classification
Title: test_stack_overflow (test.test_faulthandler.FaultHandlerTests) is stuck with GCC10
Type: behavior Stage: resolved
Components: Tests Versions: Python 3.9, Python 3.8, Python 3.7
process
Status: closed Resolution: fixed
Dependencies: Superseder:
Assigned To: Nosy List: Martin Liška, cstratak, miss-islington, vstinner
Priority: normal Keywords: patch

Created on 2019-12-04 08:33 by Martin Liška, last changed 2019-12-06 22:48 by vstinner. This issue is now closed.

Pull Requests
URL Status Linked Edit
PR 17462 closed Martin Liška, 2019-12-04 08:35
PR 17467 merged vstinner, 2019-12-04 20:05
PR 17468 merged miss-islington, 2019-12-04 20:10
PR 17469 merged miss-islington, 2019-12-04 20:11
Messages (8)
msg357778 - (view) Author: Martin Liška (Martin Liška) * Date: 2019-12-04 08:33
The test-case is stuck after update to GCC 10.
I've got a patch for that.
msg357824 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2019-12-04 20:10
New changeset 8b787964e0a647caa0558b7c29ae501470d727d9 by Victor Stinner in branch 'master':
bpo-38965: Fix faulthandler._stack_overflow() on GCC 10 (GH-17467)
https://github.com/python/cpython/commit/8b787964e0a647caa0558b7c29ae501470d727d9
msg357825 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2019-12-04 20:14
Thanks Martin Liška for the bug report and your PR 17462!

I chose to use volatile instead, since it's already the second time that we have to fix this function for a specific compiler. You abandonned your PR and confrmed that my PR fix your issue with GCC 10, so I merged my PR. Thanks for testing ;-)

I backported the fix to 3.7 and 3.8 branches.
msg357826 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2019-12-04 20:14
The cool thing is that GCC now also implements tail call optimization!
msg357827 - (view) Author: miss-islington (miss-islington) Date: 2019-12-04 20:29
New changeset 5044c889dfced2f43e2cccb673d889a4882f6b3b by Miss Islington (bot) in branch '3.7':
bpo-38965: Fix faulthandler._stack_overflow() on GCC 10 (GH-17467)
https://github.com/python/cpython/commit/5044c889dfced2f43e2cccb673d889a4882f6b3b
msg357828 - (view) Author: miss-islington (miss-islington) Date: 2019-12-04 20:30
New changeset f4a21d3b239bf4f4e4e2a8a5936b9b040645b246 by Miss Islington (bot) in branch '3.8':
bpo-38965: Fix faulthandler._stack_overflow() on GCC 10 (GH-17467)
https://github.com/python/cpython/commit/f4a21d3b239bf4f4e4e2a8a5936b9b040645b246
msg357943 - (view) Author: Charalampos Stratakis (cstratak) * Date: 2019-12-06 20:26
Would it make sense to also backport this fix to the 3.6 release?

People compiling 3.6 with gcc 10 will stumble upon that.
msg357950 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2019-12-06 22:48
> Would it make sense to also backport this fix to the 3.6 release?

Sadly, 3.6 no longer accept bugfixes:
https://devguide.python.org/#status-of-python-branches
History
Date User Action Args
2019-12-06 22:48:54vstinnersetmessages: + msg357950
2019-12-06 20:26:23cstrataksetnosy: + cstratak
messages: + msg357943
2019-12-04 20:30:35miss-islingtonsetmessages: + msg357828
2019-12-04 20:29:30miss-islingtonsetnosy: + miss-islington
messages: + msg357827
2019-12-04 20:14:43vstinnersetmessages: + msg357826
2019-12-04 20:14:09vstinnersetstatus: open -> closed
versions: + Python 3.7, Python 3.8, Python 3.9
messages: + msg357825

resolution: fixed
stage: patch review -> resolved
2019-12-04 20:11:54miss-islingtonsetpull_requests: + pull_request16948
2019-12-04 20:10:27miss-islingtonsetpull_requests: + pull_request16947
2019-12-04 20:10:27vstinnersetmessages: + msg357824
2019-12-04 20:05:28vstinnersetpull_requests: + pull_request16946
2019-12-04 09:18:26xtreaksetnosy: + vstinner
type: behavior
components: + Tests
2019-12-04 08:35:26Martin Liškasetkeywords: + patch
stage: patch review
pull_requests: + pull_request16942
2019-12-04 08:33:20Martin Liškacreate