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.

Title: assertRaises increases reference counter
Type: behavior Stage: resolved
Components: Library (Lib) Versions: Python 3.7, Python 3.6, Python 3.5
Status: closed Resolution: fixed
Dependencies: Superseder:
Assigned To: Nosy List: Vjacheslav.Fyodorov, docs@python, ezio.melotti, michael.foord, ncoghlan, r.david.murray, rbcollins, subho, vstinner
Priority: normal Keywords:

Created on 2015-04-08 18:37 by Vjacheslav.Fyodorov, last changed 2022-04-11 14:58 by admin. This issue is now closed.

File name Uploaded Description Edit Vjacheslav.Fyodorov, 2015-04-08 18:37 unit test demo for assertRaises
Pull Requests
URL Status Linked Edit
PR 193 merged vstinner, 2017-02-20 10:45
PR 858 merged vstinner, 2017-03-27 22:57
PR 2228 merged vstinner, 2017-06-15 22:33
Messages (12)
msg240280 - (view) Author: Vjacheslav Fyodorov (Vjacheslav.Fyodorov) Date: 2015-04-08 18:37
Sometimes unittest's assertRaises increases reference counter of callable. This can break tests in tricky cases. Not seen in 2.X version. Demo file attached.
msg240293 - (view) Author: R. David Murray (r.david.murray) * (Python committer) Date: 2015-04-09 00:18
This is presumably a result of the exception object being saved on the context manager object and there being a circular reference chain that doesn't immediately get GCed.  This is just how python works.

I don't know if it would be sensible/useful to use the new lightweight traceback stuff in assertRaises.  If so it would probably have to be optional for backward compatibility reasons.
msg240309 - (view) Author: Vjacheslav Fyodorov (Vjacheslav.Fyodorov) Date: 2015-04-09 07:09
It seems, as a minimum it must be noticed in docs.
msg261720 - (view) Author: Robert Collins (rbcollins) * (Python committer) Date: 2016-03-14 03:10
I don't think we make any guarantees for or against references, but - we attempt to free references already (see how we call clear_frames), so this is a bug, pure and simple.
msg288076 - (view) Author: Nick Coghlan (ncoghlan) * (Python committer) Date: 2017-02-18 12:46
As Robert noted, this is a straight up error, where the clear_frames() call and deleting the exception traceback from the saved object aren't enough to drop all references to `callable_obj`.

The trick is that the cleanup code isn't accounting for __cause__ and __context__: it's only clearing and dropping the traceback for the topmost exception in the chain.

In Vjacheslav's example, there are *two* tracebacks to worry about: both the top level one (which is getting cleaned up) and the inner one from exc.__context__ which isn't being touched.

We could likely use a "traceback.clear_all_frames()" helper that walks an exception tree clearing *all* the traceback frames, both on the original exception, and on all the __cause__ and __context__ attributes.

(We can't just clear __cause__ and __context__, since those can be accessed and tested when using the context manager form and the `exception` attribute)
msg288108 - (view) Author: Nick Coghlan (ncoghlan) * (Python committer) Date: 2017-02-19 05:55
I've been looking into this further, and a reproducer that's independent of assertRaises() is to just bind the function to a local variable name in an outer function:

    def outer():
        callable_obj = f
        except Exception as exc:
            return exc

That should make it easier to test a basic recursive "clear_all_frames" operation like:

    def clear_all_frames(exc):
        cause = exc.__cause__
        if cause is not None:
        context = exc.__context__
        if context is not None:
msg288449 - (view) Author: Nick Coghlan (ncoghlan) * (Python committer) Date: 2017-02-23 15:20
Victor's PR takes a more pragmatic approach to address this problem: rather than adding the ability to clear the frames for an arbitrary exception tree, it just uses a try/finally in a couple of key unittest function definitions to clear the offending circular references by setting them to None.
msg290664 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-03-27 22:56
New changeset bbd3cf8f1ef1e91a8d6dac6411e18b4b9084abf5 by Victor Stinner in branch 'master':
Fix ref cycles in TestCase.assertRaises() (#193)
msg290666 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-03-27 23:04
It seems like Python 2.7 is not affected.
msg296132 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-06-15 22:18
New changeset 50dbf577e10f806056d60ac956db0748d2cc8257 by Victor Stinner in branch '3.6':
bpo-23890: Fix ref cycle in TestCase.assertRaises (#858)
msg296135 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-06-15 22:51
New changeset 3dc573c8d19dc42ed786ca3237afdad183c41ca0 by Victor Stinner in branch '3.5':
Fix ref cycles in TestCase.assertRaises() (#193) (#2228)
msg296136 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-06-15 22:52
Thank you Vjacheslav Fyodorov for your bug report! The bug should now be fixed in 3.5, 3.6 and master (future 3.7) branches.

Python 2.7 is not affected by the bug.
Date User Action Args
2022-04-11 14:58:15adminsetgithub: 68078
2017-06-15 22:52:29vstinnersetstatus: open -> closed
resolution: fixed
stage: commit review -> resolved
2017-06-15 22:52:21vstinnersetmessages: + msg296136
2017-06-15 22:51:26vstinnersetmessages: + msg296135
2017-06-15 22:33:48vstinnersetpull_requests: + pull_request2273
2017-06-15 22:18:17vstinnersetmessages: + msg296132
2017-03-27 23:04:57vstinnersetmessages: + msg290666
2017-03-27 22:57:43vstinnersetpull_requests: + pull_request761
2017-03-27 22:56:31vstinnersetmessages: + msg290664
2017-02-23 15:20:20ncoghlansetmessages: + msg288449
2017-02-23 15:15:29ncoghlansetstage: needs patch -> commit review
2017-02-20 10:45:50vstinnersetpull_requests: + pull_request160
2017-02-19 05:55:16ncoghlansetmessages: + msg288108
2017-02-18 12:46:34ncoghlansetassignee: docs@python ->
type: enhancement -> behavior
components: - Documentation
versions: + Python 3.7
nosy: + ncoghlan

messages: + msg288076
2017-02-18 12:41:21subhosetnosy: + subho
2016-03-17 17:04:01vstinnersetnosy: + vstinner
2016-03-14 03:10:29rbcollinssetmessages: + msg261720
2016-01-03 06:35:40ezio.melottisetversions: + Python 3.6
nosy: + docs@python

assignee: docs@python
components: + Documentation
stage: needs patch
2015-04-09 07:17:02ezio.melottisetnosy: + rbcollins, ezio.melotti, michael.foord
2015-04-09 07:09:47Vjacheslav.Fyodorovsetmessages: + msg240309
2015-04-09 00:18:42r.david.murraysetversions: + Python 3.5, - Python 3.4
nosy: + r.david.murray

messages: + msg240293

type: behavior -> enhancement
2015-04-08 18:37:05Vjacheslav.Fyodorovcreate