Open Hellseher opened 7 months ago
I can't reproduce this with a standard test run. Is it possible this was a temporary failure driven by the parallel run?
$ pytest astroquery/esa/xmm_newton --remote-data
============================================================================================================= test session starts =============================================================================================================
platform darwin -- Python 3.10.12, pytest-7.4.2, pluggy-1.3.0
Running tests with astroquery version 0.4.7.dev9202_testrun_testrun.
Running tests with astropy_helpers version 2.0.9.
Running tests in astroquery/esa/xmm_newton.
Date: 2024-03-18T13:39:05
Platform: macOS-13.6-arm64-i386-64bit
Executable: /Users/adam/mambaforge/envs/py310forge/bin/python3.10
Full Python Version:
3.10.12 | packaged by conda-forge | (main, Jun 23 2023, 22:41:52) [Clang 15.0.7 ]
encodings: sys: utf-8, locale: UTF-8, filesystem: utf-8
byteorder: little
float info: dig: 15, mant_dig: 15
Package versions:
Numpy: 1.26.2
Matplotlib: 3.7.2
Astropy: 6.1.dev589+g2390bf8f8f
regions: 0.8
pyVO: 1.5.dev228+ga035576
mocpy: not available
astropy-healpix: 1.0.0
vamdclib: not available
Using Astropy options: remote_data: any.
rootdir: /Users/adam/repos/astroquery
configfile: setup.cfg
plugins: anyio-4.0.0, remotedata-0.4.1, asdf-3.0.0, doctestplus-1.2.1, astropy-header-0.2.2
collected 34 items
astroquery/esa/xmm_newton/tests/test_xmm_newton.py .......................... [ 76%]
astroquery/esa/xmm_newton/tests/test_xmm_newton_remote.py ........ [100%]
======================================================================================================== 34 passed in 84.28s (0:01:24) ========================================================================================================
It's totally possible to have bugs in these tests e.g. one assumes a file is around that a previous test was creating and that hasn't yet been cleaned up. Any further investigation and PRs to fix it are more than welcome!
@keflavich as @bsipocz mentioned it could be a race condition during parallel run of the tests which I tried to enable. It passes just fine in single thread invocation.
(fwiw, I do see these failures locally, in fact I see a couple more of the same kind for the esa.xmm_newton module (6 instead of 2))
Hi,
During scheduled update of astroquery in Guix I tried to speed up unit tests with
pytest-xdist
. It's minimized the wait time especially on multi threaded CI farm. Unfortunately by enabling it I've got 2 tests failed: