V0.6.x#11
Open
tristanalderson wants to merge 31 commits into
Open
Conversation
…tore, flight recorder Replaces the flash-table design with AimFileSystem (littlefs over AimBlockDevice drivers), AimConfigStore (atomic JSON load/save), and AimFlightRecorder (RDES-compressed row logging with dump protocol). Updates AimNetwork to 0.4.x packet layout and examples. Baseline commit of in-flight work prior to fw/config-ownership changes. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…load - AimFlightRecorder::closeLog(): sync+close the log handle and reset RDES state so callers can safely format the filesystem; next writeRow() starts a fresh log with a raw origin row. Destructor now reuses it. - AimFlightRecorder::isLogging() accessor. - AimConfigStore::loadDetailed(): distinguishes NOT_PRESENT (normal — first boot or post-reset) from READ_FAILED / PARSE_FAILED / STORAGE_NOT_READY. bool load() now delegates, behavior unchanged. Additive only; both ESP32 and STM32 consumers build clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Device (AimFlightRecorder::serviceDump): pad the final block to the fixed 512 B block size with 0xFF at EOF. The host reads exactly blockSize per block and trims to totalBytes afterwards, so a short final block left it in an infinite resend loop (the ~13 min wedge seen on hardware). Host (extract_tool/main.py): - enter_console_root(): console entry needs a DOUBLE Ctrl-A within the firmware's 750 ms guard window; the old single-press path silently failed and the tool hung waiting for the dump handshake. - Ack every block with 'N' (the final ack is what tells the board the dump is complete; previously the device stayed wedged in dump mode). - Bounded waits: 5 s for the '#' handshake, max 5 'L' resend retries per block — protocol faults now raise IOError instead of hanging. Verified on bench ESP32-S3: full dump end-to-end including a padded 422 B final block, 19/19 console regression checks. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
startDump now saves and zeroes the global logger filter mask (stopDump
restores it) so every dump consumer gets a clean binary stream instead of
each caller re-implementing the mute. The dump wire-protocol bytes
('#'/'N'/'L') become named constants in AimFlightRecorder.h, mirrored in
extract_tool/main.py as the shared host contract.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…etwork Clean/aim network
Both were orphaned by the console refactor — the console now goes through AimFlightRecorder/AimConfigStore and nothing in any consumer (Upper_Control, Lower_Control_test, examples) calls them. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Fixed-size periodic and latched-dirty TX slots; driven by a single service() call per loop tick. Replaces scattered per-module timers and eliminates the need for a sendTimedPktEx method in the transport. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…etwork Clean/aim network
…onsole soon to be applied eslewhere), fix bug in config load
…nsolidate CAN-ID layout and Frame struct
…lify-pass Clean/simplify pass
tristanalderson
requested review from
RealLordzombieboy,
eat7890-stick and
jeevan9s
aim_catalog.h is the single source of truth for classes, subjects, scaling, and fixed priorities; it lands in the same commit as the protocol doc per the catalog/header invariant. kSchemaVersion=1. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Rename AimLogger -> aim_logger (logger.cpp/.h unchanged) and replace the
single example/ with paired examples/{esp32,stm32}_logger/ PlatformIO
projects.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Rename AimNetwork -> aim_network and rewrite the protocol from v0.4.2 (Pkt, 11-bit IDs) to v0.6.x: Msg with 29-bit extended IDs, class-mask filtering, new aim_job.h timer helper and aim_catalog.h message catalog, updated ESP32 (TWAI) and STM32 (bxCAN) backend cores. Rework the canbus examples: fix swapped listener/publisher header banners, drive the 1 Hz time sync off g_tick1000, and drop dead locals, an unused aim_job.h include, and a duplicate trailing comment block. Move the protocol spec to the repo root as the workspace ground-truth doc (aim_protocol_v0.6.x_draft.md -> aim_protocol_v0.6.0.md) and fix its title version to match kNetworkVersionString = "0.6.0". Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Rename AimFlashStorage -> aim_flash_storage and migrate the littlefs submodule to the new path. Relocate the flash debug console out of av-prop/Upper_Control into this shared library as a hook-based skeleton (aim_console), and replace the SerialFlash dependency with a direct SpiNorFlashDriver that auto-detects any W25Q-compatible chip from its JEDEC ID (covers both BY25Q128ES and W25Q128JV). Fix a dump-stall bug: the console DUMP state now peek()s for the q/b abort key instead of read()ing it, so the host's N/L block-control bytes reach serviceDump(). Verified end-to-end on ESP32 hardware (29,870 rows). Add esp32_flash and stm32_flash examples, redo the extract_tool, and rewrite the README. Delete AimConfigStore and AimNodeConfig — node identity and the telemetry schema are build-time constants, so nothing needs persisting at runtime (drops the ArduinoJson dependency). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Update the root README "What's Inside" entries and the Quick-Start example path to the new snake_case library names, and correct the CAN design note: v0.6.x uses content-based addressing (consumers filter by message class/subject) — there is no destination field. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…ES constants Renames aim_console_init/service/is_active to lowerCamelCase (aimConsoleInit, aimConsoleService, aimConsoleIsActive) to match workspace naming convention. Extracts _isGeometryValid() helper in AimFileSystem to de-duplicate three identical geometry guard checks. Replaces bare 0x80/0xC0/0xE0 RDES bit-prefix literals with named constexpr constants (kRdesLvl2Prefix, kRdesLvl3Prefix, kRdesRaw32Prefix). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Rename valve/PT subjects to AV-tag names (Av204/Av203/Av205, AvSpare), add power-FET subjects (PwrPtUcm, PwrSolLcm, PwrPtLcm), raise priority for LCM sensors. Bump kSchemaVersion 1 -> 2 (carried in every HEARTBEAT).
service() now takes the loop's local millis() as its first argument instead of calling millis() internally, so liveness/heartbeat timing uses the same clock the caller schedules on. All six nodes are updated to match.
Concurrent LOG_* calls (e.g. the ESP32 UCM logging from both cores) could interleave mid-line on the shared stream. Guard the single output write with a std::mutex, gated on _GLIBCXX_HAS_GTHREADS so it is a real lock only where the toolchain has threads and a no-op on single-threaded bare-metal nodes.
…ff recovery, console blocking read
STM32 bxCAN switches to interrupt-driven TX/RX with static mailbox queues.
ESP32 TWAI gains bounded bus-off recovery with cooldown. aim_console adds
aimConsoleWaitRead() for multi-byte hook parsing. Job drops default member
init in favor of explicit {period, 0U} at all call sites.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.