The refactor no longer calls Serial1.setPins() on nRF52, so Serial1 uses
PIN_SERIAL1_RX/TX directly. Those were swapped (the GPS transmits on
P0.20), so no data was received. Fix them and define PIN_GPS_TX/RX by the
gps module's side, as other variants do.
The ATGM336H keeps its NMEA output config across power cycles, and other
firmware may leave every sentence disabled. Re-enable GGA + RMC after each
power-up.
I used an LLM to review what changes the likelihood of a PR getting merged. I then asked it to select potential changes based on its findings. I selected from those options and carefully guided the LLM to produce several pull requests. I fed the subset of PRs that failed back into the skill and discovered additional missing factors.
Document the stats-core output fields and decode the `errors` bitmask
(shown as "Debug flags" in the app), from the ERR_EVENT_* flags in
Dispatcher.h, and note that the flags stay set until reboot or
`clear stats`.
Fixes#3039
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds the new variant to the PR build matrix, next to the existing STM32
entry, so it is covered by CI going forward.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Adds repeater, room_server, companion_radio_usb and kiss_modem
environments for the NUCLEO-WL55JC1 (STM32WL55JC, integrated SubGHz radio).
Notes on the board setup:
- Uses the upstream nucleo_wl55jc board. PlatformIO's boards_remap maps it
to the core's NUCLEO_WL55JC1 variant, so the LEDs, user buttons and the
LPUART1 -> ST-LINK VCP serial mapping all come from the core rather than
being redefined here.
- The RF switch table is built from the core's LORAWAN_RFSWITCH_* macros
(SP3T front end on PC3/PC4/PC5) so it tracks upstream if those are ever
revised. Both RFO_LP and RFO_HP are registered, so RadioLib selects HP
above 14 dBm and LP below it.
- board_build.extra_flags pins the device to STM32WL55xx. The stock board
manifest declares product_line STM32WL55xx but extra_flags -DSTM32WLE5xx;
that only works because every #elif chain happens to test WL55xx first.
- The Nucleo's LD1..LD3 are driven active HIGH, so the TX LED hooks are
inverted relative to STM32Board's active-LOW default.
- 32 KB reserved at the top of flash for LittleFS, as on the other
STM32WL variants.
Verified on hardware: all four environments build; repeater and room_server
flash and run over ST-LINK; identity and prefs persist across reboot and
across a firmware swap; and GPIOC reads back the documented RX levels at
idle, confirming the front-end switch is driven correctly.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The room server gated its filesystem paths on NRF52_PLATFORM only, so an
STM32 build fell through to the ESP32 branch and failed to compile:
MyMesh.cpp:147: no matching function for call to
'Adafruit_LittleFS::open(const char*&, const char [2], bool)'
main.cpp:69: 'store' was not declared in this scope
STM32 uses the same InternalFS / Adafruit_LittleFS API as nRF52, including
append semantics (FILE_O_WRITE opens LFS_O_RDWR|LFS_O_CREAT and then seeks
to the end), so the nRF52 branch is already correct for it. Add
STM32_PLATFORM alongside NRF52_PLATFORM in the four remaining gates;
saveIdentity() already did this.
This unblocks the room_server example on the existing STM32WL boards
(wio-e5-mini, wio-e5-dev, rak3x72, tiny_relay) as well as new ones.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>