For running tests in the test suite you have two options. You may go the easy way and just run the test suite in docker. But for some tasks you could also need to install the test suite natively, which requires a little bit more setup since PHP and some dependencies need to be installed.
Both ways to run tests with the test suites are described here.
Let’s see what is available. Invoke the following command from within the root of the oCIS repository.
make -C tests/acceptance/docker help
Basically we have two sources for feature tests and test suites:
At the moment both can be applied to oCIS since the api of oCIS is designed to be compatible to ownCloud.
Since we have to offer an migration path to existing users of ownCloud, you can use your existing ownCloud as storage backend for oCIS. As another storage backend we offer oCIS native storage, also called “oCIS”. This stores files directly on disk. Which storage backend is used is also reflected in the tests, there are always different tests for oCIS storage and ownCloud storage.
You can invoke two types of test suite runs:
- run a full test suite, which consists of multiple feature tests
- run a single feature test
The names of the full test suite make targets have the same naming as in the CI pipeline.
make -C tests/acceptance/docker localApiTests-apiBugDemonstration-ocis runs the same tests as the
localApiTests-apiBugDemonstration-ocis CI pipeline, which runs the oCIS test suite “apiBugDemonstration” against an oCIS with oCIS storage.
make -C tests/acceptance/docker Core-API-Tests-owncloud-storage-3runs the same tests as the
Core-API-Tests-owncloud-storage-3 CI pipeline, which runs the third (out of ten) ownCloud test suite against an oCIS with owncloud storage.
The single feature tests can also be run against the different storage backends. Therefore multiple make targets with the schema test--feature- exists. For selecting a single feature test you have to add an additional
BEHAT_FEATURE=... parameter when invoking the make command:
make -C tests/acceptance/docker test-ocis-feature-ocis BEHAT_FEATURE='tests/acceptance/features/apiBugDemonstration/apiAuthOcs-ocsDELETEAuth.feature'
This must be pointing to a valid feature definition.
By default the tests will be run against docker image built from your current working state of the oCIS repository. For some purposes it might also be handy to use a oCIS image from Docker Hub. Therefore you can provide the optional flag
OCIS_IMAGE_TAG=... which must contain an available docker tag of the owncloud/ocis registry on Docker Hub (eg. ‘latest’).
make -C tests/acceptance/docker localApiTests-apiBugDemonstration-ocis OCIS_IMAGE_TAG=latest
While a test is running or when it is finished, you can attach to the logs generated by the tests.
make -C tests/acceptance/docker show-test-logs
The log output is opened in
less. You can navigate up and down with your cursors. By pressing “F” you can follow the latest line of the output.
During testing we start an redis and oCIS docker container. These will not be stopped automatically. You can stop them with:
make -C tests/acceptance/docker clean
We are using the ownCloud 10 acceptance test suite against oCIS.
All you need to do to get the acceptance tests is check out the core repo:
git clone https://github.com/owncloud/core.git
To start ocis:
PROXY_ENABLE_BASIC_AUTH=true bin/ocis server
PROXY_ENABLE_BASIC_AUTH will allow the acceptance tests to make requests against the provisioning api (and other endpoints) using basic auth.
First we will need to clone the testing app in owncloud which contains the skeleton files required for running the tests. In the ownCloud 10 core clone the testing app with the following command:
git clone https://github.com/owncloud/testing apps/testing
Then run the api acceptance tests with the following command from the root of the ownCloud 10 core repository:
make test-acceptance-api \ TEST_SERVER_URL=https://localhost:9200 \ TEST_OCIS=true \ SKELETON_DIR=apps/testing/data/apiSkeleton \ DELETE_USER_DATA_CMD='rm -rf /var/tmp/ocis/storage/users/nodes/root/* /var/tmp/ocis/storage/users/nodes/*-*-*-*' \ BEHAT_FILTER_TAGS='~@notToImplementOnOCIS&&~@toImplementOnOCIS'
Make sure to adjust the settings
OCIS_REVA_DATA_ROOT according to your environment.
This will run all tests that are relevant to oCIS.
To run a single test add
As a lot of scenarios are written for oC10, we can use those tests for Behaviour driven development in ocis.
Every scenario that does not work in oCIS with “owncloud” storage, is listed in
tests/acceptance/expected-failures-on-OWNCLOUD-storage.md with a link to the related issue.
Every scenario that does not work in oCIS with “ocis” storage, is listed in
tests/acceptance/expected-failures-on-OCIS-storage.md with a link to the related issue.
Those scenarios are run in the ordinary acceptance test pipeline in CI. The scenarios that fail are checked against the
expected failures. If there are any differences then the CI pipeline fails.
Similarly, scenarios that do not work in oCIS with EOS storage are listed in
Additionally, some issues have scenarios that demonstrate the current buggy behaviour in ocis(reva).
Those scenarios are in this ocis repository in
Have a look into the documentation to understand why we are writing those tests.
If you want to work on a specific issue
adjust the core commit id to the latest commit in core so that CI will run the latest test code and scenarios from core. For that change
# The test runner source for API tests CORE_COMMITID=38c91e5cf5fc4ffdc0536ba1d147a2a618ef83b5 CORE_BRANCH=master
locally run each of the tests marked with that issue in the expected failures file.
make test-acceptance-api \ TEST_SERVER_URL=https://localhost:9200 \ TEST_OCIS=true \ DELETE_USER_DATA_CMD='rm -rf /var/tmp/ocis/storage/users/nodes/root/* /var/tmp/ocis/storage/users/nodes/*-*-*-*' \ BEHAT_FEATURE='tests/acceptance/features/apiComments/comments.feature:123'
the tests will fail, try to understand how and why they are failing
fix the code
go back to 2. and repeat till the tests are passing.
remove those tests from the expected failures file
run each of the local tests that were demonstrating the buggy behavior. They should fail.
delete each of the local tests that were demonstrating the buggy behavior.
make a PR that has the fixed code, relevant lines removed from the expected failures file and bug demonstration tests deleted.