Closed liamhuber closed 1 month ago
:point_left: Launch a binder notebook on branch _pyiron/pyiron_workflow/retargetci
Coverage variation | Diff coverage |
---|---|
:white_check_mark: +0.00% (target: -1.00%) | :white_check_mark: ∅ |
Codacy will stop sending the deprecated coverage status from June 5th, 2024. Learn more
Totals | |
---|---|
Change from base Build 9068914836: | 0.0% |
Covered Lines: | 3524 |
Relevant Lines: | 3946 |
pympipool
is running into trouble on windows, but from the trace I don't yet see where in the tests it's happening:
..Traceback (most recent call last):
File "C:\Users\runneradmin\miniconda3\envs\my-env\Lib\site-packages\pympipool\backend\serial.py", line 65, in <module>
main(argument_lst=sys.argv)
File "C:\Users\runneradmin\miniconda3\envs\my-env\Lib\site-packages\pympipool\backend\serial.py", line 31, in main
input_dict = interface_receive(socket=socket)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\runneradmin\miniconda3\envs\my-env\Lib\site-packages\pympipool\shared\communication.py", line 167, in interface_receive
return cloudpickle.loads(socket.recv())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ModuleNotFoundError: No module named 'test_macro'
E......Traceback (most recent call last):
File "C:\Users\runneradmin\miniconda3\envs\my-env\Lib\site-packages\pympipool\backend\serial.py", line 65, in <module>
main(argument_lst=sys.argv)
File "C:\Users\runneradmin\miniconda3\envs\my-env\Lib\site-packages\pympipool\backend\serial.py", line 31, in main
input_dict = interface_receive(socket=socket)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\runneradmin\miniconda3\envs\my-env\Lib\site-packages\pympipool\shared\communication.py", line 167, in interface_receive
return cloudpickle.loads(socket.recv())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ModuleNotFoundError: No module named 'test_node'
Should be able to find it by searching "test_node" and "test_macro"
This windows failure is actually really weird. I can't find test_macro
as a standalone string anywhere in my test suite, and test_node
is only a function name in a couple places, so my guess is that it's referring to the entire tests/unit/test_X.py
module. Why would it lose these?
There is no ----- aaahhhh, right, adding the test directories to the python path, that's what's changed. I was initially suspicious of the runner, but I saw in the last merged PR that the runner version, OS, runner image, and runner image provisioner were all identical, so I knew it had to be my fault somehow.
Ok, this should be easy enough to fix by decoupling the test directory input from the add-to-path input.
EDIT: interesting to note that this is no longer necessary for linux, but only for windows.