dangernoodle-io / TaipanMiner / 28695847962

90%

push

github

feat(sse): SSE 1-connection cap on no-PSRAM boards + drop dead pool.notify topic (#616)

* feat(sse): SSE 1-connection cap on no-PSRAM boards (B1-561)

Per the B1-561 SSE-first-class design: cap CONFIG_BB_EVENT_ROUTES_MAX_CLIENTS
to 1 (from the fleet default 2) on the four no-PSRAM boards (c3-supermini,
s2-mini, wroom32, tdongle-s3). None of the four boards enable CONFIG_SPIRAM.

Verified CONFIG_MQTT_TASK_STACK_SIZE=5120 and CONFIG_BB_PUB_WORKER_STACK=8192
are pre-existing, deliberate, HW-verified-margin values (sdkconfig.defaults
comments) — left unchanged. CONFIG_MBEDTLS_DYNAMIC_BUFFER=y is already set
fleet-wide.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* refactor(sse): drop dead pool.notify topic, retain block.found (B1-561)

pool.notify had zero producers (audit-confirmed; stratum posts
mining.notify separately) -- delete the attach + topic var, frees ~2.2 KB.
Flip block.found to retained: it's a rare discrete event where
last-found-block-on-reconnect is a strict improvement over empty state,
frees ~2 KB.

Verified the periodic bb_sink_event topics (sys.mem, net.detail, fan,
power, thermal) are already registered retained=true with
bb_sink_event_seed_all() publishing the required initial snapshot right
after bb_pub_add_sink succeeds -- no change needed there.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

2876 of 3613 branches covered (79.6%)

Branch coverage included in aggregate %.

5387 of 5543 relevant lines covered (97.19%)

149015.13 hits per line