Troubleshooting

Gradle cannot resolve CubicLib #

Check that the repository URL ends in /CubicLib/maven/ and the dependency is dev.cubiclib:cubic-ftc:0.2.0. This preview must be built locally; the hosted repository currently has 0.1.0 only. Add the local build/maven repository from Installation. Keep google() and mavenCentral(). Verify the project uses FTC SDK 12.0.0 and Gradle is not in offline mode.

After updating a repository configuration, sync again. Published release versions should not be overwritten; teams may still have the previous artifacts cached.

The dashboard does not open #

Connect to the robot’s network and check its actual IP. Use HTTP on port 8087. The team website and public docs are not the robot dashboard.

Confirm the Robot Controller app is running. Initializing a CubicOpMode installs the server as a fallback if automatic startup was unavailable. Check Robot Controller logs for a CubicLib startup error, such as a port conflict.

Position or heading is wrong #

Verify starting field pose, encoder direction, decoded counts, wheel diameter, offsets and IMU mounting orientation. Pure rotation should not produce substantial translation.

Keep the Pinpoint stationary during calibration. A non-finite software-localizer IMU reading latches localization invalid until resetPose is called with a known pose. Unknown movement during sensor failure cannot be reconstructed.

Setup changes are rejected #

Stop the initialized or running CubicOpMode before saving. The new configuration is used on the next Init. Validate wheel count, unique hardware names and non-singular wheel geometry. Pinpoint requires matching pod scales and forward/left wheel orientation.

No route suggestions #

Confirm field geometry and robot dimensions only after checking them. Record pickup-confirmed and score-confirmed events with valid, continuous pose data. Suggestions require meaningful distance reduction; not every cycle will produce one.