Link proof signature, TCP proof routing, and split-packet handling
were causing intermittent message failures with Python LXMF clients.
Route >254B messages through link delivery, add proof retry for LoRa.
TCP: Move 2KB stack-allocated packet rewrite buffers in
send_outgoing() to a PSRAM-allocated member (_wrapBuffer).
The ESP32-S3 loop task has 8KB stack — a single 2KB VLA used
25%, risking overflow under deep call chains during TCP load.
WiFi AP: Cap concurrent client connections at 4. Previously
unbounded — any number of clients could connect and exhaust
memory. New clients beyond the limit get a clean TCP RST.
Outbound (Ratdeck→Python) was broken because Identity::recall() could
never find the recipient. Root cause: OS::time() returns seconds since
boot on ESP32, but persisted known_destinations entries carried timestamps
from the previous session. New announces got timestamp ~31s while persisted
entries had ~5000s, so the LRU cull immediately removed the new entry.
The microReticulum Identity.cpp fix (timestamp normalization on load) is
in .pio/libdeps and must be upstreamed to ratspeak/microReticulum separately.
Changes:
- known_destinations cap 256→512 (PSRAM pool was 1% used, plenty of room)
- Fix link delivery destHash: onLinkEstablished callback was passing link_id
instead of LXMF destination hash, corrupting conversation routing
- Add diagnostic logging: [LXMF-DIAG], [TCP-DIAG], [HEART-DIAG], [DIAG-PROOF]
for tracing link establishment, proof routing, and interface status
- Tighten transport announce rate limit from 5/sec to 2/sec
- Reduce RNS loop frequency from 200Hz to 100Hz
- Throttle LVGL rendering to 30 FPS (was unthrottled)
- Time-box TCP frame loop to 3 frames / 8ms per client
- Add 12ms global TCP budget across all clients
- Replace blocking WiFi announce delay(1500) with deferred non-blocking check
- Rotate persistData() across 3 cycles to spread file I/O
- Reduce LvNodesScreen rebuild frequency (2s → 5s, skip small transient changes)
- Bump version to 1.5.9
4-layer defense against 100+ node announce storms from rns.ratspeak.org:
- Transport-level rate limiter (5/sec) via filter_packet callback, before Ed25519 verify
- TCP frame processing time-boxed to 15ms per loop iteration
- Global announce rate limit (3/sec) in AnnounceManager
- UI rebuild throttled to once per 2 seconds
Also fixes message status not persisting to disk on queue drain.
- Fix LXMF wire format: standardize on opportunistic [src:16][sig:64][content]
- Fix LXMF signature: sign(dest||src||packed) per spec, remove message_hash
- Fix MsgPack interop: handle both str and bin types for title/content
- Radio defaults changed to Balanced preset (SF9/BW250k/CR5/TX14)
- Messages screen: sorted by most recent, preview with You:/Them:, green unread dot, timestamps
- Status bar: replace LoRa/BLE/WiFi text with signal bars (green=connected, red=offline)
- Home screen: remove Unread info (shown in Messages tab)
- Contacts screen added (Friends tab)
- Identity manager: multi-slot identity support with per-slot display names
- Message store: fix peer hash truncation, SD directory creation, .bak file leak
- Settings: check for updates, active identity display, info diagnostics
- Migrate all screens to LVGL v8.4 widget system
- Non-blocking radio TX (async endPacket via LoRaInterface)
- Live TCP server switching with transient node cleanup
- Fix UI freeze during radio transmit
- Trackball long-press delete, deferred click with debounce
- Pin microReticulum to 392363c, fix list_directory API
- Fix CI build: portable include path, remove hardcoded local path
- Rewrite MessageView with word-wrapped chat bubbles, timestamps,
incoming/outgoing alignment, pixel-based scrolling
- Add IdentityManager for multi-identity support (create/switch/delete,
8 slots max, per-identity display names)
- Fix transport mode: default to endpoint (no rebroadcast) — was acting
as transport node, rebroadcasting every TCP announce over LoRa causing
5-11 second UI freezes per TX. Loop time: 11146ms → 6ms
- Add Transport Node toggle in Settings > Network
- TCP interface now drains up to 10 frames per loop (was 1)
- Reduce LoRa debug polling from 5s to 30s
DeFiDude