Running the tests in a random order (using pytest-randomly) causes a different set of tests to fail each time in test_magicbot_sm, indicating there is some global state being kept around and racing.
This changes the tunables component names to be unique between each test to hotfix the tests.
A proper fix would change setup_tunables to accept a NetworkTableInstance to attach tunables to, which we should probably do anyway, but changing these names was an easy and simple fix.
Running the tests in a random order (using pytest-randomly) causes a different set of tests to fail each time in
test_magicbot_sm
, indicating there is some global state being kept around and racing.This changes the tunables component names to be unique between each test to hotfix the tests.
A proper fix would change
setup_tunables
to accept aNetworkTableInstance
to attach tunables to, which we should probably do anyway, but changing these names was an easy and simple fix.