-
-
Notifications
You must be signed in to change notification settings - Fork 1.7k
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
Errors with erfa when running the tests twice (with 1.2rc1) #5082
Comments
Curious... Can you give platform/version details, @bsipocz? I can't reproduce this on py 3.5 on a Mac (using the builtin pytest)... |
Ohhh, turns out it's a py 2.x-only situation. I do get these errors on 2.7, but not 3.x. I haven't the slightest idea why this would be happening, though. Maybe something weird with the sorcery that pytest and some of our pytest plugins use to make the tests run in a "safe" environment? I think this isn't really a blocker because in the past this has been the case (invoking the tests twice in the same session cause weird pytest-caused errors). As long as actual real use cases don't do this I think we're OK. But it would certainly be useful to try to figure out the root cause of this! |
@eteq - I totally agree that this is not a real use case, and only came up as I tend to run the tests twice after one another, first without Also you're right that it is a py2 only issue, I can confirm that I can't see it on python3 (lesson learnt, do not have many test runs on several machines at the same time to avoid mess in the reporting.) I've updated the wiki). Probably unrelated, but I wonder whether it is normal t have different number of total tests when running them twice (e.g. 8330 vs 8187 for py3)? |
I definitely wouldn't call it "normal" ;) there has long been some weirdness in the twice-running situation that I really have never understood... I'm labeling this priorty-low because of the easy work around of "quit and re-run the tests", but it would be really good if someone could figure out what's going on just so we can be sure there aren't some subtle problems this is a sign of... |
Same issue with py2.6 on rhel6 _without_ running 'tests' twice: { time python test_astropy_package_tests.py ; } &> test_astropy_package_tests.log test_astropy_package_tests.py: test_astropy_package_tests.log: Running tests with Astropy version 1.1.2. Date: 2016-08-29T14:37:36 Platform: Linux-2.6.32-431.29.2.el6.x86_64-x86_64-with-redhat-6.4-Santiago Executable: /usr/bin/python Full Python Version: encodings: sys: ascii, locale: UTF-8, filesystem: UTF-8, unicode bits: 20 Numpy: 1.7.2 collecting ... collected 46 items rhel6/python-2.6/lib/python2.6/site-packages/astropy/tests/test_logger.py::test_warnings_logging_disable_no_enable PASSED =================================== FAILURES ===================================
rhel6/python-2.6/lib/python2.6/site-packages/astropy/tests/tests/test_imports.py:67: /usr/lib64/python2.6/pkgutil.py:113: in walk_packages
rhel6/python-2.6/lib/python2.6/site-packages/astropy/coordinates/builtin_frames/icrs_cirs_transforms.py:17: ImportError real 0m3.582s |
Please note also in previous log the presence of garbled characters in test_run_tests.py::test_doctest_float_replacement test result |
Is Python 2.6 still relevant? We already dropped Python 2.6 support. |
I agree with @pllim. py2.6 fails is still relevant only for the LTS, for any other cases please upgrade the installs to get rid of the errors. |
Answering pllim's "Is Python 2.6 still relevant?" question: Good question ! The longer answer is: But since recent CASA versions ship with their own "private" py2.7 (older versions had py2.6),
So I reverted to install Astropy with RHEL6 py2.6, which required to downgrade Astropy version to 1.1.2.
|
@ernest-bart - just to be clear:
I would therefore recommend using the 1.0.x release series in your quality assurance environment since we can still fix issues there. |
Thank you for the clarification. Following your recommendation here is an update on rhel6 + astropy 1.0.10
astropy.test(test_path='/opt/share/python-packages/rhel6/python-2.6/lib/python2.6/site-packages/astropy/wcs/tests/test_wcsprm.py',verbose=True) Running tests with Astropy version 1.0.10. Platform: Linux-2.6.32-504.el6.x86_64-x86_64-with-redhat-6.4-Santiago Executable: /usr/bin/python Full Python Version: encodings: sys: ascii, locale: UTF-8, filesystem: UTF-8, unicode bits: 20 Numpy: 1.7.2 collected 99 items ... : Note: deleted first passed tests results =========================================================== 97 passed, 2 xpassed in 0.24 seconds =========================================================== |
I'm not sure whether this has been fixed or just worked around, but I don't see the issue any more on current master. |
The errors appear when running the tests twice from the same python session (e.g.
python -c 'from astropy import test; test(); test()'
)These show up on all systems I've tried so far. Also the whole test suite needs to be run, it's not enough to run the tests twice for either the coordinates or time modules.
The text was updated successfully, but these errors were encountered: