Willo3092
ViX Beta Tester
What is the command to clean the build?
I'm sure I've done it before but I must have lost the instructions.
I'm sure I've done it before but I must have lost the instructions.
Superb quality and spec AB-Com PULSe 4K SE. Crazy offer! Only £129! FREE UK DELIVERY! 4K UHD, Enigma 2, Multiboot 4 images & more!...
Superb quality and spec AB-Com PULSe 4K Rev II Twin Satellite tuner only £179! FREE UK DELIVERY! 4K UHD, Enigma 2, SATA HDD facility, Multiboot 4 images & more!...
Build my own Vix image
- Thread starter Willo3092
- Start date
@Willo3092 have you read this thread: https://world-of-satellite.com/threads/openvix-build-server-guide-non-guest-section.48085/
That abort is not related with birdman's scripts.I'm using the meta-local recipes from @birdman but still getting build failed when trying to build OEA-5.6 on Ubuntu 26.04.
Attached is the terminal output
Let's look at the error message:
Code:
WARNING: openvix-image-6.8-r008.006 do_rootfs: shadow.postinst returned 1, marking as unpacked only, configuration required on target.
WARNING: openvix-image-6.8-r008.006 do_rootfs: util-linux-flock.postinst returned 1, marking as unpacked only, configuration required on target.
ERROR: openvix-image-6.8-r008.006 do_rootfs: Postinstall scriptlets of ['shadow', 'util-linux-flock'] have failed. If the intention is to defer them to first boot,
then please place them into pkg_postinst_ontarget:${PN} ().
Deferring to first boot via 'exit 1' is no longer supported.That is telling us:
The fails are shadow and util-linux-flock.
In those recipes there is something like:
Code:
pkg_postinst() {
...
exit 1
}That needs to be changed to:
Code:
pkg_postinst_ontarget:${PN} () {
...
}Because the 'exit 1' is no longer supported.
Last edited:
Willo3092
ViX Beta Tester
I'll have a look later. Got to go out today 
birdman
Moderator
I'll see what happens when I try a build.
The error looks strangely familiar... (but that might just mean I've seen a lot of error messages in my time).
The error looks strangely familiar... (but that might just mean I've seen a lot of error messages in my time).
birdman
Moderator
Odd that mine built OK.
I'll have a look at the recipes on my system....
But this is not a new "feature".
Here's one report of that from 2021:
https://forums.openpli.org/topic/41198-serviceapp-gstplayer-and-exteplayer3/?view=findpost&p=1409142
birdman
Moderator
Sorry for the delay. have been upgrading all of my systems to 26.04.
When I try an ax61 build of OE5.6 on 26.04 it fails in
dtc-native-1.7.2-r0 do_compile
The reason is that this is compiled with -Werro set (a good idea - amongest other things it makes sure you've got all of your declarations correct).
Except in this case the author hasn't.
(S)He has mixed up char and const char and it seems that gcc, via the header files, now sets the protoype for some string.h calls (memchr, strchr, etc...) based on the input parameters. So if you pass in a const char* then you must assign the result to a const char*. But that isn;t happening in two places.
So the build fails.
It isn't an issue for m,my et8000,as the build for that doesn't use dtc.
Here's my latest recipes-local including a patch for dtc.
Last edited:
Willo3092
ViX Beta Tester
Just tried the new meta-local files but still getting an error message:
Is it worth starting from scratch with it?
Is it worth starting from scratch with it?
At least clean openvix-image.
bitbake -c cleanall openvix-image
Then rebuild.
bitbake -c cleanall openvix-image
Then rebuild.
birdman
Moderator
I haven't got that far yet.
I'm just building 3607 of 18999 at the moment. (Going in stages as problems hit...)
Willo3092
ViX Beta Tester
Which directory do I need to be in to run the bitbake clean command?
From build-enviroment I get this:
Willo3092
ViX Beta Tester
I managed the clean but still getting an error:
I guess thats the flock/shadow issue? Which is what I get for any broadcom box on 26.04/ OE-A 5.6
birdman
Moderator
But it's odd that this is nothing to do with the OS versions, but a bitbake config.
And even odder that I don't see if for an et8000 build, which must also build rootfs for openvix-image.bb.
Anyway, my ax61/hd61 is now at 18978 of 18999. So I should hit it soon.
Last edited:
birdman
Moderator
OK. My build has finished.
It did report an error:
The result is that the build actually completed. I do have:
Now, I''ll wipe it all out and do a complete build from scratch (with the latest patch in place) to see what happens then.
It did report an error:
However, Taskhash mismatches, although reported as ERRORs, are ignored (so why are they even done?).
The result is that the build actually completed. I do have:
openvix-6.8.008.006.developer-ax61_recovery_emmc.zip
openvix-6.8.008.006.developer-ax61_multi.zip
Now, I''ll wipe it all out and do a complete build from scratch (with the latest patch in place) to see what happens then.
Not until there are no new changes that force full rebuilds. Nothing worse then completely rebuilding 50 boxes every few weeks.
birdman
Moderator
There is s fix in git for this:
https://git.kernel.org/pub/scm/utils/dtc/dtc.git/commit/?id=9a1c801a1a3c102bf95c5339c9e985b26b823a21
but that is from mid-January, and there hasn't been a new tar release (which OE uses) since 19 months ago.
