mirror of
https://github.com/meshtastic/firmware.git
synced 2026-09-15 15:59:40 -04:00
Ben MeadorsandQuency-D
bfd1e1a231
* refactor(graphics): select Arduino_GFX panels with a capability flag TFTDisplay tested `defined(HACKADAY_COMMUNICATOR)` in a dozen places to mean "this panel is driven by Arduino_GFX rather than LovyanGFX". Every new Arduino_GFX board had to be appended to all of them. Move the decision into the variant as USE_ARDUINO_GFX so the display code stops naming individual boards. No behaviour change: the Hackaday Communicator is still the only board that sets it. * feat(boards): add Heltec RC32, RC52 and RCC6 Three boards around the same 128x220 NV3001B panel: RC32 (ESP32-S3), RCC6 (ESP32-C6) and RC52 (nRF52840). They differ only in how the panel bus is wired, so they share one branch in TFTDisplay behind TFT_NV3001B. RC32 and RC52 also carry a rotary encoder on a TCA6408 I2C expander. That lands as its own input source rather than as board conditionals inside i2cButton, which is the M5Stack UnitC6L button driver and stays untouched. On RC52 and RCC6 the panel is an add-on module, so probe it before reporting a screen. The probe reuses the bit-banged SPI helper that already backs the T114 ST7789 check. Arduino_GFX is pinned to the upstream commit that added the NV3001B driver; it has not shipped in a tagged release yet. Co-Authored-By: Quency-D <55523105+Quency-D@users.noreply.github.com> * feat(gps): detect and configure the LC760CA GNSS module The LC760CA is another Unicore part, so it joins the $PDTINFO probe family and reuses the CM121 message-rate setup. It answers with CC1161W. GNSS_MODEL_LC760CA goes immediately before GNSS_MODEL_GENERIC_NMEA: the sentinel has to stay last because isValidGnssModel() uses it as the exclusive upper bound on values the probe cache may hold. Placing the new model after it would leave LC760CA permanently uncacheable. Co-Authored-By: Quency-D <55523105+Quency-D@users.noreply.github.com> * fix(graphics): re-init the NV3001B after the panel rail comes back DISPLAYOFF de-asserts VTFT_CTRL, which cuts power to the panel, so the controller loses MADCTL, COLMOD and gamma. displayOn() only sends sleep-out and cannot restore them, leaving the panel dark or in the wrong format after wake. Re-run begin() once the rail has settled, and repaint in full since the re-init leaves display RAM undefined. Also stop the TCA6408 rotary polling from two threads at once. Registering as an InputPollable meant InputBroker's pollSoon task could call pollOnce() while runOnce() was mid-transfer on the main thread, with nothing serialising Wire or the decoder state. Drop InputPollable and have the interrupt wake the thread instead, the way ButtonThread does, so the bus and the decode stay on one thread. * fix(graphics): skip the NV3001B wake when re-init fails begin() reports whether the bus came up. Ignoring it meant a failed re-init still lit the backlight and drove a full-screen repaint at a panel that was never initialised. * chore(boards): ship the Heltec RC boards at release level release is the normal level for a variant; the matrix generator still builds each of these in this PR because they add a new platformio.ini. --------- Co-authored-by: Quency-D <55523105+Quency-D@users.noreply.github.com>
30 lines
654 B
JSON
30 lines
654 B
JSON
{
|
|
"build": {
|
|
"arduino": {
|
|
"partitions": "default_16MB.csv"
|
|
},
|
|
"core": "esp32",
|
|
"f_cpu": "160000000L",
|
|
"f_flash": "80000000L",
|
|
"flash_mode": "qio",
|
|
"hwids": [["0x303A", "0x1001"]],
|
|
"mcu": "esp32c6",
|
|
"variant": "heltec_rcc6"
|
|
},
|
|
"connectivity": ["bluetooth", "wifi", "lora"],
|
|
"debug": {
|
|
"openocd_target": "esp32c6.cfg"
|
|
},
|
|
"frameworks": ["arduino", "espidf"],
|
|
"name": "Heltec RCC6",
|
|
"upload": {
|
|
"flash_size": "16MB",
|
|
"maximum_ram_size": 327680,
|
|
"maximum_size": 16777216,
|
|
"require_upload_port": true,
|
|
"speed": 921600
|
|
},
|
|
"url": "https://heltec.org/",
|
|
"vendor": "Heltec"
|
|
}
|