[hal] Change usage reporting to string-based

[PeterJohnson]

TODO:

• Add documentation for usage conventions
• Remove the Count entry point (std::to_string / String.valueOf is easy)
• Change CameraServerShared to generic reporting as well
• Update code to match documentation
• Cleanup formatting

Fixes #301.

[github-actions]

This PR modifies commands. Please open a corresponding PR in Python Commands and include a link to this PR.

[sciencewhiz]

Fixes #301

[KangarooKoala]

Why is there a mix of adding one to the port and not adding one to the port? (Notably, AnalogAccerelometer and DigitalInput have different behavior depending on the language)

Files that started adding one: None
Files that still add one: hid.lang.jinja, pwm_motor_controller.lang.jinja, AnalogAccelerometer.cpp, DigitalInput.cpp
Files that never added one: ADXL345_I2C.java (didn't report port before), I2C.lang (used to report just device address, now also reports port), PowerDistribution.lang (didn't port module before), UpDownCounter.java
Files that no longer add one: AddressableLED.lang, AnalogAccelerometer.java, AnalogGyro.java, CAN.lang, Compressor.lang, DigitalInput.java, DigitalOutput.lang, DoubleSolenoid.lang, DutyCycle.lang, Encoder.lang (part of changing from using FPGA index to using a and b channels), Joystick.lang, PWM.lang, SerialPort.lang, Servo.lang, Solenoid.lang, Tachometer.lang, UpDownCounter.cpp, NidecBrushless.lang

[PeterJohnson]

Why is there a mix of adding one to the port and not adding one to the port? (Notably, AnalogAccerelometer and DigitalInput have different behavior depending on the language)

We shouldn't be adding 1 anywhere; I'll fix the remaining instances (I realized I needed to fix that partway through doing these updates which is why some were missed). They should have always added 1 in the "old" implementation because 0 had a special meaning in the old implementation.