For people saying there's no such thing as memory leak in python

memoryleak.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

import functools

def memoize(func):
    CACHE = dict()
    @functools.wraps(func)
    def wrapper(*args, **kwargs):
        key = (tuple(args), frozenset(kwargs.items()))
        try:
            return CACHE[key]
        except KeyError:
            ret = func(*args, **kwargs)
            CACHE[key] = ret
            return ret

def fib(n):
    if n < 2:
        return 1
    return fib(n - 1) + fib(n - 2)

for i in range(50):
    print(i, ':', fib(i))


# Do some other stuff...
# See how the memory usage evolves

I found out about this issue with functools.wraps as well. Have you figured out why it causes problem? Thanks!

@xieshuaix this is not a functools.wraps issue. The issue is that CACHE is caught in the closure.

It's not about the functools.wraps but more about how it's used.