Hi there! Just a quick question. Let's say I want my code to support Python 3.8-3.10. I set up GitHub workflow to run with these versions. What should I set in

pants.toml

to get all tools (most importantly mypy) and tests running with the specific version of Python? Is that possible? I would like to catch problems that are specific for a specific version of Python. From docs I'm not sure whether this is possible with mypy. The repo is https://github.com/5G-ERA/era-5g-client

Hey if you search for “pants python multiple version” on this channel, you see some previous discussions e.g. https://pantsbuild.slack.com/archives/C046T6T9U/p1625087433097500 https://pantsbuild.slack.com/archives/C046T6T9U/p1670362194370779

I defined multiple

python_sources

with different interpreter_constraints, which, as I believe, works fine. But I'g getting warnings like "Pants cannot safely infer a dependency because more than one target owns this module, so it is ambiguous which to use". Is it safe to ignore, or should I do it differently?

Hey as posted in the shared thread, you need to do this with the tests.

# pants-plugins/macros.py
def python_tests_global(name, **kwargs):
    kwargs.pop("interpreter_constraints", None)

    python_tests(
        name=f"{name}_py369",
        interpreter_constraints=["==3.6.9"],
        **kwargs,
    )

    python_tests(
        name=f"{name}_py385",
        interpreter_constraints=["==3.8.5"],
        **kwargs,
    )

You can first test it without macro like you have done with the

python_sources

currently and later move it to a macro

fwiw you can use

parametrize

instead of the macro https://www.pantsbuild.org/docs/targets#parametrizing-targets Snippet from the docs:

# Creates four targets:
#
#    example:tests@interpreter_constraints=py2,resolve=lock-a
#    example:tests@interpreter_constraints=py2,resolve=lock-b
#    example:tests@interpreter_constraints=py3,resolve=lock-a
#    example:tests@interpreter_constraints=py3,resolve=lock-b

python_test(
    name="tests",
    source="tests.py",
    interpreter_constraints=parametrize(py2=["==2.7.*"], py3=[">=3.6"]),
    resolve=parametrize("lock-a", "lock-b"),
)

Yeah, you probably want to parameterize here

I tried

parametrize

, and it works, but... 🙂 I probably need to specify dependencies everywhere then...

no, not dependencies, but resolve(s).. which you can do using

__defaults__

https://www.pantsbuild.org/docs/targets#field-default-values