-
Notifications
You must be signed in to change notification settings - Fork 276
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BUG: segfault on macOS (amr64) #4953
Comments
Yeah, I'm seeing a segfault in that log. I guess it could be some cython code issue, but if it's only showing up on this one mac/python combo, that does kinda smell like some other issue. I'm honestly getting pretty annoyed with code dev on my mac |
Wait do you see this segfault locally too ? |
Not personally, but I haven't been able to run anything with openMP (at least through python) on my mac. That's the annoyance I was thinking of. I'm using python 3.11 on my laptop, so at least a 2021 14-inch M1 Mac + python3.11 with no attempt to use multi-threading doesn't seem to have this issue. |
I might be getting a similar issue in this test for PR #4939, which fails specifically on python3.12 with macOS. full copy of the error report in case new tests invalidate the link:
|
Huh, I just got another error on a macOS-lastest, python 3.12 test, which occurred during the exact same test ( |
It also looks like test_variable_dx is the culprit in Furthermore, this time it happened on Python 3.10 so it's not 3.9-specific. |
Good news, I actually tried it and I can reproduce this locally with
More notes:
|
marking this as a blocker since we basically cannot release until this is adressed |
progress: I bisected it down to numpy/numpy#26821 (which is a backport of numpy/numpy#26797) |
reported upstream as numpy/numpy#27037 |
And patched upstream numpy/numpy#27070 |
Bug report
For visibility, I think this segfault seen in CI (example logs) is recurrent but not deterministic. It may be an upstream bug and I'm not even sure that it's platform-dependent or Python-version-dependent, so I'm reporting this occurrence and will start paying attention next time around so we can gather some more clues. If it turns out it's really just about Python 3.9 it may not be worth trying to fix it since it might soon be time for us to drop support for it completely.
The text was updated successfully, but these errors were encountered: