Configuration of a custom test runner?

[sumpfralle]

The hatch documentation describes how to configure custom commands for using an alternative test runner (e.g. unittest).

But the hatch test command seems to be tightly coupled with pytest, due to hatch applying pytest-specific arguments (e.g. -p no:randomly). This breaks custom runners, since they do not understand these arguments.

Currently I work around this issue by adding ramdonize = true to my pyproject.toml. This prevents hatch from adding the above argument (which is understood only by pytest). But oviously this is just a fragile hack.

I would appreciate, if I could disable any pytest-specific argument injection via configuration.

Do you think, that it is within the scope of hatch to configure a custom test runner?
(similar to the way custom formatters and linters can be configured within hatch)

Do you have an idea, how you would implement this? (maybe I could provide a PR)

Thanks for your time!