This bug was originally reported by Peter here: https://lists.openembedded.org/g/openembedded-core/message/197092. But since the right solution is not found yet it might be better to track this as a bug. Following workflow is broken when built from sstate-cache: https://docs.yoctoproject.org/singleindex.html#setting-up-the-extensible-sdk-environment-directly-in-a-yocto-build This is already broken in kirkstone. Reproducer: $ bitbake meta-ide-support && bitbake build-sysroots -c build_native_sysroot $ ls -1 build/tmp/sysroots/*/*gdb-cross* && echo OK || echo FAIL #succeeds from scratch build/tmp/sysroots/x86_64/manifest-x86_64-gdb-cross-x86_64.populate_sysroot OK $ rm -rf build/tmp $ bitbake meta-ide-support && bitbake build-sysroots -c build_native_sysroot $ ls -1 build/tmp/sysroots/*/*gdb-cross* && echo OK || echo FAIL #fails from sstate-cache ls: cannot access 'build/tmp/sysroots/*/*gdb-cross*': No such file or directory FAIL build-sysroot populates sysroot with recipes which run populate_sysroot or populate_sysroot_setscene. Dependency on its own populate_sysroot does not seem to work when building from sstate-cache as this task is not executed. This seem to becaused by inheriting nopackages class which deletes tasks thus messing the setscene dependencies.
Does not appear to be fixed yet.
Bulk move to 5.2-M3.
Bulk move to 5.2-M3 and this time I'll actually specify M3!
Still happening on poky/master 331aef4bcf 2025-03-14 scripts: Add clean-hashserver-database script
Bulk move of most 5.2-M4 unassigned bugs to 5.3-M2.
Still fails on master: ❯ git log --oneline -1 f9e0afedea (HEAD -> master, origin/master, origin/HEAD) oe-selftest: add wic.Wic.test_grub_install_pcbios Bump to M4, CC Sundeep.