Use sys.executable rather than "python" to invoke pytest

[musicinmybrain]

This supports environments where the Python interpreter is not named "python". For example, in Fedora it is "python3" unless the python-unversioned-command package is installed.

[15r10nk]

Thank you. The reason for the build failures is probably on my side. I had the brilliant/stupid idea to use UV inside pytest to test different dependency versions for specific tests. I think I will revert this change because it causes more problems than it solves.

[musicinmybrain]

Ok! Thanks for following up. For now, this PR is working for me in Fedora.

As the primary maintainer of the python-inline-snapshot package in Fedora, please let me know if there’s anything specific (testing, commentary on distribution packaging practices, etc.) that you need from me.

[15r10nk]

I reverted my changes in #168, your patch works now also for me.

This is my first project which gets packaged. Let me know if I can do something which makes it easier for people to add inline-snapshot packages into distributions.

Are there any common guidelines which I can follow?

[musicinmybrain]

Overall, things are pretty good with respect to downstream packaging!

The migration to dependency groups for “dev” dependencies is actually really nice, because – while I do have to filter out coverage-analysis and linting dependencies – this puts all the testing dependencies in one place that I can use to generate RPM BuildRequires for running the package’s tests.

It would be even nicer if there were one place I could get a list of all test dependencies without linting and coverage dependencies, but the situation is already much better than it was when the complete list of test dependencies was scattered around various hatch-specific configurations.

In case you’re curious, python-inline-snapshot is packaged in Fedora as a test dependency for python-fastapi, python-pydantic-core, and python-rich-toolkit. I expect it will also soon find its way into EPEL10 along with python-pydantic/python-pydantic-core.