Bug Report

Mypy seems to treat int as a subtype of float even though python doesn't think so and in some cases they are not interchangeable. We hit this multiple times: with FFI, using some strict runtime type assertions, and comparing json output of "equal" inputs where the inputs differed in float-ness and therefore the strings ended up having .0 only some of the time.

While I understand why, in some less-strict codebases this may be desireable, it would be good to at least be able to opt out of this.

To Reproduce

def foo(x: float) -> None:
    assert isinstance(x, float)

foo(10) 

Expected Behavior

mypy rejects the call to foo in the example

Actual Behavior

mypy accepts the call to foo in the example, but the program fails at runtime.

Your Environment

• Mypy version used: mypy==0.910
• Mypy command-line flags: mypy example.py
• Mypy configuration options from mypy.ini (and other config files):

[mypy]
exclude=.*_pb2[.]py
check_untyped_defs=True
disallow_any_generics=True
disallow_incomplete_defs=True
disallow_subclassing_any=True
disallow_untyped_defs=True
ignore_missing_imports=True
no_implicit_optional=True
strict_optional=True
warn_return_any=True
warn_unused_ignores=True
plugins = pydantic.mypy

[pydantic-mypy]
init_typed = True

• Python version used: 3.8.10
• Operating system and version: Ubuntu 20.04