pytest-services 2.2.2

Infrastructure fixtures

• worker_id

  Id of the worker if tests are run using pytest-xdist. It is set to local if tests are not run using pytest-xdist (with –dist command line option set to load). Has a deprecated alias slave_id which will be removed in a future version.

• session_id

  Test session id. Globally unique, and of course also guaranteed to be different for potentially multiple test sessions running on same test node via pytest-xdist.

• watcher_getter

  Function to instantiate test service watcher (popen object). Includes automatic finalization (exiting) of the service process, and testing the service before returning the watcher from the function. Example of usage for memcached service:

@pytest.fixture(scope='session')
def memcached(request, run_services, memcached_socket, watcher_getter):
    """The memcached instance which is ready to be used by the tests."""
    if run_services:
        return watcher_getter(
            name='memcached',
            arguments=['-s', memcached_socket],
            checker=lambda: os.path.exists(memcached_socket),
            # Needed for the correct execution order of finalizers
            request=request,
        )

• services_log

  Logger used for debug logging when managing test services.

• root_dir

  Parent directory for test service artifacts (disk based). Set to /tmp by default.

• base_dir

  Base directory for test service artifacts (disk based), unique subdirectory of root_dir. Automatically removed recursively at the end of the test session.

• temp_dir

  Temporary directory (disk based), subfolder of the base_dir. Used for strictly temporary artifacts (for example - folder where files are uploaded from the user input).

• memory_root_dir

  Parent directory for test service artifacts (memory based). Main idea of having memory base directory is to store performance-critical files there. For example - mysql service will use it to store database file, it speeds up mysql server a lot, especially database management operations. Set to /var/shm by default, with a fallback to ‘root_dir`. Note that if apparmor is running on your system, most likely it will prevent your test service to use it (for example - mysql has it’s apparmor profile). You you’ll need to disable such profile in apparmor configuration. Example of disabling apparmor for mysqld:

sudo ln -s /etc/apparmor.d/usr.sbin.mysqld /etc/apparmor.d/disable/
sudo /etc/init.d/apparmor restart

• memory_base_dir

  Base directory for test service artifacts (memory based), unique subdirectory of memory_root_dir. Automatically removed recursively at the end of the test session.

• memory_temp_dir

  Temporary directory (memory based), subfolder of the base_dir.

• lock_dir

  Lock files directory for storing locks created for resource assignment (ports, display, etc). Subfolder of memory_root_dir.

• run_dir

  Process id and socket files directory (like system-wide /var/run but local for test session). Subfolder of memory_root_dir.

• port_getter

  Function to get unallocated port. Automatically ensures locking and un-locking of it on application level via flock.

• display_getter

  Function to get unallocated display. Automatically ensures locking and un-locking of it on application level via flock.

• lock_resource_timeout

  Used in function lock_resource. A maximum of total sleep between attempts to lock resource.

Service fixtures

• memcached

  Start memcached instance. Requires pylibmc installed or memcache indicated as an extra (pip install ‘pytest-services[memcached]’).

• memcached_socket

  Memcached unix socket file name to be used for connection.

• memcached_connection

  Memcached connection string.

• do_memcached_clean

  Determine if memcached should be cleared before every test run. Equals to run_services fixture by default. Requires pylibmc installed or memcache indicated as an extra (pip install ‘pytest-services[memcached]’).

• memcached_client

  A pylibmc.Client instance bound to the service. Requires pylibmc installed or memcache indicated as an extra (pip install ‘pytest-services[memcached]’).

• mysql

  Start mysql-server instance.

• mysql_database_name

  MySQL database name to be created after initialization of the mysql service system database.

• mysql_database_getter

  Function with single parameter - database name. To create additional database(s) for tests. Used in mysql_database fixture which is used by mysql one.

• mysql_connection

  MySQL connection string.

• xvfb

  Start xvfb instance.

• xvfb_display

  Xvfb display to use for connection.

• xvfb_resolution

  Xvfb display resolution to use. Tuple in form (1366, 768, 8).

Utility functions

Django settings

In some cases, there’s a need of switching django settings during test run, because several django projects are tested whithin the single test suite. pytest_services.django_settings simplifies switching of django settings to a single function call:

• setup_django_settings

  Override the enviroment variable and call the _setup method of the settings object to reload them.

Example of usage:

conftest.py:

from pytest_services import django_settings

django_settings.clean_django_settings()
django_settings.setup_django_settings('your.project.settings')

Note that the nice project pytest-django doesn’t help with the situation, as it’s single django project oriented, as well as standard django testing technique. Single project approach works fine, as long as there are no fixtures to share between them, but when there are fixtures to share, then you can get benefit of joining several django projects tests into a single test run, because all session-scoped fixtures will be instantiated only once for all projects tests. The benefit is only visible if you have big enough test suite and your fixtures are heavy enough.