tooling: Wire bq27441 into the smoke-test sketch (src/main.cpp)

[nedseb]

Context

PR #171 (feat(bq27441): Add driver for the on-board Li-Po fuel gauge.) intentionally does NOT wire the new driver into src/main.cpp. PR #172 (mcp23009e) is open in parallel and would touch the same file, so doing the integration inside either feature PR turns the second merge into a conflict — sequencing it after both drivers land is cheaper.

What to do

Edit src/main.cpp after PR #171 is merged to add the BQ27441 alongside the existing on-board drivers:

1. #include <BQ27441.h> next to the other driver headers.
2. Instantiate on the internal bus:

   BQ27441 bq27441(internalI2C);

3. Add a probe section in setup() matching the style of the existing drivers:

   // --- BQ27441 (Li-Po fuel gauge) ---
   if (bq27441.begin()) {
       Serial.print("BQ27441 detected, DEVICE_ID = 0x");
       Serial.println(bq27441.deviceId(), HEX);
       Serial.print("BQ27441 first read: V = ");
       Serial.print(bq27441.voltageMv());
       Serial.print(" mV, SOC = ");
       Serial.print(bq27441.stateOfCharge());
       Serial.println(" %");
   } else {
       Serial.println("BQ27441 not detected");
   }

4. The STeaMi variant header does not expose a GPOUT wake pin for the gauge, so leave the constructor default (gpout_pin = -1). If a future board revision wires it, pass it as the 4th constructor argument and begin() will pulse the wake sequence automatically.

Acceptance

• make build (steami env) green.
• Smoke firmware flashed on a wired STeaMi prints BQ27441 detected with a plausible voltage (3000–4300 mV) and SOC (0–100 %) on a charged board.
• No regression in the other drivers' output.

Why this is a follow-up

Doing this inside #171 would conflict with #172, which also extends src/main.cpp. Sequencing the smoke-test edit after both feature PRs merge means whichever feature lands second only rebases its own diff — not also a src/main.cpp integration edit.