-qemu-Add-addition-environment-space-to-boot-loader-q.patch \
file://0004-qemu-disable-Valgrind.patch \
file://0005-qemu-native-set-ld.bfd-fix-cflags-and-set-some-envir.patch \
file://0006-chardev-connect-socket-to-a-spawned-command.patch \
file://0007-apic-fixup-fallthrough-to-PIC.patch \
file://0008-linux-user-Fix-webkitgtk-hangs-on-32-bit-x86-target.patch \
file://0009-Fix-webkitgtk-builds.patch \
file://0010-configure-Add-pkg-config-handling-for-libgcrypt.patch \
file://CVE-2019-15890.patch \
file://0100-stime-call-gone.patch \
"
UPSTREAM_CHECK_REGEX = "qemu-(?P<pver>\d+(\.\d+)+)\.tar"

SRC_URI[md5sum] = "cdf2b5ca52b9abac9bacb5842fa420f8"
SRC_URI[sha256sum] = "656e60218689bdeec69903087fd7582d5d3e72238d02f4481d 8dc6d79fd909c6"

COMPATIBLE_HOST_mipsarchn32 = "null"
COMPATIBLE_HOST_mipsarchn64 = "null"


I've started it building again will check back later to see if it's succeeded this time

[ccs]

So did I, kind of, but you seem to have version 4 of qemu, and I've got version 3. :confused:

[bbbuk]

I did have QEMU mismatch type errors so I started scratch again and apart from odd warnings, no issues at all. You do, occasionally, get failed errors but normally re-running them or later that day re-running, resolves them.

Also, previously I'd use a VPS but it had become unbelievably slow (days vs hours) so have now gone back to doing this locally and it's hours again. This, for me, I suspect is VPS provider becoming really busy over last few months.

[birdman]

So did I, kind of, but you seem to have version 4 of qemu, and I've got version 3. :confused:I have 4.1.
But if you look at openemdedcore on github it should be 4.2.
Hence my question as to why I wasn't getting 4.2 (which I think has the fix in it already). It would seem that there is something wrong with the configuration such that OE changes don't get reflected in the build set-up.

[ccs]

The nearest I came up with on github was openembedded-core, but no clues beyond that.

[birdman]

openembeded-core actually upgraded to qemu 5.0 on Jun 23.


http://git.openembedded.org/openembedded-core/log/meta/recipes-devtools/qemu/qemu-native_5.0.0.bb

I can only assume that somewhere within the build tree a specific commit tag is being requested for openembeded-core.*

But that can't explain why I get 4.1, while you get 3.0. I'm presuming that you're building from OE 4.4, not 4.3?

*EDIT: Which is, indeed, what is happening. It's checking out tag b6abf7c201f7c9668bdf3c6e87c7dbc70c6427f9, which is:

commit b6abf7c201f7c9668bdf3c6e87c7dbc70c6427f9
Author: Richard Purdie <richard.purdie@linuxfoundation.org>
Date: Wed Oct 9 14:13:43 2019 +0100

build-appliance-image: Update to master head revision

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

This tag (and others for other "sub-gits" is saved in oe-alliance/.git/modules/openembedded-core/HEAD (and similar). What configures this remains a mystery, as no other file in the dev setup contains that string.

[ccs]

I'm presuming that you're building from OE 4.4, not 4.3?

I used 4.3, because, if I understand correctly, OpenViX 5.3 uses the OE-A 4.3 branch.
.
Is that why it's going wrong, I was just trying to build the current 5.3 image?

[ccs]

This is what I get using 4.4, the error happens soon after the build starts...


ERROR: OE-core's config sanity checker detected a potential misconfiguration.
Either fix the cause of this error or at your own risk disable the checker (see sanity.conf).
Following is the list of potential problems / advisories:

Fetcher failure for URL: 'https://www.example.com/'. URL https://www.example.com/ doesn't work.
Please ensure your host's network is configured correctly,
or set BB_NO_NETWORK = "1" to disable network access if
all required sources are on local disk.

[lincsat]

Well the good news is that I complete the qemu OK, but now get fails on stb-kodi-vuultimo4k_18 and qtbase_git

I think I'll try from scratch again

[birdman]

I used 4.3, because, if I understand correctly, OpenViX 5.3 uses the OE-A 4.3 branch.
.
Is that why it's going wrong, I was just trying to build the current 5.3 image?That would explain the qemu version differences - I was building 5.4.xxx.xxx.
The same patch (or at least its logic) should apply to v3 of qemu.

[birdman]

This is what I get using 4.4, the error happens soon after the build starts...


ERROR: OE-core's config sanity checker detected a potential misconfiguration.
Either fix the cause of this error or at your own risk disable the checker (see sanity.conf).
Following is the list of potential problems / advisories:

Fetcher failure for URL: 'https://www.example.com/'. URL https://www.example.com/ doesn't work.
Please ensure your host's network is configured correctly,
or set BB_NO_NETWORK = "1" to disable network access if
all required sources are on local disk.

I got this back in March.
openembedded-core/meta/conf/distro/include/default-distrovars.inc includes a default setting for CONNECTIVITY_CHECK_URIS of https://www.example.com/ (which will never work). So something is supposed to set a usable value. But nothing ever does.

My comments at the time were:

Now, I can set this globally for my set-up in openembedded-core/meta/conf/conf/local.conf (which looks like a bug, as the local.conf.example file is in the first conf/, but the code only ever looks in the conf/conf/ location).

How do others set this?


The pathnames are as I gave them. However, having actually made that fix and started a build I can now remove that fix and the build still starts. Perhaps something has been added in that last few days that only gets used on building a specific, common part.

I've also done a completely fresh build of 4.4 since and not seen this problem.

None of which is really all that helpful....

[ccs]

Thanks again, I'm not see the filestore structure you're describing......


openvixbuilder@Unix:~/openvix/build-enviroment/openembedded-core/meta/conf$ ls -l
total 188
-rw-rw-r-- 1 openvixbuilder openvixbuilder 306 Aug 2 11:52 abi_version.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 189 Aug 2 11:52 bblayers.conf.sample
-rw-rw-r-- 1 openvixbuilder openvixbuilder 35720 Aug 2 11:52 bitbake.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 34 Aug 2 11:52 ccache.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 463 Aug 2 11:52 conf-notes.txt
drwxrwxr-x 3 openvixbuilder openvixbuilder 4096 Aug 2 11:52 distro
-rw-rw-r-- 1 openvixbuilder openvixbuilder 52454 Aug 2 11:52 documentation.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 501 Aug 2 11:52 image-uefi.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 3080 Aug 2 11:52 layer.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 7836 Aug 2 11:52 licenses.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 10166 Aug 2 11:52 local.conf.sample
-rw-rw-r-- 1 openvixbuilder openvixbuilder 16299 Aug 2 11:52 local.conf.sample.extended
drwxrwxr-x 3 openvixbuilder openvixbuilder 4096 Aug 2 11:52 machine
drwxrwxr-x 2 openvixbuilder openvixbuilder 4096 Aug 2 11:52 machine-sdk
-rw-rw-r-- 1 openvixbuilder openvixbuilder 32 Aug 2 11:52 migrate_localcount.conf
drwxrwxr-x 2 openvixbuilder openvixbuilder 4096 Aug 2 11:52 multiconfig
-rw-rw-r-- 1 openvixbuilder openvixbuilder 1381 Aug 2 11:52 multilib.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 22 Aug 2 11:52 prexport.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 22 Aug 2 11:52 primport.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 336 Aug 2 11:52 sanity.conf
-rw-rw-r-- 1 openvixbuilder openvixbuilder 1208 Aug 2 11:52 site.conf.sample

[ccs]

I commented out the last line CONNECTIVITY_CHECK_URIS ?= "https://www.example.com/" in ~/openvix/build-enviroment/openembedded-core/meta/conf/distro/include/default-distrovars.inc

The build has now got past that hurdle.

[ccs]

The build has now got past that hurdle.

... but hit "stime" on the second lap. :)

Started again after applying patch.

[lincsat]

Looks like qemu-native also needs patching, I'm seeing this in the stb-kodi-vuultimo4k failure log


CC mips64el-linux-user/trace/control-target.o
LINK mips-linux-user/qemu-mips
CC ppc-linux-user/gdbstub-xml.o
CC mips64el-linux-user/trace/generated-helpers.o
CC ppc-linux-user/trace/generated-helpers.o
LINK mips64el-linux-user/qemu-mips64el
CC aarch64-linux-user/trace/generated-helpers.o
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/hosttools/ld.bfd: linux-user/syscall.o: in function `do_syscall1':
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/x86_64-linux/qemu-native/qemu-native-4.1.0-r0/qemu-4.1.0/linux-user/syscall.c:7657: undefined reference to `stime'
collect2: error: ld returned 1 exit status
make[1]: *** [Makefile:209: qemu-mips] Error 1
make: *** [Makefile:472: mips-linux-user/all] Error 2
LINK mips64-linux-user/qemu-mips64
LINK ppc-linux-user/qemu-ppc
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/hosttools/ld.bfd: linux-user/syscall.o: in function `do_syscall1':
/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/vuultimo4k/tmp/work/x86_64-linux/qemu-native/qemu-native-4.1.0-r0/qemu-4.1.0/linux-user/syscall.c:7657: undefined reference to `stime'
collect2: error: ld returned 1 exit status
make[1]: *** [Makefile:209: qemu-ppc] Error 1
make: *** [Makefile:472: ppc-linux-user/all] Error 2


I've tried adding the same patch to qemu-native.inc but it fails

[BrianTheTechieSnail]

I've built a release build for the H7 successfully (openvix-5.4.000.release-zgemmah7_usb)

Me too.
Using build-enviroment 4.4 on Debian 10.4 in a VirtualBox VM with 4GB RAM and an 80GB virtual hard drive.

Haven't worked out what one can do about the the lack of feeds for 5.4.
But it loaded everything somehow, except for the picons.

[twol]

Me too.
Using build-enviroment 4.4 on Debian 10.4 in a VirtualBox VM with 4GB RAM and an 80GB virtual hard drive.

Haven't worked out what one can do about the the lack of feeds for 5.4.
But it loaded everything somehow, except for the picons.

In the builds directory you can find 2 sub directories - images and ipk (the feeds). So you can change the feeds url in /meta-oe/conf/distro/openvix to point to this directory and when you build the image it will pick them up

[birdman]

Thanks again, I'm not see the filestore structure you're describing......If you're saying that you have no conf directory under conf, that's correct. There isn't one.

But if you run strace on a build to find out where it actually looks for local.conf you will find that that is where it actually looks.

[birdman]

Looks like qemu-native also needs patchingThe config for qemu-native includes qemu.inc, so it shouldn't need separate handling.

[lincsat]

The config for qemu-native includes qemu.inc, so it shouldn't need separate handling.

I had a look at the syscall.c file and it had been patched, so don't know why that error keeps appearing. I've tried deleting the patch so it doesn't try to apply it again and it appears to have got passed that stage now

[ccs]

I'll try again when time/the weather permits.

Spoke too soon.......


ERROR: enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0 do_install: oe_runmake failed
ERROR: enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0 do_install: Execution of '/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/temp/run.do_install.3110726' failed with exit code 1:
Making install in po
make[1]: Entering directory '/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/git/po'
make[2]: Entering directory '/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/git/po'
make[2]: Nothing to be done for 'install-exec-am'.
xgettext -L python --add-comments="TRANSLATORS:" -d YahooWeather -s -o YahooWeather.pot ../*.py
for lang in de it tr; do \
msgmerge --no-location -s -N -U $lang.po YahooWeather.pot; \
done
msgfmt -o de.mo de.po
msgfmt -o it.mo it.po
../xml2po.py ../ >> YahooWeather.pot
........ done.
/bin/sh: ../xml2po.py: /usr/bin/python: bad interpreter: No such file or directory
make[2]: *** [Makefile:478: YahooWeather.pot] Error 126
make[2]: *** Waiting for unfinished jobs....
........ done.
........ done.
make[2]: Leaving directory '/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/git/po'
make[1]: *** [Makefile:357: install-am] Error 2
make[1]: Leaving directory '/home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/git/po'
make: *** [Makefile:388: install-recursive] Error 1
WARNING: /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/temp/run.do_install.3110726:1 exit 1 from 'exit 1'

ERROR: Logfile of failure stored in: /home/openvixbuilder/openvix/build-enviroment/builds/openvix/release/et10000/tmp/work/mips32el-oe-linux/enigma2-plugin-extensions-yahooweather/enigma2-plugin-extensions-yahooweather-1.2.+gitAUTOINC+f04f398ad7-r0/temp/log.do_install.3110726

[BrianTheTechieSnail]

Me too.
Using build-enviroment 4.4 on Debian 10.4 in a VirtualBox VM with 4GB RAM and an 80GB virtual hard drive.

Haven't worked out what one can do about the the lack of feeds for 5.4.
But it loaded everything somehow, except for the picons.

I noticed the 5.4 developer feeds seem to be in place already so I tried setting DISTRO_TYPE=developer but it failed saying something was missing.
Maybe it just ran out of disk space.

[ccs]

Regarding post #188, do I need to create a hard/soft link for /usr/lib/python. If so, what does it point to??


ls -ld /usr/lib/py*
drwxr-xr-x 26 root root 20480 Jul 31 17:33 /usr/lib/python2.7
drwxr-xr-x 3 root root 4096 Apr 23 08:32 /usr/lib/python3
drwxr-xr-x 30 root root 20480 Jul 31 17:12 /usr/lib/python3.8

[ccs]

Maybe that should have been bin .....


ccs@Unix:~$ ls -ld /usr/bin/pyt*
lrwxrwxrwx 1 root root 9 Mar 13 12:31 /usr/bin/python2 -> python2.7
-rwxr-xr-x 1 root root 3694632 Apr 7 13:05 /usr/bin/python2.7
lrwxrwxrwx 1 root root 9 Jul 31 16:44 /usr/bin/python3 -> python3.8
-rwxr-xr-x 1 root root 5453504 Jul 16 15:00 /usr/bin/python3.8
lrwxrwxrwx 1 root root 33 Jul 16 15:00 /usr/bin/python3.8-config -> x86_64-linux-gnu-python3.8-config
lrwxrwxrwx 1 root root 16 Mar 13 12:20 /usr/bin/python3-config -> python3.8-config
-rwxr-xr-x 1 root root 384 Mar 28 02:39 /usr/bin/python3-futurize
-rwxr-xr-x 1 root root 388 Mar 28 02:39 /usr/bin/python3-pasteurize
ccs@Unix:~$

[birdman]

Regarding post #188, do I need to create a hard/soft link for /usr/lib/python. [/code]As I recall, yes.
If this is Ubuntu install the python-is-python2 package.
(There is also a python-is-python3 package - they just create the symlink...).

[lincsat]

I had a look at the syscall.c file and it had been patched, so don't know why that error keeps appearing. I've tried deleting the patch so it doesn't try to apply it again and it appears to have got passed that stage now

Success - the builds now complete, I have ZGemmaH7 and Vuultimo4k images for openvix-5.4.000.release and openvix-5.4.000.005.developer

[BrianTheTechieSnail]

I noticed the 5.4 developer feeds seem to be in place already so I tried setting DISTRO_TYPE=developer but it failed saying something was missing.
Maybe it just ran out of disk space.

I tried again with the virtual disk expanded to 100GB and it worked.
But I don't think it really needed the extra space.
Something else must have stopped it working before.

[ccs]

Not my day today, yahoo!!

Is it easy to remove this plugin completely?

(Or run interactively and answer "y" when prompted.)



xgettext -L python --add-comments="TRANSLATORS:" -d YahooWeather -s -o YahooWeather.pot ../*.py
for lang in de it tr; do \
msgmerge --no-location -s -N -U $lang.po YahooWeather.pot; \
done
../xml2po.py ../ >> YahooWeather.pot
........ done.
........ done.
........ done.
File "../xml2po.py", line 55
print '#: ' + arg
^
SyntaxError: Missing parentheses in call to 'print'. Did you mean print('#: ' + arg)?
make[2]: *** [Makefile:478: YahooWeather.pot] Error 1

[twol]

https://github.com/oe-alliance/oe-alliance-core/tree/4.4/meta-oe/recipes-oe-alliance/enigma2-plugins —- remove the .bb
https://github.com/oe-alliance/oe-alliance-core/blob/4.4/meta-oe/recipes-oe-alliance/image/oe-alliance-feeds.bb - remove the dependency

You might find the build blows out because of some other dependency - just remove it.

Although this is a bit like a permanent self destruct - something is wrong with your build environment

[ccs]

https://github.com/oe-alliance/oe-alliance-core/tree/4.4/meta-oe/recipes-oe-alliance/enigma2-plugins —- remove the .bb
https://github.com/oe-alliance/oe-alliance-core/blob/4.4/meta-oe/recipes-oe-alliance/image/oe-alliance-feeds.bb - remove the dependency

You might find the build blows out because of some other dependency - just remove it.

Although this is a bit like a permanent self destruct - something is wrong with your build environment

Sorry, no idea what you're suggesting here. Are you saying I should edit the relevant source files, I've so far found it impossible to find such files in the build setup I'm running.

All I'm doing is following the basic steps highlighted in this thread, I'm getting errors which are already known about, and now a syntax error in the source of a plugin.

I haven't the slightest idea why my build environment could now be wrong.

[Huevos]

Me too.
Using build-enviroment 4.4 on Debian 10.4 in a VirtualBox VM with 4GB RAM and an 80GB virtual hard drive.

Haven't worked out what one can do about the the lack of feeds for 5.4.
But it loaded everything somehow, except for the picons.You need a webserver running on your build server.

So let's say the root folder of the webserver is /var/www ...

Create a folder for feeds... /var/www/feeds

In site.conf add the domain name of your feeds (and your enigma repo if you want). e.g...

SCONF_VERSION = "1"
BB_NUMBER_THREADS = "32"
PARALLEL_MAKE = "-j 16"
BUILD_OPTIMIZATION = "-O2 -pipe"
DL_DIR = "/home/openvix/sources"
INHERIT += "rm_work"
ENIGMA2_URI = "git://github.com/yourRepo/enigma2.git;protocol=git;branch=Dev"
DISTRO_FEED_URI = "http://your.feeds.domain/feeds/${DISTRO_NAME}/${DISTRO_TYPE}/${DISTRO_VERSION}/${MACHINE}"

Note: in the above file my sources folder is not in the build folder. So if I trash a build I retain all the sources files for other build envs.

Now build your image.

Once you image is built, symlink the webserver to the IPK folder. e.g...

mkdir -p /var/www/feeds/openvix/developer/5.4/h7
ln -s /home/openvix/5.4/builds/openvix/developer/h7/tmp/deploy/ipk/* /var/www/feeds/openvix/experimental/5.4/h7

[twol]

Sorry, no idea what you're suggesting here. Are you saying I should edit the relevant source files, I've so far found it impossible to find such files in the build setup I'm running.

All I'm doing is following the basic steps highlighted in this thread, I'm getting errors which are already known about, and now a syntax error in the source of a plugin.

I haven't the slightest idea why my build environment could now be wrong.

of course edit your local files - you should have a meta-oe directory and inside that a recipes-oe-alliance directory - inside that you have the files I was highlighting

[ccs]

of course edit your local files - you should have a meta-oe directory and inside that a recipes-oe-alliance directory - inside that you have the files I was highlighting

Eventually found the .bb files but the build got worse, so starting again from scratch.

Found the files around about here....

/home/openvixbuilder/openvix/build-enviroment/meta-oe-alliance/meta-oe/recipes-oe-alliance/enigma2-plugins

I couldn't find the .py files, any clues much appreciated.

[birdman]

Not my day today, yahoo!!That error message is what you'd get when running a python2 script (all of the xml2po.py ones are) with python3.

[Valiant]

Been a while since I built 5.3 for my miraclebox after vix support was dropped , so tried to build again and encountered an error straight away, miraclebox is not supported by the build process.

My environment is ok, I built vuuno4k and osmini images without error over the weekend after updating ubuntu.

If there is no way to build a later stable version for the miraclebox then I will leave it on 5.3.001 or 4 , I just thought there might be a local parameter or list I could change to allow the build to progress.

[twol]

Been a while since I built 5.3 for my miraclebox after vix support was dropped , so tried to build again and encountered an error straight away, miraclebox is not supported by the build process.

My environment is ok, I built vuuno4k and osmini images without error over the weekend after updating ubuntu.

If there is no way to build a later stable version for the miraclebox then I will leave it on 5.3.001 or 4 , I just thought there might be a local parameter or list I could change to allow the build to progress.
Its still in the makefile mbtwinplus, so no reason that it will not build - whether it will have enough to run the image is another story.

[ccs]

That error message is what you'd get when running a python2 script (all of the xml2po.py ones are) with python3.

Thanks, as always, that sorted it when I opted for python2. I was expecting python3 to be used now, is that not the case?

Can't get passed this (repeated, always the same) checksum error....


File: '/home/openvixbuilder/openvix/build-enviroment/sources/ntp-4.2.8p13.tar.gz' has md5 checksum 19bb20214e0a88629e17a5b6c605d45d when ea040ab9b4ca656b5229b89d6b822f13 was expected
File: '/home/openvixbuilder/openvix/build-enviroment/sources/ntp-4.2.8p13.tar.gz' has sha256 checksum 53a4b4b6569e8ef8b0aa6b7adbf736563467dba67aa77c6f4a 37b7387b01508f when 288772cecfcd9a53694ffab108d1825a31ba77f3a8466b0401 baeca3bc232a38 was expected

It suggests changing the checksums, maybe it's not advisable, but where do I find the relevant file? I've spent ages trying to find it, but failed miserably.


If this change is expected (e.g. you have upgraded to a new version without updating the checksums) then you can use these lines within the recipe:
SRC_URI[md5sum] = "19bb20214e0a88629e17a5b6c605d45d"
SRC_URI[sha256sum] = "53a4b4b6569e8ef8b0aa6b7adbf736563467dba67aa77c6f4a 37b7387b01508f"
Otherwise you should retry the download and/or check with upstream to determine if the file has become corrupted or otherwise unexpectedly modified.

[Valiant]

Thanks twol, ill have another go and post the error that I get.

[ccs]

Post #204.... md5 checksum confirmed as ok - 19bb20214e0a88629e17a5b6c605d45d after downloading the file onto a windows laptop.

[ccs]

For reference, the file is
http://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ntp-4.2/ntp-4.2.8p13.tar.gz

[lincsat]

How would I get the build to use my own passwd file so I don't have to manually set if a settings restore fails?

[twol]

Thanks, as always, that sorted it when I opted for python2. I was expecting python3 to be used now, is that not the case?

Can't get passed this (repeated, always the same) checksum error....


File: '/home/openvixbuilder/openvix/build-enviroment/sources/ntp-4.2.8p13.tar.gz' has md5 checksum 19bb20214e0a88629e17a5b6c605d45d when ea040ab9b4ca656b5229b89d6b822f13 was expected
File: '/home/openvixbuilder/openvix/build-enviroment/sources/ntp-4.2.8p13.tar.gz' has sha256 checksum 53a4b4b6569e8ef8b0aa6b7adbf736563467dba67aa77c6f4a 37b7387b01508f when 288772cecfcd9a53694ffab108d1825a31ba77f3a8466b0401 baeca3bc232a38 was expected

It suggests changing the checksums, maybe it's not advisable, but where do I find the relevant file? I've spent ages trying to find it, but failed miserably.


If this change is expected (e.g. you have upgraded to a new version without updating the checksums) then you can use these lines within the recipe:
SRC_URI[md5sum] = "19bb20214e0a88629e17a5b6c605d45d"
SRC_URI[sha256sum] = "53a4b4b6569e8ef8b0aa6b7adbf736563467dba67aa77c6f4a 37b7387b01508f"
Otherwise you should retry the download and/or check with upstream to determine if the file has become corrupted or otherwise unexpectedly modified.

If you look in the build log file you will see at what stage it failed and be able to find the .bb where you need to update the checksums

[ccs]

If you look in the build log file you will see at what stage it failed and be able to find the .bb where you need to update the checksums

If ~/openvix/build-enviroment/builds/openvix/dev/et10000/tmp/log/cooker/et10000/ is were the logs are stored, then I've found the relevant .bb file, with the wrong? checksum.

/home/openvixbuilder/openvix/build-enviroment/meta-openembedded/meta-networking/recipes-support/ntp/ntp_4.2.8p13.bb

Took a bit of finding is an understatement.

Why am I the only one seeing this checksum error? Am I picking up the wrong version of ntp?

I'll start again tomorrow when 5.4.000 is released, and dev/master are in sync. I started the build yesterday when 5.4 was just being finalised.

[Valiant]

Its still in the makefile mbtwinplus, so no reason that it will not build - whether it will have enough to run the image is another story.

Therein lies the answer, it helps if I get the machine name right! I've made notes rather than trusting to fickle memory. Successful build of 5.3.039, one warning, now to see if the mbtwinplus can accomodate it. Thanks twol.

[birdman]

Took a bit of finding is an understatement.You know it's likely to be an ntp*.bb file, so:


cd ..../oe-alliance
find meta* open* -iname 'ntp*.bb'

will list the likely config file candidates.

[ccs]

I've checked versions p9 thru' to p15 and none have the expected md5 checksum. I feel like I'm hitting my head against a brick wall.


File: '/home/openvixbuilder/openvix/build-enviroment/sources/ntp-4.2.8p13.tar.gz' has md5 checksum 19bb20214e0a88629e17a5b6c605d45d when ea040ab9b4ca656b5229b89d6b822f13 was expected
File: '/home/openvixbuilder/openvix/build-enviroment/sources/ntp-4.2.8p13.tar.gz' has sha256 checksum 53a4b4b6569e8ef8b0aa6b7adbf736563467dba67aa77c6f4a 37b7387b01508f when 288772cecfcd9a53694ffab108d1825a31ba77f3a8466b0401 baeca3bc232a38 was expected

Only possible clue is that their .md5 file format has changed between p12 and p13...


1522d66574bae14abb2622746dad2bdc ntp-4.2.8p12.tar.gz

MD5 (ntp-4.2.8p13.tar.gz) = 19bb20214e0a88629e17a5b6c605d45d

[Huevos]

I've checked versions p9 thru' to p15 and none have the expected md5 checksum. I feel like I'm hitting my head against a brick wall.

Only possible clue is that their .md5 file format has changed between p12 and p13...


1522d66574bae14abb2622746dad2bdc ntp-4.2.8p12.tar.gz

MD5 (ntp-4.2.8p13.tar.gz) = 19bb20214e0a88629e17a5b6c605d45dSo why don't you just change the checksum in the .bb file?

[ccs]

So why don't you just change the checksum in the .bb file?

Obviously I can, but I'm trying to work out why the checksum is wrong, I appear to be the only one hitting this problem.

I could easily have made mistakes following the instructions in this thread, If I have, then I need to find out what I've done wrong.

The one thing I wonder about is I changed python from version 3 to version 2 half way thru' a build to get yahooweather sorted.

Could that explain anything?

I'm starting again from scratch soon, but at the moment, nothing will change from the last attempt.

[Huevos]

I just checked and Ea040ab9b4ca656b5229b89d6b822f13 is the correct hash. So if your file doesn't match that hash your file must be corrupt.

Unzip the attached file and put the contents in your sources folder and try again.

[ccs]

I just checked and Ea040ab9b4ca656b5229b89d6b822f13 is the correct hash. So if your file doesn't match that hash your file must be corrupt.

Unzip the attached file and put the contents in your sources folder and try again.
Thanks.

I've worked out what is happening, no corrupt files.

Your file comes from
https://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ with the checksum the build expects.

My file (in my build) comes from
https://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ntp-4.2/ with the same filename and a different checksum. A checksum which is correct for that file, but not what the build expects.

[Joe_90]

@ccs - are you moving source files between different systems (PC's) in your setup? I recently had issues when moving scripts between my mutant box and my main desktop (which is linux mint). A new install of FileZilla had helpfully changed the line endings of .sh files to ms-dos format :eek:.
It's just a left-of-field suggestion, but if you're ftp'ing files between systems it's worth checking.


edit - ok, belay that! You've found your difference...


further edit - a different md5 checksum means that the files are different, though. You need to compare the sources.

[BrianTheTechieSnail]

Thanks.

I've worked out what is happening, no corrupt files.

Your file comes from
https://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ with the checksum the build expects.

My file (in my build) comes from
https://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ntp-4.2/ with the same filename and a different checksum.

The contents are different too!
It doesn't look like tampering though, looks like somebody updated some stuff in the file without changing the file name.
I'll just call that definitely not clever at all.

[ccs]

@ccs - are you moving source files between different systems (PC's) in your setup? I recently had issues when moving scripts between my mutant box and my main desktop (which is linux mint). A new install of FileZilla had helpfully changed the line endings of .sh files to ms-dos format :eek:.
It's just a left-of-field suggestion, but if you're ftp'ing files between systems it's worth checking.


edit - ok, belay that! You've found your difference...

No, everything has been done on a fresh install of Ubuntu 20.04, which will be even fresher tomorrow when I replace the hd and start again.

[Joe_90]

Ok, but check the differences in the sources.

[ccs]

My file (in my build) comes from ......

Probably a bad choice of words, the build I'm running is the one mentioned a number of times in this thread, with no changes whatsoever.

Can anyone confirm that python2 is the correct choice, (bearing in mind that yahooweather doesn't like python3).

[twol]

Probably a bad choice of words, the build I'm running is the one mentioned a number of times in this thread, with no changes whatsoever.

Can anyone confirm that python2 is the correct choice, (bearing in mind that yahooweather doesn't like python3).

For current builds use python2. Enigma and python3 is a pain and python3 has very exacting requirements

[BrianTheTechieSnail]

From the point of view of which one you need for a particular situation Python2 and Python3 are effectively different languages.
Almost no software written in one version will work in the other version unless you make changes.

[birdman]

I've checked versions p9 thru' to p15 and none have the expected md5 checksum. I feel like I'm hitting my head against a brick wall.This is what is in my build tree:


[parent]: md5sum ntp-4.2.8p13.tar.gz
ea040ab9b4ca656b5229b89d6b822f13 ntp-4.2.8p13.tar.gz

[birdman]

The ntp4/ntp-4.2 one was updated in mid-March at the same time as p14 was created.
Looks like someone messed up creating the releases.

EDIT: Looking at the p13 -> p14 diffs they are almost all version-stamp related.

So it looks like someone pushed all of the updates in and created a new p13 package, then realized that the version should have been updated, updated it (along with a few minor changes, which might have been what triggered noticing the version needed to change) then created a new p14. There's just over 1 hour between the production times.

So the ntp*.bb build file should be changed to refer to the file one level up....

EDIT2:
But there's little point in submitting it as a fix to github, as the latest version on github refers to the p15 file.

[ccs]

...thanks again, so I need to update the ntp*.bb build file to reflect the correct version p13?

How come I picked up the one a level down from everybody else? Is it an alternative, and I got it when the default one happened to be unavailable?

I'll have a look in the logs to see if that was the case.

[ccs]

..... The file meta-openembedded/meta-networking/recipes-support/ntp/ntp_4.2.8p13.bb has the line.....

SRC_URI = "http://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ntp-4.2/ntp-${PV}.tar.gz

[ccs]

.... and has always used that folders structure (as far as I can tell as a total novice). :confused:


https://github.com/openwrt/packages/commits/master/net/ntpd

[birdman]

..... The file meta-openembedded/meta-networking/recipes-support/ntp/ntp_4.2.8p13.bb has the line.....

SRC_URI = "http://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ntp-4.2/ntp-${PV}.tar.gzYes, you could edit that (remove the ntp-4.2), run a build then undo the file edit.

[birdman]

How come I picked up the one a level down from everybody else?You probably didn't. Everyone else picked it up before the middle of March....until then it was OK.

My suspicion is that the "correct" one was a restore, but wasn't put back to the correct location.

EDIT: I've mailed the maintainer to ask him to move them....

[bbbuk]

I had a mismatch when building ubuntu 20 (can't remember on what) so went back to 18.04 (or something similar) and didn't have an issue.

[twol]

Ubuntu 18.04 is as solid as a rock for building current OE-A versions (4.3, 4.4) and I would suggest that anyone wanting to build their own images should stick with this version.......... maybe we wouldn‘t then see so many of the reported issues.

[Willo3092]

Ubuntu 18.04 is as solid as a rock for building current OE-A versions (4.3, 4.4) and I would suggest that anyone wanting to build their own images should stick with this version.......... maybe we wouldn‘t then see so many of the reported issues.

It's only the fact that mine seems to be working fine that stopped me updating to the latest :thumbsup:

[Valiant]

Ubuntu 18.04 is as solid as a rock for building current OE-A versions (4.3, 4.4) and I would suggest that anyone wanting to build their own images should stick with this version.......... maybe we wouldn‘t then see so many of the reported issues.

I did three different machine builds with with no errors and a handful of warnings with Ubuntu 18.4.04.

[birdman]

It's only the fact that mine seems to be working fine that stopped me updating to the latest :thumbsup:I suspect that OpenVix4.xxx (or even Hades) was working fine as well, but people still updated to the latest version.

[ccs]

Finally got thru' with no errors, but loads of warnings.

Here's a summary.....


(1) Ubuntu 18.04 defaults to python2, which is needed to build yahooweather.

(2) Comment out the last line

CONNECTIVITY_CHECK_URIS ?= "https://www.example.com/"

FILE: ~/openvix/build-enviroment/openembedded-core/meta/conf/distro/include/default-distrovars.inc

(3) Get correct checksum by removing ntp-4.2/ in the line

SRC_URI = "http://www.eecis.udel.edu/~ntp/ntp_spool/ntp4/ntp-4.2/ntp-${PV}.tar.gz \

FILE: ~/openvix/build-enviroment/meta-openembedded/meta-networking/recipes-support/ntp/ntp_4.2.8p13.bb

(4) Put patch 0100-stime-call-gone.patch into ~/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu/qemu

Add reference to patch in ~/openvix/build-enviroment/openembedded-core/meta/recipes-devtools/qemu/qemu.inc

Post #165 https://www.world-of-satellite.com/showthread.php?61295-Build-my-own-Vix-image&p=500915&viewfull=1#post500915

Thanks to all that helped, not convinced by all the warnings, mind.:confused:

[birdman]

Thanks to all that helped, not convinced by all the warnings, mind.:confused:There have always been a lot of warnings.
No doubt some of them turn into errors from time to time, but their presence doesn't seem to provide any impetus to those who add the code that introduced them to actually fix it (I think they're all config or patch-fuzz issues).

Personally I'd go with the Linux kernel build rule, which is that it has to build without any warnings when all warnings are turned on, but I'm not the person in charge,

[BrianTheTechieSnail]

FWIW during the last couple of it seems to have been possible to build good working 5.4.001 (Release) images that work really well, at least for me, on my Zgemma H7S.

The "buglet" I asked about here:
https://www.world-of-satellite.com/showthread.php?62932-H7S-boots-up-with-weird-video-mode-about-half-the-time&p=498554&viewfull=1#post498554
even seems to be fixed (so far).

[lincsat]

I still get the occasional fail before the build completes.

Don't know if it's a bug or just me but using the EPG-Refresh plugin, when changing the services, ie editing the xml file, it crashes the box. Also when using the X-Streamity plugin and internally editing the settings, ie editing the xml file, it doesn't save the changes. It's probably just the way I updated, I'll get around to doing another flash and restore at some stage before posting an official bug

[BrianTheTechieSnail]

I still get the occasional fail before the build completes.
I was getting lots of fails yesterday. I think they were working on things it needs to reach out and download.



Don't know if it's a bug or just me but using the EPG-Refresh plugin, when changing the services, ie editing the xml file, it crashes the box.
Oh drat. Yes I get that too. I did spot it earlier, forgot to check again recently.
I've never tried X-Streamity so I'm sure you're right about that too.

[BrianTheTechieSnail]

The "buglet" I asked about here:
https://www.world-of-satellite.com/showthread.php?62932-H7S-boots-up-with-weird-video-mode-about-half-the-time&p=498554&viewfull=1#post498554
even seems to be fixed (so far).
... I just had a build were it seemed virtually impossible to boot up in what I regard as the correct state (with the nice sharp OSD characters) :(

[lincsat]

I was getting lots of fails yesterday. I think they were working on things it needs to reach out and download.




Don't know if it's a bug or just me but using the EPG-Refresh plugin, when changing the services, ie editing the xml file, it crashes the box.


Oh drat. Yes I get that too. I did spot it earlier, forgot to check again recently.
I've never tried X-Streamity so I'm sure you're right about that too.

Looks like a problem with the keymap.xml, using the file from the 5.3 build is OK. I've posted this issue in here rather than starting a new thread as the 5.4 isn't officially released yet. Here is the section of the crash log showing the problem


<235890.5967> [InfoBarGenerics] Key: 352 (Make) KeyID='KEY_OK' Binding='('OK',)'.
<235890.5971> [ActionMap] Keymap 'OkCancelActions' -> Action = 'ok'.
<235890.5987> [Screen] Showing screen 'SetupSummary'.
<235890.6008> [Screen] Showing screen 'EPGRefreshServiceEditor'.
<235890.8001> [eInputDeviceInit] 0 160 (352) 1
<235890.8008> [InfoBarGenerics] Key: 352 (Break) KeyID='KEY_OK' Binding='('OK',)'.
<235893.6735> [eInputDeviceInit] 1 18f (399) 1
<235893.6742> [InfoBarGenerics] Key: 399 (Make) KeyID='KEY_GREEN' Binding='('GREEN',)'.
<235893.6744> [ActionMap] Keymap 'ConfigListActions' -> Unknown action 'save'! (Typo in keymap?)
<235893.6745> [ActionMap] Keymap 'ConfigListActions' -> Action = 'save'.
<235893.7021> [Screen] Showing screen 'SetupSummary'.
<235893.7050> [Screen] Showing screen 'EPGRefreshConfiguration'.
<235893.7074> Traceback (most recent call last):
<235893.7075> File "/usr/lib/enigma2/python/mytest.py", line 240, in processDelay
<235893.7077> callback(*retval)
<235893.7078> TypeError: editServicesCallback() takes exactly 2 arguments (1 given)
<235893.7078> [ePyObject] (CallObject(<bound method Session.processDelay of <__main__.Session instance at 0xb1fdb8a0>>,()) failed)



[EDIT]

After a couple of quick tests, removing the
<map context="ConfigListActions"> section from the new keymap.xml fixes the problem. I don't know what that section is for but the image doesn't like it

[ccs]

After a couple of quick tests, removing the
<map context="ConfigListActions"> section from the new keymap.xml fixes the problem. I don't know what that section is for but the image doesn't like it

Couple of recent dev commits...


https://github.com/OpenViX/enigma2/commit/bf4d7b3d4af938e2b7d7cad9577c02dd63ba00d5#diff-edfd085aebeda59b55c2945039de3f03

https://github.com/OpenViX/enigma2/commit/7c9209386420afd5f1bab70452c0dc16762afc2b#diff-edfd085aebeda59b55c2945039de3f03

I reckon you're asking for trouble trying to build before the first official release of 5.4 is available.

[lincsat]

Couple of recent dev commits...


https://github.com/OpenViX/enigma2/commit/bf4d7b3d4af938e2b7d7cad9577c02dd63ba00d5#diff-edfd085aebeda59b55c2945039de3f03

https://github.com/OpenViX/enigma2/commit/7c9209386420afd5f1bab70452c0dc16762afc2b#diff-edfd085aebeda59b55c2945039de3f03

I reckon you're asking for trouble trying to build before the first official release of 5.4 is available.

I'm building to try it out, not as a main use image.

I seem to have the commit from 13 Days ago but not the one from 7 Days ago despite using an image built Yesterday

[Huevos]

Thanks for reporting. This bug is a known issue.

[lincsat]

Since doing a make-update a few Days ago, I've not been able to build any updates. Always fails at compiling the serviceapp plugin. Is anyone else having that problem or is it just me?

[abu baniaz]

Since doing a make-update a few Days ago, I've not been able to build any updates. Always fails at compiling the serviceapp plugin. Is anyone else having that problem or is it just me?
Are you building Dev or Release branch? For which boxes? Its been fixed in Dev, but some boxes still have an issue.

[ccs]

Maybe read the comments on here....


https://github.com/OpenViX/enigma2/commit/1213ead63cbc1f2c9779cc4be83d4c4d6bf75315#comments

[lincsat]

Maybe read the comments on here....


https://github.com/OpenViX/enigma2/commit/1213ead63cbc1f2c9779cc4be83d4c4d6bf75315#comments

Yep, that looks like the issue and answers the question, it's not just me :)

I'm currently using the release branch, I'll try dev next as it's fixed but from the comments, sounds like something else may be broken