jscholes, 6 days ago to python

Needed to rename a test fixture in a #Python file, and find/replace wasn't up for the job. So I decided to give #VSCode a go:

I started by pressing Ctrl+F2, for "Change All Occurrences". I think that is basically find/replace, and hence didn't do what I wanted.

Instead, I installed the recommended Python extension, and pressed F2 for "Rename Symbol". That claims to have only made one change, and the references to the function are still using the old name.

So, consider me confused. I'm using #pytest, whereby the test fixtures are referenced as function arguments rather than being called directly. Maybe that's what VS Code is struggling with? Either way, I've now spent more time on this than just manually editing the text.

#VisualStudioCode

SebastianM6L, 21 days ago to jenkins

Anyone had an issue with from django.test import Client causing a high memory usage in a docker container?

I run the tests in a pipeline using a #jenkins agent in #docker.

When this test runs:

def test_login_user(self):  
 c = Client()  
 response = c.post('/user/login/', {'email': 'test1@test.de', 'password': 'test123'})  
self.assertEqual(response.status_code, 302)  

The memory usage goes thru the roof. Literally using up all memory until it crashes.

#django #pytest

adamchainz, 24 days ago to python

A #pytest pattern I’ve recently enjoyed using:

1. Change some function in a way that needs fixes to call sites.
2. Run the whole test suite with pytest.
3. Copy the list of failing test IDs from pytest’s summary into a scratch file.
4. Fix up the found callers.
5. Verify everything fixed by saving the list of tests with: pytest $(cat scratch/testids.txt)

#Python #Testing

arildsen, 25 days ago to python

I have used numpy.allclose to test for approximate equality in Python for years, but I recently found pytest.approx better, because it lets Pytest interpret the result. For example, with numpy.allclose:

> assert np.allclose(result, 3.061, atol=1e-3, rtol=1e-3)
E assert False
E + where False = <function allclose at 0x7f8fea7efa60>(1.4872, 3.061, atol=0.001, rtol=0.001)
E + where <function allclose at 0x7f8fea7efa60> = np.allclose

#Python #pytest

SebastianM6L, 27 days ago to django

Any #django cracks here who can explain me why when I run my tests with the django built in manage.py test (runs unittest) I have 51 tests run.

When I run them with pytest It will only run 38 tests.

#python #pytest

hugovk, 1 month ago to python

Next month, @the_compiler is organising a pytest sprint in Austria, next to the Swiss border.

There's also a possibility for paid travel/accommodation.

See https://github.com/pytest-dev/sprint for more info and signup.
#Python #pytest #sprint

jtmoulia, 1 month ago to emacs

A guide on integrating #emacs + #python + #hatch using [mostly] builtin tooling and project local .dir-locals.el variables.

Specifically, this setup uses #Eglot + #pyright for live syntax checking, #pytest for code running, #MyPy for type-checking, and #IPython as the shell

Feedback would be welcome -- trying to get it robust + idiomatic.

https://jtmoulia.srht.site/guides/emacs-python-hatch/

HelenJoScott, 1 month ago to random

Our pytest video series to learn how to combine #pytest and PyCharm has started so you can deliver quality code even faster. Watch now 📺 https://jb.gg/ls23kz 📺

Let me know what you think! 😀

adamchainz, 1 month ago to python

✍️ New post covering a “diffing” unit test in #Python, which checks copy-paste-tweaked code is in sync with the upstream library.

I used this technique recently on a #Django upgrade and I’m curious if anyone else has done similar.

https://adamj.eu/tech/2024/04/26/python-diffing-unit-test/

jugmac00, 1 month ago (edited 1 month ago) to random German

One of the best things about any conference... The hallway track.

@the_compiler, @jugmac00 and @ramgarlic chatting about @brianokken's #pytest book at #pyconde in Berlin.

mgorny, 2 months ago to gentoo Polish

Ostatnio zastosowałem prostą sztuczkę w standardowej inwokacji PyTesta w #Gentoo, by rzucało błąd, kiedy wykryje nieużywane funkcje async. Miało to na celu zwiększenie szansy zauważenia, jeśli w którejś paczce przypadkiem pominięto zależność od dev-python/pytest-asyncio (albo równoważnej wtyczki), albo przy wyłączonym automatycznym ładowaniu wtyczek, pominięto ładowanie takiej wtyczki.

Dziś otrzymałem pierwsze zgłoszenie, dotyczące paczki dev-python/ipython. Przeszukałem źródła, i potwierdziłem zależność — tyle, że z jakiegoś powodu przypiętą do wersji < 0.22. No cóż, nie mamy już takiej, ale warto sprawdzić, czy mimo to nie zadziała. Tak więc dodałem zależność, dodałem -p asyncio… a #PyTest dalej jej nie widzi. Podrapałem się po głowie, spróbowałem PYTEST_PLUGINS — dalej nie działa. Co do…?

Tak więc sklonowałem repozytorium git, spróbowałem ze starszą wersją wtyczki — testy działają. Zaktualizowałem do 0.23.6, przestały działać. Sprawdziłem historię, i okazało się, że starą wersję wymuszono ze względu na błąd w wydaniu 0.22.0. Tyle że ten błąd już poprawiono, wydanie usunięto, a mój problem był zupełnie inny.

Przyjrzałem się sprawie bliżej. Z jakiegoś powodu, w kodzie #IPython testy nie są bezpośrednio oznaczone markerem pytest.mark.asyncio. Zamiast tego, conftest.py automatycznie dodaje ten marker do wszystkich testów, będących współprogramami. Ze starszymi wersjami to działało, z nową przestało. Wygląda na to, że test zostaje poprawnie oznaczony, ale wówczas przestaje być rozpoznawany jako współprogram. Przygotowałem minimalny przykład i zgłosiłem problem.

Istotne w tej historii jest to: (potencjalny) problem nie został zauważony przez dłuższą chwilę, bo autorzy IPythona przedwcześnie wymusili starą wersję ze względu na chwilowy błąd, a następnie nie usunęli tego ograniczenia.

https://gitweb.gentoo.org/repo/gentoo.git/commit/eclass/python-utils-r1.eclass?id=1f35acddca01e91d4477f3d0340c47329517f474
https://bugs.gentoo.org/928411
https://github.com/ipython/ipython/commit/e27ee203ad54df0431d817abad09ec1caafde4d6
https://github.com/pytest-dev/pytest-asyncio/issues/810

#Python

opensuse, 3 months ago to devops

Simplify your #DevOps life! Discover how pytest_container can transform container testing process in this #FOSDEM video. #openSUSE #pytest #containers https://fosdem.org/2024/schedule/event/fosdem-2024-3747-testing-containers-with-python-and-pytest/

hugovk, 3 months ago to python

@coveragepy can now use Python 3.12's new sys.monitoring module with much lower overhead.

On 3.12, it's about the same as if you were running tests without coverage enabled!

https://nedbatchelder.com/blog/202312/coveragepy_with_sysmonitoring.html

With 7.4.2, you can set COVERAGE_CORE=sysmon globally on your CI, and it'll only use it where available (Python 3.12 and 3.13 alpha), and use the default for 3.11 and older.

For example, @pillow is 9% - 27% faster!

https://github.com/python-pillow/Pillow/pull/7820

#Python #coverage #CoveragePy #pytest #testing #CI

mblayman, 3 months ago to python

🐍 This #pytest 101 talk explains what automated tests are and how to get started with pytest for your #Python projects. https://www.youtube.com/watch?v=etosV2IWBF0

SnoopJ, 3 months ago to random

#pytest is such a nice library. The pytest_addoption(parser) hook speaks argparse pretty naturally, so you can pass action=argparse.BooleanOptionalAction instead of action="store_true"

This option (new in Python 3.9) gives you a pair of feature flags (i.e. --feature/--no-feature) for free.

https://git.snoopj.dev/SnoopJ/playground/commit/7a735adc9a2131fd8a51aaa78a734ffaeb3fffab

foxmask, 3 months ago to random French

#Pytest tuto with cool features by #PyCharm

kjaymiller, 4 months ago to python

#python #pytest folks

If you're stuck pinning pytest<8.0 - There was a fork of pytest-lazy-fixtures that brings support.

https://github.com/dev-petrov/pytest-lazy-fixtures

kevinbowen, 4 months ago to python

Hurrah for #pytest getting to 8.0.0!

Boo for it looking like the #pytest-sugar plugin breaking due to "PytestRemovedIn8Warning: The (startdir: py.path.local) argument is deprecated, please use (start_path: pathlib.Path)"

https://docs.pytest.org/en/latest/deprecations.html#py-path-local-arguments-for-hooks-replaced-with-pathlib-path

I'm not confident in my competance in fixing this issue; but, I guess I will try to poke around the plugin this week.

I like those little progress histograms, darn it... #python

brianokken, 4 months ago to python

Instead of adding a timeout to pytest-repeat, I built a small plugin just for the suite-timeout functionality: pytest-suite-timeout
It aborts a test run between tests if a timeout expires
Works great with pytest-repeat, but also can be used anywhere.
https://github.com/okken/pytest-suite-timeout
#Python #pytest

brianokken, 4 months ago (edited 4 months ago) to random

--stepwise is the new -x

-x stop at the first failure
--stepwise first time, just like -x
--stepwise again, restart where we left off and continue until a failure.
#pytest

nedbat, 5 months ago to random

I think this #pytest plugin exists, but I can't find it: run just 20 randomly selected tests from my test suite.
I don't need splitting into groups, I don't need "the whole suite but in a random order". How can I run a randomly selected subset of my tests?

EngineerDaryl, 5 months ago to python

Gifted myself the book about PyTest by @brianokken (because I was not busy enough with all of my personal projects 😂 ). Hope this test doesn't fail: assert is_good_book.

#python #pytest #coding #tdd