➜
This issue tracker has been migrated to GitHub,
and is currently read-only.
For more information,
see the GitHub FAQs in the Python's Developer Guide.
Created on 2017-06-09 15:29 by vstinner, last changed 2022-04-11 14:58 by admin. This issue is now closed.
| Pull Requests | |||
|---|---|---|---|
| URL | Status | Linked | Edit |
| PR 2129 | merged | matrixise, 2017-06-12 12:30 | |
| Messages (3) | |||
|---|---|---|---|
| msg295543 - (view) | Author: STINNER Victor (vstinner) *  |
Date: 2017-06-09 15:29 | |
haypo@selma$ ./python -m test -R 3:3 -m test_recursive_repr test_xml_etree_c Run tests sequentially 0:00:00 [1/1] test_xml_etree_c beginning 6 repetitions 123456 ...... test_xml_etree_c leaked [4, 4, 4] references, sum=12 1 test failed: test_xml_etree_c Total duration: 179 ms Tests result: FAILURE [64152 refs] |
|||
| msg295550 - (view) | Author: STINNER Victor (vstinner) * |
Date: 2017-06-09 16:21 | |
The leak is in the element_repr() function of Modules/_elementtree.c. Py_ReprEnter() creates an internal list to track objects, and so requires that Py_ReprLeave() is always called, even on error, to destroy these objects. Is it the case here? ;-) |
|||
| msg295780 - (view) | Author: STINNER Victor (vstinner) * |
Date: 2017-06-12 13:30 | |
New changeset 41af942e57f9a450cf273ef6b72a21380c6ba8d1 by Victor Stinner (Stéphane Wirtel) in branch '2.7': bpo-30615: Fix the leak reference in Modules/_elementtree.c (#2129) https://github.com/python/cpython/commit/41af942e57f9a450cf273ef6b72a21380c6ba8d1 |
|||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2022-04-11 14:58:47 | admin | set | github: 74800 |
| 2017-06-12 13:50:23 | matrixise | set | status: open -> closed stage: resolved |
| 2017-06-12 13:30:50 | vstinner | set | messages: + msg295780 |
| 2017-06-12 12:30:10 | matrixise | set | pull_requests: + pull_request2183 |
| 2017-06-09 16:21:29 | vstinner | set | keywords:
+ easy (C) messages: + msg295550 |
| 2017-06-09 15:29:04 | vstinner | create | |