Question

Why/how does this create a seemingly infinite loop? Incorrectly, I assumed this would cause some form of a stack overflow type error.

i = 0

def foo () :
    global i

    i += 1
    try :
        foo()
    except RuntimeError :
        # This call recursively goes off toward infinity, apparently.
        foo()

foo()

print i

Answer 1

If you change the code to

i = 0
def foo ():
    global i
    i += 1
    print i
    try :
        foo()
    except RuntimeError :
        # This call recursively goes off toward infinity, apparently.
        foo()
    finally:
        i -= 1
        print i

foo()

You’ll observe that the output oscillates short below 999 (1000 being Python’s default recursion limit). That means, when the limit is hit (RuntimeError) that last call of foo() is terminated, and another one is set off to replace it immediately.

If you raise a KeyboardInterrupt you’ll observe how the entire trace is being terminated at once.

---

UPDATE

Interestingly the second call of foo() is not protected by the try ... except-block anymore. Therefore the application will in fact terminate eventually. This becomes apparant if you set the recursion limit to a smaller number, e.g. the output for sys.setrecursionlimit(3):

$ python test.py
1
2
1
2
1
0
Traceback (most recent call last):
  File "test.py", line 19, in <module>
    foo()
  File "test.py", line 14, in foo
    foo()
  File "test.py", line 14, in foo
    foo()
RuntimeError