On 8/4/2026 12:18 PM, Alexander Kanavin wrote: > CAUTION: This email comes from a non Wind River email account! > Do not click links or open attachments unless you recognize the sender and > know the content is safe. > > On Tue, 4 Aug 2026 at 16:18, Mathieu Dubois-Briand > <[email protected]> wrote: >> Problem 2: conflicting requests >> - nothing provides pkgconfig(libxml-2.0) needed by >> libcomps-dev-0.1.24-r0.core2_64 from oe-repo > > I looked into this a bit. First of all, this is easily reproducible > locally. Set up qemux86-64 poky, run 'bitbake core-image-sato-sdk' and > it will trigger. > > Second, some observations: > > 1. Looking at > tmp/work/x86-64-v3-poky-linux/libcomps/0.1.24/temp/log.do_package_write_rpm > I see: > > Processing files: libcomps-dev-0.1.24-r0.x86_64_v3 > ... > Requires: pkgconfig(expat) pkgconfig(libxml-2.0) > > So why does libcomps-dev rpm require both of these, but only > libxml-2.0 is mentioned in the error? > > 2. Let's first look at > tmp/work/x86-64-v3-poky-linux/expat/2.8.2/temp/log.do_package_write_rpm > > Processing files: libexpat-dev-2.8.2-r0.x86_64_v3 > ... > Provides: expat-dev = 2.8.2 libexpat-dev = 2.8.2-r0 > libexpat-dev(x86-64) = 2.8.2-r0 pkgconfig(expat) = 2.8.2 > > 3. Then let's look at > tmp/work/x86-64-v3-poky-linux/libxml2/2.15.3/temp/log.do_package_write_rpm > > Processing files: libxml2-dev-2.15.3-r0.x86_64_v3 > ... > Provides: libxml2-dev = 2.15.3-r0 libxml2-dev(x86-64) = 2.15.3-r0 > > 4. So it seems that for expat rpm does write the needed pkgconfig() > provides entry, and for libxml2 it does not. Both recipes however do > install and package those files into their respective -dev packages: > > $ find tmp/work/x86-64-v3-poky-linux/expat/2.8.2/packages-split/ |grep pc > tmp/work/x86-64-v3-poky-linux/expat/2.8.2/packages-split/expat-dev/usr/lib/pkgconfig/expat.pc > $ find tmp/work/x86-64-v3-poky-linux/libxml2/2.15.3/packages-split/ |grep > \\.pc > tmp/work/x86-64-v3-poky-linux/libxml2/2.15.3/packages-split/libxml2-dev/usr/lib/pkgconfig/libxml-2.0.pc > > 5. So the next question is why is rpm picking up one as a provides > entry, but not the other. This is where I stopped, but I hope you > Sumanth can take it from here. > > Cheers, > Alex
Hi Mathieu, I have resolved the reported issues and successfully tested the following test cases. If you encounter any failures, please verify them locally and provide detailed reproduction steps. For validation, I used the following configuration: bitbake-setup with poky, without sstate, and qemux86 bitbake core-image-sato -c populate_sdk bitbake core-image-sato-sdk -c rootfs oe-selftest -r debuginfod.Debuginfod.test_debuginfod_qemu Alex, thanks for the quick verification and for providing the reproduction steps. They were helpful in identifying and resolving the issue. -- # Sumanth Gavini # Wind River Linux
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#243104): https://lists.openembedded.org/g/openembedded-core/message/243104 Mute This Topic: https://lists.openembedded.org/mt/120589443/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
