Skip to content
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

Github workflow failures #144

Closed
roryyorke opened this issue Jan 30, 2021 · 1 comment
Closed

Github workflow failures #144

roryyorke opened this issue Jan 30, 2021 · 1 comment

Comments

@roryyorke
Copy link
Collaborator

See https://github.com/python-control/Slycot/actions/runs/522944389

The failure is for Python 3.9, Generic BLAS. The diagnostic is a low-level error that occurs during the python-control tests:

Windows fatal exception: code 0xc0000374
control\tests\xferfcn_test.py .......................................... [ 97%]
Windows fatal exception: code 0xc0000374

Current thread 0x00000730 (most recent call first):
  File "<string>", line 1 in <lambda>
  File "C:\Miniconda\envs\test-env\lib\tokenize.py", line 590 in _tokenize
  File "C:\Miniconda\envs\test-env\lib\site-packages\coverage\phystokens.py", line 147 in generate_tokens

Something similar happens in the Python 3.6 Generic BLAS, which is "FAILOK":

2021-01-30T08:10:18.8133613Z control\tests\nichols_test.py ..                                         [ 25%]
2021-01-30T08:10:19.1146867Z Windows fatal exception: code 0xc0000374
2021-01-30T08:10:19.1147451Z 
2021-01-30T08:10:19.1148013Z Current thread 0x00000bc4 (most recent call first):
2021-01-30T08:10:19.1150102Z   File "C:\Miniconda\envs\test-env\lib\site-packages\matplotlib\transforms.py", line 116 in __init__

Comparing the most recent 3.9 with the same build at https://github.com/bnavigator/Slycot/actions/runs/507101266, there are minor differences (different Github Windows env version, different git version, different build of python 3.9.1, pip 21.0, ruamel_yaml, different version of tzdata, matplotlib, coverage, pytest, urllilb.

The number of tests collected is different, probably due to changes in python-control:

Succesful build in PR:

============================= test session starts =============================
platform win32 -- Python 3.9.1, pytest-6.2.1, py-1.10.0, pluggy-0.13.1
rootdir: D:\a\Slycot\Slycot\python-control, configfile: setup.cfg
plugins: cov-2.11.1
collected 2339 items / 2 deselected / 2337 selected

Failed build in master:

============================= test session starts =============================
platform win32 -- Python 3.9.1, pytest-6.2.2, py-1.10.0, pluggy-0.13.1
rootdir: D:\a\Slycot\Slycot\python-control, configfile: setup.cfg
plugins: cov-2.11.1
collected 2351 items / 2 deselected / 2349 selected

For now I suggest we make Windows Python 3.9 Generic FAILOK too - we're passing Windows Python 3.9 with OpenBLAS and Intel10_64lp.

@bnavigator
Copy link
Collaborator

A rerun failed the same job but during a different test:
https://github.com/python-control/Slycot/runs/1797192943?check_suite_focus=true#step:8:208

Looks like a problem with the windows matplotlib package on conda to me. Possibly still caused by different BLAS libs for numpy and slycot.

I would say we FAILOK every windows build except unset Blas.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants