For example: bitbake lib32-core-image-minimal Both lib32-opkg-utils and opkg-utils have been built, this is incorrect, only lib32-opkg-utils should be built.
And opkg-utils would cause python bash, bc, openssl and other pgks to be built, this makes the build very slow.
*** Bug 8882 has been marked as a duplicate of this bug. ***
The issue is that images need to contain run-postinsts which depends on initscripts which depends on update-alternatives which comes from opkg-utils. opkg-utils depends on python since it contains python scripts and python has a large set of dependencies. The only real way to solve this would be to separate out the build of update-alternatives from opkg-utils itself, much as we used to do from opkg: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=44b538eedab7c255051fa3375f9f2439cd2db3dd [dropped by Marcin back in 2008]. Torn on whether this is worth while or we just accept that lib32 builds build a bit more...
I submitted a patch trying to solve the issue: https://patchwork.openembedded.org/series/24442/# However, I am afraid that I need a help with this one. I tried to generate dependency graphs: - update-alternatives has almost no deps at all - update-alternatives without bash dependency still does have quite a big bunch of them, with e.g. python3-native included - update-alternatives with bash dependency - I was not even able to display the graph by xdot... but a fast dep file check shows that e.g. python3 dependent tasks are still present there Any suggestions welcome.
Bulk move from 4.99 or 0.00 to 5.99
Robert, is this fixed?
My hunch is it is not. I have this in a local branch that I never got around to finishing/testing: https://git.yoctoproject.org/poky-contrib/commit/?h=ross/ua&id=ac03d59a0b6d30515159ab7f3dfe9f8c41474263 I can't recall if I didn't send it because it wasn't finished, or if it was just a bad time in the cycle to submit.