Both languages: MASTER yaml statics on a shared MASTER, per-stream RX on one slot, voice
sequence handling, bridge reset on the last peer, OpenBridge DNS anchoring and shared
passphrases, OBP_PROXY/ALIASES keys and reload rules, TGID_ACL on OPENBRIDGE, report v2
topology fields, plugins (voice_slot_for_tg, voice-announcements, slot sharing), server
prompts on TS2, TG 4000 leg release and voice reply, mesh_* modules.
LOOP_GUARD now defaults to true: an echo coming back through a transcoding bridge is
dropped unless a system sets log (write a *LoopGuard* line and let it through) or false.
The example yaml explains what the guard catches that stream-ID loop control cannot.
A remote caller's next over may reach this server first over another
OpenBridge link, so within the hold window it looked like an echo (8 of
3262 call starts on 2131, all OBP -> OBP re-keys). The mesh is now one
ingress: a real loop is still caught where the caller is local (HBP vs a
bridge peer, or HBP vs the mesh), and so is a local bridge peer echoing a
remote caller (the mesh vs HBP).
Show GLOBAL.LOOP_GUARD_HOLD and SYSTEM LOOP_GUARD with their defaults in
adn-server.example.yaml.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JnjAx7G27d9q88j9PFLj11
- playback_use_cases: _PLAYBACK_DELAY_S becomes public PLAYBACK_DELAY_S
(now used outside the module); the two echo tests that import it follow.
- loop_guard: MAX_LOOP_GUARD_HOLD_S = PLAYBACK_DELAY_S, so changing the
parrot's delay moves the cap with it. playback_use_cases only imports
from domain: no cycle.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JnjAx7G27d9q88j9PFLj11
- The parrot replays an over 2.0 s after it ends (_PLAYBACK_DELAY_S),
under whatever system/TG name the sysop chose. A bridge echo starts
while the original is still on air, so the hold only has to cover very
short overs: default 1.0 s, and the validator rejects 2.0 s or more.
The runtime falls back to the default for anything out of range.
- LOOP_GUARD: false stays as an extra option.
- Test: a replay starting 2.0 s after the over ends is not flagged even
with LOOP_GUARD: true on that system.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JnjAx7G27d9q88j9PFLj11
- application/routing/loop_guard.py: when a group voice stream starts,
the same rf_src on the same TG from another ingress (system + peer; an
OBP link is one ingress), with another stream ID, while the original is
active or ended < GLOBAL.LOOP_GUARD_HOLD (2 s), is a loop echo.
Stream-ID LoopControl cannot see it (bridges start a new stream ID) and
the TG-busy check lets the same caller through on purpose.
- Bridges that relay to another TG are left alone (same TG required).
- Per system LOOP_GUARD: "log" (default, one *LoopGuard* line per echo),
true (drop), false (neither checked nor recorded: the ECHO parrot).
- Called from HBP and OBP ingress and from the MASTER local REPEAT, one
verdict per stream cached by stream ID; bound streams never reach it.
Server voice IDs and plugin/announcement frames are exempt.
- Validator: LOOP_GUARD true/false/"log", LOOP_GUARD_HOLD > 0.
- Docs: configuration.md and behaviour-and-timers.md (en/es); the example
ECHO sets LOOP_GUARD: false.
Closes#118
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JnjAx7G27d9q88j9PFLj11
* fix: raise UDP SO_RCVBUF on voice listeners (C-LOCAL)
Apply a 4 MB receive buffer on system and proxy UDP sockets to reduce
kernel RcvbufErrors under OBP load; size is configurable via GLOBAL.UDP_RCVBUF.
* fix: exclude byte-identical duplicates from HBP rate counter (A.2)
Check lastData before incrementing the ingress packet counter so compressed
duplicate bursts do not trigger legitimate RATE DROP on call start.
* fix: duplicate-safe TA embed phase and REPEAT VHEAD DMRA (B)
Ignore byte-identical B-E embed bursts so duplicate uplinks do not desync the
TA phase machine, and re-emit DMRA on every VHEAD on the REPEAT path.
* chore: document echo point-to-point and logged_in reconciliation (EN/ES)
Document multi-hotspot echo/service delivery via RX_PEER, the lst_seen
logged_in reconcile loop, and cross-links between user and dev guides.
Add a dedicated public page explaining routing, contention, timers,
SINGLE mode, static/dynamic TGs, silent activation, slot mapping and
OBP parity. Cross-link from behaviour-and-timers, bridges-and-talkgroups,
introduction and special-numbers, and register it in both mkdocs navs.
Document 2.x routing model with a concrete TG 52090 example, operator-focused
performance gains (indexes, reporting, integrated proxy), and MkDocs Mermaid
support across server and monitor architecture pages.
Document ADN-report-proxy so adn-server 2.x can feed v1 legacy monitors,
and sync EN/ES MkDocs with DATABASE, dynamic TGs, monitoring, and related topics.
Replace internal bridge terminology with routing (RoutingUseCases, AclRouter,
routing_table export). SubscriptionStore remains runtime authority with O(1)
indexes for router and downlink filters. Fix STATIC TG parity on OPTIONS/RPTO,
parrot in-band edge cases, and remove per-packet routing_table export from the
hot path that caused high CPU under multi-hotspot OBP load.