“Cannot infer type of lambda” using default parameter to recapture loop variable

Bug Report

One of the most common gotchas in Python is that a loop reassigns its iteration variable rather than creating a new binding for each iteration, and a lambda closure created in the loop observes this reassigned value rather than the value from each iteration. A typical workaround is to add a “redundant” default parameter to the lambda to recapture the value from each iteration—for example, the i=i from the above link (reproduced below). But mypy rejects such a lambda with error: Cannot infer type of lambda [misc].

To Reproduce

from typing import Callable

def create_multipliers() -> list[Callable[[int], int]]:
    return [lambda x, i=i: i * x for i in range(5)]

(mypy playground)

Expected Behavior

mypy ought to be able to realize that since the lambda is coerced to Callable[[int], int], its second parameter i=i can only ever receive the default value, so its type should be successfully inferred as (x: int, i: int = ...) -> int.

Actual Behavior

main.py:4: error: Cannot infer type of lambda  [misc]
Found 1 error in 1 file (checked 1 source file)

Your Environment

• Mypy version used: 1.3.0
• Mypy command-line flags: none
• Mypy configuration options from mypy.ini (and other config files): none
• Python version used: 3.11.3

Start by running the main.py reproduction with mypy and reviewing how lambda parameter and default-expression types are inferred for the Callable[[int], int] context. Trace the relevant type-inference entry point and add a regression test for the example; done means mypy accepts the code without Cannot infer type of lambda while preserving the inferred callable signature.