Proper way to test existing routes

unittest.TestCase

I have read the docs. However, I couldn’t quite figure out how to write test cases with the unittest.TestCase class.

I am trying to use the autofixture feature here: https://docs.pytest.org/en/latest/unittest.html and also to try to organize my code better.

Proper route testing

I am slightly confused by the following from the docs http://pytest-sanic.readthedocs.io/en/latest/fixtures.html

@pytest.yield_fixture
def app():
    app = Sanic("test_sanic_app")

    @app.route("/test_get", methods=['GET'])
    async def test_get(request):
        return response.json({"GET": True})

If I have an existing sanic app that requires testing, why am I writing routes here? Are these test routes or the app’s routes?

Using your documentation, I have been writing the test pages like this but am unsure if this is the proper way to do it.

from cr.app import app as webapp

@pytest.yield_fixture
def app():
    _webapp = webapp
    yield _webapp

@pytest.fixture
def test_cli(loop, app, test_client):
    return loop.run_until_complete(test_client(app))

async def test_index(test_cli):
    resp = await test_cli.get('/')
    assert resp.status == 200

async def test_player(test_cli):
    resp = await test_cli.get('/player')
    assert resp.status == 200

Any help would be appreciated.

[garyo]

The example in this issue should replace the current quickstart example! It’s much more realistic & useful than the standalone one. I was confused for hours, not sure why the test code was creating routes, and couldn’t figure out how the real routes were getting into the test. (It shouldn’t and they weren’t; that’s a toy example that mixes test code with real code.) This is what is working for me:

# define create_app() in your server.py
sys.path.append(str(pytest.config.rootdir)) # your app's toplevel dir
from app.server import create_app

@pytest.yield_fixture
def testapp():
    app = create_app()
    yield app

# ... test as usual

[yunstanford]

because it starts Sanic in async way, https://github.com/channelcat/sanic/blob/master/sanic/app.py#L719.

As for unit tests, we need a loop to run all async unit tests under the hood.(async def test_)

Normally, when you write your application, if you don’t explicitly pass loop, that means you just use the default loop from asyncio.get_event_loop(). it doesn’t mean you don’t need event loop.

Furthermore, you can simply instantiate the sanic app with app = Sanic(__name__), we just hide the loop part:

https://github.com/channelcat/sanic/blob/master/sanic/app.py#L743 https://github.com/channelcat/sanic/blob/master/sanic/server.py#L563 https://github.com/channelcat/sanic/blob/master/sanic/worker.py#L45