2017-11-16 05:28:03 +03:00
# This used to live in test-basic.sh, but it's now shared with test-basic-unified.sh
basic_test( ) {
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
if ostree --repo= ${ repo } ls -R ${ treeref } /usr/etc/passwd-; then
2017-11-16 05:28:03 +03:00
assert_not_reached "Found /usr/etc/passwd- backup file in tree"
fi
2018-10-08 00:46:31 +03:00
echo "ok passwd no backups"
validate_passwd( ) {
f = $1
shift
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } cat ${ treeref } /usr/lib/$f | grep -v '^root' | sort > $f .tree
cat config/$f | while read line; do
2018-10-08 00:46:31 +03:00
if ! grep -q " $line " " $f .tree " ; then
echo "Missing entry: %line"
fi
done
}
validate_passwd passwd
validate_passwd group
2017-12-12 17:26:20 +03:00
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } cat ${ treeref } /usr/etc/default/useradd > useradd.txt
2019-01-05 23:27:08 +03:00
assert_file_has_content_literal useradd.txt HOME = /var/home
2019-02-06 20:04:39 +03:00
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } cat ${ treeref } \
2019-02-06 20:04:39 +03:00
/usr/etc/selinux/targeted/contexts/files/file_contexts.homedirs > homedirs.txt
2019-02-06 23:54:15 +03:00
assert_file_has_content homedirs.txt '^/var/home'
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } cat ${ treeref } \
2019-02-06 23:54:15 +03:00
/usr/etc/selinux/targeted/contexts/files/file_contexts.subs_dist > subs_dist.txt
assert_not_file_has_content subs_dist.txt '^/var/home \+'
assert_file_has_content subs_dist.txt '^/home \+/var/home$'
2019-01-05 23:27:08 +03:00
echo "ok etc/default/useradd"
2017-12-12 17:26:20 +03:00
for path in /usr/share/rpm /usr/lib/sysimage/rpm-ostree-base-db; do
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls -R ${ treeref } ${ path } > db.txt
2017-12-12 17:26:20 +03:00
assert_file_has_content_literal db.txt /Packages
done
2020-07-03 17:50:02 +03:00
ostree --repo= ${ repo } ls ${ treeref } /usr/lib/sysimage/rpm >/dev/null
2017-12-12 17:26:20 +03:00
echo "ok db"
2017-11-16 05:28:03 +03:00
2021-02-08 22:59:12 +03:00
ostree --repo= ${ repo } cat ${ treeref } /usr/lib/rpm/macros.d/macros.rpm-ostree > rpm-ostree-macro.txt
assert_file_has_content_literal rpm-ostree-macro.txt '%_dbpath /usr/share/rpm'
echo "ok rpm macro"
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } show --print-metadata-key exampleos.gitrepo ${ treeref } > meta.txt
2017-11-16 05:28:03 +03:00
assert_file_has_content meta.txt 'rev.*97ec21c614689e533d294cdae464df607b526ab9'
assert_file_has_content meta.txt 'src.*https://gitlab.com/exampleos/custom-atomic-host'
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } show --print-metadata-key exampleos.tests ${ treeref } > meta.txt
2017-11-16 05:28:03 +03:00
assert_file_has_content meta.txt 'smoketested.*e2e'
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } show --print-metadata-key rpmostree.rpmmd-repos ${ treeref } > meta.txt
assert_file_has_content meta.txt 'id.*cache.*timestamp'
ostree --repo= ${ repo } show --print-metadata-key foobar ${ treeref } > meta.txt
2019-07-08 17:34:59 +03:00
assert_file_has_content meta.txt 'bazboo'
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } show --print-metadata-key overrideme ${ treeref } > meta.txt
2019-07-08 17:34:59 +03:00
assert_file_has_content meta.txt 'new val'
2017-11-16 05:28:03 +03:00
echo "ok metadata"
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls -R ${ treeref } /usr/lib/modules | grep -v /kernel/ > ls.txt
assert_file_has_content ls.txt '/vmlinuz$'
assert_file_has_content ls.txt '^-00644 .*/initramfs.img$'
echo "ok kernel and initramfs"
2017-11-16 05:28:03 +03:00
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls ${ treeref } /usr/share > share.txt
assert_not_file_has_content share.txt /usr/share/man
# test-misc-tweaks tests the docs path
echo "ok no manpages"
2017-11-16 05:28:03 +03:00
2018-06-22 22:15:04 +03:00
# https://github.com/projectatomic/rpm-ostree/pull/1425
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls ${ treeref } /usr/etc > ls.txt
assert_not_file_has_content ls.txt 'machine-id'
# test-misc-tweaks tests the machine-id compat path
echo "ok no machine-id"
2018-06-22 22:15:04 +03:00
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls ${ treeref } usr/etc/systemd/system/multi-user.target.wants/chronyd.service > preset.txt
2017-12-14 01:17:03 +03:00
assert_file_has_content_literal preset.txt '-> /usr/lib/systemd/system/chronyd.service'
echo "ok systemctl preset"
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls -X ${ treeref } usr/bin/docker > docker.txt
2018-01-02 21:51:18 +03:00
assert_file_has_content_literal docker.txt 'system_u:object_r:container_runtime_exec_t:s0'
echo "ok container-selinux"
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls ${ treeref } /usr/bin/su > su.txt
2017-12-08 22:21:04 +03:00
assert_file_has_content su.txt '^-04[71][0-7][0-7]'
echo "ok setuid"
2020-10-14 03:58:43 +03:00
ostree --repo= ${ repo } ls -X ${ treeref } /usr/bin/arping > arping.txt
assert_file_has_content_literal arping.txt "b'security.capability', [byte"
2017-12-14 05:42:59 +03:00
echo "ok fcaps"
2017-11-16 05:28:03 +03:00
# https://github.com/projectatomic/rpm-ostree/issues/669
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls ${ treeref } /tmp > ls.txt
assert_file_has_content ls.txt 'd01777 0 0 0 /tmp'
2017-11-16 05:28:03 +03:00
echo "ok /tmp"
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls ${ treeref } /usr/share/rpm > ls.txt
2017-11-16 05:28:03 +03:00
assert_not_file_has_content ls.txt '__db' 'lock'
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } ls -R ${ treeref } /usr/etc/selinux > ls.txt
2017-11-16 05:28:03 +03:00
assert_not_file_has_content ls.txt 'LOCK'
echo "ok no leftover files"
2017-12-05 22:51:26 +03:00
2020-09-10 03:33:11 +03:00
ostree --repo= ${ repo } ls ${ treeref } /usr/share/rpm > ls.txt
assert_file_has_content ls.txt /usr/share/rpm/Packages
assert_not_file_has_content ls.txt rpmdb.sqlite
echo "ok rpmdb is bdb"
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } show ${ treeref } \
2017-12-05 22:51:26 +03:00
--print-metadata-key rpmostree.rpmdb.pkglist > pkglist.txt
assert_file_has_content pkglist.txt 'systemd'
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
assert_file_has_content_literal pkglist.txt 'foobar'
assert_not_file_has_content pkglist.txt 'foobar-rec'
2017-12-15 17:42:10 +03:00
echo "ok compose pkglist"
2019-02-26 18:51:14 +03:00
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } cat ${ treeref } /usr/share/rpm-ostree/treefile.json > treefile.json
2019-02-26 18:51:14 +03:00
assert_jq treefile.json '.basearch == "x86_64"'
echo "ok basearch"
2019-07-18 18:25:24 +03:00
tests/compose: Target FCOS 31, move off of PAPR
Again, a lot going on here, but essentially, we adapt the compose tests
to run either privileged or fully unprivileged via supermin, just like
cosa.
I actually got more than halfway through this initially using `cosa
build` directly for testing. But in the end, we simply need more
flexibility than that. We want to be able to manipulate exactly how
rpm-ostree is called, and cosa is very opinionated about this (and may
also change from under us in the future).
(Another big difference for example is that cosa doesn't care about
non-unified mode, whereas we *need* to have coverage for this until we
fully kill it.)
Really, the most important bit we want from there is the
unprivileged-via-supermin bits. So we copy and adapt that here. One
obvious improvement then is sharing this code more easily (e.g. a
`cosa runasroot` or something?)
However, we still use the FCOS manifest (frozen at a specific tag). It's
a realistic example, and because of the lockfiles and pool, we get good
reproducibility.
2019-12-22 01:42:09 +03:00
ostree --repo= ${ repo } rev-parse ${ treeref } ^ > parent.txt
2019-07-18 18:25:24 +03:00
assert_file_has_content parent.txt 01ba4719c80b6fe911b091a7c05124b64eeece964e09c058ef8f9805daca546b
echo "ok --parent"
2021-04-20 19:17:20 +03:00
# Check symlinks injected into the rootfs.
ostree --repo= " ${ repo } " ls " ${ treeref } " /usr/lib/alternatives /usr/lib/vagrant | grep '^d00755' > symlinks.txt
assert_file_has_content_literal symlinks.txt '/usr/lib/alternatives'
assert_file_has_content_literal symlinks.txt '/usr/lib/vagrant'
ostree --repo= " ${ repo } " ls " ${ treeref } " /var/lib/alternatives /var/lib/vagrant /usr/local | grep '^l00777' > symlinks.txt
assert_file_has_content_literal symlinks.txt '/usr/local -> ../var/usrlocal'
assert_file_has_content_literal symlinks.txt '/var/lib/alternatives -> ../../usr/lib/alternatives'
assert_file_has_content_literal symlinks.txt '/var/lib/vagrant -> ../../usr/lib/vagrant'
echo "ok symlinks"
2019-05-08 16:46:05 +03:00
}
2019-07-18 18:25:24 +03:00
