How to make hatch test install the whl instead of directly using it?

1 day ago 5

ARTICLE AD BOX

With the AI answers, I was recommended to do:

[tool.hatch.envs.hatch-test] dev-mode = false # This is the line recommended by AI to have it installed instead of having it editable/local use mode dependencies = [ "pytest" ]

But when I try it, hatch test

I still get

=========================================== FAILURES =========================================== _________________________________________ test_binary __________________________________________ def test_binary() -> str | None: if True: root = files('del3') children = [f for f in get_files_recursively(root)] > raise AssertionError(f"{children}") E AssertionError: [WindowsPath('E:/SourceCode/AndroidStudioProjects/Base/del3/src/del3/__about__.py'), WindowsPath('E:/SourceCode/AndroidStudioProjects/Base/del3/src/del3/__init__.py'), WindowsPath('E:/SourceCode/AndroidStudioProjects/Base/del3/src/del3/__pycache__/__init__.cpython-312.pyc')]

As you can see, it's still using the local files instead of installing the whl then use from env.

What's the correct way to handle it? [install my package as whl, not directly use it] Is hatch really the best tool? Or do other better/easier build systems exist?

My goal is to bundle a binary and then on test install the whl to confirm its working. I like the hatch build hook, but the test is unsatisfactory as it tries to use the files directly instead of installing the whl then using it.