I don't understand the test here. I think that the test should be:

def test_is_in_bool() -> None:
    vals = [True, None]
    df = pl.DataFrame({"A": [True, False, None]})
    assert df.select(pl.col("A").is_in(vals)).to_dict(False) == {
        "A": [True, None] # No FALSE here
    }

Or I am missing wormething ?

There should be a False as False is not in vals.

I don't understand, if I do .is_in([True, None]), it should be the same as .eq(True) | .is_null(), but I don't want False here

The second value is False, which is not True and is not null, so the the result is False?

Note that I am not filtering.

Oh I did not see that you changed the filter by select !

Also fixes #9247.