Hi,
My assumption here is that the /bin vs /usr/bin split is effectively pointless.
"The Case for the /usr Merge”[1] covers a lot of the historic rationale for
this split, but the tl;dr is that historically / might have been mounted before
/usr, so binaries that are needed for early boot before /usr is mounted should
be in /bin. This is entirely redundant now, we have initramfs for early boot
that can mount the rest of the file systems.
I propose running through oe-core (at least) and removing any recipe tweaks to
move files around between bindir and base_bindir, for example libusb1[2].
These changes serve no purpose beyond making the recipes more complex and prone
to issues, for example the libusb1 recipe sets libdir to ${base_libdir} but
then also has to move the .pc file back into ${libdir}.
This process has been happening slowly for some time already (see sysfsutils
from 2023[3]), I just want to clean the rest of oe-core up in one hit.
Note that I am not talking about enabling usrmerge by default, but simply
removing obsolete OE-specific changes to recipes.
Cheers,
Ross
[1] https://www.freedesktop.org/wiki/Software/systemd/TheCaseForTheUsrMerge/
[2]
https://git.openembedded.org/openembedded-core/tree/meta/recipes-support/libusb/libusb1_1.0.29.bb
[3]
https://git.openembedded.org/openembedded-core/commit/?id=a81dc665e89ccc6fe9ed025e679f22956767271b
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#2202):
https://lists.openembedded.org/g/openembedded-architecture/message/2202
Mute This Topic: https://lists.openembedded.org/mt/116291466/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-architecture/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
