ipaddress: lru_cache on instance methods causes memory leak (use cached_property instead)
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 77.2k
- Forks
- 35.9k
- PR merge metrics
- PR metrics pending
Description
Bug report
Bug description:
Lib/ipaddress.py uses @functools.lru_cache() on four instance method properties:
IPv4Address.is_private(line 1321)IPv4Address.is_global(line 1343)IPv4Network.is_private(line 1552)IPv6Address.is_private(line 2089)
Using lru_cache on instance methods causes a memory leak: the global cache holds
a reference to self, preventing garbage collection even after all other references
to the instance are dropped.
This can be demonstrated with:
import ipaddress, gc, weakref
a = ipaddress.IPv4Address('192.168.1.1')
ref = weakref.ref(a)
a.is_private # trigger the cache
del a
gc.collect()
print(ref() is not None) # True — instance is NOT collected (leak!)
The same file already uses @functools.cached_property for other properties
(e.g. with_prefixlen, compressed), which correctly ties cache lifetime to
the instance. Those do not leak:
n = ipaddress.IPv4Network('192.168.1.0/24')
ref = weakref.ref(n)
_ = n.with_prefixlen
del n
gc.collect()
print(ref() is not None) # False — correctly collected
Fix: Replace @functools.lru_cache() with @functools.cached_property on
the four affected properties.
I would like to work on a fix for this issue.
CPython versions tested on:
CPython main branch
Operating systems tested on:
Linux
Linked PRs
- gh-152790
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Read the four affected properties in Lib/ipaddress.py and compare them with the existing cached_property uses in that file. Confirm the change against the weakref and garbage-collection examples so the four address and network instances can be collected after their references are dropped.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- backend
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 35/100