[Python-Dev] [Python-checkins] cpython: Issue #14065: Added cyclic GC support to ET.Element (original) (raw)

Benjamin Peterson benjamin at python.org
Fri Mar 30 21:30:59 CEST 2012

2012/3/30 eli.bendersky <python-checkins at python.org>:

http://hg.python.org/cpython/rev/0ca32013d77e changeset:   75997:0ca32013d77e parent:      75995:cf2e74e0b7d4 user:        Eli Bendersky <eliben at gmail.com> date:        Fri Mar 30 16:38:33 2012 +0300 summary:  Issue #14065: Added cyclic GC support to ET.Element

files:  Lib/test/testxmletree.py |  27 ++++++++++-  Modules/elementtree.c     |  63 +++++++++++++++++++------  2 files changed, 74 insertions(+), 16 deletions(-)

diff --git a/Lib/test/testxmletree.py b/Lib/test/testxmletree.py --- a/Lib/test/testxmletree.py +++ b/Lib/test/testxmletree.py @@ -14,9 +14,10 @@  # Don't re-import "xml.etree.ElementTree" module in the docstring,  # except if the test is specific to the Python implementation. -import sys +import gc  import html  import io +import sys  import unittest  from test import support @@ -1846,6 +1847,30 @@ self.assertRaises(TypeError, e.extend, [ET.Element('bar'), 'foo']) self.assertRaises(TypeError, e.insert, 0, 'foo') +    def testcyclicgc(self): +        class ShowGC: +            def init(self, flaglist): +                self.flaglist = flaglist +            def del(self): +                self.flaglist.append(1)

I think a nicer way to check for cyclic collection is to take a weakref to an object, call the GC, then check to make sure the weakref is broken.

+ +        # Test the shortest cycle: lst->element->lst +        fl = [] +        lst = [ShowGC(fl)] +        lst.append(ET.Element('joe', attr=lst)) +        del lst +        gc.collect()

support.gc_collect() is preferable

+        self.assertEqual(fl, [1]) + +        # A longer cycle: lst->e->e2->lst +        fl = [] +        e = ET.Element('joe') +        lst = [ShowGC(fl), e] +        e2 = ET.SubElement(e, 'foo', attr=lst) +        del lst, e, e2 +        gc.collect() +        self.assertEqual(fl, [1])

-- Regards, Benjamin

More information about the Python-Dev mailing list