python
Calling float.__add__ with an integer as first argument is valid
Calling float.__add__ with an integer as first argument is valid.
Example
float.__add__(1, 2)
Expected Behavior
Considering that the code fails at runtime, mypy should not accept this code.
Environment
- Mypy version used: 1.9
- Python version used: 3.11.2
I would like to work on this with a group for a college software engineering class.
This is probably accepted because of the special case where int is treated like a subtype of float (https://typing.readthedocs.io/en/latest/spec/special-types.html#special-cases-for-float-and-complex).
There are ways to implement this special case that would not have this issue, but such changes may have complicated consequences.
I understand https://typing.readthedocs.io/en/latest/spec/special-types.html#special-cases-for-float-and-complex) is useful for legacy code, but I think there should be a "strict" flag to disable it. In my case, since we interact with another language (julia), we need to control the exact types.
A case where this "allowance" is weird is when storing the value. Mypy does not report any error, but the dataclass field does not contain the declared type.
from dataclasses import dataclass
@dataclass
class C:
f: float = 0
print(type(C().f))
c = C(f=4)
print(type(c.f))
python prints this
<class 'int'>
<class 'int'>
