Bump the batch payload codec to schema v4 with separate meter-time and UTC anchors, then use meter seconds for sparse batch slotting and receiver reconstruction.
Update the current 868 MHz bench configuration, allow ACKs from configured receiver short IDs, improve AP-to-STA recovery, quiet the test build, and document the changed protocol in the README.
- Replace 11 per-frequency build environments with 3 role-based targets
(production, debug, test) using shared [env] base section
- Move LoRa frequency and sender-ID config from build flags into config.h
so all variants build from the same source
- Add -fstack-protector-strong, -D_FORTIFY_SOURCE=2, -Wformat-security
- Add Unity test framework to lib_deps for pio test support
- Add __pycache__/ to .gitignore
- Add LoRa idle/sleep/receive-window helpers and use short RX windows for ACK/time sync
- Schedule sender time-sync windows (fast/slow) and track RX vs sleep time in debug
- Lower sender power (80 MHz CPU, WiFi/BT off, reduced ADC sampling, unused pins pulldown)
- Make SERIAL_DEBUG_MODE a build flag, add prod envs with debug off, and document changes
acidburns