On Sun, Jul 7, 2024 at 10:00 AM, Richard Purdie wrote: > > Thanks for the patch, unfortunately this ran into issues in our CI. One > interesting one is this: > > https://valkyrie.yoctoproject.org/#/builders/40/builds/65 > https://valkyrie.yoctoproject.org/#/builders/36/builds/64/steps/17/logs/stdio > > > which isn't very clear what is going on until you look at the files: > > (buildbot-venv) [pokybuild@rocky9-vk-1 ~]$ ls > /home/pokybuild/yocto-worker/qemuarm-oecore/build/build/tmp-glibc/work/qemuarm-oe-linux-gnueabi/core-image-sato/1.0/testimage-sdk/sysroots/cortexa15t2hf-neon-oe-linux-gnueabi/etc/bluetooth/ > -la > total 32 > dr-xr-xr-x. 2 pokybuild pokybuild 4096 Jul 7 11:11 . > drwxr-xr-x. 3 pokybuild pokybuild 4096 Jul 7 11:26 .. > -rw-r--r--. 1 pokybuild pokybuild 928 Jul 7 11:11 input.conf > -rw-r--r--. 1 pokybuild pokybuild 12597 Jul 7 11:11 main.conf > -rw-r--r--. 1 pokybuild pokybuild 120 Jul 7 11:11 network.conf > > The missing write bit on the directory is stopping the files from being > deleted.
I wonder if this resolves the issue, looks exactly like something we need: https://docs.python.org/3/library/shutil.html#rmtree-example ( https://docs.python.org/3/library/shutil.html#rmtree-example ) There's a comment in the Bitbake code saying rmtree is too slow, but that's from 12 years ago. Perhaps it is fast enough today with Python 3.12. I don't think the solution is to request BlueZ to change how it installs it's files/directory.
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#201623): https://lists.openembedded.org/g/openembedded-core/message/201623 Mute This Topic: https://lists.openembedded.org/mt/107029827/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
