feat(v0.13.0): per-target KWP response deadline (1s default, 3s slow) - #153
Merged
Conversation
Slice 2 of v0.13.0 'Real Reads, Real Long' (docs/v0.13.0_plan.md).
The K+DCAN transport's request() has hardcoded a 1s response
deadline since v0.1.0. That deadline times out on every E-series
owner's first fault read: real CIC (0x01) and CAS (0x40) modules
take 1.5-3s to answer their first KWP frame after fast-init because
their boot-up sequence is longer than the DME's. Every other KWP
tool uses 3s for these targets; beemuu silently fails.
The fix:
- Add DEFAULT_RESPONSE_DEADLINE (1s) and SLOW_RESPONSE_DEADLINE
(3s) constants at the top of the file.
- Add a slow_modules: HashSet<u8> field to KdcanTransport,
populated from default_slow_modules() in open() — initially
just CIC and CAS, matching what every other KWP tool does.
- Factor the per-target lookup into resolve_deadline(slow, target)
so the resolution logic is unit-testable without a real serial
port.
- Replace the hardcoded 1s in request() with the per-target lookup.
- Expose deadline_for() and set_slow_module() as the public API
for future Settings UI to add more slow modules as users
discover them.
Tier A + Tier B (touches src-tauri/src/transport/kdcan.rs).
Single PR. The flag at the top of the PR body flags the protected
path; wait for human merge per CLAUDE.md \u00a71.
Tests (4 new, all inline in src-tauri/src/transport/kdcan.rs::tests):
- default_slow_modules_contains_cic_and_cas
Asserts the default list contains CIC (0x01) and CAS (0x40)
but NOT DME (0x12) \u2014 otherwise every healthy DME read would
wait the full 3 seconds.
- resolve_deadline_picks_slow_for_known_targets
Asserts CIC/CAS get SLOW_RESPONSE_DEADLINE, DME/unknown get
DEFAULT_RESPONSE_DEADLINE.
- resolve_deadline_with_empty_set_uses_default_for_all
The safe fallback \u2014 empty slow-modules list means every
address gets the default deadline.
- deadline_constants_are_what_e_series_needs
Locks in the actual values (1s, 3s) so a silent change to
either constant is caught at test time.
No frontend changes \u2014 this slice is invisible until a user reads
faults on an E-series. What was a TimeoutError on every first
fault read now succeeds.
Verification:
- cargo test --lib kdcan: 4 new tests pass.
- cargo test (full): 128 + 1 + 4-ignored = all green (was 124,
now 128; +4 from this slice).
- cargo test --test async_commands: still passes (we didn't add
any new #[tauri::command]).
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 62e616a494
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| /// on real E-series cars. Add new entries here as more slow modules are | ||
| /// discovered — but only after a real-car timeout on the default 1 s. | ||
| fn default_slow_modules() -> std::collections::HashSet<u8> { | ||
| [0x01 /* CIC */, 0x40 /* CAS */].into_iter().collect() |
There was a problem hiding this comment.
When reading CIC, this override will not select the raised deadline because the repository's canonical address model identifies 0x01 as ACSM (src-tauri/src/data/ecus.rs:36, also reflected by the ACSM fixture in transport/sim.rs:98), not CIC. Consequently this adds the three-second deadline to ACSM requests while CIC requests retain the one-second default, leaving one of the two module timeouts this change is intended to fix unresolved; configure the actual CIC diagnostic target instead.
Useful? React with 👍 / 👎.
ohgeeceee
pushed a commit
that referenced
this pull request
Jul 23, 2026
…sion-sync rule to CLAUDE.md README.md badge line 18 was frozen at v0.6.0 even though v0.7.0 through v0.14.0 have all shipped (or in v0.14.0's case, had their plan merged). Three changes: - README.md: bump release badge to v0.14.0, point at CHANGELOG.md. - CHANGELOG.md: backfill [0.12.0] (Fault Memory cycle, PRs #143-#149 plus the async fix), [0.13.0] (Real Reads Real Long, PRs #150-#153 including the plan correction), [0.14.0] (Live CAN plan only, PR #156). Existing [0.11.0] and older sections untouched. - CLAUDE.md: add golden rule #6 — every release PR must bump both the README badge and the CHANGELOG section in lockstep, so the badge can't drift again. Plan-only cycles use CHANGELOG `### Planned` and leave the README badge alone. pytest backend/tests/ → 151/151 green. No code change, so cargo test and the JS suites are unaffected.
github-actions Bot
pushed a commit
that referenced
this pull request
Jul 31, 2026
…moved to ✅ Done (#200) Six items in the v0.3.0 'Real Car' historical section were marked 🟢 Ready but have actually shipped: | Item | Shipped in | Reference | |-----------------------|-----------------|--------------------------| | KWP2000 slow timeout | v0.13.0 (PR #153) | commit fd9efc2 | | ISO-TP multi-frame | v0.14.x | src-tauri/src/transport/isotp.rs (enforced by CLAUDE.md, refreshed in PR #198) | | Dark/light theme | v0.7.0 (PR #109) | commit afefc32 | | Gauge theming | v0.7.0 (PR #109) | commit afefc32 | | Save/load workspace | v0.7.0 (PR #109) | commit afefc32 | | Export PNG/SVG charts | v0.11.0 (PR #136) | commit 7f92ccb | This is the same doc-rot pattern that PR #198 fixed in CLAUDE.md: items the inventory still claims as 'Ready to land' have actually shipped, but nobody re-tagged the historical section afterward. Same 'data over invention' fix as the CLAUDE.md refresh. What changed: - '⭐ Protocol & Transport' table — kept the four genuinely-🟡 items (ENET/DoIP, BLE, WiFi, CAN-bus listener), removed the two stale 🟢 Ready items. - 'UI / UX' table — kept the only genuinely-🟡 item (Mobile- responsive), removed the five stale 🟢 Ready items. - New '✅ Protocol, Transport, UI/UX — historical (shipped)' table — pins the six shipped items with their PR references and code locations, following the v0.3.0 'Decode Functions' historical pattern that already exists just above. What this PR does NOT change: - No CHANGELOG.md edits (the shipped-PR references are in CHANGELOG.md already) - No CLAUDE.md edits (the ISO-TP / keepalive / read_vin references there are already accurate post PR #198) - No code edits - No test edits (markdown only) Verified locally (docs-only, no test diff expected): - node --test src/js/**/*.test.js src/js/**/*.test.cjs: 226/226 - python -m pytest backend/tests/ -q: 166/166 - cd src-tauri && cargo test --lib --offline: 149/149 - npm run build: rc=0, both BeeEmUu_0.14.3 bundles built Tier A docs-only per CLAUDE.md. Self-merge on CI green. Cross-references: - PR #198: 'docs: refresh CLAUDE.md hardware/timing invariants' (the same doc-rot sweep on CLAUDE.md) - PR #183: 'docs(proposal): fix stale NOT YET IMPLEMENTED claims' (proposed similar fixes but never applied) Co-authored-by: ohgeeceee <ohgeeceee@users.noreply.github.com>
9 tasks
github-actions Bot
pushed a commit
that referenced
this pull request
Jul 31, 2026
…moved to ✅ Done (#205) Six items in the v0.3.0 'Real Car' historical section were marked 🟢 Ready but have actually shipped: | Item | Shipped in | Reference | |-----------------------|-----------------|--------------------------| | KWP2000 slow timeout | v0.13.0 (PR #153) | commit fd9efc2 | | ISO-TP multi-frame | v0.14.x | src-tauri/src/transport/isotp.rs (enforced by CLAUDE.md, refreshed in PR #198) | | Dark/light theme | v0.7.0 (PR #109) | commit afefc32 | | Gauge theming | v0.7.0 (PR #109) | commit afefc32 | | Save/load workspace | v0.7.0 (PR #109) | commit afefc32 | | Export PNG/SVG charts | v0.11.0 (PR #136) | commit 7f92ccb | This is the same doc-rot pattern that PR #198 fixed in CLAUDE.md: items the inventory still claims as 'Ready to land' have actually shipped, but nobody re-tagged the historical section afterward. Same 'data over invention' fix as the CLAUDE.md refresh. What changed: - '⭐ Protocol & Transport' table — kept the four genuinely-🟡 items (ENET/DoIP, BLE, WiFi, CAN-bus listener), removed the two stale 🟢 Ready items. - 'UI / UX' table — kept the only genuinely-🟡 item (Mobile- responsive), removed the five stale 🟢 Ready items. - New '✅ Protocol, Transport, UI/UX — historical (shipped)' table — pins the six shipped items with their PR references and code locations, following the v0.3.0 'Decode Functions' historical pattern that already exists just above. What this PR does NOT change: - No CHANGELOG.md edits (the shipped-PR references are in CHANGELOG.md already) - No CLAUDE.md edits (the ISO-TP / keepalive / read_vin references there are already accurate post PR #198) - No code edits - No test edits (markdown only) Verified locally (docs-only, no test diff expected): - node --test src/js/**/*.test.js src/js/**/*.test.cjs: 226/226 - python -m pytest backend/tests/ -q: 166/166 - cd src-tauri && cargo test --lib --offline: 149/149 - npm run build: rc=0, both BeeEmUu_0.14.3 bundles built Tier A docs-only per CLAUDE.md. Self-merge on CI green. Cross-references: - PR #198: 'docs: refresh CLAUDE.md hardware/timing invariants' (the same doc-rot sweep on CLAUDE.md) - PR #183: 'docs(proposal): fix stale NOT YET IMPLEMENTED claims' (proposed similar fixes but never applied) Co-authored-by: ohgeeceee <ohgeeceee@users.noreply.github.com>
ohgeeceee
added a commit
that referenced
this pull request
Jul 31, 2026
…e.rs (#201) * docs(roadmap): audit v0.3.0 historical section — stale 🟢 Ready items moved to ✅ Done Six items in the v0.3.0 'Real Car' historical section were marked 🟢 Ready but have actually shipped: | Item | Shipped in | Reference | |-----------------------|-----------------|--------------------------| | KWP2000 slow timeout | v0.13.0 (PR #153) | commit fd9efc2 | | ISO-TP multi-frame | v0.14.x | src-tauri/src/transport/isotp.rs (enforced by CLAUDE.md, refreshed in PR #198) | | Dark/light theme | v0.7.0 (PR #109) | commit afefc32 | | Gauge theming | v0.7.0 (PR #109) | commit afefc32 | | Save/load workspace | v0.7.0 (PR #109) | commit afefc32 | | Export PNG/SVG charts | v0.11.0 (PR #136) | commit 7f92ccb | This is the same doc-rot pattern that PR #198 fixed in CLAUDE.md: items the inventory still claims as 'Ready to land' have actually shipped, but nobody re-tagged the historical section afterward. Same 'data over invention' fix as the CLAUDE.md refresh. What changed: - '⭐ Protocol & Transport' table — kept the four genuinely-🟡 items (ENET/DoIP, BLE, WiFi, CAN-bus listener), removed the two stale 🟢 Ready items. - 'UI / UX' table — kept the only genuinely-🟡 item (Mobile- responsive), removed the five stale 🟢 Ready items. - New '✅ Protocol, Transport, UI/UX — historical (shipped)' table — pins the six shipped items with their PR references and code locations, following the v0.3.0 'Decode Functions' historical pattern that already exists just above. What this PR does NOT change: - No CHANGELOG.md edits (the shipped-PR references are in CHANGELOG.md already) - No CLAUDE.md edits (the ISO-TP / keepalive / read_vin references there are already accurate post PR #198) - No code edits - No test edits (markdown only) Verified locally (docs-only, no test diff expected): - node --test src/js/**/*.test.js src/js/**/*.test.cjs: 226/226 - python -m pytest backend/tests/ -q: 166/166 - cd src-tauri && cargo test --lib --offline: 149/149 - npm run build: rc=0, both BeeEmUu_0.14.3 bundles built Tier A docs-only per CLAUDE.md. Self-merge on CI green. Cross-references: - PR #198: 'docs: refresh CLAUDE.md hardware/timing invariants' (the same doc-rot sweep on CLAUDE.md) - PR #183: 'docs(proposal): fix stale NOT YET IMPLEMENTED claims' (proposed similar fixes but never applied) * feat(v0.14.4): story coverage — 52 unit tests for story.rs + anonymize.rs Two user-facing Rust modules have shipped with **zero unit tests** since the diagnostic-story + secure-snapshot-share features landed: - `src-tauri/src/story.rs` (350 LOC) — the Generate Story pipeline that powers the one-click mechanic narrative modal in `src/index.html:513` + `src/js/main.js:2961` (renderStory). - `src-tauri/src/anonymize.rs` (113 LOC) — the VIN-stripping layer that powers the Secure Snapshot Share feature (`src/js/main.js:1071`). Both are pure-Rust, deterministic, and snapshot-driven — the right shape for unit tests with fixture-built `SessionSnapshot` inputs. This PR adds 52 unit tests covering: **story.rs (32 tests):** - `Severity::from_str` bucketing (critical / warning / unknown → info). - `Severity` ordering (Critical > Warning > Info). - `priority_for` severity → priority number. - `parse_cost_range` parser: single value, tilde, hyphen, **en-dash** (the TOML files use en-dash — parser must normalise), whitespace, empty, garbage. - `format_vehicle` for empty / VIN-only / mileage-only / decoded. - `build_context` freeze-frame string assembly. - Full `generate` pipeline: - empty snapshot → Info story with no findings. - unknown DTC → generic Info finding. - n55-specific DTC (2A82) → uses engine template. - generic DTC fallback when engine-specific missing. - severity = max of all findings (sorted critical-first). - recommendations sorted ascending by priority. - cost range sums across findings. - cost-max invariant: max >= min + 50. - DTC code case-insensitive lookup (TOML keys are uppercased on load). - summary text counts critical + warning correctly. - title uses manufacturer + VIN prefix. **anonymize.rs (20 tests):** - `hash_vin` properties: 16 hex chars, stable for same input, distinct for distinct inputs, case-sensitive (current behaviour pinned). - `anonymize` pipeline: - VIN never leaks into the anonymized JSON. - VIN → fingerprint via hash_vin. - Missing VIN → "unknown" fingerprint. - engine_family preserved from suggested_profile. - engine_family defaults to "generic". - Modules / DTCs / freeze frames / ident all preserved. - Mileage (mileage_km) stripped (privacy). - Empty modules handled. - fault_count = None → 0 in the output. - recorded_at populated with "(UTC)" suffix. - live_data is always empty (current anonymizer strips it; pinning the decision). - `export_json`: - No VIN leak in pretty JSON. - No mileage leak. - Pretty-printed (multi-line + indented). - Round-trips through serde. ## Verification - `cd src-tauri && cargo test --lib --offline` — **201/201 pass** (149 existing + 52 new). - `cd src-tauri && cargo test --test async_commands --offline` — 1/1 (the CLAUDE.md invariant guard still green — no new sync commands). - `node --test src/js/**/*.test.js src/js/**/*.test.cjs` — 226/226 (no JS diff). - `pytest backend/tests/ -q` — 166/166 (no backend diff). - `npm run build` — rc=0, 2m32s; both BeeEmUu_0.14.3 bundles built. ## Tier **A** — pure additions to existing Rust modules, no `transport/**` / `protocol/**` / `commands.rs` / `.claude/**` touches. Self-merge on CI green per CLAUDE.md rule 2. ## Cross-references - The Story modal UI is `renderStory` in `src/js/main.js:2961`, the button at `src/index.html:513`. - The Secure Snapshot Share wiring is `doSecureShare` in `src/js/main.js:1071`, invoking `anonymize_snapshot` (sync, in the SYNC_ALLOWLIST). - The story knowledge base lives in `community/stories/{generic,n55}.toml`; tests load it via `story::load()`. * fix(ci): handle stacked PRs in autonomous auto-merge job (#204) * Initial plan * fix(v0.14.4): handle stacked PR auto-merge in CI --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> --------- Co-authored-by: ohgeeceee <ohgeeceee@users.noreply.github.com> Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com>
ohgeeceee
added a commit
that referenced
this pull request
Jul 31, 2026
* docs(roadmap): audit v0.3.0 historical section — stale 🟢 Ready items moved to ✅ Done Six items in the v0.3.0 'Real Car' historical section were marked 🟢 Ready but have actually shipped: | Item | Shipped in | Reference | |-----------------------|-----------------|--------------------------| | KWP2000 slow timeout | v0.13.0 (PR #153) | commit fd9efc2 | | ISO-TP multi-frame | v0.14.x | src-tauri/src/transport/isotp.rs (enforced by CLAUDE.md, refreshed in PR #198) | | Dark/light theme | v0.7.0 (PR #109) | commit afefc32 | | Gauge theming | v0.7.0 (PR #109) | commit afefc32 | | Save/load workspace | v0.7.0 (PR #109) | commit afefc32 | | Export PNG/SVG charts | v0.11.0 (PR #136) | commit 7f92ccb | This is the same doc-rot pattern that PR #198 fixed in CLAUDE.md: items the inventory still claims as 'Ready to land' have actually shipped, but nobody re-tagged the historical section afterward. Same 'data over invention' fix as the CLAUDE.md refresh. What changed: - '⭐ Protocol & Transport' table — kept the four genuinely-🟡 items (ENET/DoIP, BLE, WiFi, CAN-bus listener), removed the two stale 🟢 Ready items. - 'UI / UX' table — kept the only genuinely-🟡 item (Mobile- responsive), removed the five stale 🟢 Ready items. - New '✅ Protocol, Transport, UI/UX — historical (shipped)' table — pins the six shipped items with their PR references and code locations, following the v0.3.0 'Decode Functions' historical pattern that already exists just above. What this PR does NOT change: - No CHANGELOG.md edits (the shipped-PR references are in CHANGELOG.md already) - No CLAUDE.md edits (the ISO-TP / keepalive / read_vin references there are already accurate post PR #198) - No code edits - No test edits (markdown only) Verified locally (docs-only, no test diff expected): - node --test src/js/**/*.test.js src/js/**/*.test.cjs: 226/226 - python -m pytest backend/tests/ -q: 166/166 - cd src-tauri && cargo test --lib --offline: 149/149 - npm run build: rc=0, both BeeEmUu_0.14.3 bundles built Tier A docs-only per CLAUDE.md. Self-merge on CI green. Cross-references: - PR #198: 'docs: refresh CLAUDE.md hardware/timing invariants' (the same doc-rot sweep on CLAUDE.md) - PR #183: 'docs(proposal): fix stale NOT YET IMPLEMENTED claims' (proposed similar fixes but never applied) * feat(v0.14.4): story coverage — 52 unit tests for story.rs + anonymize.rs Two user-facing Rust modules have shipped with **zero unit tests** since the diagnostic-story + secure-snapshot-share features landed: - `src-tauri/src/story.rs` (350 LOC) — the Generate Story pipeline that powers the one-click mechanic narrative modal in `src/index.html:513` + `src/js/main.js:2961` (renderStory). - `src-tauri/src/anonymize.rs` (113 LOC) — the VIN-stripping layer that powers the Secure Snapshot Share feature (`src/js/main.js:1071`). Both are pure-Rust, deterministic, and snapshot-driven — the right shape for unit tests with fixture-built `SessionSnapshot` inputs. This PR adds 52 unit tests covering: **story.rs (32 tests):** - `Severity::from_str` bucketing (critical / warning / unknown → info). - `Severity` ordering (Critical > Warning > Info). - `priority_for` severity → priority number. - `parse_cost_range` parser: single value, tilde, hyphen, **en-dash** (the TOML files use en-dash — parser must normalise), whitespace, empty, garbage. - `format_vehicle` for empty / VIN-only / mileage-only / decoded. - `build_context` freeze-frame string assembly. - Full `generate` pipeline: - empty snapshot → Info story with no findings. - unknown DTC → generic Info finding. - n55-specific DTC (2A82) → uses engine template. - generic DTC fallback when engine-specific missing. - severity = max of all findings (sorted critical-first). - recommendations sorted ascending by priority. - cost range sums across findings. - cost-max invariant: max >= min + 50. - DTC code case-insensitive lookup (TOML keys are uppercased on load). - summary text counts critical + warning correctly. - title uses manufacturer + VIN prefix. **anonymize.rs (20 tests):** - `hash_vin` properties: 16 hex chars, stable for same input, distinct for distinct inputs, case-sensitive (current behaviour pinned). - `anonymize` pipeline: - VIN never leaks into the anonymized JSON. - VIN → fingerprint via hash_vin. - Missing VIN → "unknown" fingerprint. - engine_family preserved from suggested_profile. - engine_family defaults to "generic". - Modules / DTCs / freeze frames / ident all preserved. - Mileage (mileage_km) stripped (privacy). - Empty modules handled. - fault_count = None → 0 in the output. - recorded_at populated with "(UTC)" suffix. - live_data is always empty (current anonymizer strips it; pinning the decision). - `export_json`: - No VIN leak in pretty JSON. - No mileage leak. - Pretty-printed (multi-line + indented). - Round-trips through serde. ## Verification - `cd src-tauri && cargo test --lib --offline` — **201/201 pass** (149 existing + 52 new). - `cd src-tauri && cargo test --test async_commands --offline` — 1/1 (the CLAUDE.md invariant guard still green — no new sync commands). - `node --test src/js/**/*.test.js src/js/**/*.test.cjs` — 226/226 (no JS diff). - `pytest backend/tests/ -q` — 166/166 (no backend diff). - `npm run build` — rc=0, 2m32s; both BeeEmUu_0.14.3 bundles built. ## Tier **A** — pure additions to existing Rust modules, no `transport/**` / `protocol/**` / `commands.rs` / `.claude/**` touches. Self-merge on CI green per CLAUDE.md rule 2. ## Cross-references - The Story modal UI is `renderStory` in `src/js/main.js:2961`, the button at `src/index.html:513`. - The Secure Snapshot Share wiring is `doSecureShare` in `src/js/main.js:1071`, invoking `anonymize_snapshot` (sync, in the SYNC_ALLOWLIST). - The story knowledge base lives in `community/stories/{generic,n55}.toml`; tests load it via `story::load()`. * fix(ci): handle stacked PRs in autonomous auto-merge job (#204) * Initial plan * fix(v0.14.4): handle stacked PR auto-merge in CI --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> * Initial plan (#206) Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: ohgeecee <ohjoncurrie@gmail.com> --------- Co-authored-by: ohgeeceee <ohgeeceee@users.noreply.github.com> Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com>
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
1 participant
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.
This PR adds new methods + a new struct field to
KdcanTransportinsrc-tauri/src/transport/kdcan.rs. Per CLAUDE.md §1,transport/is aprotected path because the IPC surface is the trust boundary — any
mistake here becomes a silent data-corruption bug the user only sees
weeks later when their car behaves oddly. The change is small (one
hardcoded constant → a per-target lookup) and well-tested (4 new
inline tests), but the protected-path exposure is real.
Wait for human merge per CLAUDE.md §1. The CI auto-merge workflow
does not apply to Tier B PRs.
What
Per-target response-deadline in
KdcanTransport::request. SlowKWP2000 modules (CIC, CAS, ...) get a raised 3 s deadline; everything
else keeps the historical 1 s default.
Why
The K+DCAN transport has hardcoded a 1 s response deadline since
v0.1.0. That deadline times out on every E-series owner's first
fault read: real CIC (0x01) and CAS (0x40) modules take 1.5–3 s
to answer their first KWP frame after fast-init because their
boot-up sequence is longer than the DME's. Every other KWP tool
uses 3 s for these targets; beemuu silently fails.
This is the 🟢 Ready item on
ROADMAP.mdline 52 that has been opensince v0.3.0.
What changed
src-tauri/src/transport/kdcan.rs— +156 / -2 lines, singlefile:
DEFAULT_RESPONSE_DEADLINE = 1 s(the historical value)SLOW_RESPONSE_DEADLINE = 3 s(matches every other KWP tool)default_slow_modules()helper returning aHashSet<u8>initialized to
{0x01 /* CIC */, 0x40 /* CAS */}.slow_modules: HashSet<u8>field onKdcanTransport,populated in
open().resolve_deadline(slow_modules, target)— free helper factoredout so the lookup is unit-testable without a real serial port.
pub fn deadline_for(&self, target) -> Duration— public APIfor future introspection.
pub fn set_slow_module(&mut self, target, slow) -> bool—hook for a future Settings UI to add user-discovered slow
modules.
request()usesSelf::resolve_deadline(&self.slow_modules, target)instead of the hardcodedDuration::from_millis(1000).mod testsat the bottom of the file.Tests
All inline in
src-tauri/src/transport/kdcan.rs::tests:default_slow_modules_contains_cic_and_casresolve_deadline_picks_slow_for_known_targetsSLOW_RESPONSE_DEADLINE, DME/unknown getDEFAULT_RESPONSE_DEADLINEresolve_deadline_with_empty_set_uses_default_for_alldeadline_constants_are_what_e_series_needsThe 4 tests cover both halves of the contract: the values themselves
(
deadline_constants_are_what_e_series_needs) and the dispatchlogic (
resolve_deadline_*). The "DME should NOT be in the slowlist" assertion in test 1 is the regression guard against the most
likely future mistake — someone adding
0x12to the default listwould degrade every healthy DME read.
Verification
cargo test --lib kdcan: 4 new tests pass.cargo test(full suite): 128 + 1 + 4-ignored = all green(was 124 before this slice → +4 from kdcan tests).
cargo test --test async_commands: still passes (we did not addany new
#[tauri::command]).src-tauri/src/transport/kdcan.rs | 158 ++++++++++++++++++++++++++++++++++++++-with
-2deletions (the originalDuration::from_millis(1000)and
kline_ready: Default::default()lines being replaced) — noEOL churn.
What this slice is NOT
connection; the 1 s deadline is already 10× more than needed.
Per the plan (
docs/v0.13.0_plan.md"What we will NOT do"):no
enet.rschanges in v0.13.0.protocol/change. UDS/KWP request builders don'tmove; the change is purely in the transport deadline.
is invisible until a user reads faults on an E-series; what was
TimeoutErroron every first fault read now succeeds.std::time::Duration+std::collections::HashSet, both already in use.Tier
A + B — touches
src-tauri/src/transport/kdcan.rs. The Rustchange is small and the 4 tests are exhaustive, but the protected
path exposure is real per CLAUDE.md §1. Wait for human merge.
The CI auto-merge workflow does not apply here.
Slice 2 of v0.13.0 — 3/4 cycle slices shipped
Unrelated working-tree noise (not in this PR)
Three files are modified locally but staged only my one intended file:
CLAUDE.md— your own rewrite (Tier C, yours to land)frontend/index.html,frontend/schematics.html— the recurring third-party "network bar" injection (untrusted, not propagated)