Landmines hidden in Python class variables

https://docs.python.org/3/tutorial/classes.html#class-and-instance-variables

Class variables Python are strange. If they are mutable, they act like STATIC PROPERTY in PHP; If they are immutable, when they reassigned, the act like DYNAMIC PROPERTY in PHP. It seems to do with reference.

class X:
    array: list = []
    immutable_array: tuple = ()
    static_v: str = 'ori'

    def __init__(self, name: str):
        self.name = name

    def init_ext(self):
        self.array = []  
        # or even use [self.name], just let it be another id

    def work(self):
        self.array.append(self.name)
        self.immutable_array = (len(self.array),)
        self.static_v = f"[{self.name}]"


if __name__ == '__main__':
    x1 = X("A")
    x2 = X("B")

    for i in range(3):
        print("x1 array: ", x1.array, 'tuple: ', x1.immutable_array, "v: ", x1.static_v),
        print("x1[id] array: ", id(x1.array), 'tuple: ', id(x1.immutable_array), "v: ", id(x1.static_v))
        print("x2 array: ", x2.array, 'tuple: ', x2.immutable_array, "v: ", x2.static_v)
        print("x2[id] array: ", id(x2.array), 'tuple: ', id(x2.immutable_array), "v: ", id(x2.static_v))
        x1.work()
        x2.work()

    print("finally")
    print("x1 array: ", x1.array, 'tuple: ', x1.immutable_array, "v: ", x1.static_v),
    print("x1[id] array: ", id(x1.array), 'tuple: ', id(x1.immutable_array), "v: ", id(x1.static_v))
    print("x2 array: ", x2.array, 'tuple: ', x2.immutable_array, "v: ", x2.static_v)
    print("x2[id] array: ", id(x2.array), 'tuple: ', id(x2.immutable_array), "v: ", id(x2.static_v))

Output:

x1 array:  [] tuple:  () v:  ori
x1[id] array:  4477451776 tuple:  4475306048 v:  4477483952
x2 array:  [] tuple:  () v:  ori
x2[id] array:  4477451776 tuple:  4475306048 v:  4477483952
x1 array:  ['A', 'B'] tuple:  (1,) v:  [A]
x1[id] array:  4477451776 tuple:  4476906464 v:  4477621936
x2 array:  ['A', 'B'] tuple:  (2,) v:  [B]
x2[id] array:  4477451776 tuple:  4476375872 v:  4477622000
x1 array:  ['A', 'B', 'A', 'B'] tuple:  (3,) v:  [A]
x1[id] array:  4477451776 tuple:  4476751488 v:  4477622064
x2 array:  ['A', 'B', 'A', 'B'] tuple:  (4,) v:  [B]
x2[id] array:  4477451776 tuple:  4476906464 v:  4477621936
finally
x1 array:  ['A', 'B', 'A', 'B', 'A', 'B'] tuple:  (5,) v:  [A]
x1[id] array:  4477451776 tuple:  4476375872 v:  4477622128
x2 array:  ['A', 'B', 'A', 'B', 'A', 'B'] tuple:  (6,) v:  [B]
x2[id] array:  4477451776 tuple:  4476751488 v:  4477622064

If you execute init_ext after initialization, the result changed.

x1 = X("A")
x2 = X("B")

x1.init_ext()
x2.init_ext()

# the same test codes as above

The output would be :

x1 array:  [] tuple:  () v:  ori
x1[id] array:  4327900992 tuple:  4325756992 v:  4327930800
x2 array:  [] tuple:  () v:  ori
x2[id] array:  4327899904 tuple:  4325756992 v:  4327930800
x1 array:  ['A'] tuple:  (1,) v:  [A]
x1[id] array:  4327900992 tuple:  4327849648 v:  4328077168
x2 array:  ['B'] tuple:  (1,) v:  [B]
x2[id] array:  4327899904 tuple:  4327372112 v:  4328077232
x1 array:  ['A', 'A'] tuple:  (2,) v:  [A]
x1[id] array:  4327900992 tuple:  4326822720 v:  4328077296
x2 array:  ['B', 'B'] tuple:  (2,) v:  [B]
x2[id] array:  4327899904 tuple:  4327849648 v:  4328077168
finally
x1 array:  ['A', 'A', 'A'] tuple:  (3,) v:  [A]
x1[id] array:  4327900992 tuple:  4327372112 v:  4328077232
x2 array:  ['B', 'B', 'B'] tuple:  (3,) v:  [B]
x2[id] array:  4327899904 tuple:  4326822720 v:  4328077296