-
-
Notifications
You must be signed in to change notification settings - Fork 2.8k
rm_rf fails for a set of tests #5974
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
Comments
IIRC correctly the previous implementation would silently ignore the error... when refactoring that part we decided to add the warning. Should we hide this warning by default then? This at least allows users to selectively show the warning if they are investigating why a particular folder is not being deleted... |
@nicoddemus we should ensure this particular warning never turns into a error |
I've also seen this lately.. in that case the call was |
i did actually rf it, will do the ls next time i seethe issue |
|
It should warn about this once at most per test run btw. It might have been triggered by ctrl-c'ing a previous test run, i.e. there is code that does not behave correctly in this case when creating the temporary directory, since it should never end up with those permissions. |
When
|
Note: we should remove the filter added by #5993 once this is sorted out. |
It is bad btw that it loses the exceptions context (original source):
|
Hi guys. Did this get fixed recently? I don't understand it much but this diff made it go away:
|
No, I am seeing this myself from time to time. Your patch suggests that it might be due to a broken test? (I've thought this would be some race with pytest itsef - wasn't paying enough attention apparently) |
@tarpas |
My hypothesis is that these two tests change the permissions of directories in tmpdir so that the standard teardown is not able to delete them. On MacOS I wasn't even able to rm -Rf them from command line easily. Why the warning seems to come out during some other test execution I don't know. I tested my solution very briefly so I'm not sure it's the right fix. Please go ahead with my patch, thanks! |
Cool, #6136. The warning only happens when pytest cleans them up, which only happens later (it keeps the most recent ones). |
i started to observe the following on the features branch:
together with the pytest warning configuration this means the "world" breaks very regular
The text was updated successfully, but these errors were encountered: