Re: [PATCH 0/2] handle DPKG_ARCH=all case for transitive deps

["'MOESSBAUER, Felix' via isar-users" <isar-users@googlegroups.com>]

On Fri, 2025-10-24 at 05:34 +0000, Hombourger, Cedric (FT FDS CES LX)
wrote:
> On Wed, 2025-08-20 at 14:41 +0200, Felix Moessbauer wrote:
> > This series fixes an issue with DPKG_ARCH=all packages that have
> > dependencies to other DPKG_ARCH=all packages. Hereby, the transitive
> > packages were built for the wrong architecture. The issue became
> > apparent when testing the SBOM series by Christoph, which builds
> > packages to be used during the ISAR build (not for the final target).
> > 
> > While fixing this, another issue surfaced where files are missing in
> > dpkg-raw packages.
> 
> In my testing with downstream layers, this change caused some packages
> to be built for architectures they were not targeted for. For instance,
> I am now seeing the k3-rti-wdt firmware package [1] being built for
> both amd64 and arm64. I also had a dpkg-prebuilt recipe for a package
> supported on arm64 and amd64: it is now tried for armhf/i386.

Where is [1]? The question hereby is if the dependency chains are
actually modeled correctly or if the patch is not behaving as expected.
Without in-tree reproducers that's really hard to check.

> 
> Then there was this issue with dpkg-customization packages now leading
> to a "multiple .bb files are due to be built" error. My work-around was
> to add DPKG_ARCH=any to the dpkg-customization class since
> customization packages are anyhow decorated with both DISTRO and
> MACHINE (it is therefore ok and possibly better to have them explicitly
> built for the arch of the MACHINE).

Apart from if we have a bug there or not, these customization packages
should be arch specific (so DPKG_ARCH=any is the right choice).

> 
> Note: Jan had suggested that I also pull [2] but that did not help
> 
> I evidently need to provide you more details. I have had a short
> discussion with Felix and we suspect that multiconfig builds may be
> causing these issues.

Likely. We don't use mc in isar-cip-core and the isar layer itself
seems to be too simple to surface the remaining bugs.

> 
> Anyhow, I hope that I can create minimal reproducers and get them added
> to the CI test suite.

That would be great.

Felix

> 
> > 
> > Best regards,
> > Felix
> > 
> > Felix Moessbauer (2):
> >   dpkg-raw: add files to source package
> >   handle DPKG_ARCH=all case for transitive deps
> > 
> >  meta/classes/dpkg-raw.bbclass  |  4 ++--
> >  meta/classes/multiarch.bbclass | 18 +++++++++++++++++-
> >  2 files changed, 19 insertions(+), 3 deletions(-)
> > 
> 
> -- 
> Cedric Hombourger
> Siemens AG
> www.siemens.com

-- 
Siemens AG
Linux Expert Center
Friedrich-Ludwig-Bauer-Str. 3
85748 Garching, Germany

-- 
You received this message because you are subscribed to the Google Groups "isar-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to isar-users+unsubscribe@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/isar-users/8f24b7a52dd2fb330ba677850343ea4ed6f37d7f.camel%40siemens.com.