[BUG]: tests intermittently timing out robotInit

[auscompgeek]

Problem description

we saw this time out in GitHub Actions today

=================================== FAILURES ===================================
________________________________ test_disabled _________________________________
...
            # If your robotInit is taking more than 2 seconds in simulation, you're
            # probably doing something wrong... but if not, please report a bug!
>           assert self._cond.wait_for(lambda: self._robot_initialized, timeout=2)
E           AssertionError

.venv/lib/python3.14/site-packages/pyfrc/test_support/controller.py:75: AssertionError

https://github.com/thedropbears/pyrebuilt/actions/runs/21539711405/job/62072162572?pr=35#step:5:138

I don't think we're doing anything particularly unusual in our code

Operating System

Linux

Installed Python Packages

Package                  Version
------------------------ ---------
attrs                    24.2.0
bcrypt                   4.2.0
cffi                     2.0.0
coverage                 7.13.1
cryptography             43.0.3
flexcache                0.3
flexparser               0.4
hypothesis               6.119.3
importlib-resources      6.5.2
iniconfig                2.0.0
librt                    0.7.8
mypy                     1.19.1
mypy-extensions          1.0.0
numpy                    2.3.5
opencv-python            4.10.0.84
packaging                23.2
paramiko                 3.5.0
pathspec                 1.0.3
phoenix6                 26.1.0
photonlibpy              2026.1.1
pint                     0.24.4
pip                      25.1.1
platformdirs             4.3.6
pluggy                   1.5.0
pycparser                2.22
pyfrc                    2026.0.1
pynacl                   1.5.0
pynetconsole             2.0.4
pyntcore                 2026.2.1
pytest                   8.3.3
pytest-integration       0.2.3
pytest-reraise           2.1.2
robotpy                  2026.2.1
robotpy-apriltag         2026.2.1
robotpy-cli              2026.0.1
robotpy-cscore           2026.2.1
robotpy-ctre             2026.1.0
robotpy-hal              2026.2.1
robotpy-halsim-gui       2026.2.1
robotpy-installer        2026.0.1
robotpy-native-apriltag  2026.2.1
robotpy-native-ntcore    2026.2.1
robotpy-native-wpihal    2026.2.1
robotpy-native-wpilib    2026.2.1
robotpy-native-wpimath   2026.2.1
robotpy-native-wpinet    2026.2.1
robotpy-native-wpiutil   2026.2.1
robotpy-rev              2026.0.1
robotpy-wpilib-utilities 2026.0.1
robotpy-wpimath          2026.2.1
robotpy-wpinet           2026.2.1
robotpy-wpiutil          2026.2.1
sleipnirgroup-choreolib  2026.0.1
sortedcontainers         2.4.0
tomli                    2.1.0
tomlkit                  0.13.2
typing-extensions        4.12.2
wpilib                   2026.2.1

Reproducible example code

[thomas302]

I've been running into the same issue. It seems to be an issue with how the tests interact with the phoenix6 motor controllers (and or devices generally).

It seems like once a certain threshold of devices are created it happens almost every time (I got it to pass once, and then the tests failed ~15 times after that with no changes to code).

A standard ctre swerve base + 5 or so motors seems to cause it to fail almost every time (1 pigeon, 4 cancoders, ~13 motors). With fewer motors the test failures are become quite few (like once out approximately 10 runs, and often only one or two test will fail).

[auscompgeek]

I've also seen no-isolation tests occasionally hang for the past year; I wouldn't be surprised if there's a deadlock somewhere in phoenix6 simulation.

[virtuald]

If you modify the timeout (by changing the python code) to something larger, does the issue go away? Or is it a deadlock in phoenix.

We could change it to 10 seconds by default -- but really, if it's longer than that, there's something broken that needs to be fixed.

[thomas302]

I just double checked with a 10 second timeout and it still fails.

Edit: I modfied my test controller incorrectly, so I may be mistaken. Ill try again in a minute.

Edit2: It seems to be passing consistently with the 10 second timeout.