Commit Graph

65 Commits

Author SHA1 Message Date
89e6eed105 Bump sync_settings_ui to 0.6.1
Some checks are pending
dependency freshness / dependencies on newest stable (push) Waiting to run
md-naming / markdown naming convention (push) Waiting to run
pre-commit / pre-commit (${{ matrix.group }}) (mypy, mypy, ) (push) Waiting to run
pre-commit / pre-commit (${{ matrix.group }}) (pylint, pylint, ) (push) Waiting to run
pre-commit / pre-commit (${{ matrix.group }}) (rest, , mypy,pylint,free-days-interop) (push) Waiting to run
Tests / test (push) Waiting to run
Tests / free-days-interop (push) Waiting to run
release-apk / release (APK) (push) Waiting to run
repo-contract / repo contract (push) Waiting to run
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-11-08 18:28:23 +01:00
06b70ab312 Bump crdt_sync to v0.12.0
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-11-08 14:24:46 +01:00
537a3a79e4 Retry session refresh; sign next_step
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WzxdWZd3smuxDGmEgXkYP8
2026-11-07 08:12:01 +01:00
13f6e3f5d2 Use google_account_token for sign-in
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WzxdWY1k4muZfdVJfb7YGJ
2026-11-02 11:27:35 +01:00
6033e3815f Home: fit one screen, details behind tiles
Stakes and history move behind two tiles that open bottom sheets;
the page no longer scrolls at Pixel 6a geometry (home_fit_test).

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WzxdWWcbzoov1KFZJSFcSs
2026-11-01 22:06:43 +01:00
fd65c454e3 chore: bump deps to newest stable
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WzxdWXTUM9dpnCS2wtFkx4
2026-11-01 20:13:43 +01:00
05738db15a feat: JSON export/import on device
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WzxdWXTUM9dpnCS2wtFkx4
2026-11-01 19:55:04 +01:00
69be24d4c0 Session: tag catch-up + pre-deadline buzz
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-11-01 18:25:15 +01:00
1c2197a3cd deps: flutter 3.47.7 (workflows), gradle 9.8.2, coverage 7.17.0, ruff 0.17.2 (newest stable)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017k9j5dbjenpfZj7dh2faPN
2026-10-23 18:09:14 +02:00
2defe09e6a Launcher: label WakeAlarm, icon fitted to the shared 340px ink circle
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012kQ5PCKqBHAkAVkPSvp6dk
2026-10-23 14:36:09 +02:00
597901dc08 phone: sick-day interop both ways with keep; docs
check.sh now runs Dart->Python (keep, unsorted/duplicate input, a cleared
keep day, a keep-only edit that must not re-stamp state) and
Python->Dart (0.2.0 keep sets plus a 0.1.0 record with no keep, a stale
keep and a malformed one), Python 0.2.0 as the oracle with explicit
paths and HOME pointed at a scratch dir. DOCS-sick-days.md and the
SickDays.kt contract describe the alarm-only native cache.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PAhbNszs5JKdtnD4uBi3xp
2026-10-16 15:31:11 +02:00
ee7e380278 phone: per-locker sheet on the sick-day calendar
Tapping today or a later normal day opens a sheet with one switch per
locker (Workout lock, LeetCode, Book, Diet, Wake alarm), all off, and
"Mark sick"; a sick day reopens with its switches, "Save" (keep-only
edit, nothing written when unchanged) and "Clear sick day". Past days
stay inert. A sick day keeping any locker carries a dot in the calendar
and a "Kept: ..." line in the list. DayCell moves to day_cell.dart for
the 250-line cap.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PAhbNszs5JKdtnD4uBi3xp
2026-10-16 15:30:11 +02:00
7fe6108df3 phone: sick-day keep set (sickdays 0.2.0) in the Dart pool; alarm cache filtered
SickDayPool gains keep: mark/clear write state+actor+keep on one tick,
setKept re-stamps only keep+actor, read rules 1-5 in sick_day_keep.dart.
The native SickDays cache now holds only days on which the alarm stands
down, so a sick day that keeps the alarm rings like a normal day.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PAhbNszs5JKdtnD4uBi3xp
2026-10-16 15:30:11 +02:00
2e594f041d phone: pin the clock in the stageLocalSettings enforcement test
The test staged a 2026-10-01 event through stageLocalSettings, which called
stageEnforcementData on the real clock; on 2026-10-16 the event aged past
boundEvents' 14-day window and the test went red for everyone.
stageLocalSettings now forwards an optional `now` to stageEnforcementData
(callers unchanged), and the test passes its fixed clock. phone_app suite:
402 passed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Xfj1sxKH9VBPqpcAni8Jry
2026-10-16 14:55:26 +02:00
4f149e67a4 phone: read the native sick-day cache back on the Sick days screen
A release build gives no way to see the SharedPreferences cache that
AlarmReceiver reads at ring time (no run-as, no log of its contents), so
"the Dart UI shows it" could not prove "the alarm will stay silent".

- SickDays.store/dates log the sorted set under WakeAlarmSickDays.
- AlarmChannel getSickDays returns the cached set (null if unreadable).
- The Sick days screen reads it on open (before the sync rewrites it) and
  after every sync: "Alarm cache: <dates>" / "none" / "unreadable".

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Xfj1sxKH9VBPqpcAni8Jry
2026-10-16 14:38:25 +02:00
34037e55c1 phone: sick days — picker, silent ring, stick stand-down, pre-alarm sync
- Sick days screen (own name, sick icon, cat-2 fill): tap today or a later
  day to mark it sick, tap again to clear; past days inert. Writes the
  sickdays wire format (state sick/cleared + actor, one HLC per write) to
  sickdays-sync/devices via crdt_sync syncLog.
- Mutual exclusion in both pickers: each syncs the other pool on open and
  refuses a date it holds, naming the pool; an unreadable pool refuses all.
- Native SickDays cache (fail closed via DayCache); AlarmReceiver skips the
  ring with reason "sick_day" (no markFired, no ring_started); StickCheck
  verdict sick_day, streak skips sick days; stakes card says so.
- Pre-alarm pool sync: allow-while-idle alarm at 21:00 and A-30 (request
  code 1006) runs a headless engine (poolSyncMain) that pulls both pools
  from RTDB and rewrites both native caches, so a PC mark arrives with the
  app closed. The ring decision stays a cache read.
- JVM unit tests (JUnit 4.13.2) for DayCache, RingDecision,
  PoolSyncSchedule, StickCheck; Dart tests; Dart->Python interop check.

pylint hook skipped: verify.py imports sickdays, not in the hook env until
the sickdays-v0.1.0 tag exists; run by hand in a sickdays venv (10.00/10).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Xfj1sxKH9VBPqpcAni8Jry
2026-10-16 13:22:40 +02:00
ffe7666f49 deps: Flutter 3.47.7, shared_preferences 2.5.6
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018RHwVPyiiQMPh4M3JQwQku
2026-10-09 19:27:56 +02:00
57a6c6ba0e dependency freshness: bump pylint 4.1.2, url_launcher 6.3.3, github_device_auth v0.3.2 (CI red fix)
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-10-04 17:39:28 +02:00
9325103f5d Flutter 3.47.6; dismiss re-arm test no longer passes only on Tue-Thu
The expectation resolved the next alarm without the earned ramp set, while
the app (correctly) re-arms with the minute the dismiss just earned. On a
non-workday that is one minute apart, so the test failed Fri-Mon on any
Flutter version. It now loads the policy exactly as the app does.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FBbkN4ZB4aUPJPzR2prRan
2026-10-02 14:28:55 +02:00
728e00cde7 Wake enforcement v2: PC fallback ring + RTC self-wake, stacking Tue-Thu phone lockdown, escape hatch until 2026-10-15, stakes in the app, daily incident audit
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-10-01 12:17:06 +02:00
19f23f35a6 Free days: a month calendar of the shared pool; tap a future date to book it, tap a booked one to release it
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-27 20:53:46 +02:00
7f7fd916f7 Earned ramp: a dismissed Mon/Fri/Sat/Sun morning earns the minute; workdays hold, misses and free days earn nothing; workdays from freedays 0.3.0
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-27 20:13:22 +02:00
702f37a7ba Bump ruff 0.16.8 -> 0.16.9 and Gradle 9.7.1 -> 9.8.0: newest stable dependency-freshness gate
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-25 13:27:37 +02:00
d8b07b2be4 Cook the frog: lower the wake baseline a minute a day from 08:00 to 07:00
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-25 13:15:44 +02:00
bfe8bab58a Add the Tue/Wed/Thu workday stick: signal focus-owner and write tomorrow's penalty
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-23 21:20:58 +02:00
c3c129b068 Make the morning session the carrot: earn 11:00 by finishing it in time
On 2026-09-19 the phone was powered off after the bathroom tag and nothing
on the PC noticed: every finished session was published and nothing read
it. Now wake-alarm distils the phone's morning into an HMAC-signed
~/.local/state/wake_alarm/morning_session.json (wake-alarm-session.timer,
07:00-11:00 every 2 min) that screen-locker and leetcode-guard read to
defer their locks -- provisionally until min(start + 90 min, 09:30) while
the session is live, until 11:00 once it completed inside that cutoff,
and not at all for a session finished later (the phone switched on at
noon completes the legs then; that must un-lock nothing).

The phone publishes the session at start as well as at finish
(current_session, cleared with the finish stamp), so the PC can tell
"started, not finished" from "never started". A required 1-5 energy tap
joins the weigh-in and ships as latest_energy with a PC reader mirroring
_weight (the shared half is _field_reader).

Also fixes the vmbox harness, red since the 2026-09-11 home-tidy: the
guest tree lives at ~/wake-alarm, and the INSTALL-OK grep matched vm's
echo of the command line. Takes Flutter 3.47.5.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tp54ewdXRLaimUuT8LgYyn
2026-09-19 16:40:51 +02:00
af293413c0 Keep the morning session reachable after leaving the app; drop the still-phone watch
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-17 09:01:22 +02:00
6bfb059309 Take gatelock 0.8.2, flutter_secure_storage 11.2.0, coverage 7.16.1 and ruff 0.16.8
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 22:36:33 +02:00
f31f195187 Keep the morning going after the ring: weigh-in, shower, dress, desk, or it rings again
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 22:35:40 +02:00
84d6cbdc17 Take gatelock 0.8.1, Flutter 3.47.4 and flutter_secure_storage 11.1.1
The freshness gate's three findings. gatelock 0.8.1 restructured
LockConfig, which this repo does not use (only log_integrity), so the
pin moves without a code change; pylint stays at 10 against it.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HHCbkJWXHjtTUaqvnHRAqA
2026-09-12 20:02:31 +02:00
07f2408f7e Re-point paths for the ~ reorganisation (home-tidy)
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AwtowrowDXx6ccmWVPkCaF
2026-09-11 18:46:34 +02:00
758b2976b8 Gate the free-day rules against the Python package, and fix two picker paths
showDatePicker asserts that initialDate satisfies selectableDayPredicate,
so with today already free the picker asserted in debug and, in release,
pre-selected a disabled day. The cursor now starts on the first non-free
day. And the reason dialog had no Cancel: a barrier tap marked the day
anyway. Both covered by widget tests; the first fails without the fix.

The cross-language check that proved the Dart port was a one-off probe,
deleted after it ran. It is now scripts/free_days_interop/check.sh, a
pre-commit hook on both inputs: Dart writes one pool through every policy
branch, Python replays the same edits into a scratch pool, and every
semantic reading has to agree. The Python package is the oracle -- Dart's
own count compared to what Python reads from Dart's file cannot see a
drift in what Dart writes. Proved by dropping the sticky consumed flag
from release: remaining dart=33 python=32.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019pBgaXSJKyY3DarYpDatDF
2026-09-11 13:53:56 +02:00
146c427d6d Add the free-day picker, and stop the phone ringing on a free day
The shared pool (~/utils/freedays) was CLI-only: the gates on the PC stood
down, but the phone kept ringing at 08:00 and there was nowhere to mark a
day from bed. This is the one picker for every app.

FreeDayPool is a Dart port of the Python rules with the identical record
shape (id = ISO date; state/consumed/actor/reason on one tick; un-mark is
a state flip, never a tombstone), verified by a real cross-language round
trip: a log written here reads back through freedays.is_free_day and
remaining() with the same verdicts and count, and the PC's live pool reads
here as the same 27 of 35 that `freedays status` prints.

The alarm fires from Kotlin with no engine running, so the pool is
mirrored into a native SharedPreferences set after every mark, release and
sync (channel method setFreeDays), and AlarmReceiver consults it before
starting AlarmService. Fail closed: a missing or corrupt cache rings. All
three branches exercised on the API 36 AVD by broadcasting the receiver.

Firebase only for this namespace -- no GitHub mirror; the alarm record
keeps its own because un-migrated installs depend on it.

Known gap, documented rather than papered over: a day marked on the PC
reaches the phone at the next app launch or picker open, not overnight.
The ring path takes no background dependency by design, so mark on the
phone, or open the app once after marking on the PC.

Verified on the phone (release build, 23181JEGR08034): picker shows the
budget, marks 2026-09-30 with a reason, pushes it to
freedays-sync/devices/<uuid>, releases it with a full refund back to 35.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019pBgaXSJKyY3DarYpDatDF
2026-09-11 13:45:42 +02:00
f709fb8028 Bring Flutter, flutter_secure_storage and ruff up to latest stable
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-11 12:37:33 +02:00
b5701fb359 Sync the NFC tags and home location so a reinstall gets them back
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-11 12:34:50 +02:00
2904d97b2c Bring the phone app's analyzer up to latest stable
very_good_analysis ^10.3.0 -> ^11.0.0, which the freshness gate flagged
against phone_app/pubspec.yaml the moment the pre-push CI mirror ran
--all-files. Pre-existing; it blocked the push rather than being caused
by it.

The major bump newly flags two async_return_with_no_await sites, both now
`return await` rather than suppressed. The rule is correct in both: a
future returned without awaiting escapes the enclosing async function, so
a failure surfaces at the caller with a stack trace that no longer names
the code that made the call. firebase_backend.dart:91 is the one that
matters -- a rejected credential fetch should fail inside the function
that describes itself as recovering from exactly that.

flutter analyze clean, 146 phone-app tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-07 15:30:00 +02:00
7523cb4225 Ring the away profile when the phone is not at home
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-31 11:45:11 +02:00
60f79774c9 Fix the race that stopped the phone publishing at all
The PC has been firing the morning routine at 07:00 instead of 08:00, off a
record the phone published on 2026-08-09 and never superseded. The phone was
connected to Firebase the whole time and the home screen still said "Connect a
sync backend in Settings first."

initState fired three futures concurrently and the last one read what the
first two were still writing:

    unawaited(_applyPolicy());     // sets _hour/_minute
    unawaited(_loadSyncState());   // sets _service
    unawaited(_syncToPc());        // reads BOTH

`publish` takes the service as a *parameter*, so it captured `_service` by
value while it was still null. Assigning it moments later could never reach
the call. The startup publish therefore passed null every time and bailed out
before touching the network -- and `_syncToPc` had exactly one call site, that
one, so nothing published, ever. `_applyPolicy` has the same shape: a publish
that did win the race would have sent the baseline instead of an active
override. The bootstrap is now strictly sequential.

Three things conspired to hide it for three weeks:

- `publish` returned "Connect a sync backend in Settings first." for BOTH a
  missing backend and a null service, so a broken store looked like a setup
  step the user had never done. The advice pointed at something that was not
  broken. The two causes now report differently, with a regression test.
- `_loadSyncState` swallowed its exception with `on Exception catch (_)`, so
  if the store genuinely failed to open there was no trace anywhere. It logs
  now.
- `_pickEarlier` never published at all, so "Wake me earlier tomorrow" -- the
  one schedule edit the app offers -- never reached the PC either. It does now.

home_screen.dart went over the 250-line cap, so its body moves to
`screens/home_body.dart` as a stateless widget taking plain values. The screen
keeps the state and the ordering rules, which is where the bugs lived.

Verified on the Pixel 6a: home screen now reads "Synced to PC ✓", and the PC's
`_sync --show --no-cache` returns `08:00 (remote)` stamped seconds earlier,
against `07:00` from 2026-08-09 before. The phone had also minted a new device
id, so both records exist in Firebase and the HLC ranking picks the newer --
which is exactly why that reader ranks by HLC and not by listing order.

104 Dart tests pass; analyzer clean.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PcdngP688mW6rJye3LRkSN
2026-08-30 16:28:20 +02:00
150a8cd7e5 Require two tags, the walk, and three answers to dismiss
The dismiss gate was an OR — NFC tag *or* QR *or* (25 steps + 2 min) —
followed by one challenge. QR never worked at ring time (no scanner
package, no CAMERA permission, `onQrScanned` had no caller), so the real
bar was "walk 25 steps, wait 2 minutes, answer one question" and the NFC
tag was already fully bypassable.

It is now a strict, latched sequence owned by `DismissGate`:

    tag 1 -> 25 steps + 2 min -> tag 2 -> 3 challenges

Every stage is required, in that order, and a stage once passed never
reverts. That last part also fixes a real bug: `_onNfcTag` and `_trySteps`
both assigned the room-proof result straight onto `_inRoom`, so a wrong
tag tapped after the gate had passed sent the user back to the start.

Two *distinct* tags are required; the same id in both slots is refused at
registration, since one tag satisfying both stages removes the walk that
is the whole point. The existing single-tag registration migrates into
slot 1 rather than being stranded.

Answers are cumulative, not consecutive: a wrong one costs the attempt and
nothing else.

The bypass is the only fallback and is deliberately narrow. After 5 minutes
of ringing, or immediately when NFC is off or fewer than two distinct tags
are registered, the two *tag* stages are skipped and the count rises to 6.
The walk is never skipped — jumping straight to the challenges would let
someone lie in bed for five minutes and answer questions without standing
up, which is weaker than the gate this replaces. Without some such hatch a
dead tag at 06:00 leaves an alarm that cannot be switched off at all.

Also deletes the `flash` challenge outright, along with the now-orphaned
`generateCode`/`kDismissCodeLength`/`_dismissChars`. It was the only kind
that removed its prompt from the screen, which is unsolvable for someone
who has just walked in from another room.

The ring screen takes an injectable clock: `Stopwatch` reads the wall
clock, which `tester.pump()` does not advance, so without that seam no
widget test could reach the two-minute floor or the five-minute bypass.

Verified on the Pixel 6a: the existing tag migrated into slot 1 and both
rows render. 102 Dart tests pass; analyzer clean.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PcdngP688mW6rJye3LRkSN
2026-08-30 15:09:00 +02:00
5117f5a0c9 Latest stable dependencies, and the gate that keeps them there
**The two shared Python libraries were four minor versions behind and the gate
could not see it.** They are PEP 508 direct references
(`gatelock @ git+...@gatelock-v0.4.1`), which parsed as unpinned PyPI packages
that do not exist on PyPI -- an undeterminable answer, not a stale one.
utils@7a682a4 teaches the gate that shape; this repo is the first to benefit:
gatelock v0.4.1 -> v0.7.2, crdt-sync v0.5.1 -> v0.9.1. The system install was
already on 0.7.1/0.9.0, so only the manifests were adrift.

Python side: bandit 1.7 -> 1.9.4, mypy 1 -> 2.3.1, pylint 3 -> 4.0.7,
pytest 8 -> 9.1.1, pytest-cov 4 -> 7.1.0, pytest-randomly 3 -> 4.1.0,
ruff 0.8 -> 0.16.5, python-kasa 0.7 -> 0.10.2, and the rest exact-pinned.
Five majors and not one lint or type finding: `scripts/ci_mirror.sh` builds a
clean venv from requirements.txt and runs the whole CI set, 465 tests at 100%
branch coverage.

phone_app takes crdt_sync_dart v0.11.1 and the flutter_secure_storage 11 chain
(compileSdk 37 / AGP 9.3.2 / Gradle 9.7.1). 79 tests green, analyzer clean,
release APK builds.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CRykabMwPgDzFHcARoqpaK
2026-08-28 14:56:16 +02:00
1c8a40d410 Split _sinks.py and home_screen.dart under the 250-line gate
_sinks.py 341 -> 249: the pactl *parsing* moves to _sink_parsing.py. That half
is pure text handling with no subprocess and no side effects, so it is
testable without touching the host's audio; _sinks.py keeps command execution
and the activate/restore orchestration.

The parsers are re-imported into _sinks under their original names, guarded by
__all__. The suite patches `wake_alarm._sinks._parse_cards` and friends, and a
moved symbol makes patch() stop intercepting *silently* — the test then
asserts nothing and still passes. Ruff proved the point mid-split: shortening
__all__ let it delete two of the re-exports and broke an import outright.
_parse_profile_line is only imported (never patched), so its test now imports
from the new module instead.

home_screen.dart 270 -> 249. This one was mine — the automatic-alarm change
pushed it over and I claimed in TODO-push-and-gate.md that nothing this
session touched was over the limit; that was wrong. Publishing moves to
AlarmPublisher, which returns a status line instead of calling setState, so it
is testable without a widget. _loadSavedTime became _loadSyncState: it no
longer loads a *time*, because the schedule is baked in and owned by
AlarmPolicy, so a stored value could only disagree with it.

465 Python at 100% coverage, 79 Dart, APK builds, and both documented CLIs
still print real output.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 23:30:02 +02:00
95be57b575 Make the alarm fully automatic and un-weakenable (08:00, earlier-only)
The alarm was opt-in: it only armed when the user tapped "Set Alarm & Sync to
PC", so it could be defeated by simply not tapping it, or by cancelling it the
night before. It is now automatic and there is no lever a half-asleep user can
pull that buys them sleep.

- AlarmPolicy bakes the schedule in at 08:00 daily. The only edit the UI
  offers is moving the NEXT alarm earlier; setEarlier() refuses anything at or
  after the baseline, and refuses to relax an active override. Earlier-only is
  the whole design: every available change costs the user sleep.
- The ratchet resets after each fire, so an early start is one-shot. A one-way
  ratchet with no reset is a trap, not a defence — the first 05:00 would
  otherwise be permanent with no way back short of reinstalling.
- load() re-validates the stored value and ignores anything at or after the
  baseline. SharedPreferences is not a trust boundary.
- No cancel. The Kotlin handler is deleted, not just the Dart wrapper, so the
  capability genuinely does not exist rather than being one invokeMethod away.
- Arming happens in main()'s init seam and via AlarmWatchdog, a daily
  self-rescheduling check that re-arms if nothing is scheduled — closing the
  gap where a force-stop or a failed re-arm goes unnoticed until the morning
  it does not ring. BootReceiver re-establishes it, since setInexactRepeating
  does not survive a reboot.
- Publishing to the PC is automatic too; it used to need the same tap.

PC half, so the two agree with the network down:
- FALLBACK_ALARM_HOUR 13 -> 8, matching the phone's baseline. At 13:00 a
  Firebase outage silently moved the alarm five hours LATER — the one
  direction the design refuses to allow.
- _trigger._clamp_to_baseline() refuses any resolved alarm later than the
  baseline, whatever its source. Without it the PC would faithfully reproduce
  a stale or tampered remote value and the forbidden direction stays reachable
  through sync.

Tests: 465 Python at 100% branch coverage, 79 Dart. Both ratchet directions
are covered, per the standing rule about monotonic counters. test_sync.py and
test_trigger.py were split (payload, token, CLI, clamp) to stay under the
250-line gate; the split lost indirect coverage of extract_hlc/newest_alarm,
which now have direct tests instead.

`python -m wake_alarm._sync --show` and `_trigger --dry-run` both re-run and
still print real output.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 23:03:00 +02:00
a845cac883 Remove the flash challenge; revert the MethodChannel split that regressed NFC
Two things, one of them an undo of my own change.

The flash challenge is gone from the phone's rotation. It hides the code after
kDismissFlashSeconds, and on the phone that is unsolvable in practice: the ring
screen is reached by walking to another room and tapping a tag, so the code has
already vanished by the time the user is looking at it. Reported from the phone
as "invisible - impossible to actually do". Every challenge the alarm shows now
stays on screen until it is answered, and a test asserts next() never returns
one that hides itself. The constructor stays for the PC half's parity tests;
nothing routes to it.

The two-channel split is reverted. I introduced it to fix a tag reaching Kotlin
but not Dart, and pointed the push channel at the call channel — which broke
NFC *and* the step counter, both of which had been confirmed working on the
phone one build earlier. The user caught it: "they worked previously so your
changes regressed it". The split was a theory that was never shown to be
necessary; the single channel from 08b44b4 is restored unchanged.

Lesson worth keeping: when hardware has just confirmed a path works, the next
change should be the smallest possible delta on top of that exact state, not a
restructure bundled with a fix.

flutter test: 70 pass. dart analyze lib/ test/: clean.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 22:45:35 +02:00
08b44b4ca2 Fix three defects the hardware run exposed in the room-proof path
All three were invisible to the emulator (no NFC, no step sensor) and to the
widget tests (which mock the platform channel), and each one on its own made
the alarm harder or impossible to dismiss.

1. AlarmActivity.onPause called NfcReader.stop(). enableReaderMode is
   per-activity but disableReaderMode is effectively global, so anything
   briefly coming in front of the ring screen killed NFC polling for the rest
   of the ring. Measured on the Pixel: `flags: 271` immediately followed by
   `flags: 0`. Teardown moves to onDestroy, and onResume re-asserts.

2. StepCounter could never report progress from a standing start.
   TYPE_STEP_COUNTER is an on-change *cumulative* sensor: it emits nothing
   until the user has already walked, and the first emission was being
   consumed as the baseline — so the ring screen sat at 0/25 no matter how far
   you walked. TYPE_STEP_DETECTOR is now primary (one event per stride, no
   baseline to miss); the counter stays as the fallback for devices without a
   detector.

3. NFC cannot be read while the ring shows over the lockscreen. Measured
   A/B on the Pixel with the same tag: registration while unlocked →
   mScreenState=ON_UNLOCKED, tag discovered; ring screen over the keyguard →
   mScreenState=ON_LOCKED, reader mode accepted (flags 271) but no discovery.
   Android gates NFC polling on lock state. showWhenLocked draws *over* the
   keyguard without unlocking, so the alarm now calls
   requestDismissKeyguard() on fire and re-asserts reader mode from the
   dismiss callback. User's call, taken explicitly: on a secured phone this
   surfaces the PIN prompt, which is an acceptable trade for an alarm you
   must get up for.

Without (3) the only satisfiable proof at 06:00 is steps, which is exactly
the single-mechanism dependency the room-proof redundancy exists to prevent.

flutter test: 69 pass. dart analyze lib/ test/: clean.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 22:23:23 +02:00
10a594305c Bring the two files this change grew back under the 250-line gate
The gate runs --all-files at push time, so it caught two files this session
pushed over: AGENTS.md (244 -> 330) and home_screen.dart (238 -> 269). Both
are mine; the pre-existing Python overages are untouched, per the standing
rule not to split a file this change does not already have to edit.

- The phone section moves to DOCS-phone-alarm.md, following the repo's own
  pattern (DOCS-architecture.md, DOCS-vmbox.md), leaving a pointer.
- Alarm arming moves out of home_screen.dart into AlarmArming — "what it
  takes to arm an alarm" was never the picker's concern, and the extraction
  also repaired a doc comment my earlier edit had orphaned onto the wrong
  method.

flutter test: 69 pass. dart analyze lib/ test/: clean. APK builds.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 21:43:16 +02:00
dc2d54668b Phase 4/3: close the ring screen on dismiss, and make NFC registerable
Two defects that each defeated the alarm in a different way:

1. Nothing finished AlarmActivity after a correct answer. Sound stopped and
   the alarm re-armed, but the dismiss screen stayed up over the lockscreen
   forever — back is disabled in both halves by design — and the live instance
   kept _inRoom=true with the solved answer still typed, so the next ring
   could reuse an already-passed gate. AlarmActivity now handles
   closeRingScreen; the shared AlarmChannel deliberately holds
   applicationContext and cannot finish an activity, hence the split handler.

2. Nothing could write roomProof.nfcTagId — there was no registration UI — so
   checkNfc always answered "no tag registered" and checkQr was never called.
   Functionally that left steps as the only satisfiable proof, which is
   exactly the state the decision rules out ("NFC or QR qualifies; steps alone
   does NOT") and removes the redundancy whose purpose is that one failing
   mechanism at 06:00 must never block dismissal. Added RoomProofScreen
   (reader mode from MainActivity, since registration needs an Activity) and
   linked it from Settings.

Also: StepCounter kept one listener instead of registering a fresh one per
ring and never unregistering, and the ring screen drops its NFC handler on
dispose.

New tests cover what had been dead code: a registered tag satisfies the proof
and reveals the challenge, a foreign tag does not, and a correct answer
invokes BOTH dismiss and closeRingScreen. The emulator cannot reach the
dismiss path at all (no step sensor, no NFC), which is why it is covered by
widget tests rather than claimed from the AVD run.

Documented ringNow/main_ring_probe.dart in AGENTS.md as a deliberate manual
test handle, the phone-side equivalent of wake-alarm.service.

flutter test: 69 pass. dart analyze lib/ test/: clean. pre-commit: passed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 21:39:01 +02:00
cb50e58a56 Phase 4/3: NFC room proof, runtime permissions, non-blocking step reads
Three defects found by reviewing the emulator run, all on the dismiss path:

1. ACTIVITY_RECOGNITION was declared but never requested at runtime (API 29+).
   Denied, TYPE_STEP_COUNTER delivers nothing, so checkSteps() reports "sensor
   unavailable" forever. With steps the only wired proof, that made the alarm
   *undismissable*, not merely degraded. MainActivity now requests it
   alongside POST_NOTIFICATIONS.

2. checkNfc/checkQr were dead code — no reader, no channel method, no
   registration — so steps was the only reachable proof. That inverts the
   decision ("NFC or QR qualifies; steps alone does NOT") and removes exactly
   the redundancy whose purpose is that one failing mechanism at 06:00 must
   never block dismissal. Added NfcReader (hand-written on the existing
   channel, NfcAdapter.enableReaderMode so the tap does not restart the
   activity and reset the challenge) plus RoomProofSettings for the registered
   tag/QR. mobile_scanner stays the droppable extra.

3. StepCounter blocked the platform main thread up to 3s per read, called from
   the MethodChannel handler — jank-to-ANR on the ring screen. Now registers a
   listener and serves the last sample; reads never block.

Also: raiseAlarmVolume() is guarded. setStreamVolume can throw under a zen
rule, and by then startForeground() has already posted — a failed volume bump
must not take the ring down with it.

Verified on the API 36 AVD with the real app widget and route (the earlier
probe rendered its own UI, so it proved the Kotlin path but nothing about
/ring): screen Asleep -> Awake, AlarmActivity top resumed over the lockscreen,
RingScreen showing the room-proof gate. New widget tests cover the /ring route
resolving and the challenge staying gated behind the room proof.

flutter test: 57 pass. dart analyze: clean. Zero Python files touched.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 21:28:24 +02:00
af2a9b416b Phase 4/3: the phone's Dart half — ring screen, challenges, room proofs
Dart side of the alarm (items 1, 4, 3). main.dart had no init seam at all;
it now has ensureInitialized(), a navigatorKey and a /ring route.

- challenge.dart: all five kinds ported from _challenges.py so both halves
  behave identically. kDismissCodeLength/kDismissFlashSeconds are duplicated
  from _constants.py, not synced — noted in the file, since Dart cannot read
  Python.
- alarm_scheduler.dart: the Dart side of the single MethodChannel.
- ring_screen.dart: room proof first, then a challenge, then dismiss. Back is
  a no-op (PopScope) so the alarm cannot be escaped.
- room_proof.dart: NFC or QR satisfies on its own; steps NEVER do alone —
  the phone is stationary while you walk to it, so steps only accrue after
  pickup and can be faked by shaking it in bed. Steps count only together
  with elapsed time, as the fallback when both others fail.
- home_screen: arms the app's own alarm instead of asking the user to confirm
  in the Clock app (item 1's "no user interaction"). SET_ALARM stays as
  redundancy with SKIP_UI=true.

Re-arm has three hooks and needs all three: startup, BOOT_COMPLETED, and
dismiss. The plan assumed LogStore.changes would cover this; that stream does
not exist in crdt_sync_dart (verified — the library has no Stream at all), so
the picker's save is only one hook and a restart would otherwise arm nothing.

Fixed a real gap the API 36 AVD exposed: POST_NOTIFICATIONS was never
requested at runtime, so a fresh install sat at importance=NONE and the
full-screen intent was rejected outright — the alarm would ring with no
dismiss screen. MainActivity now requests it, and canUseFullScreenIntent()
also checks areNotificationsEnabled() so the check gates rather than warns.

Verified on the API 36 AVD: screen Asleep -> Awake, AlarmActivity became the
top resumed activity over the lockscreen, service isForeground=true
types=0x2 (mediaPlayback) category=alarm, no crash. flutter test: 54 pass.
dart analyze: clean.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 21:22:07 +02:00
e0cc20664c Phase 4: the phone's Kotlin ring path (item 4)
Hand-written Kotlin behind ONE MethodChannel, never a plugin: AGP 9.0.1 /
Gradle 9.1 / Kotlin 2.3.20 is bleeding edge and a plugin that will not compile
breaks the build outright rather than degrading.

- AlarmScheduler: setAlarmClock() (Doze-exempt, no SCHEDULE_EXACT_ALARM
  prompt), persists the time so boot and dismiss can re-arm.
- AlarmReceiver: hands straight to the service — a receiver gets ~10s and no
  wakelock, so nothing that must outlive it belongs there.
- AlarmService: foreground service, PARTIAL_WAKE_LOCK, USAGE_ALARM
  (= STREAM_ALARM, which is what ignores DND and silent mode), looping
  ringtone, vibration, full-screen intent over the lockscreen.
- AlarmActivity: showWhenLocked + turnScreenOn, routes to the Dart ring
  screen, back is a no-op so the alarm cannot be escaped.
- BootReceiver: re-arms on BOOT_COMPLETED; AlarmManager drops alarms across a
  reboot.
- StepCounter: TYPE_STEP_COUNTER baseline for item 3's fallback proof, also
  hand-written to keep plugin risk off this path.

Details that are a hard crash or a silent no-ring rather than a warning:
FLAG_IMMUTABLE on every PendingIntent (mandatory since API 31);
foregroundServiceType declared on the <service> AND passed to
startForeground(); explicit android:exported on every component;
USE_FULL_SCREEN_INTENT declared (missing from the plan's list) and
canUseFullScreenIntent() exposed so the Dart side can gate rather than warn.

Every permission verified present in android-36's api-versions.xml rather
than from recall. Compiles green with `flutter build apk --debug`.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 21:12:51 +02:00
738027d527 Phase 4 prereq: pin explicit min/target/compile SDK on the phone app
All three were delegated to the Flutter Gradle plugin, so there was no floor
to reason about — and full-screen-intent and foreground-service semantics both
changed at API 34/35, which is exactly the code the alarm path needs.

- compileSdk/targetSdk = 36, matching the target phone.
- minSdk = 28: the floor for android:showWhenLocked / android:turnScreenOn as
  plain manifest attributes (no legacy branch), which also clears
  flutter_secure_storage's API 23 and Flutter's own default of 24.

Verified with `flutter build apk --debug` (exit 0).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EckWmyfetYq3Mi8MAnUudh
2026-08-27 21:09:56 +02:00