The vast majority of Python memory management is handled with reference counting.
Every time an object is referenced (e.g. assigned to a variable), its reference count is automatically increased. When it is dereferenced (e.g. variable goes out of scope), its reference count is automatically decreased.
When the reference count reaches zero, the object is immediately destroyed and the memory is immediately freed. Thus for the majority of cases, the garbage collector is not even needed.
>>> import gc; gc.disable() # disable garbage collector
>>> class Track:
def __init__(self):
print("Initialized")
def __del__(self):
print("Destructed")
>>> def foo():
Track()
# destructed immediately since no longer has any references
print("---")
t = Track()
# variable is referenced, so it's not destructed yet
print("---")
# variable is destructed when function exits
>>> foo()
Initialized
Destructed
---
Initialized
---
Destructed
To demonstrate further the concept of references:
>>> def bar():
return Track()
>>> t = bar()
Initialized
>>> another_t = t # assign another reference
>>> print("...")
...
>>> t = None # not destructed yet - another_t still refers to it
>>> another_t = None # final reference gone, object is destructed
Destructed