This is the new hotness.
Add doc8 linting to the pep8 target while we're doing it. And, slim down
pep8 tox env. It doesn't need to install shade or any of the
requirements.
Change-Id: I1e73ee503c81f92b4fe84bdd384f856290b6d5dd
stestr is a fork of testrepository, designed specifically to concentrate
on being a dedicated test runner. [1] The testrepository project is
basically not active anymore and has several long standing bugs and
useablility issues. stestr is still actively maintained and fixes a
large number of those issues with testrepository.
[1] https://github.com/mtreinish/stestr
Change-Id: I9b8049fd2c78d8f90aed5a8b35c0a9b40ad8c463
The oslotest package distributes a shell file that may be used to assist
in debugging python code. The shell file uses testtools, and supports
debugging with pdb.
To enable debugging, run tox with the debug environment. Below are the
following ways to run it.
* tox -e debug module
* tox -e debug module.test_class
* tox -e debug module.test_class.test_method
Change-Id: If0b06dcf094682401c4b09dd72493c678ea2a6b0
Wow. We had futures in our test-requirements so we never noticed it
wasn't in our regular requirements.
Change-Id: I6ae3434effc778524908d009627e9c0da6258f7c
We had added a betamax fixture for create_flavor, but the requests_mock
approach is actually working out much better. Go ahead and replace it to
simplify the test suite a little bit.
Remove _by_flavor tests as they tested a behavior that's actually invalid but
worked in the test by happenstance.
Transition the rest of the file while we're in there.
Change-Id: Ic2457d7380a8af41ed7bf6b264cbdc2240780ff3
Mocking out the whole client is the wrong level. Although we'd still
like to betamax the functional tests - as we put in REST calls, let's
make the mocking of them happen at the requests level so that we test as
much of shade as we can.
Nobody look at the v2 keystone catalog. The existing clouds.yaml entry
was for a v2 cloud and I wanted to keep the patch contained to the task
at hand. We should likely do some crazy testscenarios magic at some
point to make sure _all_ the things work with both v2 and v3.
Change-Id: Ife4d95df5417d329195c737814e89d1370a0597b
The discussion around the previous patch had a simple test case which
was used to show invalid and valid behavior. Turns out, that can be a
test.
Change-Id: I39550d44f4e83f803624a833a081101bad9dc545
Instead of mocking the clients, use keystoneauth1.betamax
fixture to intercept keystoneauth.construct_session, and provide
our own recorded fixtures.
Change-Id: I7b2973b0f89b66c19d6bf10571c3c93692107aa3
We haven't been able to test os_server because we haven't had a
way to reliably determine the image name (which changes from time
to time based on version) from the playbook. A stupidly clever
hack is to just grep out the name from a shell call to the openstack
client and pass that in to the playbook from the command line.
The playbook is simple for now for the initial commit. I expect new
tasks to be added to it later.
Change-Id: I3caedfd4c805126c8da1fc8bab06ccda3fd96528
The OpenStack Release team has created a great release notes management
tool that integrates with Sphinx. Start using it. For reference on how
to use it, see http://docs.openstack.org/developer/reno/
Change-Id: I57692d720174fedb68ab2f52d5a4c496a6d993b2
We've released 1.0, so we should use it. Also, you want at least 0.11.
Update the hacking pin because we need new flake8 to deal with a
pkg_resources issue.
Change-Id: I62767ea281df94f67d453a14e9c8500aeb305e97
For obj_to_dict, we filter out keys that start with _ and values that
are not normal values. Do the same for our warlock conversion.
Change-Id: Idbf4303c1e4151494d1ea814be6ca5f86e76b16d
Adds API methods to list existing routers, create a new logical
router, update an existing router, and delete an existing logical
router by name or UUID.
It is considered an error if more than one router with the same
name exists and you attempt to delete by name.
Also... MOAR TESTS!!! ZOMG
Change-Id: Ie6ea4eb5f2322bdda07e6db87e2cdbabea492ee9