Remove unrelated ecosystem comparisons and software citations while retaining hardware links, measured model names, and required license notices.
Preserve stock-peer runners and tests in a verified local validation package. Keep shared Companion USB transport in the firmware repository for the raw-radio sweep and OTA seeder, with focused CI coverage. Restore missing hosted documentation links.
Forward --inplace-memory only when explicitly supplied, preserving authenticated layout bounds and the conservative legacy fallback.
Validation: MOTATOOL_TEST_BIN=/home/mesh/git/motatool/target/release/motatool python3 -m unittest test_lora_ota (318 passed). Both new auto-sizing regressions fail against the prior runner.
Add receive-only same-image display, SPI, modem/IQ-cache and scoped-priority experiments with checkpointed collectors. Keep rejected clock-phase correction and elevated-priority experiments out of production.
Publish sanitized timing data, Full Companion/Bluetooth/USB checks and redraw measurements with their limits. Wire the new behavioral checks into CI.
Run DHCP and renewal in a cooperative worker, bound socket commands, and retain SPI ownership until cleanup is safe. Preserve framing deadlines, response ownership and whole-frame replay, including pending-ACK queue pinning.
Budget the worker additively and retain all OTA names with an exact-checked Full-only compact index. Add full-driver, sanitizer, queue, shutdown, storage and memory regression coverage.
Preserve Reader streaming during Morse timing and block synchronous inflate/progress interleaving. Fail closed on uncertain port ownership and enforce application-only DFU packages before querying USB. Cover the guards in CI and clarify experimental benchmark scope.
Use a 960-byte constant-time Ed25519 base table, compact hash/scalar code and unused-library trims while retaining Reader, BLE P-256 and the 8 KiB stack. RAK4631 Full saves 57,496 flash bytes and 3,152 static RAM bytes versus preview3.
Support complete chapters and verse ranges with bounded app continuation replies and cooperative terminal output. Use nonblocking TCP writes under reader backpressure. Add scalar/signature comparisons and a linked ARM crypto/stack CI gate.
Validated clean RAK4631, T1000-E, T096 and V4 Full builds, host regressions, and physical RAK chapter/reconnect and signing comparisons. Physical signing averages 35.9 ms versus 14.4 ms with the original table.
Keep the deployed SD target and layout, select qualified internal delta storage when no card is usable, and retain exact SD bootloader-update gates. Add storage, capability, transfer, recipe, and release-selection regression coverage.
Also preserve mandatory runtime packet logging in canonical XIAO images when verbose build debugging is disabled, fixing the four stopped release-matrix qualification failures.
mikecarper