Differences between forced early binding methods (In CPython) and how to access values stored in function?

Question

Using the CPython implementation of Python 3.9.16 consider the following codes

def foo_a():
    x = 1
    fun = (lambda _x: lambda: print(_x))(x)
    x = 2
    return fun

def foo_b():
    x = 1
    fun = lambda _x=x: print(_x)
    x = 2
    return fun

fa = foo_a()
fb = foo_b()

where I've used two different methods for forcing early binding of x. I believe these are semantically equivalent, however I notice that the resulting functions have different closures.

print(fa.__closure__)    # (<cell at 0x0000025CED84BFD0: int object at 0x0000025CE9E36930>,)
print(fb.__closure__)    # None

In the case of fa I can access the local _x variable using fa.__closure__[0].cell_contents, but where is the local _x stored in fb if not inside its closure?

Answer 1

In your first example, _x is a free variable whose value is found in a non-local scope carried by the inner function via a closure. (Informally, we say the inner function closes over the variable _x in the outer function.)

In your second example, _x is a local variable assigned at call time by an argument to fun, or if none is provided, from a default value stored in the function itself.

Your use of both functions is the same, but if you received a reference to the inner function in the first example, it would be "harder" to change the value of _x (you'd have to poke around in the closure itself; closures are, for better or worse, mutable) than it would be to change the value of _x in the second example (you would just pass a different argument).