diff options
author | Ed Santiago <santiago@redhat.com> | 2019-04-11 16:14:13 -0600 |
---|---|---|
committer | Ed Santiago <santiago@redhat.com> | 2019-04-11 20:08:27 -0600 |
commit | 859bda1936ee67a3da7a197f5207419493fffef0 (patch) | |
tree | e844da7650ea8ec422a280388d7cec7d5c9b9703 /vendor/github.com/hpcloud/tail/winfile | |
parent | cb2b019d5debadbe29cba59e93130bd8c562771a (diff) | |
download | podman-859bda1936ee67a3da7a197f5207419493fffef0.tar.gz podman-859bda1936ee67a3da7a197f5207419493fffef0.tar.bz2 podman-859bda1936ee67a3da7a197f5207419493fffef0.zip |
new uidmap BATS test: fix
Various problems, one of which was causing the test to fail
completely (otherwise I wouldn't have caught the others):
- option is --uidmap, not --uidmapping
- run_podman cannot be piped (| grep /sys/kernel). That's
an unfortunate limitation of BATS. Any invocation of 'run'
saves results to $output, which then has to be tested
in a separate step.
- do so, using 'run' and 'grep' and 'is' to produce
readable messages on failure
- remove "$expected_rc", that looks like a copy/paste bug
from a few lines above.
Skip entire test if rootless. (The one without --net=host
passes, but it also passes with older podman as both root
and rootless. I don't think it's actually testing anything,
but agree with leaving it in to catch weird regressions).
We really need to get these tests running in CI.
Signed-off-by: Ed Santiago <santiago@redhat.com>
Diffstat (limited to 'vendor/github.com/hpcloud/tail/winfile')
0 files changed, 0 insertions, 0 deletions