-
-
Notifications
You must be signed in to change notification settings - Fork 688
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
failed tests #382
Comments
I recently ran into something similar so I pinned test dependencies. b616107 |
I went through and ran the tests with various pytest versions from 3.6.4 (the version in Fedora 29) and 3.10.0 (latest). They pass fine up until 3.7.4, then it gets messy.
|
We're using the patch from NixOS/nixpkgs#49602 in openSUSE now and it fixes the testsuite. |
Aiming to release a new update before the week ends. |
Released v1.7.0 last week. |
It's said to see you're viewing pinning the pytest version as a solution to this. |
@dotlambda I can't do anything about it with a comment like that though. What would you recommend we should do exactly? Stop pinning? Pin a specified range so it doesn't happen again? Add a compat layer to handle various versions of a test dependency? |
@jpadilla I would suggest using the patch from NixOS/nixpkgs#49602. Pinning users to an old version is a bad idea, in particular when users are getting their Python packages through their Linux distribution. |
@glaubitz That I can do, thanks. Wouldn't not pinning(at least a range) mean it's more likely for tests to break like they did at a later point? Current pytest version is now 4.0.1 |
Pinning is fine as long as you're not pinning to a fixed version and you're depending on a reasonably new version which distributions would ship in their unstable or rolling release distributions. In this case, the version you pinned on was relatively old looking at the number of releases pytest made since 3.7.0 was released. So, it's expected for Linux vendors to package versions later than 3.7.0 in their unstable or rolling release distributions. It's reasonable to pin using a range if you expect the API to break after a certain version number of your dependency. So, you can definitely use a range. Just don't depend on something old. |
I co-maintain the python-jwt package in Fedora for PyJWT. The test suite is now failing in Fedora 30 and I'm not sure why. The failing tests are:
TestJWS.test_decode_no_algorithms_verify_signature_false
TestJWT.test_decode_no_algorithms_verify_false
Expected Result
Tests to pass (just like they did in Fedora 29).
Actual Result
Full output: https://gist.github.com/carlwgeorge/563597c80e3324101d754efac99d14a1
Reproduction Steps
System Information
The text was updated successfully, but these errors were encountered: