[BUG]: CommandXboxController.getRightX() and CommandXboxController.getRightY() not working

[rayluo]

Problem description

This is a regression in robotpy 2025.

Operating System

Linux

Installed Python Packages

$ pip list
Package                  Version
------------------------ ----------
bcrypt                   4.2.1
cffi                     1.17.1
cryptography             44.0.0
flexcache                0.3
flexparser               0.4
iniconfig                2.0.0
packaging                23.2
paramiko                 3.5.0
Pint                     0.24.4
pip                      24.3.1
platformdirs             4.3.6
pluggy                   1.5.0
pycparser                2.22
pyfrc                    2025.0.0
PyNaCl                   1.5.0
pynetconsole             2.0.4
pyntcore                 2025.2.1.1
pytest                   8.3.4
pytest-reraise           2.1.2
robotpy                  2025.2.1.1
robotpy-cli              2024.0.0
robotpy-commands-v2      2025.1.1
robotpy-hal              2025.2.1.1
robotpy-halsim-gui       2025.2.1.1
robotpy-installer        2025.0.1
robotpy-wpilib-utilities 2025.0.0
robotpy-wpimath          2025.2.1.1
robotpy-wpinet           2025.2.1.1
robotpy-wpiutil          2025.2.1.1
robotpy-xrp              2025.2.1.1
tomli                    2.2.1
tomlkit                  0.13.2
typing_extensions        4.12.2
wpilib                   2025.2.1.1

Reproducible example code

from wpilib import TimedRobot
from commands2.button import CommandXboxController

class MyRobot(TimedRobot):
    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.c = CommandXboxController(0)

    def robotPeriodic(self):
        print(self.c.getLeftX(), self.c.getLeftY(), self.c.getRightX(), self.c.getRightY())
        return super().robotPeriodic()

In robotpy 2024, you can move either sticks up, down, left, right, and observe the corresponding getLeftX(), getLeftY(), getRightX(), getRightY() returning a float ranging from -1 to 1. All good.

But in robotpy 2025, only the left-hand stick provides expected readings, the right-hand stick's up-down movement results -1 to 1 being returned in getRightX(), while it should have been from getRightY() instead; and then the getRightY() always return -1.

At one point, I can reproduce this with two different models of Xbox controllers, which are an Xbox Series S/X controller and an Xbox One controller. So, it feels like a software issue.

UPDATE:
My Xbox Series S/X controller exhibits this issue, while my Xbox One controller works fine. Upon close inspection of the raw "Axis[?]" readings from the simulation UI, I noticed that the different Xbox controller's inputs were mapped to different orders IN MY ENVIRONMENT.

• Xbox One controller: The Axis[0], Axis[1], ..., Axis[5] are mapped to Left stick horizontal movement, Left stick vertical movement, Left Trigger (throttle), Right Trigger (throttle), Right stick horizontal, Right stick vertical, respectively. And this config works flawlessly.
• Xbox Series S/X controller: The Axis[0], Axis[1], Axis[2] are same as above, but Axis[3], Axis[4], Axis[5] somehow mapped to right horizontal, right vertical and right trigger. This caused them being interpreted incorrectly by the wpilib.

I chalk this up to perhaps a faulty driver in my environment. Documenting my journey for posterity.

[auscompgeek]

Please include more details. What doesn't work about it? What version of the other RobotPy packages are installed?

[sciencewhiz]

Are you running with the driver station or in simulation? If simulation, I'm going to guess you didn't select the gamepad option

[rayluo]

Are you running with the driver station or in simulation? If simulation, I'm going to guess you didn't select the gamepad option

I was indeed using simulation, but I have already selected the gamepad option.

Please include more details. What doesn't work about it? What version of the other RobotPy packages are installed?

Sorry for not including details in my first report. Now I have added more details inside.

[auscompgeek]

I'm unable to reproduce the issue in simulation on Fedora Linux 41 using my 8BitDo Pro 2 in XInput mode with robotpy==2025.2.1.1.

Mind you, I had to guess at the imports, since your code snippet lacks the import statements (so isn't a complete robot.py). But I presume I got the imports correct given AFAIK there's only one TimedRobot and one CommandXboxController class in the entirety of RobotPy.

[virtuald]

I don't have an xbox controller so I can't really test this, but have you tried this in C++ or Java? It's likely an underlying problem in WPILib, as our command button (https://github.com/robotpy/robotpy-commands-v2/blob/main/commands2/button/commandxboxcontroller.py#L222) just calls the HID class which just wraps whatever WPILib is doing.

[rayluo]

Thanks for everyone who looked into this. Based on your hints, I made some progress.

The fact that the issue being unreproducible in @auscompgeek 's environment got me testing more, and found out that the same two Xbox controllers, same test snippet, with same robotpy version (albeit different bits when I tested it on Windows) works flawlessly on Windows.

Eventually I had my findings, and updated my bug report above to reflect that for posterity. Thanks again to @virtuald , @auscompgeek and @sciencewhiz for helping!

[auscompgeek]

Sounds like there's possibly a bug in the simulation GUI then. Please open an issue against allwpilib.