setfacl -nm g:adm:rx,d:g:adm:rx /home/portage_temp/portage/sys-apps/systemd-208-r2/image//var/log/journal/ make[4]: setfacl: Command not found Makefile:14533: recipe for target 'libsystemd-journal-install-hook' failed make[4]: [libsystemd-journal-install-hook] Error 127 (ignored) setfacl -nm g:wheel:rx,d:g:wheel:rx /home/portage_temp/portage/sys-apps/systemd-208-r2/image//var/log/journal/ make[4]: setfacl: Command not found Makefile:14533: recipe for target 'libsystemd-journal-install-hook' failed make[4]: [libsystemd-journal-install-hook] Error 127 (ignored) sys-apps/systemd should have build-time dependency on sys-apps/acl. Reproducible: Always
Created attachment 368430 [details] build.log.gz Sorry for compressed log, but it's too big.
(In reply to nzqr from comment #0) > sys-apps/systemd should have build-time dependency on sys-apps/acl. The ebuild already has a DEPEND on acl: ago@stonebridge ~ $ grep acl /usr/portage/sys-apps/systemd/systemd-208-r2.ebuild acl? ( sys-apps/acl )
acl? means that dependency is conditional.
Interesting. I wonder if that makes sense at all -- I don't think we're having any guarantees that /tmp (/var/tmp) will support xattrs and that xattrs will be copied to the target fs.
Yeah, it probably doesn't make much sense to include /var/log/journal/ in the installation image. If we create it, it should probably be done in pkg_postinst.
These same ACLs get applied at boot by a tmpfiles snippet. The errors in the build log can be safely ignored.