[tfluna] Recover from transient runtime reconfiguration failures by fallenmi · Pull Request #20992 · esphome/esphome

Igor Stadnyk (fallenmi) · 2026-10-17T14:42:07Z

What does this implement/fix?

Keep TF-Luna usable after a transient I2C failure while reconfiguring after a factory reset. Runtime configuration failures now leave a warning and retry on the next update instead of permanently marking the component failed. Initial setup failures remain permanent.

Also check reset/restart writes before following them up, skip runtime writes on failed components, and clear the warning only after receiving a new measurement. The existing 100 ms factory-reset delay is retained. This does not guarantee persistence to flash after a failed save.

The shared Application lifecycle/test-harness changes were split into #21001 and merged separately, as requested. This PR now changes only the two TF-Luna production files and adds its 16 host tests. Existing maintainer merge history is preserved with a normal merge, without rebasing or force-pushing.

The three LOG_STR conversions overlap the existing cleanup in #20864; they are not claimed as a separate new fix.

Verification and limitations

The original regression A/B on base 1b4721d is retained: with the same test harness, reverting only the two TF-Luna production files changes 16/16 passing to 7/16; restoring the fix returns 16/16.
Independently renewed against dev c9ba6c3: TF-Luna 16/16 and combined TF-Luna/EmonTx/i2c 66/66 pass with ASAN/UBSAN.
Paired full host runs on that exact C9 base: pristine 2240/2261 pass; candidate 2256/2277 pass, including all 16 added tests. Both have the same 21 qualified failure names and 11 normalized UBSAN sites. This is NOT a fully green run. These are macOS runs in a network-denied sandbox.
Scoped native HOST clang-tidy 22.1.8, using the actual full-host compile database, passes for TF-Luna production/header and the test translation unit. In the separate no-sensor host configuration, three dead-store warnings also reproduce on the exact original C9 production/header; they are not reported as a passing no-sensor lint run. Checks and WarningsAsErrors remain unchanged; no new suppressions.
Official ci-custom passes on the three changed paths. Native ARM clang-format 13.0.1 dry-run passes on those same unchanged bytes. The normal unmodified pinned prek run has four passing hooks and six automatic no-matching-file skips; its bundled clang-format fails to execute on this ARM host (Bad CPU type). That normal run is not claimed fully green, and its hook binary/config was not replaced.
On this candidate, the normal component configuration driver passes for ESP32-IDF, ESP8266-Arduino and RP2040-Arduino. Normal C++ source generation also passes for each platform's generated test wrapper. No full firmware build or hardware testing.
Local runtime/lint results above cover C9, not every later dev commit. A clean native merge against dev 89afb11 was checked separately; the resulting diff is still exactly these three TF-Luna paths. Current CI for this updated head remains a separate check.

AI assistance: prepared and refined with Claude Code; independently reviewed, tested and integrated with Codex. No claim of manual hardware validation.

Types of changes

Bugfix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
New developer-facing feature (adds functionality for component developers; no end-user configuration change)
Breaking change (fix or feature that would cause existing functionality to not work as expected) — policy
Developer breaking change (an API change that could break external components) — policy
Undocumented C++ API change (removal or change of undocumented public methods that lambda users may depend on) — policy
Code quality improvements to existing code or addition of tests
Other

Related issue or feature (if applicable):

No existing issue found in the scoped public searches. Related cleanup noted above is not the runtime recovery fix.

Pull request in esphome.io with documentation (if applicable):

Not applicable; no new configuration options.

Pull request in developers.esphome.io with developer documentation (if applicable):

Not applicable.

Test Environment

Native macOS host with ASAN/UBSAN; platform configuration/source-generation checks only. The hardware boxes below are intentionally unchecked.

Example entry for `config.yaml`:

Existing configuration is unchanged; the existing test fixture includes these controls:

tfluna:
  - i2c_id: i2c_bus

button:
  - platform: tfluna
    factory_reset:
      name: "factory reset"
    restart:
      name: "restart"

Checklist:

The code change is tested and works locally.
Tests have been added to verify that the new code works (under tests/ folder).

If user exposed functionality or configuration variables are added/changed:

Documentation added/updated in esphome.io.

Not applicable: no user-facing configuration changes.

Keep setup failures permanent, but retry runtime configuration after a transient I2C failure. Add regression coverage and use one shared Application lifetime in combined host tests. Prepared with Claude Code; independently checked and integrated with Codex.

esphome · 2026-10-17T14:42:17Z

To use the changes from this PR as an external component, add the following to your ESPHome configuration YAML file:

external_components:
  - source: github://pr#20992
    components: [tfluna]
    refresh: 1h

(Added by the PR bot)

esphome · 2026-10-17T14:42:20Z

👋 Hi there! This PR modifies 2 file(s) with codeowners.

Craig Andrews (@candrews) - As codeowner(s) of the affected files, your review would be appreciated! 🙏

Note: Automatic review request may have failed, but you're still welcome to review.

codecov · 2026-10-17T14:45:37Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.95%. Comparing base (c9ba6c3) to head (e3f93b5).
⚠️ Report is 9 commits behind head on dev.

Additional details and impacted files

@@           Coverage Diff           @@
##              dev   #20992   +/-   ##
=======================================
  Coverage   93.95%   93.95%           
=======================================
  Files          90       90           
  Lines       20465    20465           
  Branches     3176     3176           
=======================================
  Hits        19227    19227           
  Misses        947      947           
  Partials      291      291

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

codspeed · 2026-10-17T14:45:55Z

Merging this PR will not alter performance

⚠️

21 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

Scheduler_SetTimeout

test_load_compiled_config_cached

Scheduler_SetInterval

Scheduler_SetTimeout_ExceedPool

Scheduler_Defer

DrawPixelsAt_RGB565_BigEndian

Scheduler_Call_TimersNotDue

Scheduler_Call_NoWork

ApplicationLoop_Empty

Scheduler_Call_5IntervalsFiring

…and 11 more

✅ 210 untouched benchmarks

_{Comparing fallenmi:codex/tfluna-currentbase-oct17 (60a41ad) with dev (635df9c)}

github-actions · 2026-10-17T14:52:57Z

Memory Impact Analysis

Components: tfluna
Platform: esp8266-ard

Metric	Target Branch	This PR	Change
RAM	29,400 bytes	29,292 bytes	📉 ✅ -108 bytes (-0.37%)
Flash	272,407 bytes	272,547 bytes	📈 🔸 +140 bytes (+0.05%)

📊 Component Memory Breakdown

Component	Target Flash	PR Flash	Change
`[esphome]tfluna`	1,199 bytes	1,429 bytes	📈 🚨 +230 bytes (+19.18%)
`app_framework`	2,526 bytes	2,522 bytes	📉 -4 bytes (-0.16%)

🔍 Symbol-Level Changes (click to expand)

Changed Symbols

Symbol	Target Size	PR Size	Change
`esphome::tfluna::TFLuna::setup()`	180 bytes	49 bytes	📉 -131 bytes (-72.78%)
`esphome::tfluna::TFLuna::update()`	65 bytes	113 bytes	📈 +48 bytes (+73.85%)
`esphome::tfluna::TFLuna::restart()`	48 bytes	59 bytes	📈 +11 bytes (+22.92%)
`esphome::tfluna::TFLuna::factory_reset()`	97 bytes	105 bytes	📈 +8 bytes (+8.25%)
`esphome::tfluna::TFLuna::read_data_timeout_()`	104 bytes	100 bytes	📉 -4 bytes (-3.85%)
`setup`	735 bytes	731 bytes	📉 -4 bytes (-0.54%)
`esphome::tfluna::TFLuna::read_data_()`	224 bytes	220 bytes	📉 -4 bytes (-1.79%)
`std::_Function_handler<void (), esphome::tfluna::TFLuna::factory_reset()::{lambda()#1}>::_M_manag...` `std::_Function_handler<void (), esphome::tfluna::TFLuna::factory_reset()::{lambda()#1}>::_M_manager(std::_Any_data&, std::_Function_handler<void (), esphome::tfluna::TFLuna::factory_reset()::{lambda()#1}> const&, std::_Manager_operation)`	17 bytes	20 bytes	📈 +3 bytes (+17.65%)
`std::_Function_handler<void (), esphome::tfluna::TFLuna::factory_reset()::{lambda()#1}>::_M_invok...` `std::_Function_handler<void (), esphome::tfluna::TFLuna::factory_reset()::{lambda()#1}>::_M_invoke(std::_Any_data const&)`	9 bytes	8 bytes	📉 -1 bytes (-11.11%)

New Symbols (top 15)

Symbol	Size
`esphome::tfluna::TFLuna::configure_()`	163 bytes
`esphome::tfluna::TFLuna::factory_reset()::__pstr__`	57 bytes
`esphome::tfluna::TFLuna::reconfigure_()`	53 bytes
`esphome::tfluna::TFLuna::read_data_timeout_()::__pstr__`	27 bytes
`esphome::tfluna::TFLuna::configure_()::__pstr__`	21 bytes

Removed Symbols (top 15)

Symbol	Size
`esphome::tfluna::TFLuna::setup()::__pstr__`	21 bytes

Note: This analysis measures static RAM and Flash usage only (compile-time allocation).
Dynamic memory (heap) cannot be measured automatically.
⚠️ You must test this PR on a real device to measure free heap and ensure no runtime memory issues.

This analysis runs automatically when components change. Memory usage is measured from a representative test configuration.

ESPHome Bot (esphbot) · 2026-10-17T17:09:50Z

~~Previous review~~ — superseded by a newer review below.

ESPHome Bot (esphbot)

Tip

No blocking issues found — ready to merge.

Copilot

🔵 Needs a closer look

Shared Application lifetime changes affect all host suites, and reported suite failures lack a clean matching-base comparison.

No open issues 🎉

What changed in this PR

Improves TF-Luna recovery from transient runtime reconfiguration failures while preserving permanent setup failures.

Changes:

Retries runtime configuration and retains warnings until a new measurement arrives.
Adds 16 host tests covering recovery and failure handling.
Centralizes shared Application construction and updates affected test fixtures.

File	Description
tests/components/xiaomi_body_scale/common.h	Removes fixture-local Application construction.
tests/components/tmp102/test_tmp102.cpp	Pairs Application destruction and reconstruction.
tests/components/tfluna/test_tfluna.cpp	Adds runtime recovery regression tests.
tests/components/snapshot/snapshot_test.cpp	Uses centrally constructed Application.
tests/components/pd_pioneer_ir/pd_pioneer_ir_tests.cpp	Removes duplicate Application construction.
tests/components/mistral_ir/mistral_ir_tests.cpp	Removes duplicate Application construction.
tests/components/main.cpp	Constructs Application before tests run.
tests/components/i2c/test_i2c_chip_id.cpp	Destroys Application before replacing it.
tests/components/haier_water_heater/water_heater/test_hon_controller.cpp	Removes local construction and unused include.
tests/components/filter_lifetime/test_filter_lifetime.cpp	Preserves Application lifetime between tests.
tests/components/emontx/common.h	Makes the legacy construction helper a no-op.
tests/components/display/test_test_card.cpp	Removes redundant fixture setup.
tests/components/daikin/daikin_tests.cpp	Removes local construction and unused include.
tests/components/daikin/daikin_regression_tests.cpp	Removes local construction and unused include.
esphome/components/tfluna/tfluna.h	Declares reconfiguration helpers and retry state.
esphome/components/tfluna/tfluna.cpp	Separates setup failures from recoverable runtime failures.

_{Review effort: Balanced}

J. Nick Koston (bdraco) · 2026-10-17T17:12:18Z

The shared test-harness change belongs in its own PR
tests/components/main.cpp:35-38

Please split this PR

esphome · 2026-10-17T21:38:46Z

This pull request has merge conflicts. Please merge the latest dev branch into your branch to resolve them.

ESPHome Bot (esphbot) · 2026-10-18T00:24:31Z

J. Nick Koston (@bdraco): > 2. The shared test-harness change belongs in its own PR

Split done. Harness landed in #21001. Diff now three TF-Luna paths. construct_app() no-op removed.

ESPHome Bot (esphbot) · 2026-10-18T00:24:35Z

PR Review — [tfluna] Recover from transient runtime reconfiguration failures

Mergeable. Two non-blocking suggestions remain.

The PR is now split as J. Nick Koston (@bdraco) asked: the shared App harness work landed in #21001, so this diff is just the two TF-Luna production files plus their tests. The no-op construct_app() from the previous review is gone too. The fix itself is sound:

configure_() is shared by two paths. A boot failure is permanent (mark_failed). A runtime failure leaves a warning and sets reconfigure_pending_, and the next update() retries.
factory_reset() clears the pending flag only after the reset write succeeds.
The is_failed() guards keep the buttons and update() off the bus once the component is dead.
The RegisterBus fake injects one NACK at each configuration step. Tests cover a double press, a reset while a reconfigure is pending, and a reset NACK while one is pending.

Suggestions:

A retry after a SAVE NACK sees trigger mode already set and skips the save. A later Restart or hung-device reboot then loses trigger mode, and the SAVE test does not check save_writes.
StagedForeignTimeoutBlocksOwnership tests fixture plumbing, not TF-Luna.

✅ Resolved since last review (1)

Previously-flagged issues verified fixed

tests/components/emontx/common.h:47 Remove the no-op construct_app() instead of keeping it

Checklist

No hardcoded secrets
Error paths handled (I2C NACKs at boot vs runtime)
New branches covered by tests
Test isolation / shared state
Scope matches PR description
Tests target component behaviour
Repo C++ conventions (LOG_STR, this->, constexpr ids)

ℹ️ Triage summary

2 pre-existing finding(s) on unchanged code suppressed (freeze).

Automated review by Kōan (Claude) HEAD=e3f93b5 36s

ESPHome Bot (esphbot)

Tip

No blocking issues found — ready to merge.

Copilot

🟢 Approval recommended

The focused recovery changes have regression coverage and no unresolved blocking findings; firmware and hardware validation remain unverified.

No open issues 🎉

_{Review effort: Balanced}

J. Nick Koston (bdraco)

Thanks Igor Stadnyk (@fallenmi)

esphome Bot added bugfix code-quality component: tfluna has-tests medium-pr PR < 100 lines labels Oct 17, 2026

Merge branch 'dev' into codex/tfluna-currentbase-oct17

60a41ad

J. Nick Koston (bdraco) requested a balanced review from Copilot October 17, 2026 17:08

Copilot started reviewing on behalf of J. Nick Koston (bdraco) October 17, 2026 17:08

View session

ESPHome Bot (esphbot) approved these changes Oct 17, 2026

View reviewed changes

Copilot AI reviewed Oct 17, 2026

View reviewed changes

Igor Stadnyk (fallenmi) mentioned this pull request Oct 17, 2026

[core] Construct App once in the C++ unit test main #21001

Merged

4 of 20 tasks

esphome Bot added the has-conflicts label Oct 17, 2026

Merge dev after extracting TF-Luna test harness (esphome#21001)

e3f93b5

J. Nick Koston (bdraco) requested review from ESPHome Bot (esphbot) and a balanced review from Copilot October 18, 2026 00:22

Copilot started reviewing on behalf of J. Nick Koston (bdraco) October 18, 2026 00:23

View session

ESPHome Bot (esphbot) approved these changes Oct 18, 2026

View reviewed changes

Copilot AI reviewed Oct 18, 2026

View reviewed changes

J. Nick Koston (bdraco) approved these changes Oct 18, 2026

View reviewed changes

J. Nick Koston (bdraco) merged commit bb9067a into esphome:dev Oct 18, 2026
49 checks passed

J. Nick Koston (bdraco) mentioned this pull request Oct 18, 2026

[components] Use Component::status_comm_result for communication failures #20864

Merged

2 of 19 tasks

github-actions Bot locked and limited conversation to collaborators Oct 19, 2026

Conversation

Igor Stadnyk (fallenmi) commented Oct 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does this implement/fix?

Types of changes

Test Environment

Example entry for config.yaml:

Checklist:

Uh oh!

esphome Bot commented Oct 17, 2026

Uh oh!

esphome Bot commented Oct 17, 2026

Uh oh!

codecov Bot commented Oct 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

codspeed Bot commented Oct 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Uh oh!

github-actions Bot commented Oct 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Memory Impact Analysis

Changed Symbols

New Symbols (top 15)

Removed Symbols (top 15)

Uh oh!

ESPHome Bot (esphbot) commented Oct 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ESPHome Bot (esphbot) left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🔵 Needs a closer look

Uh oh!

J. Nick Koston (bdraco) commented Oct 17, 2026

Uh oh!

esphome Bot commented Oct 17, 2026

Uh oh!

ESPHome Bot (esphbot) commented Oct 18, 2026

Uh oh!

ESPHome Bot (esphbot) commented Oct 18, 2026

PR Review — [tfluna] Recover from transient runtime reconfiguration failures

✅ Resolved since last review (1)

Checklist

Uh oh!

ESPHome Bot (esphbot) left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟢 Approval recommended

Uh oh!

J. Nick Koston (bdraco) left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Igor Stadnyk (fallenmi) commented Oct 17, 2026 •

edited

Loading

Example entry for `config.yaml`:

codecov Bot commented Oct 17, 2026 •

edited

Loading

codspeed Bot commented Oct 17, 2026 •

edited

Loading

github-actions Bot commented Oct 17, 2026 •

edited

Loading

ESPHome Bot (esphbot) commented Oct 17, 2026 •

edited

Loading