Documented mechanism for calling black from Python code #425
Description
I tried to write a unit test that uses "black --check" to confirm that the other code in the repository is black-approved.
I couldn't see an easy way to do this by importing black directly, so instead I used the click.CliRunner class to programatically execute it using CLI arguments. Here's my implementation of that:
import black
from click.testing import CliRunner
from pathlib import Path
code_root = Path(__file__).parent.parent
def test_black():
runner = CliRunner()
result = runner.invoke(
black.main,
[str(code_root / "tests"), str(code_root / "sqlite_utils"), "--check"],
)
assert result.exit_code == 0, result.output
This works! Or rather, it works under Python 3.6. Python 3.7 in Travis is giving me this surprising result:
__________________________________ test_black __________________________________
def test_black():
runner = CliRunner()
result = runner.invoke(
black.main,
[str(code_root / "tests"), str(code_root / "sqlite_utils"), "--check"],
)
> assert result.exit_code == 0, result.output
E AssertionError:
E assert -1 == 0
E + where -1 = <Result KeyError(<Task finished coro=<BaseEventLoop.run_in_executor() done, defined at /opt/python/3.7-dev/lib/python3.7/asyncio/base_events.py:653> result=False>)>.exit_code
I'm still trying to figure out what went wrong there.
At any rate though: it would be really useful if there was an easy, documented mechanism for programatically invoking black.