-
Notifications
You must be signed in to change notification settings - Fork 15
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
Test suite showing some ageing cracks #35
Comments
The current release doesn't use tox. It runs tests with pip install .[test]
nosetests You can see all tests green here: https://github.com/wireservice/agate-sql/actions/runs/1034818027 Whether |
The difference between I'm pretty sure the issue is going to turn out to be that your tests require old versions of dependencies. Note from your passing test log:
I'm testing on a distro that has has up to date versions of all dependencies. Just because your version of testing with pinned versions of everything pass doesn't mean all is well. That will help people that run venv's and know the right things to hold back. However an argument can be made that if it doesn't work on current stable released versions of dependencies that is at least an aging crack if not a bug. In this case the actual library seems to work fine still, just the tests fail, but being able to run tests is very valuable for distro packages that need to check if everything is found in the right places etc. |
Just for background, this is issue is being raised at all because I'm trying to get |
parsedatetime introduced a breaking change. That is why we exclude specific versions. See: certbot/certbot#8042 We are not the only project to reject those versions of parsedatetime. If, for Arch Linux, it is necessary to use the latest version of parsedatetime, even if it breaks lots of things like certbot, then I guess we will just have to hope that parsedatetime's maintainers un-break things. |
OpenSUSE seems to have written a patch bmwiedemann/openSUSE@6ba73c5 to revert the change in #33 As I mention in #33, the issue seems to be that, depending on the version of SQLAlchemy and MySQL, either the |
The tests should pass now! |
Great, thanks! Will there be a release soon to reflect this? Or should I try applying that SUSE patch for now? I'll hold off a few days if a release is immanent. |
0.5.8 should appear on PyPi shortly 🚀 |
I'm the packager for this on Arch Linux and I'm starting to run into some problems. I realize you use Tox here for testing and that makes sense for your upstream use case, but it does not allow testing in place on a system to confirm that distro packages are serving their functions.
First, using setuptools as a test runner stopped working a while back because datetime as used in this project is no longer compatible with the current released version of datetime.
Second, switching to pytest as a runner, I get the following test failure. I presume this is again a time parsing issue related to Python upstream datetime changes.
Additionally it's worth noting that some other deprecations warnings are in play and Python 3.10 is going to add another failure, although it looks like that might be in a transitive dependency.
It is probably worth updating this project to use current APIs and doing a minor release so that it can be used more robustly than inside a private venv bubble with old releases of stuff.
The text was updated successfully, but these errors were encountered: