From: "Mathieu Dubois-Briand" <mathieu.dubois-briand@bootlin.com>
To: "Sumanth Gavini" <sumanth.gavini@windriver.com>,
<openembedded-core@lists.openembedded.org>
Cc: <alex.kanavin@gmail.com>, <mattware@cisco.com>,
<esparlin@cisco.com>, <peter.marko@siemens.com>,
<randy.macleod@windriver.com>
Subject: Re: [PATCH v13 09/10] rpm: Fix pkgconfig() Provides missing in cross-built RPMs
Date: Wed, 12 Aug 2026 13:40:55 +0200 [thread overview]
Message-ID: <DKMXR2LKXWBZ.3C4RKC617CK0H@bootlin.com> (raw)
In-Reply-To: <DKM6IAUNZOOF.11QH0FJ6NRSQ2@bootlin.com>
On Tue Aug 11, 2026 at 4:19 PM CEST, Mathieu Dubois-Briand wrote:
> On Mon Aug 10, 2026 at 6:52 AM CEST, Sumanth Gavini wrote:
>> * Added 0002-pkgconfigdeps.sh-Unset-cross-compilation-PKG_CONFIG-v.patch
>> - rpm 6's internal dependency generator runs pkgconfigdeps.sh during
>> rpmbuild, but OE's cross-build environment exports PKG_CONFIG_SYSROOT_DIR
>> which causes pkg-config --print-provides to silently fail on .pc files
>> in the package buildroot.
>>
>> Unset cross-compilation PKG_CONFIG_* variables in pkgconfigdeps.sh so
>> that pkgconfig() Provides/Requires are correctly generated.
>>
>> Without this fix, do_rootfs fails with errors like:
>> nothing provides pkgconfig(ogg) needed by flac-dev
>>
>> Signed-off-by: Sumanth Gavini <sumanth.gavini@windriver.com>
>> ---
>
> Hi Sumanth,
>
> Thanks for the new version.
>
> This looks better, at least for the dnf install errors. Yet we still
> have errors when reusing sstate-cache, as recipes generating pkgconfig
> should be rebuilt, but their signatures did not change:
>
> https://autobuilder.yoctoproject.org/valkyrie/#/builders/4/builds/4377
>
> I made some horrible hack in my branch to change the PR of many
> recipes, and the build now succeeds. I only tested one configuration,
> but we can expect similar results on the other ones.
>
> https://autobuilder.yoctoproject.org/valkyrie/#/builders/4/builds/4386
>
> So, depending on how you address Ross comment, you might need to make
> sure recipes signature is going to change.
>
> We also had some selftest errors: I will investigate a bit further and
> come back to you.
> https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/4525
>
> Thanks,
> Mathieu
So, talking about failed selftests. We can see three failing tests on
the autobuilder:
2026-08-11 15:19:21,651 - oe-selftest - INFO - debuginfod.Debuginfod.test_debuginfod_qemu (subunit.RemotedTestCase)
2026-08-11 15:19:21,652 - oe-selftest - INFO - ... FAIL
...
2026-08-11 16:53:01,397 - oe-selftest - INFO - sdk.SDKTests.test_sdk_manifests (subunit.RemotedTestCase)
2026-08-11 16:53:01,401 - oe-selftest - INFO - ... FAIL
...
2026-08-11 16:55:44,191 - oe-selftest - INFO - sdk.SDKTests.test_sdk_runqemu (subunit.RemotedTestCase)
2026-08-11 16:55:44,191 - oe-selftest - INFO - ... FAIL
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/4537
The first one is easy to reproduce locally. Basically, I have this in my
local.conf:
SDKMACHINE = "x86_64"
PACKAGE_CLASSES = "package_rpm package_deb package_ipk"
INHERIT += 'image-buildinfo'
IMAGE_BUILDINFO_VARS:append = ' IMAGE_BASENAME IMAGE_NAME'
SANITY_TESTED_DISTROS = ''
OEQA_TESTDISPLAY = ':1'
EXTRA_IMAGE_FEATURES ?= 'allow-empty-password empty-root-password allow-root-login'
OE_FRAGMENTS += 'distro/poky'
Then you can run:
oe-selftest -v -r debuginfod.Debuginfod.test_debuginfod_qemu
I was not able to reproduce the two other ones locally so far. It looks
like another pkgconfig issue. I suggest we focus on other issues right
now, hopefully the SDKTests failures will be gone with your next
changes.
Thanks,
Mathieu
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
next prev parent reply other threads:[~2026-08-12 11:41 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-10 4:52 [PATCH v13 00/10] rpm: 4.20.1 -> 6.0.2 Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 01/10] package_rpm.bbclass: Drop external dependency generator to support rpm 6 Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 02/10] package_rpm.bbclass: Define _lib and _libdir for rpmbuild Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 03/10] lib/oe/package.py: Don't add ldconfig_postinst_fragment for glibc or musl Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 04/10] lib/oe/package.py: Don't redirect stderr Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 05/10] target-sdk-provides-dummy: Add pkg-config to DUMMYPROVIDES Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 06/10] rpm: 4.20.1 -> 6.0.2 Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 07/10] libarchive: Make it work with rpm 6.0.2 Sumanth Gavini
2026-08-10 4:52 ` [PATCH v13 08/10] rpm: filter /usr/bin/pkg-config from RPM auto-generated Requires Sumanth Gavini
2026-08-10 9:43 ` [OE-core] " Ross Burton
2026-08-12 0:54 ` Gavini, Sumanth
2026-08-10 4:52 ` [PATCH v13 09/10] rpm: Fix pkgconfig() Provides missing in cross-built RPMs Sumanth Gavini
2026-08-11 14:19 ` Mathieu Dubois-Briand
2026-08-12 11:40 ` Mathieu Dubois-Briand [this message]
2026-08-12 23:58 ` Gavini, Sumanth
2026-08-10 4:52 ` [PATCH v13 10/10] libarchive: Add RPM format reader to support rpm 6 Sumanth Gavini
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=DKMXR2LKXWBZ.3C4RKC617CK0H@bootlin.com \
--to=mathieu.dubois-briand@bootlin.com \
--cc=alex.kanavin@gmail.com \
--cc=esparlin@cisco.com \
--cc=mattware@cisco.com \
--cc=openembedded-core@lists.openembedded.org \
--cc=peter.marko@siemens.com \
--cc=randy.macleod@windriver.com \
--cc=sumanth.gavini@windriver.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.