404 error on second client.get call

[Nicolo]

My client get call always fails in the second test. Any idea what I'm doing wrong?

from app import create_app
import pytest
from config import TestConfig


@pytest.fixture
def app():
    app = create_app(TestConfig)
    return app


class TestSomething:
    def test_index(self, client):
        res = client.get("/")
        assert res.status_code == 200

    def test_index_again(self, client):
        res = client.get("/")
        assert res.status_code == 200

I get the following error.

========================================================================================================================= FAILURES 
______________________________________________________________________________________________________________ TestSomething.test_index_again ______________________________________________________________________________________________________________

self = <conftest.TestSomething object at 0x10b2f9710>, client = <FlaskClient <Flask 'app'>>

    def test_index_again(self, client):
        res = client.get("/")
>       assert res.status_code == 200
E       AssertionError: assert 404 == 200
E        +  where 404 = <<class 'pytest_flask.plugin.JSONResponse'> streamed [404 NOT FOUND]>.status_code

conftest.py:19: AssertionError
================================================================================================================= short test summary info 
FAILED conftest.py::TestSomething::test_index_again - AssertionError: assert 404 == 200
=============================================================================================================== 1 failed, 1 passed in 0.26s 

[Cassius Close]

I'm having this issue too. Checking app.url_map shows that on the second test, the routes I define are not registered. So I have no idea why this is.

The same thing also happens without using pytest-flask, however - even if I call create_app twice in a python shell.

[Cassius Close]

Manually adding in the routes with app.add_url_rule() in the app factory function does end up working for me.

Update: I don't know how you're defining your routes, but my hunch is that it's because I'm using @current_app.route() for routes outside of blueprints, and creating multiple Flask instances doesn't play nicely with that.

[Nicolo]

@CassiusClose Thanks for leaving these comments. It helped me figure out a solution.

I'm not sure I fully understand why this happening, but the issue stemmed from using a create_app factory and importing routes. Moving to a proper blueprint implementation solved the issue.

I still don't understand why the first call of create_app would initialize the routes via the import, but subsequent calls did not.

[Cassius Close]

@nicolo Glad I could help! Could I ask what your solution was?

[Nicolo]

I switched my create_app method in my app/__init__.py to use blueprints
So from:

def create_app(config_class=None):
    app = Flask(__name__)
   
    app.config.from_object(config_class)

    with app.app_context():
        from . import routes 

        return app

to using

def create_app(config_class=None):
    app = Flask(__name__)
   
    app.config.from_object(config_class)

    from app.main import bp as main_bp
    app.register_blueprint(main_bp)

    return app

For some reason the from . import routes wouldn't map the urls the second call of the app fixture. I'm pretty new to Flask so the issue here might be obvious to someone else.

[Cassius Close]

Oh yes, you did say that above, my bad. Thanks!