Possible reference leak in error paths of update_bases() and __build_class__

[pablogsal]

BPO	44856
Nosy	@ambv, @pablogsal, @miss-islington
PRs	bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ #27647 [3.9] bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ (GH-27647) #27651 [3.8] bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ (GH-27647) #27652 [3.10] bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ (GH-27647) #27653

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2021-08-07.14:12:16.906>
created_at = <Date 2021-08-07.00:25:24.565>
labels = ['3.8', '3.9', '3.10', '3.11']
title = 'Possible reference leak in error paths of update_bases() and __build_class__'
updated_at = <Date 2021-08-07.14:12:16.905>
user = 'https://github.com/pablogsal'

bugs.python.org fields:

activity = <Date 2021-08-07.14:12:16.905>
actor = 'lukasz.langa'
assignee = 'none'
closed = True
closed_date = <Date 2021-08-07.14:12:16.906>
closer = 'lukasz.langa'
components = []
creation = <Date 2021-08-07.00:25:24.565>
creator = 'pablogsal'
dependencies = []
files = []
hgrepos = []
issue_num = 44856
keywords = ['patch']
message_count = 7.0
messages = ['399161', '399163', '399172', '399174', '399175', '399180', '399182']
nosy_count = 3.0
nosy_names = ['lukasz.langa', 'pablogsal', 'miss-islington']
pr_nums = ['27647', '27651', '27652', '27653']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue44856'
versions = ['Python 3.8', 'Python 3.9', 'Python 3.10', 'Python 3.11']

[pablogsal]

Here:

cpython/Python/bltinmodule.c

Lines 60 to 88 in 17c2316

	new_base = PyObject_CallOneArg(meth, bases);
	Py_DECREF(meth);
	if (!new_base) {
	goto error;
	}
	if (!PyTuple_Check(new_base)) {
	PyErr_SetString(PyExc_TypeError,
	"__mro_entries__ must return a tuple");
	Py_DECREF(new_base);
	goto error;
	}
	if (!new_bases) {
	/* If this is a first successful replacement, create new_bases list and
	copy previously encountered bases. */
	if (!(new_bases = PyList_New(i))) {
	goto error;
	}
	for (j = 0; j < i; j++) {
	base = args[j];
	PyList_SET_ITEM(new_bases, j, base);
	Py_INCREF(base);
	}
	}
	j = PyList_GET_SIZE(new_bases);
	if (PyList_SetSlice(new_bases, j, j, new_base) < 0) {
	goto error;
	}
	Py_DECREF(new_base);
	}

Seems that new_base is not properly cleaned on the error paths

[pablogsal]

Oh, damn, turns out that there is a ton of reference leaks in the error path of __build_class__ as well.

These leaks have been since forever!

Curiously I found about them when debugging https://bugs.python.org/issue44524

[ambv]

New changeset a40675c by Pablo Galindo Salgado in branch 'main':
bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ (GH-27647)
a40675c

[ambv]

New changeset 0a42309 by Miss Islington (bot) in branch '3.8':
bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ (GH-27647) (GH-27652)
0a42309

[ambv]

New changeset ed718e9 by Miss Islington (bot) in branch '3.9':
bpo-44856: Possible reference leak in error paths of update_bases() and __build_class__ (GH-27647) (GH-27651)
ed718e9