#64569 closed defect (fixed)
pyXY-spyder @5.2.2: generates segmentation fault on exit
Reported by: | tgederberg | Owned by: | reneeotten (Renee Otten) |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | ports | Version: | 2.7.1 |
Keywords: | Cc: | ||
Port: | py-spyder |
Description (last modified by reneeotten (Renee Otten))
I am running py39-spyder (version 5.2.2_0) and whenever I exit spyder (via the Python -> Quit Python menu), I get a segmentation fault. Attached is the error log.
Attachments (2)
Change History (10)
Changed 3 years ago by tgederberg
Attachment: | py39-spyder-segmentation-fault.txt added |
---|
comment:1 Changed 3 years ago by jmroot (Joshua Root)
Cc: | reneeotten@… removed |
---|---|
Owner: | set to reneeotten |
Port: | py-spyder added; py39-spyder removed |
Status: | new → assigned |
comment:2 Changed 3 years ago by reneeotten (Renee Otten)
comment:3 follow-up: 6 Changed 3 years ago by Gandoon (Erik Hedlund)
This is not contained to Python 3.9, it happens with the most recent Spyder under 3.10 as well.
It should be noted that a complaint that is common to both versions is: fromIccProfile: failed minimal tag size sanity
on startup.
I enclose the segfault log for the Python 3.10 Spyder dying gracelessly on MacOS 10.15 (could it have to do with older systems? May be worth looking into).
Changed 3 years ago by Gandoon (Erik Hedlund)
Attachment: | segfault_spyder_python310.log added |
---|
Spyder segfaulting under Python 3.10 on MacOS 10.15
comment:4 Changed 3 years ago by tgederberg
I am running latest version of MacOS (Monterey version 12.2) so I don't think it is due to running on a older system.
comment:5 Changed 3 years ago by reneeotten (Renee Otten)
Description: | modified (diff) |
---|---|
Summary: | py39-spyder generates segmentation fault on exit → pyXY-spyder @5.2.2: generates segmentation fault on exit |
comment:6 Changed 3 years ago by reneeotten (Renee Otten)
Replying to Gandoon:
This is not contained to Python 3.9, it happens with the most recent Spyder under 3.10 as well.
Yes, that's not really surprising as it will install the same version of packages for both Python versions. It shows that the issue is not specific to the python39
port, but that wasn't really the expectation anyway.
As I said above it seems to me the segfault is generated somewhere in Qt5, but I am not familiar enough with these crash reports to understand enough of the details to have something actionable.
comment:7 Changed 3 years ago by reneeotten (Renee Otten)
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
I do observe the same indeed and it doesn't happen when installing through
pip
in avirtualenv
, so it appears to be MacPorts related. Of course we do override a few dependency version restrictions that upstream lists, because it's impractical for MacPorts to do such version pinning as we don't want to/cannot maintain several version for the same port. Having said that, a quick attempt by installing the latest, officially unsupportedipython
versions does not reproduce the segmentation fault.Unfortunately, I have no experience in reading/understanding the crash report you added, other than that the crash appears to happen somewhere in Qt5 if I'm not mistaken. If you know what the issue is and how to fix it, please let us know. I have no immediate plans to work on this given that the program works and *just* gives the segmentation fault on exit (even though I admit that that isn't ideal).