On Wed, May 19, 2021 at 6:46 PM Konrad Weihmann <kweihm...@outlook.com> wrote: If I am not mistaken this breaks wide areas of meta-freescale + meta-qt5-extra. I have seen the manifest warning using meta-freescale for long time (and many licensed not found in do_rootfs) but now builds fail. Most recipes failing are native recipes inheriting meta-qt5's cmake_qt5-bbcalss - have many of them in meta-qt5-extra. To analyse I did:
[superandy@thanks-buchmuellers oe-core]$ bitbake -e extra-cmake-modules-native | grep ^DEPENDS DEPENDS="qtbase qtbase-native cmake-native qtbase-native qtbase-native cmake-native qttools-native python3-native python3-native ninja-native" As soon as I remove (target) qtbase from DEPENDS, build is fine. Have the strong feeling that the dynamic-packagearch magic [1] is the culprit confusing a native recipe finding target recipe's manifest if that recipe is marked as SO-SPECIFIC - as qtbase [2] is. Wrote here because I hope the audience is wider - will open an issue at meta-freescale tomorrow after zzz... Cheers, Andreas [1] https://github.com/Freescale/meta-freescale/blob/master/classes/fsl-dynamic-packagearch.bbclass [2] https://github.com/Freescale/meta-freescale/blob/733e1683ce8895a960c01a3ee1fce40118d82982/conf/machine/include/imx-base.inc#L246
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#152454): https://lists.openembedded.org/g/openembedded-core/message/152454 Mute This Topic: https://lists.openembedded.org/mt/82932967/21656 Group Owner: openembedded-core+ow...@lists.openembedded.org Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
