[ainyan03]

Summary

When every PMIC ID probe fails (an AXP192 / AXP2101 board whose PMIC does not answer), the charge and output setters already refuse the unconfirmed register map, and getChargeStateCaps() returns an empty set. getChargeControlCaps() and getPowerOutputCaps() still advertised their paths, so a caller could be offered a path whose setter always returns false. They now return 0 in that state.

The cross check between getChargeState() and getChargeStateCaps() (a state the capability set does not advertise) also runs in device builds with a debug log level (CORE_DEBUG_LEVEL >= 4 or CONFIG_LOG_DEFAULT_LEVEL >= 4). On a device a mismatch is logged with ESP_LOGE and the returned state is unchanged; the PC build keeps the assertion. Device builds that define M5UNIFIED_CHECK_CHARGE_STATE_CAPS now log instead of stopping at an assertion.

Testing

• every capability bit cleared in this state was matched against its setter (setBatteryCharge, setChargeCurrent, setChargeVoltage, setExtOutput, setUsbOutput, setVibration, setExtPortBusConfig): each refuses the unconfirmed PMIC first
• the debug check is compiled into an ESP32-S3 build with CORE_DEBUG_LEVEL=5
• On hardware: 40 boards start with the same detected board as develop
• Builds: ESP-IDF 5.1 to 6.1 and Arduino-ESP32 2.x / 3.x (ESP32, S2, S3, C3, C5, C6, C61, H2, P4), SDL, fork CI passed