A refused state change is now remembered as owed and forced by every pass
until it lands, instead of recording the opposite state, which could hide
the next real boundary. Any refusal, in the rule rewrite or the state
change, pulls the slot in to a retry, and the slot is armed in a finally
so a throwing step can't leave it unarmed. The wake-up that ends the mode
is armed allow-while-idle (one such fire a night), so Do Not Disturb no
longer lingers into the morning behind a doze deferral.
The manifest filter for the access-changed broadcast is gone: the
platform only delivers it to runtime-registered receivers, and the
screen's resume check covers it. A backup written before bedtime existed
now keeps the current bedtime settings, and a restore goes through the
engine's own field comparison, so a reminder-only difference leaves a
hand-switched mode alone.
Refs #1
A removal the system refuses while turning bedtime off now keeps the slot
armed fifteen minutes out as its retry, instead of cancelling it and
leaving Do Not Disturb on until some unrelated launch. A refused state
change records the opposite state so the next pass retries it, and a
rule that was rewritten (which clears its condition on Android 10-14) is
switched regardless. A refused rewrite is remembered and retried by every
pass until it lands, rather than resting on the one edit that asked.
Refs #1
Without exact-alarm access the slot falls back to a ten-minute window
instead of plain set(), whose slack could end an eight-hour night hours
late. A freshly opened bedtime screen reads whether bedtime is on from the
store, not from its still-loading state, so access granted while the
screen was closed creates the rule. A restore that rolled back no longer
forces the mode.
Refs #1
Android rations allow-while-idle wake-ups per app in doze, and the timer's
expiry slot and the alarm's fallback share that ration, so a 07:00
wake-up fire could hold back a timer due at 07:04. The bedtime slot is
now a plain exact alarm that doze may defer, which can move a mode
boundary but never a ring. A restore of unchanged bedtime settings no
longer forces the mode, and the screen skips its rule check while bedtime
is off.
Refs #1
Only a bedtime before 06:00 now counts as the small hours after the
chosen evening, so a day sleeper's 08:00 bedtime runs on the days picked.
A rule whose removal was refused while turning bedtime off in the daytime
is removed by the next pass, and a restored reminder length the picker
doesn't offer stays choosable.
Refs #1
A bedtime before noon now belongs to the evening before, so Sunday to
Thursday nights at 00:30 run early Monday to early Friday, not Saturday
night through Wednesday night. Turning bedtime off clears the "last
applied" state only once the rule is really gone, so a refused removal
is retried instead of leaving Do Not Disturb on. Reopening the screen
after granting access elsewhere creates the rule even if the system's
broadcast never arrived. The empty-mask rule lives in one place.
Refs #1
Whether the wake time is the same night or the next morning is now
decided on the local times, so a bedtime in a spring-forward gap no
longer turns a 45-minute window into a day-long one. Screen edits start
undispatched so they reach the engine's lock in the order they were made,
and an edit that rewrites the rule also re-asserts the mode, in case the
platform reset its state.
The launch-time bedtime pass runs even if an earlier engine pass failed,
a restore that never applied settings no longer forces the mode, and a
user who never turned bedtime on pays nothing for it at launch or boot.
The screen says when equal bedtime and wake-up times mean it never starts.
Refs #1
A rule recreated by a pass (deleted in Settings, access re-granted) now
gets its state set even when no boundary was crossed, and the applied
state is recorded only once the system accepted it. Edits that can't move
a boundary (reminder, effects) no longer re-assert the mode over a manual
override, and the screen's writes run on the application scope so leaving
right after a change can't cut it short. A late reminder counts the
minutes actually left, and one slot fire reads the zone once.
An equal bedtime and wake time is now no window instead of a day-long one,
which on every night meant Do Not Disturb never ended, and an empty days
mask in prefs reads as every night, as a backup's does. The rule-id extras
are translated to the bedtime action in MainActivity, keeping the routing
table keyed on the action alone.
Refs #1
Settings opens a rule's configuration activity with no action, only the
component and the rule id, so MainActivity now routes on the rule-id
extras. Each engine pass reads the clock once, so the mode's state and the
next slot can't disagree around a boundary, and non-edit passes switch the
mode only when the schedule crossed a boundary since Clockula last set it
(persisted), which also catches up a wake-up missed during a force stop
while leaving a mode switched on by hand alone. A reminder fire that also
ends a window now turns the mode off, and a reminder the schedule moved
away from is withdrawn.
The rule is only rewritten when a field it carries changes, and a rewrite
keeps the existing policy if it still allows alarms, so Android 10-14 no
longer loses the user's Do Not Disturb edits. The bedtime receiver, screen
and fallback alarm call are guarded, the boot-path resync runs even if an
earlier engine failed, and a backup with no nights restores every night.
Refs #1
Picking "Timer expired" as the default alarm or timer sound showed "Custom" in Settings, because the summary knew only the device default and silence. It now names the bundled sound, as the editors already do.
ARCHITECTURE.md now names the exact bundled sentinel instead of a prefix the code does not recognise, lists clockula://media values in the alarm column, gives the timer column the same values, and says RingtoneCatalog lists the bundled sounds too.
Refs #14
Clockula can now turn on a Do Not Disturb mode of its own at bedtime and
off at the wake-up time, on the nights the user picks, with an optional
reminder before it starts. Digital Wellbeing's Bedtime mode can't be
driven by a third-party app, so this is a separate rule: a typed mode
with grayscale, dimmed wallpaper, dark theme and hidden always-on display
on Android 15+, and a plain DND schedule rule on Android 10-14.
The rule's policy always allows alarms (and media). BedtimeEngine owns one
AlarmManager slot of its own, on its own receiver and request code, and
every bedtime call on the boot, time-change and launch paths runs after
the alarm and timer passes, guarded. The rule id is never stored: the
platform lists an app's own rules, so restores and reinstalls find or
recreate it. Turning bedtime off removes the rule. The screen walks the
user through granting DND access and explains the Wellbeing overlap.
Settings are part of the JSON backup; older files restore bedtime as off.
Refs #1
ARCHITECTURE.md still said every non-silent ringtone_uri is a content URI and listed the picker's rows without the bundled "Timer expired" sound, so a reader could Uri.parse the clockula://tone/ sentinel directly. It now names the sentinel, BundledTones as its only resolver, and where the tone sits in the picker.
A view-model test now covers a device whose only listed tone is the bundled one, which must still show the "no alarm sounds" banner. The existing test only checked an empty list, so it passed even if the bundled tone were counted as a device sound.
Refs #14
A new setting in Settings > Alarms: while the ring screen shows, the
volume buttons change the volume (the default, and what they did
before), snooze, dismiss, or do nothing. It is app-wide, stored next to
the clock defaults, and included in the backup; a backup without it
restores "change the volume".
A press is made without looking, so VolumeButtonPolicy keeps it from
ending an alarm more finally than the ring screen would. Dismiss with a
dismiss challenge set snoozes instead, so the challenge still stands.
With no snoozes left, Snooze (and Dismiss behind a challenge) falls
back to changing the volume, never to the engine's refused-snooze path,
which dismisses. The key's snooze also asks the engine to keep ringing
if the limit turns out to be reached after all (AlarmEngine.snooze's new
dismissIfRefused). The ring screen decides whether a press is its own
on key down and keeps that for the release, leaves the keys alone once
another alarm has taken over the ring, and ignores a second snooze or
dismiss while the first is still running.
Only the ring screen handles the keys. While the ring is a heads-up over
another app, a background app gets no key events, and the one way
around that does not hold up: a remote-playback MediaSession whose
VolumeProvider takes the keys. On Android 13 and later
MediaSessionRecord.canHandleVolumeKey refuses such a session from an app
without a MediaRouter2 routing session, which Clockula's would be, so on
current Pixels the keys never reach it. Where it does work, it loses to
whatever app is in front or played last, shows a cast-volume slider and
a phantom media player, and takes media buttons while the alarm rings.
AOSP DeskClock and BlackyHawky's Clock also handle the keys in their
alarm activity only. ARCHITECTURE.md records the limitation.
Fixes#18
The notice is local-only, so its Dismiss is never bridged to a watch where the phone's unlock does not apply. The closing and skip-spending checks land a second past the fire grace, since onFire still accepts a fire at exactly the grace.
Fixes#16
The upcoming check that spends a skip dismissed in advance now wakes once the skipped occurrence's fire grace is over rather than at the occurrence itself, so it can never get ahead of a fire still in flight for it.
Fixes#16
Keeping a skip armed through its fire grace let the editor's switch or another dismissal inside those two minutes ring the skipped alarm late, or be absorbed. The resolver goes back to spending the skip at its occurrence and recording it as handled, and the upcoming check wakes at an armed skip's occurrence so the list stops saying "skipping next" on time.
Fixes#16
The resolver now keeps a skip armed until its occurrence is out of the fire grace too, so a reschedule in those two minutes cannot spend it early and let a fire already in flight ring the dismissed alarm; onFire refuses that fire by the armed watermark. The notifier also cancels notices it posted itself, which the platform may not list as active yet.
Fixes#16
The upcoming check now also wakes once a noticed occurrence's fire grace has run out, so a notice cannot outlive a late fire, and once an armed skip's has, so the list stops saying "skipping next" about an occurrence already gone. onFire is back to checking only the armed skip: the resolver deliberately re-arms a handled occurrence that a clock moved backwards put back in the future, and that one must ring.
Fixes#16
The Dismiss action's unlock requirement is honoured only from Android 12, so on 10 and 11 the notice is now secret on the secure lock screen rather than dismissible by whoever holds the phone.
Fixes#16
A natural fire whose instant equals the skip or the handled-occurrence watermark is ignored, so a dismissal racing a reschedule cannot ring. Dismissing a repeating alarm persists its resolution before arming the skip, so an unspent old skip cannot absorb it. The deferred notice sync runs even if starting the ring throws, the swipe record is one synchronised read-modify-write, and it stays out of backups.
Fixes#16
A skip spent on its occurrence now becomes the handled-occurrence watermark, so a second resolution inside the fire grace cannot ring the skipped alarm after all. The upcoming notice records a swipe through its delete intent rather than inferring one, so a reboot or a blocked channel no longer hides it for good; its Dismiss asks for an unlock, and a Dismiss for a switched-off alarm clears its notice.
Fixes#16
A fire broadcast already in flight when a skip was armed on its occurrence is now ignored, and a Dismiss that lands once the occurrence is due is stale. The notices are synced after the ring starts on the fire path, and a swiped-away notice stays gone across a process restart. A switched-off alarm answers NotFound.
Fixes#16
Two hours before a natural occurrence (AOSP DeskClock's lead), a silent
"Upcoming alarm" notification offers Dismiss, and the alarm's row shows
the same Dismiss as a chip. Dismissing is dismissUpcoming's own branches:
a repeating alarm skips that occurrence, a one-shot is switched off.
The notices are synced at the end of every reschedule, after the real
registration, and a third AlarmManager slot on its own request code
wakes when the next window opens. The sync is wrapped so a notification
failure can never stop a ring. Dismiss names the occurrence and is
refused once the alarm is ringing or has moved, so it can't get round
the dismiss challenge. No challenge is asked before the ring, as in
Google Clock.
Fixes#16
A stored value of 60 minutes or more now reads as Never, which rings that
long anyway. The picker no longer shows both "60 minutes" and "Never" for
the same window. Also drops a redundant null check in TimerRingPolicy.
Refs #17
Fifth round of review fixes for the silence-after setting.
A ring whose start the clock has moved back behind is now re-anchored to
now, once, and stored. Before, the negative age lingered, and every later
re-measure gave the ring a fresh window. With that fixed, onAutoSilence
can safely re-arm again when the ring is still inside the window in
force, so a backstop queued behind a raised setting no longer ends the
ring early.
The timers' setting is now written through
TimerEngine.setTimerSilenceAfter, which records a session that has
already outrun the old window as lapsed, so raising the setting cannot
restart a timer that went quiet while no service pass was running. The
lapse rule and the lapsed anchor live in TimerRingPolicy, shared by the
policy and the engine.
Refs #17
Fourth round of review fixes for the silence-after setting.
onAutoSilence no longer re-arms a ring it measures as inside the window.
After a clock moved back, that check could push the backstop out again
and stretch the ring well past its window. Every writer of the window and
every clock change already re-arms the backstop under the engine lock, so
a backstop that fires has had its window, and it closes the ring as in v1.
A one-shot alarm the resolver switches off as missed also clears a
pending skip flag, as the resolver's other one-shot disable does.
SNOOZE_GRACE is its own ten-minute constant again, not tied to the
setting's default.
Refs #17
Third round of review fixes for the silence-after setting.
A backstop that fires while the clock is still behind the ring's start
now closes the ring instead of re-arming in a loop until the wall clock
catches up. A clock change re-measures a ring in progress before it
reschedules: a clock moved back re-arms the backstop at most one window
from the moved clock, and one moved forward past the window closes the
ring.
The engine now owns the alarm setting's write. setAlarmSilenceAfter
stores it and re-measures the ring under one lock and without
cancellation, so a write can no longer land without the re-measure.
A snooze the resolver drops takes its "Snoozed until" notification with
it. SNOOZE_GRACE is derived from the default window, and the silence-after
picker moves to ui/settings, since it is not a per-alarm override.
Refs #17
Second round of review fixes for the silence-after setting.
A fire now clears any stale ringingSince before it re-resolves the slot.
Before, the resolver's new one-shot close could switch off, or delete, the
very alarm that was about to ring.
A one-shot alarm displaced by another alarm's fire is switched off, as on
every other close of an undismissed cycle, instead of being armed for the
same time tomorrow.
onAutoSilence, onSilenceAfterChanged and the reboot resume now share one
rule. A ring still inside the window has its backstop re-armed, measured
from when it started, or from now if the clock has moved back behind that
start, so a clock change cannot stretch a ring past one window.
Refs #17
Review fixes for the silence-after setting.
Changing the alarm's silence-after now re-measures a ring in progress at
once, through AlarmEngine.onSilenceAfterChanged: past the new window it
closes as missed, inside it the backstop moves. onAutoSilence also re-arms
instead of closing a ring that is still inside the current window, so a
backstop armed for a shorter window cannot end a ring early.
A cycle the resolver closes as missed (a ring past its window, or a
snooze that never fired) now disables a one-shot alarm, as the engine's
own close does. A short window made that path much easier to reach, and
the alarm was being armed for the same time the next day.
A late snooze is kept for the window or ten minutes, whichever is longer,
so a short window never makes a promised snooze easier to lose than in v1.
A timer that has gone quiet stays quiet when the setting is raised: the
lapsed anchor moves back past the hard stop. The timer's heads-up is now
keyed to the session anchor rather than the window's end, so editing the
setting mid-ring no longer counts as a new session and alerts again.
A negative stored value now reads as the default instead of Never. The
picker offers a stored value that matches no choice, and the wake-lock
timeout is one shared constant.
Refs #17
A ringing alarm and a finished timer used to stop after a fixed ten
minutes. Settings > Alarms and Settings > Timers now each have a "Silence
after" row with Google Clock's choices: 1, 5, 10, 15, 20 or 25 minutes,
or Never. Ten minutes stays the default, so nothing changes on upgrade.
The value is app-wide; a per-alarm override would need a new column, a
migration and an editor row, and is left out.
"Never" rings for an hour, then stops. The backstop is also what keeps a
phone left at home from ringing its battery flat, what bounds the ring
services' wake locks, and what stops a reboot hours later from resuming a
ring nobody was there for. A truly unbounded ring would give all of that
up, so Never maps to a one-hour hard stop, longer than any finite choice,
and the settings hint says so. Stored values clamp to 0..60 on read, on
write and from a backup.
The window is read everywhere the fixed constant was: the resolver's
ringing and late-snooze checks (which now take it without a default), the
backstop onFire arms, how late a snooze fire may arrive, whether an
interrupted ring is resumed after a reboot, upcoming(), the timer ring
policy, the lapsed-session check and the timer ring screen. Both ring
services' wake-lock timeouts now derive from the hard stop instead of a
hard-coded fifteen minutes.
When the resolver closes a ring (the window was shortened mid-ring, the
clock jumped, the alarm was switched off), the engine now also stops the
ring service and cancels the backstop. Before, it only cleared
ringingSince and left the sound to a backstop armed for the old window.
Both values are backed up as default_alarm_silence_after_minutes and
default_timer_silence_after_minutes; a file without them imports as ten
minutes.
Fixes#17
The snackbar appended the codec's or the exception's own English text,
class names and all. Failures now carry a category (not a backup, from a
newer version, invalid data, too large, unreadable) that the screen maps
to a string resource; the detail stays in the result for tests.
A backup without the timer_presets field decoded it as an empty list,
which means "the user removed them all", so restoring it wiped the
built-in presets. A missing field now keeps them. BACKUP.md describes
this and the per-alarm clamps.
A timer whose ring was held back by a ringing alarm still popped a
heads-up over the alarm's screen, then alerted again once the alarm
ended. The heads-up now waits with the ring.
The widget worked out "Today" when it was drawn, and nothing redrew it
when the date changed, so a widget drawn on Monday night still named
Tuesday on Tuesday morning. It now shares the analog faces' midnight
redraw, renamed to MidnightRefresh, and redraws on a clock or zone change.
The hub read the self-check once, so after fixing something in the
self-check screen it kept showing the old number of problems. It now
re-reads on every resume, as the self-check screen does.
A sticky restart of the timer or stopwatch service has no user action
behind it, so Android 12+ can refuse its foreground start, and the crash
repeated with every restart. The same could happen to the alarm ring
after an inexact fire, after its state was written. Refused starts now
stop the service, or post the ring screen's notification, instead of
crashing. The ring service is stopped with stopService, which cannot
throw from the background.
Granting "Alarms & reminders" again left the next alarm on the inexact
fallback until something else re-registered it. The system's permission
broadcast now reschedules alarms and timers.
A MediaPlayer that errored after it started left the alarm or timer
silent, with no fallback and no forced vibration. The ring now moves on
to the next sound in the chain, and vibrates when none is left.
From Android 14 a foreground service's notification can be swiped away on
an unlocked phone, which left a sounding alarm with no controls. The ring
notification now reposts itself when it is deleted, and tapping a ringing
alarm in the list opens its ring screen.
The system reports only the earliest alarm clock of any app, so an older
alarm left in the stock clock showed up as a problem the offered fix
couldn't solve. Another app's alarm at or before Clockula's next fire is
now fine; a missing registration, or anything later than ours, still is
a problem.
Backup import clamped the app-wide defaults but passed an alarm's own
snooze length, snooze limit and volume ramp through unchanged, so a
snooze of 0 re-rang at once and a negative limit turned snoozing off.
Both import and the database read now clamp them to the ranges the
pickers offer, and imported timer lengths to the setup range.
The fire broadcast's wake lock ended with the receiver, and the ring
services only took their own after a database read, so a dozing device
could suspend in between and ring late. Both receivers now hold a short
timed wake lock across the hand-off, and the alarm service takes its own
before the read.
makiolaj