Repository navigation
Thermometer prepared send v0 - #9
Merged
Merged
Conversation
Remove the git submodule so the prepared-send integration tracks the branch with prepared packet support. Co-authored-by: Cursor <cursoragent@cursor.com>
…hernetio/temperature-sensor into thermometer-prepared-send-v0
NikolayChirkov
force-pushed
the
thermometer-prepared-send-v0
branch
from
August 27, 2026 22:58
865b068 to
a02a845
Compare
BartolomeyKant
force-pushed
the
thermometer-prepared-send-v0
branch
from
August 28, 2026 09:19
0b6b28a to
865b068
Compare
Harness measures init_to_release_ms; control run keeps SPIFFS state and proves Connect(nullopt) restores ~3.5s cycles on ESP32-C6. Co-authored-by: Cursor <cursoragent@cursor.com>
Hardware median ~3.6s for 10 FULL cycles without SPIFFS erase; reboot 3 cycles stay in the same band. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Extract SendPreparedOnce for shared hot-path encoding, add ESP harness (AE_EXP_PREPARED_MESSAGE_E2E), desktop receiver, and document the ESP32-C6 coordinated run with timings in experiments/.
Split runtime Wi-Fi cleanup from cache invalidation, use BSSID/channel/static-IP fast connect, and measure 5 FULL + 100 prepared sends via binary Æther payloads with no UART.
Export Wi-Fi/IP/gateway-MAC from the live FULL connection so first prepared sends stay fast; install static ARP before sendto. Hardware 5x20 still shows poor delivery (29/100); hold=600 control did not help. Co-authored-by: Cursor <cursoragent@cursor.com>
Prove losses track per-message Wi-Fi teardown/reassociation, not prepared packet path. Co-authored-by: Cursor <cursoragent@cursor.com>
Fix harness hang (4MB flash, deferred work after Update, no nested Write) so B1 smoke and the full silent one-factor table can complete. Co-authored-by: Cursor <cursoragent@cursor.com>
Measures ESP32-C6 reconnect cycles (channel+static IP+ARP baseline), finds WPA2 SAE-off + PRE/POST 200/300 as the most reliable long-run path (198/200 at 710ms vs ~850ms), and documents callback/auth/delay results without changing production SendPreparedOnce. Co-authored-by: Cursor <cursoragent@cursor.com>
Register esp_wifi_set_tx_done_cb immediately before sendto, wait first callback (no fingerprint), drop fixed POST=300; PRE=25 shortest screen. Co-authored-by: Cursor <cursoragent@cursor.com>
PRE=10/0 fail delivery on N50; keep late TX-done callback and PRE=25 winner. Co-authored-by: Cursor <cursoragent@cursor.com>
AE_EXP_PREPARED_TX_DONE_DIAG builds skipped ExperimentEarlyAppEntry, so every wake looked like a non-deepsleep reset and forced FULL. Add FullReason/boot snapshots, clamp hot_index on Final, and document the loop diagnosis. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
One-flash MRT1 deep-sleep run calls esp_wifi_internal_set_retry_counter on the hot path only; CONTROL leaves it unset. Results do not confirm that RF/tx_done tails are driven by that counter. Co-authored-by: Cursor <cursoragent@cursor.com>
Measure D/G/H/E runtime knobs (30 HOT each) plus combined winners validation without changing TX power or MAC retry policy. Co-authored-by: Cursor <cursoragent@cursor.com>
Wire AE_EXP_PREPARED_AP_AETHERNETIO_3X10 bench (D1 path, 1 s deep sleep), extend DsPayload with association fields, and add orchestrator plus report from the first hardware run on SSID aethernetio. Co-authored-by: Cursor <cursoragent@cursor.com>
Desktop RX prints RUN/seq/recv/lost/loss%/dup with reset detection; ESP FULL path uses monotonic uint32 sequence and silent tele/sdkconfig. Co-authored-by: Cursor <cursoragent@cursor.com>
dmitriyk-aethernet
self-requested a review
September 11, 2026 17:38
dmitriyk-aethernet
previously approved these changes
Sep 18, 2026
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
3 participants
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.
Adds the battery thermometer's prepared-send path: full startup exports a prepared message block, subsequent wakeups send the temperature using that block, and failures fall back to full Aether startup. Includes the thermometer's sensor, ULP, sleep, and peripheral power changes.
The PR contains only thermometer functionality and its regression tests. Benchmark firmware, experiment runners, reports, captured results, diagnostic receivers, and experiment-specific SDK configuration overlays have been removed. Both CMake files now select the thermometer sources directly and expose only its runtime configuration settings; prepared_send no longer depends on benchmark payloads or diagnostic APIs.
Validation: ESP32-C6 firmware configured, compiled, linked, and passed the partition size check in the local ESP-IDF 6.1 environment. The STCC4 host regression test passed. Hardware behavior has not been retested for this cleanup.