At present the neutron policyd test assumes a network visible to
the demo user has already been created at the time of the test
being executed.
This may not always be the case, so let's manage the resource we
use for the test.
Add Python 3.8 env in Travis CI test matrix.
At present the pinning of flake8 disallows running of lint on
Python 3.8 systems.
Update flake8 ignore-list to ignore W504 instead of W503, the PEP
guidance is that either is ok, but there must be local consistency.
There are more occurences of binary operator before line-break
than after in this repository, and we have also chosen to ignore
W504 in most of our other repositories.
A side effect of migrating to bootstrapping Keystone as opposed
to using the admin_token is that the charm credentials is now
subject to the Keystone policy.
At present the ``list_services`` policy is used as a test of the
Policy Override feature, however revoking access to said call
will make the charm go into an error state as it attempts to use
it as part of managing the Keystone CRUD.
Change the test to use the ``list_credentials`` policy for test
instead.
Providing a policy.d test for Octavia is a bit more complicated as the
obvious policy is just to prevent listing of load balancers;
unfortunately it doesn't work in the upstream, as the code doesn't check
the policy in octavia. However, "showing" an load balancer does involve
a policy check and so the test is more complicated than 'normal' policy
tests in that a resource is created, then the policy override is checked
and then the resource is deleted.
The difference in how openstack-dashboard and every other service uses
policies continues to create special cases. This set of fixes deals
with some more of those differences.
* Addition of the (incomplete) openstack-dashboard test.
* Modification of other charm policyd override tests to support
multi-file policy overrides resource ZIP.
The current implementation may miss waiting for config change to
actually happen as it in some circumstances jumps to waiting for a
idle state prior to the model executing anything.
Add a wait for agent status `executing` immediatelly after config
change to avoid this.
Fixes#123
This adds the modified neutron-api, glance, cinder policyd
tests. In particular, the test verifies that an operation can be
overriden and removed and that the api calls works, doesn't work, and
then works again.
Liam Young
Aurelien Lourot
Frode Nordahl
David Ames
Alex Kavanagh