Why is python collecting even if gc disabled?

Question

This code

import gc
gc.disable()
print gc.isenabled()
print len(gc.get_objects())
class Foo(): pass
print len(gc.get_objects())

a=[]
print len(gc.get_objects())
for i in xrange(10000):
    a.append(Foo())

print len(gc.get_objects())
a=[] # Should not collect
print len(gc.get_objects())
gc.collect()
print len(gc.get_objects())

Produces this output

I would expect the second to last number to be 13756, because the gc is disabled and when the original a list goes out of scope, it should not drop those objects. Apparently python is collecting those objects anyway. Why ?

python is 2.7.2

Martijn Pieters · Accepted Answer · 2013-07-24 11:44:45Z

4

Because garbage collection only deals with circular references. Quoting from the gc module documention:

Since the collector supplements the reference counting already used in Python, you can disable the collector if you are sure your program does not create reference cycles.

CPython uses reference counting to clean up most objects. When you replace the list a referred to with an empty list, the reference counts for the old list dropped to 0, so it is deleted at that point. Deleting the list causes the reference counts for the Foo() instances to drop to 0 too, etc.

If you create circular references instead, and disable garbage collection, you'll see the numbers remain constant:

class Foo(object):
    def __init__(self):
         self.bar = Bar(self)

class Bar(object):
    def __init__(self, foo):
        self.foo = foo

edited Jul 24, 2013 at 11:44

answered Jul 24, 2013 at 11:28

Martijn Pieters

1.1m325 gold badges4.2k silver badges3.4k bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Stefano Borini Over a year ago

How can I get information of the unreachable objects and when they are collected?

Martijn Pieters Over a year ago

gc.garbage gives you a list of objects that could not be collected; there is no API to list anything that is unreachable and can be collected, because that would require an explicit collection run, which then breaks the cycles at that time. You can set the thresholds to influence when a cycle will be run, but you cannot query when one will run because it is not time-based.

Collectives™ on Stack Overflow

Why is python collecting even if gc disabled?

1 Answer 1

2 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Related