bitbake, oe-core and meta-yocto tags in sync? (was Re: GPIT LAVA Test Report - Job 123474: OEQA ALL PASSED (24/75) passed)

bitbake, oe-core and meta-yocto tags in sync? (was Re: GPIT LAVA Test Report - Job 123474: OEQA ALL PASSED (24/75) passed)

[Mikko Rapeli]

Hi,

On Tue, Nov 11, 2025 at 10:35:08AM -0800, ledge-bot@linaro.org wrote:
> LAVA Test Report for Job 123474
> ==================================================
> 
> GitLab Pipeline Information:
>   Project: gpit
>   Branch: main
>   Pipeline: https://gitlab.com/Linaro/trustedsubstrate/gpit/-/pipelines/2152817594
>   bitbake: 2c6c40bdab8ebc6c04fe9992524add967fb4a053
>   bitbake git describe: yocto-5.2.1-149-g2c6c40bda
>   openembedded-core: e556df28f47e754b53e1f46c97dde2b19fd8fad1
>   openembedded-core git describe: uninative-5.0-152-ge556df28f4
>   meta-yocto: 66ea31cee15309af07afeee8dec8a6fe1d92d28e
>   meta-yocto git describe: 5.3_M3-3-g66ea31c

So this is our first build and test run with separate bitbake,
oe-core and meta-yocto built using kas to replace old poky repo master
branch which no longer builds. Setup details in
https://gitlab.com/Linaro/trustedsubstrate/gpit

The "git describe" output with tags is a bit confusing. Would it be
possible to use same version/milestone names there?

Currently it's a bit hard to tell if the trees are fully in sync and if
yocto-5.2.1, uninative-5.0 and 5.3_M3 are compatible with each other
or not. From pure tags it is not easy to see if trees used same
branches, for example.

> LAVA Job Information:
>   Job ID: 123474
>   Description: gpit-kv260
>   Device: laa-00077-kv260
>   State: Finished
>   URL: https://ledge.validation.linaro.org/scheduler/job/123474
> 
> OEQA Test Results Summary:
>   Total Tests: 75
>   Passed: 24
>   Failed: 0
>   Errors: 0
>   Skipped: 51

But at least all executed oeqa rutime tests on core-image-sato
with genericarm64 machine passed on AMD kv260 which is the same
result as with poky from one week ago before its retirement.

> Passed Tests:
>   - connman.ConnmanTest.test_connmand_help
>   - connman.ConnmanTest.test_connmand_running
>   - date.DateTest.test_date
>   - df.DfTest.test_df
>   - oe_syslog.SyslogTest.test_syslog_running
>   - oe_syslog.SyslogTestConfig.test_syslog_logger
>   - oe_syslog.SyslogTestConfig.test_syslog_restart
>   - opkg.OpkgRepoTest.test_opkg_install_from_repo
>   - pam.PamBasicTest.test_pam
>   - parselogs.ParseLogsTest.test_get_context
>   - parselogs.ParseLogsTest.test_parselogs
>   - ping.PingTest.test_ping
>   - python.PythonTest.test_python3
>   - scp.ScpTest.test_scp_file
>   - ssh.SSHTest.test_ssh
>   - systemd.SystemdBasicTests.test_systemd_basic
>   - systemd.SystemdBasicTests.test_systemd_failed
>   - systemd.SystemdBasicTests.test_systemd_list
>   - systemd.SystemdJournalTests.test_systemd_boot_time
>   - systemd.SystemdJournalTests.test_systemd_journal
>   - systemd.SystemdServiceTests.test_systemd_disable_enable
>   - systemd.SystemdServiceTests.test_systemd_status
>   - systemd.SystemdServiceTests.test_systemd_stop_start
>   - xorg.XorgTest.test_xorg_running

And we now list the passing tests in this report.
I think this is a useful detail to have.

Cheers,

-Mikko
 
> LAVA Test Suite Results:
>   BITBAKE_COMMIT_HASH: 2c6c40bdab8ebc6c04fe9992524add967fb4a053
>   BITBAKE_GIT_DESCRIBE: yocto-5.2.1-149-g2c6c40bda
>   METAYOCTO_COMMIT_HASH: 66ea31cee15309af07afeee8dec8a6fe1d92d28e
>   METAYOCTO_GIT_DESCRIBE: 5.3_M3-3-g66ea31c
>   OECORE_COMMIT_HASH: e556df28f47e754b53e1f46c97dde2b19fd8fad1
>   OECORE_GIT_DESCRIBE: uninative-5.0-152-ge556df28f4
> 
> This is an automated test report.

Re: [Automated-testing] bitbake, oe-core and meta-yocto tags in sync? (was Re: GPIT LAVA Test Report - Job 123474: OEQA ALL PASSED (24/75) passed)

[Richard Purdie]

On Wed, 2025-11-12 at 11:06 +0200, Mikko Rapeli via lists.yoctoproject.org wrote:
> Hi,
> 
> On Tue, Nov 11, 2025 at 10:35:08AM -0800, ledge-bot@linaro.org wrote:
> > LAVA Test Report for Job 123474
> > ==================================================
> > 
> > GitLab Pipeline Information:
> >   Project: gpit
> >   Branch: main
> >   Pipeline:
> > https://gitlab.com/Linaro/trustedsubstrate/gpit/-/pipelines/2152817594
> >   bitbake: 2c6c40bdab8ebc6c04fe9992524add967fb4a053
> >   bitbake git describe: yocto-5.2.1-149-g2c6c40bda
> >   openembedded-core: e556df28f47e754b53e1f46c97dde2b19fd8fad1
> >   openembedded-core git describe: uninative-5.0-152-ge556df28f4
> >   meta-yocto: 66ea31cee15309af07afeee8dec8a6fe1d92d28e
> >   meta-yocto git describe: 5.3_M3-3-g66ea31c
> 
> So this is our first build and test run with separate bitbake,
> oe-core and meta-yocto built using kas to replace old poky repo
> master branch which no longer builds. Setup details in
> https://gitlab.com/Linaro/trustedsubstrate/gpit
> 
> The "git describe" output with tags is a bit confusing. Would it be
> possible to use same version/milestone names there?
> 
> Currently it's a bit hard to tell if the trees are fully in sync and
> if
> yocto-5.2.1, uninative-5.0 and 5.3_M3 are compatible with each other
> or not. From pure tags it is not easy to see if trees used same
> branches, for example.

This whole area is a mess.

The OE TSC required certain things in the past. We'd have to dig out
the minutes to remember all the details but I didn't 100% agree with
what was asked for and we went ahead and did it regardless. We probably
need to revisit this but the new TSC has been unable to work out a time
to meet or schedule its meetings. Even if we had, someone needs to look
up and summarise what decisions were made and why. Keep in mind there
is years of some people being very vocal if there were yocto references
outside of yocto 'owned' repositories (i.e. in bitbake/oe-core). I
think we're past some of that now but certainly not all of it. I
actually dread this whole topic.

We also need to work out what we are currently doing and not doing and
why. I know I recently asked release engineering to stop doing certain
things as they were completely non-sensical. We had a load of
problematic tags to remove and new ones to create to replace them to
try and ensure everything was consistent. I'd have to go and check
whether that happened and whether we're now doing the right things or
not.

Now, that all sounds fine, some people need to go and do some painful
work. It will be far from simple, take a ton of time and nobody will
actually really see it all or ever value the end result but it can be
done. Even if (when) we do this, it isn't going to quite give you what
you want due to things like the uninative tags. Those usually happen
mid release and will "corrupt" the git describe output as those tags
will be more recent.

So yes, it does need fixing but it isn't as simple as you might think.
I'm totally burnt out and the current release is stalled. Perhaps
someone else from the TSCs wants to go and resolve this?

Cheers,

Richard